text
stringlengths 66
377k
|
|---|
High expression of COPB2 predicts adverse outcomes: A potential therapeutic target for glioma
## | me thods
## | public database and bioinformatics analysis
The transcript level of COPB2 in different cancers was ascertained by the Oncomine database (https ://www.oncom ine.org/resou rce/ main.html),with a threshold set as such-top gene rank 10%, fold change >2, and P-value <1E-4. The microarray data of patients with glioma were downloaded from the GEO (https ://www.ncbi.nlm.nih. gov/geo)public database under accession number GSE16011.Gene expression profile data containing clinical information from low-grade glioma and GBM projects (HTSeq-FPKM) were obtained from TCGA database (http://cance rgeno me.nih.gov/).The data from TCGA were further analyzed for associations between COPB2 expression and clinicopathologic characteristics in glioma.
F I G U R E 1 A, COPB2' expression level in cancers in Oncomine Database: the left box in red indicated the number of datasets with COPB2 hyperexpression and the right box in blue indicated the number of datasets with COPB2 hypo expression after comparing cancerous and normal tissues. B, C, TCGA cohort and GSE16011 dataset from GEO support the findings that indicate COPB2 upregulation in glioma
## | gene set enrichment analysis and gene set variation analysis
To investigate the potential mechanisms underlying the interaction of COPB2 expression on glioma progression, a GSEA 12 was conducted to screen out whether some biological pathways showed statistically significant differences between high and low COPB2 expression groups. For each analysis, gene set permutations were implemented 1000 times. Gene sets with a false discovery rate (FDR) <0.05 and normal P-value <.05 were viewed as significantly enriched. Moreover, GSVA 13 was performed to transform gene expression values into scores for inflammatory response metagene sets, followed by the application of correlograms to further verify correlations between COPB2 and these metagenes.
# | statistical analysis
The statistical analyses were performed utilizing R software v3.5.1.
Descriptive statistics were used to summarize the molecular and clinical characteristics of patients in the TCGA database. To analyze potential relationships between COPB2 and clinicopathologic features, Mann-Whitney U and logistic regression tests were used.
The Kaplan-Meier method and Cox regression analyses were used to compare the impact of COPB2 expression on the OS of TCGA patients alongside with other clinical variables. The remaining correlations between COPB2 expression and inflammatory and immune cell types were detected by using canonical correlation analysis in GraphPad Prism 7 and SPSS 25.0. In all statistical analyses conducted, a P-value <.05 was viewed as statistically significant.
## | re sults
## | glioma copb2 transcript levels in different databases
Firstly, the transcript levels of COPB2 in different cancers were analyzed. The Oncomine database (one of the main functions of which is gene expression differential analysis) was used to explore the expression of COPB2 mRNA in different cancers , and 189 datasets, including 33 144 samples, were included. Relative to normal clinical specimens, COPB2 indicated significant hyper-expression in bladder, brain and central nervous system, breast, esophageal, head and neck, lung, lymphoma, sarcoma, and other cancers, but hypo-expressed in leukemia , suggesting that the high expression of COPB2 is common in various types of cancer. The detailed expression profile was summarized in .
In glioma, 675 glioma patients with COPB2 expression profile data were obtained from TCGA COPB2 is significantly upregulated in tumor tissues relative to nontumor tissues , P < .001). In addition, we also used the GSE16011 dataset from GEO database for the purpose of validation , P < .001). The results indicated increased transcript levels of COPB2 in glioma.
## | tcga glioma patient characteristics
As the TCGA database contains sufficient glioma samples, we only selected this database for further analysis of the association between gene and clinical characteristics. A total of 1114 cases
## | association with copb2 expression and clinicopathologic features
To explore the expression pattern of COPB2 in gliomas, mRNA expression profiles from the TCGA database were obtained and analyzed. As shown in (A-H), increased expression of COPB2 correlated significantly with tumor grade (P < .001), histological type (P < .001), age (P < .001), KPS (P = .0260), tumor status (P < .001), and vital status (P < .001). Despite the lack of significant differences, a correlated trend was observed for IDH1 mutation (P = .114) and family history of cancer (P = .219).
Univariate analysis using logistic regression revealed that COPB2 expression (ground on median expression value) was linked to poor prognostic clinicopathologic variables (
## | survival outcomes and multivariate analysis
## | copb2-related signaling pathways based on gsea
As many signaling pathways contribute to tumor initiation and progression, the poor prognosis of COPB2-high may be related to the numerous signaling pathways activated in glioma. GSEA was utilized to recognize signaling pathways involved in glioma between low and high COPB2 expression cohorts. Significant differences (normalized P < .05, FDR < 0.05) were observed in the enrichment of the MSigDB Collection (kegg.v6.2.symbols.gmt).
Several signaling pathways-especially inflammation-and immunity-related pathways-were enriched in the COPB2 high expression phenotype, including B-cell receptor, T-cell receptor, natural killer (NK) cell-mediated cytotoxicity, antigen processing and presentation, Fc gamma R-mediated phagocytosis, cytokinecytokine receptor interaction, leukocyte transendothelial migration, and other pathways in cancer (please see and .
## | copb2-related inflammatory response
To better comprehend COPB2-related inflammatory activities, seven immune system-related metagene clusters (comprising 104 genes)that serve as surrogate markers of different immunological cell types were employed (Data S1). COPB2 expression, metagenes expression, age, gender, vital status, tumor grade, and histology of samples were displayed on a heat map in . .
## | relationship between copb2 and infiltrating immune cells
Previous studies reported that tumor-infiltrating immune cells may represent a crucial pathophysiological factor in the onset and progression of glioma.We examined the relationship between
## | d iscuss i on
In our study, the quantitative results indicated that COPB2 had higher expression levels in most cancers compared with normal tissues in the Oncomine database. In addition, COPB2's high expression in patients with glioma was further validated in the TCGA and GEO databases. RNA sequencing data from TCGA were also obtained and analyzed. COPB2 expression levels were correlated In the last decade, sequencing analysis has been applied comprehensively to explore the molecular mechanisms implicated in the process of disease progression.Recently, some reports found significant differences in COPB2's expression in various tumors.
Underexpression of COPB2 could downregulate the EMT-related protein N-cadherin and vimentin which may promote breast tumor cell invasion.Knockdown of COPB2 led to cell apoptosis by inhibiting the RTK signaling cascade molecules in gastric cancer.In lung adenocarcinoma, 7 patients with COPB2-high had worse survival status than COPB2-low. In the present study, we demonstrated that the overexpression of COPB2 in glioma was correlated with advanced clinicopathologic characteristics and predicts worse outcomes.
These findings indicated that COPB2 may be regarded as a promising target for cancer gene therapy. COPB2 expression was observed to be positively associated with the interferon gene sets . Thus, it makes sense to try to combine anti-PD-1 therapy and anti-COPB2 therapy to amplify the efficacy of treatments typically used in isolation.
Some limitations were present in this study: First, the number of patients incorporated in the univariate and multivariate Cox analyses patients were reduced as many patients had missing integrated data on all variables; second, only a small number of healthy samples were used as controls, so additional studies are needed to balance sample size; and lastly, laboratory studies should be carried out to elucidate the precise mechanisms of COPB2 overexpression in human glioma and clarify its relationship with poor prognosis and immunomodulation.
## | con clus ion
As far as we are aware, at present, this is the first study to explore the prognostic value of COPB2 in patients with glioma. This study revealed that COPB2 expression was upregulated in glioma samples and was related to adverse outcomes. We also found that COPB2 may represent an important factor in the immunomodulation of the glioma immune microenvironment. Therefore, taken together, COPB2 may act as a potential biomarker of prognosis and immunotherapeutic target for glioma.
## Co n fli c t o f i nte r e s t
The authors declare no conflict of interest.
TA B L E 5 Secreted and membrane immunosuppressive molecules expressed by glioma cells
## Cytokines type function references
TGFβ IL-10 PGE2 Gangliosides
## Soluble
Suppressing T-cell activation, proliferation and differentiation into effector cells.,31
## Cd70
Membrane proteins Inducing apoptosis of T and B cells from PBMCs.
## Fasl
Inducing apoptosis of FAS-expressing T cells.HLA-G Inhibiting proliferation, cytotoxicity by interaction with inhibitory receptors expressed on effector lymphocytes.
## Ido
Inhibiting T-cell proliferation.PD-L1 PD-L1 expressed by glioma inhibits IFN-γ production of antitumor T cells.
## O rci d
Yan Zhou https://orcid.org/0000-0002-7051-1573 |
Optimal Cut-Off Points of Fasting Plasma Glucose for Two-Step Strategy in Estimating Prevalence and Screening Undiagnosed Diabetes and Pre-Diabetes in Harbin, China
To identify optimal cut-off points of fasting plasma glucose (FPG) for two-step strategy in screening abnormal glucose metabolism and estimating prevalence in general Chinese population. A population-based cross-sectional study was conducted on 7913 people aged 20 to 74 years in Harbin. Diabetes and pre-diabetes were determined by fasting and 2 hour postload glucose from the oral glucose tolerance test in all participants. Screening potential of FPG, cost per case identified by two-step strategy, and optimal FPG cut-off points were described. The prevalence of diabetes was 12.7%, of which 65.2% was undiagnosed. Twelve percent or 9.0% of participants were diagnosed with pre-diabetes using 2003 ADA criteria or 1999 WHO criteria, respectively. The optimal FPG cut-off points for two-step strategy were 5.6 mmol/l for previously undiagnosed diabetes (area under the receiver-operating characteristic curve of FPG 0.93; sensitivity 82.0%; cost per case identified by two-step strategy ¥261), 5.3 mmol/l for both diabetes and pre-diabetes or pre-diabetes alone using 2003 ADA criteria (0.89 or 0.85; 72.4% or 62.9%; ¥110 or ¥258), 5.0 mmol/l for pre-diabetes using 1999 WHO criteria (0.78; 66.8%; ¥399), and 4.9 mmol/l for IGT alone (0.74; 62.2%; ¥502). Using the two-step strategy, the underestimates of prevalence reduced to nearly 38% for prediabetes or 18.7% for undiagnosed diabetes, respectively. Approximately a quarter of the general population in Harbin was in hyperglycemic condition. Using optimal FPG cut-off points for two-step strategy in Chinese population may be more effective and less costly for reducing the missed diagnosis of hyperglycemic condition.
# Introduction
Pre-diabetes (either impaired fasting glucose or impaired glucose tolerance is a relatively high-risk state for diabetes. [bib_ref] Relation of impaired fasting and postload glucose with incident type 2 diabetes..., De Vegt [/bib_ref] Pre-diabetes and diabetes carry risk of diabetes complications and cardiovascular disease, especially in people who remain with abnormal glucose metabolism despite intensive intervention. [bib_ref] Pre-diabetes and the risk for cardiovascular disease: a systematic review of the..., Ford [/bib_ref] In 2003, American Diabetes Association revised the lower cut-point of fasting plasma glucose (FPG), which redefined IFG from 6.1 mmol/l to 5.6 mmol/l (ADA criteria).However, World Health Organization and some other organizations adopted a cut-point of 6.1mmol/l for FPG as the upper limit of normoglycaemia as before (WHO criteria).In China, the prevalences of pre-diabetes using WHO criteria and diabetes reached 15.5% and 9.7% respectively, and 60.7% of the people with diabetes had been undiagnosed. [bib_ref] Prevalence of diabetes among men and women in China, Yang [/bib_ref] Accordingly, screening for pre-diabetes and undiagnosed diabetes in an early stage should be advocated in China.
Both FPG and 2 hour post-load plasma glucose (2-h PG) after a 75-g oral glucose tolerance test (OGTT) have been used as the gold standard to identify individuals with pre-diabetes and diabetes. Since the OGTT cost more demands on participants' time, FPG is the most widely used test for detecting diabetes and pre-diabetes in some epidemiological studies and screening. However, approximately half of the participants with undiagnosed diabetes and 70% of the participants with pre-diabetes using WHO criteria had isolated increased 2-h PG in Chinese people. [bib_ref] Prevalence of diabetes among men and women in China, Yang [/bib_ref] FPG alone (a one-step strategy) [bib_ref] Prevalence of diabetes and impaired fasting glucose in the Chinese adult population:..., Gu [/bib_ref] [bib_ref] A cross-sectional study on impaired fasting glycaemia and diabetes mellitus in residents..., Zhao [/bib_ref] [bib_ref] Prevalence of diabetes and impaired fasting glucose in Korea: Korean National Health..., Kim [/bib_ref] could lead to lower estimate of prevalence and missed diagnosis. A two-step strategy (including an OGTT was conducted only in subjects with IFG) [bib_ref] Prevalence and determinants of diabetes and impaired fasting glucose among urban community-dwelling..., Wang [/bib_ref] [bib_ref] The prevalence of impaired fasting glucose and type 2 diabetes in a..., Moadab [/bib_ref] [bib_ref] Increased prevalence of diabetes mellitus in a rural Greek population, Melidonis [/bib_ref] as recommended by WHO could reduce underestimate of diabetes relatively rather than pre-diabetes. It was unclear whether two-step strategy could be used to screen pre-diabetes and/ or diabetes in Chinese people with a large proportion of isolated increased 2-h PG.
Harbin is the capital of Heilongjiang Province, which is the most northern province of China. A population-based cross-sectional study was conducted to describe the prevalences of diabetes and pre-diabetes in here, which accurately defined by performing FPG and OGTT in all participants. Meanwhile, we identified screening potential of FPG, cost per case identified by two-step strategy, and the optimal FPG cut-points for two-step strategy in screening undiagnosed diabetes and/or pre-diabetes in general Chinese population. We also assessed the accuracy of FPG alone and two-step strategy in measuring prevalences of undiagnosed diabetes and pre-diabetes using ADA criteria or WHO criteria, respectively.
# Methods
# Ethics statement
The Institutional Review Board of Public Health College, Harbin Medical University approved the study. Written informed consent was obtained from each participant before survey.
## Study population and sampling
The study was conducted between August 2008 and October 2008 in central urban area of Harbin with a population of 3,345,328. A multistage, stratified probability random sampling method was used to obtain a representative sample of the urban residents aged 20 to 74 years in Harbin. First, 5 city districts (Nangang, Daoli, Daowai, Xiangfang, and Pingfang) were typically selected from total of 8 city districts of Harbin. Those city districts which were not selected (Songbei, Hulan, and A'cheng), with a population of 1,405,970, were far away from the central area of Harbin. Second, 3 street districts were randomly selected from each of the 5 sampled city districts. The street districts were sampled based on stratified sampling according to degree of economic development status (high, middle, and low). In the third stage, 2 communities were randomly sampled from each selected street districts. In the final stage, residents aged 20 to 74 years were randomly selected from the selected communities. The sampling proportion within communities was based on the age and sex structure of the selected city district. Only residents who had lived in their current residence for at least 1 year were eligible to participate.
A total of 9,600 eligible residents were selected and invited to participate in the study; 7,939 of them (2,854 men and 5,085 women) completed the survey. The overall response rate was 82.7%: 64.2% for men and 98.7% for women. Twenty six residents, whose data on FPG or 2-h PG were missing, were excluded from analysis, and 7,913 residents (35.9% men; mean [SD] age 49.3±12.3 years) were included in the prevalence analysis. 7,464 residents (35.8% men; mean [SD] age 48.9±12.4 years) were unknown to have diabetes, and were included in the screening test analysis.
## Data collection and examination
All physicians and staff members who conducted the study were trained in the methodology and principles of the study. Data collection was conducted by the physicians in the community health care centers in the participants' residential area in the morning. For participants unable to go to health care centers, data collection was conducted at their homes. Physicians administered a standard questionnaire to obtain information on demographic characteristics, personal and family diabetes medical history, etc. Participants were asked whether, other than during pregnancy for women, a doctor had ever told them that they suffered diabetes. Height and weight were measured while subjects were not heavy clothing. The body mass index was calculated as weight (in kilograms) divided by the square of the height (in meters). The waist circumference was measured at a point immediately above the iliac crest on the midaxillary line. The obesity was defined as body mass index !25 kg/m 2 , and abdominal obesity was defined as waist circumference !85 cm for men and !80 cm for women. [bib_ref] Diabetes and impaired fasting glucose in Mongolian population, Zhang [/bib_ref] All the participants were instructed to maintain their usual diet for 3 days before survey. After 10-12 hours of overnight fasting, venous blood samples were collected in the morning for the measurement of FPG. Then each participant underwent a standard 75-g OGTT, and blood samples were drawn at 120 minutes to measure 2-h PG. Simultaneous detection of FPG and 2-h PG was applied using the hexokinase enzymatic method (Amorsino automatic biochemistry analyzer, Mol 300, China) based on blind method by technicians of Center for Disease Control and Prevention.
## Definitions
Diagnosed diabetes was identified by a positive response from the participants to the question in the interview, "Have you ever been told that you suffered diabetes by a doctor?" Participants without diagnosed diabetes until this study were classify as undiagnosed diabetes (FPG !7.0 mmol/l and/or 2-h PG !11.1 mmol/l), pre-diabetes using ADA criteria (IFG [FPG 5.6 to <7.0 mmol/l] and/or IGT [2-h PG 7.8 to <11.1 mmol/l]), pre-diabetes using WHO criteria (IFG [FPG 6.1 to <7.0 mmol/l] and/or IGT [2-h PG 7.8 to <11.1 mmol/l]), and normal, respectively.These were as golden standards in screening test analysis. Undiagnosed diabetes was stratified into three subcategories: (a) isolated fasting diabetes (FPG !7.0 mmol/l and 2-h PG <11.1 mmol/l); (b) isolated 2h post-load diabetes (FPG <7.0 mmol/l and 2-h PG !11.1 mmol/l); (c) combined fasting and post-load diabetes (FPG !7.0 mmol/l and 2-h PG !11.1 mmol/l). Pre-diabetes using ADA criteria was also stratified into three subcategories: (a) isolated IFG (FPG 5.6 to <7.0 mmol/l and 2-h PG <7.8 mmol/l); (b) isolated IGT (FPG <5.6 mmol/l and 2-h PG 7.8 to <11.1 mmol/l); (c) combined IFG and IGT (FPG 5.6 to <7.0 mmol/l and 2-h PG 7.8 to <11.1 mmol/l). The same applied to the classification of pre-diabetes using WHO criteria. Two-step strategy was that the individuals with an increased FPG (! the FPG cut-points and < the value used to define IFG or undiagnosed diabetes) [bib_ref] Optimal cut-off points for two-step strategy in screening of undiagnosed diabetes: a..., Ye [/bib_ref] were given an OGTT after all subjects first completed FPG test.
# Statistical analysis
The prevalences of diabetes and pre-diabetes were calculated for the overall study subjects and for subgroups according to age and sex. Weights that adjusted for different sampling probabilities and the deviations in sex and age between the sample and the total study subjects were routinely used in all analyses to represent the total population aged 20 to 74 years on the basis of the study sampling scheme and Harbin's urban area population data in 2008. Standard errors and confidence intervals (CI) were estimated with the Taylor series linearization.Standardized prevalences were calculated by the direct method using China adult population aged 20 to 74 from the 2005 National Sample Survey of 1% of Population as the standard population. Prevalence analysis was performed with SURVEYFREQ Procedure that was appropriate to the complex multistage survey design in SAS 9.1.3 software (SAS Institute Inc., Shanghai, China).
We used SPSS version 13.0 for screening test analysis. A P value less than 0.05 was considered statistically significant. We used the method described by Hanley and McNeil to compare the area under the receiver-operating characteristic curves (AUC) for FPG and 2-h PG. Screening potential of FPG was described by sensitivity, specificity, likelihood ratio, post-test probability, etc. Post-test probability was calculated from pre-test probability and likelihood ratio. Pre-test probability of an individual with known characteristics was estimated from the prevalence of the abnormal glucose metabolism in known characteristics. The point with maximization of the sum of sensitivity and specificity was selected as optimal cut-off point. The total cost of two-step strategy was estimate by medical and non-medical cost. The medical, non-medical and total cost for one time FPG (OGTT) test were 7.8, 8.3, 16.1 ¥, respectively. [bib_ref] Optimal cut-off points for two-step strategy in screening of undiagnosed diabetes: a..., Ye [/bib_ref] The cost-effectiveness were calculated as follows Total cost = (all subjects ×one time FPG cost + subjects with increased FPG ×one time OGTT cost)
The cost-effectiveness (cost per case identified) of two-step strategy = total cost Ä undiagnosed diabetes and/or pre-diabetes cases identified.
# Results
## Prevalence of diabetes
The prevalences of undiagnosed diabetes, diagnosed diabetes, and total diabetes were 8.3%, 4.4%, and 12.7%, respectively. The prevalences of isolated fasting diabetes and diagnosed diabetes were similar between men and women; the prevalences of isolated 2h post-load diabetes, combined fasting and post-load diabetes, undiagnosed diabetes, and total diabetes were slightly higher in men than those in women, but the differences were not statistically significant except for combined fasting and post-load diabetes (P = 0.0427). However, the prevalences of these kinds of diabetes at age 40-59 years were significantly higher in men than those in women. In contrast, the sex differences of prevalences of these kinds of diabetes in all the other age groups were not statistically significant.(S1 The prevalences of isolated 2h post-load diabetes, combined fasting and post-load diabetes, undiagnosed diabetes, diagnosed diabetes, and total diabetes increased with age and peaked at age 60-74 years in men and women (P <0.0001). However, the prevalence of isolated fasting diabetes was fluctuated with age. (S1 Awareness, treatment and control of diabetes Of those participants with diabetes, 34.8% were aware of their diabetes, 31.5% (90.7% of participants who were aware of diabetes) were taking medication or nonpharmacological interventions, and 10.8% (31.1% of those treated) were controlled (FPG <7.0 mmol/l, and 2-h PG <11.1 mmol/l). The proportions of awareness, treatment and control of diabetes were similar between men and women.(data not shown) The proportion of undiagnosed diabetes in total diabetes was 65.2%. The proportion of undiagnosed diabetes was significantly higher in men than that in women (68.4% vs. 61.4%, respectively; P = 0.0213) and decreased with age (P = 0.0080). The proportion of undiagnosed diabetes in men at age 20-39 years was significantly higher than those at other age groups and in women.(S1 Prevalence of pre-diabetes Twelve percent (6.3% with IFG and 7.8% with IGT, 12.9% for men and 11.2% for women) or 9.0% (2.2% with IFG and 7.8% with IGT, 9.3% for men and 8.8% for women) of participants were diagnosed with pre-diabetes using ADA criteria or WHO criteria, respectively. The prevalences of isolated IFG, isolated IGT, combined IFG and IGT, total pre-diabetes, and total diabetes and pre-diabetes using ADA criteria or using WHO criteria were slightly higher in men than those in women but without statistics significant except for combined IFG and IGT using WHO criteria. The prevalences of them using ADA criteria were significantly higher in men than those in women at age 40-59 years. Meanwhile, the prevalences of isolated IGT, total prediabetes, and total diabetes and pre-diabetes using WHO criteria were also higher in men than those in women at age 40-59 years (P <0.05). The prevalences of isolated IGT, pre-diabetes, and total diabetes and pre-diabetes increased with age and peaked at age 60-74 years using ADA criteria and WHO criteria (P <0.0001). However, isolated IFG and Combined IFG and IGT were fluctuated with age.(S2 Standardized prevalences
The standardized prevalences of diabetes, pre-diabetes, and total diabetes and pre-diabetes were 12.4%, 11.5%, and 23.9% using ADA criteria and 12.4%, 8.6%, and 21.0% using WHO criteria based on the 2005 National Sample Survey of 1% of Population.
## The screening test analysis
The optimal FPG cut-off points were 5.6 mmol/l for previously undiagnosed diabetes, 5.3 mmol/l for both diabetes and pre-diabetes or pre-diabetes using ADA criteria, 5.0 mmol/l for pre-diabetes using WHO criteria, and 4.9 mmol/l for IGT. The AUCs and sensitivities of these points were lower for FPG than for 2-h PG in screening both diabetes and pre-diabetes, undiagnosed diabetes, and pre-diabetes. [fig_ref] Table 1: AUCs and screening potential for FPG alone in screening undiagnosed diabetes and/or... [/fig_ref] Nevertheless, the AUCs for FPG (sensitivity, Specificity) were greater than 0.7 (60%, 70%), and Kappa values between optimal FPG cut-off points and gold standards were statistically significant (P <0.001). Therefore further OGTT (two-step strategy) should be conducted to increase specificity for screening diabetes and/or pre-diabetes. Using the optimal FPG cut-off points for screening pre-diabetes using WHO criteria or IGT alone, Kappa values (<0.4) and specificity (<80%) were lower, and OGTT alone (one-step strategy) should be better conducted. [fig_ref] Table 1: AUCs and screening potential for FPG alone in screening undiagnosed diabetes and/or... [/fig_ref] The total costs per case of these points were relatively lower. 5.6 mmol/l for previously undiagnosed diabetes and 5.3 mmol/l for both diabetes and pre-diabetes using ADA criteria were with the least medical cost per case. Medical and total cost per case of these points were both diabetes and prediabetes (ADA criteria), ¥51, ¥110, both diabetes and pre-diabetes (WHO criteria), ¥69, ¥154, , further OGTT was not conducted for subjects with FPG !5.6 mmol/l (!6.1mmol/l) in the case of screening for both diabetes and pre-diabetes using ADA criteria (WHO criteria). In , further OGTT was not conducted for subjects with FPG !7.0 mmol/l in the case of screening for undiagnosed diabetes or both diabetes and pre-diabetes or pre-diabetes alone or IGT. pre-diabetes alone (ADA criteria), ¥112, ¥258, diabetes alone, ¥117, ¥261, pre-diabetes alone (WHO criteria), ¥166, ¥399, IGT alone, ¥205, ¥502, ascendingly.(S4 The optimal FPG cut-off points for previously undiagnosed diabetes were 5.3 mmol/l (sensitivity 86.1%; specificity 81.5%) at age 60-74 years and 5.6 mmol/l at age<60 years (sensitivity 84.6%; specificity 92.1%), respectively. The optimal FPG cut-off points for IGT were 4.9 mmol/l in men (sensitivity 60.5%, specificity 73.2%) and women (sensitivity 62.2%, specificity 75.2%). The pre-test probabilities of undiagnosed diabetes (pre-diabetes using ADA criteria, prediabetes using WHO criteria) were 11.6% (16.4%, 12.7%) for obesity and 12.0% (15.9%, 11.7%) for abdominal obesity. The post-test probabilities of them were shown in [fig_ref] Table 1: AUCs and screening potential for FPG alone in screening undiagnosed diabetes and/or... [/fig_ref]. Positive likelihood ratios of different FPG cut-off points were shown in [fig_ref] Table 2: Positive likelihood ratios of different FPG cut-off points [/fig_ref].
The accuracy of FPG alone in measuring prevalence of undiagnosed diabetes and pre-diabetes Among those classified as normal by FPG, 91.2% or 89.6% were classified equally by 2-h PG using ADA criteria or WHO criteria, respectively. Of those people with IFG using WHO criteria, 32.4% had diabetes, 31.0% had combined IFG and IGT, and 36.6% had isolated IFG. [fig_ref] Table 3: Classification of FPG alone and 2-h PG alone for abnormal glucose metabolism... [/fig_ref] The propotions of isolated 2h post-load diabetes in undiagnosed diabetes (isolated IGT in total pre-diabetes) were relatively higher in all the age groups and in men and women, especially at age 60-74 years (using WHO criteria).(S1 and S2 Tables)
# Discussion
Two national prevalence surveys of diabetes suggested that there were marked geographical differences in the prevalence of diabetes in China, with much higher prevalence in urban areas and in northern China. [bib_ref] Prevalence of diabetes among men and women in China, Yang [/bib_ref] [bib_ref] Prevalence of diabetes and impaired fasting glucose in the Chinese adult population:..., Gu [/bib_ref] Accordingly, measuring prevalence and screening diabetes should be advocated more in here. The study was a large-scale population-based survey. All participants underwent an OGTT. The diabetes and pre-diabetes were defined based on performing FPG and 2-h PG from an OGTT. This was designed to provide more strong evidence that assessed the performance of screening strategy in general population rather than for high risk population and more accurate estimates of the prevalences of diabetes and pre-diabetes than previous studies in Northern China. [bib_ref] A cross-sectional study on impaired fasting glycaemia and diabetes mellitus in residents..., Zhao [/bib_ref] [bib_ref] Type 2 diabetes and impaired glucose tolerance in North-China-based rural community adults, Wei [/bib_ref] Furthermore, the standardized training and the quality-control procedures were completed to ensure the validity of the results during the study. The prevalence of diabetes in Harbin was 12.7%, which was similar to that in U.S. in 2005 when using the U.S. 2000 Census population as standard population (Harbin 12.2% vs. U.S. 12.6%) and higher than the mean prevalence in China in 2008 (standardized prevalence based on China population in 2005: 12.4% vs. 9.7%). [bib_ref] Prevalence of diabetes among men and women in China, Yang [/bib_ref] [bib_ref] Full accounting of diabetes and pre-diabetes in the U.S. population in 1988, Cowie [/bib_ref] The proportion of undiagnosed diabetes was 65.2%, higher than that in U.S. in 2005 (39.8%) [bib_ref] Full accounting of diabetes and pre-diabetes in the U.S. population in 1988, Cowie [/bib_ref] and China in 2008 (60.7%). [bib_ref] Prevalence of diabetes among men and women in China, Yang [/bib_ref] In addition, the prevalence of pre-diabetes in Harbin was lower than that in U.S. in 2005 (standardized prevalence based on the U.S. 2000 Census population using ADA criteria: Harbin 11.9% vs. U.S. 29.0%) or China in 2008 (standardized prevalence based on China population in 2005 using WHO criteria: Harbin 8.6% vs. China 15.5%). [bib_ref] Prevalence of diabetes among men and women in China, Yang [/bib_ref] [bib_ref] Full accounting of diabetes and pre-diabetes in the U.S. population in 1988, Cowie [/bib_ref] Although the awareness, treatment and control of diabetes were relatively low in total diabetes patients, 90.7% of diagnosed diabetes were treated; among those treated, 31.1% were controlled. Therefore, screening diabetes and pre-diabetes, and improving the awareness would be more urgently needed for the intervention of hyperglycemic condition than other strategies in the adult population in Northern China.
The accuracy (sensitivity, AUC of FPG) and cost-effectiveness of the optimal FPG cutpoints for two-step strategy were relatively better. Especially in screening diabetes, the accuracy of 5.6 mmol/l as cut-point was better than other tests (A1C, fasting capillary glucose, and Chinese diabetes risk score). [bib_ref] Performance of an A1C and fasting capillary blood glucose test for screening..., Zhou [/bib_ref] [bib_ref] Evaluation of two screening methods for undiagnosed diabetes in China: an cost-effectiveness..., Zhang [/bib_ref] [bib_ref] Nonlaboratory-based risk assessment algorithm for undiagnosed type 2 diabetes developed on a..., Zhou [/bib_ref] In screening pre-diabetes using WHO criteria in Chinese people, the accuracy of 5.0 mmol/l as cut-point was better than A1C and fasting capillary glucose. [bib_ref] Performance of an A1C and fasting capillary blood glucose test for screening..., Zhou [/bib_ref] Our study demonstrated higher optimal FPG cut-off points in screening undiagnosed diabetes than that in the paper by Ye et al [bib_ref] Optimal cut-off points for two-step strategy in screening of undiagnosed diabetes: a..., Ye [/bib_ref] (5.6 mmol/l vs. 5.4 mmol/l). Meanwhile, our study suggested a lower cost using optimal FPG cut-off points in screening undiagnosed diabetes than that in the paper by Ye et al [bib_ref] Optimal cut-off points for two-step strategy in screening of undiagnosed diabetes: a..., Ye [/bib_ref] (¥261 vs. ¥615), because of higher prevalence of diabetes (12.7% vs. 8.8%) and the proportion of undiagnosed diabetes in total diabetes (65.2% vs. 40.9%) probably. According to the goal of screening (diabetes alone or both diabetes and prediabetes or pre-diabetes alone or IGT alone), criteria (2003 ADA criteria or 1999 WHO criteria) and funds (medical cost or total cost), different FPG cut-points should be chosen. In China, more and more people receive health examination paid by the goverment or medical insurance system or individual, and FPG is a routine examination in health examination center or community health care center. The optimal FPG cut-off points for census of diabetes and/or pre-diabetes in general population would guide physician decision-making in health examination center or community health care center. Meanwhile positive likelihood ratio and post-test probability allow the physician to better interpret the results of FPG and predict the likelihood of a true positive result.
If diabetes or pre-diabetes were identified by FPG alone as a one-step strategy in our survey, the prevalence exhibited underestimates of 26.0% for total diabetes, 40.0% for undiagnosed diabetes, 47.2% for pre-diabetes using ADA criteria, and 75.4% for pre-diabetes using WHO criteria. If further OGTT was conducted for subjects with FPG 5.6 to <7.0 mmol/l (6.1 to <7.0 mmol/l) as a two-step strategy, the prevalence exhibited underestimates of 12.2% (17.3%) for total diabetes and 18.7% (26.5%) for undiagnosed diabetes.(S1 and S2 The similar findings were found by the DECODA, which reported that using the two-step strategy would fail to detect every fourth individual with diabetes and every second individual with IGT in Asia.Using the two-step strategy for pre-diabetes using WHO criteria (FPG 5.3 mmol/l as cut-point) or ADA criteria (FPG 5.0 mmol/l as cut-point), the underestimates of prevalence reduced to nearly 38%. (S2 Previous studies have reported a 2-3-fold increase in prevalence of IFG using the new ADA recommended criteria compared with WHO criteria. [bib_ref] Lowering the criterion for impaired fasting glucose: impact on disease prevalence and..., Tai [/bib_ref] [bib_ref] Creating a pandemic of prediabetes: the proposed new diagnostic criteria for impaired..., Borch-Johnsen [/bib_ref] In our survey, the prevalence of IFG increased from 2.2% (WHO criteria) to 6.3% (ADA criteria). The risk of diabetes and coronary heart disease for IGT was higher than that for IFG, [bib_ref] Lowering the criterion for impaired fasting glucose: impact on disease prevalence and..., Tai [/bib_ref] and IGT carried more risk of death than IFG.In our study, people with IGT had lower fasting plasma glucose level than those with IFG in men and women using ADA criteria or using WHO criteria.(S5 Therefore, it can be inferred that the coronary heart disease risk for lower normal fasting plasma glucose level is higher than that for IFG, which is in agreement with other study. [bib_ref] Coronary disease risk and fasting glucose levels in a nondiabetic population, Onat [/bib_ref] ADA reduced the lower FPG cut-point, in part to ensure that prevalence of IFG was similar to that of IGT.However, in our survey, underestimate of isolated IGT decreased only from 6.8% (WHO criteria) to 5.7% (ADA criteria), and lower FPG cut-off point (4.9 mmol/l) for two-step strategy should be used to improve sensitivity. There was also some evidence that diabetes diagnosed solely on the basis of 2-h PG was associated with a worse prognosis than diabetes diagnosed in the sight of FPG alone for mortality and retinopathy. [bib_ref] Importance of OGTT for diagnosing diabetes mellitus based on prevalence and incidence..., Ito [/bib_ref] Therefore, a two-step strategy or an OGTT alone should be necessary for screening diabetes and pre-diabetes in the Chinese population, because it could detect isolated 2h post-load diabetes and isolated IGT, which were found in large portions in our study and the other study. [bib_ref] Prevalence of diabetes among men and women in China, Yang [/bib_ref] Although the sensitivity and cost-effectiveness of two-step strategy for screening IGT (4.9 mmol/l as FPG cut-off points) was relatively lower, the optimal cut-off point in screening IGT could provide the basis for the screening of coronary heart disease risk factors. The study [bib_ref] Fasting, non-fasting glucose and HDL dysfunction in risk of pre-diabetes, diabetes, and..., Onat [/bib_ref] suggested IGT independently predicted coronary heart disease risk in women. Therefore, optimal FPG cut-off points for two-step strategy in women in screening IGT should be identified, but the sex difference in the optimal cut-off point in screening IGT have not been found.
The study suffered from some limitations. There was a lower response rate among men than among women in this survey. The proportion of those that had never measured blood glucose among participants was 65.0%, and there was no difference between men and women (63.8% vs. 66.3%); the distribution of "How long ago have you been measured blood glucose recently?" (less than 30 days, 1-6 months, 7-12 months, and more than one year) also was similar between men and women (data not shown); the proportion of undiagnosed diabetes in men was higher than that in women. These suggested that the non-response among men was not caused by awareness of their diabetes, and in some extent the low response rate in men may not affect the prevalence of study subjects. Consequently the possibility of selection bias was minimised.
In conclusion, approximately a quarter of the general adult population has hyperglycemic condition, and diabetes has become a major public health challenge in Harbin, Northern China. The prevalences of diabetes and pre-diabetes at age 40-59 years were higher in men than those in women, and special attention should be paid to men aged 40 to 59 years. No matter which criteria should be used (ADA criteria or WHO criteria), greater efforts to use optimal FPG cut-off points for two-step strategy and identify those with IGT and with isolated 2h postload diabetes in general Chinese population may be more effective and less costly for reducing the missed diagnosis of hyperglycemic condition.
Supporting Information S1 Prevalence of diabetes, proportion of undiagnosed diabetes in total diabetes, proportion of isolated 2h post-load diabetes in undiagnosed diabetes, and missed diagnosis of undiagnosed diabetes using optimal FPG cut-off point for two-step strategy by age and sex, in urban Harbin of China, 2008 (%, 95% CI). (DOC) S2 Prevalence of pre-diabetes, proportion of isolated IGT in total pre-diabetes, and missed diagnosis of pre-diabetes using optimal FPG cut-off points for two-step strategy by age and sex based on ADA and WHO criteria, in urban Harbin of China, 2008 (%, 95% CI). (DOC) S3 AUCs and screening potential for 2-h PG alone in screening undiagnosed diabetes and/or pre-diabetes with optimal cut-off points. (DOC) S4 The cost-effectiveness of two-step strategy at different fasting plasma glucose cutoff points in screening IGT, undiagnosed diabetes and/or pre-diabetes.
[table] Table 1: AUCs and screening potential for FPG alone in screening undiagnosed diabetes and/or pre-diabetes with optimal cut-off points.AUCs, the area under the receiver-operating characteristic curves; FPG, fasting plasma glucose. *Number of participants based on golden standard. Pre-diabetes includes IFG and/or IGT.† Fig 1.The total cost per case identified by two-step strategy at different FPG cut-points for further OGTT test. In [/table]
[table] Table 2: Positive likelihood ratios of different FPG cut-off points. [/table]
[table] Table 3: Classification of FPG alone and 2-h PG alone for abnormal glucose metabolism in urban Harbin of China, 2008 (%, 95% CI). [/table]
|
Huge gastric bezoar caused by honeycomb, an unusual complication of health faddism: a case report
We report a young healthy woman, who believed that the consumption of large amounts of honeycomb would lead to good health and who finally developed a huge gastric bezoar of hard consistency. The conventional endoscopic techniques failed to manage the bezoar. Using the combination of injection of hydrogen peroxide 3% solution inside the bezoar to induce disintegration and a special designed needle-knife sphincterotome (bezotome) we managed to remove the bezoar in fragments. To the best of our knowledge this is the first reported bezoar caused by honeycomb.
# Introduction
Bezoars are foreign bodies found mainly in the stomach, which are composed of plant and vegetables (phytobezoars), persimmous (diospyrobezoars), hair (trichobezoars), milk (lactobezoars) or other bezoars [bib_ref] Bezoars and foreign bodies of the stomach, Lee [/bib_ref]. Their management includes a wide spectrum of treatment options, from conservative treatment to surgery or endoscopic intervention [bib_ref] Bezoars and foreign bodies of the stomach, Lee [/bib_ref] [bib_ref] Bezoars: classification, pathophysiology and treatment, Andrus [/bib_ref].
We describe the first case of a huge bezoar of very hard consistency, made from honeycomb, which required sophisticated endoscopic techniques for its removal.
## Case presentation
A 44-year-old Greek woman was referred to our department for endoscopic treatment of a huge gastric bezoar. Past medical history of the patient revealed daily consumption of large quantities of honeycomb during the last 2 months, because she believed that the honeycomb might have beneficial effect on the irritable bowel syndrome and on her health in general. Physical examination and laboratory data were unremarkable. During last ten days, she presented episodes of epigastric pain associated with nausea, especially after eating. Despite the initiation of treatment with proton pump inhibitors, the symptoms were not relieved. Upper endoscopy performed by a private gastroenterologist disclosed a yellow coloured huge bezoar, very hard to touch with forceps.
We performed the intervention with propofol administration to achieve a well-sedated patient. Endoscopic examination of stomach confirmed previous findings . We tried to fragment the bezoar with the use of snares and baskets but only superficial pieces were removed, leaving the bezoar practically intact. Trying to disintegrate the bezoar, we injected, via a variceal needle, inside it 100 ml of hydrogen peroxide (H 2 O 2 ) 3% solution. 24 hours later, we repeated the endoscopy and using a modified needle-knife (length of cutting wire 20 mm versus 5 mm of a conventional needle-knife) we performed fragmentation and removal of the bezoar.
# Discussion
The majority of gastric bezoars occur in patients who have undergone previous gastric surgery [bib_ref] Bezoars: classification, pathophysiology and treatment, Andrus [/bib_ref]. Loss of antral and pyloric function because of partial gastric resection and reduced gastric motility following vagotomy are major causes of gastric stasis [bib_ref] Bezoars and foreign bodies of the stomach, Lee [/bib_ref] [bib_ref] Bezoars: classification, pathophysiology and treatment, Andrus [/bib_ref]. Other predisposing conditions are impaired mastication, gastroparesis/hypochlorhydria, anatomic abnormalities such as diverticula or gastric outlet obstruction, inadequate fluid intake leading to dehydration and inspissation of enteric feeding formula [bib_ref] Bezoars and foreign bodies of the stomach, Lee [/bib_ref] [bib_ref] Bezoars: classification, pathophysiology and treatment, Andrus [/bib_ref].
The clinical presentation of gastric bezoars includes abdominal pain (70%), vomiting and nausea (64%), and early satiety [bib_ref] Bezoars and foreign bodies of the stomach, Lee [/bib_ref] [bib_ref] Bezoars: classification, pathophysiology and treatment, Andrus [/bib_ref] [bib_ref] Gastrointestinal bezoars a retrospective analysis of 34 cases, Erzurumlu [/bib_ref] [bib_ref] Endoscopic management of huge bezoars, Wang [/bib_ref]. Obstructive symptoms may be intermittent, owing to a ball valve mechanism of obstruction [bib_ref] Bezoars: classification, pathophysiology and treatment, Andrus [/bib_ref]. In some cases the initial presentation may be that of iron deficiency anemia The diagnosis is made by abdominal ultrasound, computed tomography, barium meal examination or endoscopy [bib_ref] Gastrointestinal bezoars: sonographic and CT characteristics, Ripolles [/bib_ref].
Current management includes conservative treatment (meaning waiting for them to disintegrate and pass spontaneously) if the bezoars are small, which however carries the risk of small bowel obstruction in patients who have had gastrectomy; medical treatment with enzymes and prokinetic agents [bib_ref] Update on the medicinal management of phytobezoars, Walker-Renard [/bib_ref] ; endoscopic management; and surgical removal. Huge hard bezoars usually require mechanical treatment [bib_ref] Gastrointestinal bezoars a retrospective analysis of 34 cases, Erzurumlu [/bib_ref]. Operation is necessary if endoscopic removal fails. Endoscopic management includes enzymatic dissolution by injecting cellulase, use of a water jet, a drill device, tripod forceps, polypectomy snare plus diathermy, Dormia basket, mechanical lithotriptor, or neodymium-yttrium-aluminium-garnet (Nd: YAG) [bib_ref] Bezoars and foreign bodies of the stomach, Lee [/bib_ref] [bib_ref] Bezoars: classification, pathophysiology and treatment, Andrus [/bib_ref] [bib_ref] Gastrointestinal bezoars a retrospective analysis of 34 cases, Erzurumlu [/bib_ref].
Our case is very intriguing because the consumption of honeycomb has not been reported to lead to gastric bezoar formation. Moreover, the honeycomb-bezoar was very hard to be cut with a snare or basket. We injected inside the bezoar, 100ml H 2 O 2 3% solution via a variceal needle. The aim of this injection was the contribution of H 2 O 2 in disintegration of the bezoar. The endoscopy was repeated 24 hours later. Using a modified needle-knife (bezotome) and monopolar cutting current we were able to incise the bezoar into fragments, which were easily retrieved.
# Conclusion
Our case shows that even a huge solid bezoar with hard consistency does not need to be operated on. By using sophisticated endoscopic techniques the fragmentation and removal of such bezoars is feasible.
(page number not for citation purposes)
ConsentWritten informed consent was obtained from the patient for publication of this case report and accompanying image. A copy of the written consent is available for review by the Editor-in-Chief of this journal.Competing interestsThe authors declare that they have no competing interests.Authors' contributionPK performed the endoscopy and was contributor in writing the manuscript. IP, GC, and CZ analyzed and interpreted the patient data and were contributors in writing the manuscript. TK, GL and KF reviewed the relative literature. JK was major contributor in revising the manuscript critically for important intellectual content. All authors read and approved the final manuscript.Do you have a case to share? |
Molecular Dynamics Simulations of the Bacterial UraA H+-Uracil Symporter in Lipid Bilayers Reveal a Closed State and a Selective Interaction with Cardiolipin
The Escherichia coli UraA H + -uracil symporter is a member of the nucleobase/ascorbate transporter (NAT) family of proteins, and is responsible for the proton-driven uptake of uracil. Multiscale molecular dynamics simulations of the UraA symporter in phospholipid bilayers consisting of: 1) 1-palmitoyl 2-oleoyl-phosphatidylcholine (POPC); 2) 1-palmitoyl 2-oleoyl-phosphatidylethanolamine (POPE); and 3) a mixture of 75% POPE, 20% 1-palmitoyl 2-oleoyl-phosphatidylglycerol (POPG); and 5% 1-palmitoyl 2-oleoyl-diphosphatidylglycerol/cardiolipin (CL) to mimic the lipid composition of the bacterial inner membrane, were performed using the MARTINI coarse-grained force field to self-assemble lipids around the crystal structure of this membrane transport protein, followed by atomistic simulations. The overall fold of the protein in lipid bilayers remained similar to the crystal structure in detergent on the timescale of our simulations. Simulations were performed in the absence of uracil, and resulted in a closed state of the transporter, due to relative movement of the gate and core domains. Anionic lipids, including POPG and especially CL, were found to associate with UraA, involving interactions between specific basic residues in loop regions and phosphate oxygens of the CL head group. In particular, three CL binding sites were identified on UraA: two in the inner leaflet and a single site in the outer leaflet. Mutation of basic residues in the binding sites resulted in the loss of CL binding in the simulations. CL may play a role as a "proton trap" that channels protons to and from this transporter within CL-enriched areas of the inner bacterial membrane.Author SummarySymporters are proteins that are responsible for the co-transport of ions and small molecule solutes across cell membranes. UraA is an example of a symporter, and is responsible for the proton-driven uptake of uracil in bacteria like E. coli. Despite its importance as a member of a large family of nucleobase/ascorbate transporters (NAT) and the existence of PLOS Computational Biology | The funders had no role in study design, data collection structural and functional data, the mechanism by which UraA transports uracil across the bacterial membrane, and in particular the role of its diverse and complex lipid environment in the transport mechanism, remains elusive. In this study, we have used a multiscale computational methodology to examine the dynamics of UraA and to elucidate its interactions with lipids that resemble its native environment in the bacterial inner membrane. Our results demonstrate that negatively-charged lipids in the membrane (phosphatidylglycerol and cardiolipin) associate preferentially with UraA and may play a role in its function. Additionally, our simulations resulted in a closed state of UraA, a likely intermediate in the transport mechanism that may not be readily accessible by experimental methods.UraA Transporter in a Lipid Bilayer PLOS Computational Biology |
# Introduction
Crystal structures of membrane proteins, particularly from bacterial sources, are being determined at an increasing rate (http://blanco.biomol.uci.edu/mpstruc/). The membrane protein is typically isolated and crystallized in the presence of detergents and indeed, non-native detergent molecules are often found tightly associated with the membrane protein. It is therefore important to relate the crystal structure back to the state of the protein in a native lipid bilayer [bib_ref] Influence of solubilizing environments on membrane protein structures, Cross [/bib_ref] [bib_ref] Tolerance to changes in membrane lipid composition as a selected trait of..., Sanders [/bib_ref].
Molecular dynamics (MD) simulations of membrane proteins in increasingly complex lipid bilayers [bib_ref] Atom-scale molecular interactions in lipid raft mixtures, Niemelä [/bib_ref] can now be applied with confidence to build dynamic models of membrane proteins in a native milieu [bib_ref] Membrane proteins: molecular dynamics simulations, Lindahl [/bib_ref] [bib_ref] Molecular simulation approaches to membrane proteins, Stansfeld [/bib_ref]. A key question to be explored is to what extent the crystal structure determined in the presence of detergents is the same as the structure of the membrane protein in a lipid bilayer. For example, the influenza M2 protein transmembrane domain shows significant structural differences between detergent-containing crystals and the structure in lipid bilayers determined by solid state NMR [bib_ref] Influence of solubilizing environments on membrane protein structures, Cross [/bib_ref]. Furthermore, there are a number of examples of membrane protein structures that contain tightly associated lipids bound to specific sites on the protein [bib_ref] Lipids in membrane protein structures, Palsdottir [/bib_ref] [bib_ref] Specific protein-lipid interactions in membrane proteins, Hunte [/bib_ref]. MD simulations allow us to explore molecular motions and the dynamic interactions between lipids and proteins, providing a complementary approach to the temporal and spatially averaged static structures determined by X-ray crystallography. Recent studies using MD simulations have identified cardiolipin (CL) binding sites in the cytochrome bc 1 transporter [bib_ref] Evidence for cardiolipin binding sites on the membrane-exposed surface of the Cytochrome..., Arnarez [/bib_ref] and in the cytochrome c oxidase [bib_ref] Identification of cardiolipin binding sites on cytochrome c oxidase at the entrance..., Arnarez [/bib_ref] , and PIP 2 binding sites in Kir channels [bib_ref] Simulation-based prediction of phosphatidylinositol 4,5-bisphosphate binding to an ion channel, Schmidt [/bib_ref] [bib_ref] PIP2-binding site in Kir channels: definition by multiscale biomolecular simulations, Stansfeld [/bib_ref]. MD simulations were also used to identify interactions of lipids with the aquaporin family [bib_ref] Multiscale simulations reveal conserved patterns of lipid interactions with aquaporins, Phillip [/bib_ref] [bib_ref] Molecular driving forces defining lipid positions around aquaporin-0, Aponte-Santamaría [/bib_ref] and other proteins [bib_ref] Lipid exchange mechanism of the cholesteryl ester transfer protein clarified by atomistic..., Koivuniemi [/bib_ref] [bib_ref] Identification of cholesterol binding sites in the Serotonin1A Receptor, Sengupta [/bib_ref]. Recently MD simulations were also used to study the conformation of proton-driven transporters LacY [bib_ref] Flexible gates generate occluded intermediates in the transport cycle of LacY, Stelzl [/bib_ref] and UpaA [bib_ref] Identification of the substrate recognition and transport pathway in a Eeukaryotic member..., Kosti [/bib_ref]. So, how close is the structure of a membrane protein in a lipid bilayer determined by MD simulations to the original structure of the protein determined by X-ray crystallography in the presence of detergents? Can we use MD simulations to detect specific lipid-protein interactions not often found in crystal structures and study their dynamics?
We explored these questions using the E. coli UraA H + -uracil symporter, the crystal structure of which was determined with bound uracil in the detergent n-nonyl-β-D-glucopyranoside (NG; [fig_ref] Figure 1: Structure of UraA [/fig_ref]. UraA is a member of the nucleobase/ascorbate transporter NAT family of proteins (also known as nucleobase-cation symporter-2 [NCS2]) which can be found in all species [bib_ref] The nucleobase-ascorbate transporter (NAT) family: genomics, evolution, structure-function relationships and physiological role, Gournas [/bib_ref]. The majority of the NAT proteins are responsible for the uptake of xanthine, uric acid or uracil with H + in the case of most bacteria and Na + in the case of the mammalian transporters [bib_ref] Identification of the substrate recognition and transport pathway in a Eeukaryotic member..., Kosti [/bib_ref] [bib_ref] The nucleobase-ascorbate transporter (NAT) family: genomics, evolution, structure-function relationships and physiological role, Gournas [/bib_ref]. In mammals, NAT proteins are also responsible for the transport of vitamin C [bib_ref] The nucleobase-ascorbate transporter (NAT) family: genomics, evolution, structure-function relationships and physiological role, Gournas [/bib_ref]. Including UraA, there are ten members of the NCS2 family of transporters in E. coli. Extensive mutagenesis and in silico studies identified a conserved 11-residue NAT motif in the [bib_ref] Structure and mechanism of the uracil transporter UraA, Lu [/bib_ref] are shown in grey and cyan respectively. B. Snapshot from the end of one of the UraA-AT simulations. The same colors for UraA as in A are used. The POPE, POPG and CL are shown in blue, red and green respectively. The water solvent is shown in red. C. Sequence of UraA. The secondary structure is shown below the sequence. The same colors for the core and the gate domains as in A are used.
NAT/NCS2 family ([Q/E/P-N-x-G-x-x-x-x- that is crucial for the function of its members. Mutation of the conserved residues of the motif in the Aspergillus nidulans uric acid/xanthine permease, UapA and the E. coli xanthine permease XanQ (YgfO) have shown that residues within the NAT motif contribute to the substrate specificity [bib_ref] Insights to the evolution of Nucleobase-Ascorbate Transporters (NAT/NCS2 family) from the Cys-scanning..., Frillingos [/bib_ref]. Extensive cysteinescanning mutagenesis studies on XanQ have identified key functional residues, including Gln324 and Asn325 in the NAT motif within TM10 [bib_ref] Insights to the evolution of Nucleobase-Ascorbate Transporters (NAT/NCS2 family) from the Cys-scanning..., Frillingos [/bib_ref].
The 2.8 Å crystal structure of the UraA uracil transporter [fig_ref] Figure 1: Structure of UraA [/fig_ref] , the first member of the NCS2 family to be solved, revealed a monomer protein with 14 transmembrane segments with TM segments 1-7 and TMs 8-14 having an inverted topology [bib_ref] Structure and mechanism of the uracil transporter UraA, Lu [/bib_ref]. Uracil was located in a central cavity open via a water-filled passage to the cytosolic side of the membrane, bound between a pair of short anti-parallel β-strands in TM3 and 10 via a predominant interaction with an invariant Glu241 [bib_ref] Structure and mechanism of the uracil transporter UraA, Lu [/bib_ref]. The crystal structure was obtained in the detergent n-nonyl-β-D-glucopyranoside with one inverted detergent molecule associated with the protein with the glucose head group of the detergent donating a hydrogen bond to the carbonyl in uracil, thereby stabilizing its binding.
In this study a serial multiscale MD simulation approach [bib_ref] Multiscale modeling of biomolecular systems: in serial and in parallel, Ayton [/bib_ref] was used to study the dynamics of the UraA (3QE7), simulated without bound uracil or detergent in a native-like lipid bilayer, enabling us to explore its specific interactions with the lipids. To this end, a lipid bilayer containing POPC lipids was self-assembled around UraA and then simulations in pure POPC, pure POPE and a POPE (75%), POPG (20%) and CL (5%) mixture to mimic the native bacteria membrane in the coarse-grained and atomistic resolutions were performed (see Methods, [fig_ref] Table 1: Summary of the principal simulations [/fig_ref] and S1 Table for more information). Simulations in the absence of substrate (and bound detergent) resulted in a closed state of the protein. The various lipid bilayers that we have used allowed us to study the differences in the conformational states of UraA in different lipid environments. We discovered a preferential interaction of anionic lipids PG and CL with UraA. In particular, two CL sites were identified on UraA on the inner leaflet and a single CL site in the outer leaflet. The inner association involved a preferential interaction of the phosphate oxygen of the CL head group with R299 and with R4 near the cytosolic exit site of the transporter at one site, and with R265 and K109 at the second site. CL also showed a preferential interaction with UraA on the outer leaflet involving K321 and W212 in the third site. In silico mutation of these basic residues led to loss of CL binding highlighting the importance of an electrostatic interaction in the association of this anionic lipid with UraA.
# Methods
## Coarse-grained molecular dynamics simulations
The simulation protocols used in the study are summarized schematically in S1 Fig. Coarsegrained molecular dynamics (CG-MD) simulations were performed using the MARTINI force field [bib_ref] The MARTINI coarsegrained force field: extension to proteins, Monticelli [/bib_ref]. In the MARTINI CG model used [bib_ref] The MARTINI coarsegrained force field: extension to proteins, Monticelli [/bib_ref] [bib_ref] The MARTINI force field: coarse grained model for biomolecular simulations, Marrink [/bib_ref] four heavy atoms (i.e. non-hydrogen) are represented as one CG particle. Four CG particle types exist: charged (Q), non-polar (N), apolar (I) and polar (P). Additional particle subtypes exist for assigning H-bond capabilities. Each amino acid is represented by the aforementioned particles representing the alpha carbon and the various portions of the side chain [bib_ref] The MARTINI coarsegrained force field: extension to proteins, Monticelli [/bib_ref] [bib_ref] The MARTINI force field: coarse grained model for biomolecular simulations, Marrink [/bib_ref]. An elastic network was applied to all backbone particles using a cut-off distance of 7 Å to model secondary and tertiary structure. The use of an elastic network model (ENM) to maintain protein tertiary structure during the CG-MD simulations restricts any major conformational changes of the protein. However, during the subsequent atomistic simulations the ENM is removed. In the CG-MD simulations the original MARTINI water model [bib_ref] The MARTINI coarsegrained force field: extension to proteins, Monticelli [/bib_ref] [bib_ref] The MARTINI force field: coarse grained model for biomolecular simulations, Marrink [/bib_ref] was used, without addition of antifreeze particles to any of the systems. The starting structure of the UraA transporter, without the bound uracil or single detergent molecule, was taken from the protein coordinates of 2.8 Å crystal structure deposited in the PDB (3QE7). The CG protein was centered in a 14 x 14 x 12 nm 3 simulation box of sufficient size to accommodate the 429-residue protein monomer (45 kDa) and about 400 phospholipid molecules. UraA was oriented to align the transport path of the protein with the z axis of the simulation box with the bottom of the structure facing the cytosol and the top the periplamic space. UraA contains 14 transmembrane segments and both the N-and C-termini face the cytosol.
Insertion of UraA in a lipid bilayer. POPC lipids were placed randomly in the simulation box and the system was solvated with CG water molecules and NaCl ions to *150 mM were added to neutralize the system. In particular 296 Na + and 176 Clions were added to neutralize the -84e charge of the POPG lipids, the -42e charge of the CL and the +6e net charge of the protein. These ions were used to neutralize the systems, and no specific interactions between the ions and the lipids or the protein were observed. The system was energy-minimized using the steepest descent algorithm embedded in GROMACS. Subsequently, a CG-MD simulation for 100 ns was run which allowed the lipids to self-assemble forming a bilayer around the membrane protein (S1 Video). Note that during the self-assembly simulation the protein Cα atoms were restrained so the protein backbone could not move, while the side chains were mobile. At the end of the 100 ns self-assembly CG-MD simulation with UraA now in a bilayer, the protein restraints were removed and further 1 μs of production simulation with the protein free to diffuse in the bilayer was run. This allowed the protein to optimize both its position in the lipid bilayer (e.g. optimize the protein displacement and tilt angle relative to the bilayer) and its interaction with surrounding lipids.
Lipid exchange. At the end of the unrestrained 1 μs simulation in the POPC bilayer the protein was re-centered in the simulation box and the POPC lipids were exchanged for 100% POPE, or 75% POPE, 20% POPG, 5% CL [bib_ref] Mechanical Properties of Coarse-Grained Bilayers Formed by Cardiolipin and Zwitterionic Lipids, Dahlberg [/bib_ref] using a locally written script (see [fig_ref] Table 1: Summary of the principal simulations [/fig_ref] and S1 [fig_ref] doi [/fig_ref]. The script randomly replaces one POPC molecule (in the case of POPE or POPG) or two POPC molecules (in the case of CL, which consists of two diacylphosphatidic acid moieties connected by a central glycerol) with the desired number of POPE and/or CL molecules in each leaflet thus creating a new bilayer where the lipids are randomly distributed in the two leaflets in the desired concentrations. After lipid exchange the systems were energy minimized followed by 5 ns of equilibration at 323K with the protein Cα atoms restrained. It should be emphasized that in the latter bilayer the lipids were randomly distributed between the two leaflets. Seven independent repeat production simulations for 1 μs each and three independent repeat production simulations for 10 μs each were performed starting with different initial configurations (i.e. with the same initial position but with different initial velocities) to study the dynamic interactions of UraA with native lipids. For the pure POPE bilayer only 1 simulation was performed (for 1 μs).
Mutant forms of UraA. Simulations using the aforementioned system with UraA in the 75% POPE, 20% POPG, 5% CL bilayer were also performed using mutated forms of the protein to study the role of specific residues in binding lipids (see [fig_ref] Table 1: Summary of the principal simulations [/fig_ref] for more information). Additionally, removing basic residues from the CL binding sites of UraA facing the outer bilayer leaflet (UraAmut-2-CG), facing the inner bilayer leaflet (UraAmut-3-CG) or a combination of the two (UraAmut-1-CG) allowed us to probe if CL molecules bound cooperatively to the three CL binding sites. All mutations were performed using Modeller [bib_ref] Comparative protein modelling by satisfaction of spatial restraints, Sali [/bib_ref]. In particular, in UraAmut-1-CG residues R4 and R299 in site 1 and K321 in site 3 were mutated to alanines. In UraAmut-2-CG residues K157, R222, K321, K381, and K390 in or close to site 1 were mutated to alanines. These mutations removed the CL binding site in the outer leaflet of the bilayer. In UraAmut-3-CG residues R4, R11, R177, R299, R351, R357 and R362 in or close to site 1 and K62, K109, K110, R265 and K407 in or close to site 2 were mutated to alanines. These mutations removed the CL binding sites in the region of UraA facing the inner leaflet of the bilayer. Finally, simulations with mutations restricted to each of the three CL-binding sites: i) K321A, ii) K109A/R265A and iii) R4A/R299A were also performed. Note that in many cases more than one basic residue in the CL binding site was mutated to alanine in order to change the net positive charge of the CL binding site to neutral. This allowed us to prevent the CL molecules from interacting with other positive residues in the CL binding site, thus compensating for the loss of the main interacting residues. In these simulations, after the exchange of the lipids, the wild-type form of UraA was replaced by the mutated form and 10 individual repeat production simulations of 1 μs each were performed for each mutant.
CG-MD simulations. All CG-MD simulations were performed using GROMACS 4.5 (www.gromacs.org). A Berendsen thermostat (coupling constant of 1.0 ps; reference temperature 310 K for POPC and POPE lipid bilayer and 323 K for the more physiological lipid mixture) and barostat (coupling constant of 1.0 ps; reference pressure 1 bar) were used. The integration time step was 20 fs. Lennard-Jones and Coulombic interactions were shifted to zero between 9 and 12 Å, and 0 and 12 Å, respectively, following the original MARTINI parameterization.
## Atomistic molecular dynamics simulations
Coarse-grained (CG) to atomistic (AT) conversion was performed using the CG2AT protocol as described previously by Stansfeld et al. [bib_ref] From coarse grained to atomistic: a serial multiscale approach to membrane protein..., Stansfeld [/bib_ref]. In this approach the lipids are reconstructed by aligning AT lipid fragments with the corresponding CG fragments from an energy-minimized library of atomistic lipid conformations. The protein backbone is reconstructed by the CG backbone trace using PULCHRA [bib_ref] Fast procedure for reconstruction of full-atom protein models from reduced representations, Rotkiewicz [/bib_ref] and the side chains using Modeller [bib_ref] Comparative protein modelling by satisfaction of spatial restraints, Sali [/bib_ref]. Atomistic molecular dynamics (AT-MD) simulations were performed using the GROMOS96 53a6 force field that has been widely used in simulation studies of membrane proteins. The Parrinello-Rahman barostat [bib_ref] Polymorphic transitions in single crystals: a new molecular dynamics method, Parrinello [/bib_ref] and the V-rescale thermostat [bib_ref] Canonical sampling through velocity rescaling, Bussi [/bib_ref] were used for pressure and temperature coupling, respectively. The LINCS algorithm [bib_ref] LINCS: a linear constraint solver for molecular simulations, Hess [/bib_ref] was used to constrain bond lengths, and the Particle Mesh Ewald (PME) algorithm [bib_ref] Particle mesh Ewald: an Nlog(N) method for Ewald sums in large systems, Darden [/bib_ref] was used to model long-range electrostatic interactions. Lipid parameters for POPE and CL were from [bib_ref] Electroporation of the E. coli and S. Aureus membranes: Molecular dynamics simulations..., Piggot [/bib_ref] and for POPG were from [bib_ref] Lipid models for united-atom molecular dynamics simulations of proteins, Kukol [/bib_ref]. All atomistic systems were equilibrated for 2.5 ns with the protein backbone Cα atoms restrained, followed by unrestrained atomistic MD simulations from 100 to 500 ns with a time step of 2 fs. Note that the removal of uracil is necessitated by the lack of accurate simulation parameters for this ligand.
Analyses. All the analysis was performed using GROMACS [bib_ref] GROMACS 4: algorithms for highly efficient, load-balanced, and scalable molecular simulation, Hess [/bib_ref] [bib_ref] GROMACS: fast, flexible, and free, Van Der [/bib_ref] , VMD [bib_ref] VMD: visual molecular dynamics, Humphrey [/bib_ref] and locallywritten scripts. Structures were visualized using VMD [bib_ref] VMD: visual molecular dynamics, Humphrey [/bib_ref] and PyMol (http://www.pymol.org). Analysis of the lifetime of interactions between the different lipid types and the 3 CL binding sites was performed by calculating distances (d ij ) between each binding site (i) and the head group of each of the 3 different lipid types (j). To define a contact a cut-off distance (δ) of 0.7 nm for POPE and POPG was used. For the CL lipids instead of using the standard 0.8 nm cutoff distance we have evaluated the optimal cutoff distance (δ) over a range of distances from 0.8 to 1.6 nm. This was done to reflect the larger headgroup of the CL molecules and additionally to take into consideration "rattling in a cage" motion of CL molecules. This motion of CL molecules has been reported in other CG-MD studies of these lipids [bib_ref] Evidence for cardiolipin binding sites on the membrane-exposed surface of the Cytochrome..., Arnarez [/bib_ref]. Whilst the calculation of the lifetime of interactions is sensitive to the cutoff distance used, the trend that CL molecules form more prolonged interactions with CL binding site 1 is retained for all cutoff distances used. In particular, the longest lifetime of interactions of the CL molecules binding to CL site 1 for simulation 3 varies from *200 ns to *800 ns when varying δ. Similarly, for CL binding site 2 it varies from *200 ns to *450 ns, and for CL binding site 3 varies from *100 ns to *300 ns. A comparable sensitivity of the residence time has been observed to other studies which used CG-MD simulations to probe the CL interactions with membrane proteins [bib_ref] Evidence for cardiolipin binding sites on the membrane-exposed surface of the Cytochrome..., Arnarez [/bib_ref]. Note that in all Figures/Tables we show the results obtained using a cutoff distance of 1.1 nm to define a contact for CL lipids. An occupancy variable z ij (t) was determined, defined as:
[formula] z ij ðtÞ ¼ 1 if d ij ðtÞ d or z ij ðtÞ ¼ 0 if d ij ðtÞ > d [/formula]
Thus if the distance between the lipid and the CL binding site was lower that the cut-off distance the occupancy was 1 otherwise the occupancy was 0 (see S2A for this analysis of one of the 10 μs simulations of UraA).
Using this measure of occupancy the stretches of time for which a site was continuously occupied by a specific lipid type were determined. For this analysis the trajectory was sampled every 0.3 ns, and we did not perform any smoothing of our raw data to remove any fast binding/unbinding events. This was because we concerned that smoothing of our trajectories would prevent us from capturing the faster binding/unbinding events of POPE and POPG lipids compared to CL, and additionally it has been suggested that application of different smoothing times of a trajectory may yield significantly different results [bib_ref] Evidence for cardiolipin binding sites on the membrane-exposed surface of the Cytochrome..., Arnarez [/bib_ref]. For all lipid types we have observed some fast binding/unbinding events (shown in S2A , but for our analysis we have focused on the most extended continuous periods of interactions for each lipid (S3B . Finally, note that we present all of our results without any normalization in terms of the number of lipids in the bilayer. To define the CL binding sites we have used two residues in each site for which our contact analysis had revealed that were predominantly involved in the interactions with the lipids (site 1: R4 and R299; site 2: K109 and R265; site 3: K321 and W212; see also S2 Video). Note that using a two-residue representation of the CL binding site often allows more than one lipid to occupy the CL binding site. After we have extracted the periods of time for which a binding site was continuously occupied by a specific lipid type, we have used these data to calculate how many changes of single lipids of the same lipid type occurred in the CL binding site during the time of which the site was continuously occupied by a specific lipid type. This analysis is shown and discussed in detail for one of the simulations in S2 Fig. It should be noted, however, that the nature of the CG-MD simulations is such that calculated lifetimes of interactions should be viewed as indicative rather than absolute predictions.
In order to quantify the enrichment of anionic lipids (i.e. POPG and CL) around the protein we have merged the trajectories from the three extended CG-MD simulations, and have then calculated the mean density of each of the lipid types on the bilayer xy plane (S4 . The density was calculated for each leaflet separately. A region with a radius of 3 nm from the protein center of mass (representing the region adjacent to the protein) and a region with a radius of 3.5 to 6.0 nm from the protein center of mass (representing the bulk region of the bilayer) were selected (see S4 . The change in the density of each lipid type between the region adjacent to the protein and the bulk region of the bilayer was calculated by dividing the mean density of the lipids in the annulus close to the protein over the mean density of the lipids in the annulus representing the bulk region:
[formula] S ¼ mean densityð0 À 3 nmÞ mean densityð3:5 À 6 nmÞ [/formula]
Therefore if the value of S is above 1 then it means that there is enrichment of the specific lipid type around the protein whereas if S is < 1 it suggests that the density of the lipids in the bulk region is higher compared to the density of the lipids close to the protein. Note that for the CL molecules, because of their preference for the CL binding sites, we have also calculated S using only the mean density of CL molecules in the CL binding sites.
# Results
## Cardiolipin (cl) binding sites on the uraa transporter
To study the preferential interaction of UraA with the anionic lipids in a membrane, the crystal structure of UraA (3QE7) was assembled in a lipid bilayer that resembles its native environment in the bacterial inner membrane. In particular, UraA in a lipid bilayer containing POPE (75%), POPG (20%) and CL (5%) was subjected to MD simulations at coarse-grained and atomistic resolutions (see Methods and [fig_ref] Table 1: Summary of the principal simulations [/fig_ref] for more information). Note that in all simulations the uracil molecule was absent, as was the detergent molecule present in the crystal structure, producing an empty carrier. A key goal of this study was to explore the conformation of the empty carrier, which is an intermediate in the transport mechanism. Analysis of the lipid densities around the protein during the CG-MD simulations revealed a high density of acidic lipids in the vicinity of the protein [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref]. In particular, we have identified three main preferential sites in UraA that interact with the CL. In all CL binding sites lysine or arginine residues interact with the phosphate groups of the lipid head groups. These sites are shown in [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref] and [fig_ref] Figure 3: CL binding sites [/fig_ref] and were reproducible in the repeat CG-MD simulations performed with the wild-type UraA. One of these sites (site 3) is located in the extracellular side of the protein at residue K321, which forms the main interactions with the phosphate group of CLs (formed mainly by TM helix 11, 12 and the unstructured region between helices 7 and 8). Other residues in this CL binding site that form significant interactions with the CL headgroup are V202 to T204 and I208 to W212 [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref]. The remaining two CL binding sites (sites 1 and 2) are located in the cytosolic side of the protein, which unlike the UraA extracellular part, has many lysine and arginine residues in the lipid-exposed area of the protein creating a net positive surface (S5 . CL binding site 1 is located around residues R4, A5, I6, P13, F176, R177, G178 and A296 to V300 (TM helices 1, 7 and the loop region between helices 10 and 11 that contains R299). CL binding site 2 is located around residues C61, K62, I107 to A111 and R265 (formed by TM helices 2, 4 and 9). Other dynamic interactions of CLs with other parts of the protein were also observed (see S6 for detailed analysis of all the interactions). Calculation of the average number of CLs in contact with the protein during the simulations suggested that 2-3 CLs were in contact most of the time with the protein (S7 . [fig_ref] Table 1: Summary of the principal simulations [/fig_ref]. Three different views are shown: a side view (left) showing the three CL binding sites, a cytosolic view facing the inner leaflet (middle) showing CL sites 1 and 2, and an extracellular view facing the outer leaflet of the bilayer (right) showing CL site 3. The distribution of all lipids separately is shown in . Normalized average number of contacts (using a cut-off distance of 7 Å for POPE and POPG and 8 Å for CL molecules) between the UraA and the head groups of POPG Visualization of the final snapshots from all the individual repeat CG-MD simulations [fig_ref] Figure 3: CL binding sites [/fig_ref] showed that in the vast majority of the simulations one CL lipid was associated with each CL binding site (see also S3 Video). To test if the interactions between the CL molecules and UraA and CL lipids in the bilayer (across all repeats of the extended UraA-CG simulations; see [fig_ref] Table 1: Summary of the principal simulations [/fig_ref]. For the normalization, the number of contacts of a residue with a lipid type was divided by the number of lipids, the number of frames and the ratio of cutoff volumes. C. The number of contacts from one of the CG-MD simulations was mapped on the UraA crystal structure. Blue indicates a low number, white indicates a medium number and red a large number of contacts.
doi:10.1371/journal.pcbi.1004123.g002 are retained on longer timescales, three simulations (to 10 μs) have also been performed (see [fig_ref] Table 1: Summary of the principal simulations [/fig_ref]. Analysis of the interactions (see S5B showed that the CL binding sites were retained over this time period. [fig_ref] Figure 3: CL binding sites [/fig_ref] shows the CL interactions with arginine and/or lysine residues in the CL binding sites identified above. In the CG-MD simulations the interactions of CL molecules with UraA are mediated by interactions of the negatively charged phosphate atom of the CL headgroup with polar part of the protein and hydrophobic interactions of the acyl chains with the non-polar belt of the membrane protein. The most common residues involved in the interaction with the phosphates are arginine and lysine as has been found in other simulations [bib_ref] Multiscale simulations reveal conserved patterns of lipid interactions with aquaporins, Phillip [/bib_ref]. In the CL binding site 1, CL interacts with R299 located in a short loop connecting TM helix 10 and 11 and/or R4 in the mobile N-terminus of the protein near the exit site of the transporter. The CL interaction site in the outer leaflet involves K321 located at the N-terminus of the short helix connecting TM helices 11 and 12 and W212 at the N-terminus of TM helix 7. Hydrophobic interactions also play a role in creating lipid binding grooves in UraA. CL tails interact with the hydrophobic part of the protein located below the arginine and lysine residues identified above. The surface of UraA contains a number of hydrophobic grooves that accommodate the lipid acyl chains. The surface created by TM helices 2, 4, 9 and 11 is quite smooth. There are hydrophobic grooves between TM helices 1 and 6, TM helices 5 and 6, and between TM helices 13 and 14 and the rest of the protein. The simulations (S3 Video) indicate that while the head groups of the lipids are immobilized by interaction with basic residues in the protein, the acyl chains are quite dynamic.
Interestingly, in all individual CG-MD simulations a high density of POPG molecules was also observed around the protein. The POPG lipids interact with the majority of the protein lipid exposed surface however basic residues form the highest number of interactions with POPG [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref]. This creates an acidic annulus around the protein, which may be essential in UraA function by hoarding protons required for transport. Similar annuli of anionic lipids have also been observed around TM helices [bib_ref] Local lipid reorganization by a transmembrane protein domain, Koldsø [/bib_ref] and around the TM domains of receptors e.g. EGFR [bib_ref] Interactions of the EGFR juxtamembrane domain with PIP2-containing lipid bilayers: Insights from..., Halim [/bib_ref]. Simulations studies have also shown that the mobility of lipids in an annulus up to *3 nm from the membrane surface is smaller compared to the bulk region of the bilayer [bib_ref] Membrane proteins diffuse as dynamic complexes with lipids, Niemela [/bib_ref] [bib_ref] Reduced lateral mobility of lipids and proteins in crowded membranes, Goose [/bib_ref]. POPE can bind indiscriminately to also form a complete annulus of lipids around the protein (S8C . All the interacting residues with the CL, POPG and the POPE lipids are shown in [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref] of the mean number of lipids in contact with the protein reveals that 2 -3 CLs and * 8 POPG lipids are in contact with the protein (S7 .
To quantify the enrichment of anionic lipids around the protein we have calculated the ratio of the mean density of each lipid type in each leaflet in an annulus close to the protein and in the bulk region of the bilayer (S; see Methods and S4 . If S > 1 then there is enrichment of the specific lipid type around the protein whereas if S < 1 it indicates that the mean density of the lipids in the bulk region is higher compared to the density of the lipids close to the protein. For POPG S = 1.16 for the outer leaflets and 1.33 for the inner leaflet demonstrating an enrichment of POPG around the protein. For CL, S = 1.14 for the inner leaflet, and 0.59 for the outer leaflet. However if we calculate S using only the density of the CL lipids in the CL binding sites (divided by the mean density in the bulk region) then the local value is S = 5 for the inner leaflet and S = 2.3 for the outer leaflet. This confirms the raised density of CL molecules in the discrete CL binding sites rather than the overall annulus, and thus demonstrates the preferential interaction of CL molecules with the three CL binding sites. For reference, for the zwitterionic lipid POPE S = 0.63 for the outer leaflet and 0.71 for the inner leaflet.
To examine in more detail the association of the different lipid types with the aforementioned CL binding sites we have approximated the lifetime of interactions between the different lipids and the 3 CL binding sites (see Methods, S2 Table, [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref]. This analysis suggests that whilst all 3 lipid types are able to bind to the CL binding sites, the CL molecules form prolonged interactions (see . Therefore, when a CL binding site is not occupied by CL it is likely to be occupied mainly by POPG lipids or POPE lipids to a lesser extent. Additionally, this analysis suggests that all lipid types form more stable interactions with the CL binding site 1, followed by CL binding sites 2 and 3 (S2 [fig_ref] Figure 3: CL binding sites [/fig_ref]. In particular, in the CL binding site 1 CL molecules form the longest interactions, followed by POPG and POPE (S2 [fig_ref] Figure 3: CL binding sites [/fig_ref]. In CL binding sites 2 and 3 the same pattern of interactions is observed however the time for which a lipid type continuously occupies a CL binding site is lower (S2 [fig_ref] Figure 3: CL binding sites [/fig_ref]. Interestingly, when we calculated how many changes of single lipids of the same lipid type occurred in the CL binding site when the site is continuously occupied by a lipid type, we found that in most cases only one CL molecule interacts with the CL binding site whereas in the case of POPG and POPE lipids 3 to 7 changes of single lipids occurred in the CL binding site (S2B [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref]. This analysis suggests a faster exchange of the POPG and POPE lipids between the unbound and unbound state despite the fact that for this analysis we have used only those periods when a single lipid from the same lipid type occupies the CL binding site (see S2 . Note that the stretches of times we have obtained in our study for CL binding sites 2 and 3 are short compared to those in a recent study by Arnarez et al. on the interactions of CL molecules with cytochrome bc 1 [bib_ref] Evidence for cardiolipin binding sites on the membrane-exposed surface of the Cytochrome..., Arnarez [/bib_ref]. The lifetime of interactions for CL binding site 1 are comparable with the residence times at the lower occupancy CL sites of cytochrome bc 1 , but not with the times for those sites with prolonged CL interactions with the cytochrome protein. Besides the fact that we have used a different protein, the lower residence times we have observed may be due to the fact that in our study we have POPE, POPG and CL whereas in the previous study only CL and POPC lipids were present, which may induce competition between the lipids especially between the anionic CL and POPG lipids. However, the CL binding sites identified in the cytochrome bc 1 study contain multiple lysine and arginine residues which is similar to the sites we have identified on UraA. Poyry et al. also studied the cytochrome bc 1 /CL interactions, using atomistic simulations, and also demonstrated that the interactions between CL lipids and cytochrome bc 1 are mainly mediated by the interactions between the lipids and positively charged residues [bib_ref] Atomistic simulations indicate cardiolipin to have an integral role in the structure..., Pöyry [/bib_ref]. Other studies have shown that other anionic lipids interact with specific sites on proteins e.g. PIP 2 lipids to ion channels [bib_ref] Simulation-based prediction of phosphatidylinositol 4,5-bisphosphate binding to an ion channel, Schmidt [/bib_ref].
We have also used the extended 10 μs simulations to examine the effect of the protein on the lateral diffusion of the POPG lipids and the CLs in the membrane (S2C . It is possible that the strong interactions between UraA and the aforementioned acidic lipids would influence the lipid diffusion rates. Indeed, the diffusion of the CL lipids in a bilayer with UraA decreases compared to a bilayer with the same lipid concentration but with no protein component (e.g. the calculated diffusion coefficient for the simulation with UraA in [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref] 2.5 ± 0.1 x10 −7 cm 2 /s in the inner leaflet and 3 ± 0.1 x10 −7 cm 2 /s in the outer leaflet compared to 4.5 ± 0.2 x10 −7 cm 2 /s in the simulations without any UraA present). The diffusion of the POPG molecules is also somewhat slower in the system where the protein is present compared to a bilayer with the same lipid concentration but with no protein component (S2C , possibly due to their interactions with the protein. Interestingly, calculation of the mean diffusion coefficient for the lipids in the system without any protein component (pcpgcl-CG in S1 [fig_ref] doi [/fig_ref] reveals diffusion coefficients of 4.5 ± 0.2 x10 −7 cm 2 /s for CL, 5.9 ± 0.3 x10 −7 cm 2 /s for POPG and 6.2 ± 0.2 x10 −7 cm 2 /s for POPE. This suggests that the four acyl chains of the CL and its larger cross-sectional diameter slow its diffusion about 1.5-fold relative to POPE or POPG. [fig_ref] Table 1: Summary of the principal simulations [/fig_ref]. An occupancy index of 1 is used when the lipids are in contact with the protein and 0 when the lipids are not in contact. See S2 Comparison of the diffusion coefficients of POPE and POPG with those in a CG-MD study with a similar mixture of lipids (but not with the inclusion of CL lipids and with a higher number of proteins in the bilayer) suggests a somewhat lower diffusion coefficient for the POPG and POPE lipids in our system (5.9 ± 0.3 x10 −7 cm 2 /s for POPG in our system compared to 8.5 ± 0.3 x10 −7 cm 2 /s for POPG in Goose et al. [bib_ref] Reduced lateral mobility of lipids and proteins in crowded membranes, Goose [/bib_ref]. Short atomistic simulations by Rog et al. [bib_ref] Molecular dynamics simulations of charged and neutral lipid bilayers: treatment of electrostatic..., Róg [/bib_ref] and by Murzyn et al. [bib_ref] Phosphatidylethanolamine-Phosphatidylglycerol Bilayer as a Model of the Inner Bacterial Membrane, Murzyn [/bib_ref] , again with a POPE/POPG mixture, have suggested short timescale diffusion coefficients for POPG of *3.5 x10 −7 cm 2 /s. Recent microsecond scale simulations of a POPE/POPG mixture showed that the lipid diffusion coefficient is dependent on the NaCl concentration present, with a lipid lateral diffusion coefficient of 3.1 x10 −8 cm 2 /s in the presence of 0.15 M NaCl [bib_ref] Microsecond molecular dynamics simulations of lipid mixing, Hong [/bib_ref].
To test the essential role of the basic residues in UraA in CL binding and to examine whether CL molecules bind cooperatively to the three CL binding sites, three different sets of in silico mutations (mutating basic residues in the CL binding sites of UraA facing the outer bilayer leaflet (UraAmut-2-CG), facing the inner bilayer leaflet (UraAmut-3-CG) or a combination of the two (UraAmut-1-CG)) were performed [fig_ref] Table 1: Summary of the principal simulations [/fig_ref] and the CG-MD simulations were rerun with the mutated UraA in the lipid mixture. All the simulations with the mutated forms of UraA were initiated by replacing the wild-type protein with the mutated form of the protein after the exchange of lipids step as described in the Methods section. . In systems (i) and (ii) in addition to the strong interactions of the CL lipids with the non-mutated CL binding sites, interactions of CL molecules with R222 and K381 in (i) and R11, K62, K64 and R351 in (ii) were also observed. This augments our previous observation that basic residues in the three CL binding sides regulate the CL/UraA interactions and that CL bind to the 3 CL binding sites independently. Note that when both CL binding sides in the cytosolic part of the protein were mutated the association of the POPG lipids with UraA in the inner leaflet was also reduced (S7 . Convergence analysis of the simulations showed that all simulations were converged after 4 or 5 repeat simulations (S5C .
## Atomistic simulations of uraa
To study the possible conformational changes within the protein and the preferential interaction of UraA with the different lipids in the bilayer the final snapshots from four of the ten repeat CG-MD simulations with the WT UraA in the POPE/POPG/CL lipid mixture were converted to an atomistic representation (see Methods for more details). This is a standard protocol used in previous studies where interactions between a protein and lipid molecules were studied. In all the atomistic snapshots CLs were bound in the aforementioned CL binding sites. Atomistic simulations (2 x 100 ns and 2 x 500 ns) were then run. Calculation of the root mean square displacement (RMSD) of the protein during the simulation revealed a sharp increase of the protein RMSD at the beginning of the simulations to *0.35 nm. The RMSD for the longer AT-MD simulations subsequently increased to final values of *0.5 nm (S9A . This increase is due to large fluctuations of the unstructured regions of UraA. Indeed, calculation of the route mean square fluctuation (RMSF) showed that core secondary structure elements [fig_ref] Table 1: Summary of the principal simulations [/fig_ref] have RMSFs between *0.1 and *0.2 nm whereas unstructured regions have RMSFs between *0.4 to *0.6 nm. Alignment of the crystal structure with the final snapshot of the extended simulations demonstrates that the integrity of all 14 transmembrane helices was maintained during the simulation, however the positions and angles of the helices that form the gate domain changed from the crystal structure [fig_ref] Figure 6: Closed state of UraA [/fig_ref]. Calculation of the transport path of the protein at the end of all atomistic simulations, using the program HOLE, revealed that the protein moved from an inward facing conformation that initially bound uracil (and a single detergent) to a closed conformation collapsing the water-filled passage to the cytosol [fig_ref] Figure 6: Closed state of UraA [/fig_ref]. This is presumably due to the fact that the simulations were performed in the absence of uracil. Alignment of the gate and the core domains from the end of the four atomistic simulations with the crystal structure reveals no significant conformational changes within the domains themselves . This suggests that in the absence of the uracil molecule a rigid body movement of the cytosolic part of the gate domain toward the cytosolic part of the core domain shifts the protein conformation to a close state. Simulations of the UraA protein in a simpler POPC or POPE lipid bilayer resulted in the same shift of the protein from the inward facing to a closed conformation (S9 , showing that the conformation change was not dependent upon a specific lipid composition. During the simulation no ion flux is observed. However occasionally we observe a small flux of water molecules.
Closer examination of the interactions between CL and the protein at the atomistic level suggests stable interactions between the CL oxygen atoms and the protein for the complete duration of the simulations [fig_ref] Figure 1: Structure of UraA [/fig_ref]. The contacts made between UraA and CL in the corresponding CG-MD simulations were maintained during the atomistic simulations. In particular CL head groups can be seen to form significant interactions with the R4 and K299 residues in CL binding site 1, K109 and R265 residues in the CL binding site 2 and K321 residues in CL binding site 3 [fig_ref] Figure 3: CL binding sites [/fig_ref]. The CL tails form dynamic interactions with the hydrophobic part of the CL binding sites. Additionally, the annulus of the CL and POPG lipids around the proteins is retained on the timescale of our AT-MD simulations.
# Discussion
## Cardiolipin binding sites in uraa
The major finding of our study is the identification of three CL binding sites in UraA. That these were not seen in the crystal structure is perhaps not surprising given that the purification/crystallization procedures extracts the majority of any bound lipids and that CL is a relatively minor (5%) component of the E. coli inner membrane. Indeed, no phospholipids could be detected in purified samples (*1 mg protein/ml) of His-tagged UraA prepared using the detergent dodecylmaltoside and Ni-affinity chromatography. CL has some unique physiochemical properties [bib_ref] The physicochemical properties of cardiolipin bilayers and cardiolipin-containing lipid membranes, Lewis [/bib_ref]. It consist of two phosphatidic acid linked together by a central glycerol. The free hydroxyl on the glycerol can form a hydrogen bond with the neighboring phosphate oxygens. This results in the formation of an acid-anion and a CL molecule with a single net negative charge with one phosphate having a typical low pKa (*2-3) and the second phosphate having a much higher pKa in the physiological pH range [bib_ref] pH-dissociation characteristics of cardiolipin and its 2 0 -deoxy analogue, Kates [/bib_ref]. CL can thus act as a ready proton acceptor and donor. CL also has a very small head-group compared to the size of its hydrophobic region with four acyl chains. This allows CL molecules to pack closely together to form clusters (e.g. CL "lipid rafts") in mitochondrial and bacterial membranes. The crosssectional area of CL is decreased upon binding of counter-ions like sodium due to reduction in electrostatic repulsion. The CL binding site identified in the extracellular site of UraA is located close to the transport path implying the CL may be important for providing the proton required for transport, as suggested for other proton transporters [bib_ref] Evidence for cardiolipin binding sites on the membrane-exposed surface of the Cytochrome..., Arnarez [/bib_ref] [bib_ref] Specific roles of protein-phospholipid interactions in the yeast cytochrome bc1 complex structure, Lange [/bib_ref]. The inner CL binding sites may also have active role in the proton transport, by buffering protons as they exit the protein, however further investigation is required to identify the exact role of these CL binding sites in transport. Calculation of the binding energy of CL in cytochrome c oxidase by Marrink et al. using extended CG-MD simulations (100 μs) revealed a free energy of association of *-30 kJ/mol for the main CL binding sites [bib_ref] Identification of cardiolipin binding sites on cytochrome c oxidase at the entrance..., Arnarez [/bib_ref]. A quantitative measure of the association of lipids with specific membrane proteins may be given by EPR and fluorescence quenching methods [bib_ref] Electron spin resonance in membrane research: protein-lipid interactions, Marsh [/bib_ref]. Note that no associated lipids were reported in the crystal structure of UraA [bib_ref] Structure and mechanism of the uracil transporter UraA, Lu [/bib_ref]. Other techniques such as crosslinking often require the use of chemically-modified lipids that have different properties from native lipids used in this study. The residues which are involved in the UraA/CL interactions are expected to be exposed to the lipid environment when the protein is in the outward open state and thus it is possible that CLs also associate with UraA when the protein is in the outward open conformation. However structural data for UraA in the outward open state are needed to be able to study this computationally.
UraA is a member of the nucelobase/ascorbate transporter (NAT) family, of which there are 10 family members in E. coli. We have shown that certain arginine and lysine residues in UraA are required for CL binding. How conserved are these residues in the NAT family in E. coli? Alignment of all 10 members of the NAT family in E. coli suggests that an arginine residue at the UraA R4 position is conserved in 4 out 10 members of the family. However with the exception of YbbY all other members of the family have a conserved basic residue in the vicinity of R4 (6 residues away; S11 . Inspection of UraA suggests that this conserved positive residue is also located in the CL binding site 1 described in this study. Additionally, all the members of the NAT family found in E. coli have a basic residue in the same position as R299 or in the same loop 1-3 residues distally (S11 . Interestingly, residue R299 is found at the end of the NAT-signature motif. This is a motif which is conserved in the NAT/NC2 family of proteins
[formula] ([Q/E/P-N-x-G-x-x-x-x-[RKG] [/formula]
) and it was shown to be important in the function of UapA [bib_ref] The Nucleobase-ascorbate Transporter (NAT) signature motif in UapA defines the function of..., Koukaki [/bib_ref]. In particular R299 corresponds to R417 on UapA, the fungal uric acid/xanthine permease from Aspergillus, and mutation of this residue affect the kinetics and specificity of UapA for uric acid [bib_ref] The Nucleobase-ascorbate Transporter (NAT) signature motif in UapA defines the function of..., Koukaki [/bib_ref]. Additionally, cysteine scanning mutagenesis of the xanthine permease XanQ (YgfO) residue G333, which again corresponds to the UraA R299 residue (S11 , suggests that it plays a crucial role to the specificity of XanQ for different substrates [bib_ref] Cysteine-scanning analysis of the nucleobase-ascorbate transporter signature motif in YgfO permease of..., Karatza [/bib_ref]. Furthermore, replacement of certain basic residues located within TM segments in XanQ (K164, K249, R341, R385) with cysteine resulted in functional transporters but with significant decreases in xanthine uptake [bib_ref] Role of intramembrane polar residues in the YgfO xanthine permease: HIS-31 and..., Karena [/bib_ref]. Mutation of basic residues (R285, R286, R296) in the cytosolic loop connected TM 8 and 9 showed modest decreases in transport [bib_ref] Cysteine-scanning analysis of helices TM8, TM9a, and TM9b and intervening loops in..., Mermelekas [/bib_ref]. The K298C mutant showed *50% diminished initial xanthine uptake. K298 is located at an equivalent position one residue proximal to R265 in UraA at the beginning of TM9, while R285/286 in XanQ are lysine residues in UraA.
In CL binding site 2 a basic residue in the same position as R265 is present with the exception of UapA, XanQ, XanP and PurP (S11 . Note, however, that XanQ, XanP and PurP proteins have an adjacent positive residue to the R265 position whereas UapA has a basic amino acid 3 residues distal from R265. K109 is poorly conserved in the E. coli NAT/NCS2 family, however basic residues are located in the same loop 3-5 residues distally in all family members. Finally, K321 is conserved only in some family members (RutG, UacT) but not others. CL is not found in eukaryotic plasma membranes and thus key basic residues responsible for CL interaction in bacteria are not expected to be conserved in these NAT family members.
CL plays an essential role in the assembly and functioning of mitochondrial and bacterial membranes [bib_ref] Thematic review series: Glycerolipids. Cardiolipin synthesis for the assembly of bacterial and..., Schlame [/bib_ref]. Yeast mutants deficient in CL biosynthesis have impaired mitochondrial function [bib_ref] Absence of cardiolipin in the crd1 null mutant results in decreased mitochondrial..., Jiang [/bib_ref]. The ADP/ATP translocator is associated with CL clusters in the inner mitochondrial membrane along with respiratory sub-complexes [bib_ref] Cardiolipin defines the interactome of the major ADP/ATP carrier protein of the..., Claypool [/bib_ref] and this transporter is dependent upon CL for optimal activity [bib_ref] The reconstituted ADP/ATP carrier activity has an absolute requirement for cardiolipin as..., Hoffmann [/bib_ref]. The mitochondrial enzyme creatine kinase induces clustering of CL [bib_ref] Cardiolipin clusters and membrane domain formation induced by mitochondrial proteins, Epand [/bib_ref] and in turn, CL promotes the association of ATP synthase into ribbon-like structures in the mitochondrial membrane [bib_ref] Cardiolipin affects the supramolecular organization of ATP synthase in mitochondria, Acehan [/bib_ref].
CL exists in clusters in bacterial membranes in regions of high curvature such as in "minicells" [bib_ref] Escherichia coli minicell membranes are enriched in cardiolipin, Koppelman [/bib_ref] and at the cell poles [bib_ref] Cardiolipin microdomains localize to negatively curved regions of Escherichia coli membranes, Renner [/bib_ref] [bib_ref] A Curvature-mediated mechanism for localization of lipids to bacterial poles, Huang [/bib_ref]. ProP, a H + -proline symporter, is localized to CL clusters at the cell poles of E. coli [bib_ref] Cardiolipin and the osmotic stress responses of bacteria, Romantsov [/bib_ref]. Phospholipids interact with membrane proteins during their insertion into the membrane acting as chaperones to assist in their folding [bib_ref] Molecular genetic and biochemical approaches for defining lipid-dependent membrane protein folding, Dowhan [/bib_ref]. Indeed CL interacts tightly with the bacterial SecYEG translocon where it creates a high-affinity binding site for the SecA ATPase and localizes the complex to the cell poles [bib_ref] The action of cardiolipin on the bacterial translocon, Gold [/bib_ref]. This suggests that CL may be involved in membrane protein folding and insertion into the bilayer during biosynthesis. CL may also act as a "proton trap" through its acid-anion structure and is commonly associated with proteins involved in oxidative phosphorylation, associating with sub-complexes and providing a source of protons for the ATP synthase [bib_ref] Cardiolipin: a proton trap for oxidative phosphorylation, Haines [/bib_ref]. CL is found associated with sub-complexes of the respiratory chain such as in the bacterial photoreaction center [bib_ref] Structural details of an interaction between cardiolipin and an integral membrane protein, Mcauley [/bib_ref] [bib_ref] Is there a conserved interaction between cardiolipin and the type II bacterial..., Wakeham [/bib_ref] and the yeast cytochrome bc1 complex [bib_ref] Lipids in membrane protein structures, Palsdottir [/bib_ref] [bib_ref] Crystal structure of the yeast cytochrome bc1 complex with its bound substrate..., Lange [/bib_ref]. The association of CL with proton-driven symporters like UraA may point to a similar role by providing a source of buffered protons in the vicinity of these transporters. As in the case of ProP, CL may also provide a mechanism to cluster and localize proton-driven transporters to the cell poles in bacteria. Proton may be channeled directly between CL and the proton-binding site in the center of the transport protein via a proton highway composed in part of acidic residues. For example, CL is located close to the exit site for protons in cytochrome oxidase [bib_ref] Cardiolipin: a proton trap for oxidative phosphorylation, Haines [/bib_ref]. A similar role of CL has been also identified recently for the cytochrome c oxidases [bib_ref] Identification of cardiolipin binding sites on cytochrome c oxidase at the entrance..., Arnarez [/bib_ref].
## A closed state for uraa
The second major finding of this study is the observation of a closed state of UraA, resulting from removal of the substrate uracil and a stabilizing non-native detergent molecule. Membrane proteins in general can tolerate changes in membrane lipid composition, which occur commonly in micro-organism as they adapt to changes in their environment [bib_ref] Tolerance to changes in membrane lipid composition as a selected trait of..., Sanders [/bib_ref]. Any changes in conformation that occurred during the simulation may be due to lipid interactions or loss of the substrate, or a combination of both factors. We found a similar change in the UraA structure in three different lipid environments (POPC, POPE and the lipid mixture), suggesting that lipids did not greatly influence the protein motions in moving from an empty to a closed state conformation. This suggests that the CL binding is not required for the transition from the inward facing to the closed state.
There is a water-filled passage leading from the bound uracil to the cytosol, partly occluded by the bound detergent. The protein can respond in two ways: 1, by filling in the space with water or 2, by changing conformation to close the space. The final atomistic structure of UraA in a bilayer is in a closed state conformation. This change occurred regardless of the lipid composition and thus is due to performing the simulations with the empty carrier. Thus, the MD simulation produced a closed state of the carrier, i.e. a state that may not be readily accessible by crystallography or other structural methods and that is an intermediate in the transport mechanism.
## Membrane proteins in lipid bilayers
Molecular dynamics simulations provide a powerful tool to analyze the structure and dynamics of membrane proteins in lipid bilayers of defined composition [bib_ref] Membrane proteins: molecular dynamics simulations, Lindahl [/bib_ref]. A molecular dynamics simulation of LacY, in various lipids identified specific interactions between the lipid head-groups and sites on the protein [bib_ref] Identification of specific lipid-binding sites in integral membrane proteins, Lensink [/bib_ref]. Recent studies also identified cardiolipin (CL) binding sites in the cytochrome bc 1 transporter [bib_ref] Evidence for cardiolipin binding sites on the membrane-exposed surface of the Cytochrome..., Arnarez [/bib_ref] and in the cytochrome c oxidase [bib_ref] Identification of cardiolipin binding sites on cytochrome c oxidase at the entrance..., Arnarez [/bib_ref] using CG-MD simulations. Similar studies identified specific interactions between cholesterol [bib_ref] Identification of cholesterol binding sites in the Serotonin1A Receptor, Sengupta [/bib_ref] or PIP lipids [bib_ref] Simulation-based prediction of phosphatidylinositol 4,5-bisphosphate binding to an ion channel, Schmidt [/bib_ref] [bib_ref] PIP2-binding site in Kir channels: definition by multiscale biomolecular simulations, Stansfeld [/bib_ref] with integral membrane proteins. The inner membrane of gram-negative bacteria, like E. coli, is composed of approximately 75% PE, 20% PG and 5% CL under normal laboratory growth conditions at 37°C [bib_ref] Effect of shift-down and growth inhibition on phospholipid metabolism of escherichia coli, Ballesta [/bib_ref]. In this simulation we found an annulus of anionic lipids around the UraA protein, followed by concentric rings spaced *5 Å apart defined by the cross-sectional diameter of the lipid. Rings are discernable out to 20 Å from the protein surface. This suggests that the protein restricts the mobility of the annular lipids, a property of membrane proteins noted for some time [bib_ref] Biological membranes: the importance of molecular detail, Lee [/bib_ref] [bib_ref] Specificity of intramembrane protein-lipid interactions, Contreras [/bib_ref]. The flexible acyl chains can fill in gaps in the protein's hydrophobic surface and the head groups can mediate more specific interactions with the polar parts of the protein [bib_ref] How lipids and proteins interact in a membrane: a molecular approach, Lee [/bib_ref] [bib_ref] Principles of membrane protein interactions with annular lipids deduced from aquaporin-0 2D..., Hite [/bib_ref]. Molecular simulations indicate that the association of annular lipids, with aquaporin for example can be quite dynamic [bib_ref] Multiscale simulations reveal conserved patterns of lipid interactions with aquaporins, Phillip [/bib_ref]. Tightly-bound lipids have been identified in a number of X-ray crystal structures of membrane proteins showing that there are specific binding sites for lipids on the surface of some membrane proteins, which may assist in their folding or functioning [bib_ref] Specific protein-lipid interactions in membrane proteins, Hunte [/bib_ref]. These interactions involve tight binding of the acyl chains to the hydrophobic surface of the protein as well as polar/ionic interactions involving the head groups. The different simulations run for UraA indicate that CL is preferentially associated with the protein, although this lipid was only present at a 5% content. Indeed, positively-charged residues on hydrophilic loops of the protein can engage anionic lipids in a dynamic fashion, the lipids moving on and off the protein over the overall *40 μs time course of the simulation. PE although in excess does not interact in a stable manner with UraA, likely due to its zwitterionic nature. The dynamic nature of the lipid-protein interactions observed with UraA may allow the protein to adapt to different lipid environments that the organisms like E. coli might encounter [bib_ref] Tolerance to changes in membrane lipid composition as a selected trait of..., Sanders [/bib_ref].
# Limitations
It is important to consider possible limitations of simulations discussed in this study. The use of a CG-MD based approach implies some approximations in the protein and in the lipids. It has been shown, however, that CG-MD simulations correctly predict interactions of CL [bib_ref] Evidence for cardiolipin binding sites on the membrane-exposed surface of the Cytochrome..., Arnarez [/bib_ref] and of other lipids (e.g. PIP 2 ) with integral membrane proteins [bib_ref] Simulation-based prediction of phosphatidylinositol 4,5-bisphosphate binding to an ion channel, Schmidt [/bib_ref] [bib_ref] PIP2-binding site in Kir channels: definition by multiscale biomolecular simulations, Stansfeld [/bib_ref] [bib_ref] Multiscale simulations reveal conserved patterns of lipid interactions with aquaporins, Phillip [/bib_ref] [bib_ref] Molecular driving forces defining lipid positions around aquaporin-0, Aponte-Santamaría [/bib_ref] [bib_ref] Lipid exchange mechanism of the cholesteryl ester transfer protein clarified by atomistic..., Koivuniemi [/bib_ref] [bib_ref] Identification of cholesterol binding sites in the Serotonin1A Receptor, Sengupta [/bib_ref]. Conversion of the final snapshot of four individual CG-MD simulation to a fully atomistic representation and extended AT-MD simulation runs confirm the stability of the identified CL binding sites (S10 and gives more detail information about the positioning of the CL lipid tails. The use of an elastic network model within the CG-MD simulations restricts any possible conformational changes of the protein. The use of a serial multiscale approach, with subsequent atomistic simulations, in part addresses this limitation allowing us to observe a closed state of the protein.
Despite seen some conformational changes within the protein more extended simulations and/ or exchange sampling techniques (e.g., dynamic importance sampling (DIMS) [bib_ref] Computing ensembles of transitions from stable states: Dynamic importance sampling, Perilla [/bib_ref] or selfguided Langevin dynamics [bib_ref] Self-guided Langevin dynamics simulation method, Wu [/bib_ref] are likely to be needed to more fully address possible conformational changes (e.g. outward facing conformation). Another limitation which arises due to the intrinsic approximations of the CG representation is that quantities such as the durations of protein/lipid interactions or diffusion coefficients of lipids are likely to be influenced by the coarse-graining of the protein and to be sensitive to the exact protocol used. Such analyses, however, provide a semi-quantitative measure of the stability of protein-lipid interactions observed in the CG-MD simulations. The removal of the uracil from the binding site produced a novel closed state of UraA, which may form part of the transport cycle. In future studies it will be important both to include the uracil molecule and possibly extend the time frames to capture the binding and movement of substrate across the membrane and the various conformational state of the transport protein in real time. It will also of course be important to test e.g. the predictions concerning CL binding sites and key residues via experimental mutagenesis studies. [fig_ref] Table 1: Summary of the principal simulations [/fig_ref]. Yellow is used when the lipids are in contact with the protein and black when the lipids are not in contact. B. Total number of changes of single lipids associated with CL binding site 1 during the longest stretches of continuous interactions between the different lipid types and CL binding site 1 (sim2 in S2 [fig_ref] doi [/fig_ref]. For this analysis we have used only those periods of time during which the CL binding sites were occupied by a single lipid of the same type. For all lipids there were frames during the simulations for which the binding site was occupied by more than one lipid of the same type. These frames were excluded from our calculation. Thus we consider a change from a single lipid to another lipid only when the CL site is occupied by a single lipid and this lipid is replaced by a single lipid of the same type. S2B [fig_ref] Figure 2: Interactions of UraA with lipids [/fig_ref] the total number of the aforementioned exchanges during the longest stretches of continuous interactions between the different lipid types and CL binding sites. Overall for sim2, for which we show the results, *75% of the time a single CL lipid occupied the site and *25% of the time more than one CL lipids associated with the CL binding site 1 when CL site 1 was occupied by CL lipids (using a 1.1 nm cutoff distance to define contact; see Methods). CL site 2 was occupied *85% by a single CL lipid and CL site 3 was occupied *90% by a single CL lipid. The same analysis showed that when the CL sites were occupied by POPG lipids, CL site 1 was occupied *67% of the time by a single POPG lipid, CL site 2 was occupied *70% of the time by a single POPG lipid and CL site 3 was occupied by a single POPG lipids *65% of the time. For the rest of the time more than one POPG lipids associated with the CL binding sites. When the sites were occupied by POPE lipids CL site 1 was occupied *50% of the time by a single POPE lipid, CL site 2 was occupied *65% of the time by a single POPE lipid and CL site 3 was occupied *60% of the time by a single POPE lipid. The interaction pattern is broadly similar for the other two simulations. During the longest stretches of continuous interactions the percentage of the single CL lipids which occupy the CL sites are either slightly higher or the same as the overall percentage (see above). In contrast in the case of POPG or POPE lipids the percentage of single lipids which occupy the CL sites during the longest stretches of continuous interactions is significantly lower compared to the overall percentages and in some cases the CL sites are occupied for up to *80% of the time with more than one POPG or POPE lipids. Therefore using only the frames which a single lipid occupies the CL binding sites is a rather good approximation for calculating how many different CL lipids continuously occupy a CL binding site but for the POPG and POPE lipids we use a rather small subset of the interaction between the aforementioned lipids and the CL binding sites. We have decided, however, to use the same protocol for these lipids since this provides a rough measure of the faster exchange of these lipids compared to the CL molecules. C. Mean square displacement (MSD) relative to the time for the POPE, POPG and the CLs for the pcpgcl-CG (A) and UraA-CG (B) systems (see [fig_ref] Table 1: Summary of the principal simulations [/fig_ref] and S1 [fig_ref] doi [/fig_ref]. Blue indicates a positive surface and red a negative surface. The electrostatic calculation was performed using APBS [bib_ref] Electrostatics of nanosystems: Application to microtubules and the ribosome, Baker [/bib_ref] in PyMol. Note that the basic cluster on cytosolic side of UraA is consistent with the "inside-positive" rule [bib_ref] Membrane protein structure prediction: Hydrophobicity analysis and the positiveinside rule, Von Heijne [/bib_ref]. B,C. Convergence analysis. Occupancy of the CL molecules (green) in the simulations with the WT and a mutated form of the protein is shown by using 4, 8 and 10 repeat simulations. The systems used were chosen randomly. This analysis suggests that the simulation system converges after 4 or 5 repeat simulations. [fig_ref] Table 1: Summary of the principal simulations [/fig_ref]. On average, UraA associates with *2 or 3 CLs, *8 POPGs and *12 POPEs in the simulations with the wild type protein. The number of CLs associated with the protein reduced to 1 or 2 in the simulations with the mutated forms (TIF) [fig_ref] Table 1: Summary of the principal simulations [/fig_ref] and S1 [fig_ref] doi [/fig_ref]. The RMSDs for the repeat simulations for each system are shown in different colors. D,E,F. Pore lining surfaces along the transport axis of the structures at the end of all the atomistic simulations (UraA-AT in D, UraA-pe-AT in E and UraA-pc-AT in F). The dotted line shows the uracil binding site as shown by Lu et al. [bib_ref] Structure and mechanism of the uracil transporter UraA, Lu [/bib_ref] crystal structure. The profile was calculated using the program HOLE. The profile for each of the repeat simulations is shown in different colors. In all cases in the region between 0 to -5 nm along the transport axis the pore is closed (cytosolic side of the protein). After that region there is some variation in the profiles due to the present of more dynamic regions of the protein (e.g. dynamic loops of UraA and the end of the TM helices). The pore lining surface of the crystal structure is shown in black. G,H. [fig_ref] Table 1: Summary of the principal simulations [/fig_ref]. Yellow is used when the lipids are in contact with the protein and black when the lipids are not in contact. CL exhibit a prolonged interaction with specific residues in UraA (yellow). The positively charged residues identified in the CG-MD simulations to coordinate the CL binding to the three CL binding sites are also shown (see [fig_ref] Figure 3: CL binding sites [/fig_ref] in the main text). B. Distance between the residue 350 of the UraA crystal and the same residue of UraA in the atomistic simulations relative to time. To calculate the timescale of the changes in the position of the gate domain during the AT-MD simulation we have fitted the core domain of UraA in the extended atomistic simulations to the core domain of UraA crystal structure. Subsequently we have calculated the distance of residue 350 (Cα atom) of the crystal structure to the same residue in the simulation snapshots for all simulation frames. This residue was selected because is located on the cytosolic part of helix 12 (at the end of the helical region and close to the transportation path) and its movement would give a representative timescale for the movement of the cytosolic part of the gate domain relative to the crystal structure. This distance is shown schematically in the inset picture at the bottom right of the plot. The other inset picture shows the aforementioned distance for the first 5 ns of the two extended atomistic simulations. Note that the distance at the beginning of the atomistic simulations is not 0 because an initial movement of the gate domain relative to the core domain was observed during the CG-MD simulations. C. Number of H-bonds between the different lipids and UraA for one of the extended AT-MD simulations. (TIF) S11 Fig. Sequence alignment of UraA from E. coli and UapA from Aspergillus nidulans with the other 9 E. coli members of the NAT family. A. The basic residues on the CL binding sites are highlighted as boxes. The alignment was made using Clustal program in Jalview. The sequences are colored based on the sequence conservation using the ClystalX color scheme. (TIF) S1 Summary of all other simulations. (DOCX) S2 Lifetime of the interactions between the lipids and the 3 CL binding sites. (DOCX) S1 Video. Self-assembly of POPC lipids around the UraA transporter. In this 100 ns CG-MD simulation the protein Cα atoms were restrained but the side chains were free to rotate and move. The POPC lipids are shown in bond format and the protein in VDW format. The lipid bilayer assembled around the protein very rapidly (after the first *15 ns of the simulation), confirming that UraA contains a hydrophobic belt occupied by detergent in the crystal structure. (MP4) S2 Video. Occupancy analysis for the cardiolipin (CL; green) around UraA. The occupancy was calculated as the average occupancy over all repeat simulations of the wild type UraA (UraA-CG in [fig_ref] Table 1: Summary of the principal simulations [/fig_ref]. The UraA backbone particles are shown in bond format. (MP4) S3 Video. Lipid binding to CL binding site 1. MD simulation showing the diffusion of two single CL molecules (green) and a POPG molecule (red) in a lipid bilayer and their binding to the CL site 1 on UraA (shown using the residues R4 and R299). Note that in the video UraA is centered in the bilayer for clarity and a smoothing of the trajectory every 10 frames was performed. In the simulation the protein is free to diffuse. (MP4)
## Supporting information
[fig] Figure 1: Structure of UraA. A. The core and the gate domains as suggested by Lu et al. [/fig]
[fig] Figure 2: Interactions of UraA with lipids. A. Occupancy plots showing the probability of occurrence of cardiolipin (CL; green) and of the POPG (red) lipids around UraA. The occupancy was calculated as the average over all repeat coarse-grained simulations of the wild type UraA (UraA-CG in [/fig]
[fig] Figure 3: CL binding sites. Atomistic simulations of UraA in a 75% POPE/ 20% POPG/ 5% CL bilayer. Snapshots from the specific interactions between the CL and UraA are shown for the three CL binding sites (shown in green). The protein is represented as grey ribbons. The positively charged residues that form the main interactions with the CL in each CL binding site are shown in magenta. CL is shown in VDW format. Note that for thisFigure weshow snapshots with only CL lipids bound in the CL binding sites. As detailed in S4Fig. and S6 Fig. andthe in the text, POPG and POPE molecules to a lesser extent can also associate with the CL binding sites. doi:10.1371/journal.pcbi.1004123.g003 [/fig]
[fig] Figure 4: Interactions between the lipids and CL binding site 1. A. Histograms of binding event durations (defined as the time for which CL binding site 1 was continuously occupied by a lipid; see Methods for details). B. Interactions of UraA R299 residue with the POPG and CL molecules as a function of the simulation time for one of the CG simulations with the wild type UraA (10 μs simulation UraA-CG in [/fig]
[fig] Figure 5: Mutations on the CL binding sites. Occupancy plots for the location of cardiolipins (CL; green) around UraA. The occupancy was calculated as the average over the multiple coarse-grained simulations of the wild type and the mutated forms of UraA (Mut 1: UraAmut-1-CG, Mut 2: UraAmut-2-CG and Mut 3: UraAmut-3-CG; see [/fig]
[fig] Figure 6: Closed state of UraA. A. Alignment of the core domain of the final snapshot of one of the atomistic simulations (UraA-AT) with the core domain of the crystal structure. B. Ribbon structure of UraA showing the location of the bound uracil (green) and NG detergent (orange). C,D. Pore lining surfaces of the proteins aligned in A. Red is used when the pore radius is lower than 1.15 Å, green when the pore radius is between 1.15 and 2.3 Å and blue when the pore radius is larger than 2.3 Å. Note the loss of water cavity (blue) from the inward-facing passage during the simulation to produce a closed conformational state of UraA. doi:10.1371/journal.pcbi.1004123.g006 [/fig]
[fig] S1: Fig. Flowchart of the simulations performed for this study. Schematic representation of the inputs (white background boxes) and outputs (grey background boxes) of the simulations performed. See Methods for further information. (TIF) S2 Fig. Lipid mobility and interactions with UraA. A. Interactions of UraA with the POPE, POPG and CL molecules as a function of the simulation time for one of the CG simulations with the wild type UraA (10 μs simulation; UraA-CG in [/fig]
[fig] S8: Fig. Lipid occupancy plots around UraA. Occupancy plots showing the relative probability of occurrence of cardiolipin (CL; green), of the POPG (red) and of the POPE (blue) lipids around UraA. The occupancy was calculated as the average over all repeat coarse-grained simulations of the wild type UraA (UraA-CG in Table 1). Three different views are shown: a side view (left) showing the three CL binding sites, a cytosolic view facing the inner leaflet (middle) showing CL sites 1 and 2, and an extracellular view facing the outer leaflet of the bilayer (right) showing CL site 3. (TIF) S9 Fig. Root mean square deviation (RMSD) and the closed state of UraA. A,B,C. Root mean square deviation (RMSD) of UraA as a function of the simulation time for the atomistic systems (UraA-AT in PE, PG, CL, UraA-pc-AT in PC and UraA-pe-AT in PE in [/fig]
[table] Table 1: Summary of the principal simulations. [/table]
|
Development of Environmentally Friendly Atom Transfer Radical Polymerization
Atom transfer radical polymerization (ATRP) is one of the most successful techniques for the preparation of well-defined polymers with controllable molecular weights, narrow molecular weight distributions, specific macromolecular architectures, and precisely designed functionalities. ATRP usually involves transition-metal complex as catalyst. As the most commonly used copper complex catalyst is usually biologically toxic and environmentally unsafe, considerable interest has been focused on iron complex, enzyme, and metal-free catalysts owing to their low toxicity, inexpensive cost, commercial availability and environmental friendliness. This review aims to provide a comprehensive understanding of iron catalyst used in normal, reverse, AGET, ICAR, GAMA, and SARA ATRP, enzyme as well as metal-free catalyst mediated ATRP in the point of view of catalytic activity, initiation efficiency, and polymerization controllability. The principle of ATRP and the development of iron ligand are briefly discussed. The recent development of enzyme-mediated ATRP, the latest research progress on metal-free ATRP, and the application of metal-free ATRP in interdisciplinary areas are highlighted in sections. The prospects and challenges of these three ATRP techniques are also described in the review.
# Introduction
Since it was discovered independently by Matyjaszewski and by Sawamoto in 1995, atom transfer radical polymerization (ATRP) has become one of the most powerful tools for the preparation of well-defined polymers with controlled molecular weights, narrow molecular weight distributions, and designable molecular architectures. Typically, the ATRP system is composed of monomer, initiator and catalyst. The most important component is the catalyst, which is a key factor to the 'living'/controlled characteristics of polymerization. Up to now, a large variety of transition metal complexes-such as Cu, Ru, Mo, Rh, Ni, Pd, Co coordination compounds-have been successfully employed as catalysts in ATRP systems .
The environment-friendly aspects of ATRP have attracted a lot of interest in recent years. In an ideal situation, a good ATRP catalyst could realize: (1) synthesizing polymer with desired molecular weight and narrow polydispersity; (2) high catalytic activity for the polymerization reaction; (3) a little amount of residual catalyst in the polymer. To the best of our knowledge, the fulfillment of these requirements still remains a challenge. Since the ATRP of methyl methacrylate and styrene with iron complex as catalyst was first reported in 1997, iron coordination compounds have been one of the most promising environmentally friendly catalysts because of their low toxicity, inexpensive cost, and abundant commercial availability. Besides iron complex catalyst, enzyme-mediated ATRP has also attracted considerable interest due to its high efficiency and selectivity, mild reaction conditions, and excellent biocompatibility. Recently, metal-free ATRP has emerged as a green and sustainable strategy for precise polymer synthesis.
Matyjaszewski and coworkers have summarized iron-catalyzed ATRP recently on the basis of the structures and properties of iron ligands, the effects of ligands on the polymerization rate, and the development of initiating systems for activator regeneration. Xie et al. also reviewed iron-catalyzed ATRP based on mechanistic considerations and the types of iron complex and iron ligand. Hawker and de Alaniz presented a review on metal-free ATRP and discussed its catalysis principle, catalyst structure, monomer scope, and application in the synthesis of architecturally complex materials. As enzymes could be denatured and deactivated by heavy metal catalyst, there are only a few reports on enzyme-mediated ATRP.
Herein, this review aims to provide a comprehensive understanding of these three types of environmentally friendly ATRP technology and primarily focused on the catalytic activity, initiation efficiency, polymerization controllability, and environmental friendliness. The latest research progress on iron complex-catalyzed ATRP of a large variety of monomers, the recent developments of enzyme-mediated ATRP, and the application of metal-free catalyst mediated ATRP in interdisciplinary areas are highlighted in sections. The prospects and challenges of these three ATRP techniques are also described in the review.
## Applications of iron catalyst in various atrp systems
## Applications of iron catalyst in normal atrp
Transition metal complex and organic halide are generally used as catalyst and initiator respectively in early ATRP system. The earliest catalyst in a normal ATRP was a copper complex (CuCl/Bipyridine), which had excellent catalytic activity and good control for the polymerization of multiple monomers. Compared to copper complex, iron complex has broader application prospects because of its environmentally friendly nature. The mechanism of normal ATRP catalyzed by iron complex is shown in.
Polymers 2020, 12, x FOR PEER REVIEW 2 of 33 selectivity, mild reaction conditions, and excellent biocompatibility. Recently, metal-free ATRP has emerged as a green and sustainable strategy for precise polymer synthesis. Matyjaszewski and coworkers have summarized iron-catalyzed ATRP recently on the basis of the structures and properties of iron ligands, the effects of ligands on the polymerization rate, and the development of initiating systems for activator regeneration. Xie et al. also reviewed iron-catalyzed ATRP based on mechanistic considerations and the types of iron complex and iron ligand. Hawker and de Alaniz presented a review on metal-free ATRP and discussed its catalysis principle, catalyst structure, monomer scope, and application in the synthesis of architecturally complex materials. As enzymes could be denatured and deactivated by heavy metal catalyst, there are only a few reports on enzyme-mediated ATRP.
Herein, this review aims to provide a comprehensive understanding of these three types of environmentally friendly ATRP technology and primarily focused on the catalytic activity, initiation efficiency, polymerization controllability, and environmental friendliness. The latest research progress on iron complex-catalyzed ATRP of a large variety of monomers, the recent developments of enzyme-mediated ATRP, and the application of metal-free catalyst mediated ATRP in interdisciplinary areas are highlighted in sections. The prospects and challenges of these three ATRP techniques are also described in the review.
## Applications of iron catalyst in various atrp systems
## Applications of iron catalyst in normal atrp
Transition metal complex and organic halide are generally used as catalyst and initiator respectively in early ATRP system. The earliest catalyst in a normal ATRP was a copper complex (CuCl/Bipyridine), which had excellent catalytic activity and good control for the polymerization of multiple monomers. Compared to copper complex, iron complex has broader application prospects because of its environmentally friendly nature. The mechanism of normal ATRP catalyzed by iron complex is shown in. Matyjaszewski group first reported in 1997 that iron complex-catalyzed normal ATRP of styrene (St) and methyl methacrylate (MMA) with (1-bromoethyl)benzene (PEBr) as an initiator. They investigated the effect of different ligands such as tri-n-butylphosphine, tri-n-butylamine, and triphenylphosphine on the polymerization and found that the polymerization catalyzed by FeBr2/tri-n-butylamine had well-controlled characteristics for styrene and methyl methacrylate, but not for acrylate. Sawamoto et al. also reported an iron complex-catalyzed normal ATRP of MMA in 1997 using iron(II) bis(triphenylphosphine)dichloride [FeCl2(PPh3)2] as catalyst. The number average molecular weight of obtained poly(methyl methacrylate) (PMMA) increased linearly with the increase of monomer conversion, and the molecular weight distribution was narrow (Mw/Mn = 1.1-1.3). Later, iron complexes with halide anions as ligands were found to catalyze controlled polymerizations of acrylates by Matyjaszewski and coworkers in 2000. The residual catalysts Matyjaszewski group first reported in 1997 that iron complex-catalyzed normal ATRP of styrene (St) and methyl methacrylate (MMA) with (1-bromoethyl)benzene (PEBr) as an initiator. They investigated the effect of different ligands such as tri-n-butylphosphine, tri-n-butylamine, and triphenylphosphine on the polymerization and found that the polymerization catalyzed by FeBr 2 /tri-n-butylamine had well-controlled characteristics for styrene and methyl methacrylate, but not for acrylate. Sawamoto et al. also reported an iron complex-catalyzed normal ATRP of MMA in 1997 using iron(II) bis(triphenylphosphine)dichloride [FeCl 2 (PPh 3 ) 2 ] as catalyst. The number average molecular weight of obtained poly(methyl methacrylate) (PMMA) increased linearly with the increase of monomer conversion, and the molecular weight distribution was narrow (M w /M n = 1.1-1.3). Later, iron complexes with halide anions as ligands were found to catalyze controlled polymerizations of acrylates by Matyjaszewski and coworkers in 2000. The residual catalysts could be readily removed from the polymer products because of their ionic nature. However, this polymerization system was not applicable for ATRP of styrene probably due to the involvement of cationic polymerization.
Sawamoto et al. used iron complex [Fe(Cp)I(CO) 2 ; Cp = cyclopentadienyl] as catalyst in normal ATRP of acrylates, finding that the synthesized polymers had controlled molecular weights and narrow molecular weight distributions (M w /M n < 1.2). This catalytic system was also applicable for the synthesis of poly(methyl acrylate)-b-poly(styrene) and poly(butyl acrylate)-b-poly(styrene) block copolymers, but the controllability of the polymerization was decreased when [(CH 3 ) 2 C(CO 2 CH 3 )CH 2 C(CH 3 )(CO 2 CH 3 )Br] or [(CH 3 ) 2 C(CO 2 CH 3 )CH 2 C(CH 3 )(CO 2 CH 3 )Cl] was used as an initiator. Sawamoto and coworkers later found that the iron bromide complexed with butylphosphine had an excellent activity/controllability in the normal ATRP of MMA. The polymerization reached over 90% conversion in 5 h, producing PMMA with narrow molecular weight distributions (M w /M n = 1.20-1.32). Yan and coworkers discovered that the polymerization of styrene was controlled when using low toxic organic acid as ligand in iron mediated ATRP, but the obtained polystyrene (PS) had a relatively broad molecular weight distribution (M w /M n = 1.50).
Gibson et al. demonstrated that a four-coordinated iron(II) complex bearing α-diimine ligands had high catalytic activity in normal ATRP of styrene. A five-coordinated iron(II) complex containing tridentate nitrogen donor ligands was also found very effective for styrene. They investigated the effects of different structures of catalyst on the reaction rate of ATRP and revealed that α-diimine ligands bearing electron-donating groups increased the polymerization rate but the ligands bearing electron-withdrawing substituents decreased the polymerization rate. Zhang and coworkers reported that iron(II) ligated with N-(n-hexyl)-2-pyridylmethanimine (NHPMI) was an effective catalyst for ATRP of MMA. The molar ratio of iron/NHPMI had relatively large impacts on the controllability of the polymerization. The molecular weight distribution of produced PMMA ranged from 1.25 to 2.50. Wang et al. proved that iron(II) coordinated with tris(3,6-dioxaheptyl)amine (TDA) was an excellent catalyst in normal ATRP of styrene with (1-chloroethyl)benzene (PECl) or PEBr as an initiator. The residual catalyst could be easily removed out of the polymer product because of the good water solubility of TDA.
Xue et al. reported that using 2-[(diphenylphosphino)-methyl]pyridine (DPPMP) as ligand in an iron mediated normal ATRP of MMA had well-controlled characteristics. The molecular weight of PMMA increased linearly with the increase of monomer conversion. They investigated the effect of different solvents (e.g., p-xylene, toluene, and anisole) on the polymerization, and found that the controllability was decreased when using p-xylene or anisole as a solvent and only PMMA prepared in toluene had low polydispersities (M w /M n = 1.1-1.3). The system was not applicable for controlled polymerizations of methyl acrylate (MA) and butyl acrylate (BA). Recently, Gao et al. claimed that the iron(II) complexed with anilidoimine ligands showed excellent catalytic performances in ATRP of MMA. Nagashima and coworkers described that a trinuclear iron(II) complex with 1,4,7-trimethyl-1,4,7-triazacyclononane (Me 3 TACN) as ligand was an effective catalyst for normal ATRP of styrene. The residual catalyst in the polystyrene could be removed by a simple washing process. Nagashima et al. later also reported that the [(i-Pr)TACN]FeX 2 complex (TACN = N,N,N-substituted-1,4,9-triazanonane) had a high catalytic activity in ATRP of styrene and MMA. They further found that the iron complex [{(cyclopentyl)TACN}FeBr 2 ] n could catalyze well-controlled polymerizations of styrene, BA, and MMA. The PS-b-PMMA block polymer had been synthesized using this catalyst at a low catalyst concentration of 59 ppm.
## Applications of iron catalyst in reverse atrp
Normal ATRP usually presents some limitations such as requirement of relatively large amount of catalyst and instability of the lower oxidation state of transition metal complex. The use of oxidatively stable catalyst in ATRP is a favorable approach to solve the problem. Reverse ATRP catalyzed by air-stable higher oxidation state metal complex has been successfully applied to prepare well-designed polymers. The mechanism of iron mediated reverse ATRP is shown in. The iron mediated reverse ATRP of MMA with triphenylphosphine as ligand and azodiisobutyronitrile (AIBN) as initiator was first reported by Teyssié and coworkers in 1998. From then on, a series of iron complex mediated reverse ATRP of vinyl monomers had been investigated. Qiu et al. prepared well-defined PMMA with high molecular weight (Mn = 171,800 Da) and narrow molecular weight distribution (Mw/Mn = 1.13) by a reverse ATRP with 1,1,2,2-tetraphenyl-1,2-ethanediol (TPED)/FeCl3/PPh3 as catalyst. They later investigated the polymerization of MMA by using diethyl 2,3-dicyano-2,3-diphenylsuccinate (DCDPS)/FeCl3/PPh3 as a catalyst. The polymerization could be well-controlled even at high monomer conversions.
Zhu and coworkers conducted a reverse ATRP of MMA using iron(III)/pyromellitic acid as a catalyst and AIBN as an initiator. However, they found that this catalytic system was only effective for methacrylates but not for acrylates. Zhu et al. later reported an iron mediated reverse ATRP of MMA using 2,2-azobis(2-methylpropionamidine) dihydrochloride (V-50) as initiator and N,N-butyldithiocarbamate ferrum (Fe(S2CN(C4H9)2)3) as catalyst, but the monomer conversion and polymerization rate in this system were relatively low. Ferro and coworkers showed that the Fe(BOX)Cl3 (BOX = 1,1-bis(4,4-dimethyl-1,3-oxazolin-2-yl)ethane) was an excellent catalyst for the reverse ATRP of styrene. Ferro et al. investigated the reverse ATRP of styrene initiated by TPED and catalyzed by three iron(III) complexes-namely Fe III coordinated with BOX, 3,5-dimethyl-bispyrazolylmethane as well as 2,2′-dipyridyl-and found that only TPED/FeCl3/BOX produced polystyrene with controlled molecular weights and narrow molecular weight distributions.
Shaver and coworkers revealed that reverse ATRP of MMA and styrene could be achieved using α-diimine iron complexes as catalysts and AIBN as initiator. Based on the advantages of their iron complex, a series of monomers such as MA, stearyl methacrylate (SMA), n-hexadecyl methacrylate (HMA), 2-hydroxyethyl methacrylate (HEMA), acrylonitrile (AN), n-docosyl acrylate (DA), and methacrylonitrile (MAN) had been polymerized by reverse ATRP.
## Applications of iron catalyst in initiators for continuous activator regeneration (icar) atrp
A relatively large amount of catalyst is required in normal ATRP and reverse ATRP system. Consequently, a relatively high concentration of catalyst residue is inevitably left in the polymer and brings difficulty in the purification of polymer product after polymerization. In order to overcome this limitation, the initiators for continuous activator regeneration atom transfer radical polymerization (ICAR ATRP) had been developed. Iron catalyst has been successfully applied to ICAR ATRP. The mechanism of iron mediated ICAR ATRP is shown in. The iron mediated reverse ATRP of MMA with triphenylphosphine as ligand and azodiisobutyronitrile (AIBN) as initiator was first reported by Teyssié and coworkers in 1998. From then on, a series of iron complex mediated reverse ATRP of vinyl monomers had been investigated. Qiu et al. prepared well-defined PMMA with high molecular weight (M n = 171,800 Da) and narrow molecular weight distribution (M w /M n = 1.13) by a reverse ATRP with 1,1,2,2-tetraphenyl-1,2-ethanediol (TPED)/FeCl 3 /PPh 3 as catalyst. They later investigated the polymerization of MMA by using diethyl 2,3-dicyano-2,3-diphenylsuccinate (DCDPS)/FeCl 3 /PPh 3 as a catalyst. The polymerization could be well-controlled even at high monomer conversions.
Zhu and coworkers conducted a reverse ATRP of MMA using iron(III)/pyromellitic acid as a catalyst and AIBN as an initiator. However, they found that this catalytic system was only effective for methacrylates but not for acrylates. Zhu et al. later reported an iron mediated reverse ATRP of MMA using 2,2-azobis(2-methylpropionamidine) dihydrochloride (V-50) as initiator and N,N-butyldithiocarbamate ferrum (Fe(S 2 CN(C 4 H 9 ) 2 ) 3 ) as catalyst, but the monomer conversion and polymerization rate in this system were relatively low. Ferro and coworkers showed that the Fe(BOX)Cl 3 (BOX = 1,1-bis(4,4-dimethyl-1,3-oxazolin-2-yl)ethane) was an excellent catalyst for the reverse ATRP of styrene. Ferro et al. investigated the reverse ATRP of styrene initiated by TPED and catalyzed by three iron(III) complexes-namely Fe III coordinated with BOX, 3,5-dimethyl-bispyrazolylmethane as well as 2,2 -dipyridyl-and found that only TPED/FeCl 3 /BOX produced polystyrene with controlled molecular weights and narrow molecular weight distributions.
Shaver and coworkers revealed that reverse ATRP of MMA and styrene could be achieved using α-diimine iron complexes as catalysts and AIBN as initiator. Based on the advantages of their iron complex, a series of monomers such as MA, stearyl methacrylate (SMA), n-hexadecyl methacrylate (HMA), 2-hydroxyethyl methacrylate (HEMA), acrylonitrile (AN), n-docosyl acrylate (DA), and methacrylonitrile (MAN) had been polymerized by reverse ATRP.
## Applications of iron catalyst in initiators for continuous activator regeneration (icar) atrp
A relatively large amount of catalyst is required in normal ATRP and reverse ATRP system. Consequently, a relatively high concentration of catalyst residue is inevitably left in the polymer and brings difficulty in the purification of polymer product after polymerization. In order to overcome this limitation, the initiators for continuous activator regeneration atom transfer radical polymerization (ICAR ATRP) had been developed. Iron catalyst has been successfully applied to ICAR ATRP. The mechanism of iron mediated ICAR ATRP is shown in. In an iron-catalyzed ICAR ATRP system, radicals are generated by a conventional radical initiator such as AIBN. The polymerization possesses highly controllable characteristics even at very low concentration of catalyst. Zhu and coworkers firstly reported an iron mediated ICAR ATRP of MMA and styrene in 2010. They investigated the effects of different polymerization conditions on the ICAR ATRP, finding that the ICAR ATRP of styrene could be achieved even if the concentration of iron(III) catalyst was as low as 50 ppm. However, this system showed insufficient catalytical activity for the polymerization of MMA because of the intrinsic low activity of iron catalyst for polar monomers. Zhu et al. later demonstrated a controlled ICAR ATRP of MMA using iron(III) complexed with PPh3 as a catalyst, bifunctional 1,4-(2-bromo-2-methylpropionato)benzene (BMPB2) as an initiator, and AIBN as a thermal radical initiator. This catalytic system showed excellent activities and promoted the polymerization of MMA even at a very low catalyst concentration of 30 ppm.
Wang and coworkers performed an ICAR ATRP of MMA using ethyl 2-bromoisobutyrate (EBiB) as an initiator, AIBN as a thermal radical initiator, and FeCl3·6H2O/succinic acid as a catalyst. The polymerization could produce PMMA with narrow molecular weight distribution (Mw/Mn = 1.20-1.50) at low catalyst concentrations (30-100 ppm). However, a relatively slower polymerization rate was observed in this system, and the polymerization controllability was found to decrease with the increase of AIBN, indicating the presence of coupling side reactions of macromolecular radicals in the polymerization.
Matyjaszewski et al. reported an iron(III) complex mediated ICAR ATRP of MMA using AIBN as thermal initiator at a catalyst concentration of 100 ppm. Later, an iron(III) complex mediated ICAR ATRP of styrene using 1,1-azobis(cyclohexanecarbonitrile) (ACHN) as thermal initiator had also been reported by his group. The obtained polystyrene had a narrow molecular weight distribution with a polydispersity index (PDI, PDI = Mw/Mn) of 1.29. The polymerization rate was found to be largely depended on the amount of ACHN. Matyjaszewski and coworkers investigated ICAR ATRP of MMA and styrene catalyzed by iron-based N-heterocyclic carbene (FeX3(NHC)) complexes, finding that both FeX3(IDipp) (IDipp = 1,3-bis(2,6-diisopropyl-phenyl)-imidazol-2-ylidene) and FeX3(HIDipp) (HIDipp = 1,3-bis(2,6-diisopropyl-phenyl)imidazolidin-2-ylidene) had excellent catalytical activities and produced PMMA and PS with controlled molecular weights and narrow molecular weight distributions (Mw/Mn = 1.15-1.40).
## Applications of iron catalyst in activators generated by electron transfer (aget) atrp
Activators generated by electron transfer (AGET) ATRP was first developed by Matyjaszewski. Compared to normal ATRP, AGET ATRP has a better chance for industrial applications owing to its lower requirement of catalyst. In a typical AGET ATRP, alkyl halide and transition metal complex in its oxidatively stable state are used as initiator and catalyst respectively. The activators are produced by in situ reduction of the oxidatively stable metal complex with tin(II) 2-ethylhexanoate [Sn(EH)2], ascorbic acid (VC), or other reducing agents. Iron In an iron-catalyzed ICAR ATRP system, radicals are generated by a conventional radical initiator such as AIBN. The polymerization possesses highly controllable characteristics even at very low concentration of catalyst. Zhu and coworkers firstly reported an iron mediated ICAR ATRP of MMA and styrene in 2010. They investigated the effects of different polymerization conditions on the ICAR ATRP, finding that the ICAR ATRP of styrene could be achieved even if the concentration of iron(III) catalyst was as low as 50 ppm. However, this system showed insufficient catalytical activity for the polymerization of MMA because of the intrinsic low activity of iron catalyst for polar monomers. Zhu et al. later demonstrated a controlled ICAR ATRP of MMA using iron(III) complexed with PPh 3 as a catalyst, bifunctional 1,4-(2-bromo-2-methylpropionato)benzene (BMPB 2 ) as an initiator, and AIBN as a thermal radical initiator. This catalytic system showed excellent activities and promoted the polymerization of MMA even at a very low catalyst concentration of 30 ppm.
Wang and coworkers performed an ICAR ATRP of MMA using ethyl 2-bromoisobutyrate (EBiB) as an initiator, AIBN as a thermal radical initiator, and FeCl 3 ·6H 2 O/succinic acid as a catalyst. The polymerization could produce PMMA with narrow molecular weight distribution (M w /M n = 1.20-1.50) at low catalyst concentrations (30-100 ppm). However, a relatively slower polymerization rate was observed in this system, and the polymerization controllability was found to decrease with the increase of AIBN, indicating the presence of coupling side reactions of macromolecular radicals in the polymerization.
Matyjaszewski et al. reported an iron(III) complex mediated ICAR ATRP of MMA using AIBN as thermal initiator at a catalyst concentration of 100 ppm. Later, an iron(III) complex mediated ICAR ATRP of styrene using 1,1-azobis(cyclohexanecarbonitrile) (ACHN) as thermal initiator had also been reported by his group. The obtained polystyrene had a narrow molecular weight distribution with a polydispersity index (PDI, PDI = M w /M n ) of 1.29. The polymerization rate was found to be largely depended on the amount of ACHN. Matyjaszewski and coworkers investigated ICAR ATRP of MMA and styrene catalyzed by iron-based N-heterocyclic carbene (FeX 3 (NHC)) complexes, finding that both FeX 3 (IDipp) (IDipp = 1,3-bis(2,6-diisopropyl-phenyl)-imidazol-2-ylidene) and FeX 3 (HIDipp) (HIDipp = 1,3-bis(2,6-diisopropyl-phenyl)imidazolidin-2-ylidene) had excellent catalytical activities and produced PMMA and PS with controlled molecular weights and narrow molecular weight distributions (M w /M n = 1.15-1.40).
## Applications of iron catalyst in activators generated by electron transfer (aget) atrp
Activators generated by electron transfer (AGET) ATRP was first developed by Matyjaszewski. Compared to normal ATRP, AGET ATRP has a better chance for industrial applications owing to its lower requirement of catalyst. In a typical AGET ATRP, alkyl halide and transition metal complex in its oxidatively stable state are used as initiator and catalyst respectively. The activators are produced by in situ reduction of the oxidatively stable metal complex with tin(II) 2-ethylhexanoate [Sn(EH) 2 ], ascorbic acid (VC), or other reducing agents. Iron catalysts are highly promising in AGET ATRP system because of their environmental friendliness, naturally abundant features, and biocompatibilities. A descriptive mechanism of AGET ATRP catalyzed by iron complex is shown in.
Polymers 2020, 12, x FOR PEER REVIEW 6 of 33 catalysts are highly promising in AGET ATRP system because of their environmental friendliness, naturally abundant features, and biocompatibilities. A descriptive mechanism of AGET ATRP catalyzed by iron complex is shown in. The AGET ATRP of MMA using FeCl3 complexed with iminodiacetic acid (IDA) as catalyst and VC as reducing agent had been reported by Zhang and his colleagues in 2008. The polymerization was well-controlled even in the presence of a limited amount of air, producing PMMA with narrow molecular weight distributions (Mw/Mn = 1.31-1.44). Zhang et al. used FeCl3/PPh3 complex as a catalyst and VC as a reducing agent in AGET ATRP of MMA. However, the above mentioned two catalytic systems showed relatively low catalytic performance due to low initiator efficiency. Zhang and coworkers later reported an iron(III) mediated AGET ATRP of styrene using TDA as a ligand and 1,3,5-(2′-bromo-2′-methylpropionato)benzene (BMPB) as an initiator.
The iron-based AGET ATRP of styrene derivatives including 4-methylstyrene (MS), 4-acetoxystyrene (AS), and 4-tert-butylstyrene (tBS) have been explored by Sen and his coworkers using PEBr as an initiator, FeBr3/tributylamine as a catalyst, and Sn(EH)2 as a reducing agent. Well-defined PS-b-PMMA block copolymer had been prepared using the catalytic system. The monomer conversion was revealed to be determined by the activity of the reducing agent and the yield would be dramatically decreased if a weak reducing agent was used in the polymerization. As the requirement of a large amount of reducing agent actually limits the industrial application of iron(III) mediated AGET ATRP, it is of great importance to decrease the amount of reducing agent and find low-cost and commercially available reducing agent. Xue and coworkers recently found that the iron(III) mediated ATRP of MMA could be successfully achieved using trimethylphosphite (TMP) and tributylphosphine (TBP) as ligands even in the absence of reducing agent.
Zhu and his colleagues reported an iron mediated AGET ATRP of styrene in the presence of Fe(OH)3 using commercially available tetra-n-butylphosphonium bromide (TBPBr) or tetrabutylammonium bromide (TBABr) as the ligands. They claimed that the polymerization could occur at the catalyst concentration as low as ppm level. However, as the inorganic base Fe(OH)3 was practically insoluble in most organic solvents, it was difficult to calculate the exact amount of the iron catalyst participated in the reaction. Yan and coworkers used 1-butyl-3-methyl imidazolium hydroxide as an additive to enhance the iron mediated AGET ATRP of MMA. Compared to Fe(OH)3, 1-butyl-3-methyl imidazolium hydroxide was readily soluble in MMA and other organic solvents. Therefore, it was feasible to quantify the amount of 1-butyl-3-methyl imidazolium hydroxide. They investigated the effect of the molar ratio of 1-butyl-3-methyl imidazolium hydroxide to iron complex on the polymerization, finding that the polymerization was controlled only in appropriate ratios of 10.
Cellesi and coworkers discovered that the iron(III) complexed with commercial porphyrin ligand was an excellent catalyst for the AGET ATRP of poly(ethylene glycol) methyl ether methacrylate (PEGMA). Recently, Bai et al. found that iron(III) ligated with The AGET ATRP of MMA using FeCl 3 complexed with iminodiacetic acid (IDA) as catalyst and VC as reducing agent had been reported by Zhang and his colleagues in 2008. The polymerization was well-controlled even in the presence of a limited amount of air, producing PMMA with narrow molecular weight distributions (M w /M n = 1.31-1.44). Zhang et al. used FeCl 3 /PPh 3 complex as a catalyst and VC as a reducing agent in AGET ATRP of MMA. However, the above mentioned two catalytic systems showed relatively low catalytic performance due to low initiator efficiency. Zhang and coworkers later reported an iron(III) mediated AGET ATRP of styrene using TDA as a ligand and 1,3,5-(2 -bromo-2 -methylpropionato)benzene (BMPB) as an initiator.
The iron-based AGET ATRP of styrene derivatives including 4-methylstyrene (MS), 4-acetoxystyrene (AS), and 4-tert-butylstyrene (tBS) have been explored by Sen and his coworkers using PEBr as an initiator, FeBr 3 /tributylamine as a catalyst, and Sn(EH) 2 as a reducing agent. Well-defined PS-b-PMMA block copolymer had been prepared using the catalytic system. The monomer conversion was revealed to be determined by the activity of the reducing agent and the yield would be dramatically decreased if a weak reducing agent was used in the polymerization. As the requirement of a large amount of reducing agent actually limits the industrial application of iron(III) mediated AGET ATRP, it is of great importance to decrease the amount of reducing agent and find low-cost and commercially available reducing agent. Xue and coworkers recently found that the iron(III) mediated ATRP of MMA could be successfully achieved using trimethylphosphite (TMP) and tributylphosphine (TBP) as ligands even in the absence of reducing agent.
Zhu and his colleagues reported an iron mediated AGET ATRP of styrene in the presence of Fe(OH) 3 using commercially available tetra-n-butylphosphonium bromide (TBPBr) or tetrabutylammonium bromide (TBABr) as the ligands. They claimed that the polymerization could occur at the catalyst concentration as low as ppm level. However, as the inorganic base Fe(OH) 3 was practically insoluble in most organic solvents, it was difficult to calculate the exact amount of the iron catalyst participated in the reaction. Yan and coworkers used 1-butyl-3-methyl imidazolium hydroxide as an additive to enhance the iron mediated AGET ATRP of MMA. Compared to Fe(OH) 3 , 1-butyl-3-methyl imidazolium hydroxide was readily soluble in MMA and other organic solvents. Therefore, it was feasible to quantify the amount of 1-butyl-3-methyl imidazolium hydroxide. They investigated the effect of the molar ratio of 1-butyl-3-methyl imidazolium hydroxide to iron complex on the polymerization, finding that the polymerization was controlled only in appropriate ratios of 10.
Cellesi and coworkers discovered that the iron(III) complexed with commercial porphyrin ligand was an excellent catalyst for the AGET ATRP of poly(ethylene glycol) methyl ether methacrylate (PEGMA). Recently, Bai et al. found that iron(III) ligated with 1-butyl-3-methylimidazolium hexafluorophosphate (BMIMPF 6 ) was an excellent catalyst for the AGET ATRP of MMA. Fe(0) wire was used as a reducing agent in this polymerization and it could be recycled and reused.
## Applications of iron catalyst in generation of activators by monomer addition (gama) atrp
The GAMA ATRP is usually conducted without the use of conventional free radical initiators or reducing agent. The Fe(II) complex is generated by the reaction of Fe(III) complex and monomer due to the oxidizing power of FeX 3 . The mechanism of GAMA ATRP catalyzed by iron complex is shown in.
Polymers 2020, 12, x FOR PEER REVIEW 7 of 33 1-butyl-3-methylimidazolium hexafluorophosphate (BMIMPF6) was an excellent catalyst for the AGET ATRP of MMA. Fe(0) wire was used as a reducing agent in this polymerization and it could be recycled and reused.
## Applications of iron catalyst in generation of activators by monomer addition (gama) atrp
The GAMA ATRP is usually conducted without the use of conventional free radical initiators or reducing agent. The Fe(II) complex is generated by the reaction of Fe(III) complex and monomer due to the oxidizing power of FeX3. The mechanism of GAMA ATRP catalyzed by iron complex is shown in. Noh and coworkers reported an iron(III) mediated ATRP of MMA using phosphorus as ligand in the absence of a conventional free radical initiator or reducing agent. The effects of iron complex and ligand on the polymerization have been investigated, finding that the FeBr3/DPPP/EBiB [DPPP = 2-(diphenylphosphino)pyridine] system showed the highest controllability and produced well-defined PMMA (Mn = 1.75 × 10 4 Da, PDI = 1.18). In addition, controlled polymerizations of butyl methacrylate, methyl acrylate, and styrene were also achieved by using FeBr3/DPPP as a catalyst. For comparison, a normal ATRP of MMA using FeBr2/phosphorus as the catalyst was also investigated. However, the molecular weights of produced PMMA were higher than theoretical values, and a relatively higher molecular weight distribution (PDI = 1.37) was obtained. These results indicated that the FeBr3/phosphorus showed a better controllability than the FeBr2/phosphorus for the polymerization of MMA. Later, a series of Fe(III)/phosphorus mediated GAMA ATRP have been reported by their group, and a large number of well-defined polymers and block copolymers have been synthesized.
Kamigaito and coworkers conducted a FeCl3/TnBP [TnBP = tri(n-butyl)phosphine] mediated polymerization of styrene without the use of conventional radical initiator and reducing agent. The polymerization was well-controlled and yielded poly(styrene) (PS) with a low polydispersity index (PDI = 1.19). Moreover, this polymerization was also successfully applied to the copolymerization of styrene with other monomers including MA, MMA, and BA. Later, the FeX3/nitrogen ligand (X = Cl, Br) catalyzed polymerizations of styrene, MMA and MA in the absence of conventional radical initiator and reducing agent were also achieved in their group. Noh and coworkers reported an iron(III) mediated ATRP of MMA using phosphorus as ligand in the absence of a conventional free radical initiator or reducing agent. The effects of iron complex and ligand on the polymerization have been investigated, finding that the FeBr 3 /DPPP/EBiB [DPPP = 2-(diphenylphosphino)pyridine] system showed the highest controllability and produced well-defined PMMA (M n = 1.75 × 10 4 Da, PDI = 1.18). In addition, controlled polymerizations of butyl methacrylate, methyl acrylate, and styrene were also achieved by using FeBr 3 /DPPP as a catalyst. For comparison, a normal ATRP of MMA using FeBr 2 /phosphorus as the catalyst was also investigated. However, the molecular weights of produced PMMA were higher than theoretical values, and a relatively higher molecular weight distribution (PDI = 1.37) was obtained. These results indicated that the FeBr 3 /phosphorus showed a better controllability than the FeBr 2 /phosphorus for the polymerization of MMA. Later, a series of Fe(III)/phosphorus mediated GAMA ATRP have been reported by their group, and a large number of well-defined polymers and block copolymers have been synthesized.
Kamigaito and coworkers conducted a FeCl 3 /TnBP [TnBP = tri(n-butyl)phosphine] mediated polymerization of styrene without the use of conventional radical initiator and reducing agent. The polymerization was well-controlled and yielded poly(styrene) (PS) with a low polydispersity index (PDI = 1.19). Moreover, this polymerization was also successfully applied to the copolymerization of styrene with other monomers including MA, MMA, and BA. Later, the FeX 3 /nitrogen ligand (X = Cl, Br) catalyzed polymerizations of styrene, MMA and MA in the absence of conventional radical initiator and reducing agent were also achieved in their group.. The phosphines could directly reduce Fe(III) to Fe(II) and could also act as a ligand coordinated with iron to form efficient ATRP catalyst.
Recently, Xue et al. reported a FeBr 3 /TPP (TPP = triphenylphosphine) catalyzed polymerization of MMA in the absence of conventional free radical initiators or reducing agent. The system could also be applied to the polymerization of butyl methacrylate (BMA) and styrene (St). The mechanism of the polymerization has been investigated, and it was found that the polymerization was initiated by Ph 3 PBr 2 or Ph 3 PBr 4 generated from the reaction between TPP and FeBr 3 .
## Applications of iron catalyst in supplemental activator and reducing agent (sara) atrp
Fe(0) is a kind of reducing agents in iron mediated AGET ATRP. It can also act as a supplemental activator and react with ATRP initiator to induce a polymerization. This system is termed as supplemental activator and reducing agent ATRP (SARA ATRP). The mechanism of SARA ATRP is as shown in.. The phosphines could directly reduce Fe(III) to Fe(II) and could also act as a ligand coordinated with iron to form efficient ATRP catalyst.
Recently, Xue et al. reported a FeBr3/TPP (TPP = triphenylphosphine) catalyzed polymerization of MMA in the absence of conventional free radical initiators or reducing agent. The system could also be applied to the polymerization of butyl methacrylate (BMA) and styrene (St). The mechanism of the polymerization has been investigated, and it was found that the polymerization was initiated by Ph3PBr2 or Ph3PBr4 generated from the reaction between TPP and FeBr3.
## Applications of iron catalyst in supplemental activator and reducing agent (sara) atrp
Fe(0) is a kind of reducing agents in iron mediated AGET ATRP. It can also act as a supplemental activator and react with ATRP initiator to induce a polymerization. This system is termed as supplemental activator and reducing agent ATRP (SARA ATRP). The mechanism of SARA ATRP is as shown in. Coelho and coworkers reported a Fe(0)/Cu(II) based SARA ATRP of 2-(dimethylamino)ethyl methacrylate (DMAEMA). The molecular weights of obtained poly(DMAEMA) (PDMAEMA) increased linearly with the increase of monomer conversions and the molecular weight distribution was narrow (PDI = 1.13). Moreover, the polymerization of DMAEMA was also realized by using bromo-telechelic mPEG (mPEG-Br) or cholesteryl-2-bromoisobutyrate (CHO-Br) as a macroinitiator, producing corresponding PEG-b-PDMAEMA and CHO-b-PDMAEMA block copolymers. They later carried out the Fe(0)/Cu(II) based SARA ATRP of MA and glycidyl methacrylate (GMA), and prepared PMA and poly(GMA) with low polydispersity index (PDI = 1.08 and 1.27 respectively).
## Developments of iron ligands in atrp
The ligand is a key factor in iron mediated ATRP. The redox potential of the metal core is controlled by the ligand around the center iron ions. Therefore, the polymerization rate and the controllability are practically dependent on the ligands. A large variety of phosphorous, nitrogen, and oxygen compounds have been used as iron ligands in ATRP, and these ligands are primarily classified as nitrogen-based ligand, phosphorous ligand, organic acid-based ligand, and onium salt-based ligand. Coelho and coworkers reported a Fe(0)/Cu(II) based SARA ATRP of 2-(dimethylamino)ethyl methacrylate (DMAEMA). The molecular weights of obtained poly(DMAEMA) (PDMAEMA) increased linearly with the increase of monomer conversions and the molecular weight distribution was narrow (PDI = 1.13). Moreover, the polymerization of DMAEMA was also realized by using bromo-telechelic mPEG (mPEG-Br) or cholesteryl-2-bromoisobutyrate (CHO-Br) as a macroinitiator, producing corresponding PEG-b-PDMAEMA and CHO-b-PDMAEMA block copolymers. They later carried out the Fe(0)/Cu(II) based SARA ATRP of MA and glycidyl methacrylate (GMA), and prepared PMA and poly(GMA) with low polydispersity index (PDI = 1.08 and 1.27 respectively).
## Developments of iron ligands in atrp
The ligand is a key factor in iron mediated ATRP. The redox potential of the metal core is controlled by the ligand around the center iron ions. Therefore, the polymerization rate and the controllability are practically dependent on the ligands. A large variety of phosphorous, nitrogen, and oxygen compounds have been used as iron ligands in ATRP, and these ligands are primarily classified as nitrogen-based ligand, phosphorous ligand, organic acid-based ligand, and onium salt-based ligand.
## Nitrogen-based ligands
As nitrogen ligand has shown excellent performances in copper mediated ATRP, it has also been largely used in iron mediated polymerizations. The first nitrogen ligand used in iron complex catalyzed ATRP was tri-n-butylamine (TnBA), reported by Matyjaszewski in 1997. The FeBr 2 /TnBA complex showed very high activity and catalyzed well-controlled polymerization of styrene and methyl methacrylate. Other typical nitrogen ligands include dibutylamine (DnBA), n-butylamine (nBA), TDA, tetramethylethylenediamine (TMEDA), pentamethyldiethylenetriamine (PMDETA), N,N,N ,N",N"-penta(methyl acylate)diethylene-triamine (MA 5 -DETA), 2,2 -bipyridine (bpy), 4,4 -dinonyl-2,2 -dipyridyl (dNbpy), tris(2-dimethylaminoethyl)amine (Me 6 TREN), NHPMI, and 1,3-bis-(dimethylamino)propane (TPDA).
Sawamoto and coworkers reported an iron mediated ATRP of MMA with a chiral compound, (R)-N,N-dimethyl-1-(2-(diphenylphosphino)phenyl)-ethanamine, as ligand. The polymerization could reach 92% conversion in 23 h. Matyjaszewski group showed an iron mediated polymerization of vinyl acetate (VAc) using PMDETA as ligand. They found that the iron complex in this system actually acted as a redox initiator but not a catalyst. Interestingly, the synthesized poly(vinyl acetate) (PVAc) was an excellent macroinitiator for ATRP of styrene and BA. They later reported a Fe(0) mediated ATRP of MA with Me 6 TREN as ligand. The polymerization was well controlled but only reached a low conversion of 16% due to the formation of inefficient deactivators.
Zhang and coworkers presented an iron mediated ATRP of MMA using MA 5 -DETA as a ligand.. Except for the nitrogen-based ligands mentioned above, the tridentate diiminopyridine (DOIEP) and diaminopyridine (DOAEP) were also used in iron mediated ATRP.
## Phosphorous ligands
Phosphorous ligands exhibit excellent performances in iron mediated ATRP. The first phosphorous ligand used in iron mediated ATRP was TPP, reported by Sawamoto in 1997. A large number of phosphorus ligands such as TnBP, TMP, triethyl phosphite (TEP), TPP, tricyclohexyl phosphine (TCHP), tris(4-methoxyphenyl)phosphine (TMPP), trichlorophenyl phosphine (TCPP), tris(2,4,6-trimethoxyphenyl)phosphine (TTMPP), DPPP, DPPMP, 2-(diphenylphosphino) benzaldehyde (DPPB), diphenyl-(2-methoxyphenyl)phosphine (DPMPP), N-(2-diphenyl-phosphinobenzylidene)aniline (DPPBA), N,N-dimethyl-(2-(diphenylphosphino)phenyl)-methanamine (DMDPM), (R)-N,N-dimethyl-1-(2-(diphenylphosphino)-phenyl)ethanamine (DMDPE), N,N-dimethyl-(2-(dicyclohexylphosphino)-phenyl)methanamine (DMDCM), bis(diphenylphosphino)-methane (BDPPM), and 1,2-bis(diphenylphosphino)ethane (BDPPE) have been developed for iron mediated ATRP.
Sawamoto group reported an iron mediated ATRP of PEGMA using TPP as a ligand and prepared corresponding polymers with controlled molecular weights and narrow molecular weight distributions (PDI < 1.20). They investigated the effect of different ligands such as TPP, TMPP, and TCPP on the polymerization rate of PEGMA and found that the catalytic activity of iron complex could be markedly enhanced by the introduction of electron donating groups into the TPP ligand.
Yamamoto and coworkers found that the iron mediated ATRP with TPP as a ligand could be successfully applied to grafting methyl methacrylate to polyethylene (PE). Ying et al. conducted an iron mediated ATRP of AN with TPP as a ligand. Noh and his colleagues performed iron mediated ATRP of styrene and MMA using DPPP and DPPMP as ligands and showed that FeBr 2 /DPPP had high catalytic activity in the ATRP of MMA but poor catalytical performance in the ATRP of styrene. In order to promote the polymerization rate of iron mediated ATRP, Matyjaszewski et al. investigated the iron mediated ATRP of styrene using a series of phosphines such as TPP, TMPP, and TTMPP as ligands. They revealed that Fe(III)X 3 /TTMPP was very active in the ATRP of styrene, and a high monomer conversion of 92% could be achieved in 21 h. Noh and coworkers found that the iron mediated ATRP of MMA using TPP, DPPP, or TEP as a ligand was well controlled, but the polymerization of styrene showed poor controllability. They later also investigated iron mediated ATRP using 2-(diphenylphosphino)-N,N -dimethyl-[1,1 -biphenyl]-2-amine (DPPDMA), DPPB, DPPBA, DPMPP, BDPPM, BDPPE, and 1,3-bis(diphenylphosphino)propane (BDPPP) as ligands].
## Organic acid-based ligands
The nitrogen-based ligands and the phosphorous ligands have been widely used in iron mediated ATRP. However, many these ligands are practically toxic and harmful to human health. Therefore, finding 'green' ligands for iron mediated ATRP attracts considerable attentions of chemical researchers. Organic acids seem to be excellent ligands for iron mediated ATRP because of their low toxicity and inexpensive cost characteristics. These acids include pyromellitic acid, IDA, succinic acid (SA), acetic acid, isophthalic acid (IA), and ethylenediamine tetraacetic acid (EDTA).
Yan and coworkers carried out an iron mediated ATRP of styrene using SA as ligand and obtained polystyrene with controlled molecular weights and low polydispersities (M w /M n = 1.23-1.53). The iron mediated ATRP of MMA with IA as ligand and ethyl 2-bromopropionate (EBP) as initiator was found to be well controlled in polar solvents such as N,N-dimethylformamide. However, the polymerization could not be successfully finished in bulk or in nonpolar solvents because the solubility of the catalyst and ligand in monomer or nonpolar solvents were very limited.
Ji and coworkers reported an iron mediated ATRP of acrylonitrile and prepared a poly(acrylonitrile)-g-poly(styrene) copolymer using IDA as ligand. The copolymer could be further modified by NH 2 OH·HCl to produce amidoxime poly(acrylonitrile)-g-poly(styrene) beads, which had an excellent adsorption selectivity for Hg 2+ . Similar poly(acrylonitrile) had also been synthesized with IDA as ligand by Chen group in 2011. Hou et al. used iron mediated ATRP of acrylonitrile with IDA or SA as ligand to yield poly(acrylonitrile) (PAN) with well-controlled molecular weights. Recently, an iron mediated ARGET ATRP of AN using IA as ligand and VC as reducing agent was reported by Chen group, and PAN with narrow molecular weight distributions (M w /M n = 1.14-1.38) were obtained. Iron mediated reverse ATRP of MMA with pyromellitic acid as ligand had been explored by Zhu and coworkers in 2003. A series of (di)picolinic acids including 2,6-pyridine dicarboxylic acid (PDA 1), 2,3-pyridine dicarboxylic acid (PDA 2), 3,5-pyridine dicarboxylic acid (PDA 3), 2,5-thiophene dicarboxylic acid (PCA), and their derivatives have been developed as ligands for iron mediated ATRP in their group. They concluded that the polymerization rate and the polymer polydispersity were largely depended on the structure of ligands.
EDTA was also an excellent ligand for iron mediated ATRP. Malmström and coworkers reported an iron mediated ATRP of styrene using EDTA as ligand. Similarly, Wu and coworkers reported a Fe(0) mediated single electron transfer-living radical polymerization (SET LRP) of MMA using EDTA as ligand and obtained PMMA with controlled molecular weights and low polydispersities (M w /M n = 1. in the presence of a limited amount of air.
## Onium salt-based ligands
Onium salt is also a type of widely used ligand in iron mediated ATRP. Matyjaszewski and coworkers explored a large variety of onium salts as iron ligands and found that FeBr 3 /onium salts catalyzed well-controlled reverse ATRP of MMA and MA, but the catalysts were not efficient for the polymerization of styrene. Zhu et al. investigated the iron mediated ATRP of styrene using onium salt as ligand and discussed the effects of different onium salts such as tetrabutylammonium triflate (TBAOTf), triphenylamino phosphonium bromide (TPAPB), dimethyl diallylammonium chloride (DMDAAC), hexadecyl trimethyl ammonium chloride (HDTMAC), and hexadecyl trimethyl ammonium bromide (HDTMAB) on the polymerization rate and polydispersities of obtained polymers.
Matyjaszewski group investigated iron mediated ATRP of MMA in the presence of a series of salts including TBAOTf, tetrabutylammonium perchlorate (TBAClO 4 ), tetrabutylammonium with BF 4 anion (TBABF 4 ), and tetrabutylammonium with PF 6 anion (TBAPF 6 ). They found that iron/TBAOTf could catalyze ATRP of MMA with excellent controllability not only in anisole but also in nonpolar solvents. The weakly coordinating triflate anions were beneficial to the dissolution of Fe II Br 2 . The iron mediated ATRP of MMA with phosphazenium salts (PZN-X; X = Cl, Br, I) as ligands had been studied by Inoue and coworkers in 2009. They found that the phosphazenium halide was an excellent cocatalyst, and the in situ formed iron halide/phosphazenium halide complex had an excellent catalytic performance in ATRP of alkyl and functionalized methacrylates.
Ionic liquids (ILs) have been used in iron mediated ATRP due to their low volatility and high stability.
A series of ILs such as 1-butyl-3-methylimidazolium bromide (MIBR), 1-butyl-3-methylimidazolium chloride (MICH), 1-butyl-3-methylimidazolium dodecyl sulfate (MICDDS), and 1-butyl-3-methylimidazolium carbonate (MICar) have been investigated by in iron mediated ATRP. The polymerization produced PMMA with controlled molecular weights and narrow molecular weight distributions (M w /M n = 1.. A poly(methacrylonitrile)-b-poly(styrene) block copolymer had been synthesized in the presence of 1-methylimidazolium acetate using PMAN as a macroinitiator.
## 24-1.55) in the presence of ils without any other organic ligands. the iron mediated reverse atrp of man in the presence of ionic liquids such as 1-methylimidazolium acetate ([mim][at]), 1-methylimidazolium caproate ([mim][ct]), 1-methylimidazolium butyrate ([mim][bt]), and 1-methylimidazolium heptylate ([mim][ht]) had been inspected by chen and coworkers
## Miscellaneous ligands
Generally, iron ions could complex with any ligand that has a coordination site. Therefore, a number of polar solvents such as acetonitrile (MeCN) and N,N-dimethylformamide (DMF) may be potential ligands for iron mediated ATRP.
Xue and coworkers conducted iron mediated ATRP of MMA in polar solvents such as N-methylpyrrolidone, DMF, and MeCN using Sn(EH) 2 as a reducing agent. They studied the effects of solvents and different initiators on the polymerization of MMA and found that most of the polymerizations showed excellent controllability. Xue et al. later also succeeded in iron mediated ATRP of MMA in polar solvents using alcohols (e.g., methanol, ethanol, ethylene glycol, glycerol) as reducing agents.
As a summary,
## The enzyme mediated atrp system
Though iron mediated ATRP has been successfully developed and applied to controlled polymerization of a large number of vinyl monomers in recent two decades, the residual iron catalyst in polymer products still presents a challenge to the industrialization of iron mediated ATRP and practically limits its application in biomaterials or microelectronics. Enzymes are usually non-toxic, highly selectable, biodegradable, and environmentally friendly biocatalysts and have been used to synthesize a large number of polymers under mild reaction conditions due to their biocompatible characteristic and high catalytic efficiency. Similarly, enzymes have been utilized as highly efficient catalysts in ATRP system.
Di Lena and coworkers reported that the laccase derived from fungus Trametes versicolor (LTV) could induce ATRP of methacrylic monomers in the presence of alkyl halide initiators and VC. A couple of alkyl halides including EBiB, 2-bromopropionitrile (BPN), ethyl iodoacetate (EIAc), methyl 2-chloropropionate (MCP), and 2-cyano-2-propyl dithiobenzoate (CPDB) were explored in the polymerization of PEGMA. Among these initiators, BPN showed the highest initiation efficiency and produced corresponding polymers with relatively low polydispersity index (PDI = 1.94). The catalytical system was also applied to the polymerization of hydrophobic monomers such as MMA with EBiB as an initiator in the presence of LTV. Later, di Lena et al. performed the polymerization of poly(ethylene glycol) methyl ether acrylate (PEGA) using catalase derived from bovine liver (CBL) as catalyst. The molecular weights of obtained poly(PEGA) increased linearly with the increase of monomer conversions when using BPN as an initiator, and the molecular weight distribution of poly(PEGA) was relatively narrow (PDI = 1.60). Moreover, the polymerization of PEGA in the presence of LTV or horseradish peroxidases (HRP) was also conducted, and the molecular weight of poly(PEGA) was found to increase with the increase of monomer conversion.
Bruns et al. conducted the ATRP of N-isopropylacrylamide (NIPAAm) using HRP as a catalyst and 2-hydroxyethyl-2-bromoisobutyrate (HEBiB) as an initiator and prepared poly(NIPAAm) (PNIPAAm) with a high molecular weight (M n = 9.99 × 10 4 Da) and low polydispersity index (PDI = 1.44). The polymerization kinetics and the effect of different pH (5.2 to 10.5) on the polymerization were investigated. It was found that the highest monomer conversion (78%) was obtained at pH = 7.0. Bruns and coworkers also reported other enzymes (e.g., hemoglobin and red blood cells) mediated ATRP of NIPAAm, PEGMA, and PEGA using HEBiB as an initiator and VC as a reducing agent. All the polymerizations showed a first order kinetic characteristic but a relatively poor controllability. Only the polymerization of PEGA using BPN as an initiator produced poly(PEGA) with a low polydispersity index (PDI < 1.11). The polymerization of PEGA in a polymersome, poly(dimethylsiloxane)-block-poly(2-methyl-2-oxazoline), using HRP as a catalyst and HEBiB as an initiator had also been reported by their group. Kadokawa and coworkers carried out ATRP of NIPAAm using an enzyme mimetic (hematin) as catalyst and found that the number average molecular weight (M n ) of the synthesized poly(NIPAAm) increased linearly with the increase of monomer conversion, but the molecular weight distribution of poly(NIPAAm) was relatively broad (PDI = 1.8-2.1).
Matyjaszewski and coworkers reported polymerization of oligo(ethylene oxide) methyl ether methacrylate (OEOMA) using hemin or its modified products [hemin-(PEG 1000 ) 2 and mesohemin-(MPEG 550 ) 2 , MPEG = methoxy PEG] as catalyst. The hemin mediated ATRP showed poor controllability due to its low halidophilicity. The controllability could be improved by using hemin-(PEG 1000 ) 2 as catalyst. Correspondingly, the obtained poly(OEOMA) had a low polydispersity index (PDI = 1.32) and its molecular weight increased linearly with the increase of monomer conversion. The polymerization could produce poly(OEOMA) with more narrow molecular weight distribution (PDI = 1.19) when mesohemin-(MPEG 550 ) 2 was used as a catalyst.
Deuterohemin-β-Ala-His-Thr-Val-Glu-Lys (DhHP-6) is a synthesized heme-containing peroxidase mimic showing high catalytic performance. Tang and coworkers carried out the DhHP-6 mediated ARGET ATRP of PEGMA and GMA. They found that the molecular weights of corresponding polymers increased linearly with the increase of monomer conversions. Well-defined poly(PEGMA) (M n = 6.02 × 10 3 Da, PDI = 1.08) and poly(GMA) (M n = 8.43 × 10 3 Da, PDI = 1.38) have been successfully synthesized. Poly(ε-caprolactone) (PCL) had been synthesized via enzymatic ring-opening polymerization (eROP) using novozyme 435 as a catalyst and HEBiB as an initiator. The synthesized PCL-Br was further used as a macroinitiator to prepare amphiphilic copolymers such as PCL-PHEMA and PCL-PMAA (PMAA = poly(methacrylic acid)) using DhHP-6 as a catalyst. The integration of eROP and enzyme mediated ATRP was a promising environmentally benign process for the preparation of biomaterials. Tang et al. later reported the polymerization PEGMA using DhHP-6@ZIF-8 (DhHP-6@ZIF-8 = DhHP-6 embedded in zeolite imidazolate framework-8) as a catalyst and BPN as an initiator. This catalyst could be applied to the polymerization of PEGA and NIPAAm.
As oxygen is an undesirable radical inhibitor, it is important to develop a polymerization system that can be conducted in oxygen-rich environments. Matyjaszewski and coworkers reported a well-controlled aqueous ATRP conducted in the open air. This ATRP was realized by continuous conversion of oxygen (O 2 ) to carbon dioxide (CO 2 ) using glucose oxidase (GOx) as catalyst in the presence of sodium pyruvate. In the first step of the polymerization, the glucose and oxygen were converted into d-glucono-1,5-lactone and hydrogen peroxide (H 2 O 2 ). Therefore, the inhibition of O 2 to the polymerization was eliminated. Then, the toxic H 2 O 2 produced in the first step was consumed via the reaction between H 2 O 2 and sodium pyruvate to yield CO 2 , water and acetate. In this case, the ATRP of OEOMA could be successfully achieved, producing poly(OEOMA) with low dispersity (1.09 ≤ PDI ≤ 1.29). In addition, block copolymer poly(OEOMA)-b-poly(OEOMA) (M n = 7.94 × 10 4 Da, PDI = 1.28) had also been prepared by using poly(OEOMA) (M n = 4.23 × 10 4 Da, PDI = 1.23) as a macroinitiator. This GOx mediated approach was also used to polymerize BA and BMA using ethyl α-bromophenylacetate (EBPA) as an initiator, and well-defined poly(BA) (M n = 2.53 × 10 4 Da, PDI = 1.24) and poly(BMA) (M n = 3.24 × 10 4 Da, PDI = 1.16) have been successfully prepared.
Matyjaszewski et al. also reported HRP mediated ATRP of OEOMA using α-bromophenylacetic acid (BPAA) as an initiator. The effect of HRP concentration on the polymerization was investigated. It was found that the polymerization rate increased with the increase of HRP concentration. The monomer conversion reached 94% in 30 min at a HRP concentration of 1130 nM, producing poly(OEOMA) with a low polydispersity index (PDI = 1.17). The chain extension was realized by using poly(OEOMA 500 )-Br (M n = 3.82 × 10 4 Da, PDI = 1.13) as a macroinitiator to copolymerize with OEOMA 300 . Moreover, this system was also applied to the copolymerization of OEOMA with bovine serum albumin (BSA) or human serum albumin (HSA). Well-defined BSA-b-poly(OEOMA) (M n = 6.31 × 10 4 Da, PDI = 1.38) and HSA-b-poly(OEOMA) (M n = 4.01 × 10 4 Da, PDI = 1.25) bioconjugates have been synthesized.
Polymer brush can be used as surface coating due to its ability to endow an interface with a large number of useful properties such as biocompatibility, lubrication and protein-resistance. The surface-initiated atom transfer radical polymerization (SI-ATRP) is a commonly used approach for the preparation of polymer brush. Zauscher and coworkers reported the synthesis of biomedically relevant polymer brushes such as poly(oligo(ethylene glycol) methacrylate) (POEGMA), poly(sulfobetaine methacrylate) (PSBMA), poly(2-dimethylaminoethyl methacrylate) (PDMAEMA), and poly(2-(methylsulfinyl)ethyl acrylate) (PMSEA) via enzyme (GOx) mediated SI-ATRP in an open air environment. The presence of GOx improved the fouling resistance of the polymer materials. Bruns et al. found that the enzyme mediated surface-initiated biocatalytic atom transfer radical polymerization (SI-bioATRP) could be used to prepare PNIPAAm brushes. This method provided a new way for the translation of bioadhesion into a controlled functionalization of materials.
Enzyme catalyzed ATRP has attracted considerable attentions due to the high efficiency, excellent selectivity, mild reaction conditions, and good biocompatibility of enzymes. Representative experimental data of enzyme catalyzed ATRP are summarized in. The polymerizations show typical 'living'/controlled characteristics and produces a lot of well-defined homo-(
## The metal-free catalyst mediated atrp system
Enzyme mediated ATRP has the advantages of high efficiency, mild reaction conditions, and synthesizing biocompatible polymers, which have potential valuable applications in materials science and biomedical engineering areas. However, it suffers from the problems of narrow range of polymerizable monomers and less applicable enzymes. Chemists always envisage developing metal-free catalyst and constructing new environmentally friendly technology for ATRP. The organic photoredox catalyst has been proved to be an excellent candidate.
## Organic photocatalyst mediated metal-free atrp
## Phenothiazines mediated atrp
Hawker and coworkers firstly reported in 2014 a photoinduced metal-free ATRP of methacrylates using 10-phenylphenothiazine (Ph-PTZ) as an organic photocatalyst under UV light irradiation (380 nm). The polymerization showed good controllabilities for MMA, benzyl methacrylate (BnMA), and DMAEMA. The produced PMMA, poly(benzyl methacrylate) (PBnMA) and PDMAEMA had narrow molecular weight distribution of 1.18, 1.25, and 1.11 respectively. Well-defined block copolymer PMMA-b-PBnMA (M n = 2.59 × 10 4 Da, PDI = 1.31) was also prepared via metal-free ATRP by using PMMA as a macroinitiator and Ph-PTZ as a photocatalyst.
Matyjaszewski group performed a photoinduced metal-free ATRP of AN using Ph-PTZ as an organic photocatalyst and EBPA as an initiator, and obtained PAN with M n = 6.20 × 10 3 Da and PDI = 1.60. They also tried the polymerization of AN using 10-(4-methoxyphenyl)-phenothiazine (4-MeOPh-PTZ) or 10-(1-naphthalenyl)-phenothiazine (Nap-PTZ) as a photocatalyst and BPN as an initiator. The molecular weights of produced PAN increased with the increase of monomer conversion at the early stage of the polymerization, but were much higher than theoretical values, indicating that the initiation efficiency of BPN was lower than that of EBPA.
Matyjaszewski et al. also conducted a metal-free ATRP of MMA using Ph-PTZ as a photocatalyst and EBPA as an initiator. The produced PMMA had a molecular weight (M n,GPC = 2.07 × 10 3 Da) close to its theoretical value (M n,th = 1.80 × 10 3 Da) and the molecular weight distribution was relatively narrow (PDI = 1.50). A number of photocatalysts including 10-methylphenothiazine (Me-PTZ), benzo[b]phenothiazine (Ph-benzoPTZ), 9-phenylcarbazole (Ph-CBZ), thianthrene (TH), and N-aryl phenothiazine derivatives (Nap-PTZ) have been investigated in the polymerization of MMA. The results indicated that the Ph-benzoPTZ and Nap-PTZ mediated polymerizations showed better controllability than Ph-PTZ. Moreover, a mechanism investigation of the polymerization catalyzed by phenothiazine derivatives revealed that all the selected catalysts were involved in the activation step, but only part of them were efficiently participated in the deactivation step, leading to different controllability of the polymerization.
Tran et al. reported metal-free ATRP of methacrylate monomers such as MMA, HEMA, and DMAEMA using 4-(10H-phenothiazin-10-yl)-N,N-diphenylaniline (PDPA) as a photocatalyst and phenyl 2-bromo-2-methylpropionate (PhBMP) as an initiator under UV irradiation. The polymerization reached a high monomer conversion of 94.6% and produced PMMA with molecular weight close to theoretical value, indicating that the polymerization was controlled. They also investigated the influence of solvents on the polymerization of MMA and found that the polymerization performed in THF gave higher yield (94.6%) than in DMF (85%) and toluene (43%).
## Aromatic hydrocarbons mediated atrp
Miyake and coworkers developed a metal-free ATRP of methacrylates using perylene as an organic photocatalyst and EBPA as an initiator and prepared PMMA and PBA with narrow molecular weight distributions (PDI < 1.30). Yilmaz group later reported metal-free ATRP of methacrylates and other vinyl monomers using pyrene or anthracene as catalyst and alkyl halides such as EBiB, 1-bromoethyl benzene (BEB), EBP as initiators. They found that EBiB and BEB showed higher initiation efficiency than EBP in the polymerization of MMA and produced PMMA with lower polydispersity index (PDI = 1.38, 1.37 respectively). Other vinyl homopolymers such as poly(tert-butyl acrylate) (M n = 1.07 × 10 5 Da, PDI = 1.32) and PS (M n = 2.00 × 10 3 Da, PDI = 1.32) and PMMA-b-PMMA (M n = 1.92 × 10 5 Da, PDI = 1.40) as well as PMMA-b-PS (M n = 1.85 × 10 4 Da, PDI = 1.50) block copolymers had also been prepared using the same catalyst system.
## Fluorescein mediated atrp
Fluorescein has good chemical stability, visible region absorbance, and favorable redox potential, and can activate alkyl bromide and induce metal-free ATRP by a reductive quenching pathway in the presence of electron donors. Zhang et al. reported a metal-free ATRP of MMA using fluorescein (FL) as an organic photocatalyst in the presence of triethylamine (TEA). The polymerization was controlled and produced PMMA with relatively narrow molecular weight distribution (PDI = 1.46). In order to expand the scope of monomer, styrene, GMA, PEGMA, BnMA, and AN have been polymerized using fluorescein (FL) as a photocatalyst. These polymerizations presented lower controllability than MMA.
Yagci and coworkers conducted metal-free ATRP of MMA using eosin Y or erythrosin B as a photocatalyst in the presence of electron donor amines. The polymerization was completed using EBP as initiator and PMDETA as electron donor under visible light irradiation. The results demonstrated that eosin Y and erythrosin B had higher catalytical activity and controllability than fluorescein and could produce PMMA with narrow molecular weight distributions (PDI = 1.33, 1.20 respectively). The system was also applied to the homopolymerization of other vinyl monomers including styrene, tert-butyl acrylate (t-BA), and HEMA and block copolymerization to prepare PMMA. In addition, PhenN-CF 3 mediated metal-free ATRP of MMA was also achieved under sunlight. The M n of produced PMMA increased linearly with the increase of monomer conversion and the molecular weight distribution was very narrow (PDI = 1.10), indicating that the polymerization was well-controlled. Miyake and coworkers also synthesized a series of N-aryl phenoxazines and successfully used them as catalysts to mediate ATRP of MMA, isobutyl methacrylate (IBMA), BnMA, and isododecyl methacrylate (IDMA).
## Carbazoles mediated atrp
Zhang and coworkers performed a metal-free ATRP of MMA using 1,2,3,5-tetrakis(carbazol-9-yl)-4,6-dicyanobenzene (4CzIPN) as a photocatalyst and EBPA as an initiator under the irradiation of blue light emitting diode. The effect of photocatalyst concentration (5 to 1500 ppm) on the polymerization has been investigated. It was found that the molecular weight distribution of PMMA became broader with the increase of photocatalyst concentration and the polymerization could be achieved without the initiator. The polymerization reached a high monomer conversion (90%) and was well-controlled even at a low concentration of photocatalyst (15 ppm), producing PMMA with a relatively narrow molecular weight distribution (PDI = 1.50). The initiation efficiency of the polymerization was as high as 95.2%.
## Benzaldehyde derivative mediated atrp
Yang et al. reported that the metal-free ATRP of methacrylates could be mediated by benzaldehyde derivative photocatalyst. Three benzaldehyde derivatives including p-anisaldehyde (E 0 * = −2.42 V), p-cyanobenzaldehyde (E 0 * = −2.19), and 2,4-dimethoxy benzaldehyde (E 0 * = −2.60) were investigated in the polymerization. The effects of different initiator including EBiB, EBPA, and perfluoro-1-iodohexane (CF 3 (CF 2 ) 5 -I) on the polymerization have been studied. The results showed that the polymerization of PEGMA could be controlled by using p-anisaldehyde as a catalyst while the polymerization of MMA could be controlled by using p-cyanobenzaldehyde as a catalyst. The polymerization of MMA showed a longer induction period due to the relatively higher oxidation reduction potential of p-cyanobenzaldehyde. The effect of initiator concentration on the polymerization of PEGMA was investigated using 2,4-dimethoxy benzaldehyde as a catalyst and (CF 3 (CF 2 )conducted metal-free ATRP of methacrylates under visible LED light irradiation. N,N-bis(tert-butyloxycarbonyl)-quinacridone (TBOC-QA), N,N-bis(tert-butyloxycarbonyl) -thiophenediketopyrrolopyrrole (TBOC-DPP), and N,N-bis(tert-butyloxycarbonyl)indigo (TBOC-Indigo) were developed as organic photocatalysts. The effects of fluorescence quantum yield, photostability, and reduction potential of these photocatalysts on the polymerization were investigated by using MMA as a monomer and alkyl bromide as an initiator. They found that the polymerization with TBOC-QA as photocatalyst showed excellent controllability, but the polymerization with TBOC-DPP as photocatalyst had a low initiation efficiency of 6.5% due to the poor photostability and electrochemical stability of TBOC-DPP. TBOC-indigo had the lowest fluorescence quantum yield and did not initiate any polymerization. These results indicated that TBOC-QA was a promising photocatalyst for light-controllable ATRP. To investigate the scope of polymerizable monomers in TBOC-QA mediated ATRP, other vinyl monomers including BMA, 2-(diisopropylamino) ethyl methacrylate (DPA), styrene, or OEGMA were also tested. All the monomers except styrene could be well polymerized with TBOC-QA as a photocatalyst.
## Applications of metal-free atrp in the preparation of composite materials
Amphiphilic block copolymer attracts considerable attention due to its self-assemble ability, and has been widely used in drug and gene delivery areas. Son et al. reported that an amphiphilic diblock copolymer could be prepared through metal-free ATRP by using Ph-PTZ as organic photocatalyst under the irradiation of LED light (380 nm). This method was also successfully applied to the polymerization of MMA, GMA, BA, 2-diethylaminoethyl methacrylate (DEAEMA), and allyl methacrylate (AMA). The PMMA-b-PBA (M n = 7.5 × 10 3 Da, PDI = 1.50) block copolymer and poly(ethylene glycol)-b-poly(glycidyl methacrylate) (PEG-b-PGMA) amphiphilic block copolymer were prepared by using PMMA-Br and PEG-Br respectively as a macroinitiator and Ph-PTZ as a photocatalyst. The epoxide groups of PEG-b-PGMA could react with polyethylenimine to produce a cationic polymer bearing oligoamine side chains, which could be applied to gene delivery. They later investigated visible light-mediated metal-free ATRP of MMA using N-trifluoromethylphenyl phenoxazine derivatives as organic photocatalysts. They found that the polymerization was significantly affected by the visible light absorption efficiency, excited state reduction potential, and spatially separated singly occupied molecular orbitals (SOMOs) of the catalyst. The visible light absorption could be enhanced by introducing a biphenyl ring or phenyl with electron-withdrawing groups into the phenoxazine core. On the other hand, a strong excited state reduction potential and spatially separated SOMOs were beneficial for preparing well-defined polymers.
SiO 2 hollow spheres (HS) have attracted much attention due to their wide potential applications in electrical materials and catalysis, biomacromolecule delivery, controlled drug-release carriers, and optical devices. However, the application of SiO 2 HS is largely limited because of its low physical loading ability and poor hydrophobicity. Therefore, surface modification of SiO 2 HS to improve its solubility and physicochemical properties is essential to expand its application scope. Wang and coworkers prepared an amphiphilic diblock copolymer poly(methyl methacrylate)-b-poly(N-isopropylacrylamide) grafted HS (HS-g-PMMA-b-PNIPAAm) hybrid material via metal-free surface-initiated ATRP using Ph-PTZ as a photocatalyst and α-bromoisobutyryl bromide (BIBB) as an initiator. They investigated the dispersions of HS, HS-g-PMMA, and HS-g-PMMA-b-PNIPAAm in inorganic (H 2 O) and organic solvent (THF) and found that the HS was dispersed in water but aggregated in THF while the HS-g-PMMA showed a complete opposite dispersibility to HS, indicating that the grafted PMMA improved the surface hydrophobicity of HS. The HS-g-PMMA-b-PNIPAAm could be dispersed in both THF and H 2 O, implying that the PNIPAAm chains were beneficial for the increase of surface amphiphilicity.
Wang et al. later prepared poly(DEAEMA) (PDEAEMA) grafted silica nanoparticles (SNPs) (SNPs-g-PDEAEMA) and used them for quercetin (Qu) controlled-release. The SNPs-g-PDEAEMA was a kind of pH-sensitive material which was dispersed in acid but aggregated in neuter and alkaline solutions. The self-assembled Qu-loaded microcapsules formed a tight structure under normal physiological conditions (pH = 7.4) with drug entrapped in the core, but the microcapsules became swollen under a weak acid owing to the protonation of the amine groups of PDEAEMA, resulting in drug release from the inner cores. They also evaluated the in vitro cytotoxicity of SNPs-g-PDEAEMA to L929 cells and found that the cell viabilities kept 90.18% and 92.43% on the first and second day respectively, and the cytotoxicity was completely disappeared on the third day. These results showed that the SNPs-g-PDEAEMA had excellent biocompatibility and could be served as drug carriers.
Cellulose is a kind of abundant, inexpensive and renewable biopolymeric materials, and is largely used in daily life. Compared to synthetic polymers derived from petroleum resources, cellulose shows poor solubility in organic solvents, low dimensional stability, and insufficient crease resistance. Wang and coworkers reported a method to prepare cellulose-grafted copolymers using biomass-based monomers such as lauryl methacrylate (LMA), furfuryl methacrylate (FMA), and rosin monomer (DAGMA) via photoinduced metal-free ATRP with Ph-PTZ as a photocatalyst and bromated ethyl cellulose (EC-Br) as an initiator. A series of EC grafted copolymers including EC-g-PLMA (M n = 1. Wang et al. reported a recyclable and sustainable flexible thermoset elastomer derived from fatty acid, furfural, and cellulose via the combination of metal-free ATRP with Diels-Alder reaction. Firstly, the thermoplastic ethyl cellulose (EC) grafted copolymer, EC-g-poly(lauryl methacrylate-co-furfuryl methacrylate) (EC-g-P(LMA-co-FMA)), was synthesized via metal-free ATRP using Ph-PTZ as catalyst. Then, a modified epoxidized soybean oil containing 6-maleimidohexanoic group (ESOM) was employed to conduct Diels-Alder reaction with furfural groups in the chain of EC-g-P(LMA-co-FMA) to prepare the recyclable thermoset elastomers. The formation of dynamic crosslinking structure by Diels-Alder reaction provided an excellent self-healing and recyclability for the thermoset elastomers.
Hydrogels have been widely applied in the fields of drug-controlled release, biosensors, tissue engineering and adsorbents. However, the service life of conventional hydrogel is short due to its poor mechanical strength. The mechanical strength and self-healing properties of hydrogel can be enhanced by cellulose nanocrystals (CNCs) due to their highly crystalline and nontoxic nanorods characteristics. Bai et al. reported the preparation of self-healing nanocomposite hydrogels based on modified CNCs. 4-vinylpyridine (4VP) was surface-initiated onto the surface of CNCs via metal-free ATRP using Ph-PTZ as a photocatalyst to form poly(4-vinylpyridine) CNCs hybrid material (CNCs@P4VP). The CNCs@P4VP was an excellent reinforcement for self-heal poly(acrylic acid) (PAA) hydrogels. The reversible electrostatic interaction between the carboxyl group of PAA and the pyridyl group of CNCs@P4VP acted as a dynamic reversible supramolecular interaction to heal and crosslink the PAA hydrogels. The prepared nanocomposite hydrogels showed an excellent self-healing (85.9% after repairing 6 h) and mechanical properties (6.6 MPa at a strain of 921.6%).summarizes typical experimental data of metal-free catalyst mediated ATRP. A large number of organic compounds including phenothiazines, Entry 1, 3, 4, and 8), polynuclear aromatic hydrocarbons, Entry 9 and 10), fluorescein, Entry 11), N-aryl phenoxazines, Entry 12), carbazoles, Entry 13), and benzaldehyde derivative, Entry 14), have been developed as photocatalyst and a lot of homopolymers, block polymers, star polymers, hyperbranched polymers, and composite materialshave been prepared via metal-free ATRP.
## Summary and future perspectives
Though copper complex is the most commonly used ATRP catalyst, the application of copper catalyst at industrial scale is restricted due to its biological toxicity and environmentally unsafety. Using iron complex as ATRP catalyst has attracted considerable interest in recent decades due to the negligible toxicity, low cost, and environmental friendliness of the iron catalyst. This review summarized the applications of iron catalyst in normal ATRP, reverse ATRP, ICAR ATRP, AGET ATRP, GAMA ATRP, and SARA ATRP in view of the catalytic activity, initiation efficiency, and polymerization controllability. A large variety of homo-, block-, graft-, brush-, star-, and hyperbranched polymers have been prepared via iron complex mediated ATRP and summarized in the review. As the catalytic activity of iron catalyst is largely depended on the ligand of the complex, the development of the catalyst ligand has also been discussed in this review.
Despite significant success and progress in iron complex-catalyzed ATRP, there are still challenges for iron catalysis. For instance, it is necessary to establish a structure-reactivity relationship for iron catalyst, especially the dependence of the activation and deactivation rate constant on the redox potential of iron complex. On the other hand, the polymerization of monomers containing strong polar functional groups is still troublesome due to the interaction between the polar groups and the iron catalyst. Therefore, developing more stable iron catalyst would be benefit to the development of iron-catalyzed ATRP. Moreover, developing new iron catalyst derived from biological resources for the preparation of biocompatible polymers would also be conducive to the iron complex catalyzed ATRP.
Though enzyme mediated ATRP has the advantage of synthesizing biocompatible polymers and bioconjugates, it suffers from limited polymerizable monomers and less applicable enzymes. It would be very advantageous to develop other enzymes such as hydrogenases and chlorophyl for ATRP. In addition, incorporating this method in more important reaction systems and fabricating multifunctional enzyme-containing materials with outstanding performances are also expected to be a direction of the enzyme mediated ATRP.
As metal-free ATRP is regarded as a green and sustainable process for precise polymer synthesis, a lot of desirable homo-, block-, star-, and hyperbranched polymers have been prepared via metal-free ATRP. However, a relatively large amount of photocatalyst is generally required in the polymerization. In addition, the use of photocatalyst often causes a problem of discoloration of the polymer products due to the highly colored performance of some photocatalysts. Further research is recommended to improve the catalyst activity, develop new photocatalysis strategy for cost-effective production of various polymeric materials and solve the problem of the discoloration. As many biocompatible materials such as HS-g-PMMA-b-PNIPAAm, SNPs-g-PDEAEMA, EC-g-PLMA, and CNCs@P4VP have been successfully prepared via metal-free ATRP, the application of metal-free ATRP in medicine, electricity, and other interdisciplinary areas would have a more promising perspective. |
Treatment patterns of prostate cancer with bone metastasis in Beijing: A real‐world study using data from an administrative claims database
Purpose: To explore treatment patterns among patients with prostate cancer and bone metastasis and to compare clinical outcomes following use of different hormone therapies including combined androgen blockade (CAB), nonsteroidal antiandrogen (NSAA) monotherapy, and castration monotherapy.Methods:We conducted a population-based cohort study using data from the Urban Employee Basic Medical Insurance database (2011)(2012)(2013)(2014) in Beijing. We identified 475 patients with newly diagnosed bone metastatic prostate cancer with at least one prescription for hormone therapy and described their treatment patterns over a median follow-up of 20.7 months. Cox proportional hazards model was used to compare time to chemotherapy initiation between patients starting on different hormone therapies.Results: Hormone therapy and/or bisphosphonate therapy with zoledronic acid were the initial treatments in the majority of patients (87.8%); chemotherapy, radiotherapy, and surgery were usually given later in the treatment pathway. CAB was the most common hormone treatment (73.7%). For time to chemotherapy initiation, hazard ratios (95% confidence intervals) were 2.43 (1.08-5.44) for NSAA alone vs CAB and 1.29 (0.78-2.13) for castration alone vs CAB.Conclusions:Our findings show that while a wide range of therapies are used to treat patients with prostate cancer and bone metastasis in Beijing, hormone therapy and bisphosphonate therapy are the most commonly prescribed, and use of CAB was seen to be advantageous in delaying time to chemotherapy initiation over NSAA monotherapy. Future studies should explore longer-term treatment patterns, including use of newly approved treatments.---This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
## Keywords
bone metastasis, claims database, combined androgen blockade, drug utilization, pharmacoepidemiology, prostate cancer, treatment pattern
# | introduction
China has seen a rapid increase in the incidence and prevalence 1 of prostate cancer in recent decades (ranking sixth in all male cancers), which is likely due to the ageing population and changes in diagnostic practices.In the absence of early diagnosis and primary treatment, prostate cancer commonly progresses to an advanced stage, frequently with bone metastases that have a profound impact on patients' quality of life and place a great burden on healthcare resources.We have previously shown that bone metastasis in prostate cancer patients in China are common , and higher than in other countries, and related treatment costs have increased.Continuous androgen deprivation therapy (ADT) as the main hormone therapy is the standard of care for patients with metastatic prostate cancer, providing symptom relief and delaying disease progression.However, ADT can lead to loss of bone mineral density and an increased risk of osteoporotic fragility fractures. The combined effects of ADT and bone metastasis can result in skeletal complications such as pain, ineffective haematopoiesis, and skeletalrelated events (SREs).Thus, the management of patients with bone metastatic prostate cancer often requires use of bone-targeted agents (BTAs) such as bisphosphonates, denosumab, and radium-233, 4,7-9 as well as palliative radiation, bone surgery, and pain relief. After initial ADT treatment, most patients eventually become unresponsive to castration 10 -metastatic castration-resistant prostate cancer (mCRPC). First-line treatment for this patient population has been docetaxel-based chemotherapy although a few newer agents have been approved in recent years.There are three main types of hormone therapy for patients with prostate cancer.The first type is antiandrogen alone, usually a nonsteroidal antiandrogen (NSAA) such as bicalutamide, flutamide, and nilutamide. The second type is castration alone, including luteinizing hormone-releasing hormone (LHRH) agonists or antagonists (ie, medical castration) and bilateral orchiectomy (ie, surgical castration).
The third type is medical or surgical castration combined with an NSAA-combined androgen blockade (CAB). Comparative effectiveness studies of the three methods have shown mixed findings, and international guidelines are conflicting.In China, there are limited data regarding the treatment pathways of patients with prostate cancer and bone metastasis in routine clinical practice, yet this is important knowledge to gain to evaluate whether patients are currently receiving the best available medical treatment. Furthermore, little is also known about the effectiveness of CAB compared with hormone monotherapies. We therefore aimed to explore treatment patterns among this patient population, including frequencies and sequences of therapies used and a comparison of different hormone therapies on the clinical outcomes of time to chemotherapy initiation and SRE occurrence. The study was set in Beijing, which has the nation's most advanced health care service.
## | materials and methods
## | study design and data source
We conducted a population-based cohort study using data from a large medical claims database in Beijing-the Urban Employee Basic Medical Insurance (UEBMI) database. The UEBMI is one of the three main national health insurance schemes in China and is a mandatory program covering urban and retired employees.The UEBMI contains
## Key points
- We used a claims database to summarize the observed treatment patterns of prostate cancer with bone metastasis in real-world clinical practice in Beijing; thus, our results will have better generalizability compared with those from clinical trials.
- Hormone therapy and bisphosphonate therapy were the most commonly prescribed treatments for prostate cancer with bone metastasis, in accordance with clinical guidelines.
- Chemotherapy was seldom used as initial therapy, but when used, the most common regimens were docetaxel and estramustine. Radiotherapy and surgery were used less frequently, but when used were often given in a late phase of the treatment pathway.
- Combined androgen blockade potentially has an advantage in delaying chemotherapy initiation over
## | study population
We included patients meeting all the following criteria between 1
July
## | measurements
We evaluated seven classes of treatments according to the 2014 version of the Chinese Urology Association (CUA) guideline for diagnosis and treatment of prostate cancer and following clinical expert consultation: surgery, radiotherapy, hormone therapy, chemotherapy, bisphosphonate therapy, pain treatment, and traditional Chinese medicine (TCM).Hormone therapy was classed into three main groups: NSAA alone, castration alone, or CAB. If NSAA treatment period was less than 30 days and the interval between start of LHRH and NSAA treatments was less than 14 days, the patient was included in the castration alone group rather than the CAB group. This was because all patients with planned use of LHRH agonists should be prescribed short-term antiandrogens to prevent disease flares.SREs were defined as clinical manifestations of pathological fracture, spinal cord compression, hypercalcemia, or surgery involving bone. Comorbidities were identified using ICD-10 codes and/or Chinese text in the medical claim records (any time from 1 January 2011 to the index date) and included hypertension, cardiovascular diseases, cerebrovascular disease, type 2 diabetes, and respiratory diseases (asthma and chronic obstructive pulmonary disease). Drug prescriptions were identified using Anatomical Therapeutic Chemical classification codes, and medical procedures were identified using a unified coding system developed by the data owner.
## | treatment sequence and clinical outcomes
In addition to identifying the initial treatment prescribed, we described the sequence of treatment over time, in particular noting the first three distinct treatments and the time interval between adjacent treatments. A Sankey plot was produced to visualize the treatment pathways. To compare clinical outcomes after treatment with CAB, NSAA alone, or castration alone, we followed patients from the first date of the respective hormone therapies to identify the occurrence of SREs and the initiation of chemotherapy (as a proxy for disease progression). To do this, we conducted two time-to-event analyses, one for each outcome. In these two analyses, patients who did not receive any of the three exposure treatments of interest (eg, patients who only received oestrogens as hormone therapy) and those who already had an outcome event (chemotherapy) before the relevant exposure were excluded. Patients without an outcome of interest were censored when they were last known to be event-free during the observation period.shows the numbers of patients prescribed specific hormone therapy and chemotherapy agents, and the number of medical/hospital visits made for use of each of these therapies.
# | statistical analysis
## | treatments at any time during follow-up
## | prostate cancer-related hospital visits
## | commonly used drugs
Among the hormone therapies, NSAAs (predominantly bicalutamide followed by flutamide) and LHRHs (mostly leuprorelin or goserelin followed by triptorelin) were the most commonly prescribed drugs.
Oestrogens (mainly megestrol) were less frequently prescribed.
Among chemotherapy drugs, docetaxel was the most commonly prescribed, followed by estramustine, platinum-based agents, and fluorouracil. to radiotherapy was 171 days, and to surgery was 226 days.
## | treatment pathways
## | time to initiation of chemotherapy and first sre
Adjusted HRs and median progression time from first hormone therapy to initiation of chemotherapy/first SRE are presented in. Direct adjusted survival curves are shown in. After adjusting for covariates, patients treated with NSAA alone had more than double the likelihood of subsequently receiving chemotherapy compared with those treated with CAB (HR, 2.43; 95% CI, 1.08-5.44; P = .031), while no significant difference was found between other hormone therapy treatment groups for chemotherapy initiation.
In the time to first SRE analysis, the median progression time to SRE was not reached in each group and there was no evidence of difference in time to first SRE between hormone groups.
# | discussion
In this population-based cohort study, we have shown that hormone therapy and bisphosphonate therapy with zoledronic acid are the most commonly used for patients in Beijing with prostate cancer and bone metastasis, with the majority of patients starting treatment with one or both of these therapies, in accordance with current guidelines.We have also shown that time to chemotherapy initiation was longer 93.5% of patients received castration in our study, the proportion is higher than that reported by Flaig et al (65%). Among them, medical castration (85.9%) was a more common form of castration than surgery (7.6%), though they appear equally effective with orchiectomy proved even safer recently.Possible reason is that medical castration is potentially reversible and avoids the physical and psychological discomfort associated with orchiectomy.Bone-targeting agents reduce bone reabsorption by primarily targeting osteoclasts.They are important in the treatment of patients with bone metastatic prostate cancer, and if not used, approximately half of patients will experience one or more SREs within 2 years.Data from clinical trials suggest that conventional bone-targeting treatments, such as bisphosphonate therapy with zoledronic acid, which was prescribed to most patients in our study, does not increase overall survival.Radium-233-one of several newer agents available-has, however, been shown to prolong survival in addition to reducing symptomatic bone complications.Forty-two percent of patients in our study were prescribed chemotherapy, which is very similar to that reported in a claims database study in the United States (45.4%).Docetaxel is the standard of care in patient with CRPC and was the most commonly used chemotherapy drug in our study cohort, followed by estramustine, which is no longer recommended by the American guidelinebut still retained in the Chinese guidelines.Since 2015, the use of chemotherapy together with ADT as initial therapy for mHSPC patients has been supported by clinical guidelines.In our study, we found that chemotherapy was rarely prescribed as the initial therapy and therefore was likely a sign of disease progression, as used as a secondary endpoint in a recent clinical trial.Radiotherapy and surgery were less frequently used and usually given late in the treatment pathway. Similar treatment patterns were observed by Seal et al in a hospital database study.However, for the specific drugs used in each class of treatment, patterns found in our study differed from that in a US population,evidence is required, especially from prospective randomized trials, to be confident about a survival advantage with CAB. According to ASCO and CUA, NSAA alone may be discussed as an alternative in patients at certain stage, while EAU is more neutral to its use and NCCN does not recommend its use.The results of our study moderately add to the evidence favouring CAB over NSAA alone but not over castration alone on delaying chemotherapy initiation in Chinese bone metastatic prostate cancer patients. We did not find any differences in occurrence of SREs between CAB and hormone monotherapies, although it should be noted that it is BTAs rather than hormone therapies that specifically aim to reduce these outcomes.
Our study has several strengths. To our knowledge, it is the first to look at treatment patterns among patients with prostate cancer and bone metastasis in China and thus provides valuable clinically important information describing the medical management of these patients.
Without data such as ours to use as a benchmark, it would be difficult to describe use and effectiveness of newer approved medications as they become approved and prescribed in China, through both future descriptive and analytical epidemiological studies. We used a large database representative of urban employees in Beijing enabling a reasonable sample size to be acquired. Our study also has some limitations. Firstly, the UEBMI database does not provide results of mortality data or detailed clinical and pathologic information; hence, we were unable to directly identify patients with mCRPC, and we used time to chemotherapy initiation and first SRE as surrogate outcome measures. Secondly, the sample size might still be insufficient to detect small to moderate differences, and the observation period was short leading to many censored observations in the survival analysis especially for the SRE endpoint, which to some extent undermines the observed results.
Finally, Beijing is one of the few first tier cities in China with better health care resources; therefore, treatment patterns identified in this study cannot necessarily be generalized to other regions of China.
# | conclusions
Our findings show that while a wide range of therapies are used to
# Ethics statement
This study was approved by the Ethical Review Board of Peking University Health Science Center (No. IRB00001052-16027-Exempt), and informed consent of participant was exempted. |
Breast cancer patients with hormone neoadjuvant bridging therapy due to asymptomatic Corona virus infection. Case report, clinical and histopathologic findings
# Introduction
COVID-19 is a respiratory disease that spreads from human to human. Due to its dramatical worldwide spread, WHO (World Health Organization) declared the pandemic on March 11, 2020. This Pandemic poses a challenge to health care system which had to face simultaneously the infectious emergency and other pathologies. Due to this emergency, oncological diagnostic procedures and treatments suffered a significant slowdown in order to save resources towards COVID-19 [bib_ref] Tor Vergata University-Hospital in the beginning of COVID-19-Era: experience and recommendation for..., Buonomo [/bib_ref]. Several scientific societies have released consensus recommendations for management of cancer patients [bib_ref] Ethics and resource scarcity: ASCO recommendations for the oncology community during the..., Marron [/bib_ref] [bib_ref] International Oncology Panel and European Cancer Patient Coalition collaborators, Summary of international..., Mauri [/bib_ref]. Proposal are driven by the common aim to prioritize cancer treatments and preserve resources for COVID-19 emergency [bib_ref] Tor Vergata University-Hospital in the beginning of COVID-19-Era: experience and recommendation for..., Buonomo [/bib_ref]. Multidisciplinary treatment of these patients has changed. The risk associated with surgery (e.g. staff and patient's exposure to virus and the need of resources) were balanced against the risk of cancer progression while the patient was receiving alternative therapy (e.g. systemic therapy, radiotherapy) [bib_ref] Case 22-2020: a 62-year-old woman with early breast cancer during the Covid-19..., Spring [/bib_ref]. In case of COVID-19 patients with simultaneous neoplasia it was strongly recommended to treat Sars-CoV-2 infection firstly due to its mortality, morbidity and contagiousness [bib_ref] International Oncology Panel and European Cancer Patient Coalition collaborators, Summary of international..., Mauri [/bib_ref] [bib_ref] Case 22-2020: a 62-year-old woman with early breast cancer during the Covid-19..., Spring [/bib_ref] [bib_ref] Awake breast cancer surgery: strategy in the beginning of COVID-19 emergency, Vanni [/bib_ref]. This implied a delay on the diagnosis and treatment of cancer; Vanni et al.. In addition, COVID-19 patients' anxiety and refusal of health facilities admission could contribute to a further the delay [bib_ref] Breast cancer and COVID-19: the effect of fear on patients' decision-making process, Vanni [/bib_ref].
We report a case of woman with COVID-19 and simultaneous early breast cancer treated with neoadjuvant endocrine therapy in lieu of upfront surgery and with lymph node micrometastases at pathological examination. This work is reported by following the surgical case report (SCARE) guidelines [bib_ref] For the SCARE Group, The SCARE 2018 statement: updating consensus surgical CAse..., Agha [/bib_ref].
## Presentation of case
A 53-years-old G5P3 post-menopausal women, in good clinical condition, visited our surgical outpatient clinic on February, 14 2020, with a new-onset breast lesion noted with self-examination. Family history was negative for cardiovascular and oncological disease. Her past medical history included moderate psoriasis treated with Tofacinib [bib_ref] Tofacitinib for the treatment of moderate-to-severe psoriasis, Chiricozzi [/bib_ref]. Physical examination revealed a palpable lesion, with a diameter of about 1 cm in the lower inner quadrant of the right breast without any clinically suspicious axillary lymph nodes. She underwent mammography and breast ultrasonography that revealed a suspicious malignant breast nodule BIRAD-V, with a maximum diameter of 1.4 cm. At ultrasonography, lymph nodes appear increased in volume (about 1.1 cm) but with predominant hilar vessel architecture without any radiological suspicion of malignancy. Biopsy of the nodule was planned and the malignant nature of the lesion was confirmed. The pathological examination of the biopsy revealed invasive ductal carcinoma: ER positive 90%, PR positive 45%, Ki67 25% and c-Erb-B2 negative score 0. The case was discussed at the breast cancer multidisciplinary meeting and up-front surgery was planned (right breast quadrantectomy and sentinel lymph node biopsy). On March 7, 2020 her husband was admitted to hospital due to COVID-19. Despite she was asymptomatic she underwent SARS-CoV-2 test and resulted positive (Molecular test on nasal and throat swab). According to oncological recommendations during COVID-19 pandemic, the case was re-discussed at the multidisciplinary meeting and neoadjuvant endocrine therapy was planned (bridging therapy) in lieu of upfront surgery. Patient accepted the scheduled treatment and reported anxiety and fear regarding Coronavirus diagnosis during followup telehealth revaluation. Patient remained asymptomatic, blood tests (lymphocyte count and subpopulation, D-dimer, coagulation indices, and blood inflammatory indices) were regular with positive Sars-CoV-2 tests (totally four swabs) for two months. She did not receive any treatment for COVID-19. Eighty-seven days after SARS-CoV-2 positive tests, nasal and throat swab resulted negative. Two other swab were performed respectively fifteen and thirty days from the negative one, and both confirmed the negative result. After infectious disease resolution, the surgical procedure was planned. Totally she was treated with 20 mg/day of Tamoxifene for 90 days with no noteworthy side effects and good adherence to treatments. On July 1, 2020 the patient underwent, by breast surgeon with twenty-years of experience, right breast quandtrantectomy and sentinel lymph nodes biopsy (resulted negative at frozen section). Histological examination revealed invasive ductal carcinoma grade 2 spanning at least 1.2 cm in the greatest dimension associated with an in situ component less than 20%. Immunohistochemical staining revealed strongly and diffusely ER and PR positive in tumor cells: <95% and 40% respectively [fig_ref] Figure 1: Pathological examination of lymph node and breast surgical specimen [/fig_ref] HER2 was negative and Ki67 was 30% [fig_ref] Figure 1: Pathological examination of lymph node and breast surgical specimen [/fig_ref]. Sentinel lymph node presented a micrometastasis of 0.3 mm and showed germinal centers in reactive exhaustion and mild expansion of T-dependent paracortical zone, consistent with benign reactive lymphadenitis as a result of the viral infection (Hematoxylin-Eosin staining, 2x; 10x in the insert) [fig_ref] Figure 1: Pathological examination of lymph node and breast surgical specimen [/fig_ref]. No viral RNA of SARS-CoV-2 was found in lymph node or in the breast. On 1 month's follow-up patient was in good clinical condition with adjuvant chemotherapy on going. In the same visit, written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request.
# Discussion
SARS-CoV-2 has dramatically spread worldwide forcing the implementation of restrictive measures and changing our daily routine [bib_ref] Tor Vergata University-Hospital in the beginning of COVID-19-Era: experience and recommendation for..., Buonomo [/bib_ref]. Health system experienced a total reorganization which also involved breast units and their guidelines [bib_ref] Tor Vergata University-Hospital in the beginning of COVID-19-Era: experience and recommendation for..., Buonomo [/bib_ref] [bib_ref] Ethics and resource scarcity: ASCO recommendations for the oncology community during the..., Marron [/bib_ref] [bib_ref] International Oncology Panel and European Cancer Patient Coalition collaborators, Summary of international..., Mauri [/bib_ref] [bib_ref] Case 22-2020: a 62-year-old woman with early breast cancer during the Covid-19..., Spring [/bib_ref].
Before COVID-19 pandemic, patient such as this, with clinical stage T1N0, hormone receptors positive HER2-negative breast cancer, would have been a candidate for upfront surgery [bib_ref] New insights into the metastatic behavior after breast cancer surgery, according to..., Buonomo [/bib_ref]. Given the size of patient's breast and the small diameters of the lesion, conservative breast surgery would be recommended to preserve body image and reduce hospitalization especially in this emergency period [bib_ref] Comparative study of oncoplastic versus non-oncoplastic breast conserving surgery in a group..., Cassi [/bib_ref] [bib_ref] Day-surgical management of ductal carcinoma in situ (DCIS) of the breast using..., Buonomo [/bib_ref] [bib_ref] Radioguided-surgery of early breast lesions, Buonomo [/bib_ref].
According to modified management of breast cancer patients during COVID-19 pandemic, this patient was candidate to neoadjuvant endocrine therapy followed by conservative breast surgery and radiotherapy; Spring et al. [bib_ref] Case 22-2020: a 62-year-old woman with early breast cancer during the Covid-19..., Spring [/bib_ref] [bib_ref] Awake breast cancer surgery: strategy in the beginning of COVID-19 emergency, Vanni [/bib_ref] [bib_ref] Breast cancer and COVID-19: the effect of fear on patients' decision-making process, Vanni [/bib_ref] [bib_ref] For the SCARE Group, The SCARE 2018 statement: updating consensus surgical CAse..., Agha [/bib_ref] [bib_ref] Tofacitinib for the treatment of moderate-to-severe psoriasis, Chiricozzi [/bib_ref] [bib_ref] New insights into the metastatic behavior after breast cancer surgery, according to..., Buonomo [/bib_ref] [bib_ref] Comparative study of oncoplastic versus non-oncoplastic breast conserving surgery in a group..., Cassi [/bib_ref] [bib_ref] Day-surgical management of ductal carcinoma in situ (DCIS) of the breast using..., Buonomo [/bib_ref] [bib_ref] Radioguided-surgery of early breast lesions, Buonomo [/bib_ref] [bib_ref] Serum sE-selectin levels and carcinoembryonic antigen mRNA-expressing cells in peripheral blood as..., Ferroni [/bib_ref]. Moreover, despite the patient was asymptomatic, she resulted SARS-CoV-2 positive. In order the reduce the risk of cross-infection among health workers and other patients, the avoidance of hospitalization and the availability of a bridging therapy could be considered feasible and safe.
Neoadjuvant hormone therapy was administered for 3 months and during this time the negativization of SARS-CoV-2 was achieved [bib_ref] Case 22-2020: a 62-year-old woman with early breast cancer during the Covid-19..., Spring [/bib_ref]. Although endocrine therapy is a mainstay in the adjuvant treatment, its role in the neoadjuvant schedule is unclear [bib_ref] Case 22-2020: a 62-year-old woman with early breast cancer during the Covid-19..., Spring [/bib_ref]. Moreover, despite the lack of image guided clinical response evaluation during primary therapy, neoadjuvant regimen allowed to perform surgery safely for patient, and to avoid cross infection in health care workers and other patients [bib_ref] Tor Vergata University-Hospital in the beginning of COVID-19-Era: experience and recommendation for..., Buonomo [/bib_ref].
Pathological examination of sentinel lymph node showed germinal centers in reactive exhaustion and mild expansion of T-dependent paracortical zone, consistent with benign reactive lymphadenitis resulting from the viral infection. Unlike what reported in literature, no viral RNA was found in the lymph nodes [bib_ref] Molecular detection of SARS-CoV-2 infection in FFPE samples and histopathologic findings in..., Sekulic [/bib_ref] [bib_ref] Spontaneous immunogenicity of ribosomal P0 protein in patients with benign and malignant..., Marzocchella [/bib_ref] [bib_ref] Tumor markers as targets for selective diagnostic and therapeutic procedures, Roselli [/bib_ref]. The absence of RNA of SARS-CoV-2 in the lymph node could be probably correlated to the eradicated infection. However, the T-dependent response could be an indirect sign of SARS-CoV-2 infection. In literature was reported that SARS-CoV-2 RNA can be detected in routine histopathological samples even before COVID-19 disease development. Differently, in our case viral RNA was not found neither in the surgical specimen probably correlating with the infection eradication.
Waiting time between diagnosis and surgery in our case was about 6 months. During the pandemic, oncological diagnostic procedures and treatments experienced a significant slowdown [bib_ref] Tor Vergata University-Hospital in the beginning of COVID-19-Era: experience and recommendation for..., Buonomo [/bib_ref] [bib_ref] Ethics and resource scarcity: ASCO recommendations for the oncology community during the..., Marron [/bib_ref] [bib_ref] International Oncology Panel and European Cancer Patient Coalition collaborators, Summary of international..., Mauri [/bib_ref] [bib_ref] Case 22-2020: a 62-year-old woman with early breast cancer during the Covid-19..., Spring [/bib_ref] [bib_ref] Awake breast cancer surgery: strategy in the beginning of COVID-19 emergency, Vanni [/bib_ref] [bib_ref] Breast cancer and COVID-19: the effect of fear on patients' decision-making process, Vanni [/bib_ref] [bib_ref] For the SCARE Group, The SCARE 2018 statement: updating consensus surgical CAse..., Agha [/bib_ref] [bib_ref] Tofacitinib for the treatment of moderate-to-severe psoriasis, Chiricozzi [/bib_ref] [bib_ref] New insights into the metastatic behavior after breast cancer surgery, according to..., Buonomo [/bib_ref] [bib_ref] Comparative study of oncoplastic versus non-oncoplastic breast conserving surgery in a group..., Cassi [/bib_ref] [bib_ref] Day-surgical management of ductal carcinoma in situ (DCIS) of the breast using..., Buonomo [/bib_ref] [bib_ref] Radioguided-surgery of early breast lesions, Buonomo [/bib_ref] [bib_ref] Serum sE-selectin levels and carcinoembryonic antigen mRNA-expressing cells in peripheral blood as..., Ferroni [/bib_ref] [bib_ref] Molecular detection of SARS-CoV-2 infection in FFPE samples and histopathologic findings in..., Sekulic [/bib_ref] [bib_ref] Spontaneous immunogenicity of ribosomal P0 protein in patients with benign and malignant..., Marzocchella [/bib_ref] [bib_ref] Tumor markers as targets for selective diagnostic and therapeutic procedures, Roselli [/bib_ref] [bib_ref] Does age matter? Estimating risks of locoregional recurrence after breast-conservative surgery, Vanni [/bib_ref]. In our opinion waiting time can influence breast cancer staging [bib_ref] Tor Vergata University-Hospital in the beginning of COVID-19-Era: experience and recommendation for..., Buonomo [/bib_ref]. Bridging neoadjuvant hormone therapy could be a valid option in order to achieve a partial or complete response [bib_ref] Case 22-2020: a 62-year-old woman with early breast cancer during the Covid-19..., Spring [/bib_ref]. Monitoring cancer response rates during the lockdown, especially in patients SARS-CoV-2 positive may not be easy to achieve. In our case, histological examination of sentinel lymph nodes revealed micrometastasis. Due to node involvement patients started adjuvant chemotherapy with a strong impact on her quality of life [bib_ref] For the SCARE Group, The SCARE 2018 statement: updating consensus surgical CAse..., Agha [/bib_ref]. We cannot exclude that this lymph node involvement is correlated to the longer waiting time.
# Conclusion
Therefore, the use of bridging therapy in patients with early breast cancer, during pandemic, that could benefit from upfront surgery should be evaluated in large sample studies. If it is not feasi- ble to monitor the tumor response during the neoadjuvant therapy, different strategy should be preferred. Awake breast surgery, with personal protective equipment especially in asymptomatic COVID-19 patient could be an option in order to reduce waiting time and possible cancer progression.
## Declaration of competing interest
All the authors declare no conflict of interest.
# Funding
This publication was found with the non-conditional contribution of the Italian Ministry of Health.
# Ethical approval
For this study, ethical and ethnical approval are not required.
## Consent
Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of written consent is available for review by the Editor-in-Chief of this journal on request.
# Author contribution
## Registration of research studies
Not applicable.
## Guarantor
Marco Materazzo.
[fig] Figure 1: Pathological examination of lymph node and breast surgical specimen. A) Section of axillary lymph node showing germinal centers in reactive exhaustion and mild expansion of T-dependent paracortical zone, consistent with benign reactive lymphadenitis post viral infection (Hematoxylin-Eosin staining, 2x; 10x in the insert). B) Section of breast tumor morphologically consistent with invasive ductal carcinoma, NST according to WHO Classification, grade 2 according to the Elston-Ellis modification of the Scarff-Bloom-Richardson (SBR) grading system (Hematoxylin-Eosin staining, 2x; 10x in the insert). C) Immunostaining showing Estrogen Receptor expression in neoplastic cells (clone SP1, Ventana Roche, 10x). D) Immunostaining showing Progesterone Receptor expression in neoplastic cells (clone 16, Leica, 10x). E) Proliferative index evaluated by Ki-67 immunostaining (clone MM1, Leica, 10x). F) Immunostaining for evaluation of Her2 expression, showing mild cytoplasmic reactivity and absence of membranous staining in neoplastic cells, classified as score 0 (clone CB11, Leica, 10x). [/fig]
|
Graphene/Reduced Graphene Oxide-Carbon Nanotubes Composite Electrodes: From Capacitive to Battery-Type Behaviour
# Introduction
Besides the energy transformation, energy storage is one of the most important topics of scientific research today. Similar to the existing commercial batteries, electrochemical capacitors (often called supercapacitors (SC)) are widely studied for their commercial application in electrical cars, portable electronics, etc. The most famous capacitive materials are carbon-based compounds, particularly activated carbon (AC) that is already used for energy storage due to its large surface area and low cost . However, many carbon atoms in AC cannot be accessed by electrolyte ions, thus being wasted in terms of activating their electrochemical functions, due to a very irregular shape of AC, shown in. That decreases the capacitance of the AC electrodes as well as reducing their electrical conductivity (see .
In addition to AC, a graphene (G), being 2D sp 2 -hybridized carbon sheet, also has a large surface area as shown in . However, its electrical conductivity is significantly higher, making graphene-related materials very promising for energy storage. However, from G it is easy to form aggregates due to intensive π-π interaction or to restacking forming graphite through van der Waals interactions. If G sheets are stacked together, the electrolyte ions have difficulties gaining access to the inner layers to form electrochemical double layers. Thus, the ions could be accumulated only on the top and bottom surfaces of the sheet agglomerates that can lead to a lower specific capacitance. The problem of graphene sheet agglomerations during the preparation process is valid also for multilayer reduced graphene oxide (rGO) despite lower surface area compared to that of monolayer G. However, the lower price and properties tunable by oxidation degree as well as high. Copyright Royal Society of Chemistry, 2007) and schematic model of reduced graphene oxide (rGO)/carbon nanotubes (CNT) hybrid structure (b) with blue and yellow spheres corresponding to carbon atoms of rGO and CNT, respectively, as well as white and red spheres representing hydroxyl group hydrogen and oxygen atoms, respectively (Reproduced with permission of. Copyright Elsevier, 2015).
As one of the possible ways to prevent the sheets from sticking together is a use of 1D material like carbon nanotubes (CNT) as a spacer (see. CNT have a readily accessible surface area as well as high conductivity (see . Since the surface area is known to increase with decrease of the CNT diameter, single-wall carbon nanotubes (SWCNT) are expected to have higher specific capacitance compared to multiwall carbon nanotubes (MWCNT). However, CNT often stack in bundles and only the outermost portion of CNT can function for ion absorption, whereas the inner carbon atoms are not involved in the process. From another side, pristine CNT with preserved electronic structures can be easily dispersed in graphene oxide (GO) solution without any additives and generate clean, electrically addressable carbon-carbon interfaces. Thus, GO is a "surfactant" to directly disperse CNT, while CNT can prevent the aggregation of graphene sheets as a spacer, and has high conductivity, high surface area, and prospective mechanical properties. Moreover, non-conductive GO without any thermal or chemical reduction was found to be a promising material for SC based on a simulation, which claimed its capacitance decreases with increasing oxidation state. Currently, there are a number of publications on G/rGO-CNT composites with different final parameters as promising electrode materials for energy storage applications, many of which are reviewed and analysed in this work.
## General information on energy-storage materials
Capacitors and batteries as well as supercapacitors (electrochemical capacitors) can store a charge, possessing, however, different charge storage mechanisms. The dielectric. Copyright Royal Society of Chemistry, 2007) and schematic model of reduced graphene oxide (rGO)/carbon nanotubes (CNT) hybrid structure (b) with blue and yellow spheres corresponding to carbon atoms of rGO and CNT, respectively, as well as white and red spheres representing hydroxyl group hydrogen and oxygen atoms, respectively (Reproduced with permission of. Copyright Elsevier, 2015).
As one of the possible ways to prevent the sheets from sticking together is a use of 1D material like carbon nanotubes (CNT) as a spacer (see. CNT have a readily accessible surface area as well as high conductivity (see . Since the surface area is known to increase with decrease of the CNT diameter, single-wall carbon nanotubes (SWCNT) are expected to have higher specific capacitance compared to multiwall carbon nanotubes (MWCNT). However, CNT often stack in bundles and only the outermost portion of CNT can function for ion absorption, whereas the inner carbon atoms are not involved in the process. From another side, pristine CNT with preserved electronic structures can be easily dispersed in graphene oxide (GO) solution without any additives and generate clean, electrically addressable carbon-carbon interfaces. Thus, GO is a "surfactant" to directly disperse CNT, while CNT can prevent the aggregation of graphene sheets as a spacer, and has high conductivity, high surface area, and prospective mechanical properties. Moreover, non-conductive GO without any thermal or chemical reduction was found to be a promising material for SC based on a simulation, which claimed its capacitance decreases with increasing oxidation state. Currently, there are a number of publications on G/rGO-CNT composites with different final parameters as promising electrode materials for energy storage applications, many of which are reviewed and analysed in this work. capacitor has electrostatic storage, whereas a battery can be characterized by diffusion mechanism of storage based on reduction-oxidation (redox) processes from used faradaic materials. Moreover, there are clear differences in cyclic voltammograms (CV), providing the current response to a linearly cycled potential sweep, and galvanostatic charge and discharge (GCD) curves, defining how fast a cell is charged or discharged, for these electrodes. Typically, both battery electrodes, anode and cathode, include faradaic materials that result in strong redox peaks in CV, which are clearly visible both for a single electrode and in the full cell (battery) measurements. Moreover, the GCD process of such electrodes as well as devices has long and wide plateaus that can be clearly observed. By contrast with a battery, a dielectric capacitor that stores energy by means of a static charge presents an ideal rectangular shape of CV, the area of which increases with the potential scan rate, and triangular shape of GCD. Moreover, in dielectric capacitors, the current (i) flowing through a cell is proportional to the linear variation rate of the voltage (v) as i~v.
At the same time, existing supercapacitors, which can store the energy by electrochemical reactions, include electric double-layer capacitors (EDLC) and pseudocapacitors as can be seen in. In EDLC, pure electrostatic charge storage occurs at the electrode-electrolyte interface (seeand their almost rectangular CV increases linearly with the scan rate and has symmetric triangular GCD (see, left). In pseudocapacitors, which can involve faradaic materials such as metal oxides or conductive polymers, fast and reversible surface faradaic process such as electron charge-transfer between electrolyte and electrode occurs on/near to the electrode surface (seethat can lead to visible small bulges in CV and small deviations in GCD (see, middle). Moreover, pseudocapacitive electrodes show higher capacitance value than EDLC thanks to the faradaic-electron transfer known as reversible surface redox reactions in addition to the non-faradaic charge storage on the surface observed in EDLC. However, the excessive increase of the faradaic contribution (deep intercalation in electrode material (see can not only increase the capacitance but also dominate diffusion kinetics in the electrodes that is manifested as redox peaks shifted with scan rate in CV patterns, as plateaus in GCD curves (see, right), and in longer charge/discharge time. Thus, according to, the shape of CV can be rectangular-like for EDLC or can show small redox peaks for pseudocapacitive composites including faradaic materials but with dominant EDLC charge storage. Faradaic capacitance electrodes, being also composites of capacitive materials with faradaic material similar to that of battery electrodes, are characterized by distinct and widely separated peaks in CV with an increase of scan rate. GCD curves of single battery-type electrodes are profoundly non-linear and characterized by plateaus of nearly constant potential corresponding to the potentials, at which the faradaic reduction or oxidation of the metal centres, etc. is occurring in contrast to GCD of pseudocapacitor with slight surface redox on the top of the electrode. The total current of CV measurements under a potential sweep rate (i(v)) for composites with faradaic materials consists of two parts. One part is a current related to the double-layer charge at the electrolyte interface or to initial fast faradaic reactions on the exposed electrode surface (icap). Another part is the current related to the slow diffusioncontrolled process (idif). The capacitive contribution and the diffusion-controlled contribution can be calculated following a power-law relationship with the sweep rate (v) according to Equation (1):
[formula] i(v) = icap + idif = a × v b ,(1) [/formula]
where a and b are variable parameters. b values can be estimated from the slope of log(i) vs. log(v) plot. The value b = 1 corresponds to the presence of the fast surface redox reaction and charge/discharge process inherent to EDLC when diffusion contribution is absent and CV show linear current response dependency on the scan rate (i~v). At the same time, the peak current response of a battery-type electrode with strong redox peaks will be proportional to the square root of the scan rate (i~v 1/2 ) and in this case b = 0.5. Thus, the number of the electrodes combining capacitive and faradaic materials reveals 0.5 < b < 1, exhibiting either strong capacitive or battery-type behaviour. Moreover, as also seen from, electrodes with 0.8 < b < 1 are considered as pseudocapacitive materials having predominantly capacitive storagein contrast to the electrodes with 0.5 < b < 0.8 with dominant faradaic (battery-type) behaviour. Thus, to understand the dominant storage mechanism in composite electrode (capacitive or faradaic/battery- Thus, according to, the shape of CV can be rectangular-like for EDLC or can show small redox peaks for pseudocapacitive composites including faradaic materials but with dominant EDLC charge storage. Faradaic capacitance electrodes, being also composites of capacitive materials with faradaic material similar to that of battery electrodes, are characterized by distinct and widely separated peaks in CV with an increase of scan rate. GCD curves of single battery-type electrodes are profoundly non-linear and characterized by plateaus of nearly constant potential corresponding to the potentials, at which the faradaic reduction or oxidation of the metal centres, etc. is occurring in contrast to GCD of pseudocapacitor with slight surface redox on the top of the electrode.
The total current of CV measurements under a potential sweep rate (i(v)) for composites with faradaic materials consists of two parts. One part is a current related to the double-layer charge at the electrolyte interface or to initial fast faradaic reactions on the exposed electrode surface (i cap ). Another part is the current related to the slow diffusion-controlled process (i dif ). The capacitive contribution and the diffusion-controlled contribution can be calculated following a power-law relationship with the sweep rate (v) according to Equation (1):
[formula] i(v) = i cap + i dif = a × v b ,(1) [/formula]
where a and b are variable parameters. b values can be estimated from the slope of log(i) vs. log(v) plot. The value b = 1 corresponds to the presence of the fast surface redox reaction and charge/discharge process inherent to EDLC when diffusion contribution is absent and CV show linear current response dependency on the scan rate (i~v). At the same time, the peak current response of a battery-type electrode with strong redox peaks will be proportional to the square root of the scan rate (i~v 1/2 ) and in this case b = 0.5. Thus, the number of the electrodes combining capacitive and faradaic materials reveals 0.5 < b < 1, exhibiting either strong capacitive or battery-type behaviour. Moreover, as also seen from, electrodes with 0.8 < b < 1 are considered as pseudocapacitive materials having predominantly capacitive storagein contrast to the electrodes with 0.5 < b < 0.8 with dominant faradaic (battery-type) behaviour. Thus, to understand the dominant storage mechanism in composite electrode (capacitive or faradaic/battery-type), the parameter b needs to be calculated in addition to correct and deep analysis of CV at different scan rates, GCD curves, as well as electrochemical impedance spectra (EIS) including the slope of Nyquist plot, etc.
Moreover, since the energy storage cell consists of two electrodes, different combination of electrodes can give a device with different charge storage mechanism as can be seen in. According to, the combination of two EDLC electrodes or two pseudocapacitive electrodes results in a symmetric capacitive device. At the same time, an asymmetric design of storage device, consisting of the EDLC electrode together with the pseudocapacitive one, can gain from the enlarged voltage window covering the windows of two different electrodes, still keeping the capacitive mechanism of charge storage as a dominant one. However, in the case of combining two electrodes with different storage mechanism (EDLC and battery-type or pseudocapacitor and battery-type) this mechanism cannot be kept and the final hybrid device will store energy in a different way to capacitive materials. Based on that, energy and power densities of hybrid devices must be compared directly with that reported for batteries, since their electrochemical performance must be higher in advance compared to that of symmetric and asymmetric supercapacitors because of faradaic materials, which are dominant in the battery-type electrodes. Thus, the charge storage mechanism of single electrode needs to be clarified before fabrication of the energy storage device. The best way to understand it is the detailed analysis of their characteristics, particularly CV, b parameter, GCD, Nyquist plot. Nanomaterials 2021, 11, x FOR PEER REVIEW 5 of 30 type), the parameter b needs to be calculated in addition to correct and deep analysis of CV at different scan rates, GCD curves, as well as electrochemical impedance spectra (EIS) including the slope of Nyquist plot, etc. Moreover, since the energy storage cell consists of two electrodes, different combination of electrodes can give a device with different charge storage mechanism as can be seen in. According to, the combination of two EDLC electrodes or two pseudocapacitive electrodes results in a symmetric capacitive device. At the same time, an asymmetric design of storage device, consisting of the EDLC electrode together with the pseudocapacitive one, can gain from the enlarged voltage window covering the windows of two different electrodes, still keeping the capacitive mechanism of charge storage as a dominant one. However, in the case of combining two electrodes with different storage mechanism (EDLC and battery-type or pseudocapacitor and battery-type) this mechanism cannot be kept and the final hybrid device will store energy in a different way to capacitive materials. Based on that, energy and power densities of hybrid devices must be compared directly with that reported for batteries, since their electrochemical performance must be higher in advance compared to that of symmetric and asymmetric supercapacitors because of faradaic materials, which are dominant in the battery-type electrodes. Thus, the charge storage mechanism of single electrode needs to be clarified before fabrication of the energy storage device. The best way to understand it is the detailed analysis of their characteristics, particularly CV, b parameter, GCD, Nyquist plot.
## Composite capacitive electrodes based on graphene/reduced graphene oxide-carbon nanotubes (g/rgo-cnt)
Both G/rGO and CNT are capacitive materials and their composite electrodes also present EDLC type of energy storage with specific capacitance values varying up to 375 F/gas shown in in Supplementary Materials. Typically, rGO-CNT-based electrodes show not an ideal but almost a rectangular shape of CV with slight deviation due to an electrode resistance contribution, including ion transfer resistance or electrode material resistance, or even some surface redox associated with oxygen functional groups especially at high scan rate.
Enhancement of the electrochemical performance G/rGO-CNT electrodes and, respectively, a final device, can be undertaken not only by addition of external components, but also by modification/functionalization of G/rGO and/or CNT. Thus, the properties of graphene-based materials are known to depend on the synthesis method and reduction process. In particular, for the preparation of mixed G/rGO-CNT-based composites, carbon nanotubes were combined with:
- commercially available graphene or graphene grown previously or directly in the process; - rGO obtained after reduction of GO by high temperature, or by hydrazine hydrate (H 6 N 2 O), ammonium solution (NH 4 OH), vitamin C, acids, etc.
Moreover, rGO-CNT-based electrodes were fabricated both on substrates, such as glassy carbon electrode (GCE), Ni foams, metal foils, stainless steel; carbon-based supports as carbon cloth/carbon paper, or graphite substrates; polyvinylidene fluoride (PVDF)treated paper; Si; plastic; indium tin oxide, etc. and without support as freestanding, i.e., by vacuum filtration, frying or as fibres by injections, etc. Besides such important parameters as the annealing temperature or chemicals used to control the oxygen functional groups content, a ratio between G/rGO and CNT as well as their typology and electrolyte type are also very important. Their influence will be overviewed in the subchapters below. In addition, most of the studies involved not only G/rGO and CNT but also some conductive additives and polymer binders in the similar to commercial battery fabrication process (typically, in the 80:10:10 weight ratio between active material (AM), conductive carbonfamily-material and polymer).
## Temperature effect on g/rgo-cnt-based composite electrodes
The freeze-dried GO aerogels reduced at 180 and at 700 - C were used together with SWCNT for the fabrication and comparative study of rGO-CNT composite electrodes with PVDF binder on carbon cloth by Okhay et al.. X-ray photoelectron spectroscopy (XPS) data of the rGO aerogels shown inpresent a strongly reduced oxygen functional group after heat treatment at 700 - C that is supported by the almost ideal rectangular CV curves for the high temperature-reduced aerogel (seeby contrast with that of rGO aerogel annealed at 180 - C that still show a bulge of surface redox on EDLC (see. Moreover, the specific capacitance of 129 F/g reported for the electrodes made of 180 - C rGO is much higher than that of 41 F/g at 0.1 A/g for 700 - C rGO.
A systematic study of temperature influence on functionalized graphene nanosheets/ carbon nanotubes networks (G/CNT) synthesized through chemical oxidation of CNT followed by thermal reduction according to schematic illustration inwas reported by Ding et al.. The external walls of CNT in the G/CNT structure annealed at 200 - C were observed as unzipped and transformed into functionalized graphene nanosheets, while the inner walls were not unzipped and kept the tubular structure. Freeze-dried G/CNT mixture was heat treated at 200, 300, 600 and 800 - C for 2 h under N 2 flow and mixed with carbon black (CB) and poly (tetrafluoroethylene) (PTFE) before the covering on Ni foam. . Carbon (C1s) (a,c) and oxygen (O1s) (b,d) spectra obtained by X-ray photoelectron spectroscopy (XPS) on graphene oxide (GO) aerogel reduced at 180 °C in vacuum (a,b) and at 700 °C in Ar (c,d). Cyclic voltammetry profiles at different scan rates for the rGO-CNT-based composite electrodes on carbon cloth with 180 °C (e) and at 700 °C (f) rGO (Reproduced with permission of. Copyright Elsevier, 2020). Based on Raman spectra, shown in, the ID/IG ratio was found to increase after chemical oxidation from 0.93 to 1.26, revealing that a large number of oxygen functional groups were introduced during the unzipping process. However, after thermal reduction at 200 °C partially unstable oxygen functional groups were removed as can be also seen in, leading to an increase of the integrated area of CV in comparison with as-prepared material (see. The fact that after thermal reduction of G/CNT at 200 °C conjugated carbon networks were restored was also mentioned. However, after thermal reduction at T > 200 °C (seethe integrated area of CV was also suppressed (seedespite the fact that unstable oxygen functional groups were progressively reduced. Therefore, the highest specific capacitance of 202 F/g at 0.5 A/g was reported for electrodes with G/CNT after heat treatment at 200 °C (see. Moreover, CV curves of G/CNTs and G/CNTs-200 electrodes show obvious redox humps in, meaning that the capacitances come both from EDLC and pseudocapacitance due to the reversible redox reactions among the surface oxygen functional groups. . Carbon (C1s) (a,c) and oxygen (O1s) (b,d) spectra obtained by X-ray photoelectron spectroscopy (XPS) on graphene oxide (GO) aerogel reduced at 180 - C in vacuum (a,b) and at 700 - C in Ar (c,d). Cyclic voltammetry profiles at different scan rates for the rGO-CNT-based composite electrodes on carbon cloth with 180 - C (e) and at 700 - C (f) rGO (Reproduced with permission of. Copyright Elsevier, 2020). Based on Raman spectra, shown in, the I D /I G ratio was found to increase after chemical oxidation from 0.93 to 1.26, revealing that a large number of oxygen functional groups were introduced during the unzipping process. However, after thermal reduction at 200 - C partially unstable oxygen functional groups were removed as can be also seen in, leading to an increase of the integrated area of CV in comparison with asprepared material (see. The fact that after thermal reduction of G/CNT at 200 - C conjugated carbon networks were restored was also mentioned. However, after thermal reduction at T > 200 - C (seethe integrated area of CV was also suppressed (seedespite the fact that unstable oxygen functional groups were progressively reduced. Therefore, the highest specific capacitance of 202 F/g at 0.5 A/g was reported for electrodes with G/CNT after heat treatment at 200 - C (see. Moreover, CV curves of G/CNTs and G/CNTs-200 electrodes show obvious redox humps in, meaning that the capacitances come both from EDLC and pseudocapacitance due to the reversible redox reactions among the surface oxygen functional groups. tance due to the reversible redox reactions among the surface oxygen functional groups. , as well as G/CNT annealed at 300, 600 and 800 - C (G/CNTs-300, G/CNTs-600, and G/CNTs-800, respectively) (d). Specific capacitance of rGO, CNTs, G/CNTs, G/CNTs-200, G/CNTs-300, G/CNTs-600 and G/CNTs-800 as a function of current density (e) (Reproduced with permission of. Copyright Elsevier, 2018).
## Effect of cnt length
The size of CNT can change the value of specific capacitance at least 3 times as was reported by Zeng et al.. To increase the utilization of closed pore volumes of CNT and prevent the stacking of rGO, multiwall CNT (MWCNT) are tailored into super short CNT (SSCNT) with aspect ratio of less than 5 by an ultrasonic oxidation-cut method (see. After mixing SSCNT with rGO and PVDF, it was coated onto titanium plate. The morphologies of the composite with MWCNTs and SSCNT were found very different by scanning electron microscopy (SEM) as shown in,c. The length reduction from 5-15 µm to 10-300 nm led to rich structural features, such as nanoscale length, open ends, abundant carbon atoms on the edge, quasi-0D characteristic and so on. Moreover, the introduction of SSCNT increased the specific surface area of rGO to 370 m 2 /g that was higher than 171 m 2 /g after addition of MWCNT (see. Thus, shorter CNT can be more uniformly distributed on the surface of rGO and form 3D multilayer architecture that leads to an increase of specific capacitance from 88 F/g to 251 F/g at 50 mV/s.
phologies of the composite with MWCNTs and SSCNT were found very different by scan-ning electron microscopy (SEM) as shown in,c. The length reduction from 5-15 μm to 10-300 nm led to rich structural features, such as nanoscale length, open ends, abundant carbon atoms on the edge, quasi-0D characteristic and so on. Moreover, the introduction of SSCNT increased the specific surface area of rGO to 370 m 2 /g that was higher than 171 m 2 /g after addition of MWCNT (see. Thus, shorter CNT can be more uniformly distributed on the surface of rGO and form 3D multilayer architecture that leads to an increase of specific capacitance from 88 F/g to 251 F/g at 50 mV/s. Scanning electron microscope (SEM) images of raw multiwall CNT (MWCNT) (b) and SSCNT (c). Nitrogen sorption isotherms obtained at 77 K (d) (Reproduced with permission of. Copyright Elsevier, 2013).
## Cnt concentration in g/rgo-cnt-based composite electrodes
Based on results reported by Lu et al.wt.% CNT was the optimal concentration in the range 0 ÷ 50 wt.% for rGO-CNT freestanding electrodes to obtain the highest capacitance of 265 F/g at 0.1 A/g. A lower content of 10 wt.% CNT was chosen by Kumar et al. for dried composite pressed into Ni foam as the optimal concentration of CNT to obtain the rectangular-like CV (seeand the highest specific capacitance (see. Among the studied CNT concentrations from 0 to 66.7 wt.%, 12.5 wt.% CNT was used to get the highest capacitance for rGO-CNT obtained by vacuum filtration (VF) and pressed into Ni foam. Specific capacitance of 132 F/g was reported by Lee et al. for Scanning electron microscope (SEM) images of raw multiwall CNT (MWCNT) (b) and SSCNT (c). Nitrogen sorption isotherms obtained at 77 K (d) (Reproduced with permission of. Copyright Elsevier, 2013).
## Cnt concentration in g/rgo-cnt-based composite electrodes
Based on results reported by Lu et al.wt.% CNT was the optimal concentration in the range 0 ÷ 50 wt.% for rGO-CNT freestanding electrodes to obtain the highest capacitance of 265 F/g at 0.1 A/g. A lower content of 10 wt.% CNT was chosen by Kumar et al. for dried composite pressed into Ni foam as the optimal concentration of CNT to obtain the rectangular-like CV (seeand the highest specific capacitance (see. Among the studied CNT concentrations from 0 to 66.7 wt.%, 12.5 wt.% CNT was used to get the highest capacitance for rGO-CNT obtained by vacuum filtration (VF) and pressed into Ni foam. Specific capacitance of 132 F/g was reported by Lee et al. for rGO-CNT with 11 wt.% CNT prepared on glassy carbon electrode among other composites with CNT concentration varied from 6 to 50 wt.%. Thus, the optimal amount of CNT for rGO-CNT mixed composites was found to be from 10 to 16 wt.%. However, in the case of preparation of a CNT layered structure onto graphite paper by electrophoretic deposition, significantly higher content of 40 wt.% CNT in suspension was used to obtain the highest capacitance in that work of 87 F/g at 5 mV/s. rGO-CNT with 11 wt.% CNT prepared on glassy carbon electrode among other composites with CNT concentration varied from 6 to 50 wt.%. Thus, the optimal amount of CNT for rGO-CNT mixed composites was found to be from 10 to 16 wt.%. However, in the case of preparation of a CNT layered structure onto graphite paper by electrophoretic deposition, significantly higher content of 40 wt.% CNT in suspension was used to obtain the highest capacitance in that work of 87 F/g at 5 mV/s.
## Influence of electrolyte type and potential window
Based on the published research data, G/rGO-CNT-based electrodes were studied in both aqueous and non-aqueous electrolytes as well as in liquid and solid states. Aqueous electrolytes are less expensive, not flammable, and not/less toxic in contrast to e.g., organic electrolytes used in commercially available batteries/SC. Many researchers used aqueous liquid and solid electrolytes: acids such as H2SO4 or H3PO4 or alkaline KOH or neutral Na2SO4, KCl, LiClO4, Li2SO4. Non-aqueous electrolytes were reported in the several publications as organic electrolytes TEABF4 (tetraethylammonium tetrafluoroborate) and Et4NBF4-AN (tetraethylammonium tetrafluoroborate-acetonitrile), and as ionic liquids EMIM-BF4 (1-ethyl-3-methylimidazolium tetrafluoroborate) and EMI-TFSI (1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide). However, the results obtained, particularly CV form and size, indicate different behaviour of G/rGO-CNT in different electrolytes. Cheng et al. studied supercapacitor fabricated on freestanding rGO-CNT in neutral electrolyte KCl, organic electrolyte TEABF4 in propylene carbonate (PC) and ionic liquid EMI-TFSI. CV curve in the organic electrolyte TEABF4/PC does not exhibit rectangular geometry due to the larger resistance in the organic electrolyte, redox groups such as hydroxide group and carboxyl (seeby contrast with CV in KCl (seeand in EMI-TFSI. This corresponds to GCD curves that have shown irregular shape in organic electrolyte (seebut triangular form in KCl (see.
## Influence of electrolyte type and potential window
Based on the published research data, G/rGO-CNT-based electrodes were studied in both aqueous and non-aqueous electrolytes as well as in liquid and solid states. Aqueous electrolytes are less expensive, not flammable, and not/less toxic in contrast to e.g., organic electrolytes used in commercially available batteries/SC. Many researchers used aqueous liquid and solid electrolytes: acids such as H 2 SO 4 or H 3 PO 4 or alkaline KOH or neutral Na 2 SO 4 , KCl, LiClO 4 , Li 2 SO 4 . Non-aqueous electrolytes were reported in the several publications as organic electrolytes TEABF 4 (tetraethylammonium tetrafluoroborate) and Et 4 NBF 4 -AN (tetraethylammonium tetrafluoroborate-acetonitrile), and as ionic liquids EMIM-BF 4 (1-ethyl-3-methylimidazolium tetrafluoroborate) and EMI-TFSI (1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide). However, the results obtained, particularly CV form and size, indicate different behaviour of G/rGO-CNT in different electrolytes. Cheng et al. studied supercapacitor fabricated on freestanding rGO-CNT in neutral electrolyte KCl, organic electrolyte TEABF 4 in propylene carbonate (PC) and ionic liquid EMI-TFSI. CV curve in the organic electrolyte TEABF 4 /PC does not exhibit rectangular geometry due to the larger resistance in the organic electrolyte, redox groups such as hydroxide group and carboxyl (seeby contrast with CV in KCl (seeand in EMI-TFSI. This corresponds to GCD curves that have shown irregular shape in organic electrolyte (seebut triangular form in KCl (see. A near rectangular shape of CV and slight asymmetry in GCD curves was observed by Kumar et al. for filtered out rGO-CNT pressed into Ni foam and studied in such electrolytes as KOH, NaOH and LiOH (see. Small deviation from the ideal rectangular CV and triangular GCD curves was explained by the occurrence of some faradaic reaction at the surface ascribed to the oxygen-containing functional groups attached to rGO sheets and functionalized CNT. As shown in, the highest integral area and more rectangular CV curves were observed during the test of electrodes in KOH that can be associated with a smaller hydrated ionic radius and higher ionic conductivity of K + ion in comparison to that of Na + and Li + ions. On the other hand, the ionic mobility enhanced by a lower hydrated ionic radius of K + ion gains access to the electrode surface, resulting in an improved electrochemical performance of rGO-CNT electrode. A near rectangular shape of CV and slight asymmetry in GCD curves was observed by Kumar et al. for filtered out rGO-CNT pressed into Ni foam and studied in such electrolytes as KOH, NaOH and LiOH (see. Small deviation from the ideal rectangular CV and triangular GCD curves was explained by the occurrence of some faradaic reaction at the surface ascribed to the oxygen-containing functional groups attached to rGO sheets and functionalized CNT. As shown in, the highest integral area and more rectangular CV curves were observed during the test of electrodes in KOH that can be associated with a smaller hydrated ionic radius and higher ionic conductivity of K + ion in comparison to that of Na + and Li + ions. On the other hand, the ionic mobility enhanced by a lower hydrated ionic radius of K + ion gains access to the electrode surface, resulting in an improved electrochemical performance of rGO-CNT electrode. Cui et al. reported that the CV shape of rGO-CNT composite studied in a positive potential window is different from that tested in a negative potential window (see. CV curves of rGO-CNT coated on Ti foil with irregular shape observed from 0 to +0.8 V as well as from −0.4 V to +0.4 V both in Na2SO4 (seeand NaCl electrolyte (seeindicated the existence of EDLC and pseudocapacitance in opposite to almost rectangular CV in potential window from 0 to −0.8 V. Moreover, integrated areas of CV in Na2SO4 were obviously higher than that in NaCl. Correspondingly, the specific capacitance was also different and the highest value was obtained for the composite tested in a negative potential window from 0 to −0.8 V in Na2SO4. Thus, according to Cui et al. the rGO-CNT electrode has shown great potential to be used as a negative electrode for energy-storage devices. Cui et al. reported that the CV shape of rGO-CNT composite studied in a positive potential window is different from that tested in a negative potential window (see. CV curves of rGO-CNT coated on Ti foil with irregular shape observed from 0 to +0.8 V as well as from −0.4 V to +0.4 V both in Na 2 SO 4 (seeand NaCl electrolyte (seeindicated the existence of EDLC and pseudocapacitance in opposite to almost rectangular CV in potential window from 0 to −0.8 V. Moreover, integrated areas of CV in Na 2 SO 4 were obviously higher than that in NaCl. Correspondingly, the specific capacitance was also different and the highest value was obtained for the composite tested in a negative potential window from 0 to −0.8 V in Na 2 SO 4 . Thus, according to Cui et al. the rGO-CNT electrode has shown great potential to be used as a negative electrode for energy-storage devices. Cui et al. reported that the CV shape of rGO-CNT composite studied in a positive potential window is different from that tested in a negative potential window (see. CV curves of rGO-CNT coated on Ti foil with irregular shape observed from 0 to +0.8 V as well as from −0.4 V to +0.4 V both in Na2SO4 (seeand NaCl electrolyte (seeindicated the existence of EDLC and pseudocapacitance in opposite to almost rectangular CV in potential window from 0 to −0.8 V. Moreover, integrated areas of CV in Na2SO4 were obviously higher than that in NaCl. Correspondingly, the specific capacitance was also different and the highest value was obtained for the composite tested in a negative potential window from 0 to −0.8 V in Na2SO4. Thus, according to Cui et al. the rGO-CNT electrode has shown great potential to be used as a negative electrode for energy-storage devices.
## Modified g/rgo-cnt electrodes with faradaic contribution
## Nitrogen doping
One of the popular directions today is modification of G/rGO-CNT by nitrogen (N) due to its atomic size and strong valence bonds, which are similar to those characteristics of carbon atoms. Pyrolysis of GO with a low-cost N source is a versatile method for large-scale production of N-doped graphene with flexible control over the N-bonding configurations. N-doped G/rGO-CNT structures on different substrates were reported as high-performance supercapacitor electrode materials. Different nitrogen-containing materials such as polyacrylonitrile, acetonitrile, melamine, etc. are commonly used as the nitrogen precursor.
Significant enhancement of the specific capacitance of rGO-CNT composites on GCE after N-doping was reported by Lin et al., when initial value of 10 F/g obtained for rGO-CNT has grown to 168 F/g at 0.5 A/g for N-doped rGO-CNT (designated as NGC) after addition of urea and low-cost lignosulfonate (LS) (see. Adding only LS to the mixture of GO and CNTs (designated as LGC), the obtained graphene sheets in LGC composite were very thick (see.
## Modified g/rgo-cnt electrodes with faradaic contribution
## Nitrogen doping
One of the popular directions today is modification of G/rGO-CNT by nitrogen (N) due to its atomic size and strong valence bonds, which are similar to those characteristics of carbon atoms. Pyrolysis of GO with a low-cost N source is a versatile method for large-scale production of N-doped graphene with flexible control over the N-bonding configurations. N-doped G/rGO-CNT structures on different substrates were reported as high-performance supercapacitor electrode materials. Different nitrogen-containing materials such as polyacrylonitrile, acetonitrile, melamine, etc. are commonly used as the nitrogen precursor.
Significant enhancement of the specific capacitance of rGO-CNT composites on GCE after N-doping was reported by Lin et al., when initial value of 10 F/g obtained for rGO-CNT has grown to 168 F/g at 0.5 A/g for N-doped rGO-CNT (designated as NGC) after addition of urea and low-cost lignosulfonate (LS) (see. Adding only LS to the mixture of GO and CNTs (designated as LGC), the obtained graphene sheets in LGC composite were very thick (see. The result may be attributed to the three-dimensional structure of macromolecular LS. However, after the addition of urea into the mixture of N-doped GO and CNTs, the graphene sheets in NGC were found to be thinner and looser. Moreover, more porous Ndoped rGO-CNT (PNGC) obtained after addition of both LS and urea with further heat treatment at 800 °C have shown CV with the highest integrated area that corresponds to the highest capacitance of 246 F/g at 0.5 A/g (see in Supplementary Materials for details). During the annealing process, excess urea molecules decompose abundant gases, which open the space between graphene sheets and prevent graphene sheets from stacking tightly. Furthermore, after the addition of LS and urea into the mixture of GO and CNTs, many pores on the surface of thin graphene sheets in PNGC are observed, as shown by arrows in. In addition, the CV curve of rGO-CNT shows a pair of redox peaks, which may be attributed to the residual carboxyl and hydroxyl groups of CNT via oxidation process. CV curves of NGC and PNGC can also be seen to exhibit nearly rectangular shapes and have some peaks as well, ascribed to the combination of electrical double-layer capacitance and faradic pseudocapacitance from nitrogen doping and residual The result may be attributed to the three-dimensional structure of macromolecular LS. However, after the addition of urea into the mixture of N-doped GO and CNTs, the graphene sheets in NGC were found to be thinner and looser. Moreover, more porous N-doped rGO-CNT (PNGC) obtained after addition of both LS and urea with further heat treatment at 800 - C have shown CV with the highest integrated area that corresponds to the highest capacitance of 246 F/g at 0.5 A/g (see in Supplementary Materials for details). During the annealing process, excess urea molecules decompose abundant gases, which open the space between graphene sheets and prevent graphene sheets from stacking tightly. Furthermore, after the addition of LS and urea into the mixture of GO and CNTs, many pores on the surface of thin graphene sheets in PNGC are observed, as shown by arrows in. In addition, the CV curve of rGO-CNT shows a pair of redox peaks, which may be attributed to the residual carboxyl and hydroxyl groups of CNT via oxidation process. CV curves of NGC and PNGC can also be seen to exhibit nearly rectangular shapes and have some peaks as well, ascribed to the combination of electrical double-layer capacitance and faradic pseudocapacitance from nitrogen doping and residual carboxyl and hydroxyl groups. Close capacitance value of 176 F/g at 0.5 A/g was reported for composite electrodes on Ni foam made of N-doped rGO-CNT by addition of polydopamide (PDA), acetylene black (AB) and PVDF.
## Addition of conductive polymers
## G/rgo-cnt with polypyrrole
Polypyrrole (PPy) has been extensively studied by many research groups due to its particular advantages with regard to low cost, environmental friendliness, high capacitive capability and easy processing. Typically G/rGO-CNT-PPy composite electrodes were fabricated by the in situ polymerization method. Pseudocapacitive composites of rGO-CNT with PPy were obtained by Wang et al. as fibre electrodeand by Lu et al. as freestanding electrodeas well as a composite with PTFE onto graphite substrate. In the work of Wang et al., GO-CNT fibres (seereduced by vitamin C at 90 - C have shown specific capacitance of 10.8 F/cm 3 at 0.01 V/s in LiCl electrolyte that was increased up to 25.9 F/cm 3 after covering by PPy(see in Supplementary Materials for details). carboxyl and hydroxyl groups. Close capacitance value of 176 F/g at 0.5 A/g was reported for composite electrodes on Ni foam made of N-doped rGO-CNT by addition of polydopamide (PDA), acetylene black (AB) and PVDF.
## Addition of conductive polymers
## G/rgo-cnt with polypyrrole
Polypyrrole (PPy) has been extensively studied by many research groups due to its particular advantages with regard to low cost, environmental friendliness, high capacitive capability and easy processing. Typically G/rGO-CNT-PPy composite electrodes were fabricated by the in situ polymerization method. Pseudocapacitive composites of rGO-CNT with PPy were obtained by Wang et al. as fibre electrodeand by Lu et al. as freestanding electrodeas well as a composite with PTFE onto graphite substrate. In the work of Wang et al., GO-CNT fibres (seereduced by vitamin C at 90 °C have shown specific capacitance of 10.8 F/cm 3 at 0.01 V/s in LiCl electrolyte that was increased up to 25.9 F/cm 3 after covering by PPy(see in Supplementary Materials for details). At the same time Lu et al. measured composite made of rGO, poly(sodium 4-sterene sulfonate) (PSS) functionalized CNT and PPy prepared as freestanding electrodeand as electrode on graphite substrate mixed with CB and PTFE (see. PSS containing a hydrophilic group (-SO3) was demonstrated to be strongly and uniformly adsorbed on the surface of rGO-CNT during the modified process that leads to high stability and dispersion of the functionalized rGO-CNT within the aqueous solution. Simultaneously, the sulfonic groups with negative charges extending in the solution provide a number of coordinating sites onto rGO-CNT surface. Such coordinating sites can be used to effectively tether and absorb more monomer PPy and facilitate the following "homogeneous" deposition of PPy particles on the electrode surface. Corresponding rGO-PSS-CNT-based composite electrodes presented capacitive behaviour with specific values of 211 F/g and 361 F/g at 0.2 A/g for freestanding and graphite substrate supported electrode, respectively (see in Supplementary Materials for details). That could be explained by an increasing amount of PPy from ~40 wt.% for the freestanding electrode to more than 70 wt.% that on graphite substrate. In the case of further PPy concentration increase the specific capacitance can continue to grow to 453 F/g at 5 mV/s as reported by Aphale et al. for rGO-CNT-PPy electrode using more than 99 wt.% PPy.
## G/rgo-cnt with polyaniline
Polyaniline (PANI) is a main conductive polymer with high environmental stability, redox reversibility, electroactivity and unusual doping/de-doping chemistry. PANI as a At the same time Lu et al. measured composite made of rGO, poly(sodium 4-sterene sulfonate) (PSS) functionalized CNT and PPy prepared as freestanding electrodeand as electrode on graphite substrate mixed with CB and PTFE (see. PSS containing a hydrophilic group (-SO 3 ) was demonstrated to be strongly and uniformly adsorbed on the surface of rGO-CNT during the modified process that leads to high stability and dispersion of the functionalized rGO-CNT within the aqueous solution. Simultaneously, the sulfonic groups with negative charges extending in the solution provide a number of coordinating sites onto rGO-CNT surface. Such coordinating sites can be used to effectively tether and absorb more monomer PPy and facilitate the following "homogeneous" deposition of PPy particles on the electrode surface. Corresponding rGO-PSS-CNT-based composite electrodes presented capacitive behaviour with specific values of 211 F/g and 361 F/g at 0.2 A/g for freestanding and graphite substrate supported electrode, respectively (see in Supplementary Materials for details). That could be explained by an increasing amount of PPy from~40 wt.% for the freestanding electrode to more than 70 wt.% that on graphite substrate. In the case of further PPy concentration increase the specific capacitance can continue to grow to 453 F/g at 5 mV/s as reported by Aphale et al. for rGO-CNT-PPy electrode using more than 99 wt.% PPy.
## G/rgo-cnt with polyaniline
Polyaniline (PANI) is a main conductive polymer with high environmental stability, redox reversibility, electroactivity and unusual doping/de-doping chemistry. PANI as a component of rGO-CNT composite can initiate the pseudocapacitance from the faradaic contribution of its redox nature that, together with EDLC of rGO-CNT, leads to the electrode capacitance enhancement. Typically, PANI can be obtained by an in situ polymerization process using the dissolved aniline monomer. By this method the total surface of freestanding rGO-CNT paperand fibre electrodewas covered by PANI that resulted in the capacitance of 138 F/g at 0.2 A/g and 193 F/cm 3 at 1 A/cm 3 , respectively (see in Supplementary Materials for details). Higher specific capacitance of 359 F/g at 1 A/g was obtained by Huang et al. for electrodes prepared by mixing hydrazinereduced GO, CNT and aniline to obtain a composite with 80 wt.% PANIand with larger intercalation compared with PANI coating just mentioned above. Then slight surface redox including response from PANI can be observed in CV curves but the corresponding peaks are symmetrical and do not shift with the increasing scan rate (see. That fact together with symmetric and triangular CGD curves (seesupported the dominant capacitive behaviour in these electrodes. In addition, according to our estimation of the b parameter for these electrodes it was found to be~0.9 that is close to b = 1 associated with capacitor behaviour. component of rGO-CNT composite can initiate the pseudocapacitance from the faradaic contribution of its redox nature that, together with EDLC of rGO-CNT, leads to the electrode capacitance enhancement. Typically, PANI can be obtained by an in situ polymerization process using the dissolved aniline monomer. By this method the total surface of freestanding rGO-CNT paperand fibre electrodewas covered by PANI that resulted in the capacitance of 138 F/g at 0.2 A/g and 193 F/cm 3 at 1A/cm 3 , respectively (see in Supplementary Materials for details). Higher specific capacitance of 359 F/g at 1 A/g was obtained by Huang et al. for electrodes prepared by mixing hydrazinereduced GO, CNT and aniline to obtain a composite with 80 wt.% PANIand with larger intercalation compared with PANI coating just mentioned above. Then slight surface redox including response from PANI can be observed in CV curves but the corresponding peaks are symmetrical and do not shift with the increasing scan rate (see. That fact together with symmetric and triangular CGD curves (seesupported the dominant capacitive behaviour in these electrodes. In addition, according to our estimation of the b parameter for these electrodes it was found to be ~0.9 that is close to b = 1 associated with capacitor behaviour. At the same time, similar mixture of hydrazine reduced GO (marked as GNS in, CNT and aniline, to obtain final composite including also CB and PTFE but with fraction of PANI lowered to ~64 wt.%, was found to present significantly higher capacitance of 1035 F/g at 1 mV/s. However, the appearance of strong redox peaks in CV, shifted with the increasing scan rate (see, and far from symmetrical triangular GCD curve shapes, approaching that with plateau (see, are expected not for materials with capacitive storage mechanism but rather for faradaic materials.
Moreover, also high specific capacitance of 987 F/g at 0.5 A/g was reported by Tran et al. for rGO-CNT-PANI prepared by the hydrothermal (HT) method at 180 °C and mixed with mesoporous carbon (MC) and Nafion (with PANI content lowered to ~26 wt.% in final composite) before covering onto carbon paper, while value of 638 F/g at 0.5 A/g was measured by Liu et al. for freestanding electrodes fabricated by mixing and filtration of CNT with graphene nanosheets already covered by PANI in an autoclave at 250 °C to form nanorods (with >50 wt.% PANI in composite). However, these rGO-CNT-PANI composites with high specific capacitance have shown b value much lower than 1, being thus not associated with capacitive behaviour in pseudocapacitive materials. In addition, the slope of the EIS Nyquist plot presented by Tran et al. for rGO-CNT-PANI-MC-Nafion electrode on carbon paper was closer to 45° than to 90°which means a strong faradaic At the same time, similar mixture of hydrazine reduced GO (marked as GNS in, CNT and aniline, to obtain final composite including also CB and PTFE but with fraction of PANI lowered to~64 wt.%, was found to present significantly higher capacitance of 1035 F/g at 1 mV/s. However, the appearance of strong redox peaks in CV, shifted with the increasing scan rate (see, and far from symmetrical triangular GCD curve shapes, approaching that with plateau (see, are expected not for materials with capacitive storage mechanism but rather for faradaic materials.
Moreover, also high specific capacitance of 987 F/g at 0.5 A/g was reported by Tran et al. for rGO-CNT-PANI prepared by the hydrothermal (HT) method at 180 - C and mixed with mesoporous carbon (MC) and Nafion (with PANI content lowered to~26 wt.% in final composite) before covering onto carbon paper, while value of 638 F/g at 0.5 A/g was measured by Liu et al. for freestanding electrodes fabricated by mixing and filtration of CNT with graphene nanosheets already covered by PANI in an autoclave at 250 - C to form nanorods (with >50 wt.% PANI in composite). However, these rGO-CNT-PANI composites with high specific capacitance have shown b value much lower than 1, being thus not associated with capacitive behaviour in pseudocapacitive materials. In addition, the slope of the EIS Nyquist plot presented by Tran et al. for rGO-CNT-PANI-MC-Nafion electrode on carbon paper was closer to 45 - than to 90 - which means a strong faradaic contribution in the analysed electrodes. Thus, rGO-CNT-PANI electrodes reported by Liu et al., Tran et al., and Yan et al.(see in Supplementary Materials for details) have shown a dominant diffusion-controlled mechanism of energy storage that explains the obtained high value of specific capacitance., and Yan et al.(see in Supplementary Materials for details) have shown a dominant diffusion-controlled mechanism of energy storage that explains the obtained high value of specific capacitance.
## Influence of metal catalysts, metal oxides and hydroxides
## G/cnt grown with me-catalysts
Co, Mo, Al/Fe2O3, Au and other metals were reported as catalysts used for G or CNT growth for G/CNT electrodes. Seo et al. formed vertical graphene nanosheets (VGNS) by the plasma transformation of a commercially available natural precursor butter as illustrated in. The plasma was essential in the process to break down the carbon-containing molecules in butter and reconstruct them into ordered and vertical graphitic structures (see. The growth of CNT was then performed in a thermal chemical vapour deposition (CVD) process after the deposition of a Co/Mo catalyst on VGNS. The as-grown VGNS/CNTs hybrid structure on a flexible graphite substrate is presented in. The SEM and transmission electron microscopy (TEM) images of pure VGNS and the VGNS-CNTs obtained after the direct growth process are shown in,f. An inherently open, 3D network with dense and uniform graphene nanosheets was clearly observed to cover the entire surface of the graphite paper. Measured VGNS-CNTs electrodes have shown a specific capacitance of 278 F/g at 10 mV/s (see in Supplementary Materials for details) and CV curves with typical shape for EDLC material without redox peaks from Co and Mo used as catalysts. Moreover, neither Co nor Mo was detected by XPS analysis of this structure. GCD curves and Nyquist plot also indicated capacitive energy storage mechanism in the current electrode with the negligible electrochemical contributions of Co and Mo nanoparticles.
At the same time, Fan et al. mixed GO with Co(NO3)2 before growing vertical CNT by CVD at 750 °C with Fe/Al2O3 as catalyst (see. In this case the sandwich structure was reported with vertical CNT grown between graphene sheets as can be seen inas well as nascent plateau in GCD shown inwere visible suggesting the high pseudocapacitance of cobalt hydroxide that resulted in measured specific capacitance of 385 F/g at 10 mV/s.
## Influence of metal catalysts, metal oxides and hydroxides
## G/cnt grown with me-catalysts
Co, Mo, Al/Fe 2 O 3 , Au and other metals were reported as catalysts used for G or CNT growth for G/CNT electrodes. Seo et al. formed vertical graphene nanosheets (VGNS) by the plasma transformation of a commercially available natural precursor butter as illustrated in. The plasma was essential in the process to break down the carbon-containing molecules in butter and reconstruct them into ordered and vertical graphitic structures (see. The growth of CNT was then performed in a thermal chemical vapour deposition (CVD) process after the deposition of a Co/Mo catalyst on VGNS. The as-grown VGNS/CNTs hybrid structure on a flexible graphite substrate is presented in. The SEM and transmission electron microscopy (TEM) images of pure VGNS and the VGNS-CNTs obtained after the direct growth process are shown in,f. An inherently open, 3D network with dense and uniform graphene nanosheets was clearly observed to cover the entire surface of the graphite paper. Measured VGNS-CNTs electrodes have shown a specific capacitance of 278 F/g at 10 mV/s (see in Supplementary Materials for details) and CV curves with typical shape for EDLC material without redox peaks from Co and Mo used as catalysts. Moreover, neither Co nor Mo was detected by XPS analysis of this structure. GCD curves and Nyquist plot also indicated capacitive energy storage mechanism in the current electrode with the negligible electrochemical contributions of Co and Mo nanoparticles.
At the same time, Fan et al. mixed GO with Co(NO 3 ) 2 before growing vertical CNT by CVD at 750 - C with Fe/Al 2 O 3 as catalyst (see. In this case the sandwich structure was reported with vertical CNT grown between graphene sheets as can be seen in,c. In opposite to work by Seo et al. where Co/Mo catalyst were not detected by XPS or in an electrochemical study, Fan et al. have shown visible Co-based catalysts resided at the top of CNT (see. Moreover, strong redox peaks in CV curves shown inas well as nascent plateau in GCD shown inwere visible suggesting the high pseudocapacitance of cobalt hydroxide that resulted in measured specific capacitance of 385 F/g at 10 mV/s. Very interesting results were obtained by Li et al. for electrodes made on the coreshell structure of G grown on CNT preliminarily covered by Au nanoparticles as catalyst (CNT@Au).shows a schematic diagram and obtained structures at various stages during the formation of CNT@Au composite as a function of the deposition time. Elemental Au originating from the catalyst nanoparticles was also detected by XPS. CNT@G powder with graphene growth time of 5 min being pressed into Ni foam exhibited the largest CV with redox peaks and, correspondingly, the highest specific capacitance in comparison to other CNT@G. However, the reported value of the capacitance was strongly dependent on at least two factors such as the mass load and the width of the used potential window.illustrates that the integral area became significantly larger, but the redox peaks associated with Au catalyst became inconspicuous with increasing mass loading. Based on CV measured from −1 V to +1 V presented inthe gravimetric (C m ) and areal (C a ) capacitance values at different scanning rates for CNT@G electrodes with different mass loadings were deduced (see. The highest gravimetric (or specific) capacitance of 218 F/g was obtained for the electrode with the lowest G@CNT mass loading of 0.5 mg/cm 2 at 10 mV/s, but the highest areal capacitance of 281 mF/cm 2 was obtained for the highest studied mass loading of 5-6 mg/cm 2 also at 10 mV/s. Very interesting results were obtained by Li et al. for electrodes made on the coreshell structure of G grown on CNT preliminarily covered by Au nanoparticles as catalyst (CNT@Au).shows a schematic diagram and obtained structures at various stages during the formation of CNT@Au composite as a function of the deposition time. Elemental Au originating from the catalyst nanoparticles was also detected by XPS. CNT@G powder with graphene growth time of 5 min being pressed into Ni foam exhibited the largest CV with redox peaks and, correspondingly, the highest specific capacitance in comparison to other CNT@G. However, the reported value of the capacitance was strongly dependent on at least two factors such as the mass load and the width of the used potential window.illustrates that the integral area became significantly larger, but the redox peaks associated with Au catalyst became inconspicuous with increasing mass loading. Based on CV measured from −1 V to +1 V presented inthe gravimetric (Cm) and areal (Ca) capacitance values at different scanning rates for CNT@G electrodes with different mass loadings were deduced (see. The highest gravimetric (or specific) capacitance of 218 F/g was obtained for the electrode with the lowest G@CNT mass loading of 0.5 mg/cm 2 at 10 mV/s, but the highest areal capacitance of 281 mF/cm 2 was obtained for the highest studied mass loading of 5-6 mg/cm 2 also at 10 mV/s. Regarding the potential window effect, Li et al. studied it on CNT@G electrodes with mass loadings of 3 mg/cm 2 and 5 mg/cm 2 . A CV curve example at a scanning rate of 20 mV/s for 3 mg/cm 2 mass loading can be seen in. CV curves of both electrodes show a pair of redox peaks in the negative potential range, and an additional pair of redox peaks appears with an increasing potential window on the positive side. These peaks are related to the trace amount of Au catalyst distributed on the graphene sheets. In addition, both kinds of the specific capacitance increased with the potential window Regarding the potential window effect, Li et al. studied it on CNT@G electrodes with mass loadings of 3 mg/cm 2 and 5 mg/cm 2 . A CV curve example at a scanning rate of 20 mV/s for 3 mg/cm 2 mass loading can be seen in. CV curves of both electrodes show a pair of redox peaks in the negative potential range, and an additional pair of redox peaks appears with an increasing potential window on the positive side. These peaks are related to the trace amount of Au catalyst distributed on the graphene sheets. In addition, both kinds of the specific capacitance increased with the potential window width up to the maximum of 1.8 V corresponding to the range of −0.9 to 0.9 V (see.. CV curves of CNT@G with different mass loadings at 50 mV/s (a) and calculated capacitance of the CNT@G electrode on Ni foam as a function of mass loading (b) (Reproduced with permission of. Copyright Elsevier, 2019).
Regarding the potential window effect, Li et al. studied it on CNT@G electrodes with mass loadings of 3 mg/cm 2 and 5 mg/cm 2 . A CV curve example at a scanning rate of 20 mV/s for 3 mg/cm 2 mass loading can be seen in. CV curves of both electrodes show a pair of redox peaks in the negative potential range, and an additional pair of redox peaks appears with an increasing potential window on the positive side. These peaks are related to the trace amount of Au catalyst distributed on the graphene sheets. In addition, both kinds of the specific capacitance increased with the potential window width up to the maximum of 1.8 V corresponding to the range of −0.9 to 0.9 V (see.. CV curves at different potential windows measured for CNT@G electrodes with 3 mg/cm 2 mass loading at 20 mV/s (a) and relationship between the potential window and gravimetric (or specific) capacitance (Cm) and areal capacitance (Ca) for CNT@G electrodes with mass loading of 3 mg/cm 2 and 5 mg/cm 2 (b) (Reproduced with permission of. Copyright Elsevier, 2019).
At the same time, the detailed study of CV curves recorded in the widest potential window covering the range of −0.9 V to +0.9 V (see, covered two smaller windows such as between 0 and +0.9 V (seeand between −0.9 V and 0 V (see . CV curves at different potential windows measured for CNT@G electrodes with 3 mg/cm 2 mass loading at 20 mV/s (a) and relationship between the potential window and gravimetric (or specific) capacitance (C m ) and areal capacitance (C a ) for CNT@G electrodes with mass loading of 3 mg/cm 2 and 5 mg/cm 2 (b) (Reproduced with permission of. Copyright Elsevier, 2019).
At the same time, the detailed study of CV curves recorded in the widest potential window covering the range of −0.9 V to +0.9 V (see, covered two smaller windows such as between 0 and +0.9 V (seeand between −0.9 V and 0 V (seewith completely different forms of CV. The negative potential window electrode works obviously as EDLC (seeopposite to the positive range with a visible Faradaic response (see. There is also a correlation with the calculated values of the specific capacitance for CNT@G/Ni electrode with mass loading of 3 mg/cm 2 . This value reaches only 51.3 F/g at 1 mV/s for EDLC in a potential window of −0.9 V ÷ 0 V, achieving very high 620 F/g for battery-like behaviour in the range 0 V ÷ +0.9 V, and the middle value of 373 F/g for EDLC with a Faradaic impact in the widest potential window −0.9 V ÷ +0.9 V (see. In addition, the estimated b parameter was different for each of all three diapasons and can be presented as 0.5 < b (0 ÷ +0.9 V) < b (−0.9 V÷ +0.9 V) < b (−0.9 V ÷ 0)~1 . Moreover, as can be seen in, the faradaic impact to specific capacitance seen at a low scan rate disappeared with the rate increase. Furthermore, at a high scan rate >0.02 mV/s the value of the specific capacitance in all three measured potential windows becomes the same and does not exceed the lowest capacity of 51.3 F/g (see. Nanomaterials 2021, 11, x FOR PEER REVIEW 20 of 30. CV curves obtained at different scanning rates for CNT@G with mass loading of 3 mg/cm 2 measured in potential window from −0.9 V to +0.9 V (a), from 0 to +0.9 V (b), from 0 to −0.9 V (c), and corresponding specific capacitances (d) (Reproduced with permission of. Copyright Elsevier, 2019).
## Mno2 induced pseudocapacitance
MnO2 is widely used for energy storage because of its high theoretical pseudocapacitance, wide potential range, and low toxicity and cost (natural abundance). The fact that MnO2-based composites are widely applied with neutral aqueous electrolytes being well correlated with the current environmental requirements of "green electrolytes" in supercapacitors is also important. Indeed, all reported composite electrodes based on G/rGO-CNT with MnO2 were tested in the Na2SO4 electrolyte as seen in in Supplementary Materials. Comparing the values for the composites prepared with and without MnO2 (also shown in it can be seen that MnO2 as a redox oxide can significantly increase the specific capacitance of rGO/CNT composite.
The highest enhancement was reported by Bi et al. for the layered structure of graphene and CNT decorated by MnO2 on Cu foil. The long and complicated preparation of layered G/CNT with MnO2 structure included CVD, immersion, a thermal decomposition process, etc. However, it resulted in the specific capacitance increase from 42 F/g to 365 F/g at 1 A/g before and after MnO2 deposition, respectively.
A more popular and simple method is the use of KMnO4 to obtain MnO2 during the processing. In this way Ramezani et al. obtained the capacitance of 367 F/g at 20 mV/s for the composite of hydrazine reduced rGO, CNT, MnO2, graphite powder and PVDF covering graphite paper and it was twice higher than 150 F/g mentioned in the same work for rGO-CNT without MnO2. Electrodes on Ni foam with rGO, CNT, MnO2, AB, PTFE were fabricated by Liu et al.and Deng et al.. However, Liu et al. reported the. CV curves obtained at different scanning rates for CNT@G with mass loading of 3 mg/cm 2 measured in potential window from −0.9 V to +0.9 V (a), from 0 to +0.9 V (b), from 0 to −0.9 V (c), and corresponding specific capacitances (d) (Reproduced with permission of. Copyright Elsevier, 2019).
## Mno 2 induced pseudocapacitance
MnO 2 is widely used for energy storage because of its high theoretical pseudocapacitance, wide potential range, and low toxicity and cost (natural abundance). The fact that MnO 2 -based composites are widely applied with neutral aqueous electrolytes being well correlated with the current environmental requirements of "green electrolytes" in supercapacitors is also important. Indeed, all reported composite electrodes based on G/rGO-CNT with MnO 2 were tested in the Na 2 SO 4 electrolyte as seen in in Supplementary Materials. Comparing the values for the composites prepared with and without MnO 2 (also shown in it can be seen that MnO 2 as a redox oxide can significantly increase the specific capacitance of rGO/CNT composite.
The highest enhancement was reported by Bi et al. for the layered structure of graphene and CNT decorated by MnO 2 on Cu foil. The long and complicated preparation of layered G/CNT with MnO 2 structure included CVD, immersion, a thermal decomposition process, etc. However, it resulted in the specific capacitance increase from 42 F/g to 365 F/g at 1 A/g before and after MnO 2 deposition, respectively.
A more popular and simple method is the use of KMnO 4 to obtain MnO 2 during the processing. In this way Ramezani et al. obtained the capacitance of 367 F/g at 20 mV/s for the composite of hydrazine reduced rGO, CNT, MnO 2 , graphite powder and PVDF covering graphite paper and it was twice higher than 150 F/g mentioned in the same work for rGO-CNT without MnO 2. Electrodes on Ni foam with rGO, CNT, MnO 2 , AB, PTFE were fabricated by Liu et al.and Deng et al.. However, Liu et al. reported the increase from 35 F/g at 5 mV/s for rGO-CNT-AB-PTFE to 133 F/g for rGO-CNT-MnO 2 -AB-PTFE, using GO aerogel reduced at 800 - C. At the same time, Deng et al. used hydrazine-reduced GO, CNT, MnO 2 AB, PTFE and reported specific capacitance of 91 F/g and 126 F/g at 0.25 A/g for electrodes without and with MnO 2. Using urea for GO reduction and poly(1,5-diaminoanthraquinone) (PDAA) for functionalization of MnO 2 -CNT Lei et al. obtained 80 F/g and 193 F/g at 0.2 A/g for rGO, CNT, PDAA, CB, PTFE and rGO, CNT, MnO 2 , PDAA, CB, PTFE composites, respectively. Preparation of rGO-CNT-MnO 2 by the HT method at 150 - C was used by Li et al. for the fabrication of rGO, CNT, MnO 2 , AB, PTFE composite electrode on Ni foam with the final specific capacitance 336 F/g at 0.5 A/g. As can be seen, all the aforementioned composite electrodes with MnO 2 used additives and bindersotherwise resulting in fabrication difficulties. However, Cheng et al. were able to prepare rGO-CNT-MnO 2 freestanding electrode by simple filtration. Although for that electrode, GO was reduced by hydrazine and ammonium solutions, the specific capacitance equal to 372 F/g at 10 mV/s was measured.
It needs to be stressed here that CV and GCD curves of reported electrodes with MnO 2 presented shapes typical for materials with dominant EDLC energy storage behaviour as can be seen infor G/rGO-CNT-MnO 2 -AB-PTFE reported by Deng et al.. There are no significant redox peaks appearing in CV curves even at a high scan rate (seeand no plateau in GCD curves for these electrodes (see. Moreover, the b parameter for all aforementioned G/rGO-CNT-based composites with MnO 2 was estimated by us to be~0.8 that also corresponds to dominant capacitive type of storage in these electrode materials.. As can be seen, all the aforementioned composite electrodes with MnO2 used additives and bindersotherwise resulting in fabrication difficulties. However, Cheng et al. were able to prepare rGO-CNT-MnO2 freestanding electrode by simple filtration. Although for that electrode, GO was reduced by hydrazine and ammonium solutions, the specific capacitance equal to 372 F/g at 10 mV/s was measured. It needs to be stressed here that CV and GCD curves of reported electrodes with MnO2 presented shapes typical for materials with dominant EDLC energy storage behaviour as can be seen infor G/rGO-CNT-MnO2-AB-PTFE reported by Deng et al.. There are no significant redox peaks appearing in CV curves even at a high scan rate (seeand no plateau in GCD curves for these electrodes (see. Moreover, the b parameter for all aforementioned G/rGO-CNT-based composites with MnO2 was estimated by us to be ~0.8 that also corresponds to dominant capacitive type of storage in these electrode materials.
## Effect of other metal oxides
In addition to MnO2, the influence of other metal oxides on G/rGO-CNT-based composites has been also reportedand summarized in in the Supplementary Materials. Ramesh et al. mixed CNT, ammonium reduced GO, and cellulose fibres simultaneously with Co3O4 and SnO2, added AB, PTFE and covered Ni foam with it. CV curves of such electrodes studied in KOH electrolyte presented strong redox peaks at −0.1 V ÷ −0.2 V, which, as well as GCD curve shape, cannot be attributed to EDLC (seebut correlated well with the faradaic impact from Co3O4 and SnO2. The reported specific capacitance of 215 F/g at 0.2 A/g was obtained for electrodes studied in a negative potential window from 0 to −1.0 V.
## Effect of other metal oxides
In addition to MnO 2 , the influence of other metal oxides on G/rGO-CNT-based composites has been also reportedand summarized in in the Supplementary Materials. Ramesh et al. mixed CNT, ammonium reduced GO, and cellulose fibres simultaneously with Co 3 O 4 and SnO 2 , added AB, PTFE and covered Ni foam with it. CV curves of such electrodes studied in KOH electrolyte presented strong redox peaks at −0.1 V ÷ −0.2 V, which, as well as GCD curve shape, cannot be attributed to EDLC (seebut correlated well with the faradaic impact from Co 3 O 4 and SnO 2. The reported specific capacitance of 215 F/g at 0.2 A/g was obtained for electrodes studied in a negative potential window from 0 to −1.0 V. A similar diffusion-dominated energy-storage mechanism can be observed in CV curves reported by Trian et al. for electrodes on Ni foam made of rGO-CNT with Fe2O3 and mixed with CB and PTFE and supported by GCD curves (seeor by Chen et al. for rGO-CNT with LiMn2O4 and mixed with AB, PTFE. In all the cases, strong redox peaks in CV from metal oxides indicated the significant impact from faradaic materials as can be seen infor LiMn2O4. Moreover, GCD curves (seeas well as the EIS Nyquist plot with the slop close to 45° (seewas reminiscent the battery-type electrodes. In addition Chen et al. calculated the parameter b = 0.689 that is closer to b = 0.5 characteristic for battery-type energy storage, especially at low scan rate that can be seen in[56].
## G/rgo-cnt with ni(oh)2
Nickel hydroxide is an attractive material for supercapacitor applications because of its high theoretical specific capacitance, well-defined redox behaviour and low cost. The available data for rGO-CNT before and after modification by Ni(OH)2 are presented in. Copyright Elsevier, 2017). Electrochemical performance of LiMn 2 O 4 -CNT-graphene nanocomposite: GCD at different current densities (b), EIS Nyquist plots of the electrode after cycling (c). Contribution ratios of capacitive and diffusion-controlled processes at various scan rates (d) (Reproduced with permission of. Copyright Elsevier, 2019).
A similar diffusion-dominated energy-storage mechanism can be observed in CV curves reported by Trian et al. for electrodes on Ni foam made of rGO-CNT with Fe 2 O 3 and mixed with CB and PTFE and supported by GCD curves (seeor by Chen et al. for rGO-CNT with LiMn 2 O 4 and mixed with AB, PTFE. In all the cases, strong redox peaks in CV from metal oxides indicated the significant impact from faradaic materials as can be seen infor LiMn 2 O 4. Moreover, GCD curves (seeas well as the EIS Nyquist plot with the slop close to 45 - (seewas reminiscent the battery-type electrodes. In addition Chen et al. calculated the parameter b = 0.689 that is closer to b = 0.5 characteristic for battery-type energy storage, especially at low scan rate that can be seen in[56].
## G/rgo-cnt with ni(oh) 2
Nickel hydroxide is an attractive material for supercapacitor applications because of its high theoretical specific capacitance, well-defined redox behaviour and low cost. The available data for rGO-CNT before and after modification by Ni(OH) 2 are presented inin Supplementary Materials. The reported specific capacitance of composites with Ni(OH) 2 has significantly higher values in comparison with other electrodes described above. Moreover, all the reported electrodes based on G/rGO-CNT with Ni(OH) 2 were tested in KOH electrolyte, showing rather close capacitance values independent of G/rGO processing temperatures. According to Fan et al., simple mixing of Ni(NO) 3 ·6H 2 O with urea and with rGO-CNT aerogel reduced at 800 - C resulted in a stable electrode on Ni foam with specific capacitance of 1208 F/g at 1 A/g, although a Ni-free rGO-CNT electrode prepared in the same way showed only 149 F/g at 1 A/g. A similar value of 1320 F/g at 6 A/g was reported by Chen et al. for composite electrodes made of slurry including AB, PTFE and rGO-CNT-Ni(OH) 2 obtained in autoclave at 120 - C. A more complicated method was used by Du et al. for the preparation of vertically aligned CNT (VACNT) structure from highly ordered pyrolytic graphite at 1200 - C and G growing by pyrolysis of iron phthalocyanine (FePc) at 1000 - C (seewith the following Ni(OH) 2 coating by electrochemical deposition (see. Specific capacitance of 110 F/g at 10 mV/s for G on a CNT structure and 1384 F/g at 5 mV/s for G on CNT and covered by Ni(OH) 2 was measured. with urea and with rGO-CNT aerogel reduced at 800 °C resulted in a stable electrode on Ni foam with specific capacitance of 1208 F/g at 1 A/g, although a Ni-free rGO-CNT electrode prepared in the same way showed only 149 F/g at 1 A/g. A similar value of 1320 F/g at 6 A/g was reported by Chen et al. for composite electrodes made of slurry including AB, PTFE and rGO-CNT-Ni(OH)2 obtained in autoclave at 120 °C. A more complicated method was used by Du et al. for the preparation of vertically aligned CNT (VACNT) structure from highly ordered pyrolytic graphite at 1200 °C and G growing by pyrolysis of iron phthalocyanine (FePc) at 1000 °C (seewith the following Ni(OH)2 coating by electrochemical deposition (see. Specific capacitance of 110 F/g at 10 mV/s for G on a CNT structure and 1384 F/g at 5 mV/s for G on CNT and covered by Ni(OH)2 was measured. with Ni(OH)2 (c,d). Electrochemical impedance spectra (EIS) Nyquist plots of VACNT-graphene electrode before (e) and after (f) modification by Ni(OH)2. CV at different scan rates (g) and GCD curves at various discharge current densities (h) (Reproduced with permission of. Copyright American Chemical Society, 2011).
However, all these electrodes with Ni(OH)2 presented CV and CGD curves very far from those for typical pseudocapacitive and particularly for EDLC materials. As presented in, CV curves with strong redox peaks increasing and shifting with the scan rate, typical for all G/rGO-CNT electrodes with Ni(OH)2, were reported by Du et al. with Ni(OH) 2 (c,d). Electrochemical impedance spectra (EIS) Nyquist plots of VACNT-graphene electrode before (e) and after (f) modification by Ni(OH) 2 . CV at different scan rates (g) and GCD curves at various discharge current densities (h) (Reproduced with permission of. Copyright American Chemical Society, 2011).
However, all these electrodes with Ni(OH) 2 presented CV and CGD curves very far from those for typical pseudocapacitive and particularly for EDLC materials. As presented in, CV curves with strong redox peaks increasing and shifting with the scan rate, typical for all G/rGO-CNT electrodes with Ni(OH). A significant impact of faradaic contribution to G/rGO-CNT electrodes with Ni(OH) 2 is easy to detect in EIS Nyquist plots of the G-CNT electrode before (seeand after (seeNi(OH) 2 deposition. The slope changes from almost 90 - for G-CNT with pure EDLC behaviour to almost 45 - was associated with the battery. In addition, b parameters estimated by us are close to 0.6 that means the diffusion controlled mechanism (typical for battery-type electrodes) as the dominant one in these composites. Based on that and according to requirements from many research papers (i.e., references, etc.), other units and calculations associated with batteries (i.e., mAh instead F) need to be used for the characterization of such electrode materials. Moreover, comparison of these high specific capacitance values for such hybrid materials as well as other their parameters with that of really capacitive materials are incorrect and speculative.
## Specific energy and power of supercapacitors with electrodes based on g/rgo-cnt and their cycling stability
Energy density (in Wh/cm 3 ) and power density (in W/cm 3 ) are known to be among the main characteristic parameters of SC for their commercial application. Therefore, the goal of research is to achieve high energy density at high power density, although in the case of the electrodes based on G/rGO-CNT these values are rarely presented, being always substituted by specific energy and power. SC specific energy (E in Wh/kg) and specific power (P in W/kg) can be calculated by using the following expressions:
[formula] E = 1 2 × 3.6 C total ∆V 2 or E = 1 8 × 3.6 C single el. ∆V 2 ,(2)P = E ∆t ,(3) [/formula]
where C total and C single el. are the measured capacitance of full SC and that of single electrode, respectively, ∆V is the operating voltage window, ∆t is the discharge time in hours. Thus, although the values of capacitance are very important for the SC performance, the electrolyte voltage window plays also a major role for the enhancement of specific energy as well as specific power. However, in the case of devices with strong redox peaks in CV and plateaus in GCD curves (e.g., references, the calculation of the specific energy cannot be done using Equations (2) and (3) valid only for capacitive materials characterized by rectangular CV and triangular GCD. That is mainly because of the non-triangular shaped GCD that is used for the calculation of energy. Whereas the specific power and energy calculation in capacitive materials is based on the area under the triangular GCD during charge discharge time, the actual energy in the battery-type materials is the area under the curved lines with plateaus. Thus, it appears that the charging energy is larger than the discharging one, reflecting the electrode reaction being not fully reversible. Hence only a portion of the energy used during the charging period was released during the discharging period. In this case, the energy efficiency considered as the ratio of discharging energy to charging energy is far smaller than 1, in contrast to that for capacitive energy storage. Thus, the Ragone plot shown inpresents only the available data of several symmetric supercapacitors made of the capacitive materials (EDLC and pseudocapacitors) and one asymmetric device that used rGO-CNT-AB-PTFE (EDLC electrode) and rGO-CNT-MnO 2 -AB-PTFE (pseudocapacitive electrode). As a result, the highest value was calculated for the asymmetric supercapacitor supporting the importance of the enlarged voltage window according to Equationsand. However, if we consider only symmetric SC with capacitive electrodes (open circle in, a significant performance is evident to be achieved by Ding et al. for EDLC without addition of faradaic materials. It is a surprise that a network of functionalized graphene nanosheets and CNT (fG/CNT) was synthesized by chemical oxidation of KMnO 4 simultaneously with CNT showing after low temperature treatment specific energy of 11.7 Wh/kg. This value of energy density is higher than that for devices using faradaic materials, i.e., MnO 2, Fe 3 O 4 catalyst, PANI, PDA, and it is significantly higher than that for two other SCs with EDLC electrodes. Moreover, specific capacitance of 202 F/g at 0.5 A/g that was not too high was reported for these single fG/CNT electrodes in three-electrode configuration and the widest voltage window was not used for the electrochemical test. Thus, if Ding et al. did not find in the analysed composite the traces of MnO 2, for the preparation of which KMnO 4 is usually usedand with which such an improvement could be associated, this is the best result for the G/rGO-CNT-based electrodes reported to date. showing after low temperature treatment specific energy of 11.7 Wh/kg. This value of energy density is higher than that for devices using faradaic materials, i.e., MnO2, Fe3O4 catalyst, PANI, PDA, and it is significantly higher than that for two other SCs with EDLC electrodes. Moreover, specific capacitance of 202 F/g at 0.5 A/g that was not too high was reported for these single fG/CNT electrodes in three-electrode configuration and the widest voltage window was not used for the electrochemical test. Thus, if Ding et al. did not find in the analysed composite the traces of MnO2, for the preparation of which KMnO4 is usually usedand with which such an improvement could be associated, this is the best result for the G/rGO-CNT-based electrodes reported to date. Another important characteristic for the practical application of the electrodes/full SC is their stability after charging/discharging for a long time. The cycling stability can be seen from to be rather independent of the type of electrolyte or electrode substrate for all the reported devices, although several of them reported some fluctuations during the measurements. While most of the values are close to 100%, there are also relatively low values of 75% and 80.5% reported for rGO-CNT with MnO2, AB and PTFEand rGO-CNT with PANI, respectively. On the other hand, the cycle stability measurement of energy storage devices has to begin only after stable operation has been demonstrated and the abnormal results reported sometimes are not from the stabile cycling but rather from conditioning. In this case, electrode stabilization must be performed before the cycle stability measurements can be properly made. Another important characteristic for the practical application of the electrodes/full SC is their stability after charging/discharging for a long time. The cycling stability can be seen from to be rather independent of the type of electrolyte or electrode substrate for all the reported devices, although several of them reported some fluctuations during the measurements. While most of the values are close to 100%, there are also relatively low values of 75% and 80.5% reported for rGO-CNT with MnO 2 , AB and PTFEand rGO-CNT with PANI, respectively. On the other hand, the cycle stability measurement of energy storage devices has to begin only after stable operation has been demonstrated and the abnormal results reported sometimes are not from the stabile cycling but rather from conditioning. In this case, electrode stabilization must be performed before the cycle stability measurements can be properly made. . Values of the initial specific capacitance (C in ) of G/rGO-CNT-based single electrodes measured in three-electrode configuration before long-term test and capacitance retention after long cycling (ordered with number of cycles increase) reported for G/rGO-CNT-based single electrodes as well as for full devices using them.
## Conclusions and perspective of g/rgo-cnt-based composite electrodes
The high-quality monolayer of graphene shows great potential for different applications such as miniaturized and precise micro/nano electronics, while chemically or/and thermally reduced graphene oxide provides a practical route towards lower-cost production of different rGO-based devices, particularly supercapacitors. Because GO is easily dissolved in a variety of solvents and due to high solubility of CNT achieved in GO solution, the combination of G/rGO and CNT is widely studied in SC as mixed or layered electrode materials. An amount of around 10 wt.% of CNT is generally sufficient to obtain the maximum value of the specific capacitance in case of the two-component rGO-CNT electrodes. Therefore, both freestanding and substrate supported electrodes can show high specific capacitance but only under certain/individual conditions such as the method of fabrication, type of substrate, reduction temperature or chemicals, ratio between components of the electrodes, etc. Thus, the addition of faradaic materials always increases the capacitance. However, although the compatibility of G/rGO-CNT with different polymers, metals, etc. opens up a route for a wide practical realization of functional composite materials for SC fabricated using commercial, mainly slurry-based battery technology, determining the mechanism of energy storage for each individual electrode immediately before the manufacture of the final device is essential. At the same time, talking not about electrodes only but about supercapacitors in general, high specific energy and power are crucial parameters for commercialization as well as capacitance retention. Thus, until now the highest energy and power densities were associated with hybrid devices. Based on the information described above, combining G/rGO-CNT with different additives and binders, a number of perspective composite electrodes both of capacitive and battery type can be achieved. Therefore, particular requirements to the performance of electric storage devices determine the need for detailed understanding of the relationship between the fabrication, structure and final properties of composite electrodes. In turn, the understanding of the importance of the results achieved through their comparison can be greatly facilitated if the result presentation is unified/standardized. In addition, it is obvious that the future of supercapacitors is in the asymmetric configuration, because the symmetric one has more limited behaviour that does not correspond to future needs. |
The quality of reporting of randomised controlled trials in asthma: systematic review protocol
PS1 BackgroundThe randomised controlled trial (RCT) is the most robust design to assess the efficacy and effectiveness of treatments. 1 As a result of this realisation, clinical decision-making in recent years has been directed away from reliance based solely on the doctor's clinical experience towards a paradigm based on evidence derived from RCTs. The results of large RCTs have subsequently been translated into guidelines containing evidence-graded recommendations which clinicians are encouraged to use as the basis of good clinical practice. 2 If, however, the 'raw material' is flawed, the conclusions cannot be trusted, hence the need to appraise critically the quality of the underpinning trial evidence.3Quality is a multidimensional concept which relates to the design, conduct, and analysis of a trial, its clinical relevance, and its reporting. 3 In most cases, the RCT report is the only source for clinicians, guideline developers, and other researchers to judge the validity and generalisability of the results, so the quality of reporting of trials is of inherent interest. 4 It is then of considerable concern that the quality of reporting of RCTs is often sub-optimal.5In response to these concerns about the quality of reporting of RCTs, in the mid-1990s an international group developed the Consolidated Standards of Reporting Trials (CONSORT) Statement.6This was first published in 1996 and then updated in 2001. 5 After an expert meeting in January 2007, this was further revised resulting in the latest iteration -i.e. the CONSORT 2010 Statement.5The current CONSORT Statement (hereafter referred to as CONSORT) comprises a checklist of essential items that should be included in reports of RCTs and a diagram for documenting the flow of participants through a trial. It is formulated primarily for use with reports of two-arm parallel-group RCTs. Many of the CONSORT data fields are, however, also relevant to a wider spectrum of trial designs such as non-inferiority, equivalence, factorial, cluster, and crossover trials. 5 Extensions to the CONSORT checklist for reporting trials with some of these other designs have been published, 7-9 as have those for reporting particular types of data (i.e. harms 10 ), types of interventions (i.e. non-pharmacological treatments 11 and herbal interventions 12 ), and abstracts. 13 In this review we will, as appropriate, use the non-pharmacological, non-inferiority and equivalence, cluster and pragmatic extensions of CONSORT.CONSORT criteria have been used to assess the reporting quality of RCTs in several disease areas 4,14,15 and journal types.16,17There have, however, been no recent assessments of the quality of RCTs reporting in the asthma literature. The only previous study on clinical trials of asthma treatments was undertaken for the period 1984-1997 and was published in 2002 in two reports. 18,19 This initially involved a comparison between RCTs published in Spanish and English language journals, 19 and this was then followed by a secondary analysis of a subsection of the same dataset focusing solely on the quality of RCTs in English.18The first article showed poorer reporting quality of the RCTs in Spanish publications and a strong association between the type of journal, type of intervention, and the comparison measure used and reporting quality. Moreover, this study highlighted the necessity for better reporting in general in the asthma literature, leading the authors to advocate the more widespread use of a checklist by authors and editors in order to improve reporting standards.19Building on this earlier work, we will examine the quality of reporting of asthma clinical RCTs in the contemporary asthma literature. Our secondary aim is to investigate if there is an association between specific trial characteristics that have previously been identified in the literature in influencing reporting quality and the actual quality of the trial reports.4,14,17,[20][21][22]PRIMARY CARE RESPIRATORY JOURNAL www.thepcrj.org http://dx.
# Ps1 background
The randomised controlled trial (RCT) is the most robust design to assess the efficacy and effectiveness of treatments. [bib_ref] Assessing the quality of randomized controlled trials: an annotated bibliography of scales..., Moher [/bib_ref] As a result of this realisation, clinical decision-making in recent years has been directed away from reliance based solely on the doctor's clinical experience towards a paradigm based on evidence derived from RCTs. The results of large RCTs have subsequently been translated into guidelines containing evidence-graded recommendations which clinicians are encouraged to use as the basis of good clinical practice. [bib_ref] External validity of randomised controlled trials in asthma: to whom do the..., Travers [/bib_ref] If, however, the 'raw material' is flawed, the conclusions cannot be trusted, hence the need to appraise critically the quality of the underpinning trial evidence. [bib_ref] Systematic reviews in health care: assessing the quality of controlled clinical trials, Juni [/bib_ref] Quality is a multidimensional concept which relates to the design, conduct, and analysis of a trial, its clinical relevance, and its reporting. [bib_ref] Systematic reviews in health care: assessing the quality of controlled clinical trials, Juni [/bib_ref] In most cases, the RCT report is the only source for clinicians, guideline developers, and other researchers to judge the validity and generalisability of the results, so the quality of reporting of trials is of inherent interest. [bib_ref] Quality of reporting of randomized controlled trials in general endocrinology literature, Rios [/bib_ref] It is then of considerable concern that the quality of reporting of RCTs is often sub-optimal. [bib_ref] explanation and elaboration: updated guidelines for reporting parallel group randomised trials, Moher [/bib_ref] In response to these concerns about the quality of reporting of RCTs, in the mid-1990s an international group developed the Consolidated Standards of Reporting Trials (CONSORT) Statement. [bib_ref] Improving the quality of reporting of randomized controlled trials. The CONSORT statement, Begg [/bib_ref] This was first published in 1996 and then updated in 2001. [bib_ref] explanation and elaboration: updated guidelines for reporting parallel group randomised trials, Moher [/bib_ref] After an expert meeting in January 2007, this was further revised resulting in the latest iteration -i.e. the CONSORT 2010 Statement. [bib_ref] explanation and elaboration: updated guidelines for reporting parallel group randomised trials, Moher [/bib_ref] The current CONSORT Statement (hereafter referred to as CONSORT) comprises a checklist of essential items that should be included in reports of RCTs and a diagram for documenting the flow of participants through a trial. It is formulated primarily for use with reports of two-arm parallel-group RCTs. Many of the CONSORT data fields are, however, also relevant to a wider spectrum of trial designs such as non-inferiority, equivalence, factorial, cluster, and crossover trials. [bib_ref] explanation and elaboration: updated guidelines for reporting parallel group randomised trials, Moher [/bib_ref] Extensions to the CONSORT checklist for reporting trials with some of these other designs have been published, 7-9 as have those for reporting particular types of data (i.e. harms 10 ), types of interventions (i.e. non-pharmacological treatments [bib_ref] Extending the CONSORT statement to randomized trials of nonpharmacologic treatment: explanation and..., Boutron [/bib_ref] and herbal interventions 12 ), and abstracts. In this review we will, as appropriate, use the non-pharmacological, non-inferiority and equivalence, cluster and pragmatic extensions of CONSORT.
CONSORT criteria have been used to assess the reporting quality of RCTs in several disease areas [bib_ref] Quality of reporting of randomized controlled trials in general endocrinology literature, Rios [/bib_ref] [bib_ref] Quality of randomized controlled trials reporting in the primary treatment of brain..., Lai [/bib_ref] [bib_ref] Quality of reporting of randomized, controlled trials in cerebral palsy, Anttila [/bib_ref] and journal types. [bib_ref] Quality assessment of reports on clinical trials in the Journal of Hepatology, Gluud [/bib_ref] There have, however, been no recent assessments of the quality of RCTs reporting in the asthma literature. The only previous study on clinical trials of asthma treatments was undertaken for the period 1984-1997 and was published in 2002 in two reports. [bib_ref] Quality of published clinical trials on asthma, Quinones [/bib_ref] This initially involved a comparison between RCTs published in Spanish and English language journals,and this was then followed by a secondary analysis of a subsection of the same dataset focusing solely on the quality of RCTs in English. [bib_ref] Quality of published clinical trials on asthma, Quinones [/bib_ref] The first article showed poorer reporting quality of the RCTs in Spanish publications and a strong association between the type of journal, type of intervention, and the comparison measure used and reporting quality. Moreover, this study highlighted the necessity for better reporting in general in the asthma literature, leading the authors to advocate the more widespread use of a checklist by authors and editors in order to improve reporting standards.Building on this earlier work, we will examine the quality of reporting of asthma clinical RCTs in the contemporary asthma literature. Our secondary aim is to investigate if there is an association between specific trial characteristics that have previously been identified in the literature in influencing reporting quality and the actual quality of the trial reports. [bib_ref] Quality of reporting of randomized controlled trials in general endocrinology literature, Rios [/bib_ref] [bib_ref] Quality of randomized controlled trials reporting in the primary treatment of brain..., Lai [/bib_ref] [bib_ref] Quality assessment of reports on clinical trials in the Journal of Hepatology, Gluud [/bib_ref] [bib_ref] Randomized clinical trials in hepatology: predictors of quality, Kjaergard [/bib_ref] [bib_ref] The quality of reports of randomised trials in 2000 and 2006: comparative..., Hopewell [/bib_ref] [bib_ref] A systematic review of the quality of publications reporting coronary artery bypass..., Farrokhyar [/bib_ref]
## Ps2
## Objectives
The primary objective is to assess the contemporary quality of reporting of RCTs in the asthma literature for the period 2010-2012.
The secondary objectives are to identify factors associated with better reporting quality, that is: - Are trials published in general medicine journals associated with better quality than those published in specialist journals? - Is a high impact factor of the journal of publication associated with studies of better quality than those published in lower impact journals? - Are studies conducted or led by teams in high income country settings (defined using World Bank Group definitions) associated with better quality than those in middle and low income country settings? - Does the funding source have an impact on study quality? - Are trials evaluating a pharmacological intervention associated with better quality than those evaluating a non-pharmacological intervention? - Are studies with multiple participating centres associated with better quality than single-centre studies?
# Review methods
## Search strategy
We will search the electronic database MEDLINE (via Ovid) using the search terms of the Cochrane Airways Group Specialised Register for asthma and RCTs for the time period January 2010 to July 2012. We will include studies that have been published in the top 10 impact factor journals in general medicine and respiratory specialty journals using the most recent available (i.e. 2011) rankings,as long as they publish clinical trials and include articles related to asthma. Our complete search strategy is presented in Appendix 1.
## Inclusion criteria
- Types of studies: RCTs with parallel or cluster study design that involve only human subjects - Types of participants: All study populations with asthma as the only condition being examined - Types of interventions: Pharmacological and nonpharmacological interventions evaluating the clinical effectiveness of a treatment with any conceptual framework (superiority, non-inferiority, equivalence). We consider that a trial is evaluating the effectiveness of a treatment as long as it has at least one clinical outcome (primary or secondary).
## Exclusion criteria
- Reviews, systematic reviews, and meta-analyses - Non-randomised trial designs (quasi-experimental, observational studies) - Studies with crossover and factorial design, n-of-
## Review strategy
Searches will be undertaken independently by two reviewers (CN and PB) with support from AW and AS. The references will be imported into EndNote and duplicates will be deleted. Both reviewers will independently review the titles for potentially eligible studies. They will not be blinded to study details. If they are unsure or there are disagreements they will read the abstract also. Full text copies of potentially relevant studies will be obtained and CN and PB will assess their eligibility for inclusion against the criteria mentioned above. A kappa statistic will be calculated to measure the level of agreement.Where the reviewers agree, they will either include or exclude the study as appropriate. Disagreements will be resolved through discussion with AW or AS as arbiters. The studies that will be excluded after reading the full paper ('near-misses') will be reported in a table with reasons for exclusion. The whole process will be documented on a PRISMA flow chart. [bib_ref] The PRISMA statement for reporting systematic reviews and meta-analyses of studies that..., Liberati [/bib_ref] Data extraction and quality assessment strategy Data will be extracted independently by two reviewers (CN and PB) from the selected studies using an appropriate electronic customised data extraction form (see . The reviewers will not be masked to study details. There will be pilot testing of the data extraction sheet, disagreements will be discussed, and modifications will be made if required. In case of multiple reports of the same study, we will extract data directly into a single data extraction form. Disagreements will be resolved through discussion with AS as arbiter. We will extract data on general characteristics of the trials (see Appendix 2) and use a modified 38-item CONSORT-based checklist (see Appendix 3) that consists of all the CONSORT checklist items plus one additional item from the non-pharmacological treatments extension. The assessment of the adequacy of reporting will be done according to the CONSORT 2010 guidelines and its extensions. [bib_ref] explanation and elaboration: updated guidelines for reporting parallel group randomised trials, Moher [/bib_ref] [bib_ref] CONSORT statement: extension to cluster randomised trials, Campbell [/bib_ref] [bib_ref] Improving the reporting of pragmatic trials: an extension of the CONSORT statement, Zwarenstein [/bib_ref] [bib_ref] Extending the CONSORT statement to randomized trials of nonpharmacologic treatment: explanation and..., Boutron [/bib_ref] Each item can be characterised as 'yes' if it is clearly and adequately reported, or 'no' if it is partially unclear or not reported at all. If an item is not applicable to a specific study we will characterise it as 'N/A'. Each 'yes' answer will receive a score of 1 and each "no" answer will be scored as 0. The overall quality scoring of the trial will be calculated as a proportion of the 'yes' rated applicable items (possible range 0-38 points). In addition, we will score the overall quality of reporting using key parameters of internal validity summarised in the Cochrane Risk of Bias tool (see Appendix 4) and we will categorise the studies into those at (1) low risk of bias and (2) moderate/high risk of bias.The following data will be extracted: General characteristics - Journal name - Journal type (general medicine or specialty) - Journal impact factor - Country of study (high-income, middle-income, low-income) PRIMARY CARE RESPIRATORY JOURNAL www.thepcrj.org
## Ps3
- Funding source (solely industry, part industry, non-industry, none, unknown) - Trial design (parallel or cluster) - Conceptual framework (superiority, non-inferiority, equivalence) - Type of intervention (drug or non-pharmacological) - Number of participating centres (multiple or single centre). Analysis and data synthesis We will calculate the proportion of the trials that have clearly and adequately reported each CONSORT item with a 95% confidence interval (CI). An overall quality score will also be calculated for each trial as a percentage of all the adequately reported applicable items with a 95% CI, which will be used to inform a global assessment of the quality of reporting. The general characteristics data will be presented as numbers and percentages with 95% CI when categorical and as mean and SD or median and IQR with 95% CI when continuous. SPSS software will be used to identify the variables associated with 'low risk of bias' studies with Fisher's exact test, and overall quality scores for subgroups with different trial characteristics will be compared with appropriate two-sample methods (rank-based or Normality-based, depending on the distributional characteristics of the overall quality score). We will report on the quality of reporting of asthma trials and make recommendations for researchers and journal editors regarding the conduct, reporting, and publication of asthma trials. In our description of the studies we will make reference to the setting and population in which the study was undertaken. In concluding, we will consider the quality and relevance of the body of work for informing clinical decision-making.
## Handling editor anthony d'urzo
Acknowledgements We are grateful to the following colleagues for their help in conducting the study: Marshall Dozier, Senior Liaison Librarian, for her contribution to the creation and implementation of the search strategy and data management, and Ulugbek Nurmatov, Clinical Research Fellow, for his advice on the data extraction and quality assessment tools.
## Conflicts of interest
The authors declare that they have no conflicts of interest in relation to this protocol. AS is Joint Editor-in-Chief of the PCRJ, but was not involved in the editorial review of, nor the decision to publish, this article.
## Contributorship
[table] 1: trials, split body trials • Studies evaluating diagnostic tests, prevention, prognosis, costeffectiveness, pathophysiological mechanisms, pharmacokinetics, pharmacogenetics, validation of questionnaires, tolerability of drugs, and economic studies • Trials not reported as full papers (abstracts), editorials, comments, letters, case reports, audits, guidelines, historical articles • Methodological, epidemiological and qualitative studies • Study protocols • Pilot studies and phase I, II, and IV trials • Secondary analysis of trials • Studies reporting updates of previously published RCTs. [/table]
|
GLCCI1 rs37973: a potential genetic predictor of therapeutic response to inhaled corticosteroids in Chinese chronic obstructive pulmonary disease patients
Inhaled corticosteroids (ICSs) are widely prescribed in chronic obstructive pulmonary disease (COPD).However, little is known about predictors of ICSs therapeutic response. To investigate whether the variation in glucocorticoid-induced transcript 1 (GLCCI1) rs37973 is associated with ICS efficacy. A total of 204 clinically stable COPD patients were recruited and administered to inhaled fluticasone propionate/salmeterol combination (500/50 ug, twice daily) for 24 weeks. We genotyped the functional rs37973 and mainly assessed its effects on changes in lung function. In vitro, neutrophils isolated from parts of patients were incubated with various concentrations of dexamethasone (0, 10 −6 M and 10 −4 M) in the presence or absence of cigarette smoke extract, apoptosis was then assessed by flow cytometry. Patients with the homozygous GG genotype (increases of 15.3 ± 33.2 mL) had significantly poorer improvement in FEV 1 than those with the AA (92.7 ± 29.6 mL; p < 0.001) or AG (59.4 ± 26.9 mL; p < 0.001) genotypes after 24-week treatment. In vitro, dexamethasone had less inhibitory effect of neutrophil apoptosis on GG genotype, which further validated the presence of mutant allele 'G' might negatively affect glucocorticoid responsiveness irrespective of smoking status. The GG genotype of rs37973 may associated with decreased ICSs efficacy in Chinese COPD patients.Chronic obstructive pulmonary disease (COPD) is a leading and increasing cause of mortality worldwide 1 , which lead to persistent airflow limitation, airway remodeling and progressive deterioration in lung function. Cigarette smoking is one of the major risk factors for the ongoing inflammation in airways 2 , and inhaled corticosteroids (ICSs) are the mainstay of anti-inflammatory therapy. Current guidelines recommend the use of ICSs combining with long-acting β 2 -agonists for COPD patients at high risk of exacerbations 3 . However, the efficacy shows high inter-individual variability with numerous patients having insufficient response. And prolonged ICSs therapy carries the risk of local and systemic side effects such as oropharyngeal candidiasis, hoarseness, pneumonia 4,5 , osteoporosis etc 6 . Therefore, identification of patients who are prone to nonresponse is becoming increasingly important.Recently a novel pharmacogenetic variation in the glucocorticoid-induced transcript 1 gene (GLCCI1) has been studied intensively in asthma 7-11 , since Tantisira and colleagues revealed in 2011 that the functional rs37973 polymorphism, which was in complete linkage disequilibrium with rs37972, might substantially cause a decreased response to ICSs therapy in non-Hispanic white asthmatics 12 . Subsequently, Van den Berge et al. further investigated the effect of GLCCI1 variant rs37972 on glucocorticoid responsiveness in 63 COPD patients. And then extended the findings of Tantisira et al. from asthma to COPD, by showing that GLCCI1 was also
associated with changes in pulmonary function after ICS therapy in COPD [bib_ref] Genetics of glucocorticoids in asthma, Van Den Berge [/bib_ref]. However, due to small numbers of participants and lack of functional validation, the precise role of GLCCI1 in COPD was far from definable.
Neutrophilic inflammation is a prominent feature of COPD in airways as well as circulation. This persistent and abnormal inflammation is highly correlate with disease progression, and apoptosis is a crucial resolution for non-phlogistic clearance of neutrophils [bib_ref] The neutrophil in chronic obstructive pulmonary disease, Hoenderdos [/bib_ref]. However, glucocorticoids delay neutrophil apoptosis in vitro cell culture. That may partly account for relative less effectiveness of ICSs in COPD than asthma [bib_ref] Glucocorticoid insensitivity as a future target of therapy for chronic obstructive pulmonary..., Marwick [/bib_ref]. Located in GLCCI1 promoter region, the mutant allele 'G' of rs37973 was confirmed to down-regulate the expression of GLCCI1 in vitro functional analysis [bib_ref] Genomewide association between GLCCI1 and response to glucocorticoid therapy in asthma, Tantisira [/bib_ref]. GLCCI1 induction was an early maker of apoptosis in glucocorticoid-treated thymoma cells [bib_ref] Transcriptional control of steroid-regulated apoptosis in murine thymoma cells, Chapman [/bib_ref]. Hence, we postulated that neutrophils, isolated from COPD patients with different genotypes of rs37973, might also responded differently to glucocorticoids stimulation in vitro.
Unlike asthma, although widely prescribed, very few pharmacogenetics studies in COPD have focused on the role of genetic variants in ICS therapeutic response [bib_ref] Pharmacogenetics of chronic obstructive pulmonary disease, Hizawa [/bib_ref] [bib_ref] Association between CRHR1 polymorphism and improved lung function in response to inhaled..., Kim [/bib_ref]. Based on their findings, we aimed to further investigate whether the functional rs37973 variant is associated with long-term ICS therapeutic response in Chinese COPD patients. Moreover, we aimed to verify the potential predictive value of rs37973 variant in vitro cell study of dexamethasone-mediated neutrophil apoptosis.
# Results
## Participants.
A total of 209 clinically stable COPD patients were recruited, among them, 204 eligible patients were finally included in our study (4 lost to follow-up and 1 migrated). The demographic characteristics and lung function at baseline were homogeneous between groups stratified by the rs37973 genotype . The mean age was 67.0 years old. A total of 79.4% patients were men and 53.4% were current smokers with a history of smoking more than 20 pack-years. The mean post-bronchodilator forced expiratory volume in one second (FEV 1 ) was 1.22 liters, which was 46.31% of the predicted value. All of our patients were in category C or D according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria 3 , which meant high risk of COPD exacerbation.
Genotype and allele frequencies. About the rs37973 genotype, 59 (28.9%) individuals were homozygous for the major ' A' allele, 50 (24.5%) individuals were homozygous for the mutant 'G' allele and 95 (46.6%) were AG heterozygotes. The minor allele 'G' frequency was 0.478 and all genotype frequencies were in Hardy-Weinberg equilibrium (p = 0.340).
Association between rs37973 genotype and lung function changes. After 24-week treatment of ICS, patients with the homozygous GG genotype (increases of 15.3 ± 33.2 mL) had significantly poorer improvement in FEV 1 than those with the AA (92.7 ± 29.6 mL; p < 0.001) or AG (59.4 ± 26.9 mL; p < 0.001) genotypes [fig_ref] Figure 1: Changes in lung function after 24-week treatment with ICS according to GLCCI1... [/fig_ref]. As for FEV 1 % of predicted, the GG homozygotes also increased remarkably lower (0.68 ± 1.36%), compared with the AA homozygotes (3.67 ± 1.38%; p < 0.001) or AG heterozygotes (2.40 ± 1.35%; p < 0.001), while the overall mean improvement was 2.33 ± 1.73% [fig_ref] Figure 1: Changes in lung function after 24-week treatment with ICS according to GLCCI1... [/fig_ref]. In addition, the effect of per mutant allele 'G' of rs37973 was estimated by regression analysis [fig_ref] Figure 2: Effect of per minor allele 'G' of rs37973 [/fig_ref] , suggesting that each additional copy of 'G' allele corresponded to a lower ICS efficacy (R 2 = 0.476, F = 179.23, P < 0.001).
Effects of smoking status on ICS efficacy. Consistent with previous studies [bib_ref] Effect of pharmacotherapy on rate of decline of lung function in chronic..., Celli [/bib_ref] [bib_ref] Smoking cessation and lung function in mild-to-moderate chronic obstructive pulmonary disease. The..., Scanlon [/bib_ref] , we found that smoking status was significant predictor of changes in FEV 1 [fig_ref] Figure 3: Effects of smoking status and rs37973 genotype on ICS efficacy with adjustment... [/fig_ref]. Smokers had significant poor improvement compared with non-smoking patients (46.8 ± 39.0 mL versus 71.0 ± 38.3 mL; p < 0.001), after adjusting for age, sex and baseline percentage of predicted FEV 1 . As for changes in FEV 1 % of predicted, similar results were observed in smoking and non-smoking patients (unadjusted for covariates, 1.86 ± 1.61% versus 2.89 ± 1.72%; p < 0.001). Effects of rs37973 genotype on dexamethasone-mediated neutrophil apoptosis. A total of 43 voluntary patients provided an extra 8 ml venous blood for neutrophil culture in vitro, among them, the numbers of AA, AG and GG genotypes of rs37973 were 11, 14 and 18, respectively. Dexamethasone significantly inhibited spontaneous neutrophil apoptosis irrespective of genotype in a concentration-dependent (0, 10 −6 M and 10 −4 M) manner under standard culture conditions for 18 h [fig_ref] Figure 4: Effects of dexamethasone and rs37973 genotype on neutrophil apoptosis [/fig_ref]. Interestingly, we found that neutrophils isolated from patients with GG genotype didn't show corresponding reductions in apoptosis at 10 −6 M, compared with solvent control (the percentage of apoptotic cells: 72.03 ± 2.06% versus 72.27 ± 1.33%; p = 0.903).
Previous studies have reported that dexamethasone caused a concentration-dependent inhibition of neutrophil apoptosis, usually evident at 10 −8 M and maximal at 10 −6 M of clinically relevant drug concentrations [bib_ref] Oxygen levels determine the ability of glucocorticoids to influence neutrophil survival in..., Marwick [/bib_ref] [bib_ref] Glucocorticoid treatment inhibits apoptosis in human neutrophils. Separation of survival and activation..., Cox [/bib_ref] [bib_ref] budesonide and fluticasone propionate inhibit human neutrophil apoptosis, Zhang [/bib_ref]. Similar anti-apoptotic effects were also confirmed in AA and AG genotypes of rs37973 in our study. However, since dexamethasone at 10 −6 M had little effect on GG genotype, we extended our experimental concentration to 10 −4 M, higher than clinically relevant drug concentrations. Only then did we observed an expected suppression of neutrophil apoptosis by dexamethasone in GG genotype (67.03 ± 1.74% versus 72.27 ± 1.33%; p = 0.004). In addition, we observed that neutrophils isolated from patients with AA genotype had significantly decreased percentage of apoptosis compared with those of GG genotype in the presence of both 10 −6 M (59.68 ± 5.80% versus 72.03 ± 2.06%; p = 0.025) and 10 −4 M (53.64 ± 5.88% versus 67.03 ± 1.74%; p = 0.013) dexamethasone, suggesting that dexamethasone had less inhibitory effect on GG genotype [fig_ref] Figure 4: Effects of dexamethasone and rs37973 genotype on neutrophil apoptosis [/fig_ref].
Dexamethasone attenuated CSE-induced neutrophil apoptosis. Cigarette smoke extract (CSE), used in isolated neutrophils system as the underlying inflammatory milieu, could significantly promote neutrophil apoptosis comparing with sham-treated cells of all genotypes. For AA and AG genotypes, dexamethasone reduced the pro-apoptotic effect of CSE in a concentration-dependent manner (10 −6 M and 10 −4 M), compared with merely CSE-treated cells [fig_ref] Figure 5: Dexamethasone attenuated CSE-induced neutrophil apoptosis in different genotypes [/fig_ref]. However, we didn't observe the same effect of dexamethasone on GG genotype at a concentration of 10 −6 M [fig_ref] Figure 5: Dexamethasone attenuated CSE-induced neutrophil apoptosis in different genotypes [/fig_ref] , which was similar to culturing without CSE milieu. CSE-induced neutrophil apoptosis could only be slightly attenuated by a relatively higher concentration (10 −4 M) of dexamethasone (73.63 ± 3.58% versus 81.94 ± 2.45%; p = 0.032).
# Discussion
In our study, we confirmed that GLCCI1 rs37973 was an important determinant of decreased ICSs therapeutic response in COPD. Patients with homozygous mutant 'G' genotype had significantly poor improvement in lung function after 24 weeks of ICS treatment. Moreover, we validated the potential predictive value of rs37973 variant on corticosteroid responsiveness in-vitro experiments, showing that dexamethasone had less anti-apoptotic effects on neutrophils isolated from patients with the GG genotype.
Following up the findings of Tantisira et al. in asthma 12 , Van den Berge and colleagues firstly investigated the role of GLCCI1 rs37972 polymorphism in ICS therapeutic response in 63 Dutch COPD patients. After 3 and 30 months treatment of fluticasone with or without added salmeterol, the major allele homozygotes had remarkably greater improvement in FEV 1 than those with the homozygous mutant genotypes, indicating that GLCCI1 was also associated with ICS responsiveness in COPD [bib_ref] Genetics of glucocorticoids in asthma, Van Den Berge [/bib_ref]. On the basis of Van den Berge et al., a larger sample set, 402 non-Hispanic white COPD patients drawn from two GSK-sponsored clinical studies, were treated with fluticasone furoate monotherapy for 12 weeks. However the results were contrary, showing that the GLCCI1 rs37973 polymorphism, which was highly correlated with rs37972, didn't have an effect on FEV 1 response. Considering the small sample size of Van den Berge et al.'s research, the initial association might be a false positive. However, on the other side, ICS monotherapy is not recommended by guidelines 3 , for it is less effective than combination with long-acting β 2 -agonists. The subjects in GSK study had a lower baseline FEV1 % of predicted and received exclusively ICS monotherapy for only 12 weeks, which might lead to a relative false negative. In addition, the controversial conclusion might also partly result from different ethnics. To better elucidate the possible role of GLCCI1 in COPD, our research then selected a well-characterized population of 204 Chinese clinical stable COPD patients, and all the patients received inhaled fluticasone propionate added salmeterol therapy for 24 weeks. Combining with in-vitro cell functional validated, our findings further confirmed that the functional rs37973 polymorphism held the potential to be a novel genetic predictor of ICS therapeutic response in Chinese COPD patients.
Consistent with previous researches in asthma, the functional rs37973 variant in GLCCI1 was first identified to cause substantial decrements in ICS therapeutic response in 935 white non-Hispanic adults and children by Tantisira et al. [bib_ref] Genomewide association between GLCCI1 and response to glucocorticoid therapy in asthma, Tantisira [/bib_ref] , and subsequently validated in 402 European asthmatic children [bib_ref] S31 Variation at GLCCI1: Association with Increased Steroid Dose But Not Adrenal..., Thompson [/bib_ref] and 224 Japanese adult patients 7 . However, the role of rs37973 has recently been challenged by a lager well-designed replication study in 7 GSK-sponsored clinical trials including 1924 non-Hispanic white asthmatics, by showing that rs37973 didn't significantly influence changes in FEV 1 after 8 or 12 weeks of ICSs treatment 8 . Possible explanations for this discrepant evidence might be divergent end points, different treatment durations and ethnic variations. Both COPD and asthma are chronic airway inflammation characterized by airflow limitation and probably share a common genetic background according to Dutch hypothesis [bib_ref] Asthma and chronic obstructive pulmonary disease: common genes, common environments?, Postma [/bib_ref]. Notwithstanding the similarities, patients with COPD are generally less responsive to corticosteroids than asthmatics, which may partly result from the distinct difference in key inflammatory cells involved in COPD (neutrophils, macrophages, CD8 + lymphocytes) [bib_ref] The nature of small-airway obstruction in chronic obstructive pulmonary disease, Hogg [/bib_ref] and asthma (eosinophils, CD4 + lymphocytes) [bib_ref] Immunology of asthma and chronic obstructive pulmonary disease, Barnes [/bib_ref]. As previous studies have shown that glucocorticoids inhibit neutrophil apoptosis while promote eosinophil apoptosis at clinically relevant drug concentrations, usually between 10 −10 M and 10 −6 M in vitro cell culture [bib_ref] budesonide and fluticasone propionate inhibit human neutrophil apoptosis, Zhang [/bib_ref] [bib_ref] Effects of glucocorticoids on apoptosis of infiltrated eosinophils and neutrophils in rats, Nittoh [/bib_ref] [bib_ref] Opposing effects of glucocorticoids on the rate of apoptosis in neutrophilic and..., Meagher [/bib_ref]. Though pharmacogenetics may display a class effect in various disease, considering the differences between COPD and asthma, the effect of rs37973 variant on corticosteroid responsiveness should be interpreted with caution and further prospective validations will be required in both COPD and asthma.
To the best of our knowledge, only one other genetic variant, rs242941 in the corticotrophin releasing hormone receptor 1 (CRHR1) had been reported as a probable determinant of corticosteroid responsiveness in COPD [bib_ref] Association between CRHR1 polymorphism and improved lung function in response to inhaled..., Kim [/bib_ref]. However, the study of was small with only 12-week follow-up observation in a total of 87 patients and due to lack of further validation, the role of CRHR1 in COPD was still indefinable. Although ICSs are widely or even over prescribed, only very few studies demonstrated markers that might associated with corticosteroid responsiveness in COPD. And most of them are limited to short-term response or uncertain clinical features such as bronchodilator responsiveness [bib_ref] Lung function and symptom improvement with fluticasone propionate/salmeterol and ipratropium bromide/albuterol in..., Bleecker [/bib_ref] , frequent exacerbation 32 , sputum eosinophilia etc [bib_ref] Sputum eosinophilia can predict responsiveness to inhaled corticosteroid treatment in patients with..., Kitaguchi [/bib_ref]. No predictors of long-term responsiveness has been established. Thus our research herein highlights a promising new approach of pharmacogenetics to predict long-term ICSs therapeutic efficacy in COPD.
Cigarette smoking is not only the most important cause of COPD, but also associated with accelerated pulmonary function decline and decreased corticosteroid responsiveness 2,34 . Our research confirmed those of previous studies by showing that smokers had significant poor improvement in FEV 1 after 24-week ICS therapy. Therefore, smoking cessation is an essential therapeutic strategy for COPD.
In vitro functional validation, consistent with previous studies 21-23 , we observed that dexamethasone inhibited spontaneous neutrophil apoptosis in a concentration-dependent manner under standard culture conditions. Interestingly, in the presence of 10 −6 M dexamethasone, which was thought to be the maximal inhibitory concentration [bib_ref] Glucocorticoid treatment inhibits apoptosis in human neutrophils. Separation of survival and activation..., Cox [/bib_ref] , neutrophils isolated from patients with the GG genotype of rs37973 didn't show corresponding reductions in apoptosis with or without CSE milieu. Only when we extended our experimental concentration to 10 −4 M, did we observed an expected inhibitory effort. In keeping with our clinical related findings, those results further validated the negative effect of mutant allele 'G' on corticosteroid responsiveness irrespective of smoking status.
Since apoptosis helps eliminate granulocytes without releasing histotoxic mediators and limit tissue injury, to some extent, an inappropriate delay in apoptosis may be detrimental to control neutrophilic inflammation. Thus, inhibition of neutrophil apoptosis didn't result in concomitant better clinical response, our experiments could only verify neutrophils with GG genotype responded less to glucocorticoids in vitro, neutrophil apoptosis couldn't act as an indicator of changes in lung function after ICS therapy. However, Marwick et al. reported that under severe hypoxic conditions, glucocorticoids lost their ability to promote neutrophil survival and could even reduce the pro-survival effect of inflammatory mediators like GM-CSF, suggesting that glucocorticoids might not account for augmenting neutrophil survival at sites of inflamed tissues [bib_ref] Oxygen levels determine the ability of glucocorticoids to influence neutrophil survival in..., Marwick [/bib_ref]. As potent but nonspecific anti-inflammatory drugs, glucocorticoids in COPD are more likely to exert their anti-inflammatory effects through inhibiting lymphocytic inflammation [bib_ref] Effects of inhaled corticosteroids on airway inflammation in chronic obstructive pulmonary disease:..., Jen [/bib_ref] , increasing epithelial barrier function, decreasing oxidative stress etc [bib_ref] Effect of fluticasone with and without salmeterol on pulmonary outcomes in chronic..., Lapperre [/bib_ref]. The immunopathogenesis of COPD is quite intricate with various inflammatory cells, cytokines and their interactions in different surrounding environment [bib_ref] Immunopathogenesis of chronic obstructive pulmonary disease, Holloway [/bib_ref] , thus further research imitating in-vivo inflammatory environments is clearly needed to investigate the exact anti-inflammatory mechanisms of glucocorticoids in COPD.
However, there are certain limitations in our study. We recruited a selective group of homogeneous COPD patients in a Han Chinese population. Since genetic distribution varies between ethnicities, our results may not generalize to other ethnic groups. Additionally, we only investigated one single potential genetic predictor, which was far from accurately predict ICSs efficacy in COPD. As McGeachie et al. reported that even combination of just two genetic variants, GLCCI1 and CRHR1, could achieve better accuracy in predicting ICSs therapeutic response of asthmatics [bib_ref] Predicting inhaled corticosteroid response in asthma with two associated SNPs, Mcgeachie [/bib_ref]. Although little is known about predictors of corticosteroid responsiveness in COPD, our study herein shed new light on individualized treatment for COPD, indicating that pharmacogenetic may become a promising new approach to predict long-term ICSs efficacy.
Another limitation is that combination therapy with ICS and long-acting β 2 -agonists (fluticasone propionate/salmeterol) was used in our study, ICS therapeutic response might be influenced by the β 2 -agonists in some degree. The exact effect of long-acting β 2 -agonists on ICS is uncertain, but they are believed to have a synergistic interaction [bib_ref] Inhaled combination therapy with long-acting beta 2-agonists and corticosteroids in stable COPD, Cazzola [/bib_ref]. Indeed, COPD patients benefit better from combination therapy than ICS monotherapy, and it's widely used in practice as recommended by guidelines 3 . But then, we further verified the effect of rs37973 variant on corticosteroid responsiveness in vitro cell study. Despite these limitations, we are confident that our principal findings are reliable since the clinical trial and cell experiments consistently indicated the presence of mutant allele 'G' might negatively affect corticosteroid responsiveness irrespective of smoking status.
In conclusion, our study reveals that the GG genotype of rs37973 is associated with decreased ICSs efficacy in clinically stable COPD patients in a Han Chinese population. DNA extraction and genotyping. All participants were genotyped for the rs37973 polymorphism.
Genomic DNA was extracted from 2 ml peripheral EDTA-anticoagulated blood according to the manufacturer's protocol (Blood Genomic DNA Purification Kit; Tiangen Biotech, Beijing, China). Polymerase chain reaction amplification was performed in a total volume of 12.5 μ l, containing 6.25 μ l of Taqman Universal PCR Master Mix (Applied Biosystems Inc., Foster City, CA, USA), 50 ng of extracted DNA and 300 nM of specific probes and primers (Assay ID: rs37973, AHUACTL; Applied Biosystems). Reaction conditions were as follows: 95 °C for 10 min, 45 cycles at 95 °C for 15 s and 60 °C for 1 min. Genotypes were then determined by the Allelic Discrimination Program using the ABI 7900 HT TaqMan sequence detection system (Applied Biosystems). And 10% of samples were randomly genotyped in technical duplicates.
Neutrophil isolation and culture. Neutrophils were isolated from peripheral blood by hydroxyethyl starch sedimentation, Ficoll-Hypaque density gradient centrifugation (TBD, Tianjin, China) and hypotonic lysis of contaminating erythrocytes (Becton Dickinson, Franklin Lakes, NJ, USA). The isolated cells were washed twice and resuspended in RPMI 1640 medium (Life Technologies/GIBCO, Carlsbad, CA, USA) containing 10% fetal bovine serum at 1 × 10 5 cells/ml. Neutrophil viability consistently exceeded 90%, as determined by trypan blue exclusion test. The cell suspension was then transferred equally to 6-well non-adherent culture plates (Corning, NY, USA), and divided into two groups: with or without 10 μ g/ml cigarette smoke extract (Murty Pharmaceuticals, Inc., Lexington, KY, USA). Cells of both groups were stimulated with different concentrations of dexamethasone (Sigma) 0, 10 −6 M and 10 −4 M in a solvent of dimethyl sulphoxide and were incubated in a humidified incubator (95% air, 5% CO 2 , 37 °C) for 18 hours. All experiments were performed in triplicate and repeated three times.
Detection of apoptosis by flow cytometry. After incubation, the percentage of neutrophil apoptosis was quantitatively determined by Annexin V-FITC Apoptosis Detection Kit (KeyGen, Nanjing, China) according to the manufacturer's protocol. In brief, cells were harvested, washed once with ice-cold phosphate-buffered saline and then resuspended in 400 μ l Annexin V binding buffer. After incubating with 4 ul Annexin V-FITC (Ann V) for 10 min at room temperature in the dark, 4 ul propidium iodide (PI) were added and incubated for another 5 min. All the experimental procedures were strictly operated on ice. Samples were then immediately run on a Becton Dickinson LSR flow cytometer with data analyzed by CellQuest software. The normal viable cells were defined as Ann V − /PI − , early apoptotic cells Ann V + /PI − , late apoptotic or necrotic cells were Ann V + /PI + as a result of membrane damage.
Statistical analysis. Categorical variables in baseline characteristics were compared with Pearson chi-square tests and continuous variables with one-way ANOVA, Bonferroni post hoc tests. Hardy-Weinberg equilibrium was analyzed, using chi-square test to compare the observed genotype frequencies with those expected genotype frequencies. The effect of rs37973 variant on changes in lung function after 24 weeks of ICS treatment was assessed by general linear model without adjustment, because of the homogeneous baseline stratified by genotype. Linear regression was used to estimate the effect of per minor allele 'G' of rs37973 (coded as 0, 1, or 2) on changes in FEV.
In addition to rs37973 genotype, stepwise multiple linear regression was performed as an exploratory analysis of factors that could affect ICS therapeutic response, since previous studies have shown that age, sex, baseline percentage of predicted FEV 1 and smoking status affect changes in FEV 1 [bib_ref] Effect of pharmacotherapy on rate of decline of lung function in chronic..., Celli [/bib_ref] [bib_ref] Smoking cessation and lung function in mild-to-moderate chronic obstructive pulmonary disease. The..., Scanlon [/bib_ref]. Smoking status was significant predictor of changes in FEV 1 in our study, and then we repeated our analyses with an additional factorial analysis of Scientific RepoRts | 7:42552 | DOI: 10.1038/srep42552 covariance. Fixed effects were smoking status and rs37973 genotype, with three covariates: age, sex and baseline percentage of predicted FEV Analyses of dexamethasone-mediated neutrophil apoptosis were performed by two-tailed paired t-tests or independent-samples t-tests, when compared within or between the rs37973 genotype. Statistical analyses were performed using IBM SPSS Statistics 22.0 on Windows (SPSS Inc., Chicago, IL, USA) and p value less than 0.05 was considered statistically significant. Data Availability. This study was registered in the Chinese Clinical Trials Registry, http://www.chictr.org. cn/; registration number, ChiCTR-ROB-15005824. Date of registration, 2015-01-05.
[fig] Figure 1: Changes in lung function after 24-week treatment with ICS according to GLCCI1 rs37973 genotype. The GG genotype was significantly associated with poor improvement in (a) FEV 1 and (b) FEV 1 % of predicted, relative to AA and AG genotypes. Data were represented as mean ± SD. *** P < 0.001.Scientific RepoRts | 7:42552 | DOI: 10.1038/srep42552 [/fig]
[fig] Figure 2: Effect of per minor allele 'G' of rs37973 (coded as 0, 1, or 2) on changes in FEV 1 . The equation of fitted linear regression suggested that each additional copy of 'G' allele corresponded to a lower ICS efficacy. [/fig]
[fig] Figure 3: Effects of smoking status and rs37973 genotype on ICS efficacy with adjustment for age, sex and baseline percentage of predicted FEV 1 . Smokers had significant poor mean (± SD) improvement in covariateadjusted FEV 1 , compared with non-smoking patients after 24-week treatment with ICS. *** P < 0.001. Scientific RepoRts | 7:42552 | DOI: 10.1038/srep42552 [/fig]
[fig] Figure 4: Effects of dexamethasone and rs37973 genotype on neutrophil apoptosis. (a) Dexamethasone significantly inhibited neutrophil apoptosis irrespective of genotype in a concentration-dependent manner. (b) Dexamethasone had less inhibitory effect of neutrophil apoptosis on GG genotype, relative to AA and AG genotypes. Data were presented as mean ± SEM; * P < 0.05, ** P < 0.01 (compared with respective solvent control); # P < 0.05.Scientific RepoRts | 7:42552 | DOI: 10.1038/srep42552 [/fig]
[fig] Figure 5: Dexamethasone attenuated CSE-induced neutrophil apoptosis in different genotypes. Neutrophils isolated from COPD patients with the (a) AA genotype, (b) AG genotype or (c) GG genotype were incubated in the presence (+ ) or absence (− ) of Dex or CSE (10 ug/ml) for 18 h. The percentage of neutrophil apoptosis (Ann V + /PI − ) was assessed by flow cytometry. (d) Representative FACS plots of Ann V/PI staining in one sample cultured with various concentrations of Dex (0, 10 −6 M and 10 −4 M), with or without CSE. Data were presented as mean ± SEM. * P < 0.05, ** P < 0.01, *** P < 0.001; # P < 0.05, ## P < 0.01 (compared with 10 −6 M Dex-and CSE-treated cells). Dex, Dexamethasone; CSE, cigarette smoke extract. Scientific RepoRts | 7:42552 | DOI: 10.1038/srep42552 [/fig]
|
Increased CD4+ T cell lineage commitment determined by CpG methylation correlates with better prognosis in urinary bladder cancer patients
Background: Urinary bladder cancer is a common malignancy worldwide. Environmental factors and chronic inflammation are correlated with the disease risk. Diagnosis is performed by transurethral resection of the bladder, and patients with muscle invasive disease preferably proceed to radical cystectomy, with or without neoadjuvant chemotherapy. The anti-tumour immune responses, known to be initiated in the tumour and draining lymph nodes, may play a major role in future treatment strategies. Thus, increasing the knowledge of tumour-associated immunological processes is important. Activated CD4 + T cells differentiate into four main separate lineages: Th1, Th2, Th17 and Treg, and they are recognized by their effector molecules IFN-γ, IL-13, IL-17A, and the transcription factor Foxp3, respectively. We have previously demonstrated signature CpG sites predictive for lineage commitment of these four major CD4 + T cell lineages. Here, we investigate the lineage commitment specifically in tumour, lymph nodes and blood and relate them to the disease stage and response to neoadjuvant chemotherapy. Results: Blood, tumour and regional lymph nodes were obtained from patients at time of transurethral resection of the bladder and at radical cystectomy. Tumour-infiltrating CD4 + lymphocytes were significantly hypomethylated in all four investigated lineage loci compared to CD4 + lymphocytes in lymph nodes and blood (lymph nodes vs tumour-infiltrating lymphocytes: IFNG -4229 bp p < 0.0001, IL13 -11 bp p < 0.05, IL17A -122 bp p < 0.01 and FOXP3 -77 bp p > 0.05). Examination of individual lymph nodes displayed different methylation signatures, suggesting possible correlation with future survival. More advanced post-cystectomy tumour stages correlated significantly with increased methylation at the IFNG -4229 bp locus. Patients with complete response to neoadjuvant chemotherapy displayed significant hypomethylation in CD4 + T cells for all four investigated loci, most prominently in IFNG p < 0.0001. Neoadjuvant chemotherapy seemed to result in a relocation of Th1-committed CD4 + T cells from blood, presumably to the tumour, indicated by shifts in the methylation patterns, whereas no such shifts were seen for lineages corresponding to IL13, IL17A and FOXP3.
## (continued from previous page)
Conclusion: Increased lineage commitment in CD4 + T cells, as determined by demethylation in predictive CpG sites, is associated with lower post-cystectomy tumour stage, complete response to neoadjuvant chemotherapy and overall better outcome, suggesting epigenetic profiling of CD4 + T cell lineages as a useful readout for clinical staging.
Keywords: DNA methylation, CD4-positive T lymphocytes, Urinary bladder neoplasms,
# Background
Urinary bladder cancer (UBC) is the ninth most frequent cancer disease with 380,000 new cases diagnosed worldwide and about 150,000 deaths yearly [bib_ref] Epidemiology of bladder cancer, Malats [/bib_ref] [bib_ref] Global cancer statistics, Torre [/bib_ref]. Environmental factors and life style seem to play an important role for tumour development. Chronic exposure to carcinogenic substances in the urine such as smoking-derived carcinogens, rubber and certain dyes may lead to cancer development [bib_ref] Epidemiology and risk factors of urothelial bladder cancer, Burger [/bib_ref]. In addition, infection with the trematode Schistosoma haematobium leads to chronic inflammation in the urinary bladder and development of squamous cell carcinoma [bib_ref] Epidemiology of urinary bladder cancer: from tumor development to patient's death, Murta-Nascimento [/bib_ref]. Thus, chronic exposure to irritating substances, i.e. chemicals or pathogens, may lead to malignant transformation of cells and finally cancer development. Urothelial muscle invasive bladder cancer is diagnosed (defined as tumour stages T2-T4aN0M0), based on the pathologist's assessment of tumour obtained at transurethral resection of the bladder (TUR-B). Patients judged to be fit according to the Swedish national guidelines are treated with cisplatin-based neoadjuvant combination chemotherapy (NAC), prior to radical cystectomy (RC) typically MVAC (methotrexate, vinblastine, doxorubicin and cisplatin).
UBC development is highly associated with inflammation and immune cell infiltration, an association that provides a basis for immunotherapeutic strategies, such as intravesically administered BCG (Bacillus Calmette-Guerin vaccine) in treatment of high-risk non-muscle invasive bladder cancer (HR-NMIBC) [bib_ref] Bacillus Calmette-Guerin (BCG) immunotherapy for bladder cancer: current understanding and perspectives on..., Kawai [/bib_ref]. We previously demonstrated that the presence of CD3 + tumour-infiltrating T lymphocytes (TIL) is a positive prognostic factor for survival [bib_ref] FOXP3 and survival in urinary bladder cancer, Winerdal [/bib_ref] , supporting the importance of an anti-tumour T cell response. We have also demonstrated that the regional lymph nodes (LNs) contain lymphocytes that are reactive towards the tumour [bib_ref] Sentinel node lymphocytes: tumour reactive lymphocytes identified intraoperatively for the use in..., Marits [/bib_ref] [bib_ref] Detection of immune responses against urinary bladder cancer in sentinel lymph nodes, Marits [/bib_ref] , but that the inter-patient variation of responsiveness to autologous tumour antigen stimulus is highly variable.
The maturation process of T lymphocytes is localized in the thymus through a process of positive and negative selection resulting in CD4 + MHC class II-restricted T cells and CD8 + MHC class I-restricted T cells [bib_ref] Positive and negative selection of T cells, Starr [/bib_ref]. Upon encounter of intermediate affinity/concentration of self-peptides in the thymic medulla, naïve CD4 + T cells are converted to Foxp3 stably expressing regulatory T cells (Treg). CD4 + T cells emerging from the thymus pass into the periphery and circulate various tissues. Upon encountering their cognate antigen in a tumour setting, the pattern of their maturation and differentiation into separate CD4 + T cell lineages will be decided by the combined signals from the antigen-presenting cells, tumour cells and stroma cells present in this distinct environment. The main CD4 + T cell effector lineages are Th1, Th2 and Th17, as recognized by their production of effector cytokines IFN-γ, IL-13 and IL-17A, respectively [bib_ref] Differentiation of effector CD4 T cell populations (*), Zhu [/bib_ref]. Upon activation and proliferation, naïve T cells transform to differentiated effector cells with a stable phenotype that is difficult to reverse after five cell divisions [bib_ref] Reversibility of T helper 1 and 2 populations is lost after long-term..., Murphy [/bib_ref] [bib_ref] Differentiation and stability of T helper 1 and 2 cells derived from..., Sornasse [/bib_ref]. However, plasticity among committed T cell subpopulations have started to be explored [bib_ref] Functional plasticity of an antigen-specific memory CD4 T cell population, Ahmadzadeh [/bib_ref] [bib_ref] Genetic reprogramming of primary human T cells reveals functional plasticity in Th..., Sundrud [/bib_ref] [bib_ref] Effector T cell plasticity: flexibility in the face of changing circumstances, Murphy [/bib_ref]. Long-term epigenetic stability of a T cell phenotype can be evaluated using methylation markers at predictive CpG sites [bib_ref] DNA demethylation in the human FOXP3 locus discriminates regulatory T cells from..., Baron [/bib_ref] [bib_ref] FOXP3 promoter demethylation reveals the committed Treg population in humans, Janson [/bib_ref] [bib_ref] T helper cell differentiation: regulation by cis elements and epigenetics, Lee [/bib_ref] [bib_ref] Profiling of CD4+ T cells with epigenetic immune lineage analysis, Janson [/bib_ref].
We and other groups have previously investigated the methylation status of the IFNG locus in CD4 + T cells from LNs and tumours. We have also developed methods for investigating the IL13, FOXP3 and IL17A loci to make a global CD4 + T cell assessment regarding epigenetic commitment [bib_ref] FOXP3 promoter demethylation reveals the committed Treg population in humans, Janson [/bib_ref] [bib_ref] Profiling of CD4+ T cells with epigenetic immune lineage analysis, Janson [/bib_ref] [bib_ref] CpG methylation of the IFNG gene as a mechanism to induce immunosuppression..., Janson [/bib_ref] [bib_ref] DNA methylation changes at human Th2 cytokine genes coincide with DNase I..., Santangelo [/bib_ref].
Based on previous experiences and results, we performed a snapshot analysis of the in vivo epigenetic commitment of CD4 + T cell populations in samples from patients with UBC, using DNA methylation pattern of epigenetic lineage markers for Th1, Th2, Th17 and Tregs predictive for assessing CD4 + T cell subpopulation stability [bib_ref] Profiling of CD4+ T cells with epigenetic immune lineage analysis, Janson [/bib_ref]. Further, we correlate our findings with clinical response to neoadjuvant chemotherapy and pathological tumour stage post-cystectomy.
# Methods
## Patient inclusion and clinical procedure
All patients were included in this study after giving their written and oral consent to participate, in accordance with the declaration of Helsinki. The study was approved by the local ethical committee (dnr: 2007/71- . Recruitment was performed between 2014 and 2017 from nine Swedish hospitals (Umeå University Hospital, Sundsvall Hospital, Västerås Central Hospital, Linköping University Hospital, Norrköping Hospital, Skellefteå Hospital and Gävle Hospital, Uppsala Akademiska University Hospital and Östersund County Hospital). The patients in the study were included either before TUR-B with suspected urinary bladder cancer, or before RC, after an established muscle invasive bladder cancer (MIBC) (tumour stages cT2-4aN0M0), as determined by pathologist's assessment of previously performed TUR-B. Patient characteristics are presented in [fig_ref] Table 1: Patients included in this studyNAC was administered to indicated patients Age, in... [/fig_ref]. Tumour and blood samples were collected at TUR-B from 23 patients. Samples were also collected from 21 patients at RC, including blood, tumour draining sentinel lymph nodes (SN), non-draining lymph nodes (nSN) and, in cases with remaining tumour at this stage, tumour tissue. The method for sentinel node detection has previously been described [bib_ref] Lymphatic mapping and detection of sentinel nodes in patients with bladder cancer, Sherif [/bib_ref]. From six patients, additional blood samples were obtained during NAC treatment (in-between NAC cycles). The total number of specimens was as follows: blood, n = 48; regional LN (both SN and nSN), n = 76; and tumour, n = 22. Not all specimens were analysed for every parameter, due to sample limitations.
## Lymphocyte extraction
Peripheral blood mononuclear cells (PBMCs) from blood were extracted using Ficoll paque PLUS (GE Healthcare).
Tumour-infiltrating lymphocytes were extracted by cutting tumour tissue into small pieces and disassociating the dissected tumour pieces into single cells using a gentleMACS dissociator (Miltenyi Biotec). Samples were processed in GIBCO Aim V™ (Invitrogen) supplemented with collagenase/hyaluronidase (STEMCELL Technologies). Subsequently, the single cell suspension was strained through a 40-μm strainer to exclude remaining tumour cell aggregates and tissue debris.
Lymph nodes were gently homogenized by straining through a 40-μm strainer.
## Cd4 + t lymphocyte purification
CD4 + T lymphocytes from PBMC, lymph nodes and tumours were sorted in two steps: (1) pre-sorting of total CD3 + cells was performed using EasySep Human CD3-positive selection kit II (STEMCELL Technologies). CD3 + pre-sorted cells were then stained with anti-human CD4 (PerCp Cy5.5, BioLegend), anti-human CD8 (APC, BD Biosciences), anti-human CD56 (PE, BD Biosciences), anti-human CD45RA (V500, BD Biosciences) and anti-human C45RO (APC CY7 BioLegend). (2) CD4 + cells were FACS-purified according to gating strategy presented in using a BD FACSARIA I instrument (BD Biosciences). The purity of CD4 + T lymphocytes was confirmed post-sorting and was consistently ≥ 95%. Analysis of flow cytometry data was performed using the FACS Diva software (BD Biosciences) and FlowJo (version 10, FlowJo LLC). Sorted cells were pelleted and stored in − 20°C until further analysis.
Pyrosequencing of CD4 + lymphocytes DNA from CD4 + cell pellets were extracted and bisulfite converted using EZ DNA methylation Direct kit (Zymo Research). The locus-specific pyrosequencing PCRs for FOXP3, IFNG IL13 and IL17A were conducted using primers where one primer for each PCR reaction was biotinylated (Thermo Scientific and Biomers.net) (Additional file 1: [fig_ref] Table 1: Patients included in this studyNAC was administered to indicated patients Age, in... [/fig_ref]. The PCRs were performed using a Thermal cycler (Bio-Rad). The PCR product was immobilized using the PyroMark Q96 vacuum workstation (Qiagen). The sequencing reaction was performed on a Pyro Q96, using Pyromark Gold 96Q reagents (Qiagen) and PCR assay-specific sequencing primers (Additional file 1: [fig_ref] Table 1: Patients included in this studyNAC was administered to indicated patients Age, in... [/fig_ref]. Analysis of the sequence data were performed by the Pyromark Q96 ID software (Qiagen) giving individual percentage for assayed CpGs. Graphic visualization of the four loci was made using VISTA-point [bib_ref] VISTA : visualizing global DNA sequence alignments of arbitrary length, Mayor [/bib_ref]. Histograms demonstrate species conservation between human and mouse, and circles below schematically demonstrate CpG sites in the specific region. The analysed signature CpG site is marked in red with their location, in base pairs, from transcription start site (TSS) indicated.
## Cisplatin cultures
CD4 + T lymphocytes were extracted from peripheral blood of healthy donors (n = 4) as described above. Cells were put in cultures at a concentration of 1 × 10 6 cells/ml in GIBCO Aim V™ medium, research grade (Invitrogen) supplemented with L-glutamine (Sigma-Aldrich). Cells were stimulated at day 0 or day 6, with 5 μg/ml plate bound αCD3 (BioLegend) and 2 μg/ml soluble αCD28 (BioLegend). Cisplatin (Hospira, Pfizer) was added on day 0 or day 6 of culture at a concentration of 25 μM (LD 50 [bib_ref] The effects of chemotherapeutic drugs on human monocytederived dendritic cell differentiation and..., Hu [/bib_ref] , for subsequent ELISA) or 50 μM (for subsequent pyrosequencing). All cultures were incubated at 37°C in 5% CO 2 . Cells were harvested at day 12.
## 5-methylcytosine elisa
DNA from cultures exposed to 25 μM cisplatin was extracted using DNeasy blood and tissue kit (Qiagen). DNA quantity and quality were measured on a Nanodrop instrument (ThermoFisher Scientific), and 100 ng DNA/ well was utilized to perform ELISA. 5-Methylcytosine (5mC) ELISA was conducted, and each sample was run in duplicate, using 5-mC DNA ELISA kit (ZYMO research) according to the manufacturer's instructions. Sample values were normalized against either day 0 untreated samples or day 12 samples without cisplatin treatment.
# Statistical analysis
All statistical analysis was performed using GraphPad PRISM version 7.04. Non-parametric Mann-Whitney test, Kruskal-Wallis test and Friedmans test was used where applicable. Dunn's multiple comparisons test was employed when suitable. Statistical significance in graphs are shown as * if p < 0.05, ** if p < 0.01, *** if p < 0.001 and **** if p < 0.0001. Plots show mean with standard error of mean (SEM). No statistical calculations were made on data in [fig_ref] Figure 8: Investigation of sample retrieved during NAC treatment [/fig_ref] , c, d due to low sample number.
# Results
Samples from patients with urinary bladder cancer, presented in [fig_ref] Table 1: Patients included in this studyNAC was administered to indicated patients Age, in... [/fig_ref] , were analysed. Single cell suspensions from PBMC, tumour and lymph nodes were stained and sorted by flow cytometry . FACS analysis post-sorting demonstrated ≥ 95% purity . DNA was extracted and bisulfite converted for pyrosequencing and assessment of signature CpG sites for analysis of lineage commitment to the Th1, Th2, Treg and Th17 lineages, using previously identified predictive sites in the corresponding genes IFNG, IL13, FOXP3 and IL17A [bib_ref] FOXP3 promoter demethylation reveals the committed Treg population in humans, Janson [/bib_ref] [bib_ref] Profiling of CD4+ T cells with epigenetic immune lineage analysis, Janson [/bib_ref] [bib_ref] CpG methylation of the IFNG gene as a mechanism to induce immunosuppression..., Janson [/bib_ref] [bib_ref] DNA methylation changes at human Th2 cytokine genes coincide with DNase I..., Santangelo [/bib_ref].
## Evaluation of cd4 + t cell lineage commitment
The four selected loci were investigated in CD4 + T lymphocytes sorted from PBMCs, LNs and tumours obtained at TUR-B and RC. Comparisons of sentinel and non-sentinel lymph node data demonstrated no significant differences (data not shown), and thus, these specimens were analysed Sorting strategy and loci visualization. a CD3 + cells derived from PBMC, lymph nodes or tumour tissue were sorted by flow cytometry using a FACSARIA. Gating strategy was done as follows: singlets > lymphocytes > CD56 − > CD4 + . CD4 + CD56 − . Purity analysis is presented in percentage of parent. Plots show representative best. b Signature loci visualized by VISTA-plots. The conservation of the assessed gene loci between human and mouse is depicted relative to the genes. The conservation is defined as stretches of nucleotides more than 100 bp with over 70% conserved area. The magnified area schematically describes the investigated regions. Analysed CpG is marked as red circle, describing its position in relation to transcription start site (TSS). CNS1 conserved non-coding sequence 1, UTR untranslated region together as a group referred to as "lymph nodes" (LN) throughout the analysis. The IFNG locus methylation in the CpG position -4229 bp from transcription start site (TSS) (located in the conserved non-coding sequence 1 (CNS1)) was used to assess the Th1-committed CD4 + T cells [bib_ref] CpG methylation of the IFNG gene as a mechanism to induce immunosuppression..., Janson [/bib_ref]. CD4 + T cells from TILs were significantly more demethylated in the IFNG locus compared to LN (p < 0.0001), whereas the methylation in LNs were higher compared to PBMC, suggesting an increased infiltration of Th1 IFN-γ producing CD4 + T cells into the tumour [fig_ref] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer [/fig_ref]. With regard to Th2-committed CD4 + T cells, the IL13 locus was evaluated at the signature CpG position -11 bp from TSS as previously demonstrated [bib_ref] DNA methylation changes at human Th2 cytokine genes coincide with DNase I..., Santangelo [/bib_ref]. Again, we found a significantly decreased level of methylation in the IL13 locus in TILs compared to lymph nodes (p < 0.05) [fig_ref] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer [/fig_ref]. However, there were no significant changes when comparing CD4 + T cells derived from TILs with those from PBMCs. Interestingly, CD4 + T cells from LN demonstrated significantly increased methylation in the IL13 locus compared with both PBMC (p < 0.05) and TILs (p < 0.05). CD4 + Tregs were assessed at the CpG -77 bp from TSS in the FOXP3 locus as previously described [bib_ref] FOXP3 promoter demethylation reveals the committed Treg population in humans, Janson [/bib_ref]. This signature CpG was significantly hypomethylated in CD4 + T cells from lymph nodes (p < 0.05) and tumours (p < 0.01) compared to PBMC [fig_ref] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer [/fig_ref]. Finally, Th17 cells were investigated at the IL17A signature locus at the CpG -122 bp from TSS [bib_ref] Profiling of CD4+ T cells with epigenetic immune lineage analysis, Janson [/bib_ref]. CD4 + T cells from tumours demonstrated significantly decreased methylation compared to LN (p < 0.05) and PBMCs (p < 0.0001) [fig_ref] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer [/fig_ref]. No significant difference was seen when comparing CD4 + T cells from PBMC derived from either TUR-B or RC, rationalizing the equivalence between the samples obtained at the two time points (Additional file 2: [fig_ref] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer [/fig_ref].
## Analysis of lineage commitment at the time of tur-b
Patient materials obtained at the time of TUR-B were analysed separately. CD4 + T cells purified from fresh TUR-B tumour resections did not demonstrate any difference in methylation profile in the IFNG locus compared to PBMCs (p > 0.05) [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref]. Neither was there any difference in the degree of methylation in the IL13 locus [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref]. With regard to Tregs, the FOXP3 signature locus was significantly demethylated in CD4 + T cells from the tumour compared to PBMCs (p < 0.001) [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref]. Finally, we found a decreased methylation of the IL17A locus in CD4 + T cells in the tumour compared to PBMCs (p < 0.01) [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref]. One patient (no. , included at TUR-B had benign disease and was hence not included in the general data analysis. We examined the TILs from this patient and found the IFNG locus to be highly demethylated, whereas the three other loci were hypermethylated, suggesting a Th1 lineage commitment [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref]. For comparison, we examined TILs from a patient (no. 28) with hypermethylation at the IFNG locus [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref]. This patient (cT stage cT2a) displayed hypermethylation in all four loci, corresponding to an overall low lineage differentiation.
## Analysis of lineage commitment at the time of cystectomy
CD4 + T cells derived from samples obtained at the time of RC were analysed at the four signature loci. In LN, the methylation at the IFNG locus was significantly increased, compared to the corresponding cells from PBMC (p < 0.05) [fig_ref] Figure 4: Methylation analysis in specimens from radical cystectomy [/fig_ref]. Similarly, the level of methylation was increased at the IL13 locus in lymph nodes compared to blood (p < 0.05) [fig_ref] Figure 4: Methylation analysis in specimens from radical cystectomy [/fig_ref]. No differences were seen in the methylation levels of FOXP3 or IL17A loci between CD4 + T cells derived from PBMC and LN [fig_ref] Figure 4: Methylation analysis in specimens from radical cystectomy [/fig_ref].
CD4 + T cells from five lymph nodes (indicated by red circles in were selected for their high or low methylation profiles at the IFNG locus and were individually investigated for all four signature loci. The two specimens with low methylation at the IFNG locus also demonstrated a demethylated pattern in the Treg locus, while no signs of Th2 and Th17 skewing were found: i.e. IL13 and IL17A signature loci were almost completely methylated , patient no. 6 and 24). On the contrary, the samples demonstrating a high methylation pattern in the IFNG locus displayed more of a Treg/Th2 or Treg/Th17 commitment judged by the methylation profiles in signature loci , patient no. 7 and 2). One LN revealed low commitment for all four loci compared to the other four LNs investigated , patient no. 20). The patients with LNs displaying low IFNG methylation, and therefore a Th1 signature, had a lower pathological tumour staging (pT stage), pTa-TisN1 (patient 6) and pT0 (patient 24) , although the former had a node metastasis (not included in specimens). The patients with LNs highly methylated at the IFNG locus had a more advanced disease stage, pT3aN2, pT3a and pT2a, respectively (patient, 7, 2 and 20) .
In order to epigenetically stage the collective immune response in a single patient (no. 6), the CD4 + T cell compartment in three individual LNs were analysed . LN1 demonstrated a Th1/Treg pattern, whereas the other two LNs (2 and 3) displayed different degrees of Treg/ Th17 skewing, although still with a fraction of Th1commitment .
## Cd4 + t cells from ln are differentially committed when stratified over pt stage
The pT stage determined by histopathology following cystectomy is known to predict prognosis and to function as a surrogate marker for overall survival in MIBC patients undergoing NAC [bib_ref] Pathologic downstaging is a surrogate marker for efficacy and increased survival following..., Rosenblatt [/bib_ref]. Complete response (CR), i.e. . p values are indicated as *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001, using Mann-Whitney test. Bars indicate mean with error bars displaying SEM. Downward arrow along y-axis illustrate increased lineage commitment, as a result of decreased methylation pT0N0M0 at RC, corresponds to an excellent long-term survival. We evaluated the lineage commitment in the four loci according to pT staging. The IFNG locus demonstrated a demethylated pattern in the primary tumours (at TUR-B) of pT0 patients (complete NAC-responders) and in non-invasive tumours (pTa-Tis) compared to in primary tumours with muscle invasive tumour outcomes post-RC (pT2) (p < 0.0001 resp. p < 0.001) . The methylation was decreased in the perivesical infiltrating tumours (pT3) compared to the muscle invasive pT2 tumours (p < 0.05). In the IL13 locus, the CD4 + T cells demonstrated an increased methylation in muscle invasive pT2 tumours compared to the cells from patients with pT0 stage . Methylation levels at the FOXP3 locus was increased in LN CD4 + T cells from patient with muscle invasive pT2 compared to both non-muscle invasive pTa-Tis staged patients Case studies of lymph nodes with low or high methylation in IFNG locus. a Same as [fig_ref] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer [/fig_ref] , with evaluated lymph nodes circled in red. bd Bar graphs depicts percentage of demethylation levels at the four signature sites, representing increase in lineage commitment. Th1 = IFNG, Th2 = IL13, Treg = FOXP3, Th17 = IL17A. b LN from three patients with high IFNG methylation. c LNs from two patients with low IFNG methylation patterns. d Case study of three LNs from patient no. 6, chosen for a distinct IFNG demethylation in one node (from and to the cells from patients with perivesical infiltrating tumours (pT3) (p < 0.05 for both) . In IL17A, the methylation was increased in LN-derived CD4 + T cells from patients staged with pT2 compared to the pT0 as well as non-muscle invasive pTa-Tis staged patients (p < 0.05 for both) .
## Methylation patterns in lymph nodes correspond with response to neoadjuvant chemotherapy
Patients were stratified according to their clinical response to NAC, and LN-derived CD4 + T cells from patients with complete response (CR; pT0N0M0) and those with no response (NR; pT ≥ 2N0M0) were compared. Methylation was significantly lower in the group with CR compared to those with NR in all four loci, with the most pronounced difference in the IFNG locus (IFNG p < 0.0001, IL13 p < 0.0001, FOXP3 p < 0.01, IL17A p < 0.001) [fig_ref] Figure 7: Lymph nodes grouped according to NAC response [/fig_ref]. When stratifying PBMC or LN samples according to their corresponding clinical T stage, it revealed no significant differences (data not shown).
## Th1 lineage affected during nac treatment
Patients with MIBC, and in WHO class 0-1 health condition, are according to the national Swedish guidelines recommended to receive 3-4 cycles of cisplatin-based NAC prior to RC. Blood samples were obtained after first or second cycle of NAC therapy and CD4 + T cells from PBMCs were extracted. Analysis of IFNG locus methylation at TUR-B, during NAC and at the time of RC demonstrated a temporary increase in the methylation status (n = 3) during NAC treatment [fig_ref] Figure 8: Investigation of sample retrieved during NAC treatment [/fig_ref]. When comparing PBMC derived CD4 + T lymphocytes from TUR-B with corresponding cells obtained during chemotherapy for an additional three patients, we found a tendency to increase methylation in the IFNG locus, however not significant [fig_ref] Figure 8: Investigation of sample retrieved during NAC treatment [/fig_ref]. Interestingly, when investigating the FOXP3 and IL17A loci at the same time LN-derived CD4 + T cells stratified for pT stage. CD4 + T cells from lymph nodes stratified according to the patients' pT stage. Plots show percentage of methylation at signature CpG, evaluated by pyrosequencing, for every lymph node analysed. Lymph node number (n) is stated on the x-axis beneath sample type. Group pT1 was excluded throughout all statistical analysis due to low sample number. a IFNG locus (Kruskal-Wallis test p < 0.0001), b IL13 locus (Kruskal-Wallis test p < 0.001), c FOXP3 locus (Kruskal-Wallis test p < 0.05) and d IL17A locus (Kruskal-Wallis test p < 0.01). p values from Dunn's multiple comparisons test are indicated as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Bars indicate mean with error bars displaying SEM. Downward arrow along y-axis illustrates increased lineage commitment, as a result of decreased methylation points, no change in methylation status was noted [fig_ref] Figure 8: Investigation of sample retrieved during NAC treatment [/fig_ref]. To investigate if the change in the IFNG locus methylation signature was altered due to a recruitment of naïve T cells into the circulation, we calculated the ratio between naïve CD45RA + CD45RO − and memory CD45RO + T cells. We found no significant difference in the CD45RA/CD45RO ratio or in the fraction of CD45RA single positive cells between TUR-B and cystectomy (after NAC) samples [fig_ref] Figure 8: Investigation of sample retrieved during NAC treatment [/fig_ref] , suggesting that the hypermethylation of the IFNG locus is not due to an increased inflow of naïve CD4 + T cells into the circulation [fig_ref] Figure 8: Investigation of sample retrieved during NAC treatment [/fig_ref].
## Cisplatin does not affect the dna methylation
Cisplatin is the base of the neoadjuvant treatment, acting partly through cross binding of DNA (purine bases) [bib_ref] Binding of the antitumor drug cisdiamminedichloroplatinum(II) (cisplatin) to DNA, Pinto [/bib_ref] [bib_ref] The formation, isolation and characterization of DNA adducts produced by anticancer platinum..., Eastman [/bib_ref]. To investigate the possibility that cisplatin affects DNA methylation, in vitro cultures were established using CD4 + T lymphocytes from healthy donors (n = 4). 5mC ELISA showed no significant difference in the global methylation of cells treated with cisplatin (Additional file 3: . In addition, pyrosequencing of the IFNG locus revealed no significant difference in the site-specific methylation (Additional file 3: .
# Discussion
We here demonstrate that valuable information can be obtained by studying signature CpG methylation, indicating the degree of lineage commitment of CD4 + T cells in tissues from UBC patients. CD4 + T cell lineage commitment was found to be more pronounced in tumour-infiltrating lymphocytes (TILs) compared to PBMC and regional lymph nodes, indicating that differences in the tissue environments have a significant impact on CD4 + T cell destiny. In addition, we found correlations between increased lineage commitment of CD4 + T cells in LNs after neoadjuvant chemotherapy and an improved prognosis, indicating an important role for T cell immunity in the evolution of UBC towards more aggressive forms. To the best of our knowledge, this is the first time DNA methylation of CD4 + T cell lineage markers has been investigated in cells harvested from patients with UBC, and our data suggests that this type of investigation can contribute towards a deeper understanding of the role of the immune system in the UBC setting. . c FOXP3 (number of patients: CR n = 6, NR n = 2). d IL17A (number of patients: IL17A CR n = 6 NR n = 2). Mann-Whitney test was used for the statistical analysis. p values are indicated as *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. Bars indicate mean with error bars displaying SEM As a method, methylation analysis has advantages when compared to analysis of protein and mRNA expression. No stimulation of the cells is required in order to analyse the current methylation status; and thus, resting unmanipulated cells can be analysed for phenotype stability, with no risk of misinterpreting temporary, transient protein expression for stable effector lineages. This is a clear advantage when examining primary cells from clinical specimens since no tampering of the cells is needed. However, epigenetic status does not convey if the cells are active or resting, but rather indicate lineage commitment and effector capacity of the cells upon activation.
We demonstrate that TILs have a high degree of CD4 + T cells with lineage commitment [fig_ref] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer [/fig_ref] , proposing an active immune response towards the tumour. The IFNG-committed Th1 compartment is the most prominent of the four T cell lineages examined here, indicating that Th1 is the major lineage response towards the tumour, which is in agreement with the literature [bib_ref] The critical role of Th1-dominant immunity in tumor immunology, Nishimura [/bib_ref]. Furthermore, the methylation pattern in the IFNG locus has the greatest variability, compared to the other three loci, in all investigated tissues, which we interpret as both intra-and inter-patient variation [fig_ref] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer [/fig_ref]. The Treg compartment was the only lineage with a gradual decrease in methylation, with the highest methylation in blood, through lymph nodes to the lowest methylation in tumour. This could be interpreted as an effect of the immune-stimulatory environment, where high proliferation rate leads to an increase in the Treg population [bib_ref] FOXP3 and survival in urinary bladder cancer, Winerdal [/bib_ref]. In our patient samples, we assumed that a prolonged immune response towards the tumour would be present, and this seemed to lead to a co-commitment of several stable lineages, mainly Th1 and Tregs [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref]. It becomes evident, when individually examining the four lineages in separate samples, that the specimens low in IFNG methylation had a clear Th1/Treg profile in both TILs and LN (Figs. 3e top bars and 5c), whereas the highly methylated specimens inclined more towards a Th2 or Th17 profile (Figs. 3e bottom bars and 5b), or an overall low lineage commitment pattern , patient 2). Noteworthy, the specimen with the most prominent demethylation in IFNG locus had a benign tumour, suggesting a protective IFN-γ response limiting the progress of the tumour [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref].
When separating the material according to the two time points of intervention (TUR-B and RC), the CD4 + T cells were found to display a higher degree of lineage commitment in TILs from the TUR-B [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref] compared to those from blood, while the LN CD4 + T cells at the time of cystectomy showed less lineage commitment or no difference compared to their blood counterparts [fig_ref] Figure 4: Methylation analysis in specimens from radical cystectomy [/fig_ref]. We suggest that this is due to the selective migration of activated T cells away from the LN towards the tumour.
Stratifying CD4 + T cells from LNs according to the patients' pT stage revealed differences in methylation state relating to local tumour responses to NAC, which is correlated with disease progression. The level of committed hypomethylated CD4 + T cells were increased in patients with low, non-muscle invasive pT stages as compared to those from patients with muscle invasive pT2 tumours , which was most prominent in Th1 cells. These data suggest that an increase in committed Th1 effector cells in the tumour region is favourable for prognosis. Surprisingly, both Th1 and Treg commitment is increased in pT3, when the tumours have progressed to perivesical infiltration (a clear sign of a lack of response to NAC), perhaps suggesting that the balance between regulation and effector function has switched in the environment of more aggressive tumour cells.
The response to NAC in terms of histopathological tumour regression is a major positive prognostic factor for overall survival following radical cystectomy [bib_ref] Pathologic downstaging is a surrogate marker for efficacy and increased survival following..., Rosenblatt [/bib_ref] [bib_ref] EAU guidelines on muscle-invasive and metastatic bladder cancer: summary of the 2013..., Witjes [/bib_ref]. Therefore, it was striking that the lineage commitment in LNs from patients with CR after NAC was significantly higher than in those not responding as favourably [fig_ref] Figure 7: Lymph nodes grouped according to NAC response [/fig_ref]. Our group has also previously demonstrated that cisplatin-based NAC induces immune-stimulatory effects [bib_ref] The effects of chemotherapeutic drugs on human monocytederived dendritic cell differentiation and..., Hu [/bib_ref] [bib_ref] Doxorubicin enhances the capacity of B cells to activate T cells in..., Zirakzadeh [/bib_ref] and the present findings are in line with that context.
The lack of paired patient samples from TUR-B and RC prevented us from investigating the methylation pattern in pre-treatment TILs from the patients who received NAC and correlate to subsequent NAC response in corresponding LNs. Instead, we examined the methylation pattern in paired blood samples, obtained during the course of NAC treatment. During the NAC treatment (post-chemo), PBMC-derived CD4 + T cells demonstrated a tendency towards increased methylation in the IFNG locus [fig_ref] Figure 8: Investigation of sample retrieved during NAC treatment [/fig_ref] compared to paired blood samples acquired before NAC (at time of TUR-B) as well as after NAC (at time of RC). There was no indication of changes in FOXP3, IL17A or IL13 methylation between these time points (Figs. 7c and 8d, data not shown). We further establish that the changes in DNA methylation are not a direct effect of cisplatin, by performing both a whole genome 5mC ELISA to ensure that this agent does not affect methylation on a global level, and an IFNG locus-specific analysis, for comparison with our data (Additional file 3: . Since the CD45RA/CD45RO ratio between the two occasions of TUR-B and RC was not changed [fig_ref] Figure 7: Lymph nodes grouped according to NAC response [/fig_ref] , we propose that the temporary increase of methylation in the IFNG locus is not due to the recruitment of naïve cells, but rather to the migration/relocation specifically of Th1 lineage-committed T cells towards the tumour environment. This hypothesis is also supported by the lack of effect on the other lineages [fig_ref] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs... [/fig_ref] data not shown). It has been demonstrated in various cancers that the tumour microenvironment expresses CXCL10 (IP-10), which leads to Th1-specific recruitment to the site mediated by the CXCL10 receptor CXCR3 expressed on Th1 cells. [bib_ref] Interferon-inducible protein-10 identified as a mediator of tumor necrosis in vivo, Sgadari [/bib_ref] [bib_ref] The emerging role of CXCL10 in cancer (Review), Liu [/bib_ref] [bib_ref] CXCL10 expression and prognostic significance in stage II and III colorectal cancer, Jiang [/bib_ref]. Thus, the changes in the proportion of Th1 cells specifically in different compartments may be based on their unique ability for tumour infiltration.
Our data indicates that NAC plays a role in activating and steering the immune system towards the tumour. It is tempting to speculate that patients with hypomethylation in the IFNG locus have better prognosis than those with hypermethylation. The short period of time from sample collection until present day prevents us from evaluating clinical parameters such as 5-year survival and relapse rates. However, histopathological response to NAC is a positive predictor of survival and, although a cause-and-effect relationship cannot be established, the data suggests that increased proportions of committed CD4 + T cells in the tumour regional lymph nodes after NAC translates into a better outcome. Further
[fig] Figure 2: CD4 + T cell lineage commitment in urinary bladder cancer. CD4 + T cells sorted from PBMC, LN and TILs were analysed by pyrosequencing at CD4 + T cell signature loci: a IFNG for Th1, b IL13 for Th2, c FOXP3 for Treg and d IL17A for Th17. p values stated in graphs are generated from Kruskal-Wallis test. p values from Dunn's multiple comparisons test are indicated as *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. Plots show percentage of methylation for every specimen analysed (n stated on x-axis beneath sample type). Bars indicate mean with error bars displaying SEM. Downward arrow along y-axis illustrate increased lineage commitment, as a result of decreased methylation30) [/fig]
[fig] Figure 3: DNA methylation analysis of CD4 + T cells from PBMC and TILs at time of TUR-B. Methylation levels of a IFNG, b IL13, c FOXP3 and d IL17A signature loci in CD4 + T cells from PBMC and TILs were analysed by pyrosequencing at time of TUR-B. Plots show percentage of methylation for every specimen analysed (n stated on x-axis beneath sample type). p values are indicated as *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001, using Kruskal-Wallis test. Bars indicate mean with error bars displaying SEM. Downward arrow along y-axis illustrate increased lineage commitment, as a result of decreased methylation. e Bar graphs depicting demethylation levels at the four signature sites (Th1 = IFNG, Th2 = IL13, Treg = FOXP3, Th17 = IL17A) from two patients, selected for their low vs high methylation pattern in IFNG. TILs from patient no. 30 with low methylation levels in IFNG (top bars) and patient no. 28 with high methylation in IFNG [/fig]
[fig] Figure 4: Methylation analysis in specimens from radical cystectomy (RC). Methylation percentage in CD4 + T cells from PBMC and lymph nodes retrieved at time of RC. IFNG (a), IL13 (b), FOXP3 (c) and IL17A (d) were analysed. Plots show percentage of methylation for every specimen analysed (n stated on x-axis beneath sample type) [/fig]
[fig] Figure 7: Lymph nodes grouped according to NAC response. CD4 + T cells from lymph nodes sorted according to the patients' responses to neoadjuvant chemotherapy (NAC); complete response (CR; pT0N0M0) or no response (NR; pT2 or higher N0M0) Plots show percentage of methylation at signature CpG, evaluated by pyrosequencing, for every lymph node analysed. Lymph node number (n) stated on x-axis beneath sample type. a IFNG (number of patients: CR n = 8, NR n = 3). b IL13 (number of patients: CR n = 4, NR n = 2) [/fig]
[fig] Figure 8: Investigation of sample retrieved during NAC treatment. a Examination of IFNG locus in CD4 + T cells from blood of three patients, taken at three time points: before (TUR-B), during (post-chemo) and after (cystectomy) neoadjuvant chemotherapy. b Methylation pattern in IFNG locus in paired samples from two time points, before (TUR-B) and during (post-chemo) NAC treatment (n = 6). Wilcoxon test was used for statistical analysis. c Methylation pattern in FOXP3 locus in paired samples from two time points, before (TUR-B) and during (post-chemo) NAC treatment (n = 3). d Methylation pattern in IL17A locus in paired samples from two time points, before (TUR-B) and during (post-chemo) NAC treatment (n = 3). e CD45RA/RO protein expression ratio in PBMC derived CD4 + T cells from TUR-B or cystectomy, as evaluated by flow cytometry. f Percentage of CD45RA + CD45RO − (single positive) CD4 + T cells, determined by flow cytometry, at the two time points TUR-B and RC. e, f (n stated on x-axis beneath sample type) Bars show SEM [/fig]
[table] Table 1: Patients included in this studyNAC was administered to indicated patients Age, in years at time of inclusion; gender, M male, F female; sampling, T TUR-B, C radical cystectomy, Ch during chemo blood (post-chemo); NAC responder, CR complete responder, NR non-responder, PR partial responder, prog. progression; LN, number of lymph nodes obtained indicated, Tumour, specimens acquired from indicated patients. Year, years of intervention. -, no data/sample available Groups of data were consistently excluded from statistical analysis if n = < 4. [/table]
|
Early Childhood Curiosity and Kindergarten Reading and Math Academic Achievement
BACKGROUND-Although children's curiosity is thought to be important for early learning, the association of curiosity with early academic achievement has not been tested. We hypothesized that greater curiosity would be associated with greater kindergarten academic achievement in reading and math.METHODS-Sample included 6200 children in the Early Childhood Longitudinal Study, BirthCohort. Measures at kindergarten included direct assessments of reading and math, and a parentreport behavioral questionnaire from which we derived measures of curiosity and effortful control. Multivariate linear regression examined associations of curiosity with kindergarten reading and math academic achievement, adjusting for effortful control and confounders. We also tested for moderation by effortful control, sex, and socioeconomic status (SES).RESULTS-In adjusted models, greater curiosity was associated with greater kindergarten reading and math academic achievement: b reading =0.11, p<.001; b math =0.12, p<.001. This association was not moderated by effortful control or sex, but was moderated by SES (p reading =.01; p math =.005). The association of curiosity with academic achievement was greater for children with low SES (b reading =0.18, p<.001; b math =0.20, p<.001), versus high SES (b reading =0.08, p=.004; b math =0.07, p<.001).CONCLUSIONS-Curiosity may be an important, yet under-recognized contributor to academic achievement. Fostering curiosity may optimize academic achievement at kindergarten, especially for children with low SES.Users may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research, subject always to the full Conditions of use:
# Introduction
Fostering early academic achievement in young children has been a longstanding interest of pediatricians and policymakers, with a growing awareness of the importance of social emotional skills for school readiness [bib_ref] Teachers' judgments of problems in the transition to kindergarten, Rimm-Kaufman [/bib_ref] [bib_ref] Preschool and kindergarten teachers' beliefs about early school competencies: Misalignment matters for..., Abry [/bib_ref]. The socio-emotional characteristics thought to be necessary for early learning include a child's capacity for invention and imagination (i.e., curiosity), persistence and attentiveness to tasks (i.e., effortful control), the ability to form and sustain social relationships (i.e., prosocial behavior), and the capacity to manage feelings and behavior (i.e., emotion regulation) [bib_ref] An ecological perspective on the transition to kindergarten: A theoretical framework to..., Rimm-Kaufman [/bib_ref] [bib_ref] Teachers' judgments of problems in the transition to kindergarten, Rimm-Kaufman [/bib_ref] [bib_ref] Preschool and kindergarten teachers' beliefs about early school competencies: Misalignment matters for..., Abry [/bib_ref]. While each of the aforementioned skills is important for school readiness, curiosity and effortful control are believed to be especially important for fostering early academic achievement. While the association between prosocial skills, emotional regulation and academic achievement is limited (9), we have identified four studies using national datasets which demonstrated an association between curiosity in combination with effortful control (i.e., the construct of "approach to learning") and more optimal early academic achievement in school age children [bib_ref] The relationship between approaches to learning and academic achievement among kindergarten students:..., Lee [/bib_ref] [bib_ref] Children's early approaches to learning and academic trajectories through fifth grade, Li-Grining [/bib_ref] [bib_ref] Kindergarten skills and fifth-grade achievement: Evidence from the ECLS-K, Claessens [/bib_ref]. While curiosity in combination with effortful control appears to be promotive of early academic achievement, this approach does not examine whether curiosity, independent of effortful control is associated with early academic achievement, which is a gap in the literature. Interventions to foster early learning have largely focused exclusively on the cultivation of early effortful control [bib_ref] Preschool program improves cognitive control, Diamond [/bib_ref]. If higher curiosity, independent of effortful control, is associated with higher early academic achievement, this can provide preliminary support for the importance of fostering children's curiosity in the preschool years.
Curiosity is characterized by the joy of discovery [bib_ref] Interest and deprivation factors of epistemic curiosity, Litman [/bib_ref] , and the motivation to seek answers to what is unknown [bib_ref] Curiosity and the pleasures of learning: Wanting and liking new information, Litman [/bib_ref]. Piaget recognized the importance of curiosity as a foundation for early learning, referring to children as "little scientists", and accordingly, pediatric guidelines highlight the importance of promoting curiosity as a foundation for early learning (18). Curiosity is described as an approach-oriented, motivational state associated with exploration. Curiosity is thought to be a multidimensional construct that is both personspecific (i.e., trait) and situation-specific (i.e., activity-related (state)). While curiosity traits are thought to be highly heritable [bib_ref] Genetic and environmental influences on the positive traits of the values in..., Steger [/bib_ref] and include an openness to experiences, desire for novelty, and willingness to embrace the unexpected (21), the expression of curiosity is also thought to be situational (i.e., state), related to an individual's idiosyncratic interests, which can vary with activity and context. While little is known about the factors that can promote the development of trait curiosity, it is theorized that state curiosity is malleable, and can be influenced by social and individual contexts. Curiosity is thought to be enhanced when individuals are allowed to engage in activities that are personally meaningful [bib_ref] The effects of instructors' autonomy support and students' autonomous motivation on learning..., Black [/bib_ref]. As such, it is believed that interventions which promote an experience of meaningfulness of an activity might enhance a child's engagement in that activity, and help foster curiosity. While the association between curiosity and academic achievement has been examined in middle childhood and adolescence [bib_ref] Motives emanating from personality associated with achievement in a Finnish senior high..., Froiland [/bib_ref] [bib_ref] Genetically-Mediated Associations Between Measures of Childhood Character and Academic Achievement, Tucker-Drob [/bib_ref] , to our knowledge, there have been no empirical studies examining the independent association of curiosity in early childhood with early academic achievement, which is a gap in the literature. If higher curiosity is associated with higher academic achievement, this can help inform the development of interventions to cultivate curiosity in young children to foster academic achievement. Therefore, the first objective of this study was to test the hypothesis that curiosity at kindergarten is an independent predictor of kindergarten achievement in reading and math.
In addition, we consider the alternate possibility that curiosity may be associated with more optimal early learning only in certain contexts, and that the association between curiosity and academic achievement is moderated by certain characteristics of the child (e.g., effortful control or sex), or certain characteristics of the environment (e.g., socioeconomic status). We chose to examine effortful control as a potential moderator because part of the task of learning requires children to engage in activities that are sometimes not entirely aligned with their interests. Effortful control requires a child to focus attention, delay gratification and inhibit impulses to engage in the tasks that are demanded of them [bib_ref] Relating effortful control, executive function, and false belief understanding to emerging math..., Blair [/bib_ref]. It is possible that a child's capacity for curiosity and "thirst for learning" will be associated with more optimal academic achievement only if the child is able to manifest focused attention (i.e. effortful control), and that with lower (or absent) effortful control, the association of curiosity with academic achievement is attenuated. We also considered that sex might moderate the association between curiosity and academic achievement. There is some suggestion that teachers perceive curiosity, and inquisitiveness as being a "behavior problem" [bib_ref] Relationship of teachers' ratings of kindergarteners' 21st century skills and student performance, Woods-Groves [/bib_ref]. If "curious boys" are viewed negatively, the expression of curiosity in boys might be discouraged, contributing to lower academic achievement in boys with higher curiosity. Finally, we consider the possibility that differences in socioeconomic status may moderate the association between curiosity and academic achievement. There is some support that curiosity is valued differently in families of higher versus lower SES. We hypothesize that for children in higher SES environments, curiosity may be more valued, encouraged and supported, contributing to higher academic achievement. Therefore, our second objective was to examine potential moderators of the association between curiosity and academic achievement (effortful control, sex, and SES) and to test the hypotheses that the association between curiosity and academic achievement in reading and math is present in the context of higher effortful control, female sex, and higher SES.
# Methods
## Study design and sample
Data were drawn from the Early Childhood Longitudinal Study, Birth Cohort (ECLS-B), a nationally representative, population-based longitudinal study sponsored by the US Department of Education's National Center for Education Statistics (NCES) in the Institute for Education Science. The ECLS-B is based on a nationally representative probability sample of children born in the United States in 2001 (inclusive). Data were collected from over 10,000 children and their parents at age 9 months, with subsequent assessments at 24months, preschool and kindergarten timepoints, with over 77% of the sample (n= 7700) included at the Kindergarten 2006 timepoint [bib_ref] Early Childhood Longitudinal Study, Birth Cohort (ECLS-B) Kindergarten 2006 and 2007 Data..., Snow [/bib_ref]. In the ECLS-B, some children entered kindergarten for the first time in 2006, and some entered in 2007. We defined our sample as first-time kindergarten enrollees, from the 2006 and 2007 timepoints. Data collection consisted of home visits with parent interview and direct and indirect child assessments, and included information on children's cognitive, emotional, behavioral and physical development across multiple settings [bib_ref] Early Childhood Longitudinal Study, Birth Cohort (ECLS-B) Kindergarten 2006 and 2007 Data..., Snow [/bib_ref].
For this study, we assessed children at kindergarten entry, including children who first started kindergarten in 2006 (n = 4850) and children who first started kindergarten in 2007 (n = 1500). Our sample excluded children with congenital and chromosomal abnormalities, included children born at 22-41 weeks gestation, and utilized data from 3 timepoints (birth, preschool and kindergarten). Although there were 6350 children in the combined 2006/2007 kindergarten sample (excluding 100 children with unspecified school placement), our sample was restricted to children who had complete reading, math, and behavioral questionnaire data at kindergarten entry, which reduced our sample size to 6200 children. This study was considered exempt by the Institutional Review Board because the research involved the use of a publicly available dataset, in which the participants were de-identified, and data could not be linked to the participants.
## Measures outcomes
Academic Achievement in Kindergarten Reading and Math: Children were directly assessed at kindergarten age during a home visit by trained National Center for Education Statistics staff using a specialized battery of tests developed for the ECLS-B to assess early reading and math skills. The reading assessment was formulated from existing instruments including the Peabody Picture Vocabulary Test, 3 rd Edition and Preschool Comprehensive Test of Phonological and Print Processing and measured markers of early literacy including basic reading skills (letter and word recognition, understanding letter-sound relationships, phonological awareness, sight word recognition, and understanding words in the context of simple sentences). The reliability of the early reading assessment is described by the item response theory (IRT) reliability coefficient, reported as 0.92 at kindergarten. Scores provide ability estimates in a particular domain and were reported as normally distributed theta scores which demonstrated a range of −2.11 to 3.09 (mean = 0.33, SD = 0.86) at kindergarten [bib_ref] Early Childhood Longitudinal Study, Birth Cohort (ECLS-B) Kindergarten 2006 and 2007 Data..., Snow [/bib_ref]. The ECLS-B math assessment incorporated items to test the following content areas: number sense, geometry, counting numerical operations, and pattern recognition. The item response theory reliability coefficient for the early mathematics assessment was 0.92 at kindergarten. The math theta scores demonstrated a range of −2.42 to 3.12 (mean = 0.38, SD = 0.80) at kindergarten [bib_ref] Early Childhood Longitudinal Study, Birth Cohort (ECLS-B) Kindergarten 2006 and 2007 Data..., Snow [/bib_ref].
## Predictor
Curiosity and Effortful Control: Because the ECLS-B did not contain a measure of curiosity or effortful control, we derived these measures from an existing measure of child behavior available in the dataset. The ECLS-B contains a 25-item questionnaire administered to parents and teachers at preschool and kindergarten timepoints, which was designed to assess child behavior [bib_ref] Early Childhood Longitudinal Study, Birth Cohort (ECLS-B) Kindergarten 2006 and 2007 Data..., Snow [/bib_ref]. The questionnaire was formulated from existing instruments including the Preschool and Kindergarten Behavioral Scales Second Edition (PKBS-2) and Social Skills Rating System (SSRS). Respondents were asked to report the frequency of behaviors observed in the previous 3 months on a 5-point Likert scale (1, never to 5, very often). Items were reverse coded as appropriate such that higher scores indicated more positive behaviors. Five items from the PKBS-2 were chosen from the parent questionnaire at the kindergarten timepoint to generate our measure of curiosity, and two items were chosen to generate our measure of effortful control that aligned with previous behavioral descriptions of curiosity [bib_ref] Curiosity and the interested explorer, Day [/bib_ref] [bib_ref] The Psychology and Neuroscience of Curiosity, Kidd [/bib_ref] , and effortful control. Questions related to curiosity were omitted from the teacher questionnaires and the parent preschool questionnaire. As a result, a curiosity scale could only be generated from the parent at the kindergarten timepoint, and a comparable teacher curiosity scale could not be generated. A confirmatory factor analysis (CFA) was conducted to assure reliability and to calculate the appropriate loading values for deriving curiosity and effortful control factors. Standardized scoring of the curiosity and effortful control factors was conducted using PROC STANDARD, and good internal consistency was demonstrated: curiosity (5 items, α = 0.73), effortful control (2 items, α = 0.67). (Supplemental [fig_ref] Table 1: Maternal and Child Characteristics for Weighted Sample Socioeconomic indicators calculated from measures... [/fig_ref].
Covariates-Maternal and child characteristics associated with academic achievement in reading and math [bib_ref] The impact of low birth weight, perinatal conditions, and sociodemographic factors on..., Resnick [/bib_ref] [bib_ref] Cognitive ability at kindergarten entry and socioeconomic status, Larson [/bib_ref] were chosen a priori as covariates after a review of the literature.
The following covariates were ascertained from the restricted ECLS-B birth certificate data: maternal age, race/ethnicity, marital status (married/unmarried), birthweight, and gestational age. Also included were measures of maternal education (< high school; high school graduate; > high school) and poverty (< 185% federal poverty limit; ≥ 185% federal poverty line) which were incorporated into a single composite measure of household socioeconomic status (SES) created by ECLS-B at the kindergarten 2006 timepoint [bib_ref] Early Childhood Longitudinal Study, Birth Cohort (ECLS-B) Kindergarten 2006 and 2007 Data..., Snow [/bib_ref]. Because early educational experiences and child sex have been associated with kindergarten outcomeswe included enrollment in any preschool program the year prior to kindergarten entry and child sex as covariates. We also controlled for child age at kindergarten entry, and months of kindergarten experience. Because the number of months of kindergarten experience was not normally distributed, we accounted for the variability in kindergarten experience by creating a 3-category variable indicating the trimester of kindergarten at time of kindergarten assessment (1 st trimester Kindergarten: August-October; 2 nd trimester Kindergarten: November-January; 3 rd trimester Kindergarten: February-June).
## Statistical analyses
Maternal and child characteristics were examined using descriptive statistics. Multiple linear regression (using the SURVEYREG procedure in SAS) was used to examine the associations of curiosity and effortful control generated from the CFA, with academic achievement in reading and math at kindergarten entry. Two models (one model for reading and one model for math) included both curiosity and effortful control simultaneously while controlling for all covariates.
We were also interested in examining whether the association between curiosity and academic achievement was moderated by either effortful control, sex, or socioeconomic status (SES). We ran 6 additional models (one model for reading and one model for math) to test for moderation which included the interaction terms of curiosity with effortful control (models 1-2); curiosity with sex (models 3-4); and curiosity with SES (models 5-6), controlling for a priori covariates. When the p-value for the interaction was significant (p < . 05), we performed a stratified analysis of the association between curiosity and academic achievement, adjusting for covariates.
All analyses were conducted using SAS 9.4 (40) (SAS Institute Inc., Cary, NC). Because of the complex sample design, sample weights and the Jackknife method were utilized to account for stratification, clustering and unit non-response, thereby allowing the weighted results to be generalized to the population of U.S. children born in 2001. In accord with the NCES requirements for ECLS-B data usage, reported numbers were rounded to the nearest 50.
# Results
## Sample characteristics
After applying sample weights, the maternal and child characteristics were generalizable to the US population in 2001, and the sample characteristics for the weighted sample are shown in [fig_ref] Table 1: Maternal and Child Characteristics for Weighted Sample Socioeconomic indicators calculated from measures... [/fig_ref].
## Confirmatory factor analyses
The goodness of fit indices for our CFA demonstrated appropriate fit as evidenced by the Standardized RMR (SRMR) = 0.047, and adjusted GFI (AGFI) = 0.91, with good fit indicated by SRMR < 0.08, and AGFI > 0.90. Loading coefficients for each item of the effortful control and curiosity factors were calculated, with loading coefficients >0.indicating that each item contributed to the overall factor. The loading coefficients for each item of the "effortful control" factor was 0.71, and loading coefficients for each item in the 5-item curiosity factor ranged from 0.51 -0.66. The items "likes to try new things," "shows eagerness to learn new things" and "shows imagination in work and play" loaded the most strongly on the overall curiosity factor. (Supplemental [fig_ref] Table 1: Maternal and Child Characteristics for Weighted Sample Socioeconomic indicators calculated from measures... [/fig_ref].
## Association between child curiosity and effortful control and reading and math academic achievement at kindergarten
After controlling for potential confounders, higher curiosity and higher effortful control were each associated with higher reading (b reading = 0.11, p < .001 (curiosity); b reading = 0.11, p < .001 (effortful control)), and higher math academic achievement at kindergarten (b math = 0.12, p < .001 (curiosity); b math = 0.14, p < .001 (effortful control)) .
## Moderators of the association between child curiosity and effortful control and reading and math academic achievement at kindergarten
The associations between curiosity and academic achievement in reading and math were not moderated by effortful control or sex, as evidenced by non-significant interaction terms (curiosity × effortful control: p = .65 (reading academic achievement); p = .23 (math academic achievement); curiosity × sex: p = .94 (reading academic achievement); p = .75 (math academic achievement)). We did find evidence of moderation by SES, in both reading and math models: curiosity × SES: p = .01 (reading academic achievement); p = .005 (math academic achievement). Because we found evidence of moderation by SES, we then performed a covaried analysis of the association between curiosity and academic achievement, for both reading and math, stratified by low and high SES. We found differences in the standardized parameter estimates (b) for the association between curiosity and academic achievement for children from low SES environments, compared to high SES environments for reading academic achievement: (SES ≤ median): b = 0.18, p <.001; (SES > median): b = 0.08, p =.004. The same association was found for math academic achievement: (SES ≤ median): b = 0.20, p <.001; (SES > median): b = 0.07, p < .001 [fig_ref] Figure 1: Moderation [/fig_ref].
## Characteristics of "curiosity factor" and associations with reading and math academic achievement
We also performed a post hoc analysis to better understand if there were specific features of the curiosity factor driving the association with reading and math academic achievement at kindergarten that may have particular clinical relevance. The descriptive statistics and standardized parameter estimates (b) for each question comprising the "curiosity factor" are shown in . We ran two models, one for academic achievement in reading and one for academic achievement in math. Each model included effortful control, and replaced the curiosity factor in in the final model with all 5 question items in the curiosity factor simultaneously, adjusting for the a priori covariates. In these models, "shows eagerness to learn new things," (b reading = 0.09, p <.001; b math = 0.10, p <.001) "appropriately uses a variety of words to describe feelings," (b reading = 0.08, p <.001; b math = 0.07, p <.001) and "easily adjusts to a new situation," (b reading = 0.04, p = .01; b rmath = 0.07, p <.001) were each significantly associated with reading and math academic achievement. The questions "likes to try new things," and "shows imagination in work and play," were not associated with academic achievement in reading or math in these models .
# Discussion
This is the first study examining the independent association of curiosity with kindergarten reading and math academic achievement using a nationally representative sample. After controlling for effortful control and other potential confounders, we found that curiosity was significantly associated with academic achievement in kindergarten, with an effect size similar to that for effortful control. These findings suggest that curiosity is as important as effortful control for promoting reading and math academic achievement at kindergarten-age. In the absence of controlling for effortful control, previous research has found effect sizes of curiosity on school-age reading and math academic achievement to be higher, with standardized regression coefficients ranging from b = 0.16-0.23 [bib_ref] The Relationship Between Classroom Motivation and Academic Achievement in Elementary-School-Aged Children, Broussard [/bib_ref]. Although our effect sizes of the association between curiosity and academic achievement are small, and although the strongest predictors for reading and math academic achievement continue to be prior reading and math academic achievement scores (9), our results suggest that curiosity makes a small but meaningful contribution to academic achievement. At an individual level, the effect size of curiosity on academic achievement is small, however, when considered at a population level, the magnitude of effect is notable.
We found that the associations of curiosity with reading and math academic achievement at kindergarten were not moderated by effortful control or sex, but were moderated by SES. These findings suggest that even if a child manifests low effortful control, higher curiosity may be associated with more optimal academic achievement. Currently, most classroom interventions have focused on the cultivation of early effortful control and self-regulatory capacities [bib_ref] The Promotion of Self-Regulation as a Means of Enhancing School Readiness and..., Ursache [/bib_ref]. Our results suggest that an alternate message, focused on the importance of curiosity, may also be considered. Encouragingly, child sex did not moderate the association between curiosity and academic achievement, suggesting that curiosity is equally important in boys and girls for more optimal academic achievement. While we found evidence that the association of curiosity with academic achievement was moderated by SES, contrary to expectations, we found a differential benefit of curiosity for children with low SES. Our results suggest that while higher curiosity is associated with higher academic achievement in all children, the association of curiosity with academic achievement is greater in children with lower SES. Children with higher SES likely have environments with greater access to resources to foster reading and math academic achievement, whereas children with low SES likely have academic environments that are less enriching, and thus, the drive for academic achievement is related to the child's motivation to learn (i.e. curiosity). Because there is some evidence suggesting that children with low curiosity fail engage with their environments in ways that will foster their academic development [bib_ref] Self-Concepts of High-and Low-Curiosity Boys, Maw [/bib_ref] , and because the association of curiosity and academic achievement appears to have a greater magnitude of association in children from lower SES environments, our results suggest that the promotion of curiosity may be a valuable intervention target to foster early academic achievement (30), with particular advantage for children in poverty.
Because curiosity is thought to relate to the construct of intrinsic motivation, there is some theoretical support that promoting autonomy, feelings of competence and connectedness can foster intrinsic motivation, and increase curiosity. Interventions to foster curiosity in adults have focused on highlighting the personal meaningfulness of an activity to optimize engagement [bib_ref] Motivational interviewing and self-determination theory, Markland [/bib_ref]. These features may also be helpful in designing interventions to promote curiosity in young children, and is an important area for future research.
Pediatricians, early childhood educators and policy makers have grappled with the question, "what are the early social emotional skills children need to be successful?" There has been an increasing interest in the taxonomy of "the Big Five character skills" (i.e. O.C.E.A.N.: Openness to Experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism) as a foundation for future life success [bib_ref] Little Five": Exploring the Nomological Network of the Five-Factor Model of Personality..., John [/bib_ref] [bib_ref] Fostering and measuring skills: Interventions that improve character and cognition, Heckman [/bib_ref] [bib_ref] Fostering and measuring skills: Improving cognitive and non-cognitive skills to promote lifetime..., Kautz [/bib_ref]. One dimension of the Big Five, "openness to experience," aligns with the item of "shows eagerness to learn new things," in our curiosity factor, and was associated with higher academic achievement in our adjusted models. When all items were considered together, the aspect of curiosity most strongly associated with higher academic achievement was the construct "shows eagerness to learn new things." This construct is consistent with some of the earliest descriptions of curiosity as a "passion for learning", and speaks to the positive motivational drive for knowledge which underlies curiosity [bib_ref] The psychology of curiosity: A review and reinterpretation, Loewenstein [/bib_ref]. A child's eagerness to learn may be an important characteristic to nurture in the young child, particularly to promote academic achievement. This framework, consistent with recent neuroscience research, suggests that one pathway to more optimal learning may be through captivating a child's natural curiosity [bib_ref] Scientific Thinking in Young Children: Theoretical Advances, Empirical Research, and Policy Implications, Gopnik [/bib_ref].
Our study had several strengths and limitations. The study includes a nationally representative sample, the results of which are generalizable to the population, and direct child assessments of early academic achievement. One of the limitations of the study is that the ECLS-B did not contain data on maternal intelligence quotient, or family history of learning difficulties, which can be associated with early academic achievement. In addition, our construct of curiosity was derived from a parent-report behavioral measure at the kindergarten timepoint. Because the teacher-report of child behavior included different questions, a teacher curiosity factor could not be calculated, and we were not able to examine the construct of curiosity across reporters. Finally, parents generally rated their children very highly in curiosity. Question items about curiosity that generate broader variability in parental response will be an important focus for future work.
Our study provides some preliminary evidence that higher early childhood curiosity, independent of early effortful control, is associated with more optimal early academic achievement, with a greater magnitude of association for children with low SES. To foster early learning, it may be helpful to identify opportunities to cultivate and encourage curiosity in young children, especially for children from environments of economic disadvantage.
# Conclusion
We found evidence that curiosity, independent of effortful control was associated with greater reading and math academic achievement at kindergarten, with a greater magnitude of association for children with low SES. These findings suggest that although effortful control has been emphasized as an important prerequisite for early academic achievement, curiosity is also important, and may be especially important for children from environments of economic disadvantage. In addition, the aspect of curiosity most strongly associated with higher academic achievement was the construct of "shows eagerness to learn new things." Encouraging curiosity in young children and cultivating their eagerness to learn may be a potential intervention target to foster early reading and math academic achievement at kindergarten-age, and may be particularly advantageous for children with low SES.
# Supplementary material
Refer to Web version on PubMed Central for supplementary material. Descriptive Statistics and Adjusted Associations between each Curiosity Question and Academic Achievement in Reading and Math at Kindergarten
[fig] Figure 1: Moderation [/fig]
[table] Table 1: Maternal and Child Characteristics for Weighted Sample Socioeconomic indicators calculated from measures of education and income at Kindergarten: Below poverty threshold (<185% federal poverty line) 44.7% At or above poverty threshold (≥185% federal poverty line) 55.3% Child Characteristics Mean, SD, or Weighted (%) Enrollment in any Preschool Program in Year Prior to Kindergarten [/table]
|
Lipidomics reveals accumulation of the oxidized cholesterol in erythrocytes of heart failure patients
A B S T R A C TLipids play an important role in the pathogenesis of cardiovascular disease. Changes in lipids of erythrocytes are indicative of the outcome of pathophysiological processes. In the present study, we assessed whether the lipid profiles of erythrocytes from heart failure (HF) patients are informative of their disease risk. The lipidomes of erythrocytes from 10 control subjects and 29 patients at different HF stages were analyzed using liquid chromatography time-of-flight mass spectrometry. The lipid composition of erythrocytes obtained from HF patients was significantly different from that of normal controls. The levels of phosphatidylcholines (PCs), phosphatidylethanolamines (PEs), and sphingomyelins decreased in HF erythrocytes as compared with those of control subjects; however, the levels of lysoPCs, lysoPEs, and ceramides increased in HF erythrocytes. Notably, the oxidized cholesterol 7-ketocholesterol (7KCh) accumulated to higher level in HF erythrocytes than in plasma from the same patients. We further validated our findings with a cohort of 115 subjects of control subjects (n=28) and patients (n=87). Mechanistically, 7KCh promoted reactive oxygen species (ROS) formation in cardiomyocytes; and induced their death, probably through an ATF4-dependent pathway. Our findings suggest that erythrocytic 7KCh can be a risk factor for HF, and is probably implicated in its pathophysiology.
# Introduction
Cardiovascular disease is a major health problem and the leading cause of death globally. Cardiac function deterioration hampers the ability of the heart to support blood circulation, resulting in heart failure (HF). The pathogenic mechanism leading to this end stage is complicated. Myocardial infarction, hypertension, cardiomyopathy, valvular heart disease, and inflammation-induced oxidative stress are known risk factors for disease progression [bib_ref] Novel metabolic risk factors for incident heart failure and their relationship with..., Bahrami [/bib_ref] [bib_ref] Novel metabolic risk factors for heart failure, Ingelsson [/bib_ref].
Several biochemical pathways, including the pentose phosphate pathway, anaplerotic metabolism, ketone body metabolism, lipotoxic intermediate metabolism, and glycolysis, are affected in patients with HF [bib_ref] Using metabolomics to assess myocardial metabolism and energetics in heart failure, Turer [/bib_ref] [bib_ref] Energetics of the failing heart: new insights using genetic modification in the..., Ingwall [/bib_ref] [bib_ref] Hypertrophic cardiomyopathy: a paradigm for myocardial energy depletion, Ashrafian [/bib_ref]. Changes in metabolites have been identified in plasma and are associated with clinical outcomes in patients with HF [bib_ref] Increased p-cresyl sulfate level is independently associated with poor outcomes in patients..., Wang [/bib_ref] [bib_ref] Metabolic disturbances identified in plasma are associated with outcomes in patients with..., Cheng [/bib_ref]. These findings suggest that metabolic remodeling in patients may occur during HF progression, and the metabolite profile can thus be used as a biomarker panel for a variety of assessment purposes.
Lipid metabolism alterations have been increasingly demonstrated to underlie the pathogenesis of cardiovascular disease [bib_ref] Lipidomics profiling and risk of cardiovascular disease in the prospective population-based Bruneck..., Stegemann [/bib_ref]. Currently, research on lipids has focused on the analysis of plasma lipids such as cholesterol, triacylglyceride, and phospholipids [bib_ref] Lipidomics profiling and risk of cardiovascular disease in the prospective population-based Bruneck..., Stegemann [/bib_ref] [bib_ref] Discrimination and net reclassification of cardiovascular risk with lipoprotein(a): prospective 15-year outcomes..., Willeit [/bib_ref]. Reports seldom indicate specific fatty acids and total cholesterol in erythrocytes as a predictor of cardiovascular events [bib_ref] Red blood cell fatty acid patterns and acute coronary syndrome, Shearer [/bib_ref] [bib_ref] Total cholesterol content of erythrocyte membranes is increased in patients with acute..., Tziakas [/bib_ref]. Additionally, erythrocytes are involved in reverse cholesterol transport (RCT), particularly in the low high-density lipoprotein (HDL) state [bib_ref] Red blood cells play a role in reverse cholesterol transport, Hung [/bib_ref]. Given the relatively long life (approximately 120 days) of erythrocytes, any change in the lipid profile of erythrocyte membrane may reflect pathophysiologic changes associated with disease progression.
Few studies have reported on the comprehensive assessment of the metabolome and lipidome of RBCs [bib_ref] Alterations of red blood cell metabolome during cold liquid storage of erythrocyte..., Gevi [/bib_ref] [bib_ref] Alterations of red blood cell metabolome in overhydrated hereditary stomatocytosis, Darghouth [/bib_ref] [bib_ref] Pathophysiology of sickle cell disease is mirrored by the red blood cell..., Darghouth [/bib_ref] , especially in the scenario of HF. The aim of this study was to identify lipid profiles of HF erythrocytes using high-throughput liquid chromatography time-of-flight mass spectrometry (LC-TOF-MS). Our findings suggested that the erythrocyte lipid profiles of patients with HF were significantly different from those of normal subjects. The levels of phosphatidylcholines (PCs), phosphatidylethanolamines (PEs), and sphingomyelins (SMs) decreased in HF erythrocytes. However, the levels of lysoPCs, lysoPEs, and ceramides increased in these cells. Of these lipids, 7-ketocholesterol (7KCh) accumulated in the erythrocytes of patients with HF. This accumulation may be of significance as a potential discriminator and as a player in the pathogenesis of HF. At molecular level, we demonstrated that intracellular 7KCh accumulation caused reactive oxygen species (ROS) formation and cardiomyocyte death, which was probably mediated through ATP4/CHOP pathway.
# Materials and methods
## Blood sample collection and erythrocyte preparation
Patients with HF were classified as stages A, B, and C according to the American College of Cardiology and the American Heart Association Heart Failure (ACC/AHA HF) classification system [bib_ref] American Heart Association Task Force on Practice, ACC/AHA 2005 guideline update for..., Hunt [/bib_ref] : patients in stage A were at high risk and asymptomatic but did not have structural heart disease; patients in stage B had structural heart disease but were asymptomatic; and patients in stage C had been hospitalized due to acute or decompensated chronic HF. Patients aged 39-85 years were enrolled in this study. Patients with hypothyroidism, decompensated liver cirrhosis, systemic lupus erythematosus, or serum creatinine of > 3 mg/dL were excluded. Informed consent was obtained from all patients. The study was designed and performed in accordance with the principles of the Declaration of Helsinki and with approval from the Ethics Review Board of Chang Gung Memorial Hospital.
Blood samples were collected in EDTA-containing tubes before the patients were discharged from the hospital. RBCs were prepared for LC-MS analysis. The plasma level of B-type natriuretic peptide (BNP) was measured in triplicate with the Triage BNP Test (Biosite, San Diego, CA). The precision, analytical sensitivity and stability characteristics of the assay were previously described [bib_ref] Impact of age, race, and sex on the ability of B-type natriuretic..., Maisel [/bib_ref]. Measurements of kidney function and other biochemical parameters were conducted in the Central Core Laboratory of Chang Gung Memorial Hospital at Keelung.
## Lipidomics analysis by lc-ms
For human erythrocyte lipids, a modified method was employed [bib_ref] Improved procedure for the extraction of lipids from human erythrocytes, Rose [/bib_ref]. In brief, 100 μL of erythrocyte lysates was transferred to a glass tube, and 100 μL of water was added. The sample was vortexed and then placed on ice for 15 min. After 1.1 mL of isopropanol was added, the sample was vortexed 4 times for 30 s. Approximately 700 μL of chloroform was added, and the sample was vortexed 4 times for 30 s again. It was subsequently centrifuged at 700×g for 30 min at 4°C. The supernatant was transferred to a new tube, dried under nitrogen gas, and stored at −80°C. Prior to analysis, the sample was dissolved in 500 μL of isopropanol/acetonitrile/water (2:1:1).
For lipid separation, an ACQUITY CSH C18 (2.1 mm × 100 mm, 1.7 µm) column was used. The column temperature was set at 55°C and the flow rate at 400 μL/min. Mobile phase buffer A was acetonitrile/ water (60/40) with 10 mM ammonium formate and 0.1% formic acid, whereas buffer B was isopropanol/acetonitrile (90:10) with 10 mM ammonium formate and 0.1% formic acid. The initial LC gradient conditions were 40% buffer B, increasing to 70% B within 12.1 min, and then to 99% B for 6 min before re-equilibration for 2 min at 40% B. The lyophilized sample was diluted with 500 μL of isopropanol/acetonitrile/water (2:1:1). Each sample was analyzed in triplicate. MS was performed on a Waters QTOFMS (SYNAPT G1 HDMS, Waters MS Technologies, Manchester, UK) operating in the positive or negative ion mode. The desolvation gas was set at 800 L/h at a temperature of 400°C; the cone gas was set at 25 L/h; and the source temperature was set at 100°C. The capillary voltage and cone voltage were set to 3000 and 35 V, respectively. Leucine encephalin was used as the lock mass standard (an [M+H] + ion at 556.2771 Da in electrospray ionization (ESI) positive mode; an [M+H]ion at 554.2615 Da in ESI negative mode).
MS data were processed using MassLynx V4.1 and Progenesis QI software (Waters Corp., Milford, Massachusetts, USA). The intensity of each mass ion was normalized with respect to the total ion count to generate a data matrix that included the retention time, m/z value, and normalized peak area. The multivariate data matrix was analyzed using SIMCA-P software (version 13.0, Umetrics AB, Umea, Sweden).
Significant metabolites were subjected to a database search using an in-house database, the Human Metabolome Database (HMDB) (http:// www.hmdb.ca/), or the METLIN Metabolomics Database (metlin.scripps.edu/index.php). For the identification of specific metabolites, MS/MS spectra were collected and confirmed by comparison with the spectra of chemical standards or by searching the HMDB and METLIN databases.
## Quantification of erythrocyte and plasma 7kch by lc-ms-ms
Erythrocyte 7KCh was extracted using a modified method [bib_ref] Improved procedure for the extraction of lipids from human erythrocytes, Rose [/bib_ref] as described above. A sample was dissolved in isopropanol: acetonitrile:water (2:1:1 (v/v/v)) containing the deuterated forms of 7KCh-D7 and cholesterol-D6, which were employed as internal standards for quantification. For analysis of total-form sterols and free-form sterols in plasma, plasma samples (10 μL) were preincubated with or without cholesteryl ester hydrolase at 37°C for 15 min in 10 mM phosphate buffer containing taurocholic acid [bib_ref] Determination of total cholesterol in serum by cholesterol esterase and cholesterol oxidase..., Malik [/bib_ref]. The samples were extracted with 500 μL of methanol containing cholesterol-D6 and 7-KCh-D7, and then centrifuged at 12,000×g at 4°C for 30 min. The resulting supernatant was collected in a new microtube and dried with nitrogen. The sample was dissolved in 100 μL of 100% methanol and was centrifuged at 12,000×g at 4°C for 30 min for 7KCh, cholesterol, lanosterol, lathosterol, and 7-dehydrocholesterol quantification.
All the samples were analyzed using ultra-high-performance LC coupled with Xevo TQ-S MS (Waters Corp.) as previously described with modifications [bib_ref] A comprehensive method for extraction and quantitative analysis of sterols and secosteroids..., Mcdonald [/bib_ref]. MS with atmospheric pressure chemical ionization (APCI) was performed in positive-ion multiple-reaction-monitoring mode. The optimized parameters were as follows: corona discharge current at 1 μA; probe temperature at 600°C; source temperature at 150°C; and gas flow at 800 L/h. Chromatographic separation was achieved on a pentafluorophenyl (PFP) column (100 mm×2.1 mm; with particle size of 1.8 µm; Waters Corp.) at 25°C with eluent A (acetonitrile: water 25:75 (v/v) with 0.1% formic acid) and eluent B (methanol); the flow rate was set at 0.3 mL/min. The gradient profile was as follows: isogradient 70% B, 1 min; linear gradient 70-75.7% B, 8 min; linear gradient 75.7-100% B, 1 min; and 100% B, 2 min. The column was then re-equilibrated for 8 min for the next analysis.
## Cell culture and viability assay
HL-1 atrial myocytes were cultured in fibronectin-gelatin-coated flasks containing Claycomb medium, supplemented with 10% fetal bovine serum, 100 U/mL of penicillin, 100 μg/mL of streptomycin, 2 mM L-glutamine, and 0.1 mM norepinephrine in a humidified atmosphere containing 5% CO 2 at 37°C, as previously described [bib_ref] HL-1 cells: a cardiac muscle cell line that contracts and retains phenotypic..., Claycomb [/bib_ref]. For 7KCh treatment, 2 × 10 4 cells were incubated with 10, 20, and 50 µM 7KCh for the indicated periods. To mimic the delivery of 7KCh from erythrocyte to cardiomyocytes, we prepared erythrocyte ghost and loaded it with 7KCh. The packed erythrocytes were hemolyzed in hypotonic phosphate buffer (2.5 mM NaH 2 PO 4 , 2.5 mM Na 2 HPO 4 , 1 mM EDTA, pH 8) and centrifuged at 16,000 g to precipitate the ghosts. The erythrocyte membrane was incubated with 7KCh or cholesterol at 37°C for 4 h, after which the membrane-bound 7KCh or cholesterol in the ghost pellet was diluted to the indicated concentrations for a viability test. Cell viability was determined by a neutral red assay, as previously described [bib_ref] Biochemical disorders associated with antiproliferative effect of dehydroepiandrosterone in hepatoma cells as..., Cheng [/bib_ref].
## Cytometric analysis of ros
ROS formation was analyzed quantitatively by cytometric analysis. In brief, HL-1 cells were loaded with 5 μM MitoSOX red or 5 μM H 2 DCFDA for 30 min at 37°C, washed twice with PBS, and trypsinized for flow cytometric analysis as previously described [bib_ref] Enterovirus 71 induces mitochondrial reactive oxygen species generation that is required for..., Cheng [/bib_ref]. The mean fluorescence intensity (MFI) of the fluorescence of oxidized MitoSOX red or of DCF fluorescence was quantified using CellQuest Pro software (Becton Dickinson, CA, USA).
## Immunostaining for7kch
HL-1 cells were fixed and permeabilized with 4% paraformaldehyde, 0.1% Triton-X100 in PBS for 2 h. After PBS wash, cells were treated with PBS containing 0.1% Triton X-100 and 5% BSA for 1 h at room temperature. Cells were stained with primary antibody of 7KCh (1:100; Clone #35 A, Japan Institute for the Control of Aging, Shizuoka, Japan) overnight at 4°C. Secondary antibody, conjugated to Alexa-488 (Thermo Fisher Scientific Inc., Waltham, MA, USA), was added at a 1:200 dilution and incubated for 1-2 h. Hoechst 33342 (Thermo Fisher Scientific Inc., Waltham, MA, USA was used to stain cellular nuclei.
## Western blotting assay
SDS-PAGE and western blotting were performed as previously described [bib_ref] Glucose-6-phosphate dehydrogenasedeficient cells show an increased propensity for oxidant-induced senescence, Free Radic, Cheng [/bib_ref]. The cells were rinsed with cold PBS, scraped and collected by centrifugation. Cells were immediately lysed in lysis buffer (20 mM Tris. HCl (pH 8), 1% Triton X-100, 137 mM NaCl, 1.5 mM MgCl 2 , 10% glycerol, 1 mM EGTA, 50 mM NaF, 1 mM Na 3 VO 4 , 10 mM β-glycerophosphate, 1 mM PMSF, 1 μg/mL leupeptin, 1 μg/mL aprotinin). Protein concentration of the lysate was determined by the Bradford method. The sample was analyzed by SDS-PAGE and immunoblotting with antibodies to transcription factor ATF4, CHOP, LC3A/B, cleaved caspase 3, caspase 3 and actin according to manufacturers' instructions.
## Statistical analyses
To maximize the number of differentially abundant metabolites among groups, the orthogonal projection to latent structures discriminant analysis (OPLS-DA) model was applied using SIMCA-P software (version 13.0, Umetrics AB, Umea, Sweden). The variable importance in the projection (VIP) value of each variable in the model was calculated to indicate its contribution. S-plots were constructed from OPLS-DA data. Metabolites were chosen based on their contribution to the variation and correlation within the data set of the control and the stage C HF groups. The VIP values of variables greater than 1.0 are considered significantly different.
Results are expressed as the mean ± SD for continuous variables and as the number (percentage) for categorical variables. Data were compared by two-sample t-tests. A p value of < 0.05 was considered significant.
# Results
## Baseline characteristics
Initially, a total of 115 subjects were enrolled in this study, comprising 28 normal controls and 87 patients at stages A (n = 29), B (n = 29), and C (n = 29). The baseline characteristics and laboratory data are shown in [fig_ref] Table 1: Baseline characteristics of the study populations [/fig_ref]. Notable differences in these parameters were observed between patients at stages A, B, and C. Compared with the normal controls, patients at stage C had remarkably higher glucose, creatinine, and BNP levels, but they had lower total cholesterol, lowdensity lipoprotein (LDL) cholesterol, HDL cholesterol, sodium, hemoglobin and albumin levels, and lower estimated glomerular filtration rate. The percentage of male subjects was higher in the patient groups.
For the untargeted lipid profile study, sex-and age-matched samples from 10 controls and 29 patients with HF at stage A (n = 10), stage B (n = 9), and stage C (n = 10) were chosen [fig_ref] Table 1: Baseline characteristics of the study populations [/fig_ref].
## Changes in lipidomes of erythrocytes in patients with heart failure
To investigate whether the erythrocyte lipid composition differs between patients with HF and normal controls, 39 RBC samples were subjected to untargeted metabolomic analysis using LC-TOF-MS. The typical spectra of erythrocyte extracts were obtained in positive and negative ion modes. After data processing with Progenesis QI, 457 and 438 molecular features obtained in ESI positive and negative modes were extracted, respectively. The data were subjected to SIMCA-P analysis. The derived OPLS-DA score plot showed remarkable separation between the control, stage A, stage B, and stage C groups for data sets obtained in both ESI positive and negative modes . To identify lipophilic metabolites discriminating between the samples from the normal controls and those from patients in group C, we reanalyzed these data sets. The OPLS-DA plots , F) and Splots for the data sets obtained in both ESI positive and negative modes are shown herein. From the S-plots, features with a p (corr) value of > 0.5 or < −0.5 were selected. We observed that 30 and 23 metabolites acquired in both ESI positive and negative modes had VIP scores of > 1.0 and revealed significant differences (p < 0.01) between the patients in stage C and the healthy controls . Clearly, sterols, phospholipids, and ceramides are important discriminators of the normal controls and patients in stage C . Liquid chromatography time-of-flight mass spectrometry-based lipidomics analysis of heart failure (HF) erythrocytes. Erythrocytes from normal control subjects and patients with stages A, B, and C were isolated for time-of-flight mass spectrometry analysis in electrospray ionization (ESI) positive and negative modes. Basal peak chromatograms of patients with HF at different stages and normal control subjects were obtained in ESI positive (A) and negative (B) modes, respectively. The molecular features were identified in samples (n = 39) by using Progenesis QI software and further data processing and statistical analysis were performed through SIMCA-P. Orthogonal partial least squares discriminant analysis (OPLS-DA) was performed for all samples, and the score plots for data sets obtained in ESI positive (C) and negative modes (E) are shown. The data for normal control and patients with stage C were reanalyzed using OPLS-DA, and the score plots for data sets obtained in ESI positive (D) and negative modes (F) are shown. Metabolites with significant differences in abundance in ESI positive and negative modes between normal control and patients with stage C are presented in S-plots (G and I), respectively. The Venn diagrams of the features obtained in ESI positive (VIP > 1.0 and p < 0.01) and ESI negative (VIP > 1.0 and p < 0.01) modes are shown in panels H and J, respectively. The normal control (n = 10), stage A (n = 10), stage B (n = 9), and stage C (n = 10) groups are marked in green, yellow, orange, and red, respectively. The ellipse shown in the model represents the Hotelling T2 with 95% confidence. [fig_ref] Table 2: The potential biomarkers identified for discriminating patients with HF from normal controls [/fig_ref]. The levels of lysophospholipids (such as lysoPC and lysoPE), ceramides, and oxysterols (such as 7KCh) were higher in the HF erythrocytes than in the control erythrocytes. By contrast, the levels of phospholipids (such as PCs, PEs, and sphingomyelins) decreased in the HF erythrocytes, compared with those of the normal cells [fig_ref] Table 2: The potential biomarkers identified for discriminating patients with HF from normal controls [/fig_ref].
## Erythrocyte and plasma 7kch levels in patient with heart failure
The distribution of 7KCh between plasma and erythrocytes was reported for healthy volunteers [bib_ref] Quantification of oxysterols in human plasma and red blood cells by liquid..., Pataj [/bib_ref]. Of the samples from the normal controls and patients in the current study, the 7KCh level was substantially lower than the levels of lanosterol, lathosterol, and 7-dehydrocholesterol (intermediates in cholesterol biosynthetic pathway) in plasma. By contrast, the 7KCh level was significantly higher than the levels of these metabolites in erythrocytes .
We observed significant increases in the levels of erythrocyte 7KCh in patients at stages A-C, suggesting that erythrocyte 7KCh can be used as a biomarker for the early identification of subjects at risk of HF. To test such a possibility, we determined the levels of cholesterol, lanosterol, lathosterol, 7-dehydrocholesterol, and 7KCh in erythrocytes and plasma from patients with HF and normal control subjects. The levels of lanosterol, lathosterol, 7-dehydrocholesterol, and 7KCh were normalized to that of cholesterol. The level of 7KCh in the erythrocytes was higher in patients in stages A, B, and C than in the controls. The levels of 7KCh in the control subjects and in patients in stages A, B, and C were 5.75 ± 6.88, 120.79 ± 70.93, 116.64 ± 69.69, and 62.03 ± 63.11 µmole per mmole of cholesterol, respectively . Cholesterol exists as free and ester forms in plasma, and the ester form is the major storage and transport forms carried by lipoproteins. We observed no difference in the plasma level of 7KCh between the normal controls and various patient cohorts . When total cholesterol (free cholesterol plus cholesteryl ester) was considered, the plasma level of 7KCh was slightly higher in stage C patients than in the normal controls .
## Intracellular 7kch accumulation enhances ros production and reduces the viability of cardiomyocytes
To examine whether 7KCh impairs the physiology of cardiomyocytes, we tested the viability of HL-1 cells receiving 7KCh treatment. As the level of 7KCh in healthy volunteers was 1-2 μM [bib_ref] Quantification of oxysterols in human plasma and red blood cells by liquid..., Pataj [/bib_ref] and the blood levels of 7KCh in patients with HF were 10-to 30-fold higher than those of the normal controls , 7KCh was used at concentrations ranging from 10 μM to 50 μM so as to mimic the pathophysiological condition. 7KCh caused a dose-dependent reduction in the viability of HL-1 cells. As a control, cholesterol treatment did not affect the viability of the cells [fig_ref] Figure 4: Viability reduction and ROS generation in 7KCh accrued HL-1 cells [/fig_ref]. A plausible possibility is that 7KCh in erythrocytes can be released to affect the viability of cardiomyocytes. To test this possibility, we loaded the erythrocyte membrane with 7KCh and incubated cardiomyocytes with the 7KCh-laden membrane. Upon treatment, the viability of the HL-1 cells declined in a dose-dependent manner [fig_ref] Figure 4: Viability reduction and ROS generation in 7KCh accrued HL-1 cells [/fig_ref]. These results suggest that 7KCh can be delivered from erythrocytes to cause cellular damage.
To ensure whether 7KCh accumulates in HL-1 cells, we performed an immunofluorescence assay with anti-7KCh antibody to detect the intracellular level of 7KCh in HL-1 cell [fig_ref] Figure 4: Viability reduction and ROS generation in 7KCh accrued HL-1 cells [/fig_ref]. The 7KCh accumulated in cytosol of HL-1 cell upon 24 h treatment with 20 μM 7KCh.
To determine whether 7KCh causes ROS generation in HL-1 cells, we treated cells with 7KCh for 24 h, stained with H 2 DCFDA, and analyzed them using flow cytometer. As expected, the DCF fluorescence increased in cells treated with 10 μM and 20 μM 7KCh [fig_ref] Figure 4: Viability reduction and ROS generation in 7KCh accrued HL-1 cells [/fig_ref]. H 2 O 2treated HL-1 cells served as positive control. It was consistent with the results for cells stained with MitoSOX Red, a mitochondrion-specific probe for ROS. The 7KCh-treated HL-1 cells showed higher ROS production than positive control (i.e. HL-1 cells treated with 5 mM H 2 O 2 ) [fig_ref] Figure 4: Viability reduction and ROS generation in 7KCh accrued HL-1 cells [/fig_ref].
## 7kch induces cell death through activation of transcription factor 4 (atf4) pathway
ATF4 is induced by various stressors, such as endoplasmic reticulum (ER), amino acid deprivation, and oxidative stress [bib_ref] Activating transcription factor 4, Ameri [/bib_ref]. ER stress and oxidative stress accentuate each other, which have been shown to activate apoptotic signaling in vitro and in vivo models [bib_ref] Endoplasmic reticulum stress and oxidative stress: a vicious cycle or a double-edged..., Malhotra [/bib_ref]. To test the possibility that 7KCh acts through activation of ATF4, we examined the expression level of ATF4 in 7KCh-treated HL-1 cells. The ATF4 level increased in a time-dependent manner [fig_ref] Figure 5: 7KCh activates transcription factor 4 [/fig_ref]. Moreover, the levels of CHOP and active caspase 3 increased significantly at 36 h after treatment [fig_ref] Figure 5: 7KCh activates transcription factor 4 [/fig_ref] , whereas that of LC3A/B increased slightly during the same period [fig_ref] Figure 5: 7KCh activates transcription factor 4 [/fig_ref]. These findings suggest that induction of ATF4 and CHOP contribute to death of cardiomyocytes.
# Discussion
The lipidomes of erythrocytes from patients with HF were significantly different from those of normal controls. Moreover, 7KCh, lysoPCs, lysoPEs, and ceramides significantly increased in abundance in HF erythrocytes, whereas the levels of PCs, PEs, and SMs significantly decreased. Of these lipids, 7KCh was best at discriminating between patients and controls, and may thus serve as a biomarker for early identification of individuals at risk of HF. Additionally, 7KCh may be implicated in death of cardiomyocytes and HF pathogenesis.
Chronic inflammation is associated with HF progression. A number of proinflammatory cytokines, such as tumor necrosis factor α, interleukin (IL)-1, and IL-6, were implicated in this process [bib_ref] Role of Inflammation in Heart Failure, Shirazi [/bib_ref]. In general, chronic inflammation leads to increased oxidative stress and damage and probably accounts for some of the observed changes in HF erythrocytes. Oxidative stress induces phospholipase activity, which leads to a decline in phospholipid levels and an increase in lysophospholipids levels. Moreover, 7KCh, an oxidation product of cholesterol, accumulates as a consequence of oxidative stress. Previous studies have revealed that oxidative damage products, such as oxidized LDL and oxysterols, are found in patients with cardiovascular disease [bib_ref] Elevated levels of oxidized low density lipoprotein show a positive relationship with..., Ehara [/bib_ref] [bib_ref] A critical look at the evidence for the oxidation of LDL in..., Steinberg [/bib_ref]. 7KCh is considered an important metabolite for monitoring cardiovascular disease outcomes and mortality [bib_ref] Association of Plasma 7-Ketocholesterol With Cardiovascular Outcomes and Total Mortality in Patients..., Song [/bib_ref] as well as for predicting the incidence of cardiovascular disease events in general population. Accumulation of 7KCh in HF erythrocytes suggests that 7KCh is a risk factor for HF, with a potential for clinical applications.
7KCh is mainly derived from oxidation of biomembranes and lipoproteins. Dietary intake of food oxysterols are not probably an important source of 7KCh, as dietary 7KCh can be rapidly metabolized to bile acid in liver and excreted [bib_ref] Rapid hepatic metabolism of 7-ketocholesterol by 11beta-hydroxysteroid dehydrogenase type 1: species-specific differences..., Schweizer [/bib_ref] [bib_ref] Rapid hepatic metabolism of 7-ketocholesterol in vivo: implications for dietary oxysterols, Lyons [/bib_ref] [bib_ref] 7-Ketocholesterol delivered to mice in chylomicron remnant-like particles is rapidly metabolised, excreted..., Lyons [/bib_ref]. Erythrocyte 7KCh may be derived from oxidation of membrane of erythrocytes. Moreover, elevation in erythrocyte 7KCh levels may be related to the changes in its catabolism in patients with HF. CYP7A1, the rate-limiting enzyme in bile acid biosynthesis, is the only hepatic enzyme known to be involved in 7KCh catabolism [bib_ref] Extra-hepatic metabolism of 7-ketocholesterol occurs by esterification to fatty acids via cPLA2alpha..., Lee [/bib_ref]. CYP7A1 deficiency causes premature atherosclerosis in humans [bib_ref] Bile acids: regulation of synthesis, Chiang [/bib_ref] [bib_ref] Human cholesterol 7alpha-hydroxylase (CYP7A1) deficiency has a hypercholesterolemic phenotype, Pullinger [/bib_ref]. The extrahepatic metabolism of 7KCh occurs by esterification to fatty acids through cytosolic sterol O-acyltransferase (SOAT) and subsequent selective efflux to HDL [bib_ref] Extra-hepatic metabolism of 7-ketocholesterol occurs by esterification to fatty acids via cPLA2alpha..., Lee [/bib_ref]. HDL . Quantification of the free forms of 7KCh and other sterols in plasma and erythrocytes from patients with HF and controls. Erythrocyte samples were extracted for free-form sterols, and the levels of cholesterol, lanosterol, lathosterol, 7-dehydrocholesterol, and 7KCh (7-ketocholesterol) were quantified using a LC-MS-MS. The levels of 7KCh, lanosterol, lathosterol, and 7dehydrocholesterol were normalized to the level of cholesterol (per mmole). All samples from the normal control subjects (NC, n = 28) and patients with HF in stages A (n = 29), B (n = 29), and C (n = 29) are represented in panels B, C, D, and E, respectively. The free-form sterols in the corresponding plasma samples were determined, and the levels were also normalized to the level of cholesterol (per mmole) and are shown in panels F, G, H, and I. A schematic showing the cholesterol biosynthesis pathway (A). Data are expressed as 7KCh/ Cholesterol (μmole/mmole). Data are means ± SD; *p < 0.05, **p < 0.01, ***p < 0.001 for patients with stages A, B, or C vs. controls. transfers esterified 7KCh back to liver for further catabolism. Relatively low SOAT1 and SOAT2 expression in heart tissue leads to poor esterification and accumulation of free-form 7KCh. A decrease in reverse cholesterol transport (RCT) via HDL ensues [bib_ref] Extra-hepatic metabolism of 7-ketocholesterol occurs by esterification to fatty acids via cPLA2alpha..., Lee [/bib_ref]. The situation is aggravated by lower plasma HDL content in HF patients. It has been recently shown that erythrocytes play an important role in RCT. Animal study has shown that erythrocytes acquired tritiated cholesterol from subcutaneously injected foam [bib_ref] Red blood cells play a role in reverse cholesterol transport, Hung [/bib_ref]. Also, erythrocytes can exchange cholesterol with lipoproteins [bib_ref] Quantitation of the in vitro free cholesterol exchange of human red cells..., Quarfordt [/bib_ref]. It is likely that erythrocytes may take up 7KCh from peripheral tissues and/or from other lipoproteins. Erythrocytes can then transport 7KCh directly to liver for catabolism, or transfer it to HDL via the exchange mechanism. The latter process may be less effective in light of the decrease in plasma HDL in HF patients. An increase in 7KCh abundance in erythrocytes may reflect not only increased oxidative stress but also an ineffectual removal through RCT to HDL in HF patients.
7KCh causes cellular damage and induces oxidative stress in several types of cells, such as endothelial and cardiac cells [bib_ref] 7-Ketocholesterol and 5,6-secosterol induce human endothelial cell dysfunction by differential mechanisms, Luchetti [/bib_ref] [bib_ref] Iron nanoparticles increase 7-ketocholesterolinduced cell death, inflammation, and oxidation on murine cardiac..., Kahn [/bib_ref]. Exposure to 7KCh promotes inflammation, ER stress [bib_ref] 7-Ketocholesterol induces autophagy in vascular smooth muscle cells through Nox4 and Atg4B, He [/bib_ref] [bib_ref] NAD(P)H oxidase Nox-4 mediates 7-ketocholesterol-induced endoplasmic reticulum stress and apoptosis in human..., Pedruzzi [/bib_ref] , and lysosomal dysfunction [bib_ref] 7-Ketocholesterol-induced lysosomal dysfunction exacerbates vascular smooth muscle cell calcification via oxidative stress, Sudo [/bib_ref]. 7KCh triggers autophagy through inhibition of Atg4B activity [bib_ref] 7-Ketocholesterol induces autophagy in vascular smooth muscle cells through Nox4 and Atg4B, He [/bib_ref] , and induces lysosomal dysfunction in vascular smooth muscle cells through an increase in oxidative stress [bib_ref] Iron nanoparticles increase 7-ketocholesterolinduced cell death, inflammation, and oxidation on murine cardiac..., Kahn [/bib_ref]. Our findings show that 7KCh induces ROS formation and causes cardiomyocyte death, which is probably mediated by the ATF4/CHOP pathway. Previous studies have shown that ER stress leads to activation of PERK, phosphorylation of eukaryotic translation initiation factor 2, and subsequent translation of mRNAs, including ATF4. ATF4 itself can activate the transcription of CHOP, which is essential to ER stress-induced cell death [bib_ref] Protein translation and folding are coupled by an endoplasmic-reticulum-resident kinase, Harding [/bib_ref] [bib_ref] Translational control is required for the unfolded protein response and in vivo..., Scheuner [/bib_ref] [bib_ref] Regulated translation initiation controls stress-induced gene expression in mammalian cells, Harding [/bib_ref] [bib_ref] Feedback inhibition of the unfolded protein response by GADD34-mediated dephosphorylation of eIF2alpha, Novoa [/bib_ref] [bib_ref] CHOP/GADD153 is a mediator of apoptotic death in substantia nigra dopamine neurons..., Silva [/bib_ref] [bib_ref] CHOP is involved in endoplasmic reticulum stress-induced apoptosis by enhancing DR5 expression..., Yamaguchi [/bib_ref] [bib_ref] Roles of CHOP/GADD153 in endoplasmic reticulum stress, Oyadomari [/bib_ref].
Our findings may have interesting implications about the transport functions of erythrocytes. Erythrocytes may act as a transporter of substances other than oxygen. The free-form cholesterol of the erythrocyte plasma membrane can be bidirectionally exchanged with that of plasma lipoprotein and cellular plasma membrane. Erythrocytes were postulated to play a role in RCT, particularly in the low HDL state [bib_ref] Red blood cells play a role in reverse cholesterol transport, Hung [/bib_ref]. Our unpublished findings have shown an inverse correlation between erythrocyte 7KCh and plasma HDL in HF patients. Erythrocytes, in addition to HDL, may perform RCT function. Another intriguing possibility is that erythrocytes may transport 7KCh to cardiac tissue and inflict damage to cardiac cells. The proposed scheme is summarized in .
The current study has several limitations. One of the limitations pertains to the small cohort size of patients at different stages of HF. Although we demonstrated that 7KCh induces cardiomyocyte death, the . Quantification of the total forms of 7KCh and other sterols in plasma from HF patients. For total cholesterol (free cholesterol plus cholesteryl ester) quantification, plasma samples were pre-treated with cholesteryl ester hydroxylase as described in Materials and methods section. The sterols were extracted, and the levels of cholesterol, lanosterol, lathosterol, 7-dehydrocholesterol and 7KCh (7-ketocholesterol) in sample were quantified by LC-TQ mass spectrometer. The levels of 7KCh, lanosterol, lathosterol, 7-dehydrocholesterol in samples from normal control subjects (NC, n=10) and patients with HF in stage A (n=10), stage B (n=9), and stage C (n=10) were represented in panel I, J, K, and L, respectively. The free form of sterols in the corresponding plasma samples were determined, and the levels were shown in panel E, F, G, and H. For comparison, the free form of sterols in the corresponding erythrocyte samples were shown in panel A, B, C, and D. Data are expressed as 7KCh/Cholesterol (μmole/mmole). Data are means ± SD; *p < 0.05, ***p < 0.001, group of stage A, B, or C patients vs. control group. evidence supporting the causal relationship between an increase in erythrocyte 7KCh and HF development is far from complete. An animal model is needed in further study to validate our findings.
In conclusion, blood 7KCh is concentrated in erythrocytes in patients with HF. 7KCh-laden erythrocyte ghost induces cardiomyocyte death, suggesting its involvement in the pathogenesis of HF. Moreover, our findings support the importance of erythrocyte 7KCh as a risk factor for HF. HL-1 cells were treated with 10 μM 7KCh for the indicated periods (0, 12, 24, 36, 48 h), and the cell lysate was analyzed for the levels of ATF4, CHOP, and caspase 3, LC3 and actin (loading control) by immunoblotting with respective antibodies. A representative experiment out of three is shown. The relative intensities of ATF4 (C), CHOP (D), cleaved Caspase-3 (E), caspase-3 (F), and LC3A/B (G) were normalized to that of actin, and are expressed as fold change relative to those of untreated HL-1 cells. Data are mean ± SD, n=3. *p < 0.05, **p < 0.01, ***p < 0.001 vs. untreated cells.
[fig] Figure 4: Viability reduction and ROS generation in 7KCh accrued HL-1 cells. HL-1 cells (5 × 10 4 /well) were treated with the indicated concentrations (0, 10, 20, 50 μM) of 7KCh and cholesterol (Chol) for 24 h (A) and 48 h (B), respectively. In addition, cells were treated with different concentrations (0, 10, 20, 50, 75, 100, 150, 200 μM) of 7KCh-or cholesterol-loaded erythrocyte ghost for 24 h (C) and 48 h (D), respectively. The viabilities of the treated cells were determined, and are expressed as percentage of that of untreated HL-1 cells. Data are means ± SD, n = 3; *p < 0.05, **p < 0.01, ***p < 0.001 for 7KCh-treated cells vs. untreated cells. (E) Immunostaining for 7KCh in HL-1 cells. HL-1 cells were untreated (control) or treated 20 μM 7KCh, and stained with primary antibody to 7KCh and secondary antibody conjugated to Alexa-488 (green). Hoechst 33342 (blue) was used to stain cellular nuclei. A representative experiment out of three is shown. (F) 7KCh induces ROS generation. HL-1 cells were treated with 10, 20 μM 7KCh or cholesterol (Chol) for 24 h, stained with H 2 DCFDA (green) and MitoSOX red (red), and analyzed using flow cytometry. H 2 O 2 -treated cells were positive control. The MFI was determined, and is expressed as fold change relative to that of untreated HL-1 cells. Data are means ± SD, n = 3; *p < 0.05 for 7KCh-treated cells vs. untreated cells. [/fig]
[fig] Figure 5: 7KCh activates transcription factor 4 (ATF4) pathway. (A) A schematic diagram showing the proposed effect of 7KCh on ATF4/CHOP pathway. (B) [/fig]
[table] Table 1: Baseline characteristics of the study populations. [/table]
[table] Table 2: The potential biomarkers identified for discriminating patients with HF from normal controls. [/table]
|
Glucocorticoids in rheumatoid arthritis: current status and future studies
## What does this study add?
► This review presents and discusses recent literature data on glucocorticoids use in RA and provides suggestions for future research in the field of glucocorticoids in RA.
## How might this impact on clinical practice?
► Glucocorticoids in RA must be used in combination with disease-modifying antirheumatic drugs (DMARDs), notably as bridging therapy with conventional synthetic DMARDs. ► evaluation of the benefit/risk ratio must be systematic, even for low-dose glucocorticoids.
# Abstract
Since their first use for treating rheumatoid arthritis (RA) in the late 1940s, glucocorticoids (GCs) have been representing a substantial part of the therapeutic arsenal for RA. However, even if GCs are still widely prescribed drugs, their toxicity is discussed controversially, so obtaining consensus on their use in RA is difficult. Hence, the most recent european League Against Rheumatism and American College of Rheumatology recommendations on early arthritis and RA management advocate the use of GCs as adjunct treatment to conventional synthetic disease-modifying antirheumatic drugs, at the lowest dose possible and for the shortest time possible. However, the recommendations remain relatively vague on dose regimens and routes of administration. Here, we describe literature data on which the current recommendations are based as well as data from recent trials published since the drafting of the guidelines. Moreover, we make proposals for daily practice and provide suggestions for studies that could help clarifying the place of GCs in RA management. indeed, numerous items, including the benefit/risk ratio of low-dose and very low-dose GCs and optimal duration of GCs as bridging therapy, remain on the research agenda, and future studies are needed to guide the next recommendations for RA.
# Introduction
In this time of targeted therapies, therapeutic strategies and management of comorbidities in patients with rheumatoid arthritis (RA), the question of the future position glucocorticoids (GCs) may have in RA is worth asking. Indeed, although therapy with GCs was a major therapeutic advance in the 1950s for RA, the current emphasis is more on the disadvantages than the benefits of this treatment. However, in light of recent literature data, low-dose GCs still seem to have an important role in RA. [bib_ref] Current view of glucocorticoid co-therapy with DMARDs in rheumatoid arthritis, Hoes [/bib_ref] [bib_ref] Low-dose glucocorticoid therapy in rheumatoid arthritis: an obligatory therapy, Bijlsma [/bib_ref] [bib_ref] Long-term prednisone in doses of less than 5 mg/day for treatment of..., Pincus [/bib_ref] The goal of treatments for patients with RA in the short term is to decrease disease activity and achieve clinical remission, and in the medium term to limit or prevent structural progression, disability and systemic manifestations. Available disease-modifying antirheumatic drugs (DMARDs) have shown their efficacy. Conventional synthetic (cs) DMARDs have a relatively long onset of action while most of biological (b) and targeted synthetic (ts) DMARDs (bDMARDs and tsDMARDs) act faster. GCs in RA provide the advantage of a rapid onset of action, which allows waiting for the onset of csDMARDs efficacy. Moreover, even if GCs are mainly widely used whenever clinicians need rapid symptomatic relief for their patients with RA, their structural effect must not be forgotten. However, the benefit/risk ratio of GCs remains precarious and their modalities of use in RA remain controversial.
In this review, we first detail the latest European and US recommendations on GCs use in RA and discuss the use of GCs in current practice. Then, we consider recent literature data on GCs efficacy (both clinical and structural) in RA and the various ways of using GCs. Finally, we address the adverse effects potentially associated with GCs. Our work being a narrative review and not a systematic review, we have not included all the existing studies but only the ones seeming the most relevant to us. Moreover, during this review, RMD Open RMD Open RMD Open we propose topics that could be of interest to be evaluated in future studies of GCs use in RA.
## International recommendations european recommendations
Recent updates of the European recommendations for management of early arthritis and RA 4-6 placed a greater focus on the benefit of GCs therapy than the previous versions. Short-term GCs therapy should now be considered as part of the initial treatment strategy and subsequently if an initial strategy has failed, as bridging therapy if a change in a csDMARD is considered. [bib_ref] EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological..., Smolen [/bib_ref] GCs should be tapered as rapidly as is clinically feasible: long-term use of GCs should be avoided, GCs should be gradually reduced and stopped, usually within 3 months and only exceptionally by 6 months. [bib_ref] EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological..., Smolen [/bib_ref] The term 'low-dose' GCs was replaced by 'short-term' GCs to take into account several current ways of using GCs, such as parenteral bolus. [bib_ref] EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological..., Smolen [/bib_ref]
## Us recommendations
The 2015 American College of Rheumatology (ACR) guidelines for early and established RA recommend adding GCs to DMARDs during disease flares, at the lowest dose and for the shortest period possible. [bib_ref] American College of rheumatology guideline for the treatment of rheumatoid arthritis, Singh [/bib_ref] In contrast to the European recommendations, adding GCs when a csDMARD is initiated depends on disease activity.
Taken together, these international recommendations agree on the use of GCs for disease flare and possibly at the start of a new csDMARD. Specific advice concerning dosage, duration, route of administration and strategies is limited and less consensual, because reliable and detailed evidence is scarce. For US recommendations, a dose <10 mg/day is considered a low dose, and GCs should be tapered in less than 3 months, whereas for European recommendations, the threshold is 7.5 mg/day, and GCs could be prescribed in combination with csDMARDs for up to 6 months maximum, knowing that this duration is mainly expert-driven. [bib_ref] When and for how long should glucocorticoids be used in rheumatoid arthritis?..., Gaujoux-Viala [/bib_ref] Despite these differences, international guidelines underline the importance of GCs but also advocate the use of GCs at the lowest cumulative dose possible because of the high awareness of potentially associated adverse effects.
In these sets of recommendations as well as in this review, GCs dosages are expressed in prednisone equivalent.
CurrenT praCTICe A recent study of an Australian cohort of patients with RA showed that the probability of GCs use throughout follow-up has decreased over time, from 55% in 2001 to 39% in 2012 (p<0.001). [bib_ref] Factors associated with oral glucocorticoid use in patients with rheumatoid arthritis: a..., Black [/bib_ref] In this cohort, current csDMARD use but not bDMARD use was associated with increased current GCs use. In a recent analysis considering the years from 1980 up to 2004, the reduction of mean initial low-dose, for long-term GCs therapy in RA, was found from 10.3 mg/day up to 3.6 mg/day. [bib_ref] Decline of mean initial prednisone dosage from 10.3 to 3.6 mg/day to..., Pincus [/bib_ref] In contrast, another observational cohort study showed that the proportion of patients initiating GCs was higher in the group from 1995 to 2007 compared with the earlier group from 1980 to 1994 (68% vs 36%) but the cumulative dose did not differ over the first year. [bib_ref] Time trends in glucocorticoid use in rheumatoid arthritis: results from a population-based..., Makol [/bib_ref] Anyway, GCs are still widely used in RA. GCs appear to be used in approximately 50% of patients with RA, [bib_ref] QUEST-RA: quantitative clinical assessment of patients with rheumatoid arthritis seen in standard..., Sokka [/bib_ref] with varied duration and dosage among the studies. In the German CAPEA cohort of patients with early arthritis, 82% received methotrexate (MTX) within the first months, 77% received GCs and 20% of these received <7.5 mg/day prednisone but one third received >20 mg/ day. [bib_ref] High variability in glucocorticoid starting doses in patients with rheumatoid arthritis: observational..., Albrecht [/bib_ref] After 2 years of follow-up, 12% of the patients received biologics, 52% were free of GCs and 41% were receiving <5 mg/day. In the French ESPOIR cohort of patients with early arthritis, 45% started GCs during the first 6 months and more than 50% received GCs at least once over 5 years after inclusion. 14 Overall, the dose of GCs received during follow-up was very low, the mean was 3.1±2.9 mg/day. [bib_ref] Seven-Year tolerability profile of glucocorticoids use in early rheumatoid arthritis: data from..., Roubille [/bib_ref] In the Canadian CATCH cohort of patients with RA, 42% were considered GCs users and the median oral daily dose was 5 mg (IQR 2.5-10). [bib_ref] Quality assurance study of the use of preventative therapies in glucocorticoid-induced osteoporosis..., Mckeown [/bib_ref] GluCoCorTICoIds effICaCy Clinical efficacy Current knowledge For reasons of brevity, in this part, we will focus on only the most relevant and recent data on clinical efficacy of GCs in RA published during the last 6 years (table 1). In the following parts of our work, we will discuss other relevant studies published earlier.
Overall, these results agreed with previous findings and with the conclusion of the most recent systematic literature reviews published on this topic and showed a beneficial effect of GCs when added to csDMARDs. The available data primarily relate to GCs in addition to csDMARDs and not, or not specifically, to bDMARDs or tsDMARDs. Moreover, the current literature concerns mainly patients with early arthritis, and studies reporting on GCs efficacy in patients with established RA are clearly less frequent.
In the CAPRA-2 trial, a study of patients with established RA (disease duration approximately 8 years) and active disease, low-dose (5 mg/day) prednisone with a modified release formulation (chronotherapy) added to existing DMARDs significantly ameliorated disease activity at 12 weeks as compared with placebo (PBO) added to existing DMARDs. DMARDs were almost exclusively csDMARDs because only one patient in each group received a bDMARD.
In the CareRA trial, patients with early RA without poor prognosis markers were randomised to one of two treatment arms. [bib_ref] Patients lacking classical poor prognostic markers might also benefit from a step-down..., Verschueren [/bib_ref] In one arm, GCs were initially associated with MTX (30 mg/day prednisone decreased to 5 mg/ day in 6 weeks) while in the other arm, MTX was initiated without GCs. Disease Activity Score in 28 joints (DAS28) remission at 16 weeks was more frequently achieved in Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis patients with than without GCs, although not significantly (65% vs 47%, p=0.08). [bib_ref] Patients lacking classical poor prognostic markers might also benefit from a step-down..., Verschueren [/bib_ref] According to data at 1 and 2 years, the rates of remission were still higher in the MTX+GCs than MTX-only arm but again not significantly. Ten-year data from the BeSt study were published in 2016. [bib_ref] Long-Term outcomes of patients with recent-onset rheumatoid arthritis after 10 years of..., Markusse [/bib_ref] In this trial, 508 patients with early active RA were randomised to four arms: a pre-established maintenance treatment regimen prescribed as monotherapy beginning with MTX; a step-up group with sulfasalazine (SSZ) added to MTX in case of failure; a group following the COBRA scheme (SSZ, MTX and GCs initially at 60 mg/day, then progressively tapered to 7.5 mg/day in 6 weeks) and a group receiving MTX and infliximab from the beginning. [bib_ref] Clinical and radiographic outcomes of four different treatment strategies in patients with..., Goekoop-Ruiterman [/bib_ref] In the initial study, patients following the COBRA strategy showed better clinical efficacy at 3 months than patients without GCs. At 10 years, approximately 50% of patients were in remission whatever their initial group of randomisation. In the meantime, strategy-driven changes have occurred and it is thus difficult to conclude on the long-term effect of initial GCs therapy. [bib_ref] Long-Term outcomes of patients with recent-onset rheumatoid arthritis after 10 years of..., Markusse [/bib_ref] Results of the post-trial follow-up of the CAMERA-II study were published in 2017. [bib_ref] Long-term outcome is better when a methotrexate-based treatment strategy is combined with..., Safy [/bib_ref] In this study of 236 patients with early RA, the addition during 2 years of 10 mg/day prednisone to MTX was compared with a MTX and PBO group. [bib_ref] Low-dose prednisone inclusion in a methotrexate-based, tight control strategy for early rheumatoid..., Bakker [/bib_ref] Disease activity after 2 years of treatment improved more on average in the MTX-GCs arm than MTX-PBO arm, but the differences observed in the first months tended to decrease over time. In the follow-up study (median follow-up of approximately 6.6 years), significantly fewer patients of the MTX-GCs group had initiated a bDMARD than those from the MTX-PBO group (31% vs 50%, p<0.05). [bib_ref] Long-term outcome is better when a methotrexate-based treatment strategy is combined with..., Safy [/bib_ref] Data at 10 years from the BARFOT trial have been reported. [bib_ref] Low-dose prednisolone treatment of early rheumatoid arthritis and late cardiovascular outcome and..., Ajeganova [/bib_ref] In this open randomised trial of 250 patients with early RA, the addition during 2 years of 7.5 mg/ day prednisolone to csDMARDs was compared with csDMARDs alone. Clinical outcomes were better in the GCs group at all time points (3, 6, 12, 18 and 24 months). [bib_ref] Low-dose prednisolone in addition to the initial disease-modifying antirheumatic drug in patients..., Svensson [/bib_ref] A follow-up study at 4 years no longer found differences in the proportion of patients in remission between the groups. [bib_ref] Remission achieved after 2 years treatment with low-dose prednisolone in addition to..., Hafström [/bib_ref] At 10 years, bDMARD use did not differ with and without GCs. [bib_ref] Low-dose prednisolone treatment of early rheumatoid arthritis and late cardiovascular outcome and..., Ajeganova [/bib_ref] Of note, the proportion of patients who used a bDMARD was very low (15% in each group), perhaps because patients were included in the BARFOT cohort between 1995 and 1999, before the era of biologics.
In light of these and other literature data, there is little doubt that GCs are effective for reducing disease activity in patients with RA, at least in the short term. However, confirming whether the clinical benefit of GCs persists in the medium and long term is difficult. Because of their toxicity and moderate structural effect (see below), GCs cannot be considered monotherapy and should always be used together with DMARDs.
## Studies to perform
We notably lack studies evaluating the clinical effect in the medium and long term with short-term (<6 months) GCs therapy. Such studies should be RCTs including RMD Open RMD Open RMD Open patients with RA for which a specific DMARD has to be initiated. These patients should be randomised into an arm receiving the DMARD with GCs discontinued over 3 months and an arm receiving the DMARD without GCs. Clinical efficacy measures should occur at medium and long term: after several months and even years of follow-up. As stated above, it would be of particular interest to conduct such studies in patients initiating a targeted DMARD and not only in patients initiating a csDMARD. Trials addressing the advantages provided by new formulations of GCs would also be of interest. Indeed, the pharmaceutical field of GCs continues to evolve, notably with the development of modified-release or delayed-release prednisone (chronotherapy) or GC receptor agonists to improve the efficacy and/or reduce toxicity. [bib_ref] Glucocorticoids and chronotherapy in rheumatoid arthritis, Cutolo [/bib_ref] [bib_ref] Old but good: modified-release prednisone in rheumatoid arthritis, Ursini [/bib_ref] [bib_ref] PF-04171327) versus prednisone or placebo in rheumatoid arthritis: a randomised, double-blind, multicentre,..., Buttgereit [/bib_ref] Modified night-release formulation of lowdose prednisone, although administered in the evening (acting like a replacement therapy), has been developed to contrast the circadian rise in proinflammatory cytokine levels (night), that contributes to RA disease activity and might represent the way to further optimise the DMARD activity exerted by low dose GCs in RA. [bib_ref] Circadian rhythms in rheumatology--a glucocorticoid perspective, Spies [/bib_ref] structural effect Current knowledge In this part, we have included all the studies on structural effects of GCs seeming relevant to us, without any time limit. In most trials, GCs were used in combination with csDMARDs, so concluding on the specific effect of GCs is difficult.
In 2005, the BARFOT randomised trial showed the slowing at 2 years of the structural progression more important with than without GCs (0.2% vs 0.4%, p=0.02). [bib_ref] Low-dose prednisolone in addition to the initial disease-modifying antirheumatic drug in patients..., Svensson [/bib_ref] The follow-up at 4 years suggested a lower increase in total Sharp score during the 4 years with than without GCs, although not significantly (p=0.079). [bib_ref] Remission achieved after 2 years treatment with low-dose prednisolone in addition to..., Hafström [/bib_ref] In 2007, a meta-analysis of 15 studies was in favour of a significant reduction of erosion progression by GCs given in addition to csDMARDs. [bib_ref] Effects of glucocorticoids on radiological progression in rheumatoid arthritis, Kirwan [/bib_ref] In 2010, the structural impact of GCs added to csDMARDs was evaluated in another meta-analysis, showing approximately 70% reduced radiographic progression (p=0.0008). [bib_ref] Similar effects of disease-modifying antirheumatic drugs, glucocorticoids, and biologic agents on radiographic..., Graudal [/bib_ref] In this meta-analysis, the combination of GCs to csDMARDs had a similar effect as bDMARDs added to MTX (difference of 7% in radiographic progression between the two groups, p=0.44). However, the conclusions of this meta-analysis were limited by the high heterogeneity of included trials. In the 11-year follow-up study of the COBRA trial, structural damage progression was lower in the group initially treated according to the COBRA scheme (MTX+SSZ+GCs (60 mg/day to 7.5 mg/day in 6 weeks)) than in the group initially treated with SSZ monotherapy. [bib_ref] Survival, comorbidities and joint damage 11 years after the cobra combination therapy..., Van Tuyl [/bib_ref] More recently, in the CAMERA-II study, radiographic progression with GCs was significantly reduced, with 78% of patients in the MTX-GCs group remaining erosion-free after 2 years vs only 67% in the MTX-PBO group. [bib_ref] Low-dose prednisone inclusion in a methotrexate-based, tight control strategy for early rheumatoid..., Bakker [/bib_ref] Similarly, at the 2-year post-trial follow-up, the median erosion score was significantly lower in the former MTX-GCs group than in the former MTX-PBO group (p=0.002). [bib_ref] Long-term outcome is better when a methotrexate-based treatment strategy is combined with..., Safy [/bib_ref] In light of the literature, the structural effect of GCs seems clear, at least when GCs are added to csDMARDs. Moreover, this positive effect on radiographic progression seems to last even after discontinuation of GCs. Obviously, when combined with bDMARDs or tsDMARDs, showing any additional effect of GCs is difficult because these drugs already strongly inhibit structural progression.
## Studies to perform
Studies comparing different durations of GCs treatment as bridging therapy in terms of structural protection are needed to increase the strength of the recommendations concerning the optimal period of GCs prescription. The approach of these studies would be to evaluate if a shortterm (<6 months) GCs therapy added to MTX is as effective as a more prolonged GCs therapy (during 2 years, eg, as in CAMERA-II or BARFOT trials) on structural progression at medium and long term.
## Glucocorticoids use strategies
For this part, selection of the studies was based only on their relevance and not on their publication date.
## Bridging therapy current knowledge
When and how to introduce and taper GCs are important questions for rheumatologists and their patients. In view of the risks and benefits, the usual practice is to use an induction strategy followed by a progressive decrease. This is notably done when GCs are prescribed as bridging therapy along with the introduction of a csDMARD, to wait for its efficacy. This scheme was used in the COBRA study, with results published in 1997. [bib_ref] Randomised comparison of combined step-down prednisolone, methotrexate and sulphasalazine with sulphasalazine alone..., Boers [/bib_ref] In this study, 155 patients with early RA were randomised into two arms: combination treatment with SSZ, MTX and prednisolone (initially at 60 mg/day, then progressively tapered to 7.5 mg/day in 6 weeks) compared with SSZ only. The results favoured the GCs arm, both on the clinical and on the structural effects. Since this study, various strategies for the use of GCs have been evaluated, but few RCTs have been designed to specifically compare dosage, duration and tapering of GCs.
Recently, in two articles reporting on the same RCT, two different GCs strategies were compared: the COBRAlight strategy (prednisolone at 30 mg/day, tapered to 7.5 mg/day within 9 weeks) combined with MTX, and the original COBRA strategy. The outcomes were similar between the two arms in terms of reducing clinical disease activity, improving functional ability and preventing radiographic progression, showing that treatment regimens including medium-dose GCs are not inferior to regimens with initially high-dose GCs. [bib_ref] Rediscovering the therapeutic use of glucocorticoids in rheumatoid arthritis, Van Der Goes [/bib_ref] A posthoc analysis of six phase III studies evaluating tofacitinib in RA, in monotherapy (ORAL start, ORAL Solo) or combined with csDMARDs (ORAL Scan, ORAL Standard, ORAL Sync, ORAL Step), examined the Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis impact of the presence of GCs in the treatment groups. [bib_ref] Effect of glucocorticoids on the clinical and radiographic efficacy of tofacitinib in..., Charles-Schoeman [/bib_ref] Mean dose of GCs was approximately of 6 mg/day. Across all studies, the concomitant use of GCs did not affect the clinical efficacy of tofacitinib. In contrast, response was frequently better for patients receiving MTX with GCs than without GCs, but the differences were often not statistically significant.
Moreover, in an analysis presented at the 2018 ACR meeting but not yet published of pooled data from four RCTs (AMBITION, ACT-RAY, ADACTA and FUNC-TION) on intravenous tocilizumab (TCZ), concomitant GC therapy had no impact on the clinical efficacy of TCZ at 24 weeks.The authors assessed patients in the comparator arms: the clinical response of patients receiving adalimumab as well as those initiating MTX was not affected by the use of GCs. Data on the mean GCs dose were lacking in this abstract.
According to a posthoc analysis of the TOZURA trial, a phase IV study evaluating subcutaneous TCZ, the proportion of patients achieving DAS28 remission at 24 weeks was broadly similar between the groups receiving TCZ, with or without GCs (median dose 5 mg/day). [bib_ref] Effects of concomitant glucocorticoids in TOZURA, a common-framework study programme of subcutaneous..., Choy [/bib_ref] However, concerning the quoted studies on tofacitinib and TCZ, it is necessary to specify that GCs use was not submitted to randomisation and could be prior to the beginning of the study making it difficult to compare the groups with and without GCs.
## Studies to perform
The current literature mainly concerns the use of GCs as bridging therapy when a csDMARD is initiated, but we have few studies on the use of GCs as bridging therapy when a bDMARD or tsDMARD is initiated. Indeed, the onset of action might be faster for bDMARDs, and even more so for tsDMARDs, than csDMARDs, which could explain why the recent posthoc analyses of the tofacitinib and TCZ studies did not find a benefit of adding GCs. It would be useful to have RCTs designed to evaluate clinical and structural efficacy of a short-term GCs therapy prescribed when a targeted DMARD is initiated. Finally, studies on the best initiation dosage of GCs are lacking, as are those on the duration of GCs as bridging therapy and trials comparing several tapering strategies.
flare treatment Current knowledge Data on this topic are scarce, among other reasons due to lack of an agreed-upon definition of RA flare. [bib_ref] Feasibility and domain validation of rheumatoid arthritis (rA) flare core domain set:..., Bartlett [/bib_ref] In an older study of 18 patients with RA, flares were treated with GCs (prednisone 25 mg/day tapered in 5 days) or PBO. [bib_ref] Endocrine control of inflammation: rheumatoid arthritis double-blind, crossover clinical trial, Stenberg [/bib_ref] Clinical outcomes were better at 6 months when GCs were used. In the BELIRA trial, patients with RA with active disease received a total of 250 mg/day prednisolone over 1 week added to their existing csDMARDs. [bib_ref] Glucocorticoid treatment in rheumatoid arthritis: low-dose therapy does not reduce responsiveness to..., Wolf [/bib_ref] This strategy resulted in a highly significant improvement in clinical, functional and serological measures at 1 week. A RCT comparing the efficacy of three intravenous pulses of 120 mg/day dexamethasone or 1 g/day methylprednisolone suggested that the two strategies were safe and effective for RA flares at 1 month. [bib_ref] Randomized double-blind study of the effect of dexamethasone and methylprednisolone pulse in..., Sadra [/bib_ref] Studies to perform Daily management of RA would be helped by studies investigating the modalities of treating disease flares with GCs, for example, by comparing pulse therapy to lowdose oral GCs. maintenance therapy Current knowledge Despite recommendations, many patients are already being treated for months and years with low doses GCs, without apparent excessive toxicity, and there is no consensus on whether the GCs therapy should be stopped or not. In the international, multicentre SEMIRA trial, 259 patients with RA in remission and treated for more than 6 months by TCZ±csDMARDs+GCs were randomly assigned to continued GCs or GCs taper. [bib_ref] OP0030 randomized controlled 24-week trial evaluating the safety and efficacy of blinded..., Burmester [/bib_ref] Even if continuing GCs provided better disease control than tapering GCs, almost two-thirds of patients still achieved treatment success at 24 weeks and could stop GCs entirely.
## Studies to perform
Results of this trial are suggesting that tapering and discontinuation of GCs should be tested in patients in remission under TCZ and receiving long-term low-dose GCs. In case of a flare, reintroduction of GCs should be considered. Similar studies on patients with RA receiving other treatments than TCZ would be of interest in order to extrapolate the results.
oTHer admInIsTraTIon rouTes parenteral route Current knowledge In a study published in 1993, 41 patients with RA were randomised to receive methylprednisolone orally (500 mg) or intramuscularly (120 mg) at baseline and 4 and 8 weeks. [bib_ref] Intramuscular methylprednisolone is superior to pulse oral methylprednisolone during the introduction phase..., Choy [/bib_ref] At 16 weeks, the intramuscular route seemed better than the oral administration on pain and Health Assessment Questionnaire (HAQ) score. In the tREACH trial, patients with early RA received csDMARDs (MTX monotherapy or MTX+SSZ+hydroxychloroquine (HCQ)) and oral GCs started at 15 mg/day and tapered during 10 weeks or MTX+SSZ+HCQ plus an initial intramuscular pulse of GCs. [bib_ref] Randomised comparison of initial triple DMARD therapy with methotrexate monotherapy in combination..., De Jong [/bib_ref] The two groups did not differ at 1 year in clinical response, structural progression or safety.
## Studies to perform
From the literature data, concluding on the superiority of one GCs route of administration versus another is difficult and that is why current recommendations state that different dose regimens and route of administration can be used. [bib_ref] EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological..., Smolen [/bib_ref] Strategy trials, with designs similar to that of the tREACH study, comparing the efficacy of one or several parenteral (intravenous or intramuscular) GCs injections versus initiation and tapering of oral GCs as
## Rmd open rmd open rmd open
bridging therapy could answer this question. Besides efficacy issues, the parenteral form could be beneficial for preventing self-medication by patients and avoiding the risk with GCs tapering and mid-term to long-term GCs use. [bib_ref] Controversies in rheumatoid arthritis glucocorticoid therapy, Ruyssen-Witrand [/bib_ref] Intra-articular injections Current knowledge Subanalyses of the BeSt study showed that 3 months after intra-articular GCs injection, 50% of joints were no longer swollen. After 1 year, swelling had recurred in 14% of joints with initially good clinical response. [bib_ref] Intra articular injection with corticosteroids in patients with recent onset rheumatoid arthritis:..., Gvozdenović [/bib_ref] Initially, disease activity scores significantly improved, but over time, DAS and HAQ scores became similar in injected and non-injected patients. After 8 years, the two groups did not differ in joint damage. [bib_ref] Rediscovering the therapeutic use of glucocorticoids in rheumatoid arthritis, Van Der Goes [/bib_ref] In the CIMESTRA trial, patients with early RA received intra-articular GCs in any swollen joint (maximum four joints per visit) combined with step-up csDMARDs treatment over 2 years. [bib_ref] Short-and long-term efficacy of intra-articular injections with betamethasone as part of a..., Hetland [/bib_ref] Oral GCs were not allowed. At 2 years, the median cumulative number of injections for each patient was 13, and the cumulative dose of GCs was <1 mg/day. The injections had a rapid-onset anti-inflammatory action, and 2 weeks after inclusion, 39% of patients were in DAS28 remission. At the end of follow-up, 55.5% of joints injected at baseline did not show disease relapse, and intra-articular injections were well tolerated.
Recently, one small and open-label RCT showed greater clinical efficacy (ACR20/50/70 response) of csDMARDs combined with initial intra-articular GCs injections than csDMARDs alone in patients with early RA without oral GCs. [bib_ref] Comparison of intra-articular glucocorticoid injections with DMARDs versus DMARDs alone in rheumatoid..., Menon [/bib_ref] Studies to perform Larger RCTs with a design similar to the trial by Menon et al [bib_ref] Comparison of intra-articular glucocorticoid injections with DMARDs versus DMARDs alone in rheumatoid..., Menon [/bib_ref] and with compliance to blinding are needed, as are RCTs comparing intra-articular GCs injections with other routes of administration.
## Glucocorticoids safety
In this part also, the publication date was not taken into account for the selection of the studies. Due to the considerable amount of data published on this topic, we have decided to notably quote meta-analysis and literature reviews.
## Current knowledge
GCs toxicity is a major concern when higher dosages are given for a longer time. However, even low-dose to medium-dose GCs are associated with adverse effects.
In 2007, a EULAR taskforce published recommendations for the management of systemic GCs therapy in rheumatic diseases. [bib_ref] EULAR evidence-based recommendations on the management of systemic glucocorticoid therapy in rheumatic..., Hoes [/bib_ref] According to a literature review, the taskforce identified the following main adverse events of GCs: cardiovascular diseases, infections, gastrointestinal diseases, psychological disorders, endocrine pathologies, dermatological issues, musculoskeletal disorders (including osteoporosis) and ophthalmological diseases. [bib_ref] EULAR evidence-based recommendations on the management of systemic glucocorticoid therapy in rheumatic..., Hoes [/bib_ref] In a meta-analysis of trials and follow-up studies published in 2009, the rate of adverse effects linked to GCs (prescribed for RA) at <30 mg/day was 43/100 patient-years (95% CI 30 to 55). [bib_ref] Adverse events of low-to medium-dose oral glucocorticoids in inflammatory diseases: a meta-analysis, Hoes [/bib_ref] Another meta-analysis, also published in 2009, pooled the results of six RCTs comparing low-dose GCs (mean dose 5-10 mg/day) to PBO. [bib_ref] Safety of medium-to longterm glucocorticoid therapy in rheumatoid arthritis: a meta-analysis, Ravindran [/bib_ref] According to this meta-analysis, the groups did not differ in any type of event or serious adverse events. According to the 11-year follow-up study of the COBRA trial, rates of hypertension and diabetes were higher in the group initially treated according to the COBRA scheme than in the group initially treated with SSZ monotherapy but rates of cardiovascular disease were similar between the two groups, knowing that patients from this group had also received GCs during follow-up. [bib_ref] Survival, comorbidities and joint damage 11 years after the cobra combination therapy..., Van Tuyl [/bib_ref] After 23 years of follow-up, the mortality rate of patients from the COBRA trial was similar to that of the general population, whatever the initial randomisation group. [bib_ref] Normal mortality of the cobra early rheumatoid arthritis trial cohort after 23..., Poppelaars [/bib_ref] From the 10-year data of the BARFOT trial, incident cardiovascular events were evenly distributed with and without GCs (15% and 14%), but the risk of the first cerebrovascular event was almost four times higher with than without GCs (HR 3.7 (95% CI 1.2 to 11.4), with higher mortality with GCs but not significantly. [bib_ref] Low-dose prednisolone treatment of early rheumatoid arthritis and late cardiovascular outcome and..., Ajeganova [/bib_ref] Of note, 53% of the patients with GCs had stopped them after the first 2 years, and the mean dose of prednisolone for those taking GCs was 7.2±1.1 mg/day at the 2-year follow-up, 6.5±3.6 mg/day at 4 years and 4.9±3.3 mg/day at 8 years. [bib_ref] Low-dose prednisolone treatment of early rheumatoid arthritis and late cardiovascular outcome and..., Ajeganova [/bib_ref] Long-term (median follow-up almost 7 years) post-trial data from the CAMERA-II study are in agreement with these findings, showing no statistical difference in incident GCs-related comorbidities between the MTX-GCs and MTX-PBOarms, but numerical differences in cardiovascular events (13 vs 8) and death (10 vs 6). [bib_ref] Long-term outcome is better when a methotrexate-based treatment strategy is combined with..., Safy [/bib_ref] Similar to the BARFOT post-trial study, the controlled design was stopped after the first 2 years, and only half of the patients from the MTX-GCs arm had stopped GCs at 3 years. In summary, data from RCTs are quite reassuring in terms of low-dose GCs safety, but most RCTs had only a short follow-up. However, in the two long-term post-trial studies of BARFOT and CAMERA-II, in which patients had initially taken GCs during 2 years, GCs were associated with more cardiovascular events, but not significantly.
Observational studies usually showed significantly more adverse events in patients with than without GCs, whether for cardiovascular events or for other adverse effects linked to GCs such as infections 59 or osteoporosis, and even for low-dose GCs. [bib_ref] Safety of low-to medium-dose glucocorticoid treatment in rheumatoid arthritis: myths and reality..., Santiago [/bib_ref] However, data from observational studies should be considered with caution, notably because of population heterogeneity and confounding by indication. Nevertheless, cardiovascular toxicity with low-dose GCs was confirmed in a recent meta-analysis of RCTs and observational studies, noting 47% increased cardiovascular events in patients with RA receiving GCs. [bib_ref] The effects of tumour necrosis factor inhibitors, methotrexate, non-steroidal anti-inflammatory drugs and..., Roubille [/bib_ref] In 2014, a North American study reported a threshold of 8-15 mg/day for increased mortality linked to GCs Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis in RA (adjusted HR 1.78 (95% CI 1.22 to 3.60)), after adjusting for potential confounders and the propensity to receive GCs. [bib_ref] Glucocorticoid dose thresholds associated with all-cause and cardiovascular mortality in rheumatoid arthritis, Del Rincón [/bib_ref] The minimum total cumulative dose associated with all-cause mortality was 40 g (adjusted HR 1.89 (95% CI 1.25 to 2.44)). [bib_ref] Glucocorticoid dose thresholds associated with all-cause and cardiovascular mortality in rheumatoid arthritis, Del Rincón [/bib_ref] Reaching this cumulative dose threshold would take approximately 21 years with a daily dose of 5 mg/day. Patients with RA can commonly reach such a duration of exposure. [bib_ref] Glucocorticoid treatment in rheumatoid arthritis, Rau [/bib_ref] It is clear that adverse effects are related to the dose and duration of GCs but the data on the toxic effects of a low cumulative dose of GCs are scarce. In the CareRA trial, the proportion of patients with adverse events at 1 year was similar between patients who had received GCs according to the COBRA Slim scheme and patients without GCs. [bib_ref] Effectiveness of methotrexate with step-down glucocorticoid remission induction (cobra slim) versus other..., Verschueren [/bib_ref] A recent study based on data from the French cohort of early arthritis patients (ESPOIR cohort) found no significant over-risk at 7 years of follow-up, despite numerical differences in cardiovascular events, infectious diseases or osteoporotic fracture between patients who received very low-dose GCs (mean dose 3.1±2.9 mg/day) and patients who never received GCs. [bib_ref] Seven-Year tolerability profile of glucocorticoids use in early rheumatoid arthritis: data from..., Roubille [/bib_ref] studies to perform Studies comparing various strategies of GCs use in terms of the frequency of adverse effects could help practitioners to prescribe GCs in the most suitable way. The toxic effects of GCs are clearly dose-dependent but are also influenced by individual factors, and studies evaluating the toxicity of GCs should take into account the comorbidities of the patients with RA to whom GCs are prescribed so as to identify the patients for whom the risk/benefit ratio is favourable or not.
# Conclusion
GCs clearly still have an important role to play in RA management. Their clinical (and structural) efficacy is widely acknowledged. Most studies have evaluated GCs efficacy as bridging therapy, combined with csDMARDs. Consequently, available studies have mainly focused on patients with early RA, and data for established RA are scarce. Nevertheless, GCs seem useful in this population to control flares. By contrast, GCs monotherapy does not represent an acceptable therapeutic option. The benefit of adding GCs to bDMARDs and tsDMARDs is most likely low or non-existent because of the fast onset of action of most of these DMARDs. GCs are also cheap, and their combination with csDMARDs could reduce or delay the use of bDMARDs or tsDMARDs.
However, there is also evidence of GCs toxic effects, notably with moderate-dose to high-dose GCs used for a longer time, but even low-dose GCs might have adverse effects. Besides, the long-term GCs safety might be linked to the cumulative dose, and, apart from the daily dose, the duration is also crucial. GCs must be used at the lowest possible dose and for the shortest possible time. With long-term (≥6 months) GCs therapy, EULAR has defined 5 mg/day or less (if needed for controlling the disease) as the acceptable daily intake in terms of cardiovascular risk and risk of hyperglycaemia/diabetes, osteoporosis and infection for the vast majority of patients, but the individual risk of harm must also be evaluated by taking into account patient-specific characteristics. [bib_ref] Defining conditions where long-term glucocorticoid treatment has an acceptably low level of..., Strehl [/bib_ref] The ongoing trial GLORIA, a 2-year pragmatic RCT aiming to assess the safety and efficacy of GCs at 5 mg/day vs PBO added to DMARDs in elderly patients with RA (≥65 years), could provide evidence to support further this threshold of 5 mg/day. [bib_ref] Harm, benefit and costs associated with low-dose glucocorticoids added to the treatment..., Hartman [/bib_ref] In the same prospects of improving the risk/benefit ratio of GCs, studies of chronotherapy and GCs agonists seem promising. Current recommendations do not advocate a specific administration route versus another. Indeed, according to the available literature data, oral and parenteral use of GCs seem to have similar efficacy, but data are scarce and do not concern structural efficacy. However, the parenteral form might facilitate GCs withdrawal by preventing self-medication. Each GCs prescription must be preceded by an evaluation of the benefit/risk balance and by information provided to the patient in the context of a shared decision.
New studies with modern designs evaluating GCs in RA are still needed. Considering the existing literature, we have proposed in this review several leads to guide future research. Of note, future trials will also need to include cost-utility analyses and data on patient participation. [bib_ref] The use of glucocorticoids in early rheumatoid arthritis, Verschueren [/bib_ref] In continuum with these future studies, further guideline updates will need to address specific conditions and circumstances for which GCs should be prescribed in order to improve the balance between efficacy and longterm safety.
[table] Table 1: Characteristics of studies of the clinical efficacy of glucocorticoids published in the last 6 years [/table]
|
The association between obesity and quality of life: a retrospective analysis of a large-scale population-based cohort study
Background: The relationship between obesity and health-related quality of life (HRQoL) may be confounded by factors such as multimorbidity. The aim of the study was to explore this relationship, controlling for long-term conditions and other health, lifestyle and demographic factors in a general adult population. There was specific interest in the impact of high weight status, measured by body mass index (BMI) levels (obesity, morbid obesity) compared with individuals of normal weight. Methods: Health, lifestyle and demographic data were collected from 64,631 individuals aged 16 years and over registered in the Yorkshire Health Study; a long-term cohort study. Data were collected in 2 waves: from patients attending GP surgeries in the South Yorkshire region; and using online recruitment across the entire Yorkshire and Humber area. Univariable and multivariable regression methods were utilised to identify factors associated with HRQoL as measured by the EQ-5D summary score. Long-term conditions were tested as both covariates and mediating factors on the causal pathway between obesity and HRQoL. Results: Increasing levels of obesity are associated with reduced HRQoL, although this difference is negligible between those of normal weight and those who are overweight. Individuals with obesity and morbid obesity score 4.9 and 11.3 percentage points less on the EQ-5D summary scale respectively than those of normal weight. Concurrent physical, and particularly mental health-related long-term conditions are substantively related to HRQoL: those with 3 or more reported mental or physical health conditions score 29.8 and 14.6 percentage points less on the EQ-5D summary scale respectively than those with fewer conditions. Long-term conditions can be conceptualised as lying on the causal path between obesity and HRQoL, but there is weak evidence for a partial mediating relationship only. Conclusions: To conclude, in agreement with the established literature we have found a clear inverse relationship between increasing weight status and decreasing HRQoL and confirmed the mediating role of long-term conditions in the reduction of HRQoL in people with obesity. Nevertheless, a high BMI remains independently related to HRQoL, suggesting that 'healthy people with obesity' may be in transition to an unhealthy future.
# Background
Health-related quality of life (HRQoL) is a broad subjective concept that encompasses both physical and mental health, which are themselves in complex relationships with other external factors such as health, socio-economic status, the environment and other factors. Obesity is a condition of 'abnormal or excessive fat accumulation that may impair health', defined by the WHOas a body mass index (BMI) greater than 30 kg/m 2 , with a BMI of more than 40 kg/m 2 defined as morbid obesity. The aetiology of obesity is complex and multifaceted, stemming from biological, behavioural and environmental causes.
Worldwide obesity has tripled since 1975, and in 2016, 1.9 billion adults (39% of the worldwide adult population) were considered to be overweight: i.e. have a BMI in the range 25 kg/m 2 ≤ BMI < 30 kg/m 2 ; and 650 million (13% of the worldwide population) were considered to have obesity: i.e. have a BMI in the range BMI ≤ 30 kg/ m 2. In England in 2018, 63% of adults were classified as being overweight or having obesity, with 2 and 4% of men and women respectively being defined as having morbid obesity: i.e. have a BMI in the range BMI ≤ 40 kg/m 2. It has been predicted that by 2050 Britain could be a mainly obese society. Connelly reported a noticeable increase in the proportion of the United Kingdom population at very high risk of chronic disease due to their weight. Physical associations include long-term health conditions such as Type 2 diabetes, hypertension, dyslipidaemia, coronary artery disease, stroke, various cancers, reduced reproductive function, osteoarthritis, liver and gall bladder disease, chronic pain and adverse respiratory effects. The proportion of individuals reporting long-term conditions (LTCs) has been shown to increase linearly with increasing BMI, and to be independently related to BMI, after adjusting for age and gender. Similarly, the number of reported LTCs increases with BMI, with 25 and 42% of individuals with moderate and morbid obesity respectively reporting 3 or more LTCs, compared with 12% of normal weight individuals. In addition to physical disease, obesity is also associated with mental health conditions: sleep disorders, anxiety, depression low self-esteem, motivational disorders, eating disorders, impaired body imageand serious psychiatric disorders.
Obesity is associated with physical, mental and economic consequences. The economic consequences of obesity are substantial and increasing. In the UK alone it is estimated that by 2050 the societal and business costs of obesity will reach £49.9billion per year. These costs have been categorised by Seidallas direct costs from treating obesity and its related diseases; societal costs arising from loss of work due to increased absence, physical limitations, lower life expectancy and unemployment benefits; and personal costs stemming e.g. from stigmatisation and discrimination leading to lower incomes and higher healthcare costs. Physical and mental long-term conditions can impact both on each other and Health Related Quality of Life, and the relationship between obesity and HRQoL can be both mediated and confounded by the presence of comorbiditiesand other effects such as medicationand polypharmacy.
The Yorkshire Health Study (YHS) is an observational cohort study of health and lifestyle in Yorkshire and the Humbersupported by NIHR CLAHRC (Collaboration for Leadership in Applied Health Research and Care). Adults (aged 16 and over) residing in the in the Yorkshire and Humber region of England are eligible to enter.
The data, from 70,836 adults, was collected in two waves: the first 27,813 were recruited via GP surgeries in South Yorkshire between 2010 and 2012; the second wave of data collection, from 2013 to 2015 utilised online recruitment and the National Clinical Research Network to recruit 43,023 participants. The majority of participants, whether recruited in Waves 1 or 2, completed one survey only. It is well established that there is an inverse relationship between QoL and obesity. There are many research studies that demonstrate improved quality of life following both dietary and surgical weight loss.
# Methods
The aim of this study was to utilise a large, contemporary cohort from the UK to explore the relationships between obesity and HRQoL, controlling for LTCs and other health, lifestyle and demographic factors in a general adult population; considering specifically the impact of high levels of BMI (obesity and morbid obesity) in comparison to BMI levels corresponding to individuals of normal weight.
Personal (age, gender, academic history, employment status, socio-economic status, quality of life), health (history of diabetes, physical and mental long-term conditions, frequency of visits to health care professionals, frequency of visits to hospital, days off work due to sickness) and lifestyle (smoking status, weekly levels of walking and exercise) data were collected from participants who responded to either Wave 1 and/or the full version of the questionnaire administered in Wave 2 of the YHS.
HRQoL, as measured by the EQ-5D summary index (measured on a scale from 0 to 1, with higher values representing higher QoL, and derived from scores on individual EQ-5D domains of mobility, self-care, activities, pain and anxiety), was considered to be the outcome measure in the current investigation. The key predictor variable was weight status, measured using BMI, categorised for the purposes of the current investigation as Normal weight (18 kg/m 2 ≤ BMI < 25 kg/m 2 ); Overweight (25 kg/m 2 ≤ BMI < 30 kg/m 2 ), Obese (30 kg/m 2 ≤ BMI < 40 kg/m 2 ), and Morbidly obese (BMI ≥ 40 kg/m 2 ). This variable was collected in both waves of the survey. Individuals with BMI less than 18 kg/m 2 were not included in the analysis, as BMIs in this range may be indicative of illness or eating disorder. An investigation into the relationship between QoL and BMI using the first wave only of the YHSrevealed the relationship to be monotonic and approximately linear in individuals with BMI values of 18 kg/m 2 or more: inclusion of underweight individuals' results in a curvilinear effect.
Additionally, a number of variables, also collected in one or both waves of the survey, were collected and examined for potential inclusion as covariates in the analysis. The first mentioned category of the categorical variables above was considered to be the reference category in all cases.
In addition to modelling the LTC variables as covariates in a multiple regression model, these variables were assessed for their effect as mediating variables on the causal pathway between BMI and QoL; in the light of findings by Doll et al.that the proportion of individuals reporting LTCs, and the number of reported LTCs are significantly predicted by BMI in controlled models.
Physical exercise (including activities such as swimming, playing football, cycling and aerobics) and walking time (including walking to work, to shops and leisure walking) in the week preceding data collection were estimated using the mid-point of options presented as ranges of times (none; 0-1 h per week; 2-3 h per week etc.) offered to respondents as response categories.
The data set was checked before analysis for errors. Any values outside of theoretical or plausible ranges were deleted or replaced with a limiting value as appropriate, with limits for inclusion of BMI values obtained using guidelines. The extent and nature of data missingness was investigated. Missing values were assessed for nature of missingness using Little's test for data missing completely at random (MCAR) and separate variance t-tests and cross-tabulations. Data missing at random (MAR) was inferred if the MCAR test was statistically significant but missingness could be predicted from variables other than the outcome variable from separate variance t-tests and cross-tabulations. Following verification of missing data on key variables to be MCAR or MAR, complete case analysis was used with respect to both the key predictor variable (weight status as measured by BMI category) and the outcome measure (EQ-5D score) with no imputation conducted on these variables. Controlling variables with more than 5% missing values on remaining cases were dropped from further analysis. Controlling variables with less than 5% missing values that could be shown or inferred to be MCAR or MAR were imputed using expectation maximisation.
The data were summarised descriptively, by weight status (BMI category) and as a full cohort. A series of
# Results
Valid data were collected on 64,631 individuals. Data checking revealed a small proportion of certain variables with implausible or impossible data values. These were investigated on an individual basis and deleted or amended where necessary.
Calculated BMI values of the cleaned data set ranged from 8.32 to 85.9 kg/m 2 ; with a mean value of 26.7 kg/ m 2 (SD 5.50 kg/m 2 ). The BMI ranges and corresponding frequencies associated with each original and merged category are summarised inA summary of participant characteristics (by weight status) before imputation and variable deletion is summarised in; with data based on respondents from whom a valid weight status could be deduced.
While most differences across groups were statistically significant at the 5% significance level, reflecting the large sample size, few substantive differences across groups were observed. Uni-variable tests of significance revealed low effect sizes (measured by the ϕ and partialη 2 statistics) of less than 5% for most reported variables in the table above. However, some cross-group differences of non-negligible magnitude were observed with respect to gender, diabetes status and academic qualifications. A higher proportion of women than men were in the group with morbid obesity; however, overall mean male BMI (26.9 kg/m 2 ; SD 4.83 kg/m 2 ) was higher than the mean female BMI (26.6 kg/m 2 ; SD 5.84 kg/m 2 ). The proportion of those in the Normal weight group who were qualified to degree level or above was, at 12.1%, more than double that of those in the group with obesity (5.5%) and over 3 times that of those in the group with morbid obesity (3.5%). The proportion of those in the Normal weight group who suffered from 3 or more long-term mental health-related conditions was, at 6.7%, less than half that of those in the group with obesity (15.7%) and less than a third that of those in the group with morbid obesity (24.5%).
Little's test for MCAR using all quantitative variables with complete or near-complete cases revealed no evidence that missing EQ-5D scores were not MCAR (p = 0.408). Separate variance t-tests revealed no evidence that missing weight statuses were not MAR. The variables corresponding to diabetes status, employment status, IMD, exercise levels, alcohol consumption and days off work due to sickness were not carried forward for consideration due to excessive proportions of missing values on these variables.
P-values, parameter estimates, associated confidence intervals, and effect sizes (using the partial-η 2 statistic) from a series of univariable regression analyses conducted the outcome measure of EQ-5D score on an imputed data set including the key predictor variable and all controlling variables with complete or near-complete set of cases as identified inabove, are summarised in.
A mediation analysis revealed that both of the variables modelling mental or physical health-related LTCs exhibited some mediating effect on the relationship between weight status and HRQoL. All paths in the mediation models considering weight status as a predictor, and the mental or physical health-related LTCs in turn as mediators were significant. Path coefficients for weight status were revealed to be − 0.010 in a univariable regression of QoL on weight status; − 0.007 in a model including the variable modelling mental health LTCs and − 0.007 in a model including the variable modelling physical health LTCs. Hence while conditions for partial mediation were met, the conditions were full mediation were not met. The substantive mediating effect was low and weight status continued to significantly predict the outcome in the presence of the mediating variable. Hence analysis proceeded with LTCs being modelled as a controlling covariate.
The simple regression models suggested that age, presence/absence of long-term conditions, level of contact with health professions in last 3 months, number of hours per week spent walking, and number of hospital outpatient visits in previous 3 months should be included alongside a weight status category in a multiple model. As strong evidence for statistical significance was expected in most cases due to the size of the data set, assessments for inclusion were made primarily on the basis of effect sizes, with an associated partial-η 2 statistic of about 0.025 or more considered to indicate grounds for inclusion of a particular variable. As the predictor variable of key contextual interest, this did not apply to any of the weight status categories. Model parameters from this multiple model are summarised in. The R 2 and adjusted-R 2 statistics for this model were both 0.390; representing a moderately good fit to the data. No evidence for collinearity was revealed, with variance inflation factors all within tolerable limits. Analysis of residuals revealed no clear evidence for violations of regression assumptions, with normally distributed standardised residuals which exhibited no clear pattern when plotted against standardised predicted values. The model showed very good cross-validation properties, with negligible loss in correlation computed from the validation sample fitted values against predictions from the training sample model coefficients.
Hence controlling for other categorical factors and covariates, compared to individuals in the Normal weight category; HRQoL was essentially the same in individuals in the Overweight category; slightly lower (4.9 percentage points less on the EQ-5D summary index) in individuals in the Obese category and lower (11.3 percentage points less on the EQ-5D summary index) in individuals in the Morbidly obese category. Hence the effect of morbid obesity, compared to normal weight, has approximately the same impact as 3 or more physical long-term conditions or an increase in age of about 55 years. Amongst the controlling variables, those with the greatest substantive effect on QoL were mental and physical healthrelated LTCs: those with 3 or more mental health conditions scored 29.8 percentage points less on the EQ-5D summary index than those with 2 or fewer conditions; and those with 3 or more physical health conditions scored 14.6 percentage points less on the EQ-5D summary index than those with 2 or fewer conditions. Higher quality of life was also reported by younger people, by those who saw health professionals more infrequently and spent less time visiting hospital as an outpatient, and by those who spent more time walking.
# Discussion
## Key findings
The analysis has revealed a clear relationship indicating lower levels of QoL with weight status defined by categories of increasing BMI in individuals with BMIs in the range of 18 kg/m 2 and above. This monotonic decrease in QoL, recorded in groups categorised by increasing BMI, is consistent with both the findings relating to the individual EQ-5D items in the analysis by Kearns et al.of the first wave of the YHS data, and the wider literature. The effect on QoL of weight status category is substantial, particularly for those in the highest BMI category. This reduction in QoL as a result of increasing BMI is greater than that found linked to cancer, myocardial infarction and diabetes, and similar to having schizophrenia, heart failure or kidney failure (Sullivan 2001). However, the EQ-5D summary index is a highly negatively skewed measure, with about one third of our respondents scoring the maximum value of 1.00 and over half of respondents scoring 0.84 or more.
Comparing the estimates and magnitudes thereof of the weight status variables in the simple and multiple models reveals that the effect of weight status is smaller in the multiple (controlled) model. The variables corresponding to mental and physical healthrelated LTCs in the multiple model appear to be of greater effect on QoL than weight status itself. This may be due to a proportion of the residual variance ascribed to weight status in the simple model being ascribed to other variables in the multiple model; specifically, LTCs, which are already known to be related to weight status from the descriptive analysis and is reflected in the 2007 Sach analysis of BMI and quality of life. It may also reflect the status of obesity as a risk factor for many LTCs. However, there are no changes of direction of association of parameter coefficients or substantial changes in parameter estimates or inferences of significance between the models. Further work considering the impact of specific individual conditions may be beneficial.
The mediation analysis reveals that the presence of mental or physical health-related LTCs has a limited partial mediating effect on the underlying relationship between weight status and QoL. In the current analysis, LTCs are analysed as controlling factors. Nonetheless, LTCs can alternatively be conceptualised as lying on the causal path between BMI and QoL; although the direct link between BMI and QoL is stronger and more intuitive. Further model-testing work is needed to establish the existence of, and direction of associations between other constructs represented in the YHS. The unique contribution of BMI to QoL is consistent with Scottish datawhich found an independent relationship between obesity and Quality of Life. This is in contrast to the 'Healthy Obesity' hypothesis and may represent a subset of the population 'in transition' to unhealthy obesityvia metabolic syndrome, not measured in our study.
The largest unique effect in the multiple model was the presence of 3 or more mental health LTCs. This may be an artefact of the data, explained by a presumed higher likelihood of MH LTCs being related in our sample, compared to the 'independence' of the physical domains of LTC. The second biggest effect is degree of contact with a health professional, which we presume is acting as a proxy measure for general health.
# Strengths and limitations
The strengths of the YHS are its large sample size which allows for an exploration of detailed obesity categories, comprehensive examination of a wide range of variables, and the use of EQ-5D which measures HRQoL using public preferences.
Most measures captured by the YHS are self-reported and may not be completely reliable; particularly those requiring accurate recall, such as activity levels or levels of contact with healthcare professionals over an extended period of time; or the ability of respondents to distinguish between, for example, hospital visits as an out-patient or day case. The key predictor of BMI requires accurate self-reporting of both height and body weight in appropriate units. In addition, self-reported height and weight are respectively over and underestimated in both men and women . In the current study, analysis was restricted to variables which were derived from items elicited in both waves of the questionnaire.
The fit of the multiple regression model to the data, though of moderately high magnitude, may have been constrained in magnitude by uncertainties in the integrity of certain measures and the limited availability of variables for which an acceptable proportion of valid cases were available. Nonetheless, a moderately good fit was obtained and cross-validation procedures revealed that model portability is good; it should be expected that the model will perform equally well on samples other than that from which parameter coefficients were derived.
## Implications for future work
This study has demonstrated that further work is needed to establish the existence of, and direction of associations; for example, it seems plausible that not only can factors such as BMI and exercise impact on quality of life (as was assumed in this analysis), but also that variables such as exercise level and BMI are correlated with a plausible association in either direction. A number of models are required to be tested for model fit using, for example, a confirmatory factor analysis approach in order to ensure that an optimal series of relationships are tested.
# Conclusions
To conclude, in agreement with the established literature we have found a clear inverse relationship between increasing weight status and decreasing QoL, using a large regional cohort study. We have investigated the influence of other demographic, lifestyle and health related domains on this relationship and confirmed the mediating role of LTCs in the reduction of QoL in people with obesity. Nevertheless, a high weight status remains independently related to QoL, suggesting that the 'healthy obese' may be in transition to an unhealthy future. |
Is cancer-related death associated with circadian rhythm?
## Dear editor,
Biological periodicities occur based on a periodicity in a spectrum of frequencies ranging from milliseconds to years. A circadian rhythm is any biological process that displays an endogenous oscillation of about 24 h [bib_ref] Clocks in the clinic: circadian rhythms in health and disease, Kelly [/bib_ref]. Circadian rhythm is the adaption of our bodies' biological processes to the day and night cycle. The biological clock is reset daily to synchronize with the exposure to sunlight because our internal clock repeats at around 24 h [bib_ref] Clocks in the clinic: circadian rhythms in health and disease, Kelly [/bib_ref]. Because of the use of artificial light, our internal clock is influenced by multiple factors, including social interactions [bib_ref] Clocks in the clinic: circadian rhythms in health and disease, Kelly [/bib_ref].
Our bodily functions are maintained by coordination among varying variables, such as hormonal levels, body temperature, cognitive function, and sleep, that closely follow the circadian rhythm [bib_ref] Clocks in the clinic: circadian rhythms in health and disease, Kelly [/bib_ref]. Multiple studies suggested that the temporal pattern of cardiac death is not uniform during the day [bib_ref] Recent advances in circadian rhythms in cardiovascular system, Chen [/bib_ref]. Only a handful of studies have assessed the temporal pattern of death in cancer patients, two of which were conducted in palliative care units [bib_ref] Temporal distribution of deaths in cancer patients admitted to a palliative care..., Neumann [/bib_ref] [bib_ref] Circadian variation in time of death in patients with cancer, Davies [/bib_ref] , and one study included patients who died in other wards [bib_ref] Temporal distribution of deaths in cancer patients during the day in different..., Goncalves [/bib_ref].
All these studies were conducted at single institutes with small sample sizes, and their statistical analyses involved multiple comparisons, which might have jeopardized the validity of the results. If a temporal pattern of cancer death is identified, healthcare service could be optimized to better support terminally ill patients and their family. Families can be counseled appropriately and be psychologically prepared for the final hours of patient's life. Therefore, our goal was to investigate the temporal pattern of death in cancer patients using a large sample size and robust statistical methods to account for chronobiological periodicity.
The methods in the present study are detailed in the Additional file 1: Materials and methods. A total of 353,827 cases of death reported between January 2008 and December 2016 in public hospitals as in-patients in Hong Kong were identified: 58,451 (16.5%) died of cancer, 21,544 (6.1%) died of ischemic heart disease, and 91,999 (26.0%) died of pneumonia as main causes [fig_ref] Table 1: Demographic information for the 171,994 death cases due to cancer, ischemic heart... [/fig_ref]. The observed number of cancer deaths was the most frequent at 6:00-6:59 am, having the prevalence ratio of 1.24 (95% confidence intervals [CI] 1.04-1.47). The highest prevalence ratios for deaths due to ischemic heart disease (1.49, 95% CI 1.20-1.85) and pneumonia (1.35, 95% CI 1.08-1.67) were both observed at 8:00-8:59 pm [fig_ref] Table 2: Prevalence ratios of 171,994 deaths by hour per day, age, and sex... [/fig_ref] , Additional file 2: [fig_ref] Figure 1: Scatter plot of the number of deaths by time in hours [/fig_ref]. [fig_ref] Figure 1: Scatter plot of the number of deaths by time in hours [/fig_ref] shows the observed temporal pattern of cancer deaths in hours [fig_ref] Figure 1: Scatter plot of the number of deaths by time in hours [/fig_ref] and minutes [fig_ref] Figure 1: Scatter plot of the number of deaths by time in hours [/fig_ref] of the day. Even if a small increase in the number of deaths at 6:00-6:59 am and 8:00-8:59 am was observed, there was no evidence of a unimodal sinusoidal circadian rhythm (periodicity) in the time of death according to the parametric sinusoidal circadian test (Z = 2.06, P = 0.127) and the non-parametric multimode test (H0: one unique mode, excess mass = 0.011, P < 0.001). The same analyses were repeated on deaths due to ischemic heart disease and pneumonia. Based on the circadian test, a unimodal circadian rhythm in the former, but not in the latter was observed (Additional file 3: and Additional file 4: .
We did not detect a circadian pattern of cancer death. The present study evaluated the temporal pattern of death among cancer patients using trigonometric functions and with time modeled in a circular scale. This robust method used to model periodicity validated the absence of circadian rhythms. Furthermore, we considered a large sample in hospital settings. Results from the present study were in contrast with those of prior studies on the temporal pattern of death in cancer patients [bib_ref] Temporal distribution of deaths in cancer patients admitted to a palliative care..., Neumann [/bib_ref] [bib_ref] Circadian variation in time of death in patients with cancer, Davies [/bib_ref] [bib_ref] Temporal distribution of deaths in cancer patients during the day in different..., Goncalves [/bib_ref]. Neumann et al. [bib_ref] Temporal distribution of deaths in cancer patients admitted to a palliative care..., Neumann [/bib_ref] revealed a peak of cancer death between 8:00 am and 10:00 am in palliative care wards. In the study of Goncalves et al. [bib_ref] Temporal distribution of deaths in cancer patients during the day in different..., Goncalves [/bib_ref] , peaks of cancer death were observed in a public hospital's palliative care ward between 0:00 am and 2:00 am and between 8:00 am and 10:00 am, with a trough at 4:00-8:00 am. Davies et al. [bib_ref] Circadian variation in time of death in patients with cancer, Davies [/bib_ref] observed a trough in cancer death at 2:00-5:00 am, with no peak. Even if these previous studies on circadian rhythms of cancer death in other geographical areas and periods contradict our findings, the evidence was limited, and the methods used to identify circadian rhythms were not the most appropriate and robust. In contrast, our statistical approach, modeling time in a circular scale and testing parametrically for the presence of a circadian pattern, internally validated our findings. However, more evidence is needed in different populations and geographic areas to externally validate these observations.
## Open access
## Cancer communications
A potential trigger of sudden cardiac events and death in the morning hours is related to the increase in catecholamine levels and blood viscosity when assuming a switch from a supine posture to an upright posture after waking up. However, hospitalized cancer patients in general are cachexic and are sicker than patients with many other types of disease, and the upright posture may be less likely to be a contributor to cancer death. Indeed, we did not identify a morning peak in cancer death, suggesting that cancer death is not mainly driven by the above mechanisms.
The body has the ability to synchronize internal biological processes with daily external events; the circadian timing system is responsible for this ability [bib_ref] Chapter One-The suprachiasmatic nucleus and the circadian timing system, Moore [/bib_ref]. A critical function of the system is the organization and scheduling of behavioral and physiological events [bib_ref] Chapter One-The suprachiasmatic nucleus and the circadian timing system, Moore [/bib_ref]. It influences the responsiveness to stress and challenges in different hours of the day. Disturbance in the rest-activity circadian rhythm is common in cancer patients [bib_ref] Circadian rhythm disruption in cancer biology, Savvidis [/bib_ref]. The etiology of the disturbance is multifactorial; altered physiology related to the disease, symptoms, psychological stress, and cancer treatment all appear to play a role in disrupting circadian regulatory processes [bib_ref] Circadian function in patients with advanced non-small-cell lung cancer, Levin [/bib_ref] [bib_ref] Disruption of sleep, sleep-wake activity rhythm, and nocturnal melatonin production in breast..., Li [/bib_ref]. The frequent disruption of the biological circadian rhythm in cancer patients is possibly associated with the lack of circadian patterns of death in our cohort. Another possible explanation may be the influence of factors related to routine practices in the wards, such as scheduling of medication and different levels of nursing activity and care, including bathing and dressing during a 24-h period. Furthermore, it is possible that incidental cancer deaths are more often registered at 6:00-6:59 am and at 8:00-8:59 pm because these periods coincide with a change between night and day shifts.
The present study has several limitations. The pronouncement of death may have been delayed for minutes or longer after the terminal catastrophic event leading to death and the actual death due to logistic and individual disease factors. The time lag associated with discovering the cessation of vital signs, summoning the doctor, and performing procedures to certify death may vary among different hospital settings. This contributes to the uncertainty in the correlation between the recorded and actual death time. This problem is less prominent in hospitalized death cases than in-home death cases because in a hospital setting, the more frequent availability of staff and more readily available medical procedures shorten delays in the detection of a patient's death.
Additionally, active resuscitation in the hospital may have been performed upon deterioration of the cancer patient's condition, contributing to the patient's prolonged survival. Attempts to resuscitate may occur more often in patients with sudden unexpected deterioration. The proportion of patients who died after attempted resuscitation is unclear. Furthermore, the accuracy of reported causes of death may vary with respect to whether an autopsy was performed and whether comorbidities were present before death. Delay in death reporting could decrease the peaks and increase the troughs of the observed temporal pattern of death, thereby concealing the actual temporal pattern of death.
Therefore, we consider the slightly greater number of cancer deaths at 6:00-6:59 am and at 8:00-8:59 pm to be an incidental finding related to healthcare delivery rather than to a chronobiological rhythm. However, owing to the retrospective population-based nature of the present study, the circumstances around individual patients could not be clarified. Finally, our findings did not rule out the possibility that the presence of external factors could be confounding for the circadian rhythm of death. However, data from a population database preclude detailed analysis of circumstances of individual patients. Future prospective studies should be conducted to further assess whether any external factors can influence the temporal pattern of death.
To conclude, we found no evidence of a chronobiological circadian pattern in death among cancer patients by using robust statistical methods and data from a large population in a hospital setting. Increased understanding of the temporal pattern of deaths may yield important insights toward understanding external factors associated with death.
[fig] *: Correspondence: [email protected] 1 Department of Clinical Oncology, Tuen Mun Hospital, New Territories West Cluster, Hong Kong, P. R. China Full list of author information is available at the end of the article [/fig]
[fig] Figure 1: Scatter plot of the number of deaths by time in hours (a) and in minutes of the day (b), showing the temporal pattern of cancer deaths in acute oncology and cancer palliative wards in Hong Kong, 2008-2016. We found no evidence of a unimodal sinusoidal circadian rhythm (periodicity) in the time of cancer deaths according to the parametric sinusoidal circadian test (Z = 2.06, P = 0.127). Restricted cubic splines using three knots were fitted to model the number of deaths in each hour of the day. The resulting spline fit is graphed as a red line [/fig]
[table] Table 1: Demographic information for the 171,994 death cases due to cancer, ischemic heart disease, and pneumonia in Hong Kong, 2008-2016 [/table]
[table] Table 2: Prevalence ratios of 171,994 deaths by hour per day, age, and sex in Hong Kong, 2008-2016 [/table]
|
A systematic screen for protein–lipid interactions in Saccharomyces cerevisiae
[formula] 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YAR042W Swh1 PH; OBP 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBL007C Sla1 Unclassified 1 G 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 1 G 1 G 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBL060W Yel1 PH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBL085W Boi1 PH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 YBL105C Pkc1 C1; C2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBR102C Exo84 PH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBR115C Lys2 PP-binding 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBR129C Opy1 PH 0 1, E R 1, E R 0 0 0 1, E R 0 0 0 0 0 0 0 0 1 R,E,S 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBR158W Amn1 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBR200W Bem1 PX 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBR213W Met8 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YBR265W Tsc10 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1, S G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YCL034W Lsb5 ANTH/ENTH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YCR009C Rvs161 BAR 1 P 0 0 1 P 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YCR093W Cdc39 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDL052C Slc1 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDL113C Atg20 PX 0 1, G R 1, G R 1, G R 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 R 0 0 YDL142C Crd1 Enzyme 0 0 0 0 0 1 G 0 0 1 G 1 G 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 0 0 0 0 0 1 1 0 0 0 1 0 0 0 0 0 1 1 0 YDL161W Ent1 ANTH/ENTH 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDL205C Hem3 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR103W Ste5 PH 0 1 E 1 E 1 E 0 0 1 E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR105C Tms1 Serinc 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR110W Fob1 Unclassified 1 R 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR121W Dpb4 Unclassified 1 R 1 R 1 R 0 1 R 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 R 1 R 1 R YDR150W Num1 PH 0 1 G 1 G 1 G 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR153C Ent5 ANTH/ENTH 1 R,E,G 1 R,E,G 1 R,E,G 1 R,E,G 1 R 0 1 R,E 0 0 0 0 0 0 0 0 1 R,E,S 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR208W Mss4 Enzyme 0 1 G 1 G 1 G 1 0 0 0 0 0 0 0 0 0 0 1, S G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 YDR228C Pcf11 ANTH/ENTH 1 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR283C Gcn2 Lipid-regulated 1 R,G 0 1 R,E,G 1 R,G 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 R 0 0 YDR289C Rtt103 ANTH/ENTH 1 G 0 1 G 0 1 G 0 1 G 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR294C Dpl1 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR310C Sum1 Unclassified 0 0 0 1 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR311W Tfb1 PH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 G 0 0 YDR313C Pib1 FYVE 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR323C Pep7 FYVE 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YDR358W Gga1 ANTH/ENTH 1 G 1 G 1 G 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YDR388W Rvs167 BAR 0 1 P 1, L P 1 P 1 G 0 0 0 0 0 0 0 0 0 0 1 G,S,L 1 G,S 0 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 1 G 0 0 YDR425W Snx41 PX 1 R 1 R 1 R 1 R 0 0 1, G R 1, G R 0 0 0 0 0 0 0 0 0 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR490C Pkh1 Lipid-regulated 0 1, E R 1, E R 0 1 R 0 1, E R 1, E R 0 0 0 0 0 0 0 1 R,E,G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR505C Psp1 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR517W Grh1 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YDR524C Age1 Unclassified 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YEL002C Wbp1 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YER091C Met6 Unclassified 0 1 G 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 S 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 YER114C Boi2 PH 1, G R 1, G R 1, G R 1, G R 0 0 0 0 0 0 0 0 0 0 0 1 R 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YER155C Bem2 PH 0 0 0 1, G E 0 0 0 1, G E 0 0 0 0 0 0 0 1 E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YFR002W Nic96 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YFR006W Yfr006w Unclassified 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YFR021W Atg18 Lipid-regulated 1 G 1 G 1 G 1 G 1 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YFR053C Hxk1 Lipid-regulated 1 1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 YGL001C Erg26 Enzyme 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 1 G 1 G 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YGL055W Ole1 Cyt-b5 0 0 0 1 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YGL105W Arc1 Lipid-regulated 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YGL131C Snt2 Unclassified 1 E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YGL173C Kem1 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YGL251C Hfm1 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YGR080W Twf1 Lipid-regulated 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 S 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YGR097W Ask10 PH 0 1 1 1 0 0 1 G 0 0 0 0 0 0 0 0 1, S G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YGR100W Mdr1 PH 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YGR170W Psd2 C2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YGR233C Pho81 Unclassified 1 G 1 G 1 G 1 G 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YGR241C Yap1802 ANTH/ENTH 0 1 E 1 E 0 1 0 1 E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 E,G,S 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 YHL002W Hse1 ANTH/ENTH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR001W Osh7 OBP 0 0 1, G R 1, G R 0 0 1 R 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR073W Osh3 PH; OBP 0 1 E 1 E 0 0 0 1, G E 0 0 0 0 0 0 0 0 1 E,G,S 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 YHR105W Ypt35 PX 1 1 1 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR108W Gga2 ANTH/ENTH 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR131C Yhr131c PH 0 1 R 1 R 0 1 R 0 1 R 1 R 0 0 0 0 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR155W Ysp1 PH 0 1 G 1 G 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1, S G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR161C Yap1801 ANTH/ENTH 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR178W Stb5 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR193C Egd2 Lipid-regulated 0 0 0 1, G R 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YHR205W Sch9 C2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YIL007C Nas2 PDZ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YIL043C Cbr1 Unclassified 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YIL063C Yrb2 PH 1 1 1 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YIL105C Slm1 PH 0 1 G,P E,G,P,L 0 1 E,G 0 1 E,G 1 E,G 0 0 0 0 0 0 0 E,G,S,L 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YIL128W Met18 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YIL151C Yil151c Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YJL036W Snx4 PX 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YJL097W Phs1 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YJL145W Yjl145w CRAL/TRIO 0 1 E 1 E 0 1 0 1 E 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 E 0 0 0 0 0 0 0 0 0 1 1 1 YJL157C Far1 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YJL201W Ecm25 Unclassified 1 0 0 1 L 1 L 0 1 1 0 0 0 0 0 0 1 G 0 1 G 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 1 G 0 0 0 0 0 0 YJR092W Bud4 PH 1 1 1 1 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YJR130C Str2 Unclassified [/formula]
## Ykr003w
Osh6 OBP The first and second columns give the systematic ORF and gene names. The third column lists the reported interactions and the name of the interacting lipids, with the estimated Kd when available. For enzymes, a simple representation of the reaction is shown. Finally, the fourth column collects the publication (PubMed ID) where the respective interaction (or enzymatic reaction) was reported.
[formula] 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YKR031C Spo14 PH; PX 0 0 0 0 0 1 G 0 0 1 G 1 G 0 0 0 1 1 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 1 1 0 YKR054C Dyn1 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YKR078W Ykr078w PX 1 R 1 R 1 R 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YKR081C Rpf2 Unclassified 0 0 0 0 0 0 1 E 0 0 0 0 0 0 0 0 0 1 E 1 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 E 0 0 YLL001W Dnm1 Unclassified 0 0 0 1 G 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YLL038C Ent4 ANTH/ENTH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YLR100W Erg27 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1, S R 0 0 0 0 0 0 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 1 R 0 0 YLR178C Tfs1 PBP 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YLR179C Ylr179c PBP 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YLR187W Skg3 PH 0 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YLR189C Atg26 PH 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 S 0 0 0 0 0 1 S 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 1 0 YLR206W Ent2 ANTH/ENTH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YLR260W Lcb5 Enzyme 0 1 G 1 G 0 0 1 1 1 0 0 0 0 0 0 0 1 G 1 G 0 1 G 0 1 G 1 G 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 1 0 YLR298C Yhc1 Unclassified 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YLR335W Nup2 PH 1 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 YLR336C Sgd1 Unclassified 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YLR371W Rom2 PH 1, G E 1, G E 1, G E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YLR380W Csr1 CRAL/TRIO 0 1 E 1 E 0 1 0 1 E 1 E 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YLR425W Tus1 PH 0 0 1 E 0 0 0 1 E 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YML052W Sur7 Lipid-regulated 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 YML054C Cyb2 Cyt-b5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YML072C Tcb3 C2 0 1 R 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YML112W Ctk3 Unclassified 0 0 0 0 0 0 1 G 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YML126C Erg13 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 YMR004W Mvp1 PX 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR015C Erg5 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 YMR073C Ymr073c Cyt-b5 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR079W Sec14 CRAL/TRIO 0 0 0 0 0 0 1 R,E,G 0 0 0 0 0 0 0 0 0 0 0 0 0 1, S R 0 0 0 0 0 0 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 1 R 0 0 YMR083W Adh3 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR104C Ypk2 Lipid-regulated 0 0 1 G 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR108W Ilv2 Lipid-regulated 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR113W Fol3 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR127C Sas2 Unclassified 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR170C Ald2 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR208W Erg12 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 G 0 0 YMR272C Scs7 Cyt-b5 0 0 0 0 1 G 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YMR308C Pse1 Unclassified 1 G 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL047C Slm2 PH 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL087W Tcb2 C2 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL111C Cyb5 Cyt-b5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL127W Far11 Unclassified 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL135C Fpr1 Unclassified 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL144C Ynl144c PH 0 1 R 1 R 1 R 0 0 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL201C Psy2 PH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL231C Pdr16 CRAL/TRIO 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 G 1 G 0 YNL243W Sla2 ANTH/ENTH 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 1, S E 1 S 0 0 0 1 S 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL251C Nrd1 ANTH/ENTH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL264C Pdr17 CRAL/TRIO 0 1 1 0 1 0 1 1 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL278W Caf120 PH 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNL298W Cla4 PH 0 1, G R 1, G R 1, G R 1 R 0 1 R 0 0 0 0 0 0 0 0 1 R 1 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNR006W Vps27 FYVE; ANTH/ENTH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YNR043W Mvd1 Enzyme 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YOL081W Ira2 Unclassified 1 G 0 0 1, L G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 YOL100W Pkh2 Lipid-regulated 0 1 G,P,L 1 G,P,L 0 1 L 0 0 0 0 0 0 0 0 0 0 1 G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YOL113W Skm1 PH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 YOL139CCdc33ORF Gene Name Inp53 (YOR109W) Inp52 (YNL106C) Inp51 (YIL002C) Inp54 (YOL065C) Ymr1 (YJR110W) Sac1 (YKL212W) Fig4 (YNL325C) Stt4 (YLR305C) Lsb6 (YJL100W) Pik1 (YNL267W) Mss4 (YDR208W) Vps34 (YLR240W) Fab1 (YFR019W) Plc1 (YPL268W) Isc1 (YER019W) Spo14 (YKR031C) Spo22 (YIL073C) YDL113C Atg20 C; 2 YBR200W Bem1 YMR004W Mvp1 YJL036W Snx4 YDR425W Snx41 YOR132W Vps17 YKR078W Ykr078w YPR097W Ypr097w YHR105W Ypt35 YPL115C Bem3 YKR031C Spo14 N; 1 YGR097W Ask10 YLR189C Atg26 YER155C Bem2 C; 2 YBL085W Boi1 YER114C Boi2 N; 1 YJR092W Bud4 YNL278W Caf120 YAL041W Cdc24 YNL298W Cla4 N; 1 YOL149W Dcp1 YBR102C Exo84 YOR181W Las17 P; 1 YGR100W Mdr1 YDR150W Num1 N; 1 YLR335W Nup2 YBR129C Opy1 YNL201C [/formula]
## Inositol phosphate metabolism
The matrix consists of 328 described genetic interactions between 172 analyzed proteins (rows) and 96 enzymes involved in lipid metabolism (columns). The first and second columns give the systematic ORF and gene names, respectively. Enzymes are grouped by metabolic pathways. The gene names for lipid metabolic enzymes are described in brackets. Genetic interactions were collected from literature and SGD database. SGD annotation was employed to classify the different genetic interactions types. Note that for simplicity the following groups were defined and represented with a letter code: C, Chemical Genetics; D, Dosage Dependent Interaction (Dosage rescue + Dosage Growth defect + Dosage lethality); P, Positive Suppressive Genetics (Positive genetics + Synthetic rescue + Phenotypic Suppression); N, Negative Synthetic Genetics (Negative genetics + Synthetic lethality + Synthetic growth defect). Numbers refer to the literature source of the published interactions (see legend on top left corner of the matrix). Note that for reference PMID:20093466 the stringent threshold was used PtdIns(4)P 0
[formula] YML054C Cyb2 YNL111C Cyb5 C; 2 YPL170W Dap1 YGL055W Ole1 YMR272C Scs7 P; 3 C; 2 YMR073C Ymr073c C; 2 YDR105C Tms1 YKL192C Acp1 YBR115C Lys2 YGL105W Arc1 YFR021W Atg18 P; 1 N; 1 YOL139C Cdc33 YHR193C Egd2 C; 2 YDR283C Gcn2 C; 2 YFR053C Hxk1 YMR108W Ilv2 YPL004C Lsp1 YDR490C Pkh1 YOL100W Pkh2 C; 2 YML052W Sur7 YGR080W Twf1 YKL126W Ypk1 C; 4 YMR104C Ypk2 C; 2 YPL069C Bts1 YDL142C Crd1 C; 2 YDR294C Dpl1 YPL028W Erg10 N; 3 YMR208W Erg12 N; 3 YML126C Erg13 YGL001C Erg26 YLR100W Erg27 YMR015C Erg5 YPL117C Idi1 YOR171C Lcb4 YLR260W Lcb5 C; 2 YDR208W Mss4 C; D; 1; 2 YNR043W Mvd1 YJL097W Phs1 YDL052C Slc1 YBR265W Tsc10 YKL088W Ykl088w YMR083W Adh3 YDR524C Age1 YMR170C Ald2 YBR158W Amn1 YKL114C Apn1 YIL043C Cbr1 YCR093W Cdc39 YML112W Ctk3 YLL001W Dnm1 C; 2 YDR121W Dpb4 YKR054C Dyn1 N; 1 YJL201W Ecm25 YJL157C Far1 YNL127W Far11 C; 2 YDR110W Fob1 YMR113W Fol3 YNL135C Fpr1 YDR517W Grh1 YDL205C0 1 0 0 0 Sla1 PtdIns(3)P 0 0 1 0 0 0 Sla1 PtdIns(3,4,5)P3 0 0 1 0 0 0 Sla1 DHS-1P 0 0 1 0 0 0 Sla1 Phytosphingosine-1P 0 0 1 0 0 0 Sla1 Phytoceramide C18 0 0 1 0 0 0 Boi1 DHS-1P 0 1 0 0 0 0 Boi1 Digalactosyl-DAG 0 0 0 0 0 0 Opy1 PtdIns(4)P 1 1 0 0 0 0 Opy1 [/formula]
PtdIns(4,5)P2
[formula] 1 1 0 0 0 0 Opy1 Cardiolipin 1 1 0 0 0 0 Opy1 DHS-1P 1 1 0 1 0 0 Bem1 Phytoceramide C18 1 0 0 0 0 0 Tsc10 Phytoceramide C18 0 0 1 1 0 0 Rvs161 PtdIns(3)P 0 0 0 0 1 0 Rvs161 PtdIns(3,4,5)P3 0 0 0 0 1 0 Slc1 Phytoceramide C18 0 0 1 0 0 0 Atg20 [/formula]
PtdIns (4)
## Dpb4
PtdIns ( Cardiolipin (4) PtdIns(4,5)P2
[formula] 0 0 0 0 0 0 Ecm25 PE 0 0 0 0 0 0 Ecm25 Phytosphingosine-1P 0 0 1 0 0 0 Ecm25 Sphingosine 0 0 1 0 0 0 Ecm25 Sphingosine-1P 0 0 1 0 0 0 Ecm25 1,2-DA-3β-galactosylglycerol 0 0 0 0 0 0 Ecm25 Digalactosyl-DAG 0 0 0 0 0 0 Ecm25 FarnesylPP 0 0 1 0 0 0 Bud4 PtdIns(3)P 0 0 0 0 0 0 Bud4 PtdIns(4)P 0 0 0 0 0 0 Bud4 PtdIns(4,5)P2 0 0 0 0 0 0 Bud4 PtdIns(3,4,5)P3 0 0 0 0 0 0 Bud4 Cardiolipin 0 0 0 0 0 0 Bud4 PE 0 0 0 0 0 0 Bud4 DHS-1P 0 0 0 0 0 0 Bud4 Phytosphingosine-1P 0 0 0 0 0 0 Bud4 Phytoceramide C18 0 0 0 0 0 0 Bud4 Sphingosine-1P 0 0 0 0 0 0 Bud4 TAG 0 0 0 0 0 0 Bud4 Dolichol 0 0 0 0 0 0 Str2 PE 0 0 0 0 0 0 Str2 TAG 0 0 0 0 0 0 Ykl088w PtdIns0 0 0 0 0 0 Osh6 PtdIns(3,4,5)P3 0 0 0 0 0 0 Osh6 PtdIns 0 0 0 0 0 0 Osh6 PC 0 0 0 0 0 0 Osh6 Cardiolipin 0 0 0 0 0 0 Osh6 PE 0 0 0 0 0 0 Spo14 PC 0 0 1 0 0 0 Spo14 P-methylethanolamine 0 0 1 0 0 0 Spo14 P-dimethylethanolamine 0 0 1 0 0 0 Spo14 Phytosphingosine 0 0 0 0 0 0 Spo14 Dihydroceramide 0 0 0 0 0 0 Spo14 Ceramide 0 0 0 0 0 0 Spo14 Phytoceramide C8 0 0 0 0 0 0 Spo14 Phytoceramide C18 0 0 0 0 0 0 Spo14 Sphingosine 0 0 0 0 0 0 Spo14 TAG 0 0 0 0 0 0 Spo14 1,2-DA-3β-galactosylglycerol 0 0 0 0 0 0 Spo14 Digalactosyl-DAG 0 0 0 0 0 0 Spo14 Dolichol 0 0 0 0 0 0 Spo14 Ergosterol 0 0 0 0 0 0 Dyn1 [/formula]
Phytoceramide C18
[fig] Ysr3 YKR053C: Lcb3 (YJL134W) Dpl1 (YDR294C) Lcb1 (YMR296C) Lcb2 (YDR062W) Tsc10 (YBR265W) Sur2 (YDR297W) Lip1 (YMR298W) Lag1 (YHL003C) Lac1 (YKL008C) Ydc1 (YPL087W) Ypc1 (YBR183W) Scs7 (YMR272C) Aur1 (YKL004W) Sur1 (YPL057C) Csg2 ( [/fig]
[fig] Fas1 YKL182W: Mct1 (YOR221C) Hfa1 (YMR207C) Cem1 (YER061C) Fas2 (YPL231W) Faa1 (YOR317W) Faa2 (YER015W) Faa3 (YIL009W) Faa4 (YMR246W) Fat1 (YBR041W) Pox1 (Slc1 (YDL052C) Pah1 (YMR165C) Lro1 (YNR008W) Dga1 (YOR245C) Erg10 (YPL028W) Erg13 (YML126C) Hmg1 (YML075C) Hmg2 (YLR450W) Erg12 (YMR208W) Erg8 (YMR220W) Mvd1 (YNR043W) Idi1 (YPL117C) Erg20 (YJL167W) Erg9 (YHR190W) Erg1 (YGR175C) Erg7 (YHR072W) Erg11 (YHR007C) Erg24 (YNL280C) [/fig]
[table] Table S1A: List of lipids. [/table]
[table] Table S1B: List of all selected TAP-fusions. [/table]
[table] Table S2A: Lipid-array interactions: raw dataThis table summarizes protein-lipid interactions detected with lipid-arrays. It includes both the 26 replicates performed and the interactions confirmed with recombinant proteins. The first and second columns give the systematic ORF and gene names, respectively. The first row lists the 56 lipids analyzed. Replicates are shown consecutively in the table and TAP-fusions expressed in E. coli are indicated with an *. Non-physiological lipids are labeled with an #. [/table]
[table] Table S2B: Consolidated lipid-array dataset.The matrix consists of 530 protein-lipid interactions. First and second columns give the systematic ORF and gene names. The third column indicates LBDs and protein classification. The first row lists the 56 lipids analyzed. Those interactions that did not appear in two duplicates were discarded from the consolidated dataset. The interactions for those lipids that were excluded for further analysis are shown separately. Interactions reported for Slm1-PH expressed in E. coli are also shown. Non-physiological lipids are labeled with an #. [/table]
[table] Table S2C: Integrated lipid-array dataset. [/table]
[table] Table S3A: Matrix summarizing the literature-derived reference set. The matrix consists of 71 biochemically characterized protein-lipid interactions. Among them 40 are regulatory, the other 31 take place between lipid-enzymes or transporters and their substrates or products. The first and second columns give the systematic ORF and gene names. The first row lists the 44 lipids analyzed. Non-physiological lipids are labeled with an #. [/table]
[table] Table S3B: The literature-derived reference set: literature curated list of interactions biochemically characterized. [/table]
[table] Table S3C: Summary of the genetic reference set. [/table]
|
An Efficient and Recyclable Nanoparticle-Supported Cobalt Catalyst for Quinoxaline Synthesis
The syntheses of quinoxalines derived from 1,2-diamine and 1,2-dicarbonyl compounds under mild reaction conditions was carried out using a nanoparticle-supported cobalt catalyst. The supported nanocatalyst exhibited excellent activity and stability and it could be reused for at least ten times without any loss of activity. No cobalt contamination could be detected in the products by AAS measurements, pointing to the excellent activity and stability of the Co nanomaterial.Abstract:The syntheses of quinoxalines derived from 1,2-diamine and 1,2-dicarbonyl compounds under mild reaction conditions was carried out using a nanoparticle-supported cobalt catalyst. The supported nanocatalyst exhibited excellent activity and stability and it could be reused for at least ten times without any loss of activity. No cobalt contamination could be detected in the products by AAS measurements, pointing to the excellent activity and stability of the Co nanomaterial.
# Introduction
Quinoxaline derivatives are attractive N-containing heterocycles and these scaffolds have attracted much attention, not only in synthetic chemistry [bib_ref] Quinoxalines: Supplement II, Brown [/bib_ref] but also in the medicinal field [bib_ref] 3-b]quinoxalines: DNA and protein interacting scaffold for pharmacological activities, Moorthy [/bib_ref] [bib_ref] Structure-activity relationship (SAR) studies of quinoxalines as novel HCV NS5B RNA-dependent RNA..., Rong [/bib_ref] [bib_ref] Antibacterial activity of quinoxalines, quinazolines, and 1,5-naphthyridines, Parhi [/bib_ref] [bib_ref] Fragment based design of new H 4 receptor-ligands with anti-inflammatory properties in..., Smits [/bib_ref] [bib_ref] Design, synthesis, and biological evaluation of analogues of the antitumor agent, 2-{4-[(7-chloro-2-quinoxalinyl)oxy]phenoxy}propionic..., Hazeldine [/bib_ref] [bib_ref] Perturbing pro-survival proteins using quinoxaline derivatives: A structure-activity relationship study, Rajule [/bib_ref] [bib_ref] Synthesis and antiprotozoal activity of some new synthetic substituted quinoxalines, Hui [/bib_ref] [bib_ref] Synthesis and biological activity of new quinoxaline antibiotics of echinomycin analogues, Kim [/bib_ref]. These compounds exhibit diverse biological activities, such as antiviral [bib_ref] 3-b]quinoxalines: DNA and protein interacting scaffold for pharmacological activities, Moorthy [/bib_ref] [bib_ref] Structure-activity relationship (SAR) studies of quinoxalines as novel HCV NS5B RNA-dependent RNA..., Rong [/bib_ref] , antibacterial [bib_ref] Antibacterial activity of quinoxalines, quinazolines, and 1,5-naphthyridines, Parhi [/bib_ref] , anti-inflammatory [bib_ref] Fragment based design of new H 4 receptor-ligands with anti-inflammatory properties in..., Smits [/bib_ref] , antitumoral [bib_ref] Design, synthesis, and biological evaluation of analogues of the antitumor agent, 2-{4-[(7-chloro-2-quinoxalinyl)oxy]phenoxy}propionic..., Hazeldine [/bib_ref] [bib_ref] Perturbing pro-survival proteins using quinoxaline derivatives: A structure-activity relationship study, Rajule [/bib_ref] and anti-HIV properties [bib_ref] Synthesis and antiprotozoal activity of some new synthetic substituted quinoxalines, Hui [/bib_ref] [bib_ref] Synthesis and biological activity of new quinoxaline antibiotics of echinomycin analogues, Kim [/bib_ref]. Examples of quinoxalinecontaining pharmacological entities are shown in [fig_ref] Figure 1: Biologically important quinoxalines [/fig_ref]. In addition, quinoxalines have been applied as building blocks for the development of macrocyclic molecular receptors [bib_ref] Quinoxaline-Bridged Porphyrinoids, Sessler [/bib_ref] [bib_ref] ]arenecavitand-based molecular switches, Azov [/bib_ref] , semiconducting materials [bib_ref] Electronic properties and field-effect transistors of thiophene-based donor-acceptor conjugated copolymers, Champion [/bib_ref] [bib_ref] An easily synthesized blue polymer for high-performance polymer solar cells, Wang [/bib_ref] [bib_ref] Novel quinoxaline-based organic sensitizers for dye-sensitized solar cells, Chang [/bib_ref] [bib_ref] Dithienopyrrole-quinoxaline/pyridopyrazinedonor-acceptorpolymers: Synthesis and electrochemical, optical, charge-transport, and photovoltaic properties, Zhang [/bib_ref] [bib_ref] Molecular orbital energy level modulation through incorporation of selenium and fluorine into..., Zhuang [/bib_ref] [bib_ref] Synthesis and photovoltaic properties of a D-A copolymer based on the 2,3-di(5-hexylthiophen-2-yl)quinoxaline..., Wang [/bib_ref] [bib_ref] Synthesis of two conjugated polymers as TNT chemosensor materials, Chen [/bib_ref] , dyes, cavitands [bib_ref] N-fused quinoxalines and benzoquinoxalines as attractive emitters for organic light emitting diodes, Lindner [/bib_ref] and luminescent materials [bib_ref] Chromophore-labeled quinoxaline derivatives as efficient electroluminescent materials, Thomas [/bib_ref].
# Introduction
Quinoxaline derivatives are attractive N-containing heterocycles and these scaffolds have attracted much attention, not only in synthetic chemistry [bib_ref] Quinoxalines: Supplement II, Brown [/bib_ref] but also in the medicinal field [bib_ref] 3-b]quinoxalines: DNA and protein interacting scaffold for pharmacological activities, Moorthy [/bib_ref] [bib_ref] Structure-activity relationship (SAR) studies of quinoxalines as novel HCV NS5B RNA-dependent RNA..., Rong [/bib_ref] [bib_ref] Antibacterial activity of quinoxalines, quinazolines, and 1,5-naphthyridines, Parhi [/bib_ref] [bib_ref] Fragment based design of new H 4 receptor-ligands with anti-inflammatory properties in..., Smits [/bib_ref] [bib_ref] Design, synthesis, and biological evaluation of analogues of the antitumor agent, 2-{4-[(7-chloro-2-quinoxalinyl)oxy]phenoxy}propionic..., Hazeldine [/bib_ref] [bib_ref] Perturbing pro-survival proteins using quinoxaline derivatives: A structure-activity relationship study, Rajule [/bib_ref] [bib_ref] Synthesis and antiprotozoal activity of some new synthetic substituted quinoxalines, Hui [/bib_ref] [bib_ref] Synthesis and biological activity of new quinoxaline antibiotics of echinomycin analogues, Kim [/bib_ref]. These compounds exhibit diverse biological activities, such as antiviral [bib_ref] 3-b]quinoxalines: DNA and protein interacting scaffold for pharmacological activities, Moorthy [/bib_ref] [bib_ref] Structure-activity relationship (SAR) studies of quinoxalines as novel HCV NS5B RNA-dependent RNA..., Rong [/bib_ref] , antibacterial [bib_ref] Antibacterial activity of quinoxalines, quinazolines, and 1,5-naphthyridines, Parhi [/bib_ref] , anti-inflammatory [bib_ref] Fragment based design of new H 4 receptor-ligands with anti-inflammatory properties in..., Smits [/bib_ref] , antitumoral [bib_ref] Design, synthesis, and biological evaluation of analogues of the antitumor agent, 2-{4-[(7-chloro-2-quinoxalinyl)oxy]phenoxy}propionic..., Hazeldine [/bib_ref] [bib_ref] Perturbing pro-survival proteins using quinoxaline derivatives: A structure-activity relationship study, Rajule [/bib_ref] and anti-HIV properties [bib_ref] Synthesis and antiprotozoal activity of some new synthetic substituted quinoxalines, Hui [/bib_ref] [bib_ref] Synthesis and biological activity of new quinoxaline antibiotics of echinomycin analogues, Kim [/bib_ref]. Examples of quinoxalinecontaining pharmacological entities are shown in [fig_ref] Figure 1: Biologically important quinoxalines [/fig_ref]. In addition, quinoxalines have been applied as building blocks for the development of macrocyclic molecular receptors [bib_ref] Quinoxaline-Bridged Porphyrinoids, Sessler [/bib_ref] [bib_ref] ]arenecavitand-based molecular switches, Azov [/bib_ref] , semiconducting materials [bib_ref] Electronic properties and field-effect transistors of thiophene-based donor-acceptor conjugated copolymers, Champion [/bib_ref] [bib_ref] An easily synthesized blue polymer for high-performance polymer solar cells, Wang [/bib_ref] [bib_ref] Novel quinoxaline-based organic sensitizers for dye-sensitized solar cells, Chang [/bib_ref] [bib_ref] Dithienopyrrole-quinoxaline/pyridopyrazinedonor-acceptorpolymers: Synthesis and electrochemical, optical, charge-transport, and photovoltaic properties, Zhang [/bib_ref] [bib_ref] Molecular orbital energy level modulation through incorporation of selenium and fluorine into..., Zhuang [/bib_ref] [bib_ref] Synthesis and photovoltaic properties of a D-A copolymer based on the 2,3-di(5-hexylthiophen-2-yl)quinoxaline..., Wang [/bib_ref] [bib_ref] Synthesis of two conjugated polymers as TNT chemosensor materials, Chen [/bib_ref] , dyes, cavitands [bib_ref] N-fused quinoxalines and benzoquinoxalines as attractive emitters for organic light emitting diodes, Lindner [/bib_ref] and luminescent materials [bib_ref] Chromophore-labeled quinoxaline derivatives as efficient electroluminescent materials, Thomas [/bib_ref]. Generally, quinoxalines can be prepared via a double condensation of 1,2-phenylenediamines with 1,2-diketones [bib_ref] Novel succinct routes to quinoxalines and 2-benzimidazolylquinoxalines via the Ugi reaction, Ayaz [/bib_ref] [bib_ref] Efficient, convenient and reusable polyaniline-sulfate salt catalyst for the synthesis of quinoxaline..., Srinivas [/bib_ref] [bib_ref] Cerium(IV) ammonium nitrate(CAN) as a catalyst in tap water: A simple, proficient..., More [/bib_ref] [bib_ref] An efficient protocol for the synthesis of quinoxaline derivatives at room temperature..., Bhosale [/bib_ref] [bib_ref] General microwave-assisted protocols for the expedient synthesis of quinoxalines and heterocyclic pyrazines, Zhao [/bib_ref]. A number of reagents have been shown to catalyze these reactions such as acidic alumina [bib_ref] Easy access to quinoxaline derivatives using alumina as an effective and reusable..., Jafarpour [/bib_ref] , citric acid [bib_ref] Citric acid: An efficient and green catalyst for rapid one pot synthesis..., Mahesh [/bib_ref] , magnetic Fe3O4 nanoparticles in H2O [bib_ref] Magnetic Fe 3 O 4 nanoparticles as new, efficient, and reusable catalysts..., Lü [/bib_ref] , silica-bonded sulfonic Generally, quinoxalines can be prepared via a double condensation of 1,2-phenylenediamines with 1,2-diketones [bib_ref] Novel succinct routes to quinoxalines and 2-benzimidazolylquinoxalines via the Ugi reaction, Ayaz [/bib_ref] [bib_ref] Efficient, convenient and reusable polyaniline-sulfate salt catalyst for the synthesis of quinoxaline..., Srinivas [/bib_ref] [bib_ref] Cerium(IV) ammonium nitrate(CAN) as a catalyst in tap water: A simple, proficient..., More [/bib_ref] [bib_ref] An efficient protocol for the synthesis of quinoxaline derivatives at room temperature..., Bhosale [/bib_ref] [bib_ref] General microwave-assisted protocols for the expedient synthesis of quinoxalines and heterocyclic pyrazines, Zhao [/bib_ref]. A number of reagents have been shown to catalyze these reactions such as acidic alumina [bib_ref] Easy access to quinoxaline derivatives using alumina as an effective and reusable..., Jafarpour [/bib_ref] , citric acid [bib_ref] Citric acid: An efficient and green catalyst for rapid one pot synthesis..., Mahesh [/bib_ref] , magnetic Fe 3 O 4 nanoparticles in H 2 O [bib_ref] Magnetic Fe 3 O 4 nanoparticles as new, efficient, and reusable catalysts..., Lü [/bib_ref] , silica-bonded sulfonic acid [bib_ref] Silica bonded S-sulfonic acid: A recyclable catalyst for the synthesis of quinoxalines..., Niknam [/bib_ref] , among others [bib_ref] Gallium(III) triflate-catalyzed synthesis of quinoxaline derivatives, Cai [/bib_ref] [bib_ref] III)-catalyzed rapid synthesis of 2,3-disubstituted quinoxalines in water, Yadav [/bib_ref]. Other protocols to synthesize quinoxalines mainly involve the oxidative trapping of vicinal diols or α-hydroxy ketones with 1,2-diamines [bib_ref] The Synthesis of benzimidazoles and quinoxalines from aromatic diamines and alcohols by..., Hille [/bib_ref] [bib_ref] A new ruthenium-catalyzed approach for quinoxalines from o-phenylenediamines and vicinal-diols, Cho [/bib_ref] [bib_ref] An environmentally friendly, cost effective synthesis of quinoxalines: The influence of microwave..., Jeena [/bib_ref] [bib_ref] Manganese octahedral molecular sieves catalyzed tandem process for synthesis of quinoxalines, Sithambaram [/bib_ref] [bib_ref] Quinoxaline synthesis from α-hydroxy ketones via a tandem oxidation process using catalysed..., Robinson [/bib_ref] [bib_ref] Manganese(IV) dioxide-catalyzed synthesis of quinoxalines under microwave irradiation, Kim [/bib_ref] [bib_ref] Tandem oxidation processes for the preparation of nitrogen-containing heteroaromatic and heterocyclic compounds, Raw [/bib_ref] [bib_ref] Straightforward access to pyrazines, piperazinones, and quinoxalinesby reactions of 1,2-diaza-1,3-butadienes with 1,2-diamines..., Aparicio [/bib_ref] , 1,4-addition of 1,2-diamines to diazenylbutenes [bib_ref] α-Nitro epoxides in organic synthesis: development of a one-pot organocatalytic strategy for..., Ibrahim [/bib_ref] , coupling of epoxides with ene-1,2-diamines [bib_ref] Direct and catalytic synthesis of quinoxaline derivatives from epoxides and ene-1,2-diamines, Antoniotti [/bib_ref] [bib_ref] A simple and straightforward approach to quinoxalines by iron/sulfur-catalyzed redox condensation of..., Nguyen [/bib_ref] , 2-nitroanilines with phenethylamines [bib_ref] Ruthenium-catalyzed oxidation of alkynes to 1,2-diketones under room temperature and one-pot synthesis..., Xu [/bib_ref] , alkynes or ketones with 1,2-diamines via a key oxidation process [bib_ref] Gold-catalyzed synthesis of glyoxals by oxidation of terminal alkynes: one-pot synthesis of..., Shi [/bib_ref] [bib_ref] Hypervalent iodine(III)-induced oxidative [4+2] annulation of o-phenylenediamines and electron-deficient alkynes: Direct synthesis..., Okumura [/bib_ref] [bib_ref] Efficient synthesis of quinoxalines with hypervalent iodine as a catalyst, Chen [/bib_ref] [bib_ref] Et 3 N-catalyzed oxidative dehydrogenative coupling of α-unsubstituted aldehydes and ketones with..., Zhang [/bib_ref] [bib_ref] Sustainable preparation of supported metal nanoparticles and their applications in catalysis, Campelo [/bib_ref]. Therefore, the development of efficient methods for accessing quinoxalines derivatives continues to be an active area of research.
Nanoparticle-supported catalysts can offer important advantages as compared to homogeneous transition metal systems and colloidal nanoparticles. These include a good reusability coupled with high activities and specificities in different chemistries based on their excelling properties (high surface areas, degenerated density of energy states and plasmon) [bib_ref] Supported metal nanoparticles on porous materials. Methods and applications, White [/bib_ref] [bib_ref] Nanoparticles as recyclable catalysts: The frontier between homogeneous and heterogeneous catalysis, Astruc [/bib_ref] [bib_ref] An efficient synthesis of coumarin derivatives using a SBA-15 supported Cobalt(II) nanocatalyst, Rajabi [/bib_ref]. In this regard, Co/supported catalysts were previously reported to be highly active and versatile for acid and redox catalyzed processes [bib_ref] An efficient synthesis of coumarin derivatives using a SBA-15 supported Cobalt(II) nanocatalyst, Rajabi [/bib_ref] [bib_ref] Supported cobalt oxide nanoparticles as efficient catalyst in esterification and amidation reactions, Rajabi [/bib_ref].
To the best of our knowledge, there is no protocol describing the preparation of quinoxaline derivatives using a nanoparticle-supported cobalt catalyst. In view of the explained above, we decided to examine the synthesis of substituted quinoxalines by reaction of 1,2-diketones with 1,2-phenylenediamines using a nanoparticle-supported cobalt catalyst (Scheme 1).
Molecules 2015, 20, page-page 2 acid [bib_ref] Silica bonded S-sulfonic acid: A recyclable catalyst for the synthesis of quinoxalines..., Niknam [/bib_ref] , among others [bib_ref] Gallium(III) triflate-catalyzed synthesis of quinoxaline derivatives, Cai [/bib_ref] [bib_ref] III)-catalyzed rapid synthesis of 2,3-disubstituted quinoxalines in water, Yadav [/bib_ref]. Other protocols to synthesize quinoxalines mainly involve the oxidative trapping of vicinal diols or α-hydroxy ketones with 1,2-diamines [bib_ref] The Synthesis of benzimidazoles and quinoxalines from aromatic diamines and alcohols by..., Hille [/bib_ref] [bib_ref] A new ruthenium-catalyzed approach for quinoxalines from o-phenylenediamines and vicinal-diols, Cho [/bib_ref] [bib_ref] An environmentally friendly, cost effective synthesis of quinoxalines: The influence of microwave..., Jeena [/bib_ref] [bib_ref] Manganese octahedral molecular sieves catalyzed tandem process for synthesis of quinoxalines, Sithambaram [/bib_ref] [bib_ref] Quinoxaline synthesis from α-hydroxy ketones via a tandem oxidation process using catalysed..., Robinson [/bib_ref] [bib_ref] Manganese(IV) dioxide-catalyzed synthesis of quinoxalines under microwave irradiation, Kim [/bib_ref] [bib_ref] Tandem oxidation processes for the preparation of nitrogen-containing heteroaromatic and heterocyclic compounds, Raw [/bib_ref] [bib_ref] Straightforward access to pyrazines, piperazinones, and quinoxalinesby reactions of 1,2-diaza-1,3-butadienes with 1,2-diamines..., Aparicio [/bib_ref] , 1,4-addition of 1,2-diamines to diazenylbutenes [bib_ref] α-Nitro epoxides in organic synthesis: development of a one-pot organocatalytic strategy for..., Ibrahim [/bib_ref] , coupling of epoxides with ene-1,2-diamines [bib_ref] Direct and catalytic synthesis of quinoxaline derivatives from epoxides and ene-1,2-diamines, Antoniotti [/bib_ref] [bib_ref] A simple and straightforward approach to quinoxalines by iron/sulfur-catalyzed redox condensation of..., Nguyen [/bib_ref] , 2-nitroanilines with phenethylamines [bib_ref] Ruthenium-catalyzed oxidation of alkynes to 1,2-diketones under room temperature and one-pot synthesis..., Xu [/bib_ref] , alkynes or ketones with 1,2-diamines via a key oxidation process [bib_ref] Gold-catalyzed synthesis of glyoxals by oxidation of terminal alkynes: one-pot synthesis of..., Shi [/bib_ref] [bib_ref] Hypervalent iodine(III)-induced oxidative [4+2] annulation of o-phenylenediamines and electron-deficient alkynes: Direct synthesis..., Okumura [/bib_ref] [bib_ref] Efficient synthesis of quinoxalines with hypervalent iodine as a catalyst, Chen [/bib_ref] [bib_ref] Et 3 N-catalyzed oxidative dehydrogenative coupling of α-unsubstituted aldehydes and ketones with..., Zhang [/bib_ref] [bib_ref] Sustainable preparation of supported metal nanoparticles and their applications in catalysis, Campelo [/bib_ref]. Therefore, the development of efficient methods for accessing quinoxalines derivatives continues to be an active area of research.
Nanoparticle-supported catalysts can offer important advantages as compared to homogeneous transition metal systems and colloidal nanoparticles. These include a good reusability coupled with high activities and specificities in different chemistries based on their excelling properties (high surface areas, degenerated density of energy states and plasmon) [bib_ref] Supported metal nanoparticles on porous materials. Methods and applications, White [/bib_ref] [bib_ref] Nanoparticles as recyclable catalysts: The frontier between homogeneous and heterogeneous catalysis, Astruc [/bib_ref] [bib_ref] An efficient synthesis of coumarin derivatives using a SBA-15 supported Cobalt(II) nanocatalyst, Rajabi [/bib_ref]. In this regard, Co/supported catalysts were previously reported to be highly active and versatile for acid and redox catalyzed processes [bib_ref] An efficient synthesis of coumarin derivatives using a SBA-15 supported Cobalt(II) nanocatalyst, Rajabi [/bib_ref] [bib_ref] Supported cobalt oxide nanoparticles as efficient catalyst in esterification and amidation reactions, Rajabi [/bib_ref].
To the best of our knowledge, there is no protocol describing the preparation of quinoxaline derivatives using a nanoparticle-supported cobalt catalyst. In view of the explained above, we decided to examine the synthesis of substituted quinoxalines by reaction of1,2-diketones with 1,2-phenylenediamines using a nanoparticle-supported cobalt catalyst (Scheme 1). Scheme 1.General scheme of the reactions.
# Results and discussion
Initially, we chose 1,2-diphenylethanedione (1a)and 1,2-diamino-4-nitrobenzene (2a) as model substrates to establish the best conditions for this reaction and some experiments were performed to synthesize the corresponding quinoxaline 3a [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref]. We started our studies reacting 1,2-diketone 1a (1.0 mmol) with 1,2-phenylenediamine 2a (1.0 mmol) at 100 °C for 2 h, without catalyst and solvent. Under these conditions, product 3a was not obtained [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , entry 1). Good results were obtained however when the reactions of substrates 1a and 2a were carried out using H2O as solvent in the presence of Co NPs (2 mol %) as catalyst. Reactions performed at 100 °C and 50 °C gave the desired product in 87% and 57% yield, respectively [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. A similar result was obtained when the reaction was conducted at 100 °C, however using 1 mol% of Co NPs (86% yield) [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , entry 4). Good results were also found when the reactions were performed using EtOH as solvent [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , entry 5-9). Excellent yields of product 3a were achieved in reactions carried out in EtOH at 78 °C using 1 mol% of catalyst [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref]. When the amount of catalyst was reduced to 0.5 mol %, a decrease in the yield of product 3a was observed [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref]. Finally, the reaction performed using 1 mol % of Co NPs at 100 °C and in absence of EtOH yielded the quinoxaline 3a in 72% yield [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref].
Analyzing the results shown in [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , we established the best reaction conditions reacting 1,2-diphenylethanedione (1a, 1.0 mmol, 0.033 g) with 1,2-diamino-4-nitrobenzene (2a, 1.0 mmol) using supported CoNPs (1 mol %) as catalyst and EtOH (5 mL) as solvent. After that, the mixture was stirred at reflux for 90 min in open atmosphere, affording 6-nitro-2,3-diphenylquinoxaline(3a) in 92% yield after crystallization.
# Results and discussion
Initially, we chose 1,2-diphenylethanedione (1a)and 1,2-diamino-4-nitrobenzene (2a) as model substrates to establish the best conditions for this reaction and some experiments were performed to synthesize the corresponding quinoxaline 3a [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref]. We started our studies reacting 1,2-diketone 1a (1.0 mmol) with 1,2-phenylenediamine 2a (1.0 mmol) at 100˝C for 2 h, without catalyst and solvent. Under these conditions, product 3a was not obtained [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , entry 1). Good results were obtained however when the reactions of substrates 1a and 2a were carried out using H 2 O as solvent in the presence of Co NPs (2 mol %) as catalyst. Reactions performed at 100˝C and 50˝C gave the desired product in 87% and 57% yield, respectively (Table 1, entries 2 and 3). A similar result was obtained when the reaction was conducted at 100˝C, however using 1 mol% of Co NPs (86% yield) [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , entry 4). Good results were also found when the reactions were performed using EtOH as solvent [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , entry 5-9). Excellent yields of product 3a were achieved in reactions carried out in EtOH at 78˝C using 1 mol% of catalyst (Table 1, entries 7-8). When the amount of catalyst was reduced to 0.5 mol %, a decrease in the yield of product 3a was observed (Table 1, entry 9). Finally, the reaction performed using 1 mol % of Co NPs at 100˝C and in absence of EtOH yielded the quinoxaline 3a in 72% yield [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , entry 10).
Analyzing the results shown in [fig_ref] Table 1: Optimization of reaction condition a [/fig_ref] , we established the best reaction conditions reacting 1,2-diphenylethanedione (1a, 1.0 mmol, 0.033 g) with 1,2-diamino-4-nitrobenzene (2a, 1.0 mmol) using supported CoNPs (1 mol %) as catalyst and EtOH (5 mL) as solvent. After that, the mixture was stirred at reflux for 90 min in open atmosphere, affording 6-nitro-2,3-diphenylquinoxaline(3a) in 92% yield after crystallization. In order to extend the scope of the reaction, the best conditions were employed in reactions of 1,2-diamino-4-nitrobenzene (2a) with other 1,2-diketones 1b-e with different patterns of substitution and the results are summarized in [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. As it can be seen on [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] (Entries 1-5), our methodology is suitable to a range of substituted 1,2-diketones containing electron-withdrawing groups, affording excellent yields to desired products in all examples. In addition, the possibility of performing the reaction of 1,2-diketones 1a-e with o-phenylenediamine(2b) was also investigated [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] , entries 6-10). Using these substrates, a range of substituted quinoxalines was obtained in excellent yields using the nanoparticle-supported cobalt catalyst under optimized reaction conditions. In order to extend the scope of the reaction, the best conditions were employed in reactions of 1,2-diamino-4-nitrobenzene (2a) with other 1,2-diketones 1b-e with different patterns of substitution and the results are summarized in [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. As it can be seen on [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] (Entries 1-5), our methodology is suitable to a range of substituted 1,2-diketones containing electron-withdrawing groups, affording excellent yields to desired products in all examples. In addition, the possibility of performing the reaction of 1,2-diketones 1a-e with o-phenylenediamine (2b) was also investigated [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] , entries 6-10). Using these substrates, a range of substituted quinoxalines was obtained in excellent yields using the nanoparticle-supported cobalt catalyst under optimized reaction conditions. In order to extend the scope of the reaction, the best conditions were employed in reactions of 1,2-diamino-4-nitrobenzene (2a) with other 1,2-diketones 1b-e with different patterns of substitution and the results are summarized in [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. As it can be seen on [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] (Entries 1-5), our methodology is suitable to a range of substituted 1,2-diketones containing electron-withdrawing groups, affording excellent yields to desired products in all examples. In addition, the possibility of performing the reaction of 1,2-diketones 1a-e with o-phenylenediamine(2b) was also investigated [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] , entries 6-10). Using these substrates, a range of substituted quinoxalines was obtained in excellent yields using the nanoparticle-supported cobalt catalyst under optimized reaction conditions. In order to extend the scope of the reaction, the best conditions were employed in reactions of 1,2-diamino-4-nitrobenzene (2a) with other 1,2-diketones 1b-e with different patterns of substitution and the results are summarized in [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. As it can be seen on [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] (Entries 1-5), our methodology is suitable to a range of substituted 1,2-diketones containing electron-withdrawing groups, affording excellent yields to desired products in all examples. In addition, the possibility of performing the reaction of 1,2-diketones 1a-e with o-phenylenediamine(2b) was also investigated [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] , entries 6-10). Using these substrates, a range of substituted quinoxalines was obtained in excellent yields using the nanoparticle-supported cobalt catalyst under optimized reaction conditions. In order to extend the scope of the reaction, the best conditions were employed in reactions of 1,2-diamino-4-nitrobenzene (2a) with other 1,2-diketones 1b-e with different patterns of substitution and the results are summarized in [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. As it can be seen on [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] (Entries 1-5), our methodology is suitable to a range of substituted 1,2-diketones containing electron-withdrawing groups, affording excellent yields to desired products in all examples. In addition, the possibility of performing the reaction of 1,2-diketones 1a-e with o-phenylenediamine(2b) was also investigated [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] , entries 6-10). Using these substrates, a range of substituted quinoxalines was obtained in excellent yields using the nanoparticle-supported cobalt catalyst under optimized reaction conditions. In order to extend the scope of the reaction, the best conditions were employed in reactions of 1,2-diamino-4-nitrobenzene (2a) with other 1,2-diketones 1b-e with different patterns of substitution and the results are summarized in [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. As it can be seen on [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] (Entries 1-5), our methodology is suitable to a range of substituted 1,2-diketones containing electron-withdrawing groups, affording excellent yields to desired products in all examples. In addition, the possibility of performing the reaction of 1,2-diketones 1a-e with o-phenylenediamine(2b) was also investigated [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] , entries 6-10). Using these substrates, a range of substituted quinoxalines was obtained in excellent yields using the nanoparticle-supported cobalt catalyst under optimized reaction conditions. In order to extend the scope of the reaction, the best conditions were employed in reactions of 1,2-diamino-4-nitrobenzene (2a) with other 1,2-diketones 1b-e with different patterns of substitution and the results are summarized in [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. As it can be seen on [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] (Entries 1-5), our methodology is suitable to a range of substituted 1,2-diketones containing electron-withdrawing groups, affording excellent yields to desired products in all examples. In addition, the possibility of performing the reaction of 1,2-diketones 1a-e with o-phenylenediamine(2b) was also investigated [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] , entries 6-10). Using these substrates, a range of substituted quinoxalines was obtained in excellent yields using the nanoparticle-supported cobalt catalyst under optimized reaction conditions. In order to extend the scope of the reaction, the best conditions were employed in reactions of 1,2-diamino-4-nitrobenzene (2a) with other 1,2-diketones 1b-e with different patterns of substitution and the results are summarized in [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref]. As it can be seen on [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] (Entries 1-5), our methodology is suitable to a range of substituted 1,2-diketones containing electron-withdrawing groups, affording excellent yields to desired products in all examples. In addition, the possibility of performing the reaction of 1,2-diketones 1a-e with o-phenylenediamine(2b) was also investigated [fig_ref] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a [/fig_ref] , entries 6-10). Using these substrates, a range of substituted quinoxalines was obtained in excellent yields using the nanoparticle-supported cobalt catalyst under optimized reaction conditions. Reused runs were carried out under similarly optimized conditions using 5 mmol 1,2-diphenylethanedione (1a), 5 mmol of 1,2-diamino-4-nitrobenzene (2a) and supported cobalt catalyst (0.05 mmol, 0.165 g) at 78˝C in 10 mL of ethanol. The catalyst showed excellent recoverability and reusability over ten successive runs under the same conditions as the first run. It is quite remarkable that all materials discussed in this study exhibited outstanding structural stability by TGA (results not shown). The cobalt catalyst was found to be highly stable and reusable under the investigated conditions (up to 12 runs) without any significant loss of its catalytic activity [fig_ref] Table 3: Reuses of the supported CoNP catalyst in the reaction of 1,2-diphenylethanedione [/fig_ref]. Indeed, ICP analysis of both reaction filtrate and catalyst showed no detectable Co leaching (<0.5 ppm) in the reaction filtrate upon reaction completion, with an almost identical Co content for both fresh and reused catalyst (0.30 vs. 0.29 mmol of Co per gram of catalyst for fresh and 10-time reused material, respectively). The study of the scale-up reaction (from 1 to 20 mmol of substrate) was also investigated under the optimized reaction conditions. When the amount of 1a and 2a was increased to 20 mmol, the same conversion was obtained after 90 min under optimized conditions.
The catalytic performance of our system was eventually compared to reported literature data. As can be seen in [fig_ref] Table 4: Comparison of the result in the reaction of 1,2-diphenylethanedione [/fig_ref] , our recoverable catalytic system possesses remarkably improved activities as compared to those of related previously reported heterogeneous systems.
## Experimental section
## General information
Unless otherwise stated, all reagents and chemicals in this study were used as received and were not further purified (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany). Melting point recorded on a RY-1 microscopic melting apparatus (Hangzhou Chincan Trading Co., Shanghai, China) and uncorrected. 1 H-NMR and 13 C-NMR spectra were respectively recorded on 500 MHz and 125 MHz by using a Bruker Avance 500 spectrometer (Bruker BioSpin GmbH, Rheinstetten, Germany). Metal content in the materials was determined using inductively coupled plasma (ICP) in a Philips PU 70000 sequential spectrometer (Philips, Almelo, The Netherlands) equipped with an Echelle monochromator (0.0075 nm resolution). Samples were digested in HNO 3 and subsequently analyzed by ICP Nitrogen adsorption measurements (Philips) were carried out at 77 K using an ASAP 2000 volumetric adsorption analyzer from Micromeritics (Micromeritics, Norcross, GA, USA). The samples were outgassed for 24 h at 100˝C under vacuum (p b 10-2 Pa) and subsequently analyzed.
## Preparation of the supported cobalt catalyst
CoNPs was synthesized as previously reported [bib_ref] Supported cobalt oxide nanoparticles as efficient catalyst in esterification and amidation reactions, Rajabi [/bib_ref]. Briefly, salicylaldehyde (2 mmol, 0.244 g) was added to an excess of absolute MeOH, to which 3-aminopropyl(trimethoxy)silane (2 mmol, 0.352 g) was subsequently added. The color of the solution instantly changed to yellow indicating imine formation. After 3 h, cobalt (II) acetate, Co(OAc) 2¨2 H 2 O (1 mmol, 0.248 g) was added to the solution, and the mixture stirred for three additional hours to allow the new ligands to complex the cobalt. A color change from pink to olive green is observed. SBA-15 (3 g) was activated by refluxing in concentrated hydrochloric acid (6 M) and then washed thoroughly with deionized water and dried before undergoing chemical surface modification. This activation treatment readily hydrolyses the siloxane Si-O-Si bonds to Si-OH species which will be key to anchor the cobalt complex. Both complex and activated silica were then mixed and the mixture was stirred overnight. The solvent was removed using a rotary evaporator, and the resulting olive green solid dried at 80˝C overnight. The final product was washed with MeOH and water (to remove all physisorbed metal species) until the washings were colourless. Further drying of the solid product was carried out in an oven at 80˝C for 8 h. The loading of cobalt was calculated about 0.3 mmol¨g´1 and surface analysis showed cobalt oxide species well dispersed on the surface of SBA-15 with 450 m 2¨g´1 surface area and pore size of 3.6 nm with 0.77 cm 3¨g´1 mesoporous pore volume.
## General reaction procedure
To a mixture of 1,2-dicarbonyl compound 1a-e (1.0 mmol) and 1,2-diamine 2a-b (1.0 mmol) in ethanol (5 mL), supported CoNP (0.033g, 1 mol%) was added and the mixture was refluxed in an open flask for 90 min. Reactions were monitored by thin layer chromatography (TLC) until total disappearance of the starting material. After completion of the reaction, the reaction mixture was cooled to room temperature, and resulting solid was collected by filtration and dissolved in ethyl acetate (10 mL). The supported catalyst was recovered by filtration. After evaporation of solvent, the resulting solid product was purified by crystallization in ethanol.
## Selected spectroscopic data
6-Nitro-2,3-diphenylquinoxaline (
# Conclusions
In summary, we have developed an environmentally friendly and highly active cobalt nanoparticle on mesoporous SBA-15 material for the synthesis of quinoxalinesin excellent yields from 1,2-diamine and 1,2-dicarbonyl compounds. Reactions could efficiently afford the target products after short reaction times and were run under air and mild reaction conditions and require low loadings of the supported catalyst. The catalyst was found to be highly reusable for at least ten reaction runs under the investigated conditions.
[fig] Figure 1: Biologically important quinoxalines. [/fig]
[fig] Scheme 1: General scheme of the reactions. [/fig]
[table] Table 1 3 Table 1: Optimization of reaction condition a . Optimization of reaction condition a . [/table]
[table] Table 2: Generality of the reaction of 1,2-diketones with 1,2-diamines a . [/table]
[table] Table 1: Optimization of reaction condition a . [/table]
[table] Table 3: Reuses of the supported CoNP catalyst in the reaction of 1,2-diphenylethanedione (1a) with 1,2-diamino-4-nitrobenzene (2a). supported CoNPs ( 0.05 mmol, 0.165 g) in EtOH (10 mL) at reflux conditions for 90 min. b Isolated yields. [/table]
[table] Table 4: Comparison of the result in the reaction of 1,2-diphenylethanedione (1a) with 1,2-diaminobenzene (2a) with our method and the previous literature. [/table]
[bib_ref] Synthesis of two conjugated polymers as TNT chemosensor materials, Chen [/bib_ref] |
HELPing older people with very severe chronic obstructive pulmonary disease (HELP-COPD): mixed-method feasibility pilot randomised controlled trial of a novel intervention
# Introduction
Extending palliative care to people with advanced non-malignant disease is widely advocated, 1,2 but the implications in specific diseases are poorly understood. People with very severe chronic obstructive pulmonary disease (COPD) have a well-recognised burden of disabling physical symptoms, compounded by comorbidity, psychological distress and social isolation, [bib_ref] How well do we care for patients with end stage chronic obstructive..., Gore [/bib_ref] [bib_ref] Health care needs in end-stage COPD: a structured literature review, Habraken [/bib_ref] [bib_ref] Experiences of living and dying with COPD: a systematic review and synthesis..., Giacomini [/bib_ref] although they rarely have access to co-ordinated supportive care. [bib_ref] How well do we care for patients with end stage chronic obstructive..., Gore [/bib_ref] [bib_ref] Health care needs in end-stage COPD: a structured literature review, Habraken [/bib_ref] [bib_ref] Coordination of end-of-life care for patients with lung cancer and those with..., Epiphaniou [/bib_ref] The current approach, predicated on an ability to recognise a terminal phase,has been challenged in the context of COPD, [bib_ref] We need to stop looking for something that is not there…, Kendall [/bib_ref] in which prognostication for individuals is unpredictable. [bib_ref] Prediction of appropriate timing of palliative care for older adults with non-malignant..., Coventry [/bib_ref] [bib_ref] Survival of Medicare patients after enrollment in hospice programs, Christakis [/bib_ref] In addition, qualitative data suggest that a 'point of transition' to palliative care is meaningless in a life-long condition characterised by a fluctuating, unpredictable decline. [bib_ref] Living and dying with severe chronic obstructive pulmonary disease: multi-perspective longitudinal qualitative..., Pinnock [/bib_ref] [bib_ref] Palliative care in chronic illness. We need to move from prognostic paralysis..., Murray [/bib_ref] This is compounded by a tendency for people with COPD to be 'silent' about their physical and social disabilities. [bib_ref] Living and dying with severe chronic obstructive pulmonary disease: multi-perspective longitudinal qualitative..., Pinnock [/bib_ref] [bib_ref] The silence of patients with endstage COPD: a qualitative study, Habraken [/bib_ref] Recognising trigger events may be a strategy to overcome these professional and patient barriers. [bib_ref] Potential triggers for the holistic assessment of people with severe chronic obstructive..., Cawley [/bib_ref] Analogous to 'red flags' in the context of diagnosis, triggers are events (such as hospital admissions) in the life of the disease that are meaningful to patients as signifying deterioration in their condition, and visible to health and/or social care professionals so that they can trigger holistic assessment of needs. Palliation of symptoms and supportive care can then be intensified without the requirement to identify end-stage disease.
Through a programme of work, [bib_ref] Developing and Evaluating Complex Interventions: New Guidance, Craig [/bib_ref] [bib_ref] Effectiveness of holistic interventions for people with severe chronic obstructive pulmonary disease:..., Nurmatov [/bib_ref] we developed a novel intervention, HELPing older people with very severe chronic obstructive pulmonary disease (HELP-COPD), involving a proactive holistic assessment of physical, psychological, social and spiritual needs delivered by a specialist respiratory nurse shortly after an admission with an exacerbation of COPD. Our pilot trial used quantitative and qualitative methods to assess feasibility, acceptability and potential impact of the HELP-COPD intervention and to test the trial procedures.
# Materials and methods
## Participants and recruitment
People registered with Lothian general practices admitted to Edinburgh Royal Infirmary or St John's Hospital, Livingstone with a primary diagnosis of exacerbation of COPD were eligible. We excluded people with other significant lung disease, those unable to give informed consent (e.g., because of severe dementia) or to complete questionnaires in English. The trial respiratory nurse (SF) liaised with the ward staff to identify potentially eligible patients whom she approached with information about the study and obtained written informed consent.
## Questionnaires
We assessed the feasibility in this group of patients of using a range of validated questionnaires to assess outcomes. This included the Functional Assessment of Chronic Illness Therapy with the respiratory sub-scale (Functional Assessment of Cancer Therapy-Lung) that we considered might be a suitable primary outcome measure for a future trial. [bib_ref] The Functional Assessment of Chronic Illness Therapy (FACIT) Measurement System: properties, applications,..., Webster [/bib_ref] [bib_ref] Reliability and validity of the Functional Assessment of Cancer Therapy-Lung (FACT-L) quality..., Cella [/bib_ref] Other questionnaires measured health-related quality of life 20 and physical, [bib_ref] The significance of respiratory symptoms and the diagnosis of chronic bronchitis in..., Fletcher [/bib_ref] [bib_ref] Improving the process and outcome of care in COPD: development of a..., Jones [/bib_ref] [bib_ref] Development and first validation of the COPD Assessment Test, Jones [/bib_ref] psychological [bib_ref] The hospital anxiety and depression scale, Zigmund [/bib_ref] and spiritual well-being. [bib_ref] Measuring spiritual wellbeing in people with cancer: The functional assessment of chronic..., Peterman [/bib_ref] Baseline assessment and randomisation
The trial researcher (SB) visited the patient at home 2 weeks after discharge, undertook a baseline assessment (COPD-relevant history, smoking status, co-morbidity) and administered questionnaires, [bib_ref] The Functional Assessment of Chronic Illness Therapy (FACIT) Measurement System: properties, applications,..., Webster [/bib_ref] [bib_ref] Reliability and validity of the Functional Assessment of Cancer Therapy-Lung (FACT-L) quality..., Cella [/bib_ref] [bib_ref] A self-complete measure for chronic airflow limitation-the St George's Respiratory Questionnaire, Jones [/bib_ref] [bib_ref] The significance of respiratory symptoms and the diagnosis of chronic bronchitis in..., Fletcher [/bib_ref] [bib_ref] Improving the process and outcome of care in COPD: development of a..., Jones [/bib_ref] [bib_ref] Development and first validation of the COPD Assessment Test, Jones [/bib_ref] [bib_ref] The hospital anxiety and depression scale, Zigmund [/bib_ref] [bib_ref] Measuring spiritual wellbeing in people with cancer: The functional assessment of chronic..., Peterman [/bib_ref] and then randomised in the ratio of 3:1 (block size 8) using a computergenerated random sequence (Health Services Research Unit, Edinburgh) to HELP-COPD or usual care.
## Help-copd intervention
The HELP-COPD intervention is illustrated in .
The trial respiratory nurse, experienced in palliative aspects of respiratory care, arranged a home visit approximately 4 weeks post discharge, during which she used the exemplar open questions on the HELP-COPD record (see to prompt discussion about any physical, psychological, social or spiritual concerns. A checklist of agreed actions was recorded prominently on the HELP-COPD record to facilitate implementation, and referrals made through the usual channels.
The completed one-page record was retained by the patient, with copies for the primary and secondary care records. The respiratory nurse telephoned the patient at 1, 3 and 6 months to check progress with actions.
Usual care and routine clinical care Usual care provided to people in both groups included discharge planning with appropriate community or hospital follow-up, and management in accordance with national guidelines.Quantitative data collection The trial questionnaires, 18-25 plus a one-page questionnaire about health care contacts, were posted to the patient at 3 months (with postal and telephone reminders), and administered by the trial researcher during a home visit at 6 months. Details of referrals to other health/social care agencies were taken from the HELP-COPD record. Data were entered onto the trial database by the trial researcher, with 10% checked for accuracy.
## Sample size and statistical analysis
No formal sample size calculation was undertaken for this pilot work. We aimed to recruit 40 patients (randomised 30 to HELP-COPD: 10 to usual care), this being sufficient to assess feasibility of the intervention and trial procedures. In keeping with standard practice, we did not undertake any formal statistical comparisons to avoid over-interpretation of this small data set.
## Qualitative data collection
We purposively sampled up to 10 participants from the HELP-COPD group with diverse demography, co-morbidity and family/social circumstances, and scheduled interviews with them (and a carer, if preferred) at three time points:
- Shortly after the HELP-COPD assessment to capture reflections on the experience. - After 6 weeks to explore whether actions were being implemented. - After 6 months, to explore any benefit/detriment of using HELP-COPD, whether any actions were outstanding and why.
We recruited professionals from social work, primary care and community physiotherapy or nursing teams involved in providing care to people with COPD, and adopted flexible strategies to enable participation (e.g., individual interviews, telephone options, focused discussions at team meetings). The aim was to explore perceptions of the benefits (or otherwise) of the intervention. The topic guide (see Supplementary Information) was informed by the constructs of the Normalisation Process Theory (NPT), 27 (summarised in [fig_ref] Table 1: Normalisation Process Theory [/fig_ref] and refined iteratively as themes emerged. Interviews were audio-recorded with participants' written consent. . The HELP-COPD intervention. GP, general practitioner; HELP-COPD, HELPing older people with very severe chronic obstructive pulmonary disease.
## 2-6 weeks
## Additional data sources
The respiratory nurse made field notes immediately after each HELP-COPD contact (initial face-to-face visit, and then telephone follow-up calls at 1, 3 and 6 months (see ), and was interviewed twice: at the beginning and towards the end of the study. Copies of completed HELP-COPD records provided documentary evidence.
# Analysis
The fully transcribed interviews were coded by the qualitative researcher (MK) using NVivo. We used NPT, highlighted as potentially relevant during the theoretical modelling phases of our programme of work, [bib_ref] Developing and Evaluating Complex Interventions: New Guidance, Craig [/bib_ref] as a framework for our analysis 27 (see [fig_ref] Table 1: Normalisation Process Theory [/fig_ref]. At the end of the study, the trial team (HP, MK, SB, SF and Cristina Matthews) used the online NPT toolkitto aid understanding of the experience of using HELP-COPD as reflected in our data.
Interpretation and synthesis of qualitative and quantitative data Members of the team (MK, SB, SF and Cristina Matthews) met on multiple occasions to construct in-depth case studies of participants for whom we had quantitative, qualitative and documentary data, as well as field notes from the researchers and respiratory nurse. For each individual, we noted any 'unmet' needs, 'recently met' needs (e.g., referrals made during discharge planning), evidence of any impact (or not) of the HELP-COPD intervention and noted any relevant context that may have affected the outcome. Individual cases were then compared and contrasted to aid understanding of the feasibility and impact of the intervention.
As findings emerged, we arranged discussions with the NHS Lothian Respiratory Managed Clinical Network (Managed Clinical Networks are multidisciplinary, cross-sector groups that oversee provision of clinical services throughout Scotland) and palliative care colleagues.Preliminary findings were discussed in a workshop (including grant holders, invited clinicians and researchers), enabling us to site our intervention in the context of existing services. We also presented the findings to our Lay Advisory Group and gained their perspective.
Finally, we used both qualitative and quantitative data to answer our predefined key questions (see [fig_ref] Table 2: Answers to the a priori key questions for the feasibility pilot and... [/fig_ref] , which addressed feasibility of the intervention and the trial processes. for the CONSORT diagram, and Table 2 (Question 9) for further details. We had substantial attrition largely owing to illhealth/death. Of the 44 patients who provided consent during their admission, we randomised 32 patients after their baseline assessment 2 weeks post discharge. In all, 23 of the 24 allocated to HELP-COPD were well enough to participate in the intervention at 4 weeks. Further attrition meant that 19 participants completed the follow-up questionnaires.
# Results
## Flow of patients through the trial see
Recruitment to the qualitative study Eight patients (three with carers) provided 12 interviews. In addition, we collected data from 28 social or health care professionals in 18 individual interviews and two group interviews.
Baseline characteristics Baseline characteristics are given in [fig_ref] Table 3: Baseline characteristics Abbreviations [/fig_ref]. The HELP-COPD group had more co-morbid conditions and more previous hospital admissions than the usual care group. [fig_ref] Table 2: Answers to the a priori key questions for the feasibility pilot and... [/fig_ref] (Questions 1-6) gives the answers to our questions about the feasibility of the intervention and the implications for future practice. In summary, the HELP-COPD intervention was well received by patients, and the concept resonated with professionals, but we identified fewer actions than anticipated (an average of 1.6 actions/assessment, of which five were referrals: most were provision of information or contact details of support agencies (see [fig_ref] Table 4: Needs and actions resulting from the HELP-COPD assessment [/fig_ref] for further details)). Key reasons for this were the overlap with existing discharge planning, patients' adaptation to their long-term disability and their over-riding preference to retain independence.
## Feasibility of the intervention
Feasibility of the trial processes [fig_ref] Table 1: Normalisation Process Theory [/fig_ref] gives the answers to our questions about the feasibility of the trial processes and the implications for future research. Questionnaire completion proved burdensome for some patients, with seven declining to complete follow-up questionnaires citing ill-health. See Supplementary Information for a table of quantitative outcomes.
Using NPT to explore why the intervention appeared not to be as successful as hoped and to inform future iterations The NPT scoring was informed by the qualitative data, the minutes of seven steering group meetings and eight Lay Advisory Group meetings, and feedback from discussions with NHS Lothian Managed Clinical Network and the secondary care palliative care team. (See Supplementary Data for the spider plots derived from the NPT toolkit.)
Coherence: meaning and sense-making by participants The patients spoke positively when asked about their HELP-COPD assessment, although their descriptions were often interspersed with stories about other professionals whom they had seen, and thus it was not always clear how distinct the intervention was. Some described the novelty of such a comprehensive assessment: Individuals highlighted specific benefits that had followed from the discussion about spiritual or psychological needs.
'Yes, I think, possibly one thing came out of it on the psychological side. It asks 'Do you often feel anxious or panicky?' In general I would have said 'no', but I suddenly realised that 'Yes, I do when I get breathless' … I hadn't really thought about that NPT challenges researchers to consider how their novel intervention might become 'normalised' (or not) in routine practice.
HELP-COPD: a holistic assessment of need S Buckingham et al Patients were generally positive about the assessment, openly discussing concerns and coping strategies in all domains. Professionals were interested in the concept of HELP-COPD and considered that the intervention was feasible but, noted some practical issues (e.g., overlap with discharge planning and communication with GPs)
The core concept was acceptable, but communication/integration with other services needs further consideration
2. In what proportion of patients was an action identified?
Twenty-three patients received the intervention. Overall, 37 action points were noted: an average of 1.6 per assessment. Of these:
- Five were referrals (one each to pulmonary rehabilitation, the home oxygen service, carer support, social care direct and a request to the GP). - Fifteen were contacts for the patient to follow-up (e.g., social support agencies) - Seventeen people were given information about COPD, and local services.
The reasons for fewer actions than anticipated needs to be explored and addressed. Key issues were: overlap with discharge planning, the priority attached to retaining independence and adaptation to long-term disability.
3. What actions were identified/ triggered by the HELP-COPD assessment?
- Physical symptoms were discussed in 21 assessments, but only one required action (referral to GP about morphine to relieve breathless) - Psychological issues were discussed in 19 assessments, but no actions were generated. Most felt well supported by their family. - A range of social issues were discussed in most assessments, but only one referral (for help with social care). Maintaining independence was the priority. Although few actions were generated, the physical, psychological, social and spiritual aspects of the assessment resonated with the participants and should be retained as the underpinning structure of the assessment.
4. Did follow-up help ensure that referrals occurred in a timely manner?
The follow-up telephone calls generated few actions and were perceived as achieving very little. In contrast, clinicians from community teams highlighted on-going care as important for identifying problems.
Integration with existing services (e.g., primary care, community services) could enable ongoing care. There was a strong sense from professionals that existing services did not address the complex and long-term supportive care needs of people with severe COPD. The uncertain prognosis meant that palliative care services-inextricably linked in the minds of both professionals and patients to the final months of life -were not always seen as appropriate.
'I go to the [palliative care] services almost cap in hand and again … the first question they'll ask me is 'Do you think they're likely to die within 3 months?' I don't know' [TProf03:
Community nurse] 'Patients don't understand palliative care, they still think it's for cancer patients… … they think that if they're referred to the hospice they're going in there to die …' [TProf01: Community respiratory physiotherapist] The HELP-COPD intervention was perceived as potentially providing something distinct to current services, and delivery 'closer to home' was considered to be 'much less daunting'. However, the timing of the assessment meant that actions overlapped with the step-up in care that followed an admission.
A general practitioner observed that the HELP-COPD summary could be useful for colleagues less familiar with the patient than he was, but highlighted how the record could easily be overlooked in the electronic health record. Cognitive participation: commitment and engagement by participants HELP-COPD was delivered by the trial research nurse, and thus it did not involve NHS employees undertaking any new roles. However, professional buy-in to the concept of providing supportive care for people with severe COPD was evident. Members of the community teams described having attended 'introduction to palliative care' courses or perceived their role to include advanced care planning.
'I would say, even in the last, probably, week, I've spoken to at least two patients around palliative care sort of end-of-life issues to include, sort of, resuscitation, or just issues around that that' [TProf01: Community respiratory physiotherapist] The Lay Advisory Group was very positive about the intervention and supportive of patients holding their own HELP-COPD record.
Collective action: the work participants do to make the intervention function Because HELP-COPD was delivered as a research project, no specific actions were expected from NHS personnel. The
## Question
Result Implication for future iterations 9. What is the likely attrition rate?
Forty-four patients provided consent during their admission: - Twelve were unable to participate in the baseline assessment 2 weeks post discharge because of death (n = 3), readmission (n = 3), too ill (n = 3), uncontactable (n = 3). - We randomised 32 patients. Twenty-four were allocated to the HELP-COPD intervention: one was too ill to participate in the intervention at 4 weeks. - On-going attrition throughout the 6 months: 7 of 23 participants in the HELP-COPD group and five of eight in the usual care group mostly for reasons of ill-health.
This very substantial attrition, mainly related to severity of disease and the fact that we were recruiting at a time when the disease had just exacerbated, is an important consideration for researchers recruiting people with COPD at the time of an admission. [bib_ref] Prediction of appropriate timing of palliative care for older adults with non-malignant..., Coventry [/bib_ref]. What is the most appropriate duration for a trial?
Documentary evidence from the HELP-COPD records suggested that referrals to or uptake of pulmonary rehabilitation (one of the actions we identified) or other referral agencies (oxygen, social services) take time to implement.
Our experience suggests that an RCT should be at least 6 months in duration.
11. How feasible/ practical are the proposed outcome measures?
Questionnaire completion proved burdensome for some patients, with 13 unable/declining to complete follow-up questionnaires (3 had died, 3 were too ill, 2 had family illness/bereavement, 4 gave no specific reason). For those well enough to complete the questionnaire, FACT-L (the outcome we were considering as a future primary outcome) appeared to be acceptable.
In this frail group of patients use of questionnaires needs to be rationalised. FACT-L covers all domains and obviates the need for multiple questionnaires.
12. What is the expected change in the primary outcome measure?
The FACT-L showed a trend to a difference at 6 months (See . The change from baseline (85 to 97) in the intervention group is substantial and at least twice the minimum clinically important difference of 5.5
The substantial attrition means that numbers were very small, and difficult to interpret with confidence 13. Which outcomes matter to patients/carers?
The Lay Advisory Group considered that important outcomes for them were greater independence, ability to improve their social lives and to live fuller lives.
FACT-L reflects this and endorses our decision to use it.
14. What is 'usual care'?
Usual care in the context of the HELP-COPD intervention is (i) discharge planning and (ii) community services designed to help prevent admissions.
Future iterations of HELP-COPD need to integrate, not overlap, with these services.
Abbreviations: FACT-L, Functional Assessment of Cancer Therapy-Lung; GP, general practitioner; HELP-COPD, HELPing older people with very severe chronic obstructive pulmonary disease; RCT, randomised controlled trial.
participating hospitals supported identification of eligible patients and staff members participated in the workshops, suggesting active engagement (albeit to the research rather than a service). Most patients were able to find their HELP-COPD record to show the qualitative researcher, and some explained that they had shared it with a partner or carer, although it was not clear whether this had facilitated any action(s). Reflexive monitoring: participants reflect on or appraise the intervention There was a sense that the intervention was not working quite as intended, and professionals suggested how it might be improved.
The lack of an on-going relationship between the patient and professional was seen as a disadvantage.
'… giving someone ten, fifteen, twenty minutes, half an hour as a one-off is often not the way in which to address these sort of more complex patient problems that these patients present with, which is possibly why they've come up with nothing' [Community respiratory physiotherapist]
The lack of 'actions' initially frustrated the respiratory nurse, although later she recognised that providing information about COPD was enabling people to feel more in control. The Lay Advisory Group reinforced the importance of information and education.
In general, professionals were interested in the outcomes of the study. We were, for example, invited to lead a discussion with the Lothian Respiratory Managed Clinical Network who suggested that we should consider intervening earlier in the course of the disease, so that needs could be addressed when people were still relatively active. The Lay Advisory Group supported this approach.
# Discussion
## Main findings
We recruited 44 patients, but had very substantial attrition, mainly related to severity of disease and because we were recruiting at a high-risk time (during an admission). On average, 1.6 actions were identified per assessment: most were providing information about resources. Trial data collection was feasible, although ill-health meant that questionnaire completion was burdensome for some patients.
Despite the few actions, patients were generally positive about HELP-COPD, and professionals perceived HELP-COPD as addressing an important aspect of care, although the post-admission timing of the intervention resulted in overlap with existing discharge planning.
Strengths and limitations of this study Our study built on extensive qualitative work, 12 a review of existing interventions 17 and adopted a systematic approach to designing and piloting a complex intervention. [bib_ref] Developing and Evaluating Complex Interventions: New Guidance, Craig [/bib_ref] The findings of our single-centre study may not be representative of other contexts. Specifically, the majority of patients were recruited from a tertiary centre with a reputation for innovative care of people with severe COPD, although participants recruited from the other Lothian hospital had a similar profile of needs/ actions. , fatigue or weakness (n = 8), pain (n = 6), recurrent chest infections (n = 4), coughing/sputum production (n = 3), weight loss (n = 3).
Referrals to GP to discuss the use of morphine for breathlessness (n = 1), pulmonary rehabilitation (n = 1), home oxygen services (n = 1) Concerns about medication Self-management plan arranged (n = 2) rescue medication supplied (n = 1). Check the progress of pulmonary rehabilitation referral Psychological Struggling with anxiety (n = 7): Discussion of coping strategies (n = 6)
No actions were generated, as no one requested or accepted help for anxiety and/or depression Frustration at their limitations (n = 7) No one described themselves as 'depressed' Social Topics covered included loneliness, bereavement, activities of daily living (ADL), shopping, bathing, housework, housing, going out, going on holiday.
Referral for help with personal care (n = 1), carer support services (n = 1), assessment for a shower unit (n = 1)
Coping strategies discussed: pacing, accepting family help: a few had formal carers Provision of the HELP-COPD leaflet of local services and voluntary agencies that could provide a range of support (n = 12). Provision of information about specific issues (such as going on holiday, arranging insurance, volunteer drivers) was also provided. Problems with housing (e.g., inaccessible steps, n The substantial attrition between recruitment (during the admission) and delivery of the intervention at home 4 weeks later would have been a major issue in a fully powered randomised controlled trial. We achieved data saturation with respect to our key qualitative objective of understanding the feasibility of the intervention.
Researchers' attitudes influence design, data collection and analysis of qualitative studies; discussion with our multidisciplinary professional team and lay advisors ensured a balanced interpretation of the data. [bib_ref] Qualitative research: standards, challenges and guidelines, Malterud [/bib_ref] Interpretation of findings in relation to previously published work There has been a widespread assumption that supportive care for people with organ failure should be modelled on the successful cancer model, 1,2,7 predicated on identifying people who have reached a transition point in their disease trajectory when they would benefit from a supportive care approach.However, qualitative work has questioned the relevance of such a 'transition' in COPD, [bib_ref] Experiences of living and dying with COPD: a systematic review and synthesis..., Giacomini [/bib_ref] [bib_ref] Living and dying with severe chronic obstructive pulmonary disease: multi-perspective longitudinal qualitative..., Pinnock [/bib_ref] and has suggested that the concept of 'triggers' might be more appropriate. [bib_ref] Living and dying with severe chronic obstructive pulmonary disease: multi-perspective longitudinal qualitative..., Pinnock [/bib_ref] The HELP-COPD initiative used hospital admissions as a trigger, and although this had face validity there were two over-arching reasons why it seemed less successful than we had hoped.
First, the HELP-COPD intervention, in part, duplicated care already provided as part of discharge planning. Care for physical needs had already been stepped up and assessment undertaken for social support needs. Outstanding actions detected by HELP-COPD were then limited to providing information about local (often voluntary) support agencies.
Second, offers of help were frequently declined. The HELP-COPD intervention occurred 4 weeks post discharge, at a time when many people felt they were returning to (their) normal, [bib_ref] Experiences of living and dying with COPD: a systematic review and synthesis..., Giacomini [/bib_ref] [bib_ref] Living and dying with severe chronic obstructive pulmonary disease: multi-perspective longitudinal qualitative..., Pinnock [/bib_ref] [bib_ref] The silence of patients with endstage COPD: a qualitative study, Habraken [/bib_ref] and potentially needing less rather than more help. Recalibration of expectations as a result of adaptation to their slowly progressive disability 32 meant that action was not necessarily wanted despite significant symptoms and social burden. In addition, an over-riding preference to cope independently meant that family help was more acceptable than outside services. [bib_ref] Understanding the experiences of patients with advanced multimorbidity to improve care: serial..., Mason [/bib_ref] Implications for future research, policy and practice Despite adopting an overtly holistic approach with systematic assessment of physical, psychological, social and spiritual needs, HELP-COPD had limited success in identifying actions. The intervention thus requires further developmental work before progressing to a large-scale randomised controlled trial. [bib_ref] Developing and Evaluating Complex Interventions: New Guidance, Craig [/bib_ref] Future iterations of HELP-COPD should retain the holistic approach, but may seek to integrate brief assessments into the annual reviews of people with symptomatic COPD, 34 thus intervening earlier in the course of the disease. The burden imposed on family carers as patients eschew formal support needs to be considered and addressed.
# Conclusions
The HELP-COPD intervention was well received by patients and the concept resonated with professionals, but the holistic assessment delivered 4 weeks after an admission generated few actions apart from provision of information. A more appropriate approach may be to provide holistic care routinely throughout the life-long course of the disease as an integral component of good primary care management of a long-term condition.
[fig] This 6 -: month feasibility pilot was conducted during 2012/13 with ethical approval from South East Scotland Research Ethics Committee (Ref:12/SS/ 0016) and governance approval from NHS Lothian (No:2012/R/RES/02). [/fig]
[fig] ': Ive never had anything like this before. Not in-depth. You get asked basic questions [at the surgery] [T001] [/fig]
[fig] •: Fifteen spoke of pleasure in family life and/or the importance of their faith. Five had put their affairs in order recognising 'We all have to die sometime'. [/fig]
[fig] 18, Figure 2: when the respiratory nurse was not available to maintain the admission log 3-Month telephone FU (n = 22) Flow of patients through the trial. FU, follow-up; HELP-COPD, HELPing older people with very severe chronic obstructive pulmonary disease. [/fig]
[table] Table 1: Normalisation Process Theory (NPT) 27 Coherence 'Sense-making' How is the intervention understood by participants? How do they compare it with other practices? Cognitive participation How do participants come to take part in a practice? What keeps them motivated to continue taking part? Collective action How do participants make it work? How are their activities organised and structured? Reflexive monitoring How do participants evaluate a practice? How does this change over time and what are its effects? [/table]
[table] Table 2: Answers to the a priori key questions for the feasibility pilot and implications for future development [/table]
[table] Table 3: Baseline characteristics Abbreviations: HELP-COPD, HELPing older people with very severe chronic obstructive pulmonary disease; FACIT-Sp, Functional Assessment of Chronic Illness Therapy-Spiritual; FACT-L, Functional Assessment of Cancer Therapy-Lung; HADS, Hospital Anxiety and Depression Scale; MRC, Medical Research Council; SGRQ, St George's Respiratory Questionnaire. [/table]
[table] Table 4: Needs and actions resulting from the HELP-COPD assessment [/table]
|
Cysteine proteinase cathepsin H in tumours and sera of lung cancer patients: relation to prognosis and cigarette smoking
In order to evaluate the role of cysteine peptidase cathepsin H (Cath H) in human lung cancer its protein levels were determined in 148 pairs of lung tumour tissue and adjacent non-tumourous lung parenchyma using the enzyme-linked immunosorbent assay technique. Additionally, Cath H levels were determined in sera of 171 patients with malignant tumours, 34 patients with benign lung diseases and 47 healthy controls. The median level of Cath H in tumour tissue was 0.64 times that in the corresponding lung parenchyma. Relating tumour levels with histological type we found higher Cath H levels in small-cell and adenocarcinomas and lower levels in squamous cell carcinoma, large-cell carcinoma and secondary tumours. A significant difference in Cath H level between lung tumour tissue and non-tumourous lung parenchyma was associated with the group of cigarette smokers (156 vs 263 ng mg -1 protein, P < 0.001). For this group of patients Cath H tumour levels correlated with the survival rate, while for the entire patient population this was not the case. Smokers with high tumour levels of Cath H experienced poor survival. Cath H was significantly higher in sera of patients with malignant and benign lung diseases than in control sera (P < 0.001). The increase was significant for all histological types, being the highest in small-cell and squamous cell carcinomas. Our study reveals that in lung tumours there is different behaviour of Cath H compared with other cysteine peptidases, e.g. cathepsin B and cathepsin L. Variations between tissue and serum levels of Cath H indicate either reduced expression or enhanced secretion of this enzyme in lung tumours.
Irregular functioning of cysteine proteinases cathepsin (Cath) B, H and L has been proposed as being involved in the development of various diseases, including cancer [bib_ref] Immunohistochemical study of cathepsin B. Prognostic significance in human lung cancer, Sloane [/bib_ref] [bib_ref] Proteases as prognostic markers in cancer, Duffy [/bib_ref] [bib_ref] Cysteine proteinases and their endogenous inhibitors: target proteins for prognosis, diagnosis and..., Kos [/bib_ref]. Numerous studies have demonstrated a correlation of increased proteolytic activity of cysteine cathepsins with neoplastic transformation, tumour invasion and metastasis. Furthermore, the levels of these enzymes in tumours and some extracellular fluids have been shown to allow the disease-free and overall survival period to be predicted and therefore may serve as prognostic factors for cancer patients. However, the investigations were mainly focused on Cath B and Cath L, whereas the role of Cath H in malignant progression was much less studied and remains contradictory.
Increased levels of Cath H have been observed in glioblastoma and anaplastic astrocytoma when compared with normal brain tissue and low-grade gliomas [bib_ref] Expression and the role of cathepsin H in human glioma progression and..., Sivaparvathi [/bib_ref]. Higher levels were also found in tumour tissue and sera of patients with breast cancer and in sera of melanoma patients . However, in head and neck tumours the levels of Cath H were lower in tumour tissue when compared to adjacent control tissue [bib_ref] Lysosomal proteinases cathepsins D, B, H, L and their inhibitors stefins A..., Kos [/bib_ref]. Moreover, the lower levels of Cath H were associated with shorter disease-free survival period [bib_ref] Prognostic value of cathepsins B, H, L, D and their endogenous inhibitors..., Budihna [/bib_ref]. In the same study the opposite was true for Cath B and Cath L.
Cath H has also been shown to be a prognostic marker in melanoma patients. Patients with high serum levels of Cath H experienced significantly shorter overall survival than patients with low levels of the enzyme ). In addition it was shown to predict the effectiveness of chemoimmunotherapy exhibiting significantly higher levels in non-responders than in responders .
In lung carcinoma the behaviour of Cath H has not so far been studied, although increased activity and/or concentrations of Cath B [bib_ref] Plasminogen activator and cathepsin B in normal and malignant human lung tissue, Treftz [/bib_ref] [bib_ref] Increased cathepsin B activity in human lung tumors, Krepela [/bib_ref] [bib_ref] Prognostic value of increased lung tumor tissue cathepsin B, Ebert [/bib_ref] [bib_ref] Cathepsins D, B, and L in malignant human lung tissue, Ledakis [/bib_ref] [bib_ref] Lah T, Spiess E and Ebert W (1997b) Cathepsin B fraction active..., Werle [/bib_ref] , Cath L [bib_ref] Cathepsins D, B, and L in malignant human lung tissue, Ledakis [/bib_ref] [bib_ref] Lah T, Spiess E and Ebert W (1997b) Cathepsin B fraction active..., Werle [/bib_ref] and Cath S [bib_ref] Fluorometric microassays for the determination of cathepsin L and cathepsin S activities..., Werle [/bib_ref] are well documented in lung tumour cytosols and tissue sections. Higher levels of Cath B activity [bib_ref] Prognostic value of increased lung tumor tissue cathepsin B, Ebert [/bib_ref] and of Cath L protein [bib_ref] Lah T, Spiess E and Ebert W (1997b) Cathepsin B fraction active..., Werle [/bib_ref] were found to correlate with shorter survival of patients with lung carcinoma.
The aim of the present study was to determine the levels of Cath H in lung tumours and their surrounding histologically noncancerous lung parenchyma. Additionally, the levels of Cath H were determined in sera of patients with histologically distinct lung tumours and compared to controls. Results have been related to histopathological and clinical features, considering especially the correlation of individual tumour levels of Cath H with survival and with the smoking history of patients.
Cysteine proteinase cathepsin H in tumours and sera of lung cancer patients: relation to prognosis and cigarette smoking
# Materials and methods
## Patients
Lung tumour tissue and adjacent control (non-cancerous) lung parenchyma were obtained as matched paired samples from 148 lung cancer patients treated by surgery at the Thoraxhospital Heidelberg-Rohrbach. The age of patients in this group ranged from 15 to 78 (median 60). A major proportion were smokers (n = 90, 60.8%), 17 (11.5%) were ex-smokers, having stopped smoking at least 0.25 years before surgery, 41 (27.7%) were nonsmokers, or stopped smoking at least 5 years before surgery. The cell type of lung cancer was classified according to the WHO protocol and was based on predominant cell type . The tumour disease stage (pTNM) was classified according to the international staging system .
The second group of patients consisted of 171 patients with malignant tumours and 34 patients with benign diseases included in an ongoing prospective study on prognostic values of extracellular cysteine peptidases and their inhibitors in lung cancer. From patients' malignant tumours only Cath H post-therapy serum levels were included in the study. The median age and the range for these patients is comparable to the previous group. Out of 171 patients with malignant tumours 60 were small-cell, 67 adeno and 44 squamous cell type. The group of patients with benign diseases consisted of patients with pulmonary embolism, fibrosis, bronchostenosis, sarcoidosis, pneumonia, tuberculosis, pleuritis/emphysema, congestive heart failure, actinomycosis, bronchiectases and bronchial asthma. As a control group, sera from 47 healthy blood donors were collected .
## Sample collection
Tissue homogenization was carried out as described by [bib_ref] Assessment of cathepsin L activity by use of the inhibitor CA-074 compared..., Werle [/bib_ref]. Five-millilitre blood samples were collected from lung cancer patients after therapy. The blood was clotted at 4-8°C and centrifuged at 3000 rpm. The sera were stored at -30°C until analysed. Sera from blood donors were sampled in a similar way.
## Determination of cathepsin h
Cath H concentrations were determined by enzyme-linked immunosorbent assay (ELISA) (sandwich ELISA; Krka d.d., Novo mesto, Slovenia) developed at Jozef Stefan Institute (Ljubljana, Slovenia). The components were purified and characterized, and the test optimized as described . The sheep polyclonal and murine 1D10 monoclonal antibodies recognize precursor and mature forms of Cath H, as well as enzyme-inhibitor complexes .
Linearity of the ELISA was tested by serial dilution of tissue cytosols or sera to levels within the range of the assay (Schweiger et al, 1997). The measured values of diluted samples were subsequently compared with standard values. Recovery was tested by the addition of different amounts of antigen to samples with known antigen concentration and was found to vary between 87.4 and 104.9%. A microplate reader (SLT Rainbow, Austria) was used to measure absorbance in ELISA. Cath H was expressed in ng mg -1 of total protein. The detection limit of the assay was 2 ng ml -1 .
Tissue samples at 1:50 dilution and serum samples at 1:2 dilution were added to wells of a microtitre plate previously precoated with sheep polyclonal antibody to Cath H. The assay was then performed as described . After 2 h of incubation at 37°C the wells were washed and murine 1D10 monoclonal antibody, purified by affinity chromatography on Protein A Sepharose and conjugated subsequently with horseradish peroxidase (HRP) was added. After a further 2 h incubation at 37°C, peroxidase substrate 3,3,5,5-tetramethyl benzidine (TMB, Sigma, St Louis, MO, USA) in the presence of hydrogen peroxide was added. The amount of degraded substrate, as a measure of bound immuno-complexed Cath H, was measured by absorbance at 450 nm and the Cath H concentration was calculated from the calibration curve.
## Protein determination
Protein concentrations were determined according to [bib_ref] A rapid and sensitive method for the quantitation of microgram quantities of..., Bradford [/bib_ref]. Bovine serum albumin was used as standard.
# Statistical analysis
The results of Cath H measurements in the group under study are given as 5%, 50% (median) and 95% percentiles. For comparing the data of matched pairs of tumour and lung tissue, Wilcoxon's rank test was used. Differences in Cath H levels (tissue and serum) between various groups of patients were tested by Mann-Whitney and Kruskal-Wallis test. Univariate analysis of survival probability was performed by Kaplan-Meier analysis [bib_ref] Non-parametric estimation from incomplete observation, Kaplan [/bib_ref] , using the log-rank test for determining statistical significance between survival curves. Critlevel program [bib_ref] An exploratory procedure for the evaluation of quantitative prognostic factors, Abel [/bib_ref] was used for dichotomization of variables into low and high groups. In all tests, two-sided P-values below 5% were considered significant. Various statistical packages (SPSS program, SPSS Inc., Illinois, USA; PC-Statistic, TOPSOFT, Hannover, Germany; Statistica, StatSoft, Hamburg, Germany) were used.
# Results
## Distribution of cath h in lung tumours and in nontumourous lung parenchyma
The concentrations of Cath H in 148 tissue homogenates of lung tumours and the corresponding non-tumourous lung parenchyma are summarized in [fig_ref] Table 1: Cathepsin H in lung tumour homogenates [/fig_ref]. In the total patient population the median of Cath H concentration in tumour tissue was 0.64 times that in the control lung parenchyma.
The data were further analysed with respect to tumour histology, the anatomical spread out of the tumour (pTNM-stages), the degree of cell differentiation and smoking habits [fig_ref] Table 1: Cathepsin H in lung tumour homogenates [/fig_ref]. Regarding tumour histology, as seen in , we found the lowest median value of Cath H in tumour tissues, compared to their control lung counterparts, in large-cell carcinomas and in squamous cell carcinomas. The ratios for both carcinomas were significantly different (P = 0.001 and P = 0.028 respectively). In contrast, in carcinoids (n = 5) the median level of Cath H was 1.7-fold higher in tumour tissue than in lung parenchyma, but the difference was not statistically significant.
The highest median value of Cath H was detected in smallcell carcinomas (SCLC) and in adenocarcinomas (AC). While low numbers of SCLC (n = 3) did not permit reliable statistical evaluation, Cath H was significantly higher in AC than in squamous cell carcinomas (SCC), in large-cell carcinomas (LCC) and in secondary tumours (i.e. metastases to the lung). There was no statistically significant correlation between Cath H levels and cell differentiation, lymph node involvement and pTNM-staging. However, in non-tumourous lung tissue significantly lower levels of Cath H were found in patients with tumour size pT2 compared to those with sizes pT1, pT3 and pT4 (P = 0.006, p = 0.035 and P = 0.023 respectively).
Comparing Cath H levels in tumour tissue with those in lung parenchyma, significant differences in patients with tumour size pT3 and pT4 (P = 0.014 and P = 0.041 respectively), in patients with lymph node metastasis pN1 (P = 0.001) and in patients with pTNM-stage I, II, IIIa and pTNM-stage IIIb, IV (P = 0.005 and P = 0.01 respectively) are seen.
With regard to smoking habits, Cath H levels of non-tumourous lung parenchyma were significantly higher in smokers than in non-smokers (P = 0.042) and are shown in . There were also significant differences in Cath H levels between nontumourous lung parenchyma and lung tumour tissue in smokers (P < 0.001) and ex-smokers (P = 0.039), whereas for non-smokers no significant differences could be observed.
## Detection of cath h in sera of patients with lung tumours
The levels of Cath H measured in 205 serum samples from patients suffering from lung diseases and from 47 healthy humans are listed in . Compared to healthy controls, the level of Cath H was significantly increased in sera of lung cancer patients regardless of their histological type (P < 0.001). The highest levels were observed in SCLC and in SCC. The levels of both histological types were significantly higher than those from patients suffering from benign lung diseases (P = 0.019 and P = 0.0003 respectively).
# Survival analysis
Analysis of the entire patient population revealed no significant correlation between Cath H levels and overall survival probability. However, within the subgroup of smokers, Cath H levels correlated with survival probability, since 11 out of 18 patients with Cath H levels above 349 ng mg -1 died in the observation period of 4.6 years compared to 29 out of 68 patients below that cut-off level . For the subgroup of non-smokers the correlation of Cath H with overall survival probability was not significant.
# Discussion
In our study on 148 patients with lung carcinomas we found levels of Cath H concentration in tumour tissue 0.64 times those from adjacent lung parenchyma. A similar observation has been reported for head and neck carcinoma [bib_ref] Lysosomal proteinases cathepsins D, B, H, L and their inhibitors stefins A..., Kos [/bib_ref] [bib_ref] Prognostic value of cathepsins B, H, L, D and their endogenous inhibitors..., Budihna [/bib_ref] with 0.42 times lower Cath H levels in tumour tissue than in controls.
Relating tumour tissue levels of Cath H with histological cell type we found significantly higher Cath H levels in SCLC and AC than in SCC, LC and secondary tumours. There was no correlation of Cath H tumour level with tumour size, lymph node involvement, pTNM-stage and cell differentiation (grading).
Rather high levels of Cath H were found in non-tumourous lung parenchyma of patients with LC, whereas the lowest levels of Cath H were observed in pulmonary carcinoids. It should be noted that LC are high-grade neoplasms, whereas pulmonary carcinoids are low-grade neoplasms, with 5-year overall survival rates of about 11% and 70%, respectively [bib_ref] Surgical treatment of lung cancer, Mountain [/bib_ref].
Separating patients into smokers and non-smokers, a significant difference between tumour and control Cath H levels appeared only within the former group, whereas for non-smokers Cath H remained unchanged. This is the first evidence indicating association of Cath H expression or regulation with cigarette smoking in human lung cancer and is consistent with previous studies revealing an increase of Cath B and L activity in lung tissue of smokers [bib_ref] Prognostic value of increased lung tumor tissue cathepsin B, Ebert [/bib_ref] [bib_ref] Assessment of cathepsin L activity by use of the inhibitor CA-074 compared..., Werle [/bib_ref] , in their alveolar macrophages and alveolar lavage fluid [bib_ref] Increased cathepsin B-like activity in alveolar macrophages and bronchoalveolar lavage fluid from..., Chang [/bib_ref] [bib_ref] Cathepsin L activity is increased in alveolar macrophages and bronchoalveolar lavage fluid..., Takahashi [/bib_ref] as well as in alveolar macrophages and bronchoalveolar lavage fluid of rats exposed to cigarette smoke [bib_ref] Cathepsin L activity in alveolar macrophages of rats: response to cigarette smoke, Lesser [/bib_ref] [bib_ref] Emerging roles for cysteine proteases in human biology, Chapman [/bib_ref]. Furthermore, Yukio [bib_ref] Changes in immunoreactivity for cathepsin H in rat type II alveolar epithelial..., Ishii [/bib_ref] [bib_ref] Cysteine proteinases in bronchoalveolar epithelial cells and lavage fluid of rat lung, Ishii [/bib_ref] clearly demonstrated that Cath H is secreted mainly with surfactant from Type II alveolar epithelial cells and alveolar macrophages of rat lung. This report strongly indicates a different role of Cath H from those of Cath B and Cath L. In fact, contrary to Cath H, the overexpression of Cath B and Cath L has been reported in lung tumours as well as in several other studies of aetiologically different cancers (reviewed by [bib_ref] Cysteine proteinases and their endogenous inhibitors: target proteins for prognosis, diagnosis and..., Kos [/bib_ref]. Although Cath B, L and H are all lysosomal cysteine peptidases, large variations in expression of individual cathepsins in different tissues [bib_ref] The expression of cathepsin B and other lysosomal proteinases in normal tissues..., Qian [/bib_ref] different types of cells within the same tissue [bib_ref] Review: cell and tissue distribution of lysosomal cysteine proteinases, cathepsins B, H..., Uchiyama [/bib_ref] have been reported. Selective expression of individual peptidases in different types of tumour suggest that they may participate in specialized cellular functions [bib_ref] Differences in cathepsin B mRNA levels in rat tissues suggest specialized functions, Segundo [/bib_ref] [bib_ref] Emerging roles for cysteine proteases in human biology, Chapman [/bib_ref]. Distinct function of these enzymes could result from the differences in their structure. In contrast to Cath L, which is the most powerful endopeptidase, and Cath B, which possesses endopeptidase and exopeptidase activity, Cath H acts mainly as exo(amino) peptidase [bib_ref] Endo-and aminopeptidase activities of rat cathepsin H, Koga [/bib_ref] [bib_ref] Ebert W and Spiess E (1994) Imbalance between cathepsin B and cysteine..., Kirschke [/bib_ref]. The linkage between the enzymatic properties of cysteine peptidases and their differential function in human lung as well as in cancer progression is not known. With the ELISA technique, we determined not only mature Cath H but also its proform and complexed Cath H molecules. The active enzyme status of Cath H in lung tissues and lung tumours might provide further information including correlation with lung diseases and cancer.
## And even in
Since the difference in Cath H levels between non-tumourous lung parenchyma and tumours appeared in the group of smokers and not in the non-smokers, one may speculate that the level of Cath H expression is not mainly influenced by malignant progression but by the effects of cigarette smoking. A similar observation has been found for carcinoembryonic antigen (CEA) m-RNA levels in control and tumour lung tissues [bib_ref] Effect of cigarette smoke on the mRNA and protein expression of carcinoembryonic..., Ohwada [/bib_ref]. On the other hand, extracellular levels of Cath H, being significantly increased and correlating with the stage of malignancy in breast, colorectal and melanoma cancers [bib_ref] Cathepsins B, H and L in human breast carcinoma, Gabrijelcic [/bib_ref] [bib_ref] Cathepsins B, H, and L and their inhibitors stefin A and cystatin..., Kos [/bib_ref] , identify this enzyme as being active in tumour progression. Serum levels of Cath H determined in lung cancer patients clearly confirm the latter observations. Cath H levels were raised in sera of patients of all histological types, being the highest for SCLC and SCC. Patients suffering from benign lung diseases expressed lower levels of serum Cath H than patients with malignant tumours, but still significantly higher than those in healthy controls. Our results, revealing unchanged or decreased levels of Cath H in tumour tissues and increased levels in patients' sera, support the possibility that the secretion of Cath H is enhanced in lung tumours.
Lung tumours are very heterogeneous, expressing different metastatic abilities and then respond differently to chemo-and radiotherapy [bib_ref] Chemotherapie des nichtkleinzelligen Lungenkarzinoms, Manegold [/bib_ref] [bib_ref] Enzyme-linked immunosorbent assay for the detection of total cathepsin H in human..., Schraube [/bib_ref]. Overall survival probability rate is very poor for patients with SCLC [bib_ref] Lung cancer: therapeutic modalities and cytoprotection, Schiller [/bib_ref]. For patients with NSCLC the histological type and the tumour stage (pTNM) of the lung tumours are the most relevant prognostic indicators. However, within the groups of patients with low tumour stage, new prognostic factors are still needed to discriminate between high-and low-risk patients. Cath B has been shown to be a significant prognostic marker for NSCLC at enzyme activity [bib_ref] Prognostic value of increased lung tumor tissue cathepsin B, Ebert [/bib_ref] , and immunohistochemical levels . Cath L and specific cysteine peptidase inhibitor stefin B have also been shown to be prognostic factors in lung cancer patients [bib_ref] Expression of cysteine protease inhibitors stefin A, stefin B, and cystatin C..., Ebert [/bib_ref] [bib_ref] Lah T, Spiess E and Ebert W (1997b) Cathepsin B fraction active..., Werle [/bib_ref]. In the most recent study Cath B immunostaining was demonstrated as a strong, independent prognostic factor in patients with SCC [bib_ref] Immunochemical analysis of cathepsin B in lung tumours: an independent prognostic factor..., Werle [/bib_ref].
For Cath H we found no significant correlation between tumour levels and survival when in the entire patient population was analysed. However, in the subgroup of smokers, Cath H levels correlated with the survival probability rate. Thus, cigarette smoking and high level of tumour Cath H may predict high risk of death in lung cancer patients.
In conclusion, in lung tumours Cath H shows a different behaviour from other cysteine peptidases, e.g. Cath B or Cath L. Its levels were lower in tumours but significantly higher in sera of lung cancer patients, suggesting either its lower expression or enhanced secretion or both. The prognostic impact of Cath H concentration in lung cancer is rather low and this is in agreement with the lack of correlation with the clinical and pathological parameters that indicate progression of lung cancer. However, the association of Cath H and cigarette smoking revealed a significantly increased risk of death in lung cancer patients and indicates the involvement of Cath H in tobacco-induced carcinogenesis.
[fig] Figure 1, Figure 2: Cath H levels in matched pairs of lung parenchyma (s s) and tumour tissues ( ). The bold line in the box is the median value. The top and the bottom of the box represent 25th and 75th percentiles, respectively, and the ends of the bars represent the 5th and 95th percentiles Cath H levels in matched pairs of lung parenchyma (s s) and tumour tissues ( ) of non-smokers, ex-smokers and smokers. The bold line in the box is the median value. n represents the number of patients in the sub-group. The top and the bottom of the box represent 25th and 75th percentiles, respectively, and the ends of the bars represent the 5th [/fig]
[fig] Figure 3, Figure 4: Cath H levels in sera of patients with lung tumours and healthy controls. Error bars represent mean ± 2 s.Prognostic significance of Cath H within the group of smokers. A cut-off value of 349 ng mg -1 protein to divide patients into low-( Cath H < 349 ng mg -1 prot.) and high-Cath H ≥ 349 ng mg -1 prot.) groups [/fig]
[table] Table 1: Cathepsin H in lung tumour homogenates [/table]
|
CD5-negative chronic lymphocytic leukemia/small lymphocytic lymphoma in a patient with gastrointestinal mantle cell lymphoma: an unusual case report
Richter's syndrome, the development of high-grade non-Hodgkin lymphoma in patients with chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL), may be triggered by viral infections (eg, Epstein-Barr virus infection). Herein, we report an unusual case of CD5-negative CLL/SLL patient with gastrointestinal mantle cell lymphoma (MCL) and hepatitis B virus infection. CLL/SLL was diagnosed based on lymph node immunohistochemistry and bone marrow pathology. This patient was treated with seven cycles of multi-agent chemotherapy. During treatment, the hepatitis B viruses were activated. Then, after 20 months of antiviral treatment with entecavir, he developed abdominal discomfort and abdominal lymphadenopathy and was diagnosed with MCL based on intestinal biopsy. This work indicates that the hepatitis B virus in patients with CLL/SLL may accelerate the progress or transformation to MCL.
# Introduction
Chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) is common leukemia in adults and accounts for approximately 30% and 7% of lymphoid and nodal lymphomas, respectively. [bib_ref] Cyclin D1 expression by histiocytes may mimic cyclin D1-positive proliferation centres of..., Wu [/bib_ref] Second malignancies are frequent complications in CLL/SLL patients, and this process is commonly referred to as Richter's syndrome (RS). About 2-8%, 0.5%, and 0.1% of CLL/SLL patients progress to diffuse large B cell lymphoma, Hodgkin's lymphoma, and multiple myeloma, respectively. [bib_ref] Epstein-Barr virus-positive diffuse large B cell lymphoma arising from a chronic lymphocytic..., Chen [/bib_ref] Studies show that RS is commonly associated with Epstein-Barr virus (EBV), [bib_ref] Epstein-Barr virus-positive diffuse large B cell lymphoma arising from a chronic lymphocytic..., Chen [/bib_ref] [bib_ref] CLL progression after one cycle of FCR: richter's transformation versus EBV-associated lymphoproliferation, Jain [/bib_ref] [bib_ref] Concurrent classical Hodgkin lymphoma and plasmablastic lymphoma in a patient with chronic..., Foo [/bib_ref] karyotypic changes, [bib_ref] Second cancers and Richter transformation are the leading causes of death in..., Strati [/bib_ref] and gene mutations. [bib_ref] Microsatellite instability and hMLH1 promoter hypermethylation in Richter's transformation of chronic lymphocytic..., Fulop [/bib_ref] [bib_ref] IgVH mutational status and clonality analysis of richterʼs transformation, Mao [/bib_ref] [bib_ref] Aberrant somatic hypermutation in transformation of follicular lymphoma and chronic lymphocytic leukemia..., Rossi [/bib_ref] CLL/SLL usually expresses CD5 antigen, but 7-20% of CLL/SLL patients are CD5 negative. [bib_ref] Clinical and laboratory features of CD5-negative chronic lymphocytic leukemia, Demir [/bib_ref] Primary gastrointestinal mantle cell lymphoma (MCL) is a rare and progressive disorder that accounts for only 1-4% of primary gastrointestinal lymphoma. [bib_ref] Primary gastrointestinal lymphoma, Ghimire [/bib_ref] Here, we reported an unusual case that a 61-year-old patient previously diagnosed as CD5-negative CLL/SLL developed MCL after chemotherapy and antiviral treatment.
## Case report
A 61-year-old man with fever and lymph node enlargement for one month was admitted to our hospital. He had a fever (>38.5°C) for one month, but no night sweats , hepatitis B e antibody (HBeAb, +), and hepatitis B core antibody (HBcAb, +). Hepatitis B virus-deoxyribonucleic acid (HBV-DNA) and EBV were negative. Bone marrow pathology indicated that CD20, PAX-5, CD23, SIg, and Bcl-2 were positive; SOX-11, CD3, CD5, MPO, CD34, CD10, Bcl-6, MUM-1, LEF-1 or CyclinD1 were negative, and Ki-67 staining revealed a proliferative index of 10% . Immunohistochemistry (IHC) of cervical lymph node showed that the lymphocytes were mature, small, and positive for CD20, PAX-5, CD21, CD23, and Bcl-2, but negative for CD3, CD5, CyclinD1, SOX-11, CD10 or Bcl-6, Ki-67 was 15% [fig_ref] Table 1: Characteristics of the patient [/fig_ref]. Flow cytometry showed that lymphocytes accounted for 68.94% nuclear cells (35.11% of B lymphocytes); CD19, CD20, and CD23 were positive, CD22 was weakly positive; CD10, CD5, FMC-7, κ, and λ were negative. Fluorescence in situ hybridization of bone marrow did not find abnormal Bcl-2 (18q21), Bcl-6(3q27), CEP8/MYC/IGH (11q13/14q32), and API2/MALT1 (11q22/ 18q21). IgVH, IgDH, and IgK were rearranged. Karyotype analysis showed 46, XY [20]. The above examinations supported the diagnosis of CLL/SLL (CD5 negative). He was subsequently treated with seven cycles of multi-agent chemotherapy, including cyclophosphamide, vincristine, and prednisone (COP * 1), and rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (RCHOP * 6).
After that his superficial lymphadenopathy disappeared, but was found with HBsAg (+), HBsAb (-), HBeAg (-), HBeAb (+) and HBcAb (+), and HBV-DNA rose to 2.656×10^5 copy/mL, and EBV was still negative . Then,
[formula] HE CD5(-) CD23(+) CyclinD1(-) MUM(-) [/formula]
SOX-11(-) Immunohistochemistry of the patient's bone marrow: mature, and small lymphocytic (hematoxylin-eosin staining, H&E). CD20, CD23, SIg, Bcl-2, and PAX-5 were positive; CD5, LEF-1, Bcl-6, CD10, SOX-11, and CyclinD1 were negative, Ki-67 staining revealed a proliferative index of 10% (Scan 10×40).
[formula] CD5 - - + CD3 - - - CD10 - - - CD20 + + + CD23 + + - CD21 ND + + PAX-5 + + ND CyclinD1 - - + Bcl-2 + + ND Bcl-6 - - - Ki-67 10% 15% 45% CD34 - ND ND MPO - ND ND CK ND ND - SOX-11 - - + LEF-1 - ND - IRF-4( [/formula]
# Discussion
It is important to distinguish CLL/SLL from other small B-cell lymphoma, such as MCL, which is an aggressive and incurable non-Hodgkin lymphoma (NHL). The complete remission rate of MCL treated by chemotherapy is low, and the median overall survival of MCL is 4-5 years. [bib_ref] Mantle cell lymphoma: 2017 update on diagnosis, risk-stratification, and clinical management, Vose [/bib_ref] Our patient was diagnosed as gastrointestinal MCL after chemotherapy and antiviral treatment with entecavir. Although the presence of negative EBV during the treatment period was different from previous studies, 2-4 our patient was found positive HBV-DNA after chemotherapy, and then he accepted 20 months of antiviral treatment. A recent study also shows that antiviral treatment can lead to a complete remission of hepatitis C virus (HCV)-associated low-grade NHL, suggesting a causative role of HCV in these tumors. [bib_ref] Remission of follicular lymphoma after treatment for hepatitis C virus infection, Maciocia [/bib_ref] Metaanalyses further show HBV-infected patients have two-to three-fold higher risk of developing B-NHL. [bib_ref] High prevalence of hepatitis B virus infection in patients with aggressive B..., Wang [/bib_ref] [bib_ref] Hepatitis B virus infection and risk of non-Hodgkin lymphoma in South Korea:..., Engels [/bib_ref] [bib_ref] Hepatitis B infection increases the risk of non-Hodgkin lymphoma: a meta-analysis of..., Dalia [/bib_ref] HBV is a hepatotropic virus, but can also infect lymphocytes and the lymphoid system that has been shown as an important reservoir of HBV. [bib_ref] Hepatitis viruses and non-Hodgkin lymphoma: epidemiology, mechanisms of tumorigenesis, and therapeutic opportunities, Marcucci [/bib_ref] Like the scenario of EBV-driven lymphoma, HBV also directly infects B-cells, leading to the genetic alterations that contribute to tumor development. One possible mechanism is that like in HBV-induced hepatocellular carcinomas, 18 HBV DNA can integrate into the B-cell genome, directly activate oncogenes or repress tumor suppressors, leading to tumor development and progression. Gene mutation is another important cause for RS. [bib_ref] Microsatellite instability and hMLH1 promoter hypermethylation in Richter's transformation of chronic lymphocytic..., Fulop [/bib_ref] [bib_ref] IgVH mutational status and clonality analysis of richterʼs transformation, Mao [/bib_ref] [bib_ref] Aberrant somatic hypermutation in transformation of follicular lymphoma and chronic lymphocytic leukemia..., Rossi [/bib_ref] As reported, the HBV-associated gene expression signature is contributed by the enrichment of genes regulated by BCL6, FOXO1, and ZFP36L1, which contributed to HBVrelated lymphomagenesis. In this study, we detected IgVH, IgDH, and IgK rearrangement in our patient. The underlying molecular mechanisms of RS are largely unknown, and no reliable markers are available that may predict which CLL/SLL patients are prone to RS. Thus, more clinical data need to be further analyzed.
# Conclusion
Viral or other infectious complications mimicking RS in the context of CLL/SLL with HBV infection are an essential differential diagnosis to RS. Experiences from this patient suggest we also should pay attention to other viral infections, such as HBV, especially for EBV negative patients.
## Ethical approval and consent
Ethical approval for the publication of this case was obtained from the ethical committee of Anqing Municipal Hospital. The patient provided written informed consent to publish the case report and accompanying images.
## Dovepress
## 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.
Submit your manuscript here: https://www.dovepress.com/oncotargets-and-therapy-journal
[fig] Figure 2, Figure 3: HBV-DNA and EBV-DNA were detected by the patient. In October 2016, HBV-DNA of the patient was raised to 2.656×10^5 copy/mL, so far, EBV-DNA was still negative. Abbreviations: CLL/SLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; MCL, mantle cell lymphoma; HBV-DNA, hepatitis B virus-deoxyribonucleic acid; EBV-DNA, Epstein-Barr virus-deoxyribonucleic acid. Immunohistochemistry of the patient's intestinal biopsy (The arrow point to intestinal tumor tissue): CD5, CyclinD1, CD20, and SOX-11 were positive; CD10, CD23, and Bcl-6 were negative, Ki-67 staining revealed a proliferative index of 45% (Scan 10×40). [/fig]
[table] Table 1: Characteristics of the patient [/table]
|
The Effect of Bortezomib on Antibody-Mediated Rejection after Kidney Transplantation
Purpose:Recently, bortezomib has been used to treat antibody-mediated rejection (AMR) refractory to conventional treatment such as plasmapheresis, intravenous immunoglobulin, and rituximab. The authors aimed to describe their experiences when bortezomib was used to treat refractory AMR. Materials and Methods: Eleven refractory AMR episodes treated with bortezomib were included in this study. The patients received one or two cycles of bortezomib (1.3 mg/m 2 ) on days 1, 4, 8, and 11. Results: Bortezomib effectively reduced antibodies against various targets, including human leukocyte antigen (HLA) class I and II, ABO blood group antigen, and angiotensin II type 1 receptor. Antibodies were depleted or reduced significantly in eight AMR episodes. Overall, there was a significant improvement in the mean estimated glomerular filtration rate (eGFR) at 3 months after therapy (36.91±22.15 mL/min/1.73 m 2 ) versus eGFR at time of AMR diagnosis (17.00±9.25 mL/min/1.73 m 2 ; p=0.007). All six early-onset AMR episodes (within 6 months post-transplantation) showed full recovery of allograft function. Additionally, three of the five late-onset AMR episodes (>6 months post-transplantation) showed improved allograft function. Conclusion: Anti-humoral treatment based on bortezomib might be an effective strategy against refractory AMR caused by various types of antibodies. Notably, this treatment could be more effective in early-onset AMR than in late-onset AMR.
# Introduction
Despite major advances in transplant medicine, antibody-mediated rejection (AMR) remains one of the major barriers to successful long-term outcomes.Traditional treatments for AMR, such as plasmapheresis (PP), intravenous immunoglobulin (IVIg), and rituximab have provided suboptimal results. Furthermore, these strategies do not deplete plasma cells (B-lympho-cyte lineage cells) that produce antibodies. [bib_ref] Diagnosis and management of antibody-mediated rejection: current status and novel approaches, Djamali [/bib_ref] Recent studies have demonstrated that the proteasome inhibitor bortezomib reduces antibodies by depleting antibody-producing plasma cells; thus, bortezomib has been effectively used to treat AMR episodes refractory to traditional AMR therapies. [bib_ref] Review of bortezomib treatment of antibody-mediated rejection in renal transplantation, Ejaz [/bib_ref] [bib_ref] Proteasome inhibitor-based therapy for antibody-mediated rejection, Walsh [/bib_ref] [bib_ref] Early and late acute antibody-mediated rejection differ immunologically and in response to..., Walsh [/bib_ref] [bib_ref] Use of bortezomib as anti-humoral therapy in kidney transplantation, Yang [/bib_ref] [bib_ref] Outcomes of combination therapy for chronic antibody-mediated rejection in renal transplantation, Kim [/bib_ref] Bortezomib is a proteasome inhibitor that induces the apoptosis of plasma cells and was approved by the Food and Drug Administration for the treatment of multiple myeloma in 2008. [bib_ref] Improved survival in multiple myeloma and the impact of novel therapies, Kumar [/bib_ref] Recently, it has been used to reduce HLA antibodies either before transplantation or as treatment for AMR. Studies on bortezomib in the transplant field have focused on donor-specific anti-HLA antibodies (DSHA). [bib_ref] Review of bortezomib treatment of antibody-mediated rejection in renal transplantation, Ejaz [/bib_ref] [bib_ref] Proteasome inhibitor-based therapy for antibody-mediated rejection, Walsh [/bib_ref] [bib_ref] Early and late acute antibody-mediated rejection differ immunologically and in response to..., Walsh [/bib_ref] [bib_ref] Late antibody-mediated rejection in renal allografts: outcome after conventional and novel therapies, Gupta [/bib_ref] [bib_ref] Differential effect of bortezomib on HLA class I and class II antibody, Philogene [/bib_ref] However, in view of its action mechanism as a proteasome inhibitor, its effect may not be limited to anti-HLA antibodies, [bib_ref] Review of bortezomib treatment of antibody-mediated rejection in renal transplantation, Ejaz [/bib_ref] [bib_ref] Proteasome inhibitor-based therapy for antibody-mediated rejection, Walsh [/bib_ref] and in fact, bortezomib has been used to treat disorders other than multiple myeloma. [bib_ref] Immunoglobulin light chain amyloidosis: 2014 update on diagnosis, prognosis, and treatment, Gertz [/bib_ref] [bib_ref] The proteasome inhibitor bortezomib depletes plasma cells and protects mice with lupus-like..., Neubert [/bib_ref] Therefore, the aim of this study was to investigate the effect of bortezomib on AMR caused by various types of antibodies.
# Materials and methods
## Subjects
This retrospective study was conducted on ten consecutive patients diagnosed with AMR and treated with bortezomib from November 2011 to April 2014. The study was approved by the Institutional Review Board of Severance Hospital, Yonsei University Health System (4-2014-0968).
## Diagnosis and treatment of amr
Patients presented with the defining features of AMR, as described by the Banff 2011 meeting report, [bib_ref] Meeting report: new concepts in antibody-mediated rejection, Mengel [/bib_ref] and all ten were refractory to conventional treatment involving PP, IVIg (200 mg/kg administered after each PP treatment), and rituximab (single dose, 375 mg/m 2 ). Bortezomib was administered after conventional treatment had failed. Each bortezomib cycle consisted of four doses of 1.3 mg/m 2 , which was reduced to 1.0 mg/m 2 in three patients depending on toxicities, and was administered on days 1, 4, 8, and 11. The second cycle of bortezomib was decided on case by case after weighing up the advantages and disadvantages in accordance with each patient's clinical status (antibodies, serum creatinine, and side effects). Early-onset AMR was defined as an occurrence ≤6 months after kidney transplantation, and late-onset AMR was defined as an occurrence at >6 months posttransplantation.
## Renal function evaluation
Initial renal function was assessed at the time of AMR diagnosis based on serum creatinine and the estimated glomerular filtration rate (eGFR, calculated using the Modification of Diet in Renal Disease formula). Responses to therapy were also assessed using serum creatinine and eGFR at 3 months after treatment. We evaluated the most recent renal function data and followup duration.
## Detection and characterization of antibodies
DSHA were identified using a single antigen bead assay that utilized the multiplex flow-bead microarray method (Lifecodes LSA class I and II; Gen-Probe Transplant diagnostics, Inc., Stamford, CT, USA). The presence and antigen specificities of Abs to HLA-A, -B, -DR, and -DQ were determined. Results are expressed as mean fluorescence intensities (MFI). A normalized value of >1000 MFI was considered positive for DSHA. A C1q binding assay was performed on all available sera of recipients with DSHA (C1qScreen TM , One Lambda, CA, USA). Anti-ABO antibody titers were measured using standard serological techniques. 14 Anti-angiotensin II type 1 receptor (AT1R) antibodies were retrospectively evaluated in cases of biopsyproven AMR without DSHA. Levels of anti-AT1R antibodies (U/mL) were quantified using AT1R assay kits (One Lambda, CA, USA), which utilize the enzyme-linked immunosorbent assay principle.
# Statistical analysis
Statistical analysis was performed using SPSS version 20.0 (SPSS Inc., Chicago, IL, USA). Wilcoxon's signed-rank test was used to compare differences before and after treatment. p values of <0.05 were considered statistically significant.
# Results
## Baseline characteristics
Demographic data, including immunologic risk factors are presented in [fig_ref] Table 1: Baseline Characteristics of Patients LD, living donor [/fig_ref]. A total of ten patients received bortezomib for AMR. Two patients underwent deceased-donor kidney transplantation and eight underwent living-donor kidney transplantation. Five recipients had pre-formed DSHA. Of these five patients, four received preoperative desensitization, including PP, IVIg, and rituximab due to a positive crossmatch or ABO incompatibility. The remaining patient received one dose of rituximab prior to transplantation. One patient diagnosed with biopsyproven AMR in the absence of DSHA was pre-sensitized against AT1R.
Five episodes of AMR developed early (≤6 months posttransplantation), and four developed late (>6 months posttransplantation). One patient experienced recurrent AMR (at 1 and 27 months post-transplantation).
## Histologic data
Histologic data are shown in [fig_ref] Table 2: Pathologic Data of Patients [/fig_ref]. All rejection episodes manifested as pure AMR, that is, not as AMR mixed with acute cellular rejection. C4d status and biopsy scoring are presented according to the Banff 2011 classification. One of six early-onset AMR cases and four of five late-onset AMR cases exhibited transplant glomerulopathy (cg≥1).
## Clinical outcomes after bortezomib treatment
Treatment outcomes are shown in [fig_ref] Table 3: Changes of Renal Allograft Function and Antibody Levels after Bortezomib Treatment AMR,... [/fig_ref] and [fig_ref] Figure 1: Changes of renal allograft function after bortezomib treatment according to AMR onset... [/fig_ref]. Six episodes of AMR occurred within 6 months of transplantation. Patient B (refer to [fig_ref] Table 3: Changes of Renal Allograft Function and Antibody Levels after Bortezomib Treatment AMR,... [/fig_ref] for patient details) had anti-B antibody and class II HLA antibody. His initial anti-B isoagglutinin titer was low (IgM/IgG 1:32/1:16); however, this increased to 1:64/1:256 after transplantation despite postoperative PP with IVIg treatment. Patient F had no DSHA yet had a high anti-AT1R antibody titer at the time of biopsy. Four patients received a standard dose of bortezomib (1.3 mg/m 2 ), and two patients (A and F) received a reduced dose (1.0 mg/m 2 ) due to toxicities. In all early-onset AMR cases renal function fully recovered after bortezomib treatment. The DSHA of two patients (A and B) disappeared completely after treatment; however, although the DSHA of three patients (C, D, and E) declined, it was not eradicated. Patient F showed a decline of anti-AT1R antibody after treatment.
Five episodes of late-onset AMR were included in this study. Patient A experienced recurrent AMR with class I and II HLA antibodies. She received one cycle of bortezomib and recovered well. Three patients (G, H, and I) had class II HLA antibodies, and one patient (J) had class I HLA antibodies. They received one to two cycles of bortezomib. Renal function recovered in two patients (H and I) after bortezomib treatment; however, their antibody titers increased. Patient G received a reduced dose (1.0 mg/m 2 ) due to thrombocytopenia.
Overall, a significant improvement in the mean eGFR was observed at 3 months after therapy (36.91±22.15 mL/min/1.73 m 2 ) versus the mean eGFR at the time of diagnosis (17.00±9.25 mL/min/1.73 m 2 ; p=0.007).
Bortezomib-related toxicities (thrombocytopenia and peripheral neuropathy) were all transient and responded to con-servative management.
# Discussion
Traditional AMR therapeutic strategies have focused on antibody removal and B-cell depletion while not directly focusing on plasma cell depletion. However, bortezomib is a proteasome inhibitor that induces the apoptosis of plasma cells, which are the sole source of antibody production. [bib_ref] Diagnosis and management of antibody-mediated rejection: current status and novel approaches, Djamali [/bib_ref] [bib_ref] Early and late acute antibody-mediated rejection differ immunologically and in response to..., Walsh [/bib_ref] Patients in this series experienced substantial rejection episodes refractory to PP, IVIg, and rituximab. Bortezomib was effective against various antibody targets, including HLA class I and II, ABO antigen, [fig_ref] Table 1: Baseline Characteristics of Patients LD, living donor [/fig_ref] 1 Bx, biopsy; TG, transplant glomerulopathy; FSGS, focal segmental glomerulosclerosis. and angiotensin receptor. However, the antibody-removing effect of bortezomib was not sustained long-term. Furthermore, the treatment effect of bortezomib was more evident in earlyonset AMR than in late-onset AMR.
## ) (i) (v) (cg) (ct) (ci) (cv) (mm) (ah) (ptc)
In this case series, bortezomib effectively reduced antibodies against various targets, including HLA class I, HLA class II, ABO blood group antigen, and AT1R. The majority of studies conducted to date have only reported on the effectiveness of bortezomib against anti-HLA antibodies. [bib_ref] Proteasome inhibitor-based therapy for antibody-mediated rejection, Walsh [/bib_ref] [bib_ref] Early and late acute antibody-mediated rejection differ immunologically and in response to..., Walsh [/bib_ref] [bib_ref] Late antibody-mediated rejection in renal allografts: outcome after conventional and novel therapies, Gupta [/bib_ref] Only a small number of case reports have described the effect of bortezomib on anti-ABO antibody-mediated AMR. [bib_ref] Successful treatment of severe ABO antibody-mediated rejection using bortezomib: a case report, Westphal [/bib_ref] In the present study, AMR caused by anti-AT1R antibodies was included, and treatment outcomes were satisfactory. To the best of our knowledge, no previous studies have been conducted on the use of bortezomib for the treatment of AMR caused by anti-AT1R antibodies.
The effects of bortezomib on immune response are complex. Endoplasmic reticulum stress and caspase induction are considered the primary mechanisms by which bortezomib eliminates plasma cells. The inhibition of nuclear factor-kappa B activity plays a central role in the anti-humoral activity of bortezomib. In addition, it has been shown to cause apoptosis and cell-cycle arrest. [bib_ref] Review of bortezomib treatment of antibody-mediated rejection in renal transplantation, Ejaz [/bib_ref] [bib_ref] Proteasome inhibitor-based therapy for antibody-mediated rejection, Walsh [/bib_ref] With such a wide range of actions, bortezomib may also have an effect on various kinds of antibodies.
In view of costs and side-effects, the selection of the correct treatment target is important. We analyzed DSHA class (HLA class I or II) and complement binding capacity. In the present study, preformed antibodies were observed in most early onset cases, and antibodies in late onset AMR were predominantly directed at HLA class II. These distributions concur with those already described. [bib_ref] Diagnosis and management of antibody-mediated rejection: current status and novel approaches, Djamali [/bib_ref] [bib_ref] Early and late acute antibody-mediated rejection differ immunologically and in response to..., Walsh [/bib_ref] [bib_ref] Differential effect of bortezomib on HLA class I and class II antibody, Philogene [/bib_ref] Several studies reported that bortezomib has different effects on HLA class I and II antibodies. [bib_ref] Differential effect of bortezomib on HLA class I and class II antibody, Philogene [/bib_ref] Bortezomib reduces HLA class I-restricted antigen presentation by reducing cell-surface HLA class I expression. However, in the present study, most DSHAs were reduced after bortezomib treatment regardless of DSHA class. Recently, the complementbinding capacity of DSHA was found to be an important factor of graft injury. [bib_ref] Complement-binding anti-HLA antibodies and kidney-allograft survival, Loupy [/bib_ref] In the present study, we performed C1q binding assays to assess the complement binding capacity of DSHA retrospectively and found that bortezomib treatment was effective in AMR with or without C1q-fixing DSHA. However, caution is required in generalizing this result due to the small numbers of patients involved. Nevertheless, we experienced satisfactory treatment outcomes regardless of DSHA class or the comple- ment-binding capacity of DSHA.
The time between kidney transplantation and AMR onset impacts the treatment effect of bortezomib. In the present study, we divided our study group into early-onset AMR and late-onset AMR. The six early onset episodes experienced full recovery from AMR after one cycle of bortezomib treatment. However, late-onset AMR demonstrated a poorer response. Three of five patients with late-onset AMR showed improved renal allograft function, while the other two did not respond to bortezomib treatment. Prior studies have also reported different responses for early-and late-onset AMR. [bib_ref] Early and late acute antibody-mediated rejection differ immunologically and in response to..., Walsh [/bib_ref] [bib_ref] Late antibody-mediated rejection in renal allografts: outcome after conventional and novel therapies, Gupta [/bib_ref] Two explanations have been proposed. First, early-onset AMR is likely to be detected early in its course, whereas late-onset AMR is more likely to cause irreversible damage due to exposure to harmful antibodies for a longer time. [bib_ref] Early and late acute antibody-mediated rejection differ immunologically and in response to..., Walsh [/bib_ref] [bib_ref] Transplant glomerulopathy: subclinical incidence and association with alloantibody, Gloor [/bib_ref] In fact, most of our late-onset AMR patients demonstrated transplant glomerulopathy. Second, early-and late-onset AMR may differ in terms of plasma cell characteristics. [bib_ref] Early and late acute antibody-mediated rejection differ immunologically and in response to..., Walsh [/bib_ref] Future studies are required on this topic.
Despite its wide range of action, the duration of the anti-humoral response to bortezomib was not sustained. We found that certain patients experienced de novo antibody formation or antibody rebound after bortezomib treatment. According to previous studies, antibody titer after transplantation represents dynamic change and continues at varying levels thereafter. [bib_ref] Meeting report: new concepts in antibody-mediated rejection, Mengel [/bib_ref] [bib_ref] Significance of C4d Banff scores in early protocol biopsies of kidney transplant..., Loupy [/bib_ref] Accordingly, given that antibodies can reoccur, regular monitoring of antibody titers is required after bortezomib-based anti-humoral treatment.
Several limitations of our study require consideration. First, the study was limited by a small cohort size, as only eleven episodes were included. A larger, randomized controlled trial with a long-term follow-up is required. Second, given that patients received bortezomib after conventional treatments (PP+IVIg+ rituximab), it was difficult to assess the contribution of bortezomib to overall efficacy. Third, in the absence of accepted guidelines for the evaluation of therapeutic response in the setting of AMR, we evaluated renal function and antibody levels at 3 months after bortezomib treatment.
In conclusion, anti-humoral treatment based on bortezomib might be an alternative strategy against refractory AMR. Despite the small cohort size, our data suggested therapeutic effectiveness against a wide range of antibodies, including HLA antibodies, ABO blood group antibodies, and anti-AT1R antibodies. Finally, bortezomib treatment could be more effective in early-onset AMR than in late-onset AMR.
[fig] Figure 1: Changes of renal allograft function after bortezomib treatment according to AMR onset time. AMR, antibody-mediated rejection; Dx, diagnosis; mo, months. [/fig]
[table] Table 1: Baseline Characteristics of Patients LD, living donor; DD, deceased donor; CDC, complement-dependent cytotoxicity; AHG, anti-human globulin; PRA, panel reactive antibodies; AMR, antibody-mediated rejection; PP, plasmapheresis; IVIG, intravenous immunoglobulin; RIT, rituximab; KT, kidney transplantation; DSHA, donor specific anti-HLA antibodies; AT1R Ab, angiotensin II type 1 receptor antibody; N/A, not available. *Data were expressed as mean fluorescence intensities (MFI). [/table]
[table] Table 2: Pathologic Data of Patients [/table]
[table] Table 3: Changes of Renal Allograft Function and Antibody Levels after Bortezomib Treatment AMR, antibody-mediated rejection; Cr, creatinine; eGFR, estimated glomerular filtration rate; DSHA, donor specific anti-HLA antibodies; N/A, not available; AT1R Ab, angiotensin II type 1 receptor antibody; CREG, cross reactive group. *Data were expressed as mean fluorescence intensities. [/table]
|
Analytical Treatment Interruption in HIV Trials: Statistical and Study Design Considerations
Purpose of Review Analytical treatment interruption (ATI) remains an essential component in clinical studies investigating novel agents or combination treatment strategies aiming to induce HIV treatment-free remission or long-term viral control. We provide an overview on key study design aspects of ATI trials from the perspective of statisticians. Recent Findings ATI trial designs have evolved towards shorter treatment interruption phases and more frequent viral load monitoring aiming to reduce prolonged viremia risks. Criteria for ART resumption have evolved as well. Common outcome measures in modern ATI trials include time to viral rebound, viral control, and viral set point. Summary Design of the ATI component in HIV clinical trials is driven by the scientific question and the mechanism of action of the intervention being investigated.
# Introduction
Antiretroviral therapy (ART) effectively suppresses HIV viral load and reduces morbidity and mortality in people living with HIV (PLH). However, ART alone cannot eradicate the infection and lifelong treatment is needed [bib_ref] The biology of the HIV-1 latent reservoir and implications for cure strategies, Cohn [/bib_ref]. The success of the first two individuals cured of HIV [bib_ref] Long-term control of HIV by CCR5 Delta32/Delta32 stemcell transplantation, Hütter [/bib_ref] [bib_ref] Evidence for HIV-1 cure after CCR5Δ32/Δ32 allogeneic haemopoietic stem-cell transplantation 30 months..., Gupta [/bib_ref] and new discoveries of multiple broadly neutralizing antibodies (bNAbs) and therapeutic vaccines have renewed interest in developing strategies for sustaining long-term viral control without ART . The main obstacle to eradicating the virus is HIV reservoirs, cells where HIV is able to remain "latent" by being inactive [bib_ref] Targeting the Latent Reservoir for HIV-1, Sengupta [/bib_ref]. Multiple laboratory assays have been developed to measure the size of the HIV reservoir or other metrics of HIV persistence in PLH on treatment (levels of residual viremia, cell-associated HIV RNA, integrated total HIV DNA, infectious units per million cells (IUPM) determined by quantitative viral outgrowth assay, and intact proviral DNA assay (IPDA)) [bib_ref] The alphabet soup of HIV reservoir markers, Sharaf [/bib_ref] [bib_ref] Recommendations for measuring HIV reservoir size in cure-directed clinical trials, Abdel-Mohsen [/bib_ref]. These biomarkers have been primary endpoints of trials aiming to evaluate HIV remission strategies. However, studies show these reservoir measures do not correlate well with each other, and can under-or over-estimate reservoir size (e.g., the replication-competent virus population that start new rounds of infection) [bib_ref] Towards an HIV-1 cure: measuring the latent reservoir, Bruner [/bib_ref] [bib_ref] Comparative analysis of measures of viral reservoirs in HIV-1 eradication studies, Eriksson [/bib_ref]. Furthermore, these biomarker levels and the real clinical endpoint, absence of viral rebound after stopping ART, are often inconsistent [bib_ref] The critical roles of treatment interruption studies and biomarker identification in the..., Li [/bib_ref] [bib_ref] Antiretroviral therapy interruption (ATI) in HIV-1 infected patients participating in therapeutic vaccine..., Leal [/bib_ref]. Thus, ART interruption remains the essential component of clinical trials to evaluate new strategies or interventions aiming to achieve viral control without treatment. Analytical treatment interruption (ATI) makes endpoints like time to viral rebound or viral control feasible using plasma HIV RNA measurement, the only FDA-approved clinical measure, in trials assessing efficacy of interventions aimed at achieving HIV remission or viral control.
A review by Lau et al. found that 159 clinical studies (with and without interventions more than ART) have incorporated ATI from 2000 to 2017, with significant variation in duration of ATI, monitoring strategies, and thresholds for restarting ART. A 2018 forum at the Ragon Institute of MGH, MIT, and Harvard gathered clinical researchers in Cambridge, MA, to formulate recommendations for conducting ATI trials that covered scientific value, risks/benefits, and ATI methodologies, including ethical and community perspectives. Major points of discussion and consensus viewpoints achieved were published . In this current review, we will focus on key study design aspects of ATI trials from the perspective of statisticians, including choosing efficacy outcome measures, ART resumption criteria, and singlearm versus placebo-controlled design.
## Types of ati trials
When ATI was first introduced in HIV clinical trials in the 1990s and early 2000s, the only antiviral treatment was ART, and stopping treatment meant stopping all ART. Recently, the term ATI has been used in trial designs that stop ART but continue other investigational agents with potential antiviral activity, such as bNAbs. In several trials evaluating the safety and effect of one (VRC01 or 3BNC117) or multiple (3BNC117 and 1010-74) bNAbs in delaying viral rebound, participants received a first bNAb(s) infusion and then stopped ART, while continuing to receive two more doses of bNAb(s) [bib_ref] Effect of HIV antibody VRC01 on viral rebound after treatment interruption, Bar [/bib_ref] [bib_ref] HIV-1 antibody 3BNC117 suppresses viral rebound in humans during treatment interruption, Scheid [/bib_ref] [bib_ref] Combination therapy with anti-HIV-1 antibodies maintains viral suppression, Mendoza [/bib_ref]. Because trial participants are under coverage of bNAbs during the ATI phase after stopping ART, these trials assess the antiviral activity of other agents without ART and should be considered switch or maintenance studies.
Another design type assesses anti-reservoir activity or ability to induce prolonged HIV remission by stopping all interventions, including ART and other investigational agents that may have antiviral activity. These trials begin their ATI phase after all agents are cleared from participants' systems, which may require a washout period under ART coverage to allow clearance of such investigational agents. An element of ATI study design thus may involve estimation of the pharmacokinetics and half-life of the experimental agents, often based on data from prior studies but potentially reassessed in the first set of study participants before they reach the ATI time point. For example, the nonhuman primate study of bNAb ± TLR7 agonist included a 16-week period on ART after the interventions to allow bNAb washout prior to the ATI [bib_ref] Antibody and TLR7 agonist delay viral rebound in SHIVinfected monkeys, Borducchi [/bib_ref]. A washout period also helps to delineate the safety of ATI from that of investigational agents which will commonly be evaluated prior to initiating ATI. In this review, we will focus on the second type: trials that evaluate ATI after discontinuation of all interventions and if necessary, a washout period.
## Evolution of ati trial design
As initially used in late 1990s and early 2000s' HIV clinical studies, the ATI was a mechanism for examining therapeutic vaccine effects or reducing ART exposure [bib_ref] Treatment interruption as a tool to measure changes in immunologic response to..., Kutzler [/bib_ref] [bib_ref] Therapeutic vaccination for HIV: hopes and challenges, Stephenson [/bib_ref]. With these objectives, typical ATI trials in that era were designed with one or more pre-determined treatment interruption periods, but with less stringent ART resumption criteria than most current ATI studies. The primary outcome measures for these trials were viral load set point, peak viral load, and viral burden [bib_ref] Evidence that intermittent structured treatment interruption, but not immunization with ALVAC-HIV vCP1452,..., Jacobson [/bib_ref] [bib_ref] ACTG 5197: a placebo controlled trial of immunization of HIV-1 infected persons..., Schooley [/bib_ref] [bib_ref] Interruption of antiretroviral treatment in HIV-infected patients with preserved immune function is..., Skiest [/bib_ref] [bib_ref] A randomized, partially blinded phase 2 trial of antiretroviral therapy, HIV-specific immunizations,..., Kilby [/bib_ref].
The 2006 CD4-guided ART management trial called the SMART study assessed the benefits of treatment interruption and reduced ART exposure. SMART results showed that participants with prolonged treatment interruption had significantly increased risk of opportunistic disease, cardiovascular and other non-AIDS-defining events, and death [bib_ref] CD4+ count-guided interruption of antiretroviral treatment, El-Sadr [/bib_ref] , and raised concerns about ATI trials.
In 2009, Timothy Brown became the first individual to have been cured of HIV after stopping ART [bib_ref] Long-term control of HIV by CCR5 Delta32/Delta32 stemcell transplantation, Hütter [/bib_ref]. ATI trials have come back in focus as an approach to test the efficacy of strategies to control HIV viral replication without treatment. Contemporary ATI trial designs have addressed these safety concerns raised in the SMART study by evolving towards shorter treatment interruption phases and frequent participant monitoring [13 - - ], aiming to reduce prolonged viremia risks through frequent viral load monitoring and then restarting ART when viral load thresholds are reached [bib_ref] The critical roles of treatment interruption studies and biomarker identification in the..., Li [/bib_ref]. These studies use different primary outcome measures (e.g., time to viral rebound, viral control post-rebound) to understand viral rebound kinetics and identify sustained HIV suppression [bib_ref] Lessons learned from HIV antiretroviral treatment interruption trials, Wen [/bib_ref]. Their main design characteristics include a single, potentially shorter-duration ATI phase, frequent HIV RNA monitoring, and immediate ART restart when the viremia threshold is reached.
## Choices of virologic outcome measures
Outcome measures should reflect the anticipated mechanism of action of the intervention and the associated scientific questions. ATI trials have used various outcome measures as virologic endpoints, e.g., peak viral load, rate of initial viral load increase during ATI, time averaged area under the curve during viral rebound, time to viral rebound, and viral set point [bib_ref] Virological outcome measures during analytical treatment interruptions in chronic HIV-1-infected patients, Fehér [/bib_ref]. These outcome measures quantify viral rebound kinetics, and trials using them have provided valuable knowledge and experience in understanding viral rebound kinetics. Here we focus on three common outcome measures in ATI trials: time to viral rebound, viral control, and viral set point.
## Time to viral rebound
Time to viral load rebound quantifies the time an individual is off ART and maintains viral control below a prespecified HIV RNA threshold. Viral rebound is confirmed when two consecutive HIV RNA levels exceed the threshold. Several ATI studies have used time to viral rebound as a primary outcome measure [bib_ref] Panobinostat, a histone deacetylase inhibitor, for latent-virus reactivation in HIV-infected patients on..., Rasmussen [/bib_ref] [bib_ref] Safety and immunogenicity of a modified vaccinia Ankara-based HIV-1 vaccine (MVA-B) in..., Mothe [/bib_ref] [bib_ref] A randomized controlled safety/efficacy trial of therapeutic vaccination in HIV-infected individuals who..., Sneller [/bib_ref] [bib_ref] SEARCH 019 and RV254 study teams. A randomized trial of vorinostat with..., Kroon [/bib_ref]. Time to viral rebound may be the safest endpoint in an ATI trial as participants resume ART promptly once the rebound is confirmed and the endpoint is observed. This approach minimizes time off ART with viral load above the threshold.
Time to viral rebound is a time-to-event outcome measure and is observed with interval censoring, as rebound is detectable only when measuring HIV RNA during study visits. The exact rebound time will likely occur between visits. This outcome measure can be analyzed using classical approaches accommodating interval censoring [bib_ref] Time to Viral Rebound and Safety after Antiretroviral Treatment Interruption in Postpartum..., Le [/bib_ref]. However, these approaches either rely heavily on parametric assumptions difficult to verify in practice, or are computationally challenging [bib_ref] A demonstration of interval-censored survival analysis, Radke [/bib_ref]. Approaches relying on different parametric assumptions may make it challenging to compare results from different studies. Hence, analyses in most trials with a time to viral rebound outcome measure use the standard approach for right-censored data instead of interval-censored analysis approaches [bib_ref] Antibody and TLR7 agonist delay viral rebound in SHIVinfected monkeys, Borducchi [/bib_ref] [bib_ref] A randomized controlled safety/efficacy trial of therapeutic vaccination in HIV-infected individuals who..., Sneller [/bib_ref]. Comparison between treatment groups can use median time to viral rebound or proportion without rebound at certain time points. The advantage of approaches for right-censored data is that they require fewer assumptions than interval-censored approaches and are easy to implement with available statistical computing software. Participants who resume ART for reasons other than viral rebound can be censored as non-informative; i.e., their time to resume ART for other reasons is plausibly independent of their (unobserved) viral rebound. However, methods for dealing with right-censored data are prone to potential bias introduced by monitoring frequency, as less-frequent HIV RNA monitoring would cause estimation bias toward later time to viral rebound than more frequent monitoring. For example, study A has a monthly visit schedule while study B has a weekly visit schedule. Viral rebounds for some participants might be detected during the extra weekly visits between the monthly visits, and therefore study B will likely estimate a shorter time to rebound. The different HIV RNA monitoring schedules require consideration when comparing results between trials and highlights the importance of standardizing visit schedules when designing future ATI trials.
## Viral control
Viral control during ATI evaluates participants at a prespecified time point after they stop ART and with HIV RNA below a predetermined threshold, while remaining off ART. Depending on the evaluation time point and threshold level, participants can achieve viral control before or after viral rebound. The importance of this outcome measure has been recognized when interventions with immunotherapy are being investigated for viral control. Moreover, this endpoint is expected to be necessary for regulatory approval of any interventions aiming to induce viral suppression in the absence of ART. Several recently developed studies evaluating combination interventions with immunotherapy are using viral control as the outcome measure (NCT04340596; NCT04357821; NCT03588715). Because of the immunologic nature of these study treatments, where a period of viremia may be needed to generate host mechanisms of immunologic control, the primary efficacy outcome assesses viral control at a prespecified time point after discontinuation of ART instead of time to viral rebound. Viral control during ATI is a binary outcome. Participants remaining off ART but with HIV RNA above the threshold at evaluation are defined as failures. Participants meeting viral rebound criteria and resuming ART before evaluation week are also failures. Analytic approaches for this outcome are straightforward. Normally, the probability of viral control is estimated using the proportion of participants achieving viral control and confidence intervals. Comparison between arms use methods testing two independent samples, either with asymptotic approximation for trials with moderate-to-large sample sizes or exact methods for smaller trials. Direct comparisons between studies are possible. Investigators need to consider how to deal with participants who restart ART for other reasons, e.g., participant choice or pregnancy, before the evaluation time point. Counting these participants as failures might lead to underestimating the true effect of the intervention as these ART restarts might have been viral control successes if participants remained off ART. However, excluding these ART restarts will decrease the number included in the analysis, causing loss of precision and potential for bias, which may have a significant impact on early exploratory trials with relatively small sample sizes. One way to handle this missing data situation is to use methods for time-to-event data, estimating the cumulative probability of not achieving viral control at evaluation week, and assuming the other reasons for restarting ART are non-informative for the viral control at evaluation time point later. Participants who restart ART for reasons other than virologic criteria can be censored when restarting ART. [bib_ref] ACTG 5197: a placebo controlled trial of immunization of HIV-1 infected persons..., Schooley [/bib_ref] [bib_ref] Safety and immunogenicity of therapeutic DNA vaccination in individuals treated with antiretroviral..., Rosenberg [/bib_ref] [bib_ref] Dendritic cell immunotherapy for HIV-1 infection using autologous HIV-1 RNA: a randomized,..., Jacobson [/bib_ref] [bib_ref] A therapeutic HIV-1 vaccine enhances anti-HIV-1 immune responses in patients under highly..., Tung [/bib_ref]. Viral set point as a primary outcome measure may expose participants to a longer period of uncontrolled viremia, typically 12-16 weeks or longer. Studies have shown that individuals might reach viral steady state as early as 4-6 weeks after acute infection [bib_ref] Detection and treatment of Fiebig stage I HIV-1 infection in young at-risk..., Dong [/bib_ref] and participants achieved new viral set points 8-12 weeks after ATI [bib_ref] Evidence that intermittent structured treatment interruption, but not immunization with ALVAC-HIV vCP1452,..., Jacobson [/bib_ref] [bib_ref] ACTG 5197: a placebo controlled trial of immunization of HIV-1 infected persons..., Schooley [/bib_ref]. Generally, however, trial participants will still need to tolerate viral peaks as high as 100,000 copies/mL for several weeks before the endpoint is observed if viral set point is the primary endpoint.
## Viral set point
Viral set point is a continuous outcome, with comparisons between treatment groups using, for example, rank-based twosample tests. To avoid missing evaluations, the primary outcome measure is usually the average of two measurements or a single value when the other was missing. Participants with missing measurements are assigned the worst rank.
In general, clinical trial outcome measures and analysis methods are kept simple as discussed above but more complicated statistical modeling of viral rebound kinetics can also be applied [bib_ref] A flexible nonlinear mixed effects model for HIV viral load rebound after..., Wang [/bib_ref] [bib_ref] Comparison of empirical and dynamic models for HIV viral load rebound after..., Bing [/bib_ref]. However, these methods were developed using older ATI studies; substantive modeling complications might arise if applied to ATI trials with relatively short durations off ART such as studies with ART restart directly after viral rebound. Furthermore, these methods have potential for informative censoring, where ART restart censors the observations of subsequent off-ART viral loads.
## Art resumption criteria
As with ATI trial design, ART resumption criteria have evolved as knowledge of viral rebound kinetics accumulated. Early ATI trials that evaluated the therapeutic vaccine effect commonly used viral set point as a primary outcome measure and many designs specified when to resume ART. All participants would restart ART at the same predetermined study visit to allow the observation of viral set point. Participants could restart ART sooner if their CD4 counts declined substantially, if they had very high viral loads, or if they experienced clinical events. Following the SMART study results, ART resumption criteria evolved to minimize time off ART while viral load rebounded. Participants would restart ART when detectable viral load was confirmed.
A 2013 study found that elite or post-treatment controllers achieved viral control after a period of high viral load [bib_ref] Post-treatment HIV-1 controllers with a long-term virological remission after the interruption of..., Sáez-Cirión [/bib_ref]. Several successful trials in non-human primates also suggest a period of rebound viremia after ATI might be necessary to achieve viral control [bib_ref] Antibody and TLR7 agonist delay viral rebound in SHIVinfected monkeys, Borducchi [/bib_ref] [bib_ref] Combined active and passive immunization in SHIVinfected rhesus monkeys, Barouch [/bib_ref]. Expanding immune-mediated mechanisms, such as antibody-dependent cellular cytotoxicity or enhancing cytotoxic T-cell responses, can lead to sustained viral control, but requires viral replication and viral antigen expression. Thus, resuming ART promptly after confirming viral rebound may miss post-rebound viral controllers, especially if interventions depend on host-virus interaction after stopping ART.
The Ragon Institute ATI forum reached a consensus that when considering ART resumption criteria, viremia duration might be more important than the level of viremia .
## Aids clinical trials group (actg) reservoirs remission and cure transformative science group and its bnabs
Working Group developed criteria for ACTG studies with ATI components. The key virologic criterion is plasma HIV RNA ≥1000 copies/mL for ≥4 consecutive weeks and has not dropped 0.2 log 10 from the previous week. Other criteria include confirmed CD4 decline (< 350 cells/mm 3 or CD4% < 15%); acute retroviral syndrome; clinical disease progression; and not necessarily HIV disease-related reasons such as participant choice, pregnancy, a sexually transmitted infection, SARS-CoV-2, and unprotected sex.
These ART resumption criteria serve as a starting point for clinicians to consider when designing trials. The mechanism of action of the interventions under study and other design factors may require and justify modifying these criteria. For example, 12-16 weeks of uncontrolled viremia could be supported when viral set point is the primary endpoint . As studies using these criteria proceed, they will shed new light on the validity and utility of ART resumption criteria, which will likely evolve further.
## Single arm versus placebo controlled
Randomized placebo-controlled design is the gold standard for evaluating treatment efficacy in clinical research. However, single-arm trials are useful in early phase proof of concept studies, as they can assess the efficacy, safety, and tolerability of novel interventions before larger-scale randomized clinical trials [bib_ref] The design of single-arm clinical trials of combination antiretroviral regimens for treatment-naive..., Zheng [/bib_ref]. The availability of investigational products and enrollment feasibility often determine sample sizes for early phase clinical studies evaluating HIV remission strategies. A growing number of studies aim to enroll individuals treated during acute infection and require them to maintain viral suppression on ART for extended periods before entering the study. This rarer population has less diverse viral reservoirs than those who started ART during chronic infection, and may have a lower chance of viral breakthrough. Thus, single-arm design might be more feasible, with all participants receiving the investigational products. The focus of single-arm studies is to provide an accurate point estimate of the efficacy of the study treatment. Generally, the efficacy is compared with historical data. Investigators may set an efficacy threshold that serves as a go/no-go criterion for moving to the next stage of investigation. With historical data providing information comparable to a control arm, specifically for individuals who started ART during chronic infection, trials can devote more resources to the novel investigational treatment and minimize risk to participants. Early ATI trials provided ample historical data and characterized the viral rebound kinetics in chronically treated individuals [13- - , 45- - ], but much less data on acutely treated populations on modern, potent ART regimens. Recent studies suggest that ATI kinetics are similar after stopping older versus modern ART, and in the vast majority of cases, viral replication is rapidly resuppressed upon resumption of ART [bib_ref] Time to viral rebound after interuption of modern antiretroviral therapies, Li [/bib_ref] [bib_ref] Kinetics of plasma HIV rebound in the era of modern antiretroviral therapy, Sneller [/bib_ref].
The key assumption for single-arm design is that historical controls are sufficiently similar to a concurrent control arm. Potential drawbacks of single-arm studies include limited generalizability to populations not included in the study or limited comparability to other studies. Viral control rates can be due to factors other than the investigational agents. For example, a high viral control rate in a single-arm study may be due to enrollment of a population with less resistant or diverse virus and not the anti-reservoir efficacy of the investigational agent. Sneller et al. observed a spontaneous suppression rate higher than previously reported in the placebo participants following ATI [bib_ref] A randomized controlled safety/efficacy trial of therapeutic vaccination in HIV-infected individuals who..., Sneller [/bib_ref]. Other potential confounding and bias factors include different study implementation or assay testing procedures in the historical control study, and different participant characteristics, such as CD4 count at ART start, which may impair the interpretability of study results. Different ART resumption criteria, study designs, and visit schedules could also introduce confounding.
In early phase proof of concept studies, the value of placebo-controlled trials is less for well-powered randomized comparison than for descriptive understanding of the treatment effect. Definitive answers to address efficacy objectives will still need larger scale, randomized phase III clinical trials. Placebo-controlled trials also have the advantage of providing investigators with an unbiased assessment of safety by clinicians being blinded to treatment assignment when evaluating the association between adverse events and study treatment. Treatment effects between blinded study groups are compared directly and results are less prone to confounding factors, including the "white coat" effect of enrolling in the clinical trial setting.
The perspective of people living with well-controlled HIV also merits consideration. For some whose main motivation for participating in HIV remission trials is receiving an experimental agent that may reduce reservoir size or induce immune control, they might be reluctant to participate if they could be placed in a placebo group and also undergo ATI. Investigators should therefore engage community representatives to assess openness to placebo-controlled trials. An uneven allocation ratio, e.g., 2:1 to active:placebo, may be considered in these situations.
Despite their limitations, single-arm studies have a unique role in clinical trials when a randomized placebo-controlled clinical trial is not feasible or desirable. They can provide critical pilot efficacy and safety data on novel treatments aiming to reduce viral reservoirs or induce post-rebound viral control. However, if interpreting study findings properly requires a placebo group-as when there are no historical controls-then a study should use placebo-controlled design, or potentially an open-label or partially blinded randomized control arm. Ultimately, the science and study constraints should be the main factors that drive design choices.
# Ethical considerations
Ethical issues surrounding the use of ATIs in HIV clinical trials assessing strategies for treatment-free HIV remission have been extensively discussed . Recent studies have suggested that ATI is safe, and participants were able to re-suppress virus after ART re-initiation. Evidence also indicates that short-term ATI does not increase viral diversity or reservoir size [bib_ref] HIV-1 latent reservoir size and diversity are stable following brief treatment interruption, Salantes [/bib_ref] [bib_ref] Effect of short-term antiretroviral therapy interruption on levels of integrated HIV DNA, Strongin [/bib_ref] [bib_ref] Effect of analytical treatment interruption and reinitiation of antiretroviral therapy on HIV..., Clarridge [/bib_ref]. However, ATI is not without risks, which can include potential transmission to sex partners and acute retroviral syndrome . Prolonged viral exposure during ATI may also alter immune status in seronegative participants who initiated ART during acute infection. In a small trial evaluating viremic control after ATI in individuals treated during very early stage of HIV infection, four out of seven Fiebig I participants seroconverted after ATI [bib_ref] Rapid HIV RNA rebound after antiretroviral treatment interruption in persons durably suppressed..., Colby [/bib_ref].
Trial investigators must make sure participants understand the potential physical, social, financial, and psychological risks of ATI studies by carefully addressing risks in informed consent documents. Most ATI studies are early stage trials, and current interventions under investigation are not expected to lead to a conclusion of potential cure or remission. Prospective participants should be given realistic expectations of trial outcomes and experiences [bib_ref] Acceptability, motivation and the prospect of cure for people living with HIV..., Lau [/bib_ref] [bib_ref] Perceptions of HIV virologic control strategies among younger and older age groups..., Saberi [/bib_ref].
It is thus crucial that investigators determine if they can justify including ATI in an early stage trial, and they should explore alternatives or novel study designs. One option is to first show on-ART activity of the investigational agents or combinations, such as boosting anti-HIV immunity. A study design may include predefined go/no-go criteria for incorporating ATI based on evidence for treatment effects for on-ART virologic or immunologic outcomes. A staged design is another option where enrollment of additional participants is decided based on the efficacy signal of early stage in the same trial to minimize the number of participants who undergo ATI unnecessarily [bib_ref] Rapid HIV RNA rebound after antiretroviral treatment interruption in persons durably suppressed..., Colby [/bib_ref].
# Conclusion
ATI is essential in evaluating novel strategies aiming to achieve HIV treatment-free remission or long-term viral suppression. Design of the ATI component in HIV clinical trials is driven by the scientific question and the mechanism of action of the intervention being investigated, e.g., choosing outcome measures and ART resumption criteria. Single-arm design may be a viable option for proof of concept early phase studies when appropriate historical control data are available. However, investigators need to understand how a single-arm design may affect interpretation of the trial results.
At the time of authorship, the global SARS-CoV-2 pandemic has had major impacts on daily life, medical care, and clinical research. Conducting HIV clinical trials with ATI during the pandemic faces new challenges and requires reevaluating the risks and benefits together with pragmatic mitigation strategies [bib_ref] HIV cure research in the time of COVID-19 -Antiretroviral therapy treatment interruption..., Fidler [/bib_ref] [bib_ref] Operationalizing human immunodeficiency virus cure-related trials with analytic treatment interruptions during the..., Peluso [/bib_ref].
This paper focused on designing ATI studies in adult populations. The review by Lau et al. found only one ATI study out of 59 that investigated interventions beyond ART in a pediatric population. Investigators should consider whether similar design choices, with potentially more restraints, apply to trials in children and adolescents. Design challenges include limited prior pharmacologic and safety data on investigational agents, reduced sample volume for measuring HIV reservoirs, potentially more restrictive ATI eligibility criteria and low viral load thresholds for ART restart criteria [bib_ref] Early antiretroviral treatment of infants to attain HIV remission, Kuhn [/bib_ref] [bib_ref] Pharmacokinetics and safety of early nevirapine-based antiretroviral therapy for neonates at high..., Ruel [/bib_ref].
The search for treatments for HIV remission without ART continues and more trials involving ATI are proceeding. It may not be feasible to frame standardized approaches for conducting ATI trials, given the wide variety of interventions being studied. ATI trial design will thus continue to evolve to reflect the ever-changing clinical and scientific landscape for HIV remission and cure.
# Declarations
Conflicts of Interest The authors declare no competing interests.
Human and Animal Rights This article does not contain any studies with human or animal subjects performed by any of the authors.
[fig] Funding: This work was supported in part by AIDS Clinical Trials Group funded by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number UM1 AI068634 and UM1 AI068636, by Combined Immunologic Approaches to Cure HIV-1 funded under Award Number UM1 AI126603 and by the Statistical and Data Management Center of the International Maternal Pediatric Adolescent AIDS Clinical Trials Network under the National Institute of Allergy and Infectious Diseases grant No. UM1 AI068616. [/fig]
|
Recent Developments in Minimally Invasive Cardiac Surgery: Evolution or Revolution?
Intraluminal aortic clamping has been achieved until now by means of a sophisticated device consisting of a three-lumen catheter named Endoclamp, which allows at the same time occlusion of the aorta, antegrade delivering of cardioplegia, and venting through the aortic root. This tool has shown important advantages allowing aortic occlusion and perfusate delivering without a direct contact with ascending aorta reducing meanwhile the risk of traumatic and/or iatrogenic injuries. Recently, a new device (Intraclude catheter) with the same characteristics and properties has been proposed and introduced in clinical practice. The aim of this paper is to investigate the differences between Endoclamp and Intraclude catheters and to analyze the advantages advocated by this new device for intraluminal aortic occlusion since it is noticeable as these new technological tools are gaining more and more attractiveness due to their appraised clinical efficacy.
# Introduction
Since Bailey reported in 1951 the first surgical treatment of mitral valve with mitral annulus narrowing by external constriction through left thoracotomy, several approaches and techniques for mitral valve surgery have been progressively proposed, modified, and refined, especially after the introduction of CPB (cardiopulmonary bypass). LILLE-HEI and colleagues reported the first case of mitral valve repair through a right thoracotomy, using femoral artery cannulation for cardiopulmonary bypass (CPB) [bib_ref] Surgical correction of pure mitral insufficiency by annuloplasty under direct vision, Lillehei [/bib_ref]. With extensive use of CPB in 1960s, median sternotomy became the primary surgical approach to mitral valve considering the undoubted advantages related to its reliability, speed, and worthy exposure of the mitral valve as well as access to the rest of the heart compared to right thoracotomic incision. In the late 90s, the increasing interest for minimally invasive surgery due to patients demand, marketing policy, and new developing technologies stimulated the reconsideration of different left atrial and mitral approaches. At first, parasternal incision and partial sternotomy, following the work of Gillinov and Cosgrove [bib_ref] Minimally invasive mitral valve surgery: mini-sternotomy with extended transseptal approach, Gillinov [/bib_ref] , have been the most popular minimally invasive approach. More recently, in accordance with the rule of courses and historical claims, the experimental works performed in the laboratories at Stanford University and New York University have refocused the attention to right thoracotomy leading to the development of minithoracotomic videoassisted or video-guided port-access approach [bib_ref] Port Access (Thru-Port System) video-assisted mitral valve surgery, Chirichilli [/bib_ref].
This new technique exploited a peripheral perfusion and a balloon catheter for aortic occlusion allowing a less invasive procedure through a mini thoracotomy (4-6 cm) approach [bib_ref] Port-access mitral valve replacement in dogs, Pompili [/bib_ref] [bib_ref] Minimally invasive mitral valve replacement: port-access technique, feasibility, and myocardial functional preservation, Schwartz [/bib_ref] [bib_ref] Minimally invasive cardiopulmonary bypass with cardioplegic arrest: a closed chest technique with..., Schwartz [/bib_ref] , with undoubted advantages related to an overall reduction in surgical trauma, an effective improvement in patient comfort, lower morbidities, and shorter in-hospital stay, besides the remarkable cosmetic advantages especially in women [bib_ref] Techniques and results of direct-access minimally invasive mitral valve surgery: a paradigm..., Aklog [/bib_ref] [bib_ref] Minimally invasive mitral valve operations, Navia [/bib_ref] [bib_ref] Minimally invasive cardiac valve surgery improves patient satisfaction while reducing costs of..., Cohn [/bib_ref].
Due to materials and instrumentation improvements different cannulation and aortic clamping strategies are nowadays available:
(i) Full extra-thoracic CPB with external transthoracic aortic clamping.
(ii) Full extra-thoracic CPB with endoaortic clamping.
(iii) Central arterial cannulation with external transthoracic aortic clamping.
(iv) Central arterial cannulation with endoaortic clamping.
Intraluminal aortic clamping has been achieved until now by means of a sophisticated device consisting of a three-lumen catheter named Endoclamp, which allows at the same time occlusion of the aorta, antegrade delivering of cardioplegia, and venting through the aortic root. This tool has shown important advantages allowing aortic occlusion and perfusate delivering without a direct contact with ascending aorta reducing meanwhile the risk of traumatic and/or iatrogenic injuries. Therefore, other than in less invasive surgery procedures, the Endoclamp can be successfully adopted in systematic surgery especially in presence of extensive calcification of ascending aorta and in redo procedures allowing safer cross-clamping without requirement for dissection and manipulation of the ascending aorta and aortic root. Recently, a new device (Intraclude catheter) with the same characteristics and properties has been proposed and introduced in clinical practice. The aim of this paper is to analyze the differences between Endoclamp and Intraclude catheters and to analyze the advantages advocated by this new device for intraluminal aortic occlusion since it is noticeable as these new technological tools are gaining more and more attractiveness due to their appraised clinical efficacy.
## Surgical technique using endoclamp
Under general anesthesia, the patient is positioned in the supine position, with slight elevation (30 ∘ ) of the right hemithorax. All candidates for minithoracotomic video-assisted or video-guided port-access approach are ventilated with a double-lumen endotracheal tube in order to exclude, when needed, right lung ventilation. Monitoring includes double side arterial lines and use of TEE (Transesophageal Echocardiography). A small right mini thoracotomy (working port) and two additional ports are performed as previously described [bib_ref] Port Access (Thru-Port System) video-assisted mitral valve surgery, Chirichilli [/bib_ref] [bib_ref] Minimally invasive valve operations, Cosgrove [/bib_ref] [bib_ref] Evolution of mitral valve surgery: toward a totally endoscopic approach, Felger [/bib_ref] [bib_ref] A simple solution to a difficult problem: mitral pannus removal using a..., Rose [/bib_ref]. Venous drainage is generally achieved with double venous cannulation with a 14-20 Fr cannula placed percutaneously, under transesophageal echocardiographic guidance, through the internal jugular vein into the superior vena cava and a cannula into the inferior vena (Fr) cava through the femoral vein using Seldinger technique. Arterial cannulation is performed with placement under direct vision into the femoral artery of a dedicated 21-23 Fr cannula (Endoreturn), a Y-shaped device with a side branch that allows the introduction of the occlusion balloon. The Endoclamp endoaortic balloon is at this time placed from the Endoreturn cannula side branch, under TEE guidance, in the ascending aorta just above the sinotubular junction [fig_ref] Figure 1: Endoaortic balloon correct placement [/fig_ref]. The Endoclamp is a 10.5 Fr, 100 cm long, three-lumen catheter with an elastomeric balloon near its tip customized for endoluminal occlusion of the ascending aorta in order to separate the aortic root from arterial circulation. The surface contact of the balloon with the aortic wall is limited to 10 mm in length to avoid coronary occlusion during cardioplegia delivery. The large central lumen of the catheter attends two functions: delivery of cardioplegic solution during occlusion and venting from the left cardiac chambers both through the aortic root. The two remaining lumens are designed to serve as conduits for balloon inflation and aortic root pressure monitoring throughout the cardiac arrest. After proper position of the device, CPB is instituted and, under TEE monitoring to avoid balloon migration, the endoaortic balloon is progressively inflated with careful attention to its position at the level of the sinotubular junction [bib_ref] Continuous transesophageal echocardiographic (TEE) monitoring during port-access cardiac surgery, Schulze [/bib_ref]. Once its correct position is ascertained the cardioplegic BioMed Research International 3 solution is administered via an antegrade route. Using this technique blood pressure through the arterial line should be continuously monitored in order to promptly detect possible modifications that might be suggestive of partial or transient occlusion of the arterial arch vessels. In our experience the balloon should initially be inflated using an amount of saline solution proportional to the diameter of the sinotubular junction (1 : 1) to avoid unnecessary overexpansion. The balloon pressure is continuously monitored to reach and maintain a target pressure of around 350/400 mmHg. It is important to remember that during the cardiac arrest the Endoclamp pressure can progressively decrease by 10-20% due to the variation of temperature and the reduced stiffness of the aortic wall. In this case, no additional volume of intraballoon saline is needed if the heart is asystolic and the field is dry. The adhesion of the balloon with the aortic wall is crucial for steadiness of the device. Usually a balance is achieved because the balloon is pushed downstream by arterial flow from femoral arterial cannula and upstream by the pressure originated inside the root by cardioplegia delivery or by the systolic ejection from the heart before complete cardiac arrest. In presence of trivial aortic regurgitation and/or inadequate drainage of the left ventricle with unsatisfactory cardiac arrest during the antegrade cardioplegic induction, adenosine injection directly in the aortic root can be used to facilitate heart arrest and therefore facilitate the correct endoclamping function. In addition, to optimize left ventricular drainage, an Endopulmonary Vent Catheter, previously inserted by anesthesiologist through central vein access, can be used when needed. During surgery continuous TEE monitoring is recommended, to provide an optimal monitoring of venous cannula position, deairing maneuvers, and, certainly, an assessment of valve function after the operation [bib_ref] Echocardiographic monitoring of minimally invasive mitral valve surgery using an endoaortic clamp, Falk [/bib_ref] [bib_ref] Echocardiographic assessment in minimally invasive mitral valve surgery, Aybek [/bib_ref]. Other indirect monitoring tools include NIRS (INVOS) or transcranial Doppler [bib_ref] Control of endoaortic clamp position during port-access mitral valve operations using transcranial..., Schneider [/bib_ref] both able to detect any functional impairment of cerebral blood flow [bib_ref] Influence of intraoperative cerebral oximetry monitoring on neurocognitive function after coronary artery..., Colak [/bib_ref] caused by balloon migration. Once the surgical procedure has been completed, the balloon is deflated and partially withdrawn. At this time aortic venting is achieved through the same sideline of the catheter. After weaning from CPB, the device is fully removed through the side branch of the Endoreturn arterial cannula. A major concern of this technique with adoption of this device is the possible reduction of the arterial cannula lumen after introduction of the endoluminal balloon. Although this phenomenon is unusual with a 23 Fr cannula it has been described with the 21 Fr size with possible negative impact on systemic perfusion or in safety of CPB management. The reduction of the arterial cannula lumen related to the steric hindrance of the Endoclamp catheter can, in fact, result in an elevated pressure on the line of arterial perfusion (>250 mmHg), especially in cases of small and elastic femoral arteries, like in young women with small body surface area or in patients with severe atherosclerotic disease of the iliac-femoral tree. In our experience, with pressure >300 mmHg during full flow CPB, a contralateral femoral arterial cannulation, even with a small [bib_ref] Influence of intraoperative cerebral oximetry monitoring on neurocognitive function after coronary artery..., Colak [/bib_ref] [bib_ref] Port-access cardiac surgery: from a learning process to the standard, Greco [/bib_ref] Fr) cannula, is advisable by means of a double Y line perfusion to avoid malperfusion or complication on CPB lines or oxygenator.
## Pitfalls
Although most of studies showed the feasibility and safety of minimally invasive mitral valve surgery using the endoaortic clamping technique [bib_ref] Port-access cardiac surgery: from a learning process to the standard, Greco [/bib_ref] [bib_ref] Mitral valve surgery can now routinely be performed endoscopically, Casselman [/bib_ref] [bib_ref] Minimally invasive mitral valve surgery: a 6-year experience with 714 patients, Grossi [/bib_ref] , several specific issues emerged from data reported in literature. Particularly in first series, multiple severe complications have been described, such as aortic dissection or iliac artery injury [bib_ref] Minimally invasive port-access mitral valve surgery, Mohr [/bib_ref] , probably due to first generation stiffer catheters, worse monitoring techniques, and learning curve of the operator. In fact, originally Endoclamp position monitoring was performed using fluoroscopy only during the positioning of the device without any further control during the surgical procedure and, moreover, surgeons were not enough skilled to manage catheters and guide-wires. Nowadays severe vascular complications are very rare and cannot be considered a specific burden of endoluminal clamping technique itself [bib_ref] Transaortic endoclamp for mitral valve operation through right minithoracotomy in 369 patients, Glower [/bib_ref] [bib_ref] Video-assisted mitral surgery through a micro-access: a safe and reliable reality in..., Greco [/bib_ref].
Concerning generic complications some authors and especially the ISMICS (International Society for Minimally Invasive Cardiothoracic Surgery) summit claimed an augmented risk for cerebrovascular events, hypothetically due to greater use of femoral arterial cannulation for CPB. Plaque embolization during catheter introduction into the femoral artery or related to retrograde perfusion as well as traumatic injuries with consequent artery dissection or pseudoaneurysms formation and epiaortic vessel obstruction caused by balloon migration are well known complications [bib_ref] Minimally invasive mitral valve surgery: influence of aortic clamping technique on early..., Mazine [/bib_ref] [bib_ref] Avoiding vascular complications during minimally invasive, totally endoscopic intracardiac surgery, Jeanmart [/bib_ref] limited in more recent practice by the adoption of a more careful monitoring and dedicated catheter and devices.
Furthermore, another pitfall can be related to insertion of the endoluminal balloon into the arterial cannula that in some cases can induce increased resistance in the arterial line requiring a double arterial cannulation [bib_ref] Port-access surgery as elective approach for mitral valve operation in re-do procedures, Ricci [/bib_ref].
Other possible complications, described particularly in the first "era," are directly related to the endoluminal balloon device. Retrograde aortic dissection, balloon migration, balloon caught by suture for proximal anastomosis in coronary surgery, and balloon perforation during mitral valve procedure have been described and reported in literature [bib_ref] Complications of port-access cardiac surgery, Wimmer-Greinecker [/bib_ref]. Other authors, despite the overlapping results between classical sternotomy technique and port-access technique using Endoclamp, described cases that required switching to external cross-clamping to solve the unexpected problems arisen with endoluminal balloon [bib_ref] Minimally invasive port access versus conventional mitral valve surgery: prospective randomized study, Dogan [/bib_ref].
## Intraclude improvements
In this milieu a novel device, the Intraclude catheter, has been designed and approved for clinical use to overcome and solve these issues. Intraclude is a three-lumen catheter designed as an evolution with the same purposes of the Endoclamp. Innovations of this device compared to the Endoclamp are related particularly to the catheter size. New developed technology permitted, in fact, reducing the size of the device from 10.5 Fr to 9.5 Fr leading to the attainment of decreased resistances through the arterial cannula and, therefore, allowing a major blood flow at minor pressures, with consequent limitation of the stress and the so called "sand blast effect" related to the high-pressure blood jet. Some authors reported occasionally experiences of catheter "kinking," probably due to the minor caliber and softer material with respect to Endoclamp. This phenomenon is not frequent since the reduced caliber of the tip of Intraclude compared to the proximal part (hub) that maintains a diameter of 10.5 Fr is specifically designed to avoid the risk of kinking or twisting of the catheter and consequently to prevent the hazard of high pressure during cardioplegia delivering. Femoral artery injury or pseudoaneurysm formation leading to limb ischemia could be complications of femoral cannulation itself, regardless of the endoaortic balloon use, and prevention strategies are described elsewhere [bib_ref] Avoiding vascular complications during minimally invasive, totally endoscopic intracardiac surgery, Jeanmart [/bib_ref] [bib_ref] Thigh ischemia complicating femoral vessel cannulation for cardiopulmonary bypass, Gates [/bib_ref]. Concerning balloon migration, Intraclude has a different shaped balloon with wider cylindrical-shape compared to the spherical balloon of the Endoclamp with advantages in terms of surface contact that is increased from 10 mm to 18 mm. This change is supposed to ameliorate the stability of the endoluminal balloon with a better "fitting" into the aortic lumen and improved adhesion to the aortic wall allowing a more reliable sealing after inflation and therefore reducing the incidence of dislocation and/or blood leak into the ascending aorta.
To further ameliorate safety and performance of the endoluminal clamping, the Intraclude shaft is curve-shaped, allowing a better adhesion to the aortic arch and a perfect tip orientation towards the aortic valve for the cardioplegia delivery. Moreover, this curvature is supposed to avoid the "slack effect" experienced with the straight Endoclamp shaft limiting the possible migration of the balloon toward the aortic valve due to the tension generated by the catheter bended into the aortic arch.
The different balloon shape also allows the availability of a wider range of calibers, ranging from 20 to 40 mm, rather than the Endoclamp limited to a range of 20-38 mm. Dealing with balloon disruption or perforation, continuous TEE monitoring is essential to avoid malposition and prevent accidental perforation of the balloon during mitral valve surgery as well described in literature [bib_ref] Continuous transesophageal echocardiographic (TEE) monitoring during port-access cardiac surgery, Schulze [/bib_ref]. To prevent spontaneous ruptures, the inflation volume has been reduced to 35 mL (from 40 of Endoclamp), preventing in this way either possible clamp failure, or balloon migration. In this regard, it is useful to keep in mind that pressure variations throughout the surgical procedure can lead to migration of the device as well as that migration itself can determine pressure change with vicious cycle mechanism.
In a recent analysis conducted by several European surgeons [bib_ref] Minimally invasive versus conventional mitral valve surgery: a propensitymatched comparison, Svensson [/bib_ref] , the routine adoption of Intraclude device has been considered as one of the key factors leading to a significant reduction of morbidity and complications with particular emphasis concerning the stroke rate incidence.
# Discussion
Minimally invasive surgical treatment of valvular heart disease has steadily increased over the last several years becoming an established technique with high successful outcome in many specialized centers. The vast majority of larger clinical studies demonstrate that minimally invasive valve surgery using the port-access approach after an initial learning curve [bib_ref] Port-access cardiac surgery: from a learning process to the standard, Greco [/bib_ref] [bib_ref] Mitral valve surgery can now routinely be performed endoscopically, Casselman [/bib_ref] [bib_ref] Learning minimally invasive mitral valve surgery: A cumulative sum sequential probability analysis..., Holzhey [/bib_ref] [bib_ref] Different ways to repair the mitral valve with artificial chordae: a systematic..., Bizzarri [/bib_ref] [bib_ref] Accessory mitral valve tissue causing left ventricular outflow tract obstruction: case reports..., Prifti [/bib_ref] [bib_ref] Learning minimally invasive mitral valve surgery: a cumulative sum sequential probability analysis..., Holzhey [/bib_ref] is a safe and effective approach in terms of short-and long-term results, mainly for redo operations and even for elderly patients with moderately elevated perioperative risk. Furthermore this technique has shown a low morbidity and mortality achieving functional and echocardiographic outcomes comparable to those obtained with conventional surgery. Measurable patient benefits from case-matched control trials [bib_ref] Training surgeons in minimally invasive mitral valve repair: a single institution experience, Murzi [/bib_ref] [bib_ref] Minimally invasive mitral valve surgery: a systematic review and meta-analysis, Modi [/bib_ref] [bib_ref] Videoassisted cardioscopy for removal of primary left ventricular myxoma, Greco [/bib_ref] include less pain, less blood transfusions, fewer wound infections and pulmonary complications, and faster recovery as well as a better cosmetic result. Moreover recent improvements in Thru-Port systems offer excellent visualization of cardiac structures [fig_ref] Figure 3: Excellent visualization of the mitral valve by Thru-Port System [/fig_ref] through a virtually bloodless, unobstructed operative field without any increase of operative difficulty, procedure, and pump times, thus consenting to adopt successfully the same well established surgical techniques through the smallest incision possible.
In this setting the new device Intraclude undoubtedly improved safety and properness of intraluminal aortic occlusion during minimally invasive mitral surgery. The preshaped curved silhouette, the reduced diameter, and the cylindrical balloon profile have shown unquestionable advances allowing an easier and more reliable endoartic clamping with positive impact in terms of reduced stroke incidence suggesting a spreader use of this device other than in minimally invasive surgery. New catheter Intraclude has been introduced in the European market in 2012, replacing the old catheter Endoclamp. Since then, more than 2500 catheters have been used in Europe for ascending aorta occlusion and cardioplegia delivering during minimally invasive mitral valve surgery.
In our experience, we used the catheter Endoclamp from 2000 in our patients in more than 600 operations. We moved to use the new device, Intraclude, from the beginning, in our patients and we performed more than 60 cases with this device. At the beginning of the experience, we had some concerns regarding the extreme softness of the catheter with some risk of twisting and kinking. Nevertheless, we immediately appreciated, with respect to Endoclamp, the fact that its reduced size allowed a minor increase in the arterial line pressure during perfusion. Using Endoclamp we experienced at least five cases of migration of the balloon toward aortic valve and left ventricle. We did not have similar cases with Intraclude probably due to wider adhesion of this balloon to the aortic wall. The preshaped curvature of Intraclude makes its position easier in contact with the aortic arch reducing the need of additional maneuvers to avoid slack of catheter in thoracic aorta. From our experience, during introduction of the device it is mandatory to avoid any twisting of the curvature before the tip of the catheter reaches the root of ascending aorta. We had four cases of spontaneous rupture of Endoclamp but none using Intraclude. This could be related to the different shape of the balloon and/or to the material of this new device that seems to be helpful in terms of strength and fitting of the balloon with irregularities of the aortic wall. In conclusion we firmly believe all these technological developments and tools applied to minimally invasive procedure have gained, over time, more and more attractiveness due to their appraised clinical efficacy, and the acquired clinical experience in thousands of patients worldwide has led to a global improvement and to an implementation of this promising and ground-breaking surgical approach. In brief we are facing a gradual process where something changes into a different and usually more complex or better form, which simply means evolution.
## Conflict of interests
Professor Ernesto Greco has consulted for Edwards Lifesciences on minimally invasive valve surgery and holds a patent concerning minimally invasive access (no. US D701,305 S). The other authors declare that they have no competing interests.
# Authors' contribution
Ernesto Greco and Antonino G. M. Marullo conceived the paper. All authors have been involved in drafting the paper or revising it critically for important intellectual content and have given final approval of the version to be published. All authors read and approved the final paper.
[fig] Figure 1: Endoaortic balloon correct placement. [/fig]
[fig] Figure 2: TEE monitoring of balloon position. [/fig]
[fig] Figure 3: Excellent visualization of the mitral valve by Thru-Port System. [/fig]
|
Improved recording of work relatedness during patient consultations in occupational primary health care: a cluster randomized controlled trial using routine data
Background: Prolonging working careers is a key policy goal in ageing populations in Europe, but reaching this goal is complex. Occupational health services are in the best position to contribute towards prolonging working careers through preventing illnesses that cause work disability and early retirement. However, impacting on the trajectory between illness and work disability requires continuity of care and follow up, enabled through identifying patients at risk. We aimed to determine whether a combined educational and electronic reminder system in occupational health care could improve the recording and follow up of primary care visits made by patients at risk of work disability, and whether the system could impact on sickness absence rates.Methods: This study is a pragmatic, cluster-randomized controlled trial using medical record data. Twenty-two Pihlajalinna Työterveys units were randomized into an intervention group receiving education and electronic reminders or a group receiving usual care through minimization methods. Patient consultation data were extracted from routine Pihlajalinna Työterveys patient registers from 2015 to 2017. In addition, process indicators were collected from the electronic system. Data were cleaned and analysed on an intention-to-treat basis using analysis of covariance.Results: There was no significant difference between intervention and control units in terms of sickness absences of different duration. Process indicators suggested that there was a change in physicians' practice of recording patients' risk of work disability and work-relatedness of visits following the educational intervention.Conclusion: Education with an electronic reminder can change physicians' practice, but long-term follow up is needed to determine whether this impacts on patients' sickness absences.
# Background
Prolonging working careers is a key policy goal among ageing populations in Europe, but reaching this goal is complex. Economics and personal health influence decisions about whether to continue at work at pension age [bib_ref] Why work beyond 65? Discourse on the decision to continue working or..., Nilsson [/bib_ref]. A more pressing problem in European settings is the increasing number of disability pensions, which at least in Finland mostly affects young people of working age. An estimated 145000individuals are on early disability pension in Finland. Being on disability pensions affects personal finances, wellbeingand national insurance expenditure. In 2015 the disability pension expenditure in Finland was 2057 million euros.
Poor working conditions and workplace risks increase the likelihood of disability pensions [bib_ref] Predictors of disability pension over a 10-year period for men and women, Albertsen [/bib_ref]. While work can bring economic, psychological and even health benefits [bib_ref] Does employment affect health?, Ross [/bib_ref] , workplace risks and conditions can impact negatively on both physical and mental health [bib_ref] Job strain and the risk of disability pension due to musculoskeletal disorders,..., Mäntyniemi [/bib_ref] [bib_ref] Work-related primary care in occupational health physician's practice, Ikonen [/bib_ref] and exacerbate already existing conditions, such as depression. Work-related disorders, and thus work disability, can be prevented through close collaboration between workplaces and health care. In Finland, occupational health services (OHS) provide both preventive services in the workplace and curative primary healthcare services for individual employees. The key role of Finnish OHS is to provide healthcare services to organisations' employees, during which they can identify patients at risk of work-related diseases, and identify injuries and illnesses that threaten an employee's ability to work. They are in a key position in the country to implement workplace interventions that can help prevent work-related diseases and decrease rates of work disability [bib_ref] Do workplace interventions reduce disability rates, Midtsundstad [/bib_ref]. However, to date, little research has been conducted on the effectiveness of OHS in preventing work disability. In order to impact on work disability, we need to test different interventions for identifying patients at risk and to target prevention and care to them before their health worsens to the point of disability. The current OHS process for impacting on work disability mandates that a patient from primary care is directed to appropriate workplace or personal interventions following the identification of a risk to the employee's ability to work or a work-related illness. According to a Finnish survey, 25% of mens' and 32% of womens' occupational health primary care visits were work-related [bib_ref] Primary care visits to occupational health physicians and nurses in Finland, Kimanen [/bib_ref].
Therefore, accurately recording the work-relatedness of each visit and the patient's potential risk of disability is important. While recording work-relatedness is standard practice across OHS in Finland, to date no studies have been published on assessing this process, nor has there been any study of how well physician records match true risks, and how well OHS follow up patients and conduct interventions to prevent work disability after the employee's visit.
We hypothesized that clearer recording of workrelatedness at primary care visits and systematic recording of the work disability risk of individual patients, with systematic follow up and initiation of intervention to mitigate risks, could help to reduce work disability. The aim of this study was to evaluate an intervention designed to improve recording and follow up of OHS primary care visits and its impact on sickness absences.
# Methods
The intervention protocol was reported in full elsewhere (see [bib_ref] Impact of improved recording of workrelatedness in primary care visits at occupational..., Atkins [/bib_ref].
## Setting
The study was conducted in Pihlajalinna Työterveys, a large private OHS provider, which at the time of starting the study had 28 private healthcare units across Finland. Pihlajalinna Työterveys had approximately 68,370 employees on their register in 2015. The organisation went through several rounds of mergers and corporate acquisitions during the study period, which led to a substantial increase in patient and healthcare unit numbers.
## The intervention
The intervention was multifaceted and implemented sequentially. First, a notice was sent to the entire organisation informing all practitioners that the study would be conducted. The intervention consisted of two separate activities: one involved training, mentoring and follow up of physicians in intervention units on how to identify and record work-related illnesses during primary care visits and how to identify and record risk of work disability. The training sessions were conducted at each intervention unit. During the sessions the intervention and its components were introduced, and information about work-related illnesses was reinforced. This also included training on the intervention processes -actions that were to be initiated after a patient was identified during a visit as at risk of work disability. Following the training, a project physician responded to questions and followed up with training participants by telephone. Second, an organisation-wide change was made to the electronic healthcare system, clarifying the way in which work-relatedness and risk of disability pension was recorded. This change reinforced the messages given in the training. The electronic change was made to clarify language in sections where work disability risk was assessed. No specific training was conducted on the change to the electronic record system, as changes were minor and had been introduced in the intervention training. Therefore, intervention physicians were more likely to adhere to the change in the electronic health record, as they had been trained.
The trainees in the intervention sites were those occupational health physicians who were responsible for collaboration with their own patient companies, working at any of the 22 sites included in the study. The intervention and control sites had 58 physicians and 50 physicians, respectively, employed during the study. These physicians would be responsible for contacting workplaces, and would be involved in tailoring patients' work tasks or conducting other workplace interventions.
If the occupation health (OH) physician noted a patient's visit as being related to work or that the patient presented with a condition that could potentially result in work disability in the near future, they marked this onto an electronic system. Following this, a sequence of events was kicked off at the intervention sites. The OHS nurses responsible for the employer organisations used the electronic health records to collect the patients identified as at risk. Together with the physicians they then initiated the interventions that the physician recommended, either for the patients or more widely at the workplace. These interventions could include, for example, an occupational health collaborative negotiation to modify the employee's work tasks or timing, or organisational interventions focused on workplace ergonomics or teamwork counselling. Other interventions could include, for example, starting medical or vocational rehabilitation for the individual patient, involving both the workplace and the patient/employee. It was not possible to collect the number of these interventions in this study, since the individual patients that were identified as at risk of work disability or that had a work-related condition could not be associated with the interventions conducted at workplace level. A fuller description of the intervention can be found in the Template for intervention description and replication (TIDIER) reporting guide for population health interventions: tidierguide. org/#/gen/pFqrFqw3M
Information about the study was sent to all sites in April 2016. The intervention training was conducted in May 2016. The electronic change to systems was implemented on 9 March 2017. Data collection ended in December 2017.
## Randomization
We included 22 Pihlajalinna Työterveys clinics that were functional in 2016 in the study. We treated each healthcare unit as a cluster, as individual randomization in this context would have been challenging. NT, the team statistician, conducted initial simple randomization to randomize the first four clusters. After this, we used the minimization approach to randomize the remaining 18 clusters so that the following confounders were balanced across the intervention and control sites (see [bib_ref] Impact of improved recording of workrelatedness in primary care visits at occupational..., Atkins [/bib_ref] : (1) the occupational sector (e.g. industrial, service sector, public service), (2) presence of a large industry patient and (3) patient volume per site. The occupational health professionals and the research team were not blinded to the intervention.
## Outcomes
Our primary outcome was reduction of the mean number of medium-term (4-14 calendar days) of sickness absences per intervention and control centre from baseline after 1 year of follow up as measured from information in OHS patient records [bib_ref] Impact of improved recording of workrelatedness in primary care visits at occupational..., Atkins [/bib_ref]. We considered medium-length sickness absences from 4 to 14 calendar days, instead of 9 working days as indicated in the original protocol. We chose this to match our findings more closely with the Finnish Insurance Agency's definitions for sickness absences, which considers mediumlength sickness absences as absences including 9 working days (including Saturdays but not Sundays). The patient records included also weekends as sickness absence days, which differed from this approach. A similar choice was made to that in a previous report [bib_ref] The durations of past sickness absences predict future absence episodes, Laaksonen [/bib_ref].
Our secondary outcomes were:
1. Reduction in the mean number of short-term (1-3 consecutive calendar days) sickness absences from the workplace per cluster after 1 year from the start of the intervention as measured either by selfreported sickness absences recorded on the OHS system or sickness absences certified by OHS physicians working at the OHS units included 2. Reduction in the mean number of any form of work disability pensions as measured by an employee registered on the central pensions register as receiving a work disability pension from baseline to up to 2 years from the intervention 3. Reduction in the mean number of long-term more consecutive calendar days) sickness absences from the workplace per cluster from baseline to 1 year after the intervention as measured by OHS records
The follow-up time was set at 1 year due to funding and the planned study duration. This article focuses on reporting the primary outcome, medium-term sickness absences, as we deemed the follow-up period too short to report on disability pensions or longterm sickness absences. We also report short-term sickness absences and the process indicators collected on recording work relatedness and risk of work disability at consultations across the control and intervention sites.
## Power calculation
Our initial power calculation suggested that we would have 91% power to detect a 10% change in mean sickness absence rates across the intervention and control clusters, if we had 22 occupational health units with 24, 892 patients. For the trial, we retained all 22 units, with 26,804 patients recorded on the system.
## Data collection
We collected medical record data on patients' healthcare consultations at Pihlajalinna Työterveys from 2015 to 2017. The medical records included 68,370 patients in 2015 and 107,413 patients in 2017. The cohort was dynamic in that patients could be added to the cohort as the study progressed. Data were pseudonymised, and researchers had no access to patient-identifying data. All patients above the age of 18 years and whose employers had a contract with Pihlajalinna Työterveys including primary healthcare services were included in the study. The data were combined with pseudonymised data from the Finnish Centre for Pensions, where we obtained data on all participants' pensions granted for the study period.
# Data analysis
After data collection was complete, we noted that Pihlajalinna's acquisition of another large occupational health services provider impacted on our outcomes. Therefore, we used all initially randomized sites in the intention-totreat analyses and excluded them in the per-protocol analyses.
We included data on curative patient visits to OHS physicians responsible for patient organisations. OHS services have many casual workers, who deal with primary care patients but are not occupational health specialists and most of them were not exposed to training. We also excluded preventive visits such as health examinations. We analysed data 6 months before the intervention, during the intervention, and for 6 months after the intervention. After initial analysis we chose a period of 6 months after the intervention corresponding with the same season of the 6 months preceding the intervention, to ensure that seasonal effects did not confound our analysis. We analysed data using analysis of covariance (ANCOVA), setting alpha at 0.05.
We also analysed process indicators among intervention and control clinics. These indicators included whether physicians had changed their practice of recording relatedness to work, during the consultation after the intervention. The intervention required a physician to record whether or not the patient's visit was related to work or whether this was not assessed. We analysed changes after the educational intervention and after the change in the electronic system, using descriptive statistics.
# Results
The flowchart [fig_ref] Figure 1: Flowchart of the trial randomization [/fig_ref] presents the final data after randomization, divided by sex. [fig_ref] Figure 1: Flowchart of the trial randomization [/fig_ref] Flowchart for the intention-to-treat analysis (ITT). OHS, occupational health services
The baseline characteristics of the study population are shown in [fig_ref] Table 1: Baseline characteristics of intervention and control groups by sex [/fig_ref].
There were differences between women at the intervention and control sites in terms of age, proportion of registered employees visiting without sick leave, total number of visits, and medium-term and short-term sickness absences. In men, only age, visit without sick leave and short-term sick leave differed across the intervention and control sites [fig_ref] Table 1: Baseline characteristics of intervention and control groups by sex [/fig_ref].
The results of our primary outcome analysis are shown in [fig_ref] Table 2: Intention-to-treat analysis, sickness absences before and after the intervention [/fig_ref]. As can be seen, the intervention had no significant effect on short-term, long-term or medium-term sickness absences in men or women.
Short-term and long-term sickness absences decreased among men and increased among women, though none of these changes were statistically significant. The perprotocol analysis, excluding entire occupational health units, had similar results.
Our analysis of process indicators that measured how intervention and control groups recorded patient visits in practice was more promising. [fig_ref] Table 3: Process indicators [/fig_ref] shows change in physicians' practice, at baseline, after education, after the electronic information system change and 6 months after the intervention. [fig_ref] Table 3: Process indicators [/fig_ref] shows that before the intervention most visits were recorded as "not related to work", which was the default setting (89% and 85% across control and intervention units, respectively). After the institutional information presented and education conducted at intervention units, a change was observed where 75% of intervention units' records and 38% of control units' records stated "not related to work". At the same time, the rates of "not assessed" increased in both units, more in the control units (50%) than in the intervention units (9%).
However, after the electronic reminder in the system changed and the default setting changed to "not assessed" from "not related to work", we can see that while the control sites' default answers increased (from 50% to 61%), the intervention sites' default answers stayed nearly the same (from 9 to 10%), suggesting that intervention sites' recordings were actual recordings made actively by physicians more than default choices. These effects were sustained over time. As the physicians' recording improved, we can see that the percentage of visits related to work also increased, from 13% in the beginning to 15% at the end. Trends in recording possible work disability were similar across intervention and control sites. Physicians recorded similar numbers of possible future work disability for each consultation across intervention and control sites. There were slightly more records of no threat of disability at the intervention sites than there were at control sites.
# Discussion
Though our intervention showed no effect on sickness absences, it produced a promising indication of the effectiveness of education in improving occupational health professionals' practices of recording work-related visits in primary care. This effect was supported by a change in electronic information systems.
While there was no statistical difference between the intervention and control arms in the rates of sickness absence as primary and secondary outcomes, there may be a number of reasons for the lack of differences. First, while approximately 15% of visits were identified as work-related, these form a relatively small subset of the Including only those with sick leave (control group, n = 4990; intervention group, n = 5668) entire population analysed for detecting a difference in sickness absences. At the individual level, actions leading to shortening of sickness absences take time, as rehabilitative processes are gradual. In addition, initiating individual work modifications in collaboration with workplaces requires time. Second, many of the conditions that are work-related require workplace interventions starting with including workplace assessment, with subsequent commitment by employers to implement these changes. An example of such intervention could be improving workplace psychological wellbeing [bib_ref] Improving employee well-being and effectiveness: systematic review and meta-analysis of web-based psychological..., Carolan [/bib_ref] , or changing the workplace environment, for example lighting [bib_ref] Workplace lighting for improving alertness and mood in daytime workers, Pachito [/bib_ref] or disruptions from open plan offices [bib_ref] Open-plan office noise: The susceptibility and suitability of different cognitive tasks for..., Jahncke [/bib_ref]. These interventions are considerable commitments for organisations both in terms of processes and financially, and may take time to be implemented. In order for our intervention to impact on these, it should have had a workplace outreach component. The increase in sickness absences among women can be related to increasing age over time [bib_ref] Explanations for gender differences in sickness absence: evidence from middle-aged municipal employees..., Laaksonen [/bib_ref] but also to poor interpersonal relations at the workplace [bib_ref] Workplace bullying and sickness absence: a systematic review and meta-analysis of the..., Nielsen [/bib_ref]. Our linked study on frequent attenders in occupational health services similarly identified women as at risk of frequent use of services [bib_ref] Frequent attenders in occupational health primary care: a cross-sectional study, Reho [/bib_ref] , where particularly women from the service industry and public administration were at risk [bib_ref] Frequent attenders in occupational health primary care: a cross-sectional study, Reho [/bib_ref]. We also found that frequent attenders of OHS primary care are also at increased risk of sickness absences after their consultation frequency has diminished [bib_ref] Occasional and persistent frequent attenders and sickness absences in occupational health primary..., Reho [/bib_ref]. This supports the lack of impact on sickness absences, when no workplace intervention was included. Finally, a possible reason for lack of impact is changes in national rules for sickness certification, where employees could be absent from work without a certificate for a longer period (from 3 days to 7 days), which was implemented by many businesses and public organisationsduring the study period. These, and other changes in sickness certification over time are more likely to impact on sickness absence rates than our intervention. The study sites also experienced relatively high physician turnaround. This means that our intervention might not have reached all practising physicians in the intervention sites. This suggests a need for a continuous education approach in future interventions.
Our study has several strengths and limitations. As the Pihlajalinna patient register has a large, nationwide sample representing different industries, we can consider our sample generalizable to the working age population in Finland. However, our pragmatic trial approach meant that we could not control the fidelity with which physicians adhered to the educational programme, nor were we able to determine which activities increased at the workplaces after the intervention. We also could not prevent crossover of physicians from the intervention to the control arm. Nevertheless, conducting such trials using routine patient registers allowed us to evaluate the outcome of the intervention with a large sample of highquality data.
Despite the intervention not impacting on patients' sickness absences, the impact of the educational intervention is promising. Early identification of patients with work disability risk enables timely follow up by the OH team and early intervention for issues that might threaten ability to work. This can possibly improve continuity of care in primary healthcare settings. OHS physicians are seen as better positioned to evaluate the need for sickness absences than general practitioners working in other settings [bib_ref] Occupational health physicians have better work conditions for handling sickness certification compared..., Ljungquist [/bib_ref] , and early consultation with an OH physician has been found effective in reducing the total number of sickness absence days taken by individuals who are at risk of sickness absences [bib_ref] Structured early consultation with the occupational physician reduces sickness absence among office..., Kant [/bib_ref]. With a simple educational intervention combined with an electronic reminder, data indicate that occupational health physicians in 11 intervention clinics changed their practice of recording work-relatedness and potential for work disability. These effects were sustained after the intervention was concluded. As recording in electronic systems is a challenge and poorly functioning electronic referral systems can even result in occupational stress [bib_ref] Predictors of physicians' stress related to information systems: A nine-year follow-up survey..., Heponiemi [/bib_ref] , this is a positive and fairly surprising outcome. Clinicians may feel that electronic health records impact negatively on their professional satisfaction [bib_ref] Strategies for primary care stakeholders to improve electronic health records (EHRs), Olayiwola [/bib_ref] , therefore the systems need to be both meaningful and easy to use. This simple intervention succeeded in improving the accuracy and frequency of recording. This better and more accurate recording can enable better follow up, interventions and assessment. Further training can reinforce this message. While recording of workrelatedness at consultations or the patient's risk of work disability itself does not translate into reduced sickness absences, there is a possibility that improved recording can result in better reporting to employers, and better and timelier opportunities for preventive actions in the workplace and for patients.
# Conclusion
Our cluster, pragmatic, randomized controlled trial using patient registers as data did not find that an educational and electronic health information intervention had significant effects on sickness absences in the context of occupational health primary care in Finland. However, the intervention changed occupational health physicians' practice of recording the work-relatedness of patient consultations, and potentially enabled better continuity of care and follow up for patients at risk of needing disability pensions. In future, such interventions should include detailed follow up of patients, with a workplace component to ensure adequate follow up of and intervention for patients at risk.
[fig] Figure 1: Flowchart of the trial randomization [/fig]
[table] Table 1: Baseline characteristics of intervention and control groups by sex: mean (standard deviation) or percentage (%) within group [/table]
[table] Table 2: Intention-to-treat analysis, sickness absences before and after the intervention (n = 22) [/table]
[table] Table 3: Process indicators: physician registration of work-relatedness of each patient visit [/table]
|
Numerous protein-bound solutes are cleared by the kidney with high efficiency
The kidney clears numerous solutes from the plasma; however, retention of these solutes causes uremic illness when the kidneys fail. We know remarkably little about which retained solutes are toxic and this limits our ability to improve dialysis therapies. To explore this we employed untargeted mass spectrometry to identify solutes that are efficiently cleared by the kidney. High resolution mass spectrometry detected 1808 features in the urine and plasma ultrafiltrate of 5 individuals with normal renal function. The estimated clearance rates of 1082 peaks were greater than the creatinine clearance indicating tubular secretion. Further analysis identified 90 features representing solutes with estimated clearance rates greater than the renal plasma flow. Quantitative mass spectrometry with stable isotope dilution confirmed that efficient clearance of these solutes is made possible by the combination of binding to plasma proteins and tubular secretion. Tandem mass spectrometry established the chemical identity of 13 solutes including hippuric acid, indoxyl sulfate, and p-cresol sulfate. These 13 efficiently cleared solutes were found to accumulate in the plasma of hemodialysis patients, with free levels rising to more than 20-fold normal for all but two of them. Thus, further analysis of solutes efficiently cleared by secretion in the native kidney may provide a potential route to the identification of uremic toxins.
# Introduction
The kidney removes numerous waste solutes from the blood plasma. When kidney function is lost, these solutes accumulate in the body and cause uremic illness culminating in death unless renal function is partially replaced by dialysis. At present we know remarkably little about which retained solutes are toxic, and this lack of knowledge limits our ability to improve treatment [bib_ref] Dialysis cannot be dosed, Meyer [/bib_ref]. Progress has been slow in part because the number of solutes retained when the kidneys fail is very large [bib_ref] Update of uremic toxin research by mass spectrometry, Niwa [/bib_ref] [bib_ref] Metabolite profiling identifies markers of uremia, Rhee [/bib_ref] [bib_ref] Metabolomic analysis of human plasma from haemodialysis patients, Sato [/bib_ref]. The current study employed untargeted mass spectrometry to find solutes which are efficiently removed from the plasma by the kidney. It revealed that there are many waste solutes for which renal clearance rates normally exceed the renal plasma flow. Such high clearances require a combination of binding to plasma proteins and active tubular secretion. Concentrations of such solutes can rise to high levels when renal function is replaced by hemodialysis which clears solutes only by diffusion. Presuming that evolution has provided for the kidney to remove toxic substances efficiently, identification of solutes with high renal clearance rates could provide a route to the identification of uremic toxins.
# Results
Measurements were made in four men and one woman with normal renal function as reflected by an average creatinine clearance of 142±22 ml/min/1.73 m 2 . A total of 1808 features were detected in both urine and plasma ultrafiltrate by untargeted high resolution mass spectrometry. Clearance rates for these features were estimated as the urinary excretion rate divided by the concentration in plasma ultrafiltrate. Clearance values are thus expressed in terms of the "free" unbound solute concentration in plasma rather than the total solute concentration. The distribution of estimated clearance rates relative to the creatinine clearance is depicted in [fig_ref] Figure 1: Figure 1. [/fig_ref]. For 1082 features, estimated clearance rates were greater than the clearance of creatinine with a false discovery rate of q < 0.05. Because the creatinine clearance is slightly higher than the glomerular filtration rate, these features were considered likely to represent solutes secreted by the renal tubules. There were in contrast only 290 features with estimated clearance rates less than the creatinine clearance with a false discovery rate of q < 0.05.
For 163 features, estimated clearance rates were greater than seven-fold the clearance of creatinine. This suggested that their clearances exceeded the renal plasma flow, which is approximately four-fold the clearance of creatinine. Among these 163 features, 90 were considered to represent unique chemical compounds after elimination of duplicates and features considered to represent dimers, adducts, isotopes, or artifacts on manual review of chromatograms [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref]. Compounds with matching mass values were sought in standard databases and the chemical identities of 13 of these 90 features were established by comparison of their chromatographic retention times and MS/MS spectra with those of reagent standards [fig_ref] Table 1: Studies were performed in accord with the Declaration of Helsinki [/fig_ref]. For 40 of the 90 features of interest, however, no candidate compounds were found among known human metabolites with mass within 3 parts per million (ppm) (Supplementary [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref] [bib_ref] METLIN: a metabolite mass spectral database, Smith [/bib_ref]. Clearance values in excess of the renal plasma flow rate were possible because the "free" concentrations of these solutes in plasma ultrafiltrate were lower than their total plasma concentrations, presumably reflecting binding to plasma proteins. While all of the efficiently cleared solutes were protein-bound, the extent of binding varied widely, with the free fraction ranging from 2 to 52 percent of the total plasma concentration. Calculation in terms of total plasma concentration yielded much lower clearance values for the bound solutes [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref].
Measurements using LC/MS/MS with isotopically labeled standards confirmed the finding of very high clearance rates for the bound solutes hippurate, indoxyl sulfate, and p-cresol sulfate. Clearance values for these solutes along with urea and creatinine are summarized in [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref]. For comparison, clearance values were also measured for phenylacetylglutamine which previous studies had shown to be secreted by the renal tubules but largely unbound [bib_ref] Phenylacetylglutamine and hippuric acid in uremic and healthy subjects, Zimmerman [/bib_ref]. As expected, the clearance of urea was less than the creatinine clearance, reflecting tubular reabsorption after glomerular filtration. The clearance of phenylacetylglutamine in contrast averaged 455±62 ml/min/1.73m 2 , or approximately three quarters of the estimated renal plasma flow rate. Clearance values for hippurate, indoxyl sulfate, and p-cresol sulfate were much higher. The values obtained by quantitative assay with labeled standards were slightly lower than those estimated from peak areas assessed by untargeted mass spectrometry but still well above the estimated renal plasma flow. Clearance values for these solutes expressed in terms of the total plasma concentration were much lower than those expressed in terms of the free concentration and did not exceed the renal plasma flow.
Additional studies examined the accumulation in patients with renal failure of solutes found to be efficiently cleared by the native kidney. As summarized in [fig_ref] Table 1: Studies were performed in accord with the Declaration of Helsinki [/fig_ref] , free levels of all of the 13 chemically identified solutes with high native kidney clearance rates were elevated in hemodialysis patients. Of note, for all but two of these solutes, the average free concentration in pre-treatment samples from hemodialysis patients were more than 20-fold normal, and the free solute concentrations rose higher than the total solutes concentrations. The great majority of solutes characterized only by exact mass values which have high native kidney clearances were also found to accumulate in hemodialysis patients, as further summarized in [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref].
# Discussion
Metabolomic studies have found that the urine contains hundreds of solutes, the majority of which remain to be chemically identified [bib_ref] Ion-pairing reversed-phase liquid chromatography fractionation in combination with isotope labeling reversed-phase liquid..., Guo [/bib_ref] [bib_ref] Annotation of the human adult urinary metabolome and metabolite identification using ultra..., Roux [/bib_ref] [bib_ref] Evaluation of coupling reversed phase, aqueous normal phase, and hydrophilic interaction liquid..., Zhang [/bib_ref]. Presumably, most of these solutes are cleared by the kidney from the plasma, and their accumulation in the body as uremic solutes may contribute to illness when renal function is reduced. At present, however, we have little knowledge of which of these solutes are clinically important [bib_ref] Normal and pathologic concentrations of uremic toxins, Duranton [/bib_ref].
The current study was designed to identify solutes for which renal clearance rates are normally very high. For any rate of solute production, a high clearance serves to keep the solute level in the body low. We could thus expect to find toxic waste compounds among those solutes with high clearance rates. Analysis of timed urine and plasma ultrafiltrate samples by untargeted mass spectrometry revealed the presence of 90 features considered likely to correspond to solutes with renal clearance rates greater than the renal plasma flow. Of note, compounds with corresponding mass values for many of these features could not be found in standard lists of human metabolites [bib_ref] METLIN: a metabolite mass spectral database, Smith [/bib_ref]. This suggests that many substances for which evolution has provided high clearance rates remain to be identified. The compounds we were able to identify included indoxyl sulfate and p-cresol sulfate. These compounds derive from the action of gut bacteria and are known to be secreted by the renal tubules. They accumulate in the plasma when the kidneys fail and have recently received extensive consideration as uremic toxins [bib_ref] Metabolomics analysis reveals large effects of gut microflora on mammalian blood metabolites, Wikoff [/bib_ref] [bib_ref] Uremic toxins originating from colonic microbial metabolism, Evenepoel [/bib_ref] [bib_ref] The gut: the forgotten organ in uremia?, Schepers [/bib_ref] [bib_ref] Uremic solutes from colon microbes, Meyer [/bib_ref]. The other compounds identified comprised three dicarboxylic acids, six acyl glycines, and two substituted purine metabolites. All these substances have previously been found in human urine and hippurate, adipic acid, and cinnamoylglycine have also been reported to accumulate in the plasma of patients with renal failure (http://www.hmdb.ca) [bib_ref] Normal and pathologic concentrations of uremic toxins, Duranton [/bib_ref] [bib_ref] Colonic contribution to uremic solutes, Aronov [/bib_ref]. Like indoxyl sulfate and p-cresol sulfate, hippurate and 3-hydroxy hippurate have been shown to be secreted by the renal tubules [bib_ref] The Renal Clearances of Substituted Hippuric Acid Derivatives and Other Aromatic Acids..., Smith [/bib_ref]. But as far as we can discover, the possibility of a renal clearance exceeding the renal plasma flow has not been considered for any of these solutes.
As described here, the kidney can achieve clearance rates in excess of the renal plasma flow through a combination of tubular secretion and rapidly reversible solute binding to plasma proteins. For solutes which are confined to the plasma and not protein-bound, the maximum clearance is equal to the renal plasma flow rate. But for bound solutes, active secretion lowers the free plasma solute concentration in blood passing through the peritubular capillaries after it leaves the glomeruli so that the bound portion of the solute tends to dissociate from the binding proteins and becomes available for secretion. If the avidity of secretory transport into the tubular lumen is sufficient, the amount of solute secreted is greater than the renal plasma flow multiplied by the free concentration in the plasma entering the peritubular capillaries, and the clearance will rise above the plasma flow rate. The net effect is to reduce the unbound solute concentration in the systemic circulation to a lower level than would be achieved if the solute were completely removed from the plasma passing through the kidneys but were not protein-bound. Since it is the free, unbound concentration of solutes to which tissues throughout the body are exposed, the combination of protein-binding and tubular secretion can provide adaptive advantage in the removal of toxic waste compounds. In essence, the addition of reversible protein-binding to secretion allows the free level of a solute to be reduced without increasing kidney blood flow and size.
The ability of the kidney to reduce the free plasma concentration of bound solutes to very low levels was recognized by E. K. Marshall [bib_ref] Two lectures on renal physiology, Marshall [/bib_ref] who provided the first unequivocal demonstration of solute secretion by the renal tubules. Marshall did not measure any natural solutes but hypothesized their potential high clearance based on observation of the renal handling of phenol red, a protein-bound dye. Since then, however, the potential advantage of protein binding has been largely ignored in renal medicine, and clearance values for bound as well as unbound waste solutes have been expressed in terms of the total rather than the more relevant free plasma concentration. Standard practice has been different, however, in the pharmacology literature [bib_ref] Changes in plasma protein binding have little clinical relevance, Benet [/bib_ref]. With pharmaceutical agents as with other compounds, only the free portion of a bound solute is biologically active, and clearance rates for pharmaceuticals have therefore routinely been expressed in terms of their free plasma concentrations. Calculation in terms of the total concentration yields lower values and underestimates the body's ability to limit the effective, free solute level as revealed in [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref].
An obvious question for renal medicine is whether secretion can persist when glomerular filtration declines. The hope based on early morphologic observations that significant numbers of "aglomerular tubules" continue to operate in patients with glomerular disease was largely disappointed by subsequent analyses [bib_ref] Tubular injury in glomerular disease, Meyer [/bib_ref] evidence is most extensive for the clearance of para-aminohippurate, long used as a measure of renal plasma flow. On average, the clearance of para-aminohippurate declines only slightly less than the GFR, with variation among individual patients and specific diseases [bib_ref] Renal Function in Chronic Renal Disease, Bricker [/bib_ref]. A decline in secretory clearance with GFR has also been demonstrated for endogenous organic anion such as hippurate and 5-hydroxyindolacetate and for many pharmaceuticals including protein-bound compounds like furosemide [bib_ref] Plasma hippurate in renal failure: high-performance liquid chromatography method and clinical application, Igarashi [/bib_ref] [bib_ref] Plasma 5-hydroxyindoleacetic acid as an endogenous index of renal plasma flow, Hannedouche [/bib_ref] [bib_ref] Comparison of loop diuretics in patients with chronic renal insufficiency, Voelker [/bib_ref]. Similar reduction in the average secretory clearance of the tightly bound endogenous solutes indoxyl sulfate and p-cresol sulfate in parallel with eGFR have also recently been described in abstract form (Meijers B, Viaene L, Poesen R; Estimated glomerular filtration rate is a good marker of renal clearance for indoxyl sulfate and p-cresyl sulfate. PO384, ASN Kidney
Week 2012). Rising plasma concentrations as chronic kidney disease progresses thus do not serve to distinguish compounds which are cleared by secretion from those that are cleared largely by filtration. More subtle questions, such as the extent to which levels of secreted solute are affected by competition for transport molecules and altered expression of transport molecules during renal disease progression, have been less thoroughly studied.
A weakness of untargeted mass spectrometry is that it does not provide precise quantitation. Matrix effects alter the strength of ion signals from individual samples, and particularly from samples of different fluids. The clearance values in [fig_ref] Table 1: Studies were performed in accord with the Declaration of Helsinki [/fig_ref] which are based on relative peak areas in chromatograms of urine and plasma ultrafiltrate must thus be regarded as only approximate. More accurate quantitative LC/MS/MS assays were developed to measure concentrations of selected solutes for which we were able to obtain both unlabeled and isotopically labeled standards. These measurements confirmed that the clearances for hippurate, p-cresol sulfate, and indoxyl sulfate were much higher than the estimated renal plasma flow. Their clearances can be contrasted with that of phenylacetylglutamine, a solute which is actively secreted but largely unbound so that its clearance can only approach the renal plasma flow. Indoxyl sulfate, p-cresol sulfate, 3-hydroxyhippurate, 1,3,7-trimethyluric acid and 1,7-dimethyluric acid are known to be handled by the organic acid transporters OAT1 and/or OAT3 in the renal proximal tubule, and the structures of the other solutes we identified as efficiently cleared make them likely candidates for transport by the same mechanisms [bib_ref] Interactions of human organic anion as well as cation transporters with indoxyl..., Enomoto [/bib_ref] [bib_ref] Structure-affinity relationship in the interactions of human organic anion transporter 1 with..., Sugawara [/bib_ref] [bib_ref] Organic anion transporter family: current knowledge, Anzai [/bib_ref] [bib_ref] Untargeted metabolomics identifies enterobiome metabolites and putative uremic toxins as substrates of..., Wikoff [/bib_ref]. Identification of which transporters handle which solutes, however, will require studies in cultured cells or animals in which transporter activity has been genetically manipulated.
Confirmation that individual solutes are cleared by the kidney is provided by the finding of high levels in patients with renal failure. Among the compounds we identified as having native kidney clearances greater than the renal plasma flow, the total plasma concentrations of hippurate, p-cresol sulfate, indoxyl sulfate, adipic acid and cinnamoylglycine have previously been reported elevated in patients with renal failure [fig_ref] Figure 1: Figure 1. [/fig_ref]. The current study shows further that free levels of bound solutes which are efficiently cleared by the native kidney can rise to high levels in patients maintained on hemodialysis which provides clearance by passive diffusion. Among the 13 compounds we identified as having high native kidney clearances, free levels for all but two were more than 20-fold normal in hemodialysis patients. The exceptions, 1,3,7-trimethyluric acid and 1,7-dimethyluric acid are caffeine metabolites. All of our normal subjects were coffee or tea drinkers, while intake among the hemodialysis patients was restricted to a single daily cup of tea in one individual. In addition to differences in production, the non-renal clearance and volume of distribution of individual solutes can influence the degree to which their levels are elevated in dialysis patients.
Several limitations should be acknowledged. First, our list of features corresponding to solutes with high renal clearance rates is undoubtedly incomplete. Previous studies have shown that the number of features identified by mass spectrometry increases when samples are assayed using multiple chromatographic and ionization methods [bib_ref] Annotation of the human adult urinary metabolome and metabolite identification using ultra..., Roux [/bib_ref] [bib_ref] Evaluation of coupling reversed phase, aqueous normal phase, and hydrophilic interaction liquid..., Zhang [/bib_ref] [bib_ref] The human serum metabolome, Psychogios [/bib_ref] [bib_ref] Multiple ionization mass spectrometry strategy used to reveal the complexity of metabolomics, Nordstrom [/bib_ref]. Second, the current approach would fail to identify compounds that are efficiently removed from the plasma but then either degraded in the kidney or altered before excretion into the urine. Finally, we may have calculated falsely high clearance values for substances that are excreted in the urine following production in the kidney.
In summary, the current study shows that the combination of protein-binding with tubular secretion allows the kidney to clear waste solutes at rates exceeding the renal plasma flow. And it suggests that such high clearance rates are achieved for a large number of natural solutes, many of which remain to be chemically identified. Further analysis of the group of solutes efficiently cleared by the kidney could provide a route to identification of these uremic toxins.
# Methods
Blood samples were obtained at the midpoint of timed urine collections following an overnight fast in five subjects with normal renal function whose characteristics are noted in Supplementary intended to allow sub 1 ppm accuracies in the mass range of the majority of solutes excreted in the urine and thereby decrease the number of possible elemental compositions. To minimize effects of instrument drift, samples from each subject were run together in triplicate and samples from all subjects were run first in negative and then positive mode.
MZmine software v2.2 (Okinawa Institute of Science and Technology) was used to identify features characterized by retention time and m/z from each LC/MS run and to assign amplitudes to these features based on integration of the ion current values [bib_ref] MZmine 2: modular framework for processing, visualizing, and analyzing mass spectrometry-based molecular..., Pluskal [/bib_ref]. Because our goal was to identify solutes cleared by the kidney, analysis was further restricted to the 754 negative ion and 1054 positive ion features with average peak area greater than 4,000 in the triplicate runs in the plasma ultrafiltrate of at least three of five subjects and also in the urine of at least three of five subjects. Clearance rates for these 1808 features were estimated by comparing the concentrations in urine and plasma ultrafiltrate to those of creatinine. For 1082 of the 1808 features the estimated clearance rates were greater than the creatinine clearance with a false discovery rate of q < 0.05. This was considered evidence of tubular secretion. Chromatograms for the 163 of the 1082 secreted features with clearance values more than seven-fold the creatinine clearance were examined manually. The cut-off of seven fold was chosen to be well above the renal plasma flow, which is about four-fold the creatinine clearance, and to include the feature identified as p-cresol sulfate for which a deuterated standard was available allowing its high renal clearance to be confirmed by a more quantitative assay (below). Elimination of duplicates when features appeared in both positive and negative mode and of features considered to represent dimers, adducts, isotopes, or artifacts on manual review reduced the total number of features with estimated clearance rates greater than seven-fold the creatinine clearance to 90 as summarized in [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref]. Compounds with m/z values corresponding to these features were identified using the Human Metabolome and Metlin Databases (http://www.hmdb.ca, http:// metlin.scripps.edu/) and chemical standards were obtained as further summarized in Supplementary [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref]. To confirm chemical identities, selected urine samples and chemical standards were run using the same LC method coupled to an LTQ Orbitrap Velos (Thermo Fisher) for features appearing with negative mode and to an LTQ XL ion trap (Thermo Fisher) for features appearing with positive mode. Chemical identify was established by match of retention time, principle ion mass, and MS/MS spectrum (Supplementary . Among the 13 chemically identified features, the average magnitude of the mass error assigned to peaks extracted by MZmine was 0.5±0.5 ppm, encouraging confidence that correct mass values were assigned to the unidentified features listed in [fig_ref] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/fig_ref].
Pretreatment plasma from 6 hemodialysis patients and from 6 normal subjects whose characteristics are summarized in [fig_ref] Table 1: Studies were performed in accord with the Declaration of Helsinki [/fig_ref] was also analyzed by high resolution mass spectrometry as described above. Plasma was deproteinized, dried, and reconstituted in 90:10 vol:vol water:acetonitrile to one fourth the original concentration for dialysis patients and to the original concentration for normal subjects. Ultrafiltrate was reconstituted in 90:10 water:acetonitrile to the original concentration for dialysis patients and ten fold the original concentration for normal subjects. Peaks areas for plasma and ultrafiltrate obtained with MZmine software were averaged excluding values of less than 4,000 in duplicate runs of each sample. Peaks corresponding to solutes previously identified as having very high renal clearance values were identified by matching retention time and mass. Peaks corresponding to the 13 chemically identified features were reintegrated manually using Xcalibur Software (ThermoFisher).
Concentrations of p-cresol sulfate, indoxyl sulfate, hippurate, and phenylacetylglutamine in the normal samples were further assayed by stable isotope dilution LC/MC/MS using pcresol-d 8 -sulfate (synthesized from p-cresol d 8 [bib_ref] Simplified method for the preparation of aromatic sulfuric acid esters, Feigenbaum [/bib_ref] (Cambridge Isotopes)), indoxyl-2,4,5,6,7-d 5 -sulfate (Isosciences), and N-benzoyl-d 5 -glycine and Nα-(phenyl-d 5acetyl)-L-glutamine (both C/D/N Isotopes) as internal standards. Preparation was the same except that dried samples were reconstituted in water to twice the concentration used for untargeted analysis.
## Statistics
Clearance rates for the 1808 features detected in urine and in plasma ultrafiltrate were compared to creatinine clearance rates in each subject by the Wilcoxon signed-rank test using SAS Enterprise Guide 4.3 and unpaired comparisons between values for hemodialysis patients and normal subjects were performed using the Mann-Whitney U test using SPSS V20. False discovery rates were then calculated using Q-VALUE (http:// genomics.princeton.edu/storeylab/qvalue/).
# Supplementary material
Refer to Web version on PubMed Central for supplementary material. The blue line represents the distribution of estimated clearance rates relative to the creatinine clearance for the 1,808 features found in the urine and plasma ultrafiltrate of normal subjects. The red triangles represent the 13 features for which identity was confirmed by analysis of reagent standards. Values are mean±sd. Clearance/clearance cre is the average ratio of solute clearance to creatinine clearance. Free fraction is the level in plasma ultrafiltrate as a percent of the total plasma level. Hemodialysis/Normal is the ratio of the average pre-treatment concentration in hemodialysis patients to the average concentration in normal subjects. a indicates the free fraction for furoylglycine was calculated in only one subject because peaks in plasma samples from other subjects were too small to quantify.
b indicates q < 0.05 for elevation of the solute concentration in hemodialysis patients above the level in normal subjects. Values are mean±sd. Clearance/clearance cre is the average ratio of solute clearance to creatinine clearance. Free fraction is the level in plasma ultrafiltrate as a percent of the total plasma level. Clearance total is the value that would be obtained if clearance were calculated using the plasma total concentration rather than the plasma free concentration.
[fig] Figure 1: Figure 1. [/fig]
[table] Table 1: Studies were performed in accord with the Declaration of Helsinki. [/table]
[table] Table 2: Clearance Values Obtained Using LC/MS/MS Assay [/table]
|
Abstract Profiles of Structural Stability Point to Universal Tendencies, Family-Specific Factors, and Ancient Connections between Languages
This Electronic Supplementary Material (S1) includes:- more information about the primary data used in this paper and its coding (Tables S1, S3, S4 and S15); - the relationships between the language family stability profiles (Tables S2 and S5, and Figures S1-S14); - the involvement of different features in sets maximizing the correlations between geographic and stability distances (Tables S6-S13); - the combined p-values (Tables S14 and S16); - the punctuated evolution of structural features (Tables S18 and S19, and Figures S15-S18); and - the R code implementing the five methods for combining p-values (Table S17).The order of these items largely reflects the structure of the paper's main text.Dediu & LevinsonAbstract profiles of structural stability: ESM (S1)Abstract profiles of structural stability: ESM (S1) 5/67 Figure S1: The observed Ripley's K function (computed at 100 lags) for the 4 datasets using BayesLang versus 10,000 simulated Poisson processes. It can be seen that the actually observed values (black) are well outside the 95% confidence interval around the mean randomizations (and, in fact, outside their whole range) for all meaningful values of the radius r. Abstract profiles of structural stability: ESM (S1) 6/67 Figure S2: The observed Ripley's K function (computed at 100 lags) for the 4 datasets using MrBayes versus 10,000 simulated Poisson processes. It can be seen that the actually observed values (black) are well outside the 95% confidence interval around the mean randomizations (and, in fact, outside their whole range) for all meaningful values of the radius r. Abstract profiles of structural stability: ESM (S1) 7/67Figure S3: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset MBE. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 8/67 Figure S4: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset MBW. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 9/67 Figure S5: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset MBH. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 10/67 Figure S6: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset MPE. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 11/67 Figure S7: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset MPW. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 12/67 Figure S8: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset MPH. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 13/67 Figure S9: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset BBE. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 14/67 Figure S10: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset BBW. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 15/67 Figure S11: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset BBH. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 16/67 Figure S12: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset BPE. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 17/67 Figure S13: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset BPW. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. Abstract profiles of structural stability: ESM (S1) 18/67 Figure S14: MDS (left) and annotated Network (right) representations of the stability distances between language families for dataset BPH. Please note that for the MDS plot only the first two dimensions are shown and the scales and directionality of the axes are arbitrary. ,(amh,tig)), ,hau,kna),ngi),ker,(mrt,ttr)),(bej,irk,(hae,som)),tzm) 17 BBH (((tzm,shi),thv),(((gde,ttr,mrt),bdm),(lln,ker), ,(anc,sur)),(ngi,mkf))),(awn,(hae,aar,som),bej,(bds,irk)),cop,((aij,((ary,arz,apc,afb),heb)),(tig,amh))) 32 BBW ((amh,arz,heb,tig),(anc,hau,kna,sur,ngi),bej,tzm,irk,ker,(mrt,ttr),(hae,som)) 17 BPE (((arz,heb),(amh,tig)),(bej,irk,(hae,som)),(tzm,shi),(hau,ker,mrt)) 13 BPH (((tzm,shi),thv),((gde,ttr,mrt),(lln,ker),(hau,((kna,pip),(anc,sur)),(ngi,mkf))),((hae,som),bej,(bds,irk)),cop,((((ary,arz,apc,afb,mlt),heb)),(tig,amh))) 29 BPW ((amh,arz,heb,tig),bej,tzm,mdx,(hau,kna,sur,ngi),irk,ker,(mrt,ttr),(hae,som)) 17
## Mbe
(aiw,((arz,heb),(amh,tig)),((((anc,sur),hau,kna),ngi),ker,(mrt,ttr)),(bej,irk,(hae,som)),tzm) 18 MBH (((tzm,shi),thv),(((gde,ttr,mrt),bdm),(lln,ker),(hau,((kna,pip),(anc,sur)),(ngi,mkf))),(awn,(hae,aar,som),bej,(bds,irk)),cop,((aij,((ary,arz,apc,afb),heb)),(tig,amh))) 32 MBW ((amh,arz,heb,tig),(anc,hau,kna,sur,ngi),bej,tzm,irk,ker,(mrt,ttr),(hae,som)) 17
## Mpe
((aiw,mdx),((arz,heb),(amh,tig)),(bej,irk,(hae,som)),tzm,(((hau,kna,sur),ngi),ker,(mrt,ttr))) 18 MPH (((tzm,shi),thv),((gde,ttr,mrt),(lln,ker),(hau,((kna,pip),(anc,sur)),(ngi,mkf))),((hae,som),bej,(bds,irk)),cop,((((ary,arz,afb,mlt),heb)),(tig,amh))) 28 Arawakan BBE (((apu,cni),(guc,rgr))) 4
BBH ,(rgr,tae,bae,gae,ycn)),(ign,cni,apu),ame)) 13 BBW (apu,cni,guc,rgr) 4 BPE (apu,cni) 2 BPH ((plu,((guc,arw,cab),(rgr,gae)),(ign,cni,apu),ame)) 10 BPW (apu,cni,guc,rgr) 4 MBE ((((aca,rgr),guc),(apu,cni))) 5
MBH ,(rgr,tae,bae,gae,ycn)),(ign,cni,apu),ame)) 13
## Mbw
(apu,cni,guc,rgr) 4
## Mpe
((((aca,rgr),guc),(apu,cni))) 5
MPH ,(rgr,gae)),(ign,cni,apu),ame)) 10 ,(wol,fuv)),(kqs,tem)), ,ewo),(bav,agq))),(pym,bom),(((ann,efi),(gkn,ogo))),yor,enn,ibo,amo,gbr), (gry,klu),(ewe,(adj,gaa,(lef,(nko,aka)))),((((dow,mzm),mdd),(gbp,mfc,zne)),(kfz,(dga,dag),spp)))) 50 BPH (((dyo,(snf,ndv),(wol,fuv)),(kqs,tem)), ,lue,ndo,(zul,sna,sot)),ewo),bav)),(pym,bom),(ann,ogo),yor,enn,ibo,amo,(gbr,nup)),(gry,klu),(ewe,(gaa, (lef,aka))),((((dow,mzm),mdd),(liy,gbp,mfc,zne)),(kfz,(dga,dag),spp)))) 46 MBH (((dyo,(snf,ndv),(wol,fuv)),(kqs,tem)), ,ewo),(bav,agq))),(pym,bom),(((ann,efi),(gkn,ogo))),yor,enn,ibo,amo,gbr), (gry,klu),(ewe,(adj,gaa,(lef,(nko,aka)))),((((dow,mzm),mdd),(gbp,mfc,zne)),(kfz,(dga,dag),spp)))) 50 MPH (((dyo,(snf,ndv),(wol,fuv)),(kqs,tem)), ,lue,ndo,(zul,sna,sot)),ewo),bav)),(pym,bom),yor,enn,ibo,amo,(gbr,nup)),gry,(ewe,(gaa,(lef,aka))), ((((dow,mzm),mdd),(gbp,mfc,zne)),(kfz,(dga,dag),spp)))) 42 , ,rap),smo)))),(gil,pon),pma),kwd,tnl),(khl,kij,tgc),yap),(irh,mky)),tet),(iba,ind),plt,pau,tgl,bhq),pwn,dru) 31 BBH (tay,((mnb,bhq),((xbr,nni,tet), ,rap),(fut,smo)),(ton,niu))),rtm),(dhv,iai),((gil,kos,((pon,mkj),woe))),cir,pma),(aty,erg,tnl),kwd),((tgc,(hla,ksd),nak), (khl,(xsi,pss),mva),(((tbo,gvs),sbe),(meu,kij))),yap),((irh,mky),(mhz,amk)))),cha,(plt,bdl),jav,ljp,(((ace,cja),(iba,ind)),sun),((btx,bbc),nia),pau,tgl),pwn,dru,tsu) 66 BBW (tay,(btx,bbc,iba,ind),cha,(dhv,fij,haw,iai,khl,kij,gil,kwd,tnl,mri,pma,pon,rap,smo,tgc),(irh,mky),plt,pwn,pau,dru,tgl,tet,bhq,yap) 31 BPE , ,rap))))),gil,pma)),(kij,tgc),yap),mky)),ind,plt,tgl,bhq),pwn) 19 BPH (tay,((mnb,bhq),((xbr,tet), ,rap),(fut,smo)),(ton,niu))),rtm),(dhv,iai), ,woe))),cir,pma),(aty,erg,tnl),(kwd,aia)),(((tgc,ksd),nak), (jae,khl,mva),((tbo,gvs),(meu,kij))),yap),((irh,mky),mhz))),cha,(plt,bdl),(((ace,cja),(iba,ind)),sun),(btx,bbc),pau,(pam,(ceb,tgl),pag)),pwn,dru,tsu) 62 BPW ((ace,btx,bbc,ind,sun),tay,cha, ,plt,pwn,pau,dru,mky,tgl,tet,bhq,yap) 34 MBE ((ace,(iba,ind),(btx,bbc),cha, , ,rap),smo)))),(gil,pon),pma),kwd,tnl),(khl,kij,tgc),yap),(irh,mky)),tet),plt,pau,tgl,bhq),tay,pwn,dru) 32
## Mbh
(tay,((mnb,bhq),((xbr,nni,tet), ,rap),(fut,smo)),(ton,niu))),rtm),(dhv,iai),((gil,kos,((pon,mkj),woe))),cir,pma),(aty,erg,tnl),kwd),((tgc,(hla,ksd),nak), (khl,(xsi,pss),mva),(((tbo,gvs),sbe),(meu,kij))),yap),((irh,mky),(mhz,amk)))),cha,(plt,bdl),jav,ljp,(((ace,cja),(iba,ind)),sun),((btx,bbc),nia),pau,tgl),pwn,dru,tsu) 66
## Mbw
(tay,(btx,bbc,iba,ind),cha,(dhv,fij,haw,iai,khl,kij,gil,kwd,tnl,mri,pma,pon,rap,smo,tgc),(irh,mky),plt,pwn,pau,dru,tgl,tet,bhq,yap) 31 MPE ((((ace,ind),sun),(btx,bbc),cha, , ,rap),smo)))),(kos,gil,(mkj,pon)),pma,cir),kwd,tnl),(khl,kij,tgc),yap),mky),tet),plt,pau,tgl,bhq),tay,pwn,dru) 34
## Mph
(tay,((mnb,bhq),((xbr,tet), ,rap),(fut,smo)),(ton,niu)))),(dhv,iai),((gil,kos,((pon,mkj),woe))),cir,pma),(aty,erg,tnl),(kwd,aia)),(((tgc,ksd),nak), (jae,khl,mva),((tbo,gvs),(meu,kij))),yap),((irh,mky),mhz))),cha,(plt,bdl),(((ace,cja),(iba,ind)),sun),(btx,bbc),pau,(pam,(ceb,tgl),pag)),pwn,dru,tsu) 61 Central-Sudanic BBH (niy,lgg,(yul,(bmi,(myb,sba)))) 6 BPH (niy,lgg,(yul,(bmi,(myb,sba)))) 6 MBH (niy,lgg,(yul,(bmi,(myb,sba)))) 6
## Mph
(niy,lgg,(yul,(bmi,(myb,sba)))) 6 ,fra)),ita),ron)),((eng,deu),(isl,(nor,swe))),ell,(((hin,pan),kas,mar,nep,sin),((kmr,pes),pst)),(lav,lit)) 29
## Bbh
(aln,hye,(lit,lav),(((bre,cor,cym),(gla,gle))),(((nor,dan,swe),isl),(eng,deu,nld)),ell,(guj,pan,lmn,((urd,hin),(ben,mai),nep,kas,sin,mar),((oss,pst),(kmr,pes))),(ita,(fra,(cat, (spa,por)),ron)), ,bul,(ces,pol,hrv))) 44
## Bbw
(aln,hye,(bre,gle),(bul,pol,rus),(cat,fra,ita,ron,spa),(eng,deu,isl,nor,swe),ell,(hin,kas,mar,nep,pan,sin),(kmr,pst,pes),(lav,lit)) 29
## Bpe
(aln,hye,(bre,gle),(bul,pol,rus),((eng,deu),(isl,(nor,swe))),(((fra,spa),ron)),ell,((hin,kas),(pes,pst)),(lav,lit)) 22
## Bph
(aln,hye,(lit,lav),(((bre,cym),(gla,gle))),(((nor,dan,swe),isl),(eng,deu,nld)),ell,(pan,((urd,hin),(ben,mai),nep,kas,sin,mar),((oss,pst),(kmr,pes))),(ita,(fra,(cat,(spa,por)),ron)), ,bul,(ces,pol,hrv))) 41 BPW (aln,hye,(bre,gle,cym),(bul,pol,rus),(cat,fra,ita,ron,spa),(nld,eng,deu,isl,nor,swe),ell,(hin,kas,nep,pan,sin),(kmr,pst,pes),(lav,lit)) 30
## Mbe
(((afr,eng,deu),(isl,(nor,swe))),aln,hye,(bre,gle),(bul,pol,rus), ,fra)),ita),ron)),ell,(((hin,pan),kas,mar,nep,sin),((kmr,pes),pst)),(lav,lit)) 30
## Mbh
(aln,hye,(lit,lav),(((bre,cor,cym),(gla,gle))),(((nor,dan,swe),isl),(eng,deu,nld)),ell,(guj,pan,lmn,((urd,hin),(ben,mai),nep,kas,sin,mar),((oss,pst),(kmr,pes))),(ita,(fra,(cat, (spa,por)),ron)), ,bul,(ces,pol,hrv))) 44
## Mbw
(aln,hye,(bre,gle),(bul,pol,rus),(cat,fra,ita,ron,spa),(eng,deu,isl,nor,swe),ell,(hin,kas,mar,nep,pan,sin),(kmr,pst,pes),(lav,lit)) 29
## Mpe
((((afr,nld),eng,deu),(isl,(nor,swe))),aln,hye,(((bre,cym),gle)),(bul,pol,rus), ,fra)),ita),ron)),ell,(((hin,pan),kas,nep,sin),((kmr,pes),pst)),(lav,lit)) 31
## Mph
(aln,hye,(lit,lav),(((bre,cym),(gla,gle))),(((nor,dan,swe),isl),(eng,deu,nld)),ell,(pan,((urd,hin),(ben,mai),nep,kas,sin,mar),((oss,pst),(kmr,pes))),(ita,(fra,(cat,(spa,por)),ron)), ,bul,(ces,pol,hrv))) 41
## Mpw
(aln,hye,(bul,pol,rus),(eng,deu,swe),(fra,ron,spa),ell,(hin,kas),gle,(lav,lit),(pst,pes)) 19
Iroquoian BBE (chr,one) 2
## Bbh
((see,one),chr) 3
## Bph
((see,one),chr) 3
## Mbh
((see,one),chr) 3
## Mpe
(chr,(one,see)) 3
## Mph
((see,one),chr) 3
Je-Jabuti BBH (kgp,(apn,ram)) 3
## Dediu & levinson
Abstract profiles of structural stability: ESM (S1) 28/67 ,vai))),(xpe,men)))) 6 MBH ((daf,mev),(bam,(xpe,men))) 5
MPH ,(xpe,men)))) 5 ,lbe,(aqc,(lez,rut))),(bbl,(inh,che))) 9
## Mataco-guaicuru
BBW ((aqc,lez,rut),(ava,huz),(inh,bbl),lbe) 8 BPH (((gdo,ava,huz),lbe,(aqc,(lez,rut))),(bbl,(inh,che))) 10 BPW ((aqc,lez),(ava,huz),(inh,bbl),lbe) 7
## Mbh
(((ava,huz),lbe,(aqc,(lez,rut))),(bbl,(inh,che))) 9
MBW ((aqc,lez,rut),(ava,huz),(inh,bbl),lbe) 8
## Mph
(((ava,huz),lbe,(aqc,(lez,rut))),(bbl,(inh,che))) 9
## Mpw
(huz,inh,lez) 3
## Niger-congo
BBE , , ,nhu)),ibo,yor),(((dag,kfz),spp),((dow,mzm),(gbp,zne))),gry),((dyo,ndv,wol),(kqs,tem)),ijc),(bam,daf),mor) 29 BBW ((aka,ewe),bam,(bom,pym),(dag,kfz,spp),daf,(dyo,ndv,wol),(dow,gbp,mzm,sag,zne),(ewo,lue,cgg,nhu,swh,zul),gry,ibo,ijc,(kqs,tem),mor,yor) 30 BPE ((((aka,ewe),gry,(ibo,(lue,cgg,swh,zul),yor),(kfz,spp)),(dyo,wol),ijc),bam) 15 BPW ((aka,ewe),bam,bom,(dag,kfz,spp),(dyo,fuv,wol),(dow,gbp,mzm,sag,zne),(ewo,lue,cgg,swh,zul),gry,ibo,ijc,(kqs,tem),mor,yor) 27 MBE (((((abi,aka),ewe),((bom,pym),((ewo,(lue,cgg,swh,zul),nhu)),ibo,yor),(((dag,kfz),spp),((dow,mzm),(gbp,zne))),gry),((dyo,ndv,wol),(kqs,tem)),ijc),(bam,daf),mor) 30 MBW ((aka,ewe),bam,(bom,pym),(dag,kfz,spp),daf,(dyo,ndv,wol),(dow,gbp,mzm,sag,zne),(ewo,lue,cgg,nhu,swh,zul),gry,ibo,ijc,(kqs,tem),mor,yor) 30 MPE (((((abi,aka),ewe),(bom,(ewo,(lue,cgg,swh,zul)),ibo,yor),(((dag,kfz),spp),((dow,mzm),(gbp,zne))),gry),(dyo,(fuv,wol),(kqs,tem)),ijc),bam,mor) 27 MPW ((aka,ewe),bam,(dyo,wol),(gbp,sag),gry,ibo,ijc,(kfz,spp),(lue,cgg,swh,zul),yor) 17
Nilo-Saharan BBE (bmi,(lgg,niy),wti,((dip,(((laj,luo),mde)),mas,niq),ikx,(mur,nrb,kzh)),fvr,(khq,ses),kun,knc,kgo) 20 ,luo)),mde),dip),mas,niq),ikx,(mur,nrb,kzh)),(bmi,(lgg,niy)),wti,fvr,(khq,ses),kun,knc,kgo) 21
## Mbw
(bmi,wti,(dip,laj,luo,mas,niq),fvr,ikx,knc,(khq,ses),kun,lgg,mde,mur,nrb,niy,kzh) 19 ,luo)),mde),dip),(mas,tuv),niq),ikx,(mur,nrb,kzh)),(bmi,(lgg,niy)),fvr,(khq,ses),kun,knc,kgo) 21
[formula] MPE (((( [/formula]
## Mpw
(bmi,fvr,knc,(khq,ses),kun,(laj,luo,mas,niq),lgg,mde,mur,niy,kzh) 15
Nilotic BBH (bfa,(mas,tuv),(kpz,niq),(dip,(lkr,(luo,laj)))) 9 BPH (bfa,(mas,tuv),(kpz,niq),(dip,(lkr,(luo,laj)))) 9
## Mbh
(bfa,(mas,tuv),(kpz,niq),(dip,(lkr,(luo,laj)))) 9
## Mph
(bfa,(mas,tuv),(kpz,niq),(dip,(lkr,(luo,laj)))) 9
North Caucasian BBE ((abk,kbd),(aqc,(lez,rut),ava,huz,(inh,bbl),lbe)) 10 BPE (abk,(ava,huz,inh,lbe,lez)) 6 ,trc)),(maq,(ctp,zai))),((cle,(chq,cco),ote),tpx)) 11
## Mbe
## Mpw
(cle,mig,ote) 3
## Dediu & levinson
Abstract profiles of structural stability: ESM (S1) 31/67 ,lep),(lif,new)),lhu,mni,dap),(yue,cmn)) 19 BBH (cmn,(hak,yue),(bca,((lhu,mya),nbf),(((cdm,kgj),(byw,lif,dus),new),lep,((taj,ggn),bee)),(sip,(bod,lbj)),(kac,(brx,grt)),(ksw,(eky,bwe)),((((cnh,bgr),lus),ctd), (njo,nsm,mni)),raw,dap)) 35 ,(aey,ssd)), , ,gaj,kew,ygr,kpw),spl,(dni,ekg))),mtg) 19
## Bbw
BPH ,(aey,ssd)), , ,gaj,kew,ygr,kpw),spl,(dni,ekg))),mtg) 19
MBH ,(aey,ssd)), , ,gaj,kew,ygr,kpw),spl,(dni,ekg))),mtg) 19
MPH ,(aey,ssd)), , ,gaj,kew,ygr,kpw),spl,(dni,ekg))),mtg) 19 ,(kaa,kir)),uzn)),sah,tyv)) 11
BPH ,kir),uzn)),sah,tyv)) 9
MBH ,(kaa,kir)),uzn)),sah,tyv)) 11
MPH ,kir),uzn)),sah,tyv)) 9
## Dediu & levinson
Abstract profiles of structural stability: ESM (S1) 34/67
## Uto-aztecan
BBE , ,(ood,yaq))) 7
BBH ,xaw,pao),(chl,lui)),(crn,(ppl,(nhg,ncj)),(ntp,ood),yaq)) 14
BBW ,ood,yaq) 7 BPE , ,(ood,yaq))) 6
BPH ,(xaw,ute),pao),(chl,lui)),(crn,(ppl,(nhg,ncj)),(ntp,ood),yaq)) 15 BPW ((chl,lui),(com,par),hop,(nhg,ppl),ood,yaq) 9
MBE , ,(ood,yaq))) 7
MBH ,xaw,pao),(chl,lui)), ,(ntp,ood),yaq)) 14 ,ood,yaq) 7
## Mbw
## Mpe
(((chl,lui),(com,par),hop),((nhg,ppl),(ood,yaq))) 9
MPH ,(xaw,ute),pao),(chl,lui)),(crn,(ppl,(nhg,ncj)),(ntp,ood),yaq)) 15 0.66 0.67 † A combined p-value of exactly 0.00 is an artifact of combining very small randomization p-values (smaller than the lower limit given by the number of randomizations used here, 10 -4 ) and represents in fact an extremely small (but still greater than 0) p-value. ‡ For tests involving Siberia (N-E Eurasia), the first entry in the H cells represents Chukotko-Kamchatkan & Tungusic while the second entry represents Chukotko-Kamchatkan only; this extends to the combined p-value cells.
## Mpw
## Dediu & levinson
Abstract profiles of structural stability: ESM (S1) 55/67 datasets containing the family. Both Pearson's r and Spearman's ρ are given and the vast majority is very highly significant due to the very large number of observations in the posterior distribution, except for those marked with † which are non-significant at an α-level of 0.05. We considered for this test only the datasets using the Ethnologue and Harald Hammarstöm's classifications given that they have an unconstrained number of levels.
## Macro-area geo
## P-values for each dataset paired t-test
## Dediu & levinson
Abstract profiles of structural stability: ESM (S1) 62/67
[fig] Figure S1 5, Figure S2 6, Figure S3 7, Figure S4 8, Figure S5 9, Figure S6: Observed Ripley's K for BayesLang .................................................................................Observed Ripley's K for MrBayes ......................................................................................Stability distances (MDS and NeighborNet) for MBE ........................................................Stability distances (MDS and NeighborNet) for MBW .......................................................Stability distances (MDS and NeighborNet) for MBH ........................................................Stability distances (MDS and NeighborNet) for MPE ......................................................10 Figure S7: Stability distances (MDS and NeighborNet) for MPW .....................................................11 Figure S8: Stability distances (MDS and NeighborNet) for MPH ......................................................12 Figure S9: Stability distances (MDS and NeighborNet) for BBE .......................................................13 Figure S10: Stability distances (MDS and NeighborNet) for BBW ....................................................14 Figure S11: Stability distances (MDS and NeighborNet) for BBH .....................................................15 Figure S12: Stability distances (MDS and NeighborNet) for BPE .....................................................16 Figure S13: Stability distances (MDS and NeighborNet) for BPW ....................................................17 Figure S14: Stability distances (MDS and NeighborNet) for BPH .....................................................18 Figure S15: Punctuated evolution across families ...............................................................................64 Figure S16: Punctuated evolution across categories of structural features ..........................................65 Figure S17: Punctuated evolution: means across families and categories ...........................................66 Figure S18: Punctuated evolution: standard deviations across families and categories ......................67 Supplementary Tables Table S1: The 12 datasets .......................................................................................................................3 Table S2: Observed versus expected stability distances .........................................................................4 Table S3: The structural features and their coding ...............................................................................22 Table S4: The composition and structure of the language families.......................................................36 Table S5: Mantel correlations between stability and geographic distances .........................................37 Table S6: Most involved features for dataset MBE .............................................................................38Table S7: Most involved features for dataset MBW ............................................................................39 Table S8: Most involved features for dataset MPE .............................................................................40 Table S9: Most involved features for dataset MPW ............................................................................41 Table S10: Most involved features for dataset BBE ............................................................................41 Table S11: Most involved features for dataset BBW ...........................................................................42 Table S12: Most involved features for dataset BPE ............................................................................43 Table S13: Most involved features for dataset BPW ...........................................................................44 [/fig]
[table] Table S14: Correlations and concordances between methods for combining p-values .......................45 Table S15: The composition and interpretation of sets of language families........................................54 Table S16: Statistical robustness of sets of language families..............................................................56 Table S17: The R code implementing the methods for combining p-values ........................................59 Table S18: Punctuated evolution across language families ..................................................................61 Table S19: Punctuated evolution across categories of structural features ............................................59 [/table]
[table] Table S2: Comparison between the observed stability distances between the actual language families and 10,000 simulations of random language families. Diag: the maximum possible stability distance in the stability cube with the number of dimensions equal to the number of features in this case. Measure: the summary of the stability distances matrix between language families (NN is the average distance to the nearest neighbor; Mean is the mean stability distance). Observed: the actual value of the measure for the real language families. Simulated: 10,000 simulations of random language families in the current stability cube resulting in 10,000 simulated stability distance matrices; for each such simulated matrix the measures as defined above have been computed and the distribution of these 10,000 simulated measures was summarized using their Mean and standard deviation (SD); SDs from mean represents how far is the observed measure from the mean of the simulated measures in terms of the simulated SDs. P-value: the empirical probability that the observed measure is outside the simulated distribution. [/table]
[table] Table S3: The Poly and Bin cells are empty for those features not included in the stability estimation (as polymorphic or binary) for various reasons such as data coverage and coding meaningfulness. The last two rows give the lists of polymorphic and binary features ordered from the most stable to most unstable (for easiness of comparison). See[9] and especially Tables S7, S4 and S10 in there for more details and explanations. [/table]
[table] Table S4: The composition and structure of the language families can vary among cases. The structure is given as the topology of the language family in the parentheses notation (Newick format; http://evolution.genetics.washington.edu/phylip/newicktree.html). The languages are given using their ISO 639-2 three-letter codes. The number of families and their composition might differ slightly among outgroups for MrBayes. [/table]
[table] Table S5: The [/table]
[table] Table S6: Most [/table]
[table] Table S7: Most [/table]
[table] Table S8: Most [/table]
[table] Table S9: Most [/table]
[table] Table S10: Most [/table]
[table] Table S11: Most [/table]
[table] Table S12: Most involved features for dataset BPE (showing only those with involvement > 0.03). The PC1 of the 5 runs explains 99.95% of variance and the maximum Mantel correlation is 0.45. [/table]
[table] Table S13: Most [/table]
[table] Table S16: Statistical robustness of sets of language families. Actual p-values for each of the datasets and the combined p-values using the five methods (Fisher, Z-transform, Hartung, Simes and Makambi) applied to all datasets for raw (Geo is "No") and geography-corrected (Geo is "Yes") stability distances. Also showing the paired t-tests between the raw and geography-corrected p-values (bold=significant t-test, italic=positive t-test). For S we show if H0 was rejected("Yes" or "No") for α=0.05. For H and M the estimated inter-datasets correlations are also shown. The most conservative combined p-value among the methods is in bold. Significant p-values at α=0.05 are in italic. See Table S15 for the actual composition of the sets of families. [/table]
[table] Table S18: The correlations between branch length and number of nodes for all language families amalgamated (first line) and each family separately for each of the 8 [/table]
[table] Table S19: The correlations between branch length and number of nodes for all types of categories separately for each of the 4 datasets considered. Reported are the mean and standard deviation of Pearson's r across language families and outgroups. We considered for this test only the datasets using Harald Hammarstöm's classification given the computational cost. N=1 considers all families, while N=5 and N=7 only those with data for at least 5 or for all 7 categories. [/table]
|
Differences in the clinical characteristics and outcomes of COVID-19 patients in the epicenter and peripheral areas of the pandemic from China: a retrospective, large-sample, comparative analysis
Methods: The Sichuan & Wuhan CollaborationResearch Group for COVID-19 established two retrospective cohorts that separately reflect the epicenter and peripheral area during the early pandemic. The epidemiology, clinical characteristics and outcomes of patients in the two groups were compared. Multivariate regression analyses were used to estimate the adjusted odds ratios (aOR) with regard to the outcomes.Results: The Wuhan (epicenter) cohort included 710 randomly selected patients, and the peripheral (Sichuan) cohort included 474 consecutive patients. A higher proportion of patients from the periphery had upper airway symptoms, whereas a lower proportion of patients in the epicenter had lower airway symptoms and comorbidities. Patients in the epicenter had a higher risk of death (aOR=7.64), intensive care unit (ICU) admission (aOR=1.66), delayed time from illness onset to hospital and ICU admission (aOR=6.29 and aOR=8.03, respectively), and prolonged duration of viral shedding (aOR=1.64). Conclusions: The worse outcomes in the epicenter could be explained by the prolonged time from illness onset to hospital and ICU admission. This could potentially have been associated with elevated systemic inflammation secondary to organ dysfunction and prolonged duration of virus shedding independent of age and comorbidities. Thus, early supportive care could achieve better clinical outcomes.
# Background
In December 2019, an outbreak of pneumonia of unknown cause was identified in Wuhan, the capital of Hubei province in China. A novel coronavirus, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which had not been detected previously in humans, was identified subsequently by Chinese scientists as the cause [bib_ref] A novel coronavirus genome identified in a cluster of pneumonia cases -Wuhan, Wenjie [/bib_ref]. The disease was named the coronavirus disease 2019 (COVID-19) by the World Health Organization (WHO). The clinical spectrum of COVID-19 appears to be wide, and ranges from self-limited mild upper respiratory tract illness to severe pneumonia causing hospitalization or death. The clinical characteristics of some COVID-19 case series in Wuhan, the epicenter of the pandemic, have been previously reported in detail. The reports indicated that 26 to 33% of patients required intensive care and 4 to 15% died [bib_ref] Clinical features of patients infected with 2019 novel coronavirus in Wuhan, Huang [/bib_ref] [bib_ref] Clinical characteristics of 138 hospitalized patients with 2019 novel coronavirus-infected pneumonia in..., Wang [/bib_ref] [bib_ref] Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia..., Chen [/bib_ref].
After the outbreak of COVID-19 in Wuhan, the government of the Sichuan province implemented strict measures to combat COVID-19. The Health Commission of Sichuan Province (HCSP) focused on traditional public health outbreak response tactics, including isolation, quarantine, social distancing, and community containment, as recommended by the National Health Commission of China. All medical resources were allocated by the HCSP to ensure efficient use. An expert panel drawn from multidisciplinary teams was established and comprised 125 physicians who were led by Dr. Wei Min Li and Dr. Zong An Liang (the corresponding authors of this study) since January 15, 2020. This expert panel soon released emergency prevention and control guidelines for COVID-19 in the medical institutions of the Sichuan province. Furthermore, we funded two additional important expert panels with psychological counseling [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] and traditional Chinese medicine as complementary and alternative treatment options [bib_ref] Covid-19 -navigating the uncharted, Fauci [/bib_ref] [bib_ref] Pharmacologic treatments for coronavirus disease 2019 (COVID-19): a review, Sanders [/bib_ref]. Physicians caring for severely or critically ill patients could receive daily internet consultations with members of the expert panel. There were 208 designated hospitals across Sichuan Province that were accessible for SARS-CoV-2-suspected or -confirmed individuals. This arrangement resulted in improved outcomes in Sichuan province, one of the peripheral areas of the pandemic. In other peripheral areas, 2 to 10.1% of patients needing intensive care, and an approximately 1.0% mortality rate were reported in recently published studies [bib_ref] COVID-19 in a designated infectious diseases hospital outside Hubei Province, Cai [/bib_ref] [bib_ref] Clinical findings in a group of patients infected with the 2019 novel..., Xu [/bib_ref] [bib_ref] Characteristics of COVID-19 infection in Beijing, Tian [/bib_ref].
The factors underlying the significantly different clinical outcomes between the epicenter and peripheral areas affected by the pandemic remains largely unexplored. Recently, Liang et al. [bib_ref] Clinical characteristics and outcomes of hospitalised patients with COVID-19 treated in Hubei..., Liang [/bib_ref] observed the clinical characteristics and outcomes of hospitalized patients with COVID-19who were treated in Hubei (epicenter) or outside Hubei (non-epicenter). However, as theirs is a multicenter study, the possibility of selection bias for the included patients cannot be ruled out. Furthermore, hospitalized patients in Hubei but not in Wuhan, would not be well representative of the first-generation COVID-19 cases. Considering the rapidly increasing number of cases with SARS-CoV-2 infection worldwide, the existing research into the differences between the epicenter and peripheral areas of the pandemic in the clinical characteristics and outcomes of COVID-19 patients was insufficient.
This study could provide information that would be potentially helpful in formulating clinical practice recommendations to tackle the COVID-19 pandemic worldwide.
# Methods
## Study design and subjects
This was a retrospective study based on two cohorts evaluated by the Sichuan and Wuhan Collaboration Research Group for COVID-19, China. The Wuhan cohort, drawn from the epicenter area of the pandemic, was formed using a computer-generated simple random sampling method that was applied to enroll subjects from two designated hospitals, namely the Wuhan Red Cross Hospital and Renmin Hospital of Wuhan University, Wuhan, China. The Sichuan cohort, as the group of patients from the peripheral area of the pandemic, consisted of SARS-CoV-2-confirmed patients who were consecutively recruited from 41 designated hospitals until March 12, 2020. Based on the exposure history, we further divided the Sichuan cohort into two sub-cohorts, with or without Wuhan exposure history. All patients enrolled in this study were diagnosed with COVID-19 according to the interim guidance issued by the National Health Commission of China and the WHO. SARS-CoV-2 infection was confirmed by a positive result on a real-time reverse-transcriptase-polymerase-chain-reaction of nasopharyngeal, pharyngeal, throat-swab or sputum specimens. Some of these patients were included in studies reported by Wei et al. [bib_ref] Acute myocardial injury is common in patients with COVID-19 and impairs their..., Wei [/bib_ref] , Xiong et al. [bib_ref] New onset neurologic events in people with COVID-19 infection in three regions..., Xiong [/bib_ref] and Xiong et al. [bib_ref] Hypertension is a risk factor for adverse outcomes in patients with coronavirus..., Xiong [/bib_ref] ; however, their study purposes are significantly different from those of this study.
## Data collection
The medical records of patients with COVID-19 were reviewed by members of the trained research team. Epidemiological, demographic, clinical, laboratory, radiological, treatment and outcome data were collected by using standardized data collection forms (modified case record form for the clinical characterization of severe acute respiratory infection that was shared by the International Severe Acute Respiratory and Emerging Infection Consortium [ISARIC]) from the electronic medical records. The cutoff date was Mar 12, 2020. We collected details of the exposure history, clinical symptoms and signs, and laboratory findings on admission. Laboratory examinations were performed according to the clinical care needs of the patients. Data on radiological abnormalities were extracted from the selected documentation. Patients were excluded if their medical records were not available. A team of trained researchers abstracted the data and entered the structured spreadsheet. All data were cross-checked.
## Study outcomes
The primary outcomes included death or mechanical ventilation whether or not it involved intensive care unit (ICU) admission. Mechanical ventilation was performed in the ward if ICU admission was not possible due to the overwhelming numbers of COVID-19 patients. The secondary outcomes were the rate of ICU admission, time from illness onset to ICU admission and discharge, length of hospital stay, and duration of viral shedding after COVID-19 onset. Duration of viral shedding was defined to ended when two consecutive negative results with qPCR detection were obtained at time intervals greater than 24 h. The criteria for discharge were absence of fever for at least 3 days, substantial improvement in both lungs on chest computed tomography (CT), clinical remission of respiratory symptoms and comorbidities, and cessation of SARS-Cov-2.
# Statistical analysis
Continuous variables were compared using the Student's t test or the Mann-Whitney U test; categorical variables were compared by the chi-square test or Fisher's exact test as appropriate. Logistic or linear regression was performed to identify clinical variables that were associated with outcomes. The detailed statistical analysis is described in the supplementary data.
# Results
## Epidemiological and clinical characteristics at hospitalization
As of March 12, 2020, a total of 1979 cases from the two hospitals in Wuhan were identified. The Wuhan cohort included 35.9% (n=710) of all patients from Wuhan, selected using a computer-generated simple random sampling method, formed the Wuhan cohort. In the Wuhan cohort, illness onset in the first case was noted on December 24, 2019, and the first hospitalization occurred on January 5, 2020 [fig_ref] Figure 1: See legend on next page [/fig_ref].
There were 538 patients with COVID-19 who were consecutively admitted to 41 designated hospitals in Sichuan Province. The Sichuan cohort comprised 474 patients [fig_ref] Figure 1: See legend on next page [/fig_ref] , when 64 patients with inaccessible medical records were excluded. Epidemiological data indicated that the first cases of SARS-CoV-2 infection in the Sichuan cohort occurred in December 31, 2019, and the first case was admitted to the designated hospital on January 16, 2020 [fig_ref] Figure 1: See legend on next page [/fig_ref]. The daily Wuhan-related exposure cases with onset of COVID-19 in the Sichuan cohort peaked on January 23, 2020, and those without Wuhan-related exposure peaked on February 1, 2020 . The median time from illness onset to admission in the Sichuan cohort was significantly shorter than that in the Wuhan cohort The demographic and clinical characteristics of these patients are shown in [fig_ref] Table 1: Demographics and clinical characteristics of patients in the Sichuan and Wuhan cohorts [/fig_ref] Fever was the commonest symptom and was present in 61.8% of patients in the Sichuan cohort or 65.1% of patients in the Wuhan cohort, but the difference was not significant (P=0.246). The Sichuan cohort had a higher incidence of productive cough than the Wuhan cohort (P=0.012). However, the Wuhan cohort seemed to have a higher symptomatic burden with regard to the lower respiratory tract, including shortness of breath (25.4% vs. 9.0%, P< 0.001), chest distress (23.8% vs. 9.0%, P< 0.001), wheeze (13.9% vs. 4.8%, P< 0.001), and general symptomatic burden, including fatigue (36.2% vs. 22.3%, P< 0.001), hemoptysis (3.0% vs. 1.1%, P=0.028), altered consciousness (1.8% vs. 0.2%, P=0.011), and diarrhea (12.1% vs. 6.3%, P=0.001). In contrast, the Sichuan cohort was more likely to have upper respiratory symptoms, including pharyngalgia (13.9% vs. 7.5%, P< 0.001), rhinorrhea (5.0% vs. 1.4%, P< 0.001), nasal obstruction (3.4% vs. 1.1%, P=0.007), and headache (10.1% vs. 4.6%, P< 0.001) [fig_ref] Figure 3 a: Symptomatic burden of patients with COVID-19 between Sichuan cohort and Wuhan cohorts [/fig_ref]. Different severity distributions were observed between the two cohorts (P< 0.001), as assessed by CURB-65 and MuLBSTA (both P< 0.001). More than 75% of patients in both cohorts had mild or general disease, although the Sichuan cohort had a higher proportion of severe cases (17.0% vs. 8.4%) and the Wuhan cohort had more critically ill patients (13.6% vs. 6.3%). Chest CT radiographs in the Wuhan cohort were more likely to show bilateral lung involvement (P=0.012) and consolidation (P=0.006).
There was no difference in white blood cell count, lymphocyte count, prothrombin time, albumin, alanine aminotransferase, aspartate aminotransferase, procalcitonin and interleukin 6 (IL-6) between the two cohorts. The Sichuan cohort had lower neutrophil count (P< 0.001), platelet count (P< 0.001), D-dimer levels (P= 0.001), and C-reactive protein levels (P< 0.001) and higher levels of hemoglobin (P=0.015), activated partial thromboplastin time (P< 0.001), creatinine (P=0.040), and creatine kinase (P< 0.001) [fig_ref] Table 1: Demographics and clinical characteristics of patients in the Sichuan and Wuhan cohorts [/fig_ref].
## Treatments and clinical outcomes
A comparison of treatments and clinical outcomes between the two cohorts is shown in [fig_ref] Table 2: Treatments and outcomes of the patients in the Sichuan and Wuhan cohorts [/fig_ref]. Almost all patients received antiviral treatment in Sichuan (94.7%) or Wuhan (93.2%). Fewer patients in the Sichuan cohort received antibiotics (P< 0.001), corticosteroids (P< 0.001) and supplemental oxygen therapy (P< 0.001).
The case fatality rate in the Sichuan cohort was obviously lower than that in the Wuhan cohort (0.6% vs. 8.3%, P< 0.001). However, there was no significant difference in the proportion of patients receiving noninvasive mechanical ventilation or invasive mechanical ventilation between the two cohorts (5.7% vs. 5.9%, P=0.872 and 1.7% vs. 1.4%, P=0.701). With regard to the secondary outcomes, the proportion of patients who were admitted to the ICU in the Sichuan cohort was significantly lower than that in the Wuhan cohort (6.3% vs. 13.6%, P< 0.001). The time from illness onset to ICU admission and time from illness onset to discharge in the Sichuan cohort were shorter than that in the Wuhan cohort (7.0
## Logistic regression analyses
Multivariable logistic regression models were used to explore the differences in clinical outcomes between the Sichuan and Wuhan cohorts ( for time from illness onset to hospitalization, the risk of the Wuhan cohort was nearly unchanged; however, the Sichuan cohort had a lower risk for extended time from illness onset to discharge (aOR=0.46, 95% CI=[0.34, 0.63], P< 0.001).
In the overall study population of COVID-19 patients from the two cohorts, we constructed multivariable logistic regression models to detect the risk factors at admission for death, ICU admission, mechanical ventilation and duration of viral shedding after COVID-19 onset [fig_ref] Table 2: Treatments and outcomes of the patients in the Sichuan and Wuhan cohorts [/fig_ref]. After adjusting for the cohort sites, sex, age, smoking status and the CCI, we found that white blood cells (> 10× 10 9 /L), neutrophils (> 6.3× 10 9 /L), lymphocytes (> 1.0× 10 9 /L), hemoglobin (< 90 g/L), D-dimer (> 0.5 mg/L), creatine kinase (> 185 IU/L), hyper-sensitive troponin [fig_ref] Table 4: Correlation of time from illness onset to hospitalization with systemic inflammation and... [/fig_ref]. After adjusting for age, the correlations of delay in hospitalization with BUN and D-dimer did not achieve statistical significance, which indicated that the delay in hospitalization was independent of age. We further analyzed the relationship between the delay in hospitalization and ICU admission with elevated systemic inflammation after adjusting for age, sex, smoking, and steroid use. In general, the relationship between the delay in hospitalization and elevated systemic inflammation did not change, which implied that these relationships were independent of age, sex, smoking, and steroid use.
## Subgroup analyses between sichuan sub-cohorts with vs. without wuhan-related exposure
There was almost no difference in clinical characteristics and outcomes between the two sub-cohorts with and without Wuhan-related exposure in Sichuan. Detailed information is provided in Supplementary Data.
Sichuan sub-cohort with Wuhan-related exposure vs.
## Wuhan cohort
The differences in the clinical characteristics and outcomes between the Sichuan sub-cohort with Wuhan-related exposure and Wuhan cohort were similar to the differences between the Sichuan and Wuhan cohorts. The results are described in detail in the Supplementary Data .
# Discussion
To the best of our knowledge, there exists a paucity of information obtained from a comparative large-sample study on the differences in epidemiology, clinical characteristics and outcomes of patients with COVID-19 between the epicenter (Wuhan) and the peripheral areas of pandemic. This comparative study provides important insights. First, the outbreak and transmission of COVID-19 within the region of Sichuan as the peripheral epidemic area has been well contained within 2 months through the use of traditional public health outbreak response tactics. Second, the Sichuan cohort is characterized by a higher incidence of upper airway symptoms, whereas the Wuhan cohort was older, had fewer lower airway symptoms and comorbidities, and had elevated pivotal systemic inflammation indicative of organ dysfunction as well as worse clinical outcomes independent of sex, age, smoking and comorbidities. Third, the subgroup analysis indicated that, within the Sichuan cohort, the patients with Wuhan-related exposure had similar clinical features and outcomes to those with non-Wuhan-related exposure. Fourth, the Wuhan-related exposure patients in the Sichuan cohort had better clinical outcomes than those in the Wuhan cohort, although these two groups of patients had a similar Wuhanrelated exposure history. As indicated in recently published studies [bib_ref] Clinical characteristics and outcomes of hospitalised patients with COVID-19 treated in Hubei..., Liang [/bib_ref] , the COVID-19 patients in Wuhan, at the epicenter area of the epidemic, were older, had more co-existing conditions assessed by the CCI, had extended time from illness onset to hospitalization, and included more severely ill patients. However, the Sichuan cohort, as the peripheral area, had some characteristics features. First, there were fewer healthcare workers in the Sichuan cohort than in the Wuhan cohort, which could be at least partially explained by the insufficient implementation of precautions and the overwhelmed health system during the earlier stage of this outbreak in Wuhan. Second, intriguingly, there was a higher incidence of upper airway symptoms, rather than high incidence of lower airway symptoms in the Wuhan cohort at the epicenter epidemic, which was similar to the findings from exported cases in Singapore [bib_ref] Epidemiologic features and clinical course of patients infected with SARS-CoV-2 in Singapore, Young [/bib_ref]. Accordingly, the exported patients from the epicenter were usually diagnosed with a "common cold" at the beginning of the COVID-19 outbreak. The different populations, the airway proliferation location, or the evolution of SARS-CoV-2 possibly could account for these differential symptoms [bib_ref] Epidemiologic features and clinical course of patients infected with SARS-CoV-2 in Singapore, Young [/bib_ref] [bib_ref] Transmission dynamics and evolutionary history of 2019-nCoV, Li [/bib_ref] [bib_ref] Evolving epidemiology and transmission dynamics of coronavirus disease 2019 outside Hubei province,..., Zhang [/bib_ref] [bib_ref] Stability and viability of SARS-CoV-2, Rubens [/bib_ref]. Third, within the consecutively recruited cases in the Sichuan cohort as a well-defined population, the subgroup analyses indicated a higher proportion of males and older patients among the non-Wuhan-related exposure patients, which supported the theory of the propensity for SARS-CoV-2 infection in males and elders [bib_ref] COVID-19 in a designated infectious diseases hospital outside Hubei Province, Cai [/bib_ref] [bib_ref] Association between ages and clinical characteristics and outcomes of coronavirus disease 2019, Liu [/bib_ref] [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref].
Recent studies from the USA and Italy have reported that a greater proportion of elderly and male COVID-19 patients would experience more critical illness [bib_ref] Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in..., Richardson [/bib_ref] [bib_ref] Baseline characteristics and outcomes of 1591 patients infected with SARS-CoV-2 admitted to..., Grasselli [/bib_ref].
Until now no antiviral treatment for COVID-19 has proven effective, and supportive care is the mainstay of treatment is. Compared with the Wuhan cohort, the use of antibiotics (i.e. cephalosporin and quinolones) and glucocorticoids in the Sichuan cohort decreased by 26.4 and 16.1%, respectively. These results could possibly be explained as follows. First, as indicated earlier, the expert panel drawn from the multidisciplinary team established by HCSP together developed and adjusted the treatment plan for severely or critically ill patients according to the interim guidance from the National Health Commission of China and the WHO across the 208 designated hospitals in Sichuan by using the 5G network every day. Accordingly, the use of systemic corticosteroids was strictly managed and they were not routinely administered for the treatment of COVID-19 patients. Second, the COVID-19 patients in the Wuhan cohort would actually be more severe or critically ill, which was supported by the increased use of supplemental oxygen in case of acute hypoxia. In addition, prone-position ventilation, physical rehabilitation and a variety of traditional Chinese medicines were used more often in Sichuan under the guidance of the expert panel; however, this aspect needs to be investigated further in randomized controlled trials [bib_ref] Covid-19 -navigating the uncharted, Fauci [/bib_ref] [bib_ref] Pharmacologic treatments for coronavirus disease 2019 (COVID-19): a review, Sanders [/bib_ref].
In terms of clinical outcomes, several important findings were identified in this study. An epidemic outbreak provided an opportunity to obtain important information, some of which were associated with a limited window of opportunity. This study showed that there was a delay from illness onset to hospitalization in the Wuhan cohort, which might be an important risk factor for the progression of COVID-19. Multivariate regression analysis showed that the time from illness onset to hospitalization was significantly associated with mortality and ICU admission, which suggested some important implications with regard to the pathogenesis of SARS-CoV-2 and may provide insights into a unique window of opportunity for intervention [bib_ref] Covid-19 -navigating the uncharted, Fauci [/bib_ref]. Liang et al. [bib_ref] Clinical characteristics and outcomes of hospitalised patients with COVID-19 treated in Hubei..., Liang [/bib_ref] recently found that Wuhan-related exposure patients have worse clinical outcomes compared with the non-Wuhan-related exposure cases; they attributed the attenuated disease to the onward transmission of SARS-CoV-2. In fact, this is paradoxical to the findings reported from Liang et al.'s study [bib_ref] Clinical characteristics and outcomes of hospitalised patients with COVID-19 treated in Hubei..., Liang [/bib_ref] because the relationship between Wuhan-related exposure and prognosis disappeared after adjusting for confounders. Our study firstly found that COVID-19 patients in the Wuhan cohort had worse clinical outcomes including case fatality rate, ICU admission, and duration of virus shedding, independent of sex, age, smoking, comorbidities, and even time from illness onset to hospitalization. The severity of COVID-19 and the shortage of medical resources would partly account for these worse outcomes. For example, during an earlier stage of the outbreak, some patients would not have received sufficient oxygen support because of insufficient oxygen pressure.
The duration of infectious virus replication is an important factor in assessing the risk of transmission and for guiding decisions on the isolation of patients; however, the duration of SARS-CoV-2 RNA detection has not been well explored. Our study found that the Wuhan cohort in the epicenter area had the prolonged virus shedding, which may contribute to the disease severity and clinical course [bib_ref] Viral dynamics in mild and severe cases of COVID-19, Liu [/bib_ref] [bib_ref] The duration of viral shedding of discharged patients with severe COVID-19, Zhou [/bib_ref]. Furthermore, we found for the first time that the duration of virus shedding was independently associated with age and time from illness onset to hospitalization. Our findings are supported by those of other studies. Liu et al. [bib_ref] Secondary attack rate and superspreading events for SARS-CoV-2, Liu [/bib_ref] found that the viral load in severe cases was higher than that of mild cases, which had early viral shedding. Wolfel et al. [bib_ref] Virological assessment of hospitalized patients with COVID-2019, Wolfel [/bib_ref] found that virus shedding in the upper airway, which is the location of mild COVID-19, was very high during the first week of symptoms, whereas shedding of viral RNA from sputum derived from the lower airway, which is the region of general to critical illness in COVID-19, outlasts the disappearance of symptoms. Xu et al. [bib_ref] Factors associated with prolonged viral RNA shedding in patients with COVID-19, Xu [/bib_ref] found that elderly patients had prolonged virial shedding, but the correlation of age with the duration of viral shedding disappeared after adjusting for confounders, although this might be partly attributed to the small sample size.
As the pandemic evolves, mutations and natural selection of SARS-CoV-2 inevitably occur, although this virus a lower mutation rate than that of other RNA viruses [bib_ref] Genetic grouping of SARS-CoV-2 coronavirus sequences using informative subtype markers for pandemic..., Zhao [/bib_ref]. The China National Center for Bioinformation aligned 77,801 genome sequences of SARS-CoV-2 that were detected globally and identified a total of 15,018 mutations [bib_ref] Characteristics of SARS-CoV-2 and COVID-19, Hu [/bib_ref]. Studies have shown that mutations play an important role in the virulence and infectivity of SARS-CoV-2, although no significant association was found between mutations and outcomes pertaining to hospitalization or death [bib_ref] Mutations strengthened SARS-CoV-2 infectivity, Chen [/bib_ref] [bib_ref] Tracking changes in SARS-CoV-2 spike: evidence that D614G increases infectivity of the..., Korber [/bib_ref]. Thus, it is unclear whether the different clinical outcomes of patients with COVID-19 between the epicenter and peripheral areas affected by the pandemic are due to mutations in SARS-CoV-2.
This large-sample comparative study provides informative insights into the differences in epidemiology, clinical characteristics and outcomes of patients with COVID-19 between the epicenter (Wuhan) and peripheral (Sichuan) areas of the pandemic. However, there are several limitations that need to be addressed. First, due to the retrospective study design, data generation was clinically driven, and not all laboratory data were available for all patients. Accordingly, the missing data for some patients may have biased the findings. Second, the Sichuan cohort, which represented the peripheral area of the COVID-19 pandemic, was incomplete although consecutive patients accounting for 88.1% of total cases with COVID-19 were recruited from 41 designated hospitals in Sichuan. Third, we did not analyze the genetic diversity of virus strains and the evolutionary history, which might well explain the differences between the epicenter and peripheral areas affected by the pandemic.
# Conclusions
This comparative study found that there were significant differences in the epidemiology, clinical characteristics, and outcomes of patients with COVID-19 between the epicenter and peripheral areas affected by the pandemic. The worse outcomes in the epicenter could be partly explained by the overwhelming of health systems and the delayed time from illness onset to hospitalization. This was associated with elevated systemic inflammation, organ dysfunction and prolonged duration of virus shedding, independent of sex, age, smoking and comorbidities. This has potential implications that are of clinical relevance in interventions for COVID-19. The data suggests that urgent or early supportive care would achieve improved clinical outcomes, leading to a lower death rate although no proven effective therapies currently exist. No differences were found in the epidemiology, clinical characteristics, and outcomes between the first generation and secondary generation of patients in the peripheral area of pandemic. Biological differences accounting for the differences between the Wuhan-related exposure patients in the Sichuan cohort and Wuhan cohort need to be further investigated.
## Supplementary information
The online version contains supplementary material available at https://doi. org/10.1186/s12879-020-05728-7.
Additional file 1: [fig_ref] Table 1: Demographics and clinical characteristics of patients in the Sichuan and Wuhan cohorts [/fig_ref]. Detailed comorbidities of patients in Sichuan and Wuhan cohorts. [fig_ref] Table 2: Treatments and outcomes of the patients in the Sichuan and Wuhan cohorts [/fig_ref]. Regression analysis of the risk factors for death, ICU admission and mechanical ventilation in all patients from the Sichuan and Wuhan cohorts.. Demographics and clinical characteristics of patients in Sichuan sub-cohorts with vs. without Wuhan-related exposure. [fig_ref] Table 4: Correlation of time from illness onset to hospitalization with systemic inflammation and... [/fig_ref]. Outcomes of patients in Sichuan sub-cohorts with vs. without Wuhan-related exposure. . Risk of adverse outcomes in Sichuan sub-cohorts with vs. without Wuhan-related exposure. . Demographics and clinical characteristics of patients in Sichuan sub-cohort with Wuhan-related exposure vs. Wuhan cohort. . Outcomes in Sichuan sub-cohort with Wuhan-related exposure vs. Wuhan cohort. . Risk of adverse outcomes in Sichuan subcohort with Wuhan-related exposure vs. the Wuhan cohort.
[fig] Figure 1: See legend on next page.) [/fig]
[fig] Figure 3 a: Symptomatic burden of patients with COVID-19 between Sichuan cohort and Wuhan cohorts; Kaplan-Meier survival curve for time from illness onset to hospital admission (b), to ICU admission (c) and to discharge (d) of patients with COVID-19 between Sichuan and Wuhan cohorts Wang et al. BMC Infectious Diseases (2021) 21:206 I (> 0.04 ng/mL), alanine aminotransferase (> 50 IU/L), aspartate aminotransferase (> 40 IU/L), procalcitonin (> 0.5 ng/mL) and delayed hospitalization were associated with death, ICU admission and mechanical ventilation. In addition, we found that time from illness onset to hospitalization was associated with prolonged duration of virus shedding (adjusted β=0.11, 95% CI= [0.03, 0.24], P=0.009). We analyzed the relationship between the delay in hospitalization and elevated systemic inflammation and features of organ dysfunction. The time from illness onset to hospitalization was positively correlated with systemic inflammatory cells such as white blood cells (r=0.086, P= 0.004), neutrophils (r=0.089, P=0.003), eosinophils (r=0.116, P< 0.001), platelets (r=0.212, P< 0.001), and inflammatory biomarkers, such as D-dimer (r=0.101, P=0.004), procalcitonin (r=− 0.093, P=0.019), features of organ dysfunction, such as hemoglobin (r=− 0.155, P< 0.001), BUN (r= 0.10, P=0.002), creatine (r=− 0.094, P=0.003), albumin (r=− 0.263, P< 0.001), and APTT (r=− 0.247, P< 0.001) ( [/fig]
[table] Table 3: The results showed that the Wuhan cohort had higher risk of death (aOR=7.64, 95% CI=[2.31, 25.27], P=0.001), ICU admission (aOR=1.66, 95% CI=[1.05, 2.63], P=0.031), delayed time from illness onset to hospital (aOR=6.29, 95% CI= [4.70, 8.40], P< 0.001) and ICU admission (aOR=8.03, 95% CI=[1.74, 37.06], P< 0.001) admissions, prolonged duration of viral shedding after COVID-19 onset (aOR= 1.64, 95% CI=[1.15, 2.33], P=0.006), a decreased hospital stay (aOR=0.41, 95% CI=[0.32, 0.53], P< 0.001) after adjusting for age, sex, smoking status and the CCI. There was no difference in time from illness onset to discharge (aOR=0.99, 95% CI=[0.77, 1.28], P=0.968) after adjusting for these confounders. When we additionally [/table]
[table] Table 1: Demographics and clinical characteristics of patients in the Sichuan and Wuhan cohorts [/table]
[table] Table 2: Treatments and outcomes of the patients in the Sichuan and Wuhan cohorts [/table]
[table] Table 4: Correlation of time from illness onset to hospitalization with systemic inflammation and features of organ dysfunction [/table]
|
Dichlorodiphenyldichloroethane burden and breast cancer risk: a meta-analysis of the epidemiologic evidence.
The relationship of dichlorodiphenyltrichloroethane (DDT) exposure and breast cancer risk has received increasing attention since the beginning of the 1990s. Contradicting published results regarding the relationship between body burden levels of p,p´-dichlorodiphenyldichloroethane (p,p´-DDE)-the main DDT metabolite-and breast cancer, we argue that such differences stem from methodologic differences among those studies. We performed a meta-analysis of 22 articles using DerSimonian and Laird's method for random effects models. The Q-statistic was used to identify heterogeneity in the outcome variable across studies. The gradient of p,p´-DDE exposure in epidemiologic studies was homogenized to serum lipid bases (nanograms per gram). The potential for publication bias was examined by means of the Begg's test. We discuss methodologic features of the studies in an attempt to reconcile the findings. The summary odds ratio (OR) for selected studies was 0.97 (95% confidence interval, 0.87-1.09) and the gradient of exposure ranged from 84.37 to 12,948 ng/g. No overall heterogeneity in the OR was observed (χ 2 = 27.93; df = 23; p = 0.218). Neither the study design nor the lack of breast-feeding control or the type of biologic specimen used to measure p,p´-DDE levels were the causes of heterogeneity throughout the studies. Evidence for publication bias was not found (p = 0.253). Overall, these results should be regarded as a strong evidence to discard the putative relationship between p,p´-DDE and breast cancer risk. Nevertheless, the exposure to DDT during critical periods of human development-from conception to adolescence-and individual variations in metabolizing enzymes of DDT or its derivatives are still important areas to be researched in regard to breast cancer development in adulthood. Key words: breast cancer, DDT, epidemiology, evidence, meta-analysis.
The insecticide dichlorodiphenyltrichloroethane (DDT) was first synthesized in 1874 and reached worldwide use during the early 1960s to control malaria and some agricultural pests. In 1972 the use of DDT was banned in the United States, and by the beginning of the 1980s this chemical was prohibited in most developed countries. However in India, Indonesia, and Italy, DDT was still produced until 1990, and in Mexico DDT was in use until 1997 to control malaria [bib_ref] Dichlorodiphenyltrichloroethane (DDT): ubiquity, persistence, and risks, Turusov [/bib_ref]. The long persistence and environmental spreading exhibited by DDT and its metabolites, along with their estrogenic potential, are the main concerns regarding its potential role in the etiology of estrogen-related malignant tumors [bib_ref] Pesticides and breast cancer risk: a review of DDT, DDE, and dieldrin, Snedeker [/bib_ref].
In 1993, [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] first reported on the presumed positive association between p,p´-dichlorodiphenyldichloroethane (p,p´-DDE)-the main metabolite of DDTand breast cancer, and subsequently the assessment of the body burden of DDT metabolites in relation to breast cancer risk received a lot of attention. During the last decade, results from several epidemiologic studies were published but most were unable to replicate the positive association between p,p´-DDE and breast cancer risk. Among other explanations for such negative results, some suggestedbut never proved-that single studies lacked an adequate gradient of exposure to p,p´-DDE among breast cancer cases and controls, and this flaw obscured the differences [bib_ref] Organochlorine residues and breast cancer, Talbott [/bib_ref]. As demonstrated below, the gradient exposure in 22 published studies ranged from 84.37 to 12928.08 ng/g [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] , in sharp contrast to the narrower gradients achieved by most single studies.
Serum and adipose tissue were the human biologic matrices used to estimate p,p´-DDE body burden and its potential relationship with breast cancer risk. Circulating lipids influence blood levels of DDT metabolites [bib_ref] Chlorinated hydrocarbon levels in human serum: effects of fasting and feeding, Phillips [/bib_ref] , yet approaches to this condition varied greatly in the published scientific literature. Some studies reported p,p´-DDE serum levels in lipid bases [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Organochlorine exposure and breast cancer survival, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Serum organochlorine levels and breast cancer: a nested case-control study of Norwegian..., Ward [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] , whereas others performed an indirect adjustment by fitting a cholesterol term in linear regression models [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] and the rest only provided wet-based measurements [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Organochlorine exposure and breast cancer risk in Colombian women, Olaya-Contreras [/bib_ref] [bib_ref] Blood levels of DDT and breast cancer risk among women living in..., Schecter [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref]. This heterogeneity among biologic matrices and reported units of cumulative p,p´-DDE levels limited the ability to evaluate the gradient of p,p´-DDE body burden levels across the epidemiologic studies published so far.
In this article we aim to a) estimate the strength of the association between p,p´-DDE and breast cancer on the basis of published epidemiologic evidence; b) identify the gradient of exposure that was captured in the same epidemiologic studies; and c) discuss the consistency of published results in the context of their main methodologic features.
# Materials and methods
We searched for the epidemiologic evidence on p,p´-DDE and breast cancer risk in both the MEDLINE and PubMed databases (www.ncbi.nlm.nih.gov). A total of 35 analytic studies [bib_ref] Organochlorine pesticide content of breast adipose tissue from women with breast cancer..., Bagga [/bib_ref] [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] High organochlorine body burden in women with estrogen receptor-positive breast cancer, Dewailly [/bib_ref] [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] A population-based case-control study of farming and breast cancer in North Carolina, Duell [/bib_ref] [bib_ref] Pesticides and polychlorinated biphenyl residues in human breast lipids and their relation..., Falck [/bib_ref] [bib_ref] Chlororganic pesticides and polychlorinated biphenyls in breast tissue of women with benign..., Güttes [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Organochlorine exposure and breast cancer survival, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] -bis(pchlorophenyl) ethylene and polychlorinated biphenyls and breast cancer: combined analysis of five..., Laden [/bib_ref] [bib_ref] Case-control study on breast cancer and adipose tissue concentrations of congener specific..., Liljegren [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Organochlorine exposure and breast cancer risk in Colombian women, Olaya-Contreras [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Blood levels of DDT and breast cancer risk among women living in..., Schecter [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] Organochlorine compounds in the adipose tissue of deceased persons with and without..., Unger [/bib_ref] [bib_ref] Organochlorine compounds in human breast fat from deceased with and without breast..., Unger [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] in English up to February 2001 were found using the following MeSH headings, key, and text words: breast cancer, organochlorines, pesticides. The articles identified were then reviewed to determine whether they met the following inclusion criteria for statistical analyses: to be epidemiologic cohort or case-control studies; to have enrolled at least 50 cases; to have reported measures of association and confidence intervals (CIs) for breast cancer risk; to have measured p,p´-DDE levels in biologic samples (serum or adipose tissue); and to have been published in journals listed by the Journal Citation Reports-Science Edition (JCR) .
Under the former considerations, six articles were discarded because no measures of association were reported [bib_ref] High organochlorine body burden in women with estrogen receptor-positive breast cancer, Dewailly [/bib_ref] [bib_ref] Pesticides and polychlorinated biphenyl residues in human breast lipids and their relation..., Falck [/bib_ref] [bib_ref] Chlororganic pesticides and polychlorinated biphenyls in breast tissue of women with benign..., Güttes [/bib_ref] [bib_ref] Organochlorine compounds in human breast fat from deceased with and without breast..., Unger [/bib_ref] [bib_ref] Organochlorine compounds in the adipose tissue of deceased persons with and without..., Unger [/bib_ref] [bib_ref] Organochlorine compounds in neoplastic and adjacent apparently normal breast tissue, Wassermann [/bib_ref] , one more was eliminated because it lacked CIs , another article was excluded because the researchers evaluated survival instead of breast cancer risk [bib_ref] Organochlorine exposure and breast cancer survival, Hoyer [/bib_ref] , another one did not measure p,p´-DDE levels in biologic samples , one was not published in a JCR journal [bib_ref] Organochlorine exposure and breast cancer risk in Colombian women, Olaya-Contreras [/bib_ref] , two reported on < 50 breast cancer cases [bib_ref] Case-control study on breast cancer and adipose tissue concentrations of congener specific..., Liljegren [/bib_ref] [bib_ref] Blood levels of DDT and breast cancer risk among women living in..., Schecter [/bib_ref] and one study used a cross-sectional design [bib_ref] Organochlorine pesticide content of breast adipose tissue from women with breast cancer..., Bagga [/bib_ref]. These 13 epidemiologic studies were discarded in this step and 22 were kept for further analyses [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref].
From each eligible report and using a predefined review form, two independent reviewers extracted the following methodologic information: name of the author, year and place of publication, epidemiologic design, type of controls and biologic specimens, confounding variables considered in the analysis, and the measure of association estimated for the highest versus lowest category of exposure along with the corresponding CI.
After their extraction, we entered relevant data into evidence tables. We then performed a meta-analysis using the method of the inverse of variance for fixed-effects models and the DerSimonian and Laird method for random-effects models. (DerSimonian and Laird 1986). Separate odds ratios (ORs) were used in the meta-analysis for one article that reported estimates from population-based and hospital controls , and the same was done with estimates from one study in which serum samples were taken and analyzed for two different periods of time . The results are displayed as summary ORs and 95% CIs for the effect of p,p´-DDE on breast cancer, corresponding to the contrast of the highest versus the lowest level of p,p´-DDE exposure.
We plotted the outcomes for included studies for visual examination and performed meta-analysis regression using the Q-statistic to identify heterogeneity in the outcome variable across studies [bib_ref] A randomeffects regression model for meta-analysis, Berkey [/bib_ref] [bib_ref] Meta-analysis in clinical trials, Dersimonian [/bib_ref]. Potential sources of heterogeneity were evaluated, including the study design, control for breastfeeding and the kind of biologic specimen in which the DDT metabolites were measured. We assessed the potential for publication bias using a funnel plot in conjunction with the Begg's test, which is based on the fact that smaller studies tend to have larger effect size estimates and the publication bias induces a correlation between the effect estimates and their variances [bib_ref] A measure to aid in the interpretation of published clinical trials, Begg [/bib_ref] [bib_ref] Operating characteristics of a rank correlation test for publication bias, Begg [/bib_ref].
To estimate the trend of p,p´-DDE body burden evaluated by the epidemiologic studies analyzed, we determined the crude mean p,p´-DDE levels among cases and controls reported by each study and homogenized Abbreviations: BMI, body mass index; ln, natural logarithm. a In design or analysis. b Date in which blood sample was returned, time of day that blood sample was obtained, fasting status at blood sampling and for postmenopausal homone use, BMI at blood collection, history of benign breast disease. c Number and dates of blood donations, day of menstrual cycle for premenopausal women, ln height, ln (BMI) -menopausal status at blood donation interaction. d Vital statistics at time of diagnosis and weight. e Year of blood draw and history of benign disease at the time of diagnosis. f Date of blood donation and day of menstrual cycle, race, BMI at age 20 or current. g Date of examination and vital status at the time of diagnosis, weight, height, alcohol consumption, smoking, physical activity, income, marital status, and education. h Serum lipids, month in which the blood sample was returned, time of day that the blood sample was obtained, fasting status at blood sampling, postmenopausal hormone use, history of benign breast disease, BMI. i Date of examination, length of follow-up after examination, race, date of joining the Kaiser Permanente Medical Care Program, year of multiphasic examination and BMI. j Number and date of blood donation, if premenopausal women: day of menstrual cycle at the time of the first blood drawing. them to serum p,p´-DDE levels in lipid bases (nanograms per gram) as follows: the arithmetic mean serum levels of p,p´-DDE in wet bases (nanograms per milliliter) were multiplied by a factor of 129.8 to convert them to the arithmetic mean of serum levels in lipid basis (nanograms per gram), and otherwise the arithmetic means of adipose tissue levels of p,p´-DDE were divided by a factor of 4.2 to estimate the corresponding serum levels in lipid basis (nanograms per gram) (López- .
The percent of recovery of p,p´-DDE levels was not considered. Five articles did not provide mean values of p,p´-DDE and thus were not included [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] ; also not included were two others that reported adjusted mean values of p,p´-DDE [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] and three in which the p,p´-DDE levels were statistically modeled through the contents of triglycerides, serum, and total cholesterol (Dello [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref]. Therefore, we included 12 studies in this step of the analysis. We estimated the trend of the mean p,p´-DDE body burden levels in nanograms per gram according to the year when the biologic samples were collected by linear regression.
To evaluate the gradient of p,p´-DDE body burden captured by studies of interest, we plotted the middle point of p,p´-DDE levels in nanograms per gram in serum (according to the methodology already described) for each category of exposure, against the corresponding ORs reported by 17 studies. We did not include two studies because no information on the magnitude of p,p´-DDE quartile distribution was provided [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] and three studies in which p,p´-DDE was lipid-adjusted by regression methods [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref]. All the statistical analyses were performed using the software Stata release 7.0 (Stata Corp., College Station, TX, USA).
# Results
Tables 1, 2 and 3 describe the 22 studies that were included in the meta-analysis. All were case-control studies, and of these nine were prospective (nested case-control) [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] and 13 retrospective [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref]. Among the retrospective studies, four were populationbased case-control studies [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] and seven were clinic-based case-control studies [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] ; in one study only a subsample of a populationbased case-control study population was analyzed ; and another study included two types of controls: population and clinical . All the studies are presented in the tables according to decreasing date of publication and design features.
Thirteen studies were conducted in the United States [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] , two in Canada [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] , two in Denmark [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] , two in Mexico [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] , one in Italy (Dello [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] , one in Germany, Netherlands, Ireland, Switzerland, Spain [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] , and one more in Brazil ). The number of cases ranged from 58 in the study performed by [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] to 456 in the study performed by [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref].
The inclusion criteria of the referent groups varied among studies. Individuals with skin cancer were accepted in the control group by some studies, [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] whereas in others controls had been diagnosed with benign breast disease [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] , and in the remaining studies only healthy individuals and/or subjects with no cancer diagnosis made up the comparison group [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref].
In most studies, body burden levels of p,p´-DDE were measured in serum samples [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] , but in four studies, adipose tissue from the breast [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] Article | A meta-analysis of DDT and breast cancer risk Environmental Health Perspectives - VOLUME 112 | NUMBER 2 | February 2004 209 or the buttocks (van´t were the biologic matrices chosen to estimate the cumulative exposure to p,p´-DDE. The collection of biologic specimens dated back to about 10-25 years before the diagnosis of breast cancer in the prospective case-control studies [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] to the period immediately around the date of diagnosis in all retrospective studies [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref].
The results of all studies were controlled by the age of the participants. The control of other potential confounders, either in the design or the analysis, was distinct across the studies. History of breast-feeding was not considered in six studies [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] , and parity and menopausal status were controlled in most studies but not in two [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref]. History of familial breast cancer and/or benign breast disease was controlled in most of the studies [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] as well as body mass index [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref]. Other adjustment variables were fasting status [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] , day of menstrual cycle at the date of blood sampling [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] , vital and/or income status [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] , physical activity [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] , use of hormonal replacement therapy [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] , tobacco smoking [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] , and alcohol consumption [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] as well as intake of fruits, vegetables [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] , and fat [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref].
Overall, the data provided by the published studies do not support an association between p,p´-DDE body burden levels and breast cancer risk, because the summary OR was 0.97 (95% CI, 0.87-1.09) . We found no evidence for significant overall heterogeneity in the OR [χ 2 = 27.93; degrees of freedom (df) = 23; p = 0.218].
Summary ORs for prospective and retrospective population-based case-control and retrospective hospital-based case-control studies were 0.91 (95% CI, 0.74-1.12), , and 0.93 (95% CI, 0.77-1.12), respectively . Although summary ORs did not show significant heterogeneity within prospective or retrospective hospital-based case-control studies (χ 2 = 10.68; df = 9; p = 0.298, and χ 2 = 4.107; df = 7; p = 0.767, respectively), we found a borderline statistically significant test of heterogeneity within retrospective population-based case-control studies (χ 2 = 11.23; df = 5; p = 0.047) in the study performed by.
Summary ORs were not different for the 16 studies where breast-feeding was controlled as confounder [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] 1999, 2000) and the estimate was 1.01 (95% CI, 0.88-1.16), compared to the OR for studies in which breast-feeding was uncontrolled (OR = 0.87; 95% CI, 0.68-1.10) [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref]. There were no differences either for the 18 studies using blood serum as biologic specimen 1.00 (95% CI, 0.88-1.14) (Dello [bib_ref] Organochlorines and breast cancer. A study on Neapolitan women, Iacovo [/bib_ref] [bib_ref] Risk and aggressiveness of breast cancer in relation to plasma organochlorine concentrations, Demers [/bib_ref] [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] [bib_ref] Organochlorine exposure and risk of breast cancer, Hoyer [/bib_ref] [bib_ref] Repeated measurements of organochlorine exposure and breast cancer risk (Denmark), Hoyer [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] Plasma organochlorine levels and the risk of breast cancer: an extended follow-up..., Laden [/bib_ref] [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref] [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] [bib_ref] Dichlorodiphenyldichloroethene, polychlorinated biphenyls, and breast cancer among African-American and white women in..., Millikan [/bib_ref] [bib_ref] Environmental organochlorine exposure and postmenopausal breast cancer risk, Moysich [/bib_ref] [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref] [bib_ref] Organochlorine exposures and breast cancer risk in New York City women, Wolff [/bib_ref] [bib_ref] Risk of breast cancer and organochlorine exposure, Wolff [/bib_ref] [bib_ref] Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2´-bis(p-chlorophenyl)ethylene, Zheng [/bib_ref] or for studies that used adipose tissue 0.84 (95% CI, 0.62-1.13) [bib_ref] Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and..., Stellman [/bib_ref] [bib_ref] DDE and DDT in breast adipose tissue and risk of female breast..., Zheng [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref] [fig_ref] Table 4: Overall ORs for breast cancer risk and p,p´-DDE body burden levels [/fig_ref]. No evidence of publication bias was found (p = 0.253).
In we depict a significantly decreasing trend (β = -130.59; p = 0.001) of the mean levels of p,p´-DDE that were evaluated by the studies analyzed, according to the date when the biologic sample was collected. All those levels were converted to the corresponding equivalent nanograms per gram in lipid serum bases. Significantly, p,p´-DDE levels as reported by the study carried out in Mexico City by [bib_ref] Breast cancer, lactation history, and serum organochlorines, Romieu [/bib_ref] , were at great variation with the other studies performed at about the same time, including the one performed in that same city [bib_ref] Dichlorodiphenyltrichloroethane serum levels and breast cancer risk: a case-control study from Mexico, López-Carrillo [/bib_ref].
In we present the ORs for the effect of p,p´-DDE on breast cancer risk from each study, according to the gradient of the p,p´-DDE body burden levels expressed as nanograms per gram in serum lipid bases (middle point of p,p´-DDE levels for each category of exposure). The range of that gradient varied from 84.37 ng/g in the study performed by [bib_ref] Organochlorines and breast cancer: a case-control study in Brazil, Mendonca [/bib_ref] , to 12928.08 ng/g in the study by [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref]. As shown in , most studies reported p,p´-DDE body burden levels in the range of 84.37-9,000 ng/g and yielded negative results, with two exceptions [bib_ref] Blood levels of organochlorine residues and risk of breast cancer, Wolff [/bib_ref]. The studies that had the highest levels of p,p´-DDE body burden levels (9,001-12928.08 ng/g) did not show an increasing risk of breast cancer due to p,p´-DDE body burden levels [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] [bib_ref] DDT (dicophane) and postmenopausal breast cancer in Europe: case-control study, Veer [/bib_ref].
# Discussion
The results of the meta-analysis of 22 studies showed no evidence for an association between p,p´-DDE body burden levels and breast cancer risk. The summary OR reported in this manuscript was 0.97 (95% CI, 0.87-1.09), very similar to the one recently estimated from a pooled analysis of five studies (OR = 0.99; 95% CI 0.77-1.27) performed in the United States [bib_ref] -bis(pchlorophenyl) ethylene and polychlorinated biphenyls and breast cancer: combined analysis of five..., Laden [/bib_ref] and had the same covariates to be included in the logistic models. Some studies [bib_ref] Plasma organochlorine levels and the risk breast cancer, Hunter [/bib_ref] [bib_ref] -bis(pchlorophenyl) ethylene and polychlorinated biphenyls and breast cancer: combined analysis of five..., Laden [/bib_ref] apparently made repeated use of some subjects as part of their study populations in subsequent papers. Hence, one additional check was to remove in subsequent steps those studies for which we believed that such condition could be met, and the estimates of summary ORs remained almost unaltered (data not shown).
An intrinsic flaw in many environmental epidemiologic studies is the lack of an adequate gradient of exposure both within and throughout the different populations studied. On average, the difference between the highest and the lowest levels of p,p´-DDE in the 22 studies was 6928.92 ± 6414.5 ng/g. The studies that had the widest internal gradient of exposure were negative [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref] as were the studies that reported the highest levels of p,p´-DDE. [bib_ref] Serum organochlorine pesticides and PCBs and breast cancer risk: results from a..., Dorgan [/bib_ref] [bib_ref] Breast cancer and serum organochlorines: a prospective study among white, black, and..., Krieger [/bib_ref]. In this regard, evidence from occupational studies, which evaluated much higher levels of p,p´-DDE exposure, does not suggest a high risk for breast cancer [bib_ref] A prospective follow-up study of cancer mortality in relation to serum DDT, Austin [/bib_ref] [bib_ref] Cancer incidence in a cohort of licensed pesticide applicators in Florida, Fleming [/bib_ref]. Hence, we believe we can rule out the possibility that contradictory results among the 22 studies are caused by differences in p,p´-DDE levels.
Methodologic features among the studies that may explain the contradictory results include differences in the temporal relationship between the measurement of p,p´-DDE Article | A meta-analysis of DDT and breast cancer risk Environmental Health Perspectives - VOLUME 112 | Confounding is a potential explanation for inconsistent epidemiologic results. Among the confounders that might distort the relationship between p,p´-DDE and breast cancer risk are breast-feeding and diet. Lactation is a way of eliminating the body burden levels of p,p´-DDE (López- [bib_ref] Breastfeeding and serum p,p´-DDT levels among Mexican women of childbearing age. A..., Carrillo [/bib_ref] and has been found to decrease the risk of breast cancer in several studies [bib_ref] Breast cancer and lactation history in Mexican women, Romieu [/bib_ref]. However, we found no heterogeneity in our meta-analysis to assign explanatory relevance to the lack of control by this variable in some of the published studies. Yet this analysis is not enough to rule out the possibility that equivocal results might be partially explained by differences in the ranges of values for the adjustment variables across the studies; moreover, measurement error in the confounding variables is likely to result in unpredictably biased estimates of effect for the main variable of interest when adjustments are performed.
Residues of p,p´-DDE were reportedly found in several foods (fish, dairy products, meat) (Galván- [bib_ref] Food consumption and adipose tissue DDT levels in Mexican women, Portillo [/bib_ref] , and the consumption of some of them may be related to breast cancer risk. For example, meat intake is related to an excess risk for breast cancer, whereas fish intake, presumably because of the presence of omega-3 fatty acids, seems to be inversely related to breast cancer incidence . A study performed by [bib_ref] Curcumin and genistein, plant natural products, show synergistic inhibitory effects on the..., Verma [/bib_ref] showed that genistein, an isoflavonoid present in soybeans, and curcumin, a component of turmeric powder and also a widely used spice, can inhibit the action of pesticides with estrogenic activity. The great variation in breast cancer risk raises the possibility that dietary factors are related to its etiology. In this regard, dietary factors and particularly specific compounds such as phytoestrogen were scarcely or not at all taken into account as covariates in studies of p,p´-DDE and breast cancer risk; thus, the lack of adjustment by these variables might partially explain the equivocal results so far available.
Another methodologic issue of concern is the type of controls that were enrolled-i.e., hospital or population based-in that one should expect that p,p´-DDE body burden levels were not related to the diseases identified among clinical controls, and also that population controls actually constitute a representative sample of the p,p´-DDE body burden levels present in the target population. We were not able to find heterogeneity according to the study design, except for a borderline significant result within retrospective population-based case-control studies (χ 2 = 11.23; df = 5; p = 0.047). And this variation mainly arose from the study by, which assembled a referent group that combined hospital-and populationbased controls and provided no explanation for such an unusual combination. (2000) contributed data for clinical and population controls, which were considered separately. That was also the case for [bib_ref] Serum concentrations of organochlorine compounds and the subsequent development of breast cancer, Helzlsouer [/bib_ref] , in which serum samples were reported for two different moments in time: 1974 and 1989. Thus the sample size for the analysis became 24.
[fig] Figure 1, Figure 2: Accumulated meta-analysis; summary OR = 0.97 (95% CI, 0.87-1.09). a Biologic samples taken in 1974. b Biologic samples taken in 1989. c Controls are population based. d Controls are clinical based. al. (1993) Krieger et al. (1994) Hunter et al. (1997) López-Carillo et al. (1997) van't Vear et al. (1997) Hoyer et al. (1998) Moysich et al. (1998) Dello Iacovo et al. (1999) Dorgan et al. (1999) Helzlsouer et al. (1999) a Helzlsouer et al. (1999) b Mendonca et al. (1999) Zheng et al. (1999) Aronson et al. (2000) Demers et al. (2000) c Demers et al. (2000) d Hoyer et al. (2000b) Millikan et al. (2000) Romieu et al. (2000) Stellman et al. (2000) Wolff et al. (2000a) Wolff et al. (2000b) Zheng et al. (2000) Laden et al. Meta-analysis according to type of design: (A) prospective studies (nested case-control); (B) retrospective studies (population-based case-control); (C) retrospective studies (hospital-based case-control). a Biologic samples taken in 1974. b Biologic samples taken in 1989. c Controls are population based. d Controls are clinical based.Wolff et al. (1993) Krieger et al. (1994) Hunter et al. (1997) Hoyer et al. (1998) Dorgan et al. (1999) Helzlsouer et al. (1999) a Helzlsouer et al. (1999) b Hoyer et al. (2000b) Wolff et al. (2000b) Laden et al. al. (2000) c Millikan et al. (2000) Romieu et al. (2000) López-Carillo et al. (1997) Mendonca et al. (1999) Zheng et al. (1999) Aronson et al. (2000) Demers et al. (2000) d Stellman et al. (2000) Wolff et al. (2000a) Zheng et al. (2000) A B C [/fig]
[fig] Figure 3, Figure 4 p: Mean levels of p,p´-DDE (ng/g) in 12 reported studies according to the date of sample collection. ∆, fitted values. a Biologic samples taken in 1974. b Biologic samples taken in 1989. Effect of p,p´-DDE on breast cancer risk from each study according to p,p´-DDE (ng/g) body burden levels. ln, natural logarithm. a Biologic samples taken in 1974. b Biologic samples taken in 1989. c Controls are population based. d Controls are clinical based. ,p'-DDE serum levels (ng/g) ln OR [/fig]
[table] Table 1: Article | López-Cervantes et al. VOLUME 112 | NUMBER 2 | February 2004 • Environmental Health Perspectives Prospective epidemiologic studies (nested case-control) on p,p´-DDE and breast cancer risk. [/table]
[table] Table 2: Retrospective epidemiologic population-based case-control studies on p,p´-DDE and breast cancer risk.BMI, body mass index. a In design or analysis. b Hormone replacement treatment, income and race, BMI, age. c Region of residence and history of benign breast disease, BMI. d DDT serum levels, BMI, socioeconomic status. e BMI, cholesterol. f Fruit and vegetable intake, lipid serum, education, and BMI. g Hospital controls and population controls were used together. h Alcohol consumption, study site, and BMI. [/table]
[table] al: Article | López-Cervantes et al. VOLUME 112 | NUMBER 2 | February 2004 • Environmental Health Perspectives [/table]
[table] Table 3: Retrospective epidemiologic (hospital-based case-control) studies on p,p´-DDE and breast cancer risk. [/table]
[table] Table 4: Overall ORs for breast cancer risk and p,p´-DDE body burden levels. In total, 22 studies were useful for the purposes of this table; Demers et al. [/table]
|
Neural spatio-temporal patterns of information processing related to cognitive conflict and correct or false recognitions
A.1 BOLD activities mapFig. A.1shows BOLD maps from GLM analysis for representing different brain activities for POScorr > LURfalse (a) and LURfalse > POScorr (b) contrasts.
On the other hand, the use of all features available might be misleading for classification and lead to overfitting when a model starts to fit non-important features. Use of Shapley methodology makes it possible to select features that are the most informative. It suffices to build a model using all the features available (represented with leftmost points on the graphs in [fig_ref] Figure A. 4: Mean AUC values [/fig_ref] and compute their Shapley values once. Then, it is possible to select only the features with the highest absolute values to build new models. It can be seen that the best models with the highest mean AUC values use only 10-20% of all the features. For the AAL parcellation the best models use 100-200 features (out of 10 × 116 = 1160 total), while for the MMP/CA parcellation it is about 300 (out of 10 × 718 = 7180).
## A.5 poscorr-lurfalse relevant regions
Tables A.1 and A.2 show the early 0-5 TR and late 5-9 TR window ∆ Area relevant regions for the POScorr-LURfalse experiment. To compare with Shapley/GradientBoosting selected features, see. The critical values for the ∆ Area are computed with bootstrap. Inregions with the highest Shapley sums for the early and late time windows were selected. It can easily be seen that the most important regions using each approach match together. It is important to remember that in case of ∆ Area in temporal analysis approach an additional condition was for the region to have a maximum in the time 2/9 window, while the ML uses the feature values although separately in each TR, but in the context of other TRs (see and the discussion of the early and late stage of neural responses in Results section in the main of the paper). The maxima requirement is also why in.2 only Default regions are found, while in, part for the late window, several others are found. Default-72_L-Ctx PCUN L 108 -9.6 -50.7 37.2 0.45 ± 0.13 R_a24
Default-07_R-Ctx ACC R 82 5.9 39.5 -1.6 0.44 ± 0.13
## A.7 comparison of complementary regions for poscorr-lurfalse and negcorr-lurcorr contrasts
To compare the two contrasts, which are the same in terms of different levels of cognitive conflict, but differ in the context of correctness, we depicted the mean signals of the same regions in both contrasts.
[fig] Figure A. 1: BOLD activity on the contrasts: a -correct recognition of positive probe versus false recognition of lure probe, b -false recognition of lure probe versus correct recognition of positive probe A.2 Task procedure Fig. A.2 shows the example trial of the task presented to the participants. [/fig]
[fig] Figure A. 2, Figure A. 3: The sequence of stimuli in the task with an example set of abstract stimuli, POS -positive probe, LUR -lure probe, NEG -negative probe.A.3 Accuracy statisticsFig. A.3give the accuracy statistics, together with reaction times, of all the response types considered in the paper. The accuracy statistics (left side) and reaction times (right side) of all the response types which we considered in our analysis.A.4 Gradient boosting trainingThe gradient boosting training was performed using the Catboost implementation 21 .Fig. A.4shows the Area Under Curve (AUC) values (vertical axis; the higher, the better) from 5-fold cross validations. The values are for AAL (top) and MMP/CA (bottom row) data parcellations of POScorr-LURfalse (left) and NEGcor-LURcorr (right column) experiments. Several models were found by reducing the number of features used by using the Shapley importance values. These values were computed only once at the beginning, therefore the following feature selection was fair. With all features available, the gradient boosting approach (see Methods for more details) gives results comparable to other approaches (compare the untuned Gradient Boosting bar inFig. 1in the main part). [/fig]
[fig] Figure A. 4: Mean AUC values (vertical axis) from 5-fold cross validations for the AAL (top) and MMP/CA (bottom row) parcellations with gradient boosting models that use different number of features. Results are shown for the POScorr-LURfalse (left) and NEGcorr-LURcorr (right column). Best models are obtained with 100-200 features for the AAL map, and around 300 features for the MMP/CA parcellation. [/fig]
[fig] 4/ 9, Figure A. 5: Fig. A.5 showed the time courses of brain areas complementary to regions presented in the manuscript. Time courses of brain regions complementary to regions in Fig. 2a, 2b, and 4 to compare the both contrasts (POScorr-LURfalse and NEGcorr-LURcorr). The first column (on the left) is complementary to regions in Fig. 2a, 2b, and the remaining regions -to Fig. 4. [/fig]
[table] Table A 1: Regions in the MMP/CA parcellation which are statistically relevant in the early period (0-5 TR) after the POScorr and LURfalse retrieval events and the mean activations for both events have local maxima in the range 1-5 TR. ∆ Area is the area between average activations in the range 0-5 TR. The threshold of statistical significance from permutation tests is 0.5453. The errors are estimated by bootstrap. Abbreviations: ACC -anterior cingulate cortex, CB -cerebellum, IFG -inferior frontal gyrus, INS -insula, IOG -inferior occipital gyrus, IPG -inferior parietal gyrus, MOG -middle occipital gyrus, SFG -superior frontal gyrus, SMA -supplementary motor area, STG -superior temporal gyrus, L -left hemisphere, R -right hemisphere. TR late stages for NEGcorr and LURcorr retrieval events. The given regions have their ∆ Area values above appropriate critical values computed using permutation tests (see Methods). The additional requirement, to be consistent with the proposed mean signal analysis, is for the regions to have their ∆ Area local maxima in the appropriate TR window. The Shapley relevant regions for both early and late regions are given inTable A.5. Because of the maximum requirement for the mean analysis,Table A.4 is composed of only Default regions, while the late stage important regions for the gradient boosting machine learning approach using Shapley approach are composed of other regions too. This is because machine learning approach uses all features, i.e., regions in their TRs, separately, although in context of other features. [/table]
[table] Table A 2: Regions in the MMP/CA parcellation which are statistically relevant in the late period (5-9 TR) after the POScorr and LURfalse retrieval events and the mean activations for both events have local maxima in the range 5-9 TR. ∆ Area is the area between average activations in the range 5-9 TR. The threshold of statistical significance from permutation tests is 0.4138. The errors are estimated by bootstrap. Abbreviations: ACC -anterior cingulate cortex, CAL -calcarine gyrus, MCC -middle cingulate gyrus, MFGorb -medial frontal gyrus orbital part, MOG -middle occipital gyrus, PCC -posterior cingulate cortex, PCUN -precuneus, SFG -superior frontal gyrus, L -left hemisphere, R -right hemisphere. [/table]
[table] Table A 3: Regions in the MMP/CA parcellation which are statistically relevant in the early period (0-5 TR) after the NEGcorr and LURcorr retrieval events and the mean activations for both events have local maxima in the range 1-5 TR. ∆ Area is the area between average activations in the range 0-5 TR. The threshold of statistical significance from permutation tests is 0.5280. The errors are estimated by bootstrap. Abbreviations: MTG -middle temporal gyrus, INS -insula, MOG - [/table]
[table] Table A: 4. Regions in the MMP/CA parcellation which are statistically relevant in the late period (5-9 TR) after the NEGcorr and LURcorr retrieval events and the mean activations for both events have local maxima in the range 5-9 TR. ∆ Area is the area between average activations in the range 5-9 TR. The threshold of statistical significance from permutation tests is 0.4162.The errors are estimated by bootstrap. Abbreviations: ANG -angular gyrus, MCC -middle cingulate gyrus, MFG -middle frontal gyrus, MOG -middle occipital gyrus, MTG -middle temporal gyrus, PCC -posterior cingulate cortex, PCUNprecuneus, L -left hemisphere, R -right hemisphere. Table A.5. Most relevant regions for MMP/CA parcellation NEGcorr-LURcorr problem with Shapley value Sh sums given for 0-4 TR and 5-9 TR time windows. Most relevant regions for AAL coded POScorr-LURfalse problem with Shapley value Sh sums for the regions R T R found to be most important in the 0-4 TR and 5-9 TR time windows. Abbreviations: ACC -anterior cingulate cortex, ANG -angular gyrus, CAL -calcarine gyrus, DCN -diencephalon, FFG -fusiform gyrus, HIPPhippocampus, IFG -inferior frontal gyrus, INS -insula, IOG -inferior occipital gyrus, IPG -inferior parietal gyrus, MOGmiddle occipital gyrus, MTG -middle temporal gyrus, PCUN -precuneus, PRE -precentral gyrus, SFG -superior frontal gyrus, SMG -supramarginal gyrus, SPG -superior parietal gyrus, STG -superior temporal gyrus, L -left hemisphere, Rright hemisphere. [/table]
|
Need for ensuring care for neuro-emergencies—lessons learned from the COVID-19 pandemic
Background To investigate whether patients with critical emergency conditions are seeking or receiving the medical care that they require, we characterized the reality of care for patients presenting with neuro-emergencies during the first phase of the COVID-19 pandemic. Methods In this observational, longitudinal cohort study, all neurosurgical admissions that presented to our department between February 1 and April 15 during the COVID-19 pandemic and during the same time period in 2019 were identified and categorized according to the presence of a neuro-emergency, the route of admission, management, and the category of disease. Further, the clinical course of patients with aneurysmal subarachnoid hemorrhage (aSAH) and chronic subdural hematoma (cSDH) was investigated representatively for severe vascular and semi-urgent traumatic conditions that present with a wide variety of symptoms. Results During the pandemic, the percentage of neuro-emergencies among all neurosurgical admissions remained similar but a larger proportion presented through the emergency department than through the outpatient clinic or by referral (*p = 0.009). The total number of neuro-emergencies was significantly reduced (*p = 0.0007) across all types of disease, particularly in vascular (*p = 0.036) but also in spinal (*p = 0.007) and hydrocephalus (*p = 0.048) emergencies. Patients with spinal emergencies presented 48 h later (*p = 0.001) despite comparable symptom severity. For aSAH, the number of cases, aSAH grade, aneurysm localization, and treatment modality did not change but strikingly, elderly patients with cSDH presented less frequently, with more severe symptoms (*p = 0.046), and were less likely to reach favorable outcome (*p = 0.003) at discharge compared with previous years. Conclusions Despite pandemic-related restrictive measures and reallocation of resources, patients with neuro-emergencies should be encouraged to present regardless of the severity of symptoms because deferred presentation may result in adverse outcome. Thus, conservation of critical healthcare resources remains essential in spite of fighting COVID-19.
# Introduction
As the COVID-19 pandemic continues to develop, many countries have taken incisive measures to limit the spread of SARS-CoV-2, such as a restriction of social and economic life. These measures, together with the fear of infection, have changed peoples' routines in drastic fashion in a very short period of time. Against this background, fear of COVID-19 may prevent patients with critical medical or surgical emergencies from actively presenting in emergency departments and outpatient clinics. Further, reallocation measures and triaging of medical services with the purpose to ensure care for a potential surge of COVID-19 patients represent an unprecedented challenge for all specialties that remain committed to medical emergencies that require immediate attention. National and international concerns in this direction have been voiced in the field of cardiology and neurology, where it is feared that patients with acute coronary heart syndrome or cerebral stroke might avoid hospitals and emergency rooms due to social distancing or fear of acquiring COVID-19. Next to such acutely life-threatening emergencies, it seems even more conceivable that patients suffering milder symptoms may currently rather tolerate these than expose themselves to a potential risk of infection that may be associated with hospitalization. Here, neurosurgery offers the opportunity to investigate how the effects of the COVID-19 pandemic may have influenced nonelective emergency care, because neuro-emergencies occur with a wide spectrum of neurological symptoms and severity of disease. In this observational, longitudinal cohort study, we captured the development of neuro-emergencies in one of the largest European university hospitals in the first phase of the COVID-19 pandemic.
# Methods and materials
## Study design and patient population
This longitudinal cohort study was approved by the local ethics committee of the Charité University Hospital in Berlin, Germany (EA1/097/20), and is reported according to the STROBE statement (http://www.strobe-statement.org/). Informed consent was waived due to the retrospective nature of the study. To analyze the effect of the COVID-19 pandemic on neuro-emergency admissions, we identified all patients that were admitted to the Department of Neurosurgery at the Charité University Hospital in Berlin between February 1 and April 15 in the years 2019 and 2020. The time window was chosen based on the identification of the first COVID-19-positive patient in Germany on January 28, 2020, which is considered the beginning of the COVID-19 period when social life and medical operations in Germany were beginning to get affected. Social distancing in Germany was officially recommended by the federal government on March 12, 2020. Data review and analysis was performed by observers (NH, LW, and FW) who were not directly involved in the admission and triage process.
## Definition of neurosurgical emergencies
All patients treated during the pandemic were triaged according to a recently published consensus statement from the German Society of Neurosurgery (DGNC) and the Association of German Neurosurgeons (BDNC) regarding the definition of non-elective cases (https://www.dgnc.de/gesellschaft/aktuelles/ statements). After an individual review of the patients' charts and neuroimaging, neuro-emergencies were assigned to 1 of 6 disease categories. Specifically, these included the following: vascular (aneurysmal subarachnoid hemorrhage (aSAH), malignant cerebral infarction, space-occupying intracerebral hematoma, hemorrhage due to arteriovenous (AV) malformations, higher grade dural AV fistulas, procedures including revascularization in patients with evidence of relevant vascular occlusive disease and unstable aneurysms), cranial oncological (malignant primary brain tumors, brain metastases of any primary tumor type, benign or low-grade tumors with marked parenchymal compression or progressive neurological deficits and pituitary tumors with cranial nerve deficits, visual impairment or endocrine deficiency that cannot be managed conservatively), spinal (intraspinal pathologies with signs of spinal cord compression, degenerative spine conditions with acute onset of motor deficits and/or vegetative dysfunction, progressive myelopathy of cervical and/or thoracic spine, vertebral body fractures with therapyrefractory and severe pain, instability and/or signs of spinal cord compression and spinal metastases or primary tumors with therapy-refractory severe pain, instability and/or compression of the spinal cord), traumatic brain injury (acute traumatic brain injury with subdural hematoma and/or epidural hematoma, any scenario where intracranial pressure cannot be controlled by means of conservative management and chronic subdural hematoma with neurological symptoms), hydrocephalus (progressive increase of intracranial pressure with signs and symptoms suggestive of elevated intracranial pressure or shunt dysfunction), infection, or other emergencies (pain syndromes that do not respond to non-invasive therapeutic modalities, battery depletion in DBS and SCS patients, and benign or malignant peripheral nerve tumors with neurological deficits). The development of emergency admissions was then estimated by calculating the stepwise increase of emergency admissions for each of the six disease categories during the pre-specified time window in 2019 and 2020.
## Neurosurgical measures during the covid-19 pandemic
To maintain a fully operational neurosurgical service, on March 12, 2020, our department implemented pre-specified cohort formation (faculty and residents) with definition of clean cohorts at each campus, home-office, distancing and hygiene measures with outpatient re-organization to videoor telephone appointments, video conference calls for all clinical and educational purposes, business trip ban, and voluntary limitation of social interaction outside of the workplace. For all admitted patients, COVID-19 diagnostics were executed in case of a body temperature > 37.3°C, a quick Sepsis-related Organ Failure Assessment (qSOFA)score ≥ 1, in the presence of respiratory symptoms, and/or in the case of previous contact to another person with confirmed COVID-19. The protocol included three repetitive, combined deep bilateral nasal and deep oropharyngeal swabs for detection of SARS-CoV-2 RNA through PCR analysis, phlegm analysis for SARS-CoV-2 and/or bacterial infection, and chest X-ray.
# Statistical analysis
Descriptive summary statistics are presented as median and range or percentage, as appropriate. Normality was determined with the Shapiro-Wilk test. Statistics were calculated with GraphPad Prism for Mac (Version 8.1.0, GraphPad Software, San Diego, CA, USA). For contingency analysis, a chi-square test was used. For comparison of age in both populations, a two-tailed t test was performed. For comparison of the stepwise increase of neuroemergency admissions for 2019 and 2020, a Wilcoxon matchedpairs signed-rank test was used. For patients with chronic subdural hematoma (cSDH), the stepwise increase of cases between 2014 and 2019 was compared with that of 2020 with a Friedman test for matched pairs and uncorrected Dunn's test for multiple comparison. For comparison of symptom duration, a Mann-Whitney test was performed. All tests were two-tailed and statistical significance was set at p < 0.05.
# Results
## Patient characteristics
Demographic and clinical data are presented in. Between February 1 and April 15, 2020, the number of overall admissions to our department was 46% lower than during the same time period of the previous year. Also, a significantly higher percentage of emergency room admissions was noted (76% versus 61%) at the cost of outpatient admissions and referrals compared with 2019 (*p = 0.009). Twenty-one out of 352 admissions in 2020 received COVID-19 PCR diagnostics following hospitalization. None had evidence of SARS-CoV-2 RNA.
## Presentation and clinical course of neuroemergencies
Although no significant difference was noted in the frequency distribution of neuro-emergency admissions during the pandemic, a significant (40%) reduction in the total number of neuro-emergency admissions was noted across all types of diseases (*p = 0.0007;and particularly for vascular (*p = 0.036), spinal (*p = 0.007), and hydrocephalus (*p = 0.048) emergencies. To get a better idea of the of the pattern of seeking emergency care, we analyzed the frequency distribution of symptoms according to severity for spinal, vascular, hydrocephalus, and trauma emergencies and detected no difference in the distribution of the symptom. For a more detailed description of the presentation and treatment indication of a severe neuro-emergency, we analyzed patients suffering aSAH and found that the number of cases, aSAH severity, aneurysm localization, and treatment modality had not changed during the pandemic. Next, we analyzed patients suffering cSDH representative of a mild-to-moderate neuro-emergency, because cSDH can manifest with the entire range of neurological symptoms. Second, the trauma leading to cSDH most likely occurred before or at the beginning of the socio-economic restrictive measures, so that the incidence of cSDH should remain uninfluenced by pandemic-associated lifestyle changes. Third, cSDH has a high prevalence in the elderlywho are predisposed for an adverse course of COVID-19. Compared with previous years, significantly less patients received acute treatment for cSDH (*p = 0.02 for 2020 versus 2017 and *p < 0.0001 for 2020 versus any other year; and strikingly, these patients significantly more often suffered severe than mild-to-moderate symptoms compared with previous years (*p = 0.046 for 2014-2019 versus 2020;. Further, cSDH patients during the pandemic had a significantly lower likelihood of experiencing favorable outcome at discharge (*p = 0.003 for mRS 0-2 in 2014-2019 versus 2020;.
# Discussion
In this study, we show that despite the restrictive measures associated with COVID-19, patients with neuro-emergencies continue to present across the entire spectrum of disease, which highlights the importance of maintaining a fully operational (neurosurgical) service in spite of reallocation of medical resources, facilities, and staff. For the first time, our findings suggest that patients who suffer a common, semi-urgent neuro-emergency and belong to a population at risk might defer immediate presentation and risk collateral damage.
On February 1, 2020, zero COVID-19 case was registered in Berlin among 16 cases in Germany. On March 12, 2020, social distancing was officially recommended by our federal government. On April 15, the number of COVID-19 infections in Berlin had exponentially increased to 4722 among 127,584 cases in Germany (https://corona.rki.de/). The Department of Neurosurgery at the Charité with its three sites serves a catchment area of approximately 3.7 million people, for whom emergency care may currently be hampered by constraints on facilities, staffing, and protective equipment. Further, COVID-19-related collateral effects may also occur due to deferral of systemically relevant care for medical emergencies, because patients may actively avoid emergency departments due to social-economic lockdown measures and fear of infection. This concern is supported by the fact that high-volume cardiac catheterization centers in the USA and Europe are experiencing a decreased hospitalization and catheterization rate of acute coronary syndrome and myocardial infarction. In our study, we observed a similar development in the category of vascular emergencies. However, the presentation, severity, and treatment of potentially life-threatening aSAH remained unchanged. On the one hand, this reflects that we did not change our indication or perform hidden rationing or ethical triage for treatment of aSAH. On the other hand, the development of aSAH shows that patients with acute aSAH continued to seek emergency medical care regardless of COVID-19, which stands in contrast to previous concerns and needs to be considered when developing strategies for resource reallocation and ethical triage. Under the reasonable assumptions that the incidence of neuro-emergencies did not change during the first phase of the pandemic and that we experienced no access restriction to emergency medical care as other countries, the overall decline of vascular, spinal, and hydrocephalus emergencies rather suggests that less patients are currently seeking emergency medical care but regardless of the type and severity of disease. In spinal emergencies, this is supported by the delayed pattern of seeking care despite similar symptom severity as in 2019. Importantly, the collateral effect of such deferred presentation could also be mirrored by the fewer number of cSDH patients that presented with more severe symptoms and had a lower chance of favorable Aneurysmal subarachnoid hemorrhage (n) 9 8 Hunt and Hess grade (n, %) p = 0.83
[formula] I 3 (33%) 3 (39%) II 2 (23%) 1 (12%) III 1 (11%) 1 (12%) IV - 1 (12%) V 3 (33%) 2 (25%) Aneurysm localization (n, %) p = 0.99 [/formula]
Anterior circulation 7 (78%) 6 (75%) Posterior circulation 2 (22%) 2 (25%) Treatment modality (n, %) p = 0.99 Microsurgical 5 (56%) 4 (50%) Endovascular 4 (44%) 4 (50%) outcome, since cSDH represents a semi-urgent trauma emergency with high prevalence in patients most likely to defer an emergency presentation due to incapacity or fear of infection with SARS-CoV-2.
A natural limitation of our study is the single-center design, although from an epidemiological standpoint, we serve a welldefined catchment area as one of the highest volume neurosurgical centers in Europe. Of course, we cannot exclude that some patients might have presented at other hospitals, but for the majority of cases, this seems unlikely, because cranial neurosurgery in Germany is highly regionally focused. Alternatively, one might argue that we only managed to cope because we had sufficient resources and the pandemic did not hit us as hard compared with other regions in Europe, China, or the USA. However, a nation-wide reallocation of medical resources, facilities, and staff was implemented in the beginning of February 2020 in order to generate sufficient capacities for the expected surge of COVID-19 patients that may require isolation and/or intensive care. In mid-March 2020, we experienced a peak of this development as neurosurgical staff, OR staff, and nurses were reassigned to emergency departments and newly designated COVID-19 units, 25-30% of our regular hospital beds were blocked for COVID-19 patients, and our overall surgical capacity was reduced by 50%, which mirrors an international experience from other neurosurgical departments affected by COVID-19. For us, this resource reallocation translated into a reduction to 66 instead of 96 regular beds and overall, our hospital generated approximately 300 regular and intensive care unit (ICU) beds for COVID-19 patients. By the end of March, we experienced up to 75-80 COVID-19 patients requiring ICU care. The fact that patients even continued to present through our emergency department in an increased percentage and required surgery across the entire spectrum of disease underlines the importance of maintaining a fully functional service in spite of these pandemic-associated reallocation measures. Apart from cohort formation and containment, one of the most essential actions that ensured successful adjustment was an early commitment from our societies (DGNC and BDNC) to uphold care for specific non-elective neurosurgical procedures across the entire spectrum of disease. Also, we were able to maintain 15 out of 16 dedicated neurosurgical ICU beds across all of our sites. Another factor that helped us cope is that > 90% of our faculty and residents are generalists who are able to cover the entire neurosurgical spectrum rather than individual subspecialists, so that the strain on our neurosurgical workforce by cohort measures, sick leave, and the ordered reduction of OR capacity was compensated by pushing the available OR resources to run 24 h a day, 7 days a week. Therefore, despite an extremely restrictive environment regarding OR capacities, staff, and facilities, we were able to maintain care for neuroemergencies that also regularly received a high independent prioritization in the objective triage of OR resources.
In summary, our findings substantiate the following: First, patients with neuro-emergencies appear to be scared of COVID-19. Second, fear of COVID-19 resulted in a 40% reduction of neuro-emergency presentations, regardless of the disease or severity of symptoms. Third, patients that present despite COVID-19 require neurosurgical treatment, which underlines high systemic relevance and suggests that neurosurgery should be prioritized in the reallocation of critical healthcare resources. Ultimately, it will be essential to quantify collateral effects that may have resulted from deferred clinical presentation and we believe that our experience may help limit such effects in future situations requiring resource reallocation and triaging of medical services.
Funding information Open Access funding provided by Projekt DEAL.
## Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee (ethics committee of the Charité University Hospital in Berlin, Germany; EA1/097/20) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent For this type of study, formal consent is not required.
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/. |
Tailored flavoproteins acting as light-driven spin machines pump nuclear hyperpolarization
The solid-state photo-chemically induced dynamic nuclear polarization (photo-CIDNP) effect generates non-Boltzmann nuclear spin magnetization, referred to as hyperpolarization, allowing for high gain of sensitivity in nuclear magnetic resonance (NMR). Well known to occur in photosynthetic reaction centers, the effect was also observed in a light-oxygen-voltage (LOV) domain of the bluelight receptor phototropin, in which the functional cysteine was removed to prevent photo-chemical reactions with the cofactor, a flavin mononucleotide (FMN). Upon illumination, the FMN abstracts an electron from a tryptophan to form a transient spin-correlated radical pair (SCRP) generating the photo-CIDNP effect. Here, we report on designed molecular spin-machines producing nuclear hyperpolarization upon illumination: a LOV domain of aureochrome1a from Phaeodactylum tricornutum, and a LOV domain named 4511 from Methylobacterium radiotolerans (Mr4511) which lacks an otherwise conserved tryptophan in its wild-type form. Insertion of the tryptophan at canonical and novel positions in Mr4511 yields photo-CIDNP effects observed by 15 N and 1 H liquidstate high-resolution NMR with a characteristic magnetic-field dependence indicating an involvement of anisotropic magnetic interactions and a slow-motion regime in the transient paramagnetic state. The heuristic biomimetic design opens new categories of experiments to analyze and apply the photo-CIDNP effect.The solid-state photo-CIDNP (photo-chemically induced dynamic nuclear polarization) effect 1-3 allows to enhance nuclear magnetic resonance (NMR) signals by a build-up of transient nuclear spin-hyperpolarization caused by light-induced transient spin-correlated radical pairs (SCRPs) in electron-transfer proteins 4,5 . The effect was discovered by Zysmilich and McDermott in 1994 by performing 15 N magic-angle spinning (MAS) NMR studies on an isolated, frozen and quinone-blocked photosynthetic reaction center protein of the purple bacterium Rhodobacter sphaeroides 6 . Since its discovery, the effect has been observed in plenty of other photosynthetic reaction centers (RCs) of plants 7-10 , algae 11,12 , diatoms 13,14 , purple bacteria 15-18 , heliobacteria 19,20 , and green sulfur bacteria 21 . The spin-chemical machinery pumping nuclear polarization has been probed by field-dependent 22,23 , time-resolved 24-26 , and preparation-dependent 27,28 experiments, and is interpreted by the occurrence of up to three mechanisms running in parallel 22,24,29 : In the Differential Relaxation (DR) mechanism, also called "cyclic reactions mechanism", the symmetry between both decay branches is broken due to a difference in relaxation caused by the paramagnetic molecular triplet state 30-32 . In the electron-electron nuclear Three-Spin Mixing (TSM) mechanism, symmetry breaking is driven by the pseudosecular part of the hyperfine interaction during the evolution of the SCRP 33,34 . The differences in decay kinetics between the SCRP in its singlet and its triplet state cause the Differential Decay (DD) mechanism 35 . Recently, the coherent electron-electron-nuclear spindynamics has been described in terms of a unified theoretical approach, based on level crossings (LCs) and level anti-crossings (LACs) 36-38 , which allowed the establishment of CIDNP sign rules and to determine conditions optimal for polarization formation.Besides photosynthetic RCs, there is another class of proteins acting as molecular spin-machines pumping non-Boltzmann nuclear spin-state upon illumination: biological photoreceptors that contain a flavin cofactor as chromophore with characteristic absorption maxima in the UV-A and blue light regions. Merely one family of flavoproteins, phototropins, was reported to show the photo-CIDNP effect[39][40][41]. Generally, phototropins are blue-light receptors that harbor two light-sensing light-oxygen-voltage (LOV) domains, each of which incorporates one flavin mononucleotide (FMN) cofactor non-covalently. Upon excitation, the FMN in a wild-type LOV domain of phototropin undergoes adduct formation with a close-by conserved cysteine 42,43 . However, so far, no photo-CIDNP effect was observed from a wild-type LOV domain. The first successful solid-state photo-CIDNP effect observation was demonstrated for a frozen cysteine-lacking LOV domain of phototropin from the green alga Chlamydomonas reinhardtii (Crphot-LOV1-C57S)41,44. In this case, the photo-excited FMN undergoes inter-system crossing to the triplet state, causes intra-protein electron transfer from a tryptophan to the FMN and thus gives rise to formation of a transient SCRP. Spin evolution of this SCRP allows for the build-up of the solid-state photo-CIDNP selectively on FMN and tryptophan. In Crphot-LOV1-C57S, the edge-to-edge distance ( r FW ) between FMN (F) and tryptophan (W) is ~ 11 Å. Remarkably, Weber et al. first observed photo-CIDNP in a similar LOV protein, the LOV2 domain of phototropin 1 from oat, Avena sativa (Asphot1-LOV2-C450A), by 13 C liquid-state NMR 40,45,46 . The crystal structure of Asphot1-LOV2-C450A is almost identical to that of Crphot-LOV1-C57S, given that they share 47% of amino acid sequences. The r FW value is very similar for the two proteins. This is a very rare example that a nuclear-spin hyperpolarization effect generated in one system could be observed by both liquid-state and solid-state NMR. To explore the action of anisotropic mechanisms in the liquid sample that are expected to be sensitive to the magnetic field strength 44,46 , a solution-state NMR spectrometer equipped with a field-cycling device 47 was employed to study 1 H, 13 C and 15 N photo-CIDNP generated in Crphot-LOV1-C57S within a broad field range from 0.01-9.4 T 39 .The LOV domain is the blue-light mediating motif of not only phototropin but also of some other proteins, e.g., aureochromes 48 . Aureochromes represent a class of LOV proteins with an unusual inversed domain arrangement, as they carry, in contrast to the majority of LOV-domain proteins, the functional/signaling domain N-terminally to the LOV domain and not C-terminally; they exhibit great potential in controlling DNA binding 49 as a natural optogenetic module. Despite a different domain organization of the full-length protein, the LOV domain of aureochromes seem to show a high degree of conservation to canonical LOV domains with respect to sequence and secondary structure 50 . Like in other LOV domains, the optical excitation of FMN chromophore leads to formation of an adduct with a conserved cysteine residue. Furthermore, a recently discovered flavoprotein from the radiation-resistant bacterium Methylobacterium radiotolerans, Mr4511, unusually lacks the single tryptophan conserved in 75% of LOV domains 51 . Due to the absence of tryptophan to quench the triplet state of FMN, 3 FMN, after excitation by light, the cysteine-lacking mutant of Mr4511 can serve as an efficient singletoxygen generator 51 . The variety of LOV domains raises the question whether LOV domains can be generally constructed in such a way that they show a photo-CIDNP effect.The aim of this work was to extend the range of biological systems amenable to the solid-state photo-CIDNP effect. To reach this goal, we at first produced a cysteine-lacking LOV domain of aureochrome1a from the diatom Phaeodactylum tricornutum (Ptaureo1a-LOV-C287S) 52,53 and show that it generated 1 H-, 13 C-, and 15 N photo-CIDNP effects in aqueous solution. By using a field-cycling system, the magnetic field-dependencies of the 1 H, 13 C, and 15 N hyperpolarization effects have been obtained revealing that the magnetic field for maximal photo-CIDNP depends on the nuclear gyromagnetic ratio. Such a behavior is a characteristic feature of the solid-state photo-CIDNP effect 36,39 , closely resembling that of Crphot-LOV1-C57S. Based on this result, we propose that, like in phototropin, the anisotropic magnetic interactions might play a decisive role in photo-CIDNP formation in the LOV domain of aureochrome in solution.Furthermore, in the heuristic approach, biomimetic protein design is used to control the conditions for the occurrence of the photo-CIDNP effect. We employed Mr4511 for an extended protein mutation strategy allowing us to change parameters such as the distance between the partners governing the electron transfer reactions that give rise to SCRP formation and recombination, and to tune magnetic parameters of the SCRP. By doing so, we were able to probe the spin dynamics in the SCRP by magnetic-field dependent photo-CIDNP studies. We could also elucidate the role of different photo-CIDNP mechanisms that are responsible for the formation of nuclear spin-polarization. Last but not least, we discovered a non-tryptophan-induced photo-CIDNP effect generated by the cysteine-devoid Mr4511, in which tryptophan is absent.
# Results
To rationalize the key properties of molecular spin-machines that can be used to generate photo-CIDNP, we proposed a design strategy based on mutations, supported by field-dependent CIDNP studies. By using various mutations, as described below, we were able to vary the distance between the electron donor and acceptor. In this way, we affected the rate of SCRP formation and recombination, and also varied the electron-electron spin-spin interaction in the SCRP. To probe the reaction and spin dynamics in the SCRP, we used the field dependence of photo-CIDNP.
Screening LOV domains for induction of photo-CIDNP. Aiming for designed molecular spinmachines producing light-induced nuclear hyperpolarization, we have designed a series of protein mutants, which will be presented in parts (i)-(iii) .
(i) So far, the occurrence of the solid-state photo-CIDNP effect was limited to cysteine-lacking LOV domains of phototropin [bib_ref] Nuclear spin-hyperpolarization generated in a flavoprotein under illumination: experimental field-dependence and theoretical..., Ding [/bib_ref] [bib_ref] Photochemically induced dynamic nuclear polarization in a C450A mutant of the LOV2..., Richter [/bib_ref] [bib_ref] Solid-state photo-CIDNP effect observed in phototropin LOV1-C57S by magic-angle spinning nmr spectroscopy, Thamarath [/bib_ref] ; for this reason, here we explored other potential LOV-based light-induced hyperpolarization generators. Alignment and comparison of the amino-acid sequences of Asphot1-LOV2, Scientific Reports | (2020) 10:18658 | https://doi.org/10.1038/s41598-020-75627-z www.nature.com/scientificreports/ Crphot-LOV1 and Ptaureo1a-LOV show about 50% of identity and the crystal structures show almost identical tertiary structures . In particular, the distance between FMN and tryptophan , which determines the strength of the spin-spin coupling in the SCRP and is therefore central to generate photo-CIDNP, is nearly the same, being approximately 11 Å. Therefore, we used the mutant Ptaureo1a-LOV-C287S ( r FW ~ 11 Å) for the liquid state photo-CIDNP NMR experiment, in which the conserved cysteine 287 is replaced by serine. (ii) Formation of an SCRP by electron transfer to excited FMN can occur if a nearby tryptophan is present to act as the electron donor. It has been shown that the amino acid tryptophan is able to provide this function [bib_ref] Solid-state photo-CIDNP effect observed in phototropin LOV1-C57S by magic-angle spinning nmr spectroscopy, Thamarath [/bib_ref] [bib_ref] Natural abundance solution 13 C NMR studies of a phototropin with photoinduced..., Eisenreich [/bib_ref]. Therefore, in the LOV protein Mr4511, lacking the conserved tryptophan, we introduced a tryptophan at its canonical position Q112 by mutation resulting in the Mr4511-C71S-Q112W double mutant. Previously, transient absorption experiments have been used to test the function of these mutants. In Mr4511, when the cysteine residue was mutated to serine (C71S) or glycine (C71Q) and no tryptophan was present, the lifetime of 3 FMN, τ T , was around 240 μs [bib_ref] Single mutation in a novel bacterial LOV protein yields a singlet oxygen..., Consiglieri [/bib_ref]. Introduction of tryptophan to the canonical position, Mr4511-C71S-Q112W, gave rise to faster quenching of 3 FMN reducing τ T to ~ 24 μs 51 , a value very close to ~ 27 μs observed in Crphot-LOV1-C57S ( r FW ~ 11 Å) [bib_ref] Phot-LOV1: photocycle of a blue-light receptor domain from the green alga Chlamydomonasreinhardtii, Kottke [/bib_ref]. Hence, the double mutant Mr4511-C71S-Q112W ( r FW ~ 11 Å) is the second candidate for the generation of the solid-state photo-CIDNP effect. (iii) Finally, we introduced the electron-donating tryptophan at non-canonical positions. The introduction of tryptophan to a new position of the protein allows to change the distance between FMN and tryptophan, their relative orientation and chemical environments and, therefore, to affect the kinetic and magnetic parameters, critical for the formation of the solid-state photo-CIDNP effect. It is difficult to fine-tune all relevant reaction and magnetic resonance parameters simultaneously, therefore, we focused on creating mutants with different r FW . Lacking a crystal structure of Mr4511, the design relied on a structural model created using SWISS-MODEL and the crystal structure of aureochrome1a-LOV (PDB: 3UE6) from a eukaryotic photosynthetic stramenopile as template [bib_ref] Crystal structures of aureochrome1 LOV suggest new design strategies for optogenetics, Mitra [/bib_ref]. The report of the modeling parameters is provided in . Additionally, comparison of the amino-acid sequence allowed us to predict the occurrence of α-helix and β-sheet secondary structures and to reconstruct the tertiary structure of Mr4511. As targets for mutation, we considered amino acids that do not interact with the FMN 50 and also have a bulky side chain similar as tryptophan. Using these ideas, we have designed the following mutants with different positions of tryptophan with various r FW values: .
[formula] Mr4511-C71S- F130W (~ 6 Å), Mr4511-C71S-Y116W (~ 9 Å), Mr4511-C71S-Y129W (~ 11 Å) and Mr4511-C71S-K57W (~ 17 Å) ( [/formula]
Another aspect relevant for rational design of a biomimetic light-driven spin-machine for production of photo-CIDNP is the possibility to introduce isotope labels. In particular, for the measurement of the 15 N photo-CIDNP we employed 15 NH 4 Cl as the sole nitrogen source in the bacterial growth medium during protein expression and produced uniformly 15 N-labelled protein and cofactor (see "Methods"). For the 13 C photo-CIDNP experiment, different labelling strategies were previously applied, either by incorporating the 13 C-labelled FMN into a natural abundant protein moiety of the phototropin-LOV domain [bib_ref] Photochemically induced dynamic nuclear polarization in a C450A mutant of the LOV2..., Richter [/bib_ref] [bib_ref] Detecting a new source for photochemically induced dynamic nuclear polarization in the..., Kothe [/bib_ref] or by selective 13 C-labelling of the single tryptophan of the phototropin-LOV domain [bib_ref] Nuclear spin-hyperpolarization generated in a flavoprotein under illumination: experimental field-dependence and theoretical..., Ding [/bib_ref]. This enables unambiguous assignment of hyperpolarized [bib_ref] Photo-CIDNP in the reaction center of the diatom Cyclotella meneghiniana observed by..., Zill [/bib_ref] C signals and analysis of the photo-CIDNP effect generated by electron donor and acceptor separately. A complete picture of the effect, however, involving both electron donor and acceptor is still missing. Therefore, in the present work we produced a uniformly 13 C-labelled Crphot-LOV1-C57S ( r FW ~ 11 Å), aiming to compare the photo-CIDNP effect of FMN and tryptophan under the same conditions. The hyperpolarization effect in combination with isotope labelling paves the way to field-dependent NMR measurements, providing knowledge . LOV proteins employed for photo-CIDNP NMR in this work. a The FMN-Trp distance ( r FW ) is given as edge-to-edge distance and estimated based on a structural model obtained by SWISS-MODEL 55 . b The percentage identity is compared with the amino-acid sequence of Crphot-LOV1-C57S. www.nature.com/scientificreports/ of the relationship between enhancement factor and magnetic field, which might provide the key data for future theoretical analysis of the exact photo-CIDNP mechanism.
Comparison of the photo-CIDNP effect between phototropin and aureochrome. [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref] shows the 1 H photo-CIDNP effect and its field dependence observed in Crphot-LOV1-C57S ( r FW ~ 11 Å). The protein produced for the experiment initially contained all the nuclei in their natural abundance. Then the protonated buffer of the sample was exchanged to a deuterated buffer (see "Methods" section). The final protein solution may contain ~ 0.4% residual 1 H. From this sample, the effect has not been observed directly on the protons of FMN and tryptophan, however, the light-minus-dark spectra show a negative enhancement (emissive signal, i.e., opposite to the thermal polarization) of the HDO signal at 4.7 ppm, which agrees with the previous publication [bib_ref] Nuclear spin-hyperpolarization generated in a flavoprotein under illumination: experimental field-dependence and theoretical..., Ding [/bib_ref]. A closer look at the light-minus-dark spectra shows that the entire range (-2 to 10 ppm) in the proton NMR spectra exhibits hyperpolarization, in particular the aliphatic region (0 to 2.5 ppm) as well as the HDO signal. Integrating either the HDO signal or the range of 0 to 2.5 ppm or the range of -2 to 10 ppm, we obtain a field dependence with a maximum at 0.6 T, as shown in [fig_ref] Figure 3: Intensity of selected hyperpolarized signals of Crphot-LOV1-C57S [/fig_ref]. Since the position of the maximum is the same for all protons, it is plausible that the 1 H hyperpolarization has originated from the same SCRP and has been distributed over the whole protein as well as to the residual protons in the deuterated solvent. We assume that the spread of hyperpolarization under liquid-state conditions is due to cross relaxation or due to spin diffusion mediated by non-averaged proton-proton dipolar couplings 59,60 . 13 C photo-CIDNP of a uniformly 13 C-labelled Crphot-LOV1-C57S ( r FW ~ 11 Å) induced at various field strengths ranging from 0.5 to 9.4 T and detected always at 9.4 T is shown in [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref]. The hyperpolarized 13 C signals are tentatively assigned according to previous studies [bib_ref] Nuclear spin-hyperpolarization generated in a flavoprotein under illumination: experimental field-dependence and theoretical..., Ding [/bib_ref] [bib_ref] Detecting a new source for photochemically induced dynamic nuclear polarization in the..., Kothe [/bib_ref] [bib_ref] Strategy for Enhancement of 13 C-photo-CIDNP NMR spectra by exploiting fractional 13..., Eisenreich [/bib_ref] [bib_ref] 1 H and 13 C hyperfine coupling constants of the tryptophanyl cation..., Kiryutin [/bib_ref] , and are summarized in [fig_ref] Table 2: Photo-CIDNP experimental results and obtained 15 N chemical shifts from the eight... [/fig_ref] in Supplementary Information. [fig_ref] Figure 3: Intensity of selected hyperpolarized signals of Crphot-LOV1-C57S [/fig_ref] presents the integrated areas of some selected hyperpolarized carbon signals of FMN and tryptophan against the magnetic field strength. Overall, the carbons on the tryptophan indole ring seem to be stronger polarized than those on the isoalloxazine ring of FMN. The hyperpolarized carbon nuclei of tryptophan show emissive polarization at all fields and maximal polarization at around 3 T in agreement with previous results [bib_ref] Nuclear spin-hyperpolarization generated in a flavoprotein under illumination: experimental field-dependence and theoretical..., Ding [/bib_ref]. The selected 13 C atoms of the isoalloxazine ring of FMN show different enhancement patterns. As the magnetic field increases, the signal of, e.g., FMN C-8 stays negative, while FMN C-2 changes the sign of polarization from emission to enhanced absorption. Due to the difference of labelling strategies, the comparison of the 13 C photo-CIDNP spectrum of Crphot-LOV1-C57S ( r FW ~ 11 Å) with a previously published 13 C photo-CIDNP spectrum of Asphot1-LOV2-C450A ( r FW ~ 11 Å) [bib_ref] Photochemically induced dynamic nuclear polarization in a C450A mutant of the LOV2..., Richter [/bib_ref] [bib_ref] Detecting a new source for photochemically induced dynamic nuclear polarization in the..., Kothe [/bib_ref] is not straightforward. Despite that, we can compare the signal of Trp C-3 carbon in both spectra. In Asphot1-LOV2-C450A ( r FW ~ 11 Å) where the tryptophan nucleus is at natural abundance, strong hyperpolarization at Trp C-3 occurs. The chemical shift of C-3 of Crphot-LOV1-C57S (109.0 ppm) slightly differs from δ (C-3) = 113.5 ppm for Asphot1-LOV2-C450A which might result from the different protein environment of the tryptophan residues. In both cases, the sign of the signal of Trp C-3 is always emissive at all fields studied. Nevertheless, the field at which the maximum polarization at Trp C-3 occurs, B max = 3 T for Crphot-LOV1-C57S and B max = 7 T for Asphot1-LOV2-C450A, is well distinguished. We assume that the difference in the label positions causes the difference in field dependence although details are not yet understood. Furthermore, in a previous work by Kothe et al. [bib_ref] Detecting a new source for photochemically induced dynamic nuclear polarization in the..., Kothe [/bib_ref] the photo-CIDNP effect has been measured at four different magnetic fields (5.9, 7.1, 9.4, and 11.8 T), limiting comparability of data points and localization of the maximum B max , while in the present work, we overcame this problem by using a shuttle system (10 nT < B o < 9.4 T).
Figures 2C and 3C depict the field dependence of the 15 N photo-CIDNP obtained in uniformly 15 N-labelled Crphot-LOV1-C57S ( r FW ~ 11 Å) at 11 magnetic field strengths ranging from 0.1 to 9.4 T. In [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref] , the three hyperpolarized signals and their field dependencies are shown (for assignment, see [fig_ref] Table 2: Photo-CIDNP experimental results and obtained 15 N chemical shifts from the eight... [/fig_ref]. The Trp N-1 CIDNP is a low absorptive polarization at low fields, turning into strong emissive polarization at about 1 T field and having a maximum at around 5 T. In contrast, the photo-CIDNP of FMN N-5 is always absorptive with a maximum at 3 T. At increasing field, the hyperpolarization of FMN N-10 turns from enhanced absorption to emission at around 3 T, similar to the photo-CIDNP of Trp N-1 and in clear contrast to the photo-CIDNP of FMN N-5. A common feature for these three nuclei is that the high-field maximum occurs at around 4 T.
For comparison of the photo-CIDNP field dependencies for the three different NMR active nuclear isotopes, we have chosen the most enhanced signal or spectral region for each nucleus, 1 H, [bib_ref] Photo-CIDNP in the reaction center of the diatom Cyclotella meneghiniana observed by..., Zill [/bib_ref] [bib_ref] Structure of a native-like Aureochrome 1a LOV domain dimer from Phaeodactylum tricornutum, Banerjee [/bib_ref] , and the simulated structure of Mr4511 (yellow) without FMN. The simulation is performed with SWISS-MODEL based on the crystal structure of aureochrome1a-LOV from Vaucheria frigida (PDB entry: 3UE6) [bib_ref] Crystal structures of aureochrome1 LOV suggest new design strategies for optogenetics, Mitra [/bib_ref]. The information about structural modeling is listed in Supplementary Information . Five mutants of cysteine-lacking Mr4511 were generated, one with tryptophan placed at the canonical position, Q112W, r FW ~ 11 Å, the other four at non-canonical positions F130W, Y116W, Y129W and K57W with increasing r FW . [fig_ref] Table 2: Photo-CIDNP experimental results and obtained 15 N chemical shifts from the eight... [/fig_ref] and 15 N chemical shifts in [fig_ref] Table 2: Photo-CIDNP experimental results and obtained 15 N chemical shifts from the eight... [/fig_ref]. The lower trace shows stacked spectra of the CIDNP effects of the corresponding nuclei measured at five different magnetic fields. [fig_ref] Figure 4: Comparison of the field dependencies of the 1 H and 13 C... [/fig_ref] ; the corresponding 1D spectra are shown in . Hence, a high similarity in the field dependence is observed for the two proteins. All three types of nuclei in both samples exhibit emissive NMR signals at the maxima; the positions of maxima depend on the nuclear gyromagnetic ratios (the maximum is positioned at a lower field for a nucleus with a higher gyromagnetic ratio). In both cases, the 1 H hyperpolarization reaches its emissive maximum at a field around 0.6 T, while the maxima of 13 C and 15 N photo-CIDNP appear to be at higher fields. Although the primary sequence of the two proteins does not have a high identity, their tertiary structures and, as shown here, their functional mechanisms are almost identical. Therefore, the production of the photo-CIDNP does not depend crucially on the primary sequence of individual amino acids, whereas the distance between the two redox partners forming the SCRP and their mutual orientation plays of a decisive role.
Tryptophan and non-tryptophan derived photo-CIDNP effect generated in Mr4511 mutants. To generate a photo-CIDNP effect in Mr4511, the conserved cysteine was first replaced with the inactive serine resulting in the mutant Mr4511-C71S. Furthermore, tryptophan was introduced to different locations of the Mr4511-C71S protein, generating five additional mutants . The magnetic field dependence of the photo-CIDNP effect under liquid-state conditions obtained by [bib_ref] Photo-CIDNP 13 C magic angle spinning NMR on bacterial reaction centres: exploring..., Matysik [/bib_ref] N and 1 H NMR is summarized in [fig_ref] Figure 5: MagneticMr4511-C71S-Y116W [/fig_ref] , plotted in the same way as in [fig_ref] Figure 3: Intensity of selected hyperpolarized signals of Crphot-LOV1-C57S [/fig_ref]. The corresponding 15 N NMR spectra are shown in [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref] and the 15 N chemical shifts are provided in [fig_ref] Table 2: Photo-CIDNP experimental results and obtained 15 N chemical shifts from the eight... [/fig_ref]. [fig_ref] Figure 3: Intensity of selected hyperpolarized signals of Crphot-LOV1-C57S [/fig_ref] and S4). Time-resolved optical absorption analysis on F130W suggests that the triplet state 3 FMN is not formed to an observable extent in this protein, most likely, because of the ultrafast electron transfer between FMN and tryptophan (Aba Losi, personal communication). Accordingly, for the construction of light-induced artificial spin-machines pumping nuclear hyperpolarization, the information of a minimum distance of the redox partners of the SCRP is highly relevant.
Except for Mr4511-C71S-F130W ( r FW ~ 6 Å), all other Mr4511-C71S mutants generated both, the 15 N and 1 H photo-CIDNP effect under liquid-state conditions. As shown in [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref] , Mr4511-C71S-Y116W ( r FW ~ 9 Å) shows hyperpolarization for the nitrogen on FMN and tryptophan, therefore it is referred to as "tryptophanderived photo-CIDNP". This protein shows a maximal 1 H hyperpolarization at 2.4 T [fig_ref] Figure 5: MagneticMr4511-C71S-Y116W [/fig_ref] , which is a higher field than observed for any other mutant (see below).
The r FW value of ~ 11 Å in the phototropin and aureochrome LOV domains was known to generate a photo-CIDNP effect [bib_ref] Photochemically induced dynamic nuclear polarization in a C450A mutant of the LOV2..., Richter [/bib_ref] [bib_ref] Solid-state photo-CIDNP effect observed in phototropin LOV1-C57S by magic-angle spinning nmr spectroscopy, Thamarath [/bib_ref]. Here we compare the two cases when the tryptophan is at the canonical position, Mr4511-C71S-Q112W ( r FW ~ 11 Å) [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref] and the non-canonical position but with nearly the same r FW distance, Mr4511-C71S-Y129W ( r FW ~ 11 Å) [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref]. Although Mr4511-C71S-Q112W and Mr4511-C71S-Y129W exhibit a similar field maximum for the 1 H photo-CIDNP effect, their field-dependent 15 N photo-CIDNP effects are significantly different. This means that the efficiency of photo-CIDNP formation does not only depend on the spatial distance. Different orientations and different local mobility of the residues might be considered as the origin of this difference. However, the phase of hyperpolarized Trp N-1 signal is allways negative for mutants Y116W, Y129W and Q112W.
The tryptophan residue in Mr4511-C71S-K57W ( r FW ~ 17 Å) is the most remote electron donor from FMN in the studied set of mutants and, thus, the reaction rate constant of electron transfer in this protein mutant is expected to be the smallest in this series [bib_ref] Nature of biological electron transfer, Moser [/bib_ref]. The protein shows a 15 N hyperpolarization solely on the FMN N-5, while the FMN N-10 and Trp N-1 signals do not exhibit any enhancement in a wide range of magnetic fields [fig_ref] Figure 5: MagneticMr4511-C71S-Y116W [/fig_ref]. The same photo-CIDNP experiments were also performed on Mr4511-C71S in which no tryptophan was present leading to very similar results [fig_ref] Figure 5: MagneticMr4511-C71S-Y116W [/fig_ref] with somewhat weaker single emissive FMN N5 signal (Figure S2 graphs D and E). This surprising result clearly indicates that the 15 N photo-CIDNP effect reported here is not derived solely from the involvement of the added tryptophan residue. It is well-known that tyrosine can also act as an electron donor in biological systems [bib_ref] Tyrosine radicals are involved in the photosynthetic oxygen-evolving system, Barry [/bib_ref] [bib_ref] Formation of interacting spins on flavosemiquinone and tyrosine radical in photoreaction of..., Nagai [/bib_ref]. According to the amino-acid sequence and our structural model of Mr4511, there are four tyrosine residues located in proximity to FMN in the range of 9 to 12 Å. In line with this speculation is the fact that the 15 N photo-CIDNP effect is not observable for tyrosine, since tyrosine does not contain 15 N in the side chain. Therefore, the light-driven molecular spin-machines can probably also rely on SCRPs containing a tyrosine radical.
# Discussion
Here, we show that a photo-CIDNP effect originating from the SCRP of FMN and tryptophan can be produced in artificially designed flavoproteins. We employ a systematic mutation strategy to vary reaction and magnetic parameters of the paramagnetic centers generated by light. It appears that the r FW distance of ~ 6 Å between the FMN and tryptophan is too short to provide conditions suitable for photo-CIDNP formation, whereas in the range of ~ 9 Å to ~ 11 Å, the effect has been observed [fig_ref] Figure 5: MagneticMr4511-C71S-Y116W [/fig_ref] ; [fig_ref] Table 2: Photo-CIDNP experimental results and obtained 15 N chemical shifts from the eight... [/fig_ref]. The data on the field dependence of the photo-CIDNP effect generated by the designed LOV domains show complex dependencies, which are not expected for the liquid-state photo-CIDNP effect. In particular, the differences in field dependence obtained for LOV domains having the same distance of donor and acceptor suggest that anisotropic spin interactions come into play as they are expected for solids. In addition to the field dependence, a distance dependence has been documented. Apparently, further parameters are involved, presumably the relative orientation of donor and acceptor as well as their local dynamics. Both, anisotropy and relaxation effects require further studies. Furthermore, the effect of different label patterns requires a future study.
The presence of solid-state mechanisms in LOV domains in liquids implies that the transient SCRP occurs in a slow-motion regime, during which the anisotropic electron-nuclear interactions are conserved for the build-up of hyperpolarization. In contrast, on NMR time scale, all the anisotropic nuclear interactions, i.e., nuclear dipolar coupling and chemical shift anisotropy as present in solids are averaged out and thus the hyperpolarized signals in the NMR spectra exhibit no obvious anisotropic features. Such phenomenon was previously observed for a Regarding the 1 H photo-CIDNP, the region of chemical shifts ranging from − 2 to 10 ppm of the spectrum is integrated and plotted against the magnetic field (A1-F1). The three 15 N photo-CIDNP hyperpolarized signals correspond to FMN N-5 (green), FMN N-10 (yellow) and Trp N-1 (blue); the signals are plotted against the magnetic field strength (A2-F2) in the same range and scale as for the 1 H signal. For straightforward comparison, the signal strength at the maximal hyperpolarization magnitude of each graph is normalized to unity. The positive and negative signs of the y-axis indicate absorptive and emissive hyperpolarization, respectively. Since the mutant F130W shows no detectable photo-CIDNP effect, as presented in the [fig_ref] Figure 3: Intensity of selected hyperpolarized signals of Crphot-LOV1-C57S [/fig_ref] , the data points shown in (A1) and (A2) were measured the over the same spectra regions with the same noise level as in the corresponding spectra for Y129W shown in (D1) and (D2).
## Scientific reports
| (2020) 10:18658 | https://doi.org/10.1038/s41598-020-75627-z www.nature.com/scientificreports/ photosynthetic RC protein-membrane complex corresponding to ~ 1 MDa, measured by 13 C liquid-state NMR [bib_ref] A 10 000-fold nuclear hyperpolarization of a membrane protein in the liquid..., Daviso [/bib_ref]. For LOV proteins, having the molecular weight of less than 20 kDa, the occurrence of anisotropic mechanisms in liquids likely relies on the formation of dimers [bib_ref] Blue-light-induced unfolding of the Jα helix allows for the dimerization of aureochrome-LOV..., Herman [/bib_ref] [bib_ref] Structure of a native-like Aureochrome 1a LOV domain dimer from Phaeodactylum tricornutum, Banerjee [/bib_ref] [bib_ref] Photoreaction dynamics of LOV1 and LOV2 of phototropin from Chlamydomonas reinhardtii, Nakasone [/bib_ref] or higher multimers in solution. The slow tumbling rate may lead to the presence of residual proton-proton couplings which allow for the 1 H hyperpolarization transfer from the center of the photo-reaction into the environment [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref]. So far, flavoproteins and photosynthetic RCs are the only reported electron transfer protein systems that show solid-state photo-CIDNP effect. Even despite the different cofactor arrangements and spin-dynamics, they might share the same mechanisms. Consequently, similar features of CIDNP are expected regarding the sign change of nuclear spin hyperpolarization and the similar field at which the maximum polarization occurs [bib_ref] Electron spin density distribution in the special pair triplet of Rhodobacter sphaeroides..., Thamarath [/bib_ref] [bib_ref] Field-cycling NMR with high-resolution detection under magic-angle spinning: determination of field-window for..., Gräsing [/bib_ref]. LCs and LACs analysis suggested that a solid-state photo-CIDNP effect is not only field-dependent, but also strongly orientation-dependent because of the anisotropic interactions governing in spin dynamics of the SCRP in solids [bib_ref] Level crossing analysis of chemically induced dynamic nuclear polarization: towards a common..., Sosnovsky [/bib_ref] [bib_ref] Magnetic field and orientation dependence of solid-state CIDNP, Sosnovsky [/bib_ref]. To the present experiments conducted under liquid-state conditions, the same theory will be applied to understand the sign change that occurred in the LOV proteins as shown in [fig_ref] Figure 3: Intensity of selected hyperpolarized signals of Crphot-LOV1-C57S [/fig_ref] ,C as well as [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref].
Summarizing these considerations, we can propose the following interpretation of the experimental findings. By increasing r FW , we decrease two parameters: The SCRP recombination rate and the electron-electron spin-spin coupling, J SCRP , within the SCRP. When the r FW distance is too short, photo-CIDNP formation is suppressed, most probably, due to two reasons: The first reason is that J SCRP is too large, introducing an energy gap between the singlet and triplet SCRP spin states. This gap cannot be overcome by the relatively small hyperfine couplings, and singlet-triplet interconversion in the SCRP is thus suppressed. The second reason is that the spin-evolution of the SCRP requires sufficient time for photo-CIDNP formation: fast SCRP recombination interrupts this process and thus no photo-CIDNP is formed.
As r FW increases further, we enter the regime in which the SCRP lifetime is sufficiently long and J SCRP is sizeable, but not too large to suppress singlet-triplet mixing, giving rise to photo-CIDNP formation. In this situation, the TSM scenario comes into play and the maximum position, B max , in the photo-CIDNP field dependence is given 37,39 by the matching condition |γ N |B max ≈ J SCRP , with γ N being the nuclear gyromagnetic ratio; the sign of polarization of the three different kinds of nuclei ( 1 H, 13 C and 15 N) is also consistent with previous theoretical considerations [bib_ref] Magnetic field and orientation dependence of solid-state CIDNP, Sosnovsky [/bib_ref] [bib_ref] Nuclear spin-hyperpolarization generated in a flavoprotein under illumination: experimental field-dependence and theoretical..., Ding [/bib_ref]. Hence, the B max field is different for different nuclei, which is consistent with the experimental data.
When r FW increases further, J SCRP is decreased and other photo-CIDNP mechanisms 37 come into play. In this situation, polarization formation is no longer sensitive to the γ N value, i.e., to the nuclear spin isotopes, so that different kinds of nuclei exhibit a similar behavior. The sign changes of photo-CIDNP can be rationalized in terms of changing dominance of enhancement mechanisms, as it happens in RCs [bib_ref] Magnetic field and orientation dependence of solid-state CIDNP, Sosnovsky [/bib_ref].
The design strategy also leads to the discovery of a new-type of photo-CIDNP effect generated by Mr4511-C71S in which no tryptophan is present. The same effect [fig_ref] Figure 5: MagneticMr4511-C71S-Y116W [/fig_ref] also occurs in Mr4511-C71S-K57W ( r FW ~ 17 Å). Based on the present results, we are unable to unravel the origin of the new-type photo-CIDNP effect. Recent research on a designed cysteine-lacking Asphot1-LOV2 domain indicated that, without presence of the tryptophan, the FMN was reduced, however to less extend compared to the case when the tryptophan was present. Kinetic data suggested that one of the tyrosines in the LOV domain acts as counter radical [bib_ref] Electron transfer pathways in a light, oxygen, voltage (LOV) protein devoid of..., Kopka [/bib_ref]. Therefore, we proposed that tyrosine might act as electron donor in the absence of tryptophan also in our case.
With this, the present work significantly extends the class of light-driven molecular spin machines, which pump nuclear spin-hyperpolarization. The LOV systems are particularly suitable for such biomimetic design, while photosynthetic RCs due to their structural complexity allow for limited manipulations only. The biomimetic design that affects the parameters of the photo-CIDNP effect provides new categories of experiments to analyze the conditions for its occurrence.
# Methods
Protein preparation. The plasmid (i) encoding the LOV-C287S mutant of aureochrome1a from P. tricornutum comprising the flanking Jα and A'α helices (238-378) was received from Peter Kroth (University of Konstanz) [bib_ref] Blue-light-induced unfolding of the Jα helix allows for the dimerization of aureochrome-LOV..., Herman [/bib_ref]. The plasmid (ii) that encodes Mr4511 from M. radiotolerans (1-164) was generated in our own group by genome cloning [bib_ref] Single mutation in a novel bacterial LOV protein yields a singlet oxygen..., Consiglieri [/bib_ref]. All genetic manipulations were according to standard protocols. Plasmid (iii) encodes the LOV1-C57S mutant of phototropin from C. reinhardtii (16-133) carrying a 15 × His-tag at the N-terminus [bib_ref] The phot LOV2 domain and its interaction with LOV1, Guo [/bib_ref]. Further information about all the mutants employed in this work is summarized in . The protocol of heterologous overexpression and isotope-labelling of all the mutants in Escherichia coli has been reported elsewhere [bib_ref] Nuclear spin-hyperpolarization generated in a flavoprotein under illumination: experimental field-dependence and theoretical..., Ding [/bib_ref] ] glucose as the sole source in the growth medium yields a uniformly 15 N or 13 C labelled protein, while supplementation of 13 C or 15 N isotope-enriched indole as precursor to the normal medium results in a selective labelling of the tryptophan side chain. The 15 N and 13 C labelled proteins were used for corresponding [bib_ref] Photo-CIDNP 13 C magic angle spinning NMR on bacterial reaction centres: exploring..., Matysik [/bib_ref] N and 13 C NMR measurements. For 1 H NMR measurement, the employed proteins are in their natural abundance, and they were washed with deuterated phosphate buffer (300 mM NaCl, 50 mM KsPO 4 in D 2 O, pH 8.0) to a final solution containing approximately 0.4% residual protons. The final concentration of the flavoproteins were controlled at about 16 μM ( ε 450nm = 12,500 M −1 cm −1 ) before photo-CIDNP measurement.
## Scientific reports
| (2020) 10:18658 | https://doi.org/10.1038/s41598-020-75627-z www.nature.com/scientificreports/ Photo-CIDNP solution-state NMR. The field-dependent 15 N-, 13 C-, and 1 H photo-CIDNP experiments of the LOV proteins were carried out on an NMR spectrometer operating at a magnetic field of 9.4 T ( 1 H frequency of 400 MHz) (Bruker Avance III HD) equipped with a home-built field-cycling device [bib_ref] Field-cycling NMR experiments in an ultra-wide magnetic field range: relaxation and coherent..., Zhukov [/bib_ref]. It transfers the sample to variable magnetic fields within the range 10 nT < B o < 9.4 T at which the sample is illuminated and returns it for the NMR detection at 9.4 T. For the 13 C and 15 N photo-CIDNP experiments, pulse-acquire with WALTZ-16 proton decoupling was employed. The pulse sequence of the 1 H photo-CIDNP experiments starts with a pre-saturation composite pulse sequence 68 at 9.4 T, followed by the sample shuttle cycle that includes the sample transfer to the chosen magnetic field for illumination by LED (called "light") or the same cycle without illumination (called "dark") during the fixed time and the return to 9.4 T, and it ends with the detection sequence. For all photo-CIDNP experiments, the samples were measured in dark and light, respectively, with the same number of scans. The illumination source was a 400-nm 2-W LED (Chanzon, China) and the illumination time was optimized to 0.5 s. A fresh aliquot of the sample stock was taken for a measurement at each magnetic field to compensate the effect of photo-bleaching. The temperature was 289 K for all samples with the exception that Ptaureo1a-LOV-C287S was measured at 277 K. The line-broadening for [bib_ref] Photo-CIDNP 13 C magic angle spinning NMR on bacterial reaction centres: exploring..., Matysik [/bib_ref] N and 13 C NMR spectra were set to 30 Hz and for 1 H spectra was set to 1 Hz. The 15 N and 13 C NMR spectra were phased to the external standard, a mixture of 0.1 M 15 N labelled urea and 0.1 M 13 C labelled methanol in DMSO. The chemical shifts of 15 N NMR spectra were relative to liquid ammonia and referenced externally to urea 15 N at 76.3 ppm 69 . To present the field-dependence of the photo-CIDNP effect, the selected hyperpolarized signal was integrated and plotted against the field at which the sample was illuminated. For straightforward comparison, the signal strength at the hyperpolarization maximum is set to unity and the other signals are normalized to this value. The positive and negative signs of the y-axis indicate absorptive and emissive hyperpolarization, respectively. The error bars of the 1 H photo-CIDNP data represent the standard deviation of the mean value obtained from three measurements; the uncertainty of the 13 C and 15 N photo-CIDNP data represents the noise level relative to the corresponding hyperpolarized signal area. The spectra shown in [fig_ref] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1 [/fig_ref] and Supplementary Information , S2, and S4 were created with OriginPro Version 2017.
Received: 6 July 2020; Accepted: 15 October 2020
[fig] Figure 2: The figure was created by the PyMOL Molecular Graphics System, Version 1.2r3pre, Schrödinger, LLC. (C) IUPAC numbering of atomic positions in the isoalloxazine ring of FMN and the side chain of tryptophan. (A) 1 H, (B)13 C and (C) 15 N photo-CIDNP spectra of Crphot-LOV1-C57S (~ 11 Å) detected at 289 K at different magnetic field strengths by using a 9.4 T liquid-state NMR spectrometer equipped with a fieldcycling device. Hereafter, all photo-CIDNP spectra are light-minus-dark NMR spectra. The upper trace includes signal assignments (blue) with 13 C chemical shifts of hyperpolarized signals mentioned inSupplementary Information [/fig]
[fig] Figure 3: Intensity of selected hyperpolarized signals of Crphot-LOV1-C57S ( r FW ~ 11 Å), integrated and plotted against the magnetic field strength for (A) 1 H, (B)13 C and (C)15 N. For convenient comparison, the magnitude of the maximum polarization of each nucleus is arbitrarily set to − 1 and the signal enhancement is normalized to this value. The plus and minus signs at the y-axis represent the enhanced absorptive (positive) and emissive polarization (negative) relative to the thermal polarization. Uniformly 13 C-and uniformly15 N-enriched Crphot-LOV1-C57S (~ 11 Å) samples were employed for the13 C and 15 N photo-CIDNP NMR experiments, respectively. Regarding 1 H photo-CIDNP, the signal of HDO, the aliphatic region (0 to 3 ppm), and the broader region (− 2 to 10 ppm) of the 1 H photo-CIDNP spectrum were integrated for comparison. Scientific Reports | (2020) 10:18658 | https://doi.org/10.1038/s41598-020-75627-z www.nature.com/scientificreports/ and the 1 H spectrum intensity integrated from -2 to 10 ppm of Crphot-LOV1-C57S ( r FW ~ 11 Å) (Fig. 4A). The field at which the 1 H, 13 C and 15 N photo-CIDNP production reached its maximum follows the relationship: 0.6 T ( 1 H) < 3 T ( 13 C) < 5 T ( 15 N). Under similar conditions, the field maxima for photo-CIDNP enhancement of the corresponding nuclei in Ptaureo1a-LOV-C287S ( r FW ~ 11 Å) (tryptophan is selectively isotope-labelled) were obtained. The positions of the maxima are in the same order [/fig]
[fig] Figure 4: Comparison of the field dependencies of the 1 H and 13 C and 15 N photo-CIDNP effects generated by (A) Crphot-LOV1-C57S ( r FW ~ 11 Å) and (B) Ptaureo1a-LOV-C287S ( r FW ~ 11 Å). Detailed experimental parameters are provided in the "Methods" section. The hyperpolarized 1 H signal has been obtained by integrating the spectral region of − 2 to 10 ppm (black square); the 13 C signal from C-3 of tryptophan (red triangle) and the 15 N signal from indole of tryptophan (blue circle) are also shown. The dashed lines are added to guide the eye. The maximum emissive polarization of all nuclei is scaled to − 1. The minus sign of the y-axis implies that the hyperpolarization is emissive. The 1D spectra of the photo-CIDNP effect observed in Ptaureo1a-LOV-C287S are shown in Supporting InformationFigure S1. Mr4511-C71S-F130W ( r FW ~ 6 Å), which has the shortest distance between the redox partners of the SCRP, shows neither a 1 H nor a 15 N photo-CIDNP in the magnetic field range from 0.1 to 9.4 T(Fig. 5A1,A2;Supplementary Information [/fig]
[fig] Figure 5: MagneticMr4511-C71S-Y116W ( r FW ~ 9 Å), (C) Mr4511-C71S-Q112W ( r FW ~ 11 Å), (D) Mr4511-C71S-Y129W ( r FW ~ 11 Å), (E) Mr4511-C71S-K57W ( r FW ~ 17 Å) and (F) Mr4511-C71S. [/fig]
[table] Table 2: Photo-CIDNP experimental results and obtained 15 N chemical shifts from the eight LOV proteins studied in this work. [/table]
|
Selective Methylation of CpGs at Regulatory Binding Sites Controls NNAT Expression in Wilms Tumors
Aberrant expression of imprinted genes, such as those coding for the insulin-like growth factor 2 (IGF2) and neuronatin (NNAT), is a characteristic of a variety of embryonic neoplasms, including Wilms tumor (WT). In case of IGF2, it is generally accepted that loss of imprinting in a differentially methylated region of the IGF2/H19 locus results in biallelic expression and, thus, upregulation of the gene. In this study we examined methylation pattern at potential regulatory elements of the paternally expressed NNAT gene in a cohort of WT patients in order to further characterize the molecular mechanism causing overexpression of this regulatory gene. We demonstrate that transcriptional upregulation of NNAT in WT is grossly independent of the bladder cancer-associated protein (BLCAP) gene, an imprinted gene within the imprinted domain of the NNAT locus. However, expression of the BLCAP transcript isoform v2a formerly known to be selectively expressed from the paternal allele in brain was associated with high expression of NNAT. This contrasts the situation we found at the IGF2/H19 locus, which shows high overexpression of IGF2 and inversely correlated expression of the H19 gene in WT. An analysis of DNA methylation in two potential regulatory regions of the NNAT locus by pyrosequencing revealed significant hypomethylation of the tumors compared to normal kidney tissue. Interestingly, the difference in DNA methylation was highest at CpGs that were observed within three putative binding sites of the CCCTC-binding factor CTCF. Most importantly, hypomethylation of both NNAT regulatory regions is significantly associated with the upregulation of NNAT expression and the BLCAP_v2a transcript. Our data indicate that the methylation status of a not-yet-described regulatory element within the NNAT locus that contains four potential CTCF binding sites determines the expression level of NNAT and the nearby located BLCAP_v2a transcript, thereby suggesting a functional role in the aberrant upregulation of NNAT in WT.
# Introduction
Genomic imprinting is the parent-of-origin silencing that results in monoallelic expression, with the associated allelic methylation at imprinting control (ICR) or differential methylated regions (DMR) being maintained irrespective of the transcriptional output [bib_ref] How imprinting centres work, Lewis [/bib_ref]. These regions are characterized by cytosine-phosphate-guanine (CpG)-rich islands and contain binding sites for regulatory factors. High levels of methylation at these regions are suspected to prevent binding to regulatory elements such as promoters and insulators. The mechanism of genomic imprinting is known to silence developmental genes after the embryonic period [bib_ref] How imprinting centres work, Lewis [/bib_ref]. Recent studies have shown that these imprinting patterns are aberrant in different neoplasms, especially in embryonal tumors [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref] [bib_ref] Loss of imprinting of IGF2 and H19 in osteosarcoma is accompanied by..., Ulaner [/bib_ref] [bib_ref] Loss of imprinting of insulin-like growth factor-II in Wilms' tumor commonly involves..., Cui [/bib_ref] [bib_ref] Epigenetic alteration at the DLK1-GTL2 imprinted domain in human neoplasia: analysis of..., Astuti [/bib_ref].
Wilms tumor (WT), or nephroblastoma, accounts for 6% of all solid tumors in patients less than 15 years of age with an incidence of eight per one million children [bib_ref] Wilms Tumor, Davidoff [/bib_ref]. These malignant embryonal tumors are thought to develop from precursor cells at an early stage of metanephric differentiation [bib_ref] Gene expression in Wilms' tumor mimics the earliest committed stage in the..., Li [/bib_ref]. Overexpression of the insulin-like growth factor 2 (IGF2) has been reported as a common characteristic in both WT and other embryonal tumors [bib_ref] New prognostic markers revealed by evaluation of genes correlated with clinical parameters..., Wittmann [/bib_ref] [bib_ref] Loss of imprinting of the insulin-like growth factor II gene occurs by..., Nakagawa [/bib_ref] [bib_ref] Mosaic allelic insulin-like growth factor 2 expression patterns reveal a link between..., Ohlsson [/bib_ref].
The IGF2 gene is part of a cluster together with H19, a non-coding RNA gene, and the expression of both genes is regulated by a central DMR (H19DMR).To date, the IGF2/H19 cluster and its regulation is the best-investigated imprinted region, and its role has been studied in many embryonal and adult cancers [bib_ref] Loss of imprinting of insulin-like growth factor-II in Wilms' tumor commonly involves..., Cui [/bib_ref] [bib_ref] Epigenetic alteration at the DLK1-GTL2 imprinted domain in human neoplasia: analysis of..., Astuti [/bib_ref] [bib_ref] Loss of imprinting of the insulin-like growth factor II gene occurs by..., Nakagawa [/bib_ref]. In somatic cells, H19DMR is methylated only on the paternal allele that expresses IGF2 [bib_ref] Epigenetic differences between Wilms' tumours in white and east-Asian children, Fukuzawa [/bib_ref]. As a consequence, some regulatory elements, such as the CCCTC-binding factor (CTCF), are able to bind only on the maternal allele. CTCF is an 11-zinc-finger protein that acts as a cis-regulatory transcription factor with insulator activity [bib_ref] How imprinting centres work, Lewis [/bib_ref]. It is generally accepted that loss of imprinting (LOI) at the IGF2/H19 locus results in biallelic expression of the fetal growth factor IGF2 and contributes to the development of embryonal tumors. However, overexpression of IGF2 independent from LOI at the H19DMR has also been described for WT [bib_ref] Increased expression of the insulin-like growth factor-II gene in Wilms' tumor is..., Wang [/bib_ref].
We have shown in a previous study that the aberrant methylation of H19DMR is associated with the overexpression of other imprinted genes, including neuronatin (NNAT) [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref]. Those findings suggest a general overexpression of imprinted genes in WT by an as-yet-unknown superordinated mechanism. NNAT is a developmental gene involved in differentiation of neuronal tissues that may play a role in tumorigenesis [bib_ref] Gene expression in Wilms' tumor mimics the earliest committed stage in the..., Li [/bib_ref]. Gene expression is regulated by a DMR spanning the first two exons [bib_ref] The neuronatin gene resides in a ''micro-imprinted'' domain on human chromosome 20q11, Evans [/bib_ref]. NNAT is located in an imprinted domain within the intron of the bladder cancer-associated protein (BLCAP) gene [bib_ref] The neuronatin gene resides in a ''micro-imprinted'' domain on human chromosome 20q11, Evans [/bib_ref]. Recent studies described a tissue and parent-of-origin specific expression of BLCAP and see the NNAT gene as a part of a NNAT/BLCAP imprinting cluster [bib_ref] Transcript-and tissue-specific imprinting of a tumour suppressor gene, Schulz [/bib_ref].
The aim of this study was to analyze the methylation pattern at regulatory elements of the NNAT gene in a cohort of WT patients and compare its influence on gene expression of NNAT and its neighbor BLCAP.
# Materials and methods
## Patients
Forty-five WT specimens were investigated from patients undergoing surgical tumor resection in our department. The median age at time of surgery was 34.5 months (range 2 months to 17 years), with a sex ratio of 1:1.5 (f:m). Thirty-seven patients (82%) underwent neoadjuvant chemotherapy according to the International Society of Pediatric Oncology (SIOP) protocol [bib_ref] Reduction of postoperative chemotherapy in children with stage I intermediaterisk and anaplastic..., De Kraker [/bib_ref]. Eleven patients (24%) were found to have bilateral WT. The control group (n = 11) consisted of renal tissue from the healthy part of the resected specimen after tumor nephrectomy. The median age of the control group was 39 months (range 3 months to 5 years) with a sex ratio of 1:2.7 (f:m). Histological classification of the samples was performed by a pathologist. The study was approved by the ethics committee of the Ludwig-Maximilians-University of Munich. Written consent was obtained from all parents.
## Real-time reverse transcription-pcr (rt-pcr)
Tri ReagentH was used for the isolation of total RNA from native samples. Total RNA and DNA were separated and subsequently purified using DNase and an RNeasy Mini Kit, respectively (Qiagen, Hilden, Germany). Reverse transcription of total RNA was performed using random hexamers (Roche Diagnostics, Penzberg, Germany) and SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA). Intron-spanning primers were designed for the human genes IGF2, H19, BLCAP, and NNAT using Primer ExpressH v2.0 (Applied Biosystems, Foster City, CA, USA) based on the sequence information contained in the Ensembl Database. Primers were as follows (592.39 orientation):
NNAT: CGGCTGGTACATCTTCCGC, TGTCCCTGGAGGATTTCGAAA;
IGF2: CCCGCTGGGCCAATCT, GAGTCTGGTTTTGATGC-CACC;
H19: CTCACCCACCGCAATTCATT, CGTGCCGGAGCTGCC; BLCAP: AACGGAAGCCTTGCA-CAATT, ATCGGAGCAGTGGTACAGGAA; BLCAP_v1a: GTGGCGAGCTGAGGTGGA, ATCGGAGCAGTGGTA-CAGGAA; BLCAP_v2a: TTTTCTGCTGGACAGGTCGTT, TGCTCTCTGGCTGTCAGCC.
Polymerase chain reaction (PCR) amplifications were carried out with 40 ng of cDNA, 500 nM forward and reverse primer and iTaq SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA, USA) in a final volume of 20 ml. The PCR reactions were run for 40 cycles consisting of 15 sec denaturation at 95uC, primer annealing for 15 sec at 55uC, and extension for 30 sec at 72uC on a Mastercycler Realplex 2 cycler (Eppendorf, Hamburg, Germany). All PCR reactions were prepared in doublets and standardized to the reference gene TATA-Box-binding-Protein (TBP). The level of expression was calculated according to the mathematical model of Pfaffl et al. [bib_ref] A new mathematical model for relative quantification in realtime RT-PCR, Pfaffl [/bib_ref].
## Methylation analysis using pyrosequencingh
Double stranded DNA (dsDNA) was extracted from the native samples using standard procedures. Two micrograms of dsDNA were converted using the Epitec Bisulfite KitH (Qiagen) according to the user manual. Pyrosequencing primers were designed with PyroMark Assay Design Software 2.0 and were as follows (592.39 orientation):
NNAT-promoter: GGGGTAGGTATGGAAAGAGTAGA, AAACCCCCTCAAACTTACCTACAAC, GGTGGAGG-GAGGGTATTTAA (sequencing primer); NNAT-CTCF: TGTTTGGGTTTTGTTTAGTAGAGAT, AATACTAAATA-CAAAACCCTATCCTTAC, GGTTTTGTTTAGTAGAGAT-TAT (sequencing primer); BLCAP-promoter: AGTTTTT-GAGTTTTGGTTGTATGATGAA, AATCCCCTCATACA-TATACAACAA, GGGGTGGATTTTGAGTTA (sequencing primer); H19-DMR: GTATAGAGTTAGGGGGTTTTTGTA-TAGTA, CTCCCATAAATATCCTATTCCCAAATA, GGTTTTATAGTTTGGATGGT (sequencing primer). The DNA was amplified by PCR using 50 ng of dsDNA, 500 nM forward and reverse primer, and Maxima HotStart-Taq (Fermentas, Glen Burnie, USA) in a final volume of 15 ml. The PCR reactions were run for 45 cycles consisting of 20 sec denaturation at 95uC, primer annealing for 20 sec at 58.3uC, and extension for 35 sec at 72uC on a Mastercycler Realplex 2 cycler (Eppendorf, Hamburg, Germany). PCR products were sequenced on a PyroMark TM Q24 instrument (Qiagen) using sequencing primer and ProMark TM Gold reagent as recommended by the manufacturer. Quantitative analysis of methylation was accomplished using the Pyro Q-CpG Software (Qiagen).
## Statistics
A statistical analysis was carried out with SPSSH v19.0. An explorative analysis was made without correcting p-values for multiple testing. The Mann-Whitney U-Test was used for comparison of gene expression data and methylation status. The relationship between two expression levels was measured by Spearman's rank correlation.
# Results
## Gene expression at the igf2/h19 and nnat/blcap imprinted loci in wilms tumor
First, we investigated the transcriptional activity of the genes IGF2 and H19 as well as NNAT and BLCAP in a cohort of 45 WT and normal kidney tissue. IGF2 (p,0.0001) and NNAT (p = 0.0002) were significantly overexpressed, whereas expression of BLCAP (p = 0.0063) was significantly suppressed in tumors compared to healthy kidney tissue [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref] , d, e). The median H19 expression level did not differ significantly from kidney (p = 0.2701) [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref]. As different BLCAP transcripts have been described [bib_ref] Transcript-and tissue-specific imprinting of a tumour suppressor gene, Schulz [/bib_ref] , we further analyzed the expression level of the main v1a and v2a transcripts. Whereas the expression of BLCAP_v1a in tumors did not differ from kidney (p = 0.3076) [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref] , we found an overexpression of the BLCAP_v2a transcript in a subset of tumors [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref]. Of note, the BLCAP_v2a transcript was expressed at much lower levels than the BLCAP_v1a transcript.
Next, we correlated the expression level of the genes contained within the imprinted loci to each other using the Spearmaǹs rank correlation coefficient. We found that expression of IGF2 is negatively associated with that of H19 (p = 0.0003, r = 20.4747) [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref] , corroborating earlier findings that IGF2 is negatively regulated by H19 [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref] [bib_ref] Mechanisms regulating imprinted genes in clusters, Edwards [/bib_ref]. In contrast, there was no correlation between NNAT and BLCAP (p = 0.7701, r = 0.0424) [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref] as well as its v1a transcript (p = 0.1949, r = 0.1619) (not shown). Interestingly, there was a significant correlation between NNAT and BLCAP_v2a expression (p = 0.0045, r = 0.3774) [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref] as well as NNAT and IGF2 expression (p,0.001, r = 0.459) (not shown).
## Dna methylation at the igf2/h19 and nnat/blcap imprinted loci in wilms tumor
To determine the methylation pattern at the IGF2/H19 locus we established a pyrosequencing assay that analyzes six CpGs inside the H19DMR spanning one CTCF binding site [fig_ref] Figure 2: Schematic illustration of the A [/fig_ref]. When calculating the mean methylation level across all six CpGs and defining hypo-and hypermethylation as a value of more than 2 standard deviations below and above the mean of the controls, we found all normal kidney samples and 32/45 tumors to display normal methylation values [fig_ref] Figure 3: Mean DNA methylation levels at the A [/fig_ref] , thereby suggesting a monoallelic methylation pattern of the H19DMR in these cases. However, 9 of 45 tumors showed hypermethylation (mean methylation .71.8%), which implies biallelic methylation and, thus, loss of imprinting. Interestingly, we also detected four cases displaying hypomethylation (mean methylation ,34.8%). By focusing at each individual CpG, we found that the mean methylation across all tumors was comparable to that of the kidney tissues [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref] , which is plausible because of tumor cases with normal, hyper-and hypo-methylation. Thus, we repeated the analysis by including only tumors with IGF2 overexpression (n = 20 with .10-fold mean IGF2 expression of kidney), assuming that these tumors should show the strongest methylation differences. Indeed, we detected significantly higher mean methylation levels in tumors compared to normal kidney samples for the six informative CpGs [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref].
To analyze the methylation status of the NNAT/BLCAP locus, we first screened the genomic sequence around the NNAT and BLCAP coding sequence for potential imprinting control regions. We defined two regions of interest that are located inside of CpG islands and presumably lay within the NNAT and BLCAP promoter [fig_ref] Figure 2: Schematic illustration of the A [/fig_ref]. A third region covers four potential CTCF binding sites 1.5 kb upstream of exon 1 [fig_ref] Figure 2: Schematic illustration of the A [/fig_ref] of the NNAT gene [fig_ref] Figure 2: Schematic illustration of the A [/fig_ref]. By analyzing 10 CpGs of the NNAT CTCF binding site by pyrosequencing, we found a dramatic decrease in the overall methylation across all CpGs in the tumor samples (median 40%) versus normal kidney (median 60%) [fig_ref] Figure 3: Mean DNA methylation levels at the A [/fig_ref]. By looking at the individual CpGs, we found considerable variation of methylation levels, ranging from 5%-100% [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref]. Generally, mean methylation levels were higher in normal kidneys than in tumor tissues, with the most significant differences in CpGs 6-8, which span the core region of the CTCF binding sites. This difference was even more pronounced when we included only tumors with NNAT overexpression (n = 33 with .10-fold mean NNAT expression of kidney) [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref] , thereby suggesting that hypermethylation of the potential CTCF binding site inside the NNAT CpG islands might play an important role in the physiological regulation of NNAT.
At the NNAT promoter region [fig_ref] Figure 2: Schematic illustration of the A [/fig_ref] , we also found a strong reduction of the mean methylation level in tumor (median 45%) versus kidney tissues (median 63%) [fig_ref] Figure 3: Mean DNA methylation levels at the A [/fig_ref]. In addition, the mean methylation rates at the individual CpGs differed considerably between normal kidney and tumor tissue, with significantly higher methylation levels in kidney [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref]. This difference remained nearly unchanged when including only the 33 tumors with NNAT overexpression [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref]. Interestingly, we detected a high correlation between the methylation of the NNAT CTCF binding site and the NNAT promoter (p,0.0001; r = 0.7234).
In contrast to the CTCF binding site and promoter region of the NNAT gene, the mean methylation at the BLCAP promoter was very low, ranging from 1 to 9%, and showed no differences between tumors and normal kidneys [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref].
## Correlation of gene expression and dna methylation at the igf2/h19 and nnat/blcap imprinted loci in wilms tumor
To investigate the influence of the DNA methylation pattern at the IGF2/H19 and NNAT/BLCAP imprinted loci on the transcriptional activity of the genes, we correlated methylation with expression using the Spearman-rank correlation.
At the IGF2/H19 locus, we revealed a correlation between high methylation levels of the H19DMR and the overexpression of IGF2 (p = 0.0123, r = 0.3447) [fig_ref] Figure 5: Association of gene expression and mean methylation of A [/fig_ref]. In sharp contrast, the second gene of interest at this locus, H19, was significantly suppressed by a high methylation level of the H19DMR (p,0.0002, r = 20.4986), as expected at this locus [fig_ref] Figure 5: Association of gene expression and mean methylation of A [/fig_ref]. At the NNAT/BLCAP locus, high methylation levels at both the NNAT CTCF binding site (p,0.0001; r = 20.6385) and the NNAT promoter (p,0.0001; r = 20.6902) were associated with low NNAT expression levels [fig_ref] Figure 5: Association of gene expression and mean methylation of A [/fig_ref]. Although NNAT is located within the intron of BLCAP, mRNA expression of BLCAP including the BLCAP_v1a transcript was neither influenced by the methylation status of the NNAT-CTCF binding site nor of the NNAT promoter (data not shown). However, expression of the BLCAP_2a transcript was inversely correlated with methylation level at the NNAT CTCF binding site (p,0.0428; r = 20.2728) [fig_ref] Figure 5: Association of gene expression and mean methylation of A [/fig_ref] , but not the NNAT promoter (p,0.0580; r = 20.2621) (data not shown).
# Discussion
Overexpression of imprinted genes is a common phenomenon in several embryonal tumors, including WT. We previously speculated that this might be mediated by a yet unknown superordinated mechanism [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref]. Here, we report on the comparative analysis of two imprinted genomic regions, the IGF2/H19 and NNAT/BLCAP loci, in regard to gene expression and DNA methylation in a large collection of WT. Both loci were chosen due to the known overexpression of IGF2 and NNAT in embryonal tumors [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref] [bib_ref] Multiple imprinted and stemness genes provide a link between normal and tumor..., Dekel [/bib_ref] and their key role in embryonic development [bib_ref] Multiple imprinted and stemness genes provide a link between normal and tumor..., Dekel [/bib_ref] [bib_ref] Segment-specific expression of the neuronatin gene during early hindbrain development, Wijnholds [/bib_ref]. While IGF2 is part of a gene cluster with the H19DMR being responsible for the expression of both IGF2 and H19, the regulation of NNAT located within the intron of the gene BLCAP is less far understood [bib_ref] The neuronatin gene resides in a ''micro-imprinted'' domain on human chromosome 20q11, Evans [/bib_ref]. Recent studies have postulated a tissue and transcript specific regulation of NNAT and BLCAP in a cluster, at least in the brain [bib_ref] Transcript-and tissue-specific imprinting of a tumour suppressor gene, Schulz [/bib_ref].
To date, the functional relevance of CTCF binding in regulating the expression of imprinted genes has been extensively studied at the IGF2/H19 locus [bib_ref] Loss of imprinting of the insulin-like growth factor II gene occurs by..., Nakagawa [/bib_ref] [bib_ref] Novel cis-regulatory function in ICR-mediated imprinted repression of H19, Ideraabdullah [/bib_ref] , but not at the NNAT/ BLCAP locus. It is generally accepted that the methylation patterns of binding sites for regulatory proteins are responsible for the transcriptional regulation of many imprinted genes. CTCF is one of the factors that are involved in gene regulation, especially at the H19DMR and is well-studied at this locus [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref] [bib_ref] Epigenetic differences between Wilms' tumours in white and east-Asian children, Fukuzawa [/bib_ref]. By researching the Ensembl database, we detected a regulatory region of the NNAT locus that harbors four putative CTCF binding sites 1.5 kb upstream of the first exon of NNAT [fig_ref] Figure 2: Schematic illustration of the A [/fig_ref] and outside of its described promoter [bib_ref] The neuronatin gene resides in a ''micro-imprinted'' domain on human chromosome 20q11, Evans [/bib_ref]. Thus, the aim of our study was to investigate the role of this CTCF binding region in the misregulation of the NNAT gene in WT and compare it to the respective site at the IGF2/H19 locus.
In a first set of experiments we confirmed the well-known overexpression of IGF2 in WT compared to kidney tissues [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref] and 1d) [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref] [bib_ref] Multiple imprinted and stemness genes provide a link between normal and tumor..., Dekel [/bib_ref] , with the expected coordinate downregulation of H19 [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref] [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref] [bib_ref] Epigenetic differences between Wilms' tumours in white and east-Asian children, Fukuzawa [/bib_ref] [bib_ref] Mechanisms regulating imprinted genes in clusters, Edwards [/bib_ref]. Moreover, we have shown hypermethylation of the CTCF binding site in 9 WT [fig_ref] Figure 3: Mean DNA methylation levels at the A [/fig_ref] , which is in line with data described in earlier studies [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref] [bib_ref] Epigenetic differences between Wilms' tumours in white and east-Asian children, Fukuzawa [/bib_ref] [bib_ref] Mechanisms regulating imprinted genes in clusters, Edwards [/bib_ref] [bib_ref] CTCF-Dependent Chromatin Insulator Is Linked to Epigenetic Remodeling, Ishihara [/bib_ref]. As expected, there was a high correlation between hypermethylation and overexpression of IGF2 [fig_ref] Figure 5: Association of gene expression and mean methylation of A [/fig_ref] and a coordinate downregulation of H19 [fig_ref] Figure 5: Association of gene expression and mean methylation of A [/fig_ref]. Using pyrosequencing, we were also able to analyze not only general methylation, but also the methylation of each single CpG in the region of interest, thereby revealing considerable differences with regard to methylation between each single CpG [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref]. However, these differences were only relevant for those tumors with a .10-fold mean expression of normal kidney [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref]. This could be explained by the inclusion of cases with hyper-and hypomethylation at H19DMR in our calculations [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref]. Nevertheless, the most significant differences in the degree of methylation were observed at CpGs 1-3 [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref]. Altogether, these data qualified our cohort of WT patients to be representative.
Next, we analyzed the transcriptional activity of NNAT and BLCAP in our samples. We thereby replicated the previously described overexpression of NNAT [bib_ref] Altered expression of imprinted genes in Wilms tumors, Hubertus [/bib_ref] [bib_ref] Gene expression in Wilms' tumor mimics the earliest committed stage in the..., Li [/bib_ref] [bib_ref] Aberrant expression of tenascin-c and neuronatin in malignant peripheral nerve sheath tumors, Dugu [/bib_ref]. Aberrant transcriptional activity of NNAT has been reported in other tumors such as glioblastomas [bib_ref] Neuronatin in a subset of glioblastoma multiforme tumor progenitor cells is associated..., Xu [/bib_ref] , neoplasms of eccrine, apocrine and sebaceous glands [bib_ref] Abundant expression of neuronatin in normal eccrine, apocrine and sebaceous glands and..., Dugu [/bib_ref] , malignant peripheral nerve sheath tumors [bib_ref] Aberrant expression of tenascin-c and neuronatin in malignant peripheral nerve sheath tumors, Dugu [/bib_ref] , and others. Xu et al. described a correlation of NNAT overexpression and increased cell proliferation in glioblastomas associated with poorer outcome [bib_ref] Neuronatin in a subset of glioblastoma multiforme tumor progenitor cells is associated..., Xu [/bib_ref]. In contrast, BLCAP was generally suppressed in WT [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref] , when we simultaneously measured all transcript isoforms. However, by analyzing the expression of the two main BLCAP transcripts v1a and v2a that have their transcription starts either 59 or 39 of the NNAT gene we detected a more differentiated transcriptional activity. We found comparable expression levels of BLCAP_v1a in WT and normal kidneys, whereas the BLCAP_v2a transcript was overexpressed in a subset of tumors, even at very low levels [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref]. Analogous, we have seen a correlation of the NNAT and BLCAP_v2a expression [fig_ref] Figure 1: Relative expression of the genes A [/fig_ref]. This may be explained by the postulated paternal expression of both, NNAT and BLCAP_v2a, while BLCAP_v1a seems to be maternally expressed [bib_ref] Transcript-and tissue-specific imprinting of a tumour suppressor gene, Schulz [/bib_ref]. Moreover, this finding suggests that NNAT and BLCAP_v2a might be regulated together by a shared regulatory element.
Indeed, we qualified a small region containing potential CTCF sites to be highly correlated with the expression of NNAT and BLCAP_v2a. The mean level of methylation of the potential CTCF binding sites and the promoter was lower in tumors compared to kidney tissues [fig_ref] Figure 3: Mean DNA methylation levels at the A [/fig_ref]. Thus, our data suggest that the majority of WT exhibit hypomethylation of these two sites. This is contrary to findings in pituitary adenoma, which are characterized by a hypermethylation of adjacent CpG islands and a downregulation of NNAT [bib_ref] Loss of neuronatin expression is associated with promoter hypermethylation in pituitary adenoma, Revill [/bib_ref]. This might be explained by the finding that the pituitary gland is the only adult tissue with high transcriptional activity of NNAT [bib_ref] Loss of neuronatin expression is associated with promoter hypermethylation in pituitary adenoma, Revill [/bib_ref]. From the mechanistically point of view, our data are in accordance with showing a strong association of hypermethylation and gene silencing for the promoter and, even more interestingly, with the newly described potential CTCF binding site. While all of the promoter CpGs had a lower mean level of methylation in tumors [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref] , methylation especially of CpGs 6-8 containing three of the four potential CTCF binding sites, considerably differed from kidney tissue [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref]. As seen at H19DMR [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref] , these differences are intensified when we look at tumors with relevant increased expression levels [fig_ref] Figure 4: Mean DNA methylation levels delineated for each CpG at A [/fig_ref]. However, since the methylation status of both the potential CTCF binding site and the promoter are correlated with the level of expression [fig_ref] Figure 5: Association of gene expression and mean methylation of A [/fig_ref] , we cannot discriminate the influence of each individual regulatory element on gene regulation. In addition, we found no methylation of the BLCAP promoter, which is in line with data from fetal brain and other tissues [bib_ref] The neuronatin gene resides in a ''micro-imprinted'' domain on human chromosome 20q11, Evans [/bib_ref] , thereby explaining the general expression of BLCAP in both the tumors and the normal kidney tissues.
Finally, because CTCF functions as an insulator [bib_ref] CTCF-Dependent Chromatin Insulator Is Linked to Epigenetic Remodeling, Ishihara [/bib_ref] [bib_ref] More than insulator: multiple roles of CTCF at the H19-Igf2 imprinted domain, Singh [/bib_ref] and the NNAT CTCF binding site is placed inside a BLCAP intron, a potential influence on BLCAP expression can be assumed. However, there was no correlation of the total BLCAP and BLCAP_v1a expression with methylation at the potential NNAT CTCF binding site. However, the BLCAP_v2a transcript was expressed in tumors only and was correlated with the NNAT expression as well as the methylation of the NNAT CTCF binding site. These findings corroborate the results of Schulz et al. [bib_ref] Transcript-and tissue-specific imprinting of a tumour suppressor gene, Schulz [/bib_ref] , which described a transcript and tissue specific imprinting of the NNAT/BLCAP locus. Nevertheless, the reactivation of NNAT and other imprinted genes may be attributed to genome-wide altered methylation of regulatory binding sites. The question of whether the potential NNAT CTCF binding site in WT has never been methylated and mirrors the embryonic state rather than being actively demethylated during tumorigenesis cannot be answered to date.
In conclusion, our data suggest that, in contrast to the IGF2/ H19 locus, hypomethylation of regulatory elements are associated with NNAT overexpression. We have shown that differences in methylation are most prominent at a few CpGs covering potential CTCF binding sites, whereas differential methylation of the NNAT promoter region is evident over a large proportion. This might lead to the concomitant overexpression of NNAT and BLCAP_v2a transcripts in WT, while leaving the general expression of BLCAP unaffected. Thus, the process of aberrant methylation in WT seems to be a targeted and site-specific process that impacts selected regulatory sites rather than a general ongoing mechanism acting throughout the genome. This suggests that methylating enzymes may be reliant on motifs in the DNA.
[fig] Figure 1: Relative expression of the genes A) IGF2, B) H19, D) NNAT E) BLCAP, G) BLCAP_v1a, and H) BLCAP_v2a in 45 Wilms tumors and 11 normal kidney tissues, as determined by real-time PCR. The dots represent relative candidate gene expression in relation to the house-keeping gene TBP. Median expression values are given as red horizontal lines. Association of C) IGF2 and H19, F) NNAT and BLCAP, as well as I) NNAT and BLCAP_v2a gene expression in log2 scale using Spearman's rank correlation. Data from the tumor and normal kidney cases are depicted as black and orange diamonds, respectively. doi:10.1371/journal.pone.0067605.g001 [/fig]
[fig] Figure 2: Schematic illustration of the A) IGF2/H19 and B) NNAT/BLCAP locus. Transcription starts are indicated by arrows and transcript variants are numbered. The CTCF binding sites (yellow bars) within the A) H19DMR and B) CpG-rich intergenic NNAT region as well as the NNAT and BLCAP promoter used for generating the DNA methylation assays are highlighted as magnifications indicating each CpG dinucleotide under investigation. C) Consensus sequence motif of CTCF-binding sites according to Kim et al.[28]. The height of each letter represents the relative frequency of occurrence of the nucleotide at each position. CTCF motifs 1-4 within the NNAT regulatory region with scores for similarity to the consensus, as determined by the JASPAR software (jaspar.genereg.net). CpGs under investigation are highlighted in black squares. doi:10.1371/journal.pone.0067605.g002 [/fig]
[fig] Figure 3: Mean DNA methylation levels at the A) H19DMR, B) the distal NNAT-CTCF binding site, and C) the NNAT promoter in normal kidney and tumor tissues. Red horizontal lines either depict two standard deviations above or below the mean of the normal controls (A) or the median DNA methylation values (B+C). doi:10.1371/journal.pone.0067605.g003 [/fig]
[fig] Figure 4: Mean DNA methylation levels delineated for each CpG at A) and D) the H19DMR, B) and E) the distal NNAT-CTCF binding site, C) and F) the NNAT promoter, and G) the BLCAP promoter in normal kidney (red squares) and tumor tissues (blue diamonds). A) to C) depict the DNA methylation levels for all tumor and normal kidney cases, whereas D) to F) depict only tumors with .10-fold mean expression of normal kidney for D) IGF2 (n = 22) and E) to F) NNAT (n = 33). Error bars represent the standard deviation of all tumor or kidney sample measurements. doi:10.1371/journal.pone.0067605.g004 [/fig]
[fig] Figure 5: Association of gene expression and mean methylation of A) IGF2 and H19DMR, B) H19 and H19DMR, C) NNAT and NNAT-CTCF, D) NNAT and NNAT promoter, and E) BLCAP_v2a and NNAT-CTCF using Spearman's rank correlation. The data from the tumor and normal kidney cases are depicted as black and orange diamonds, respectively. F) Model of the NNAT/BLCAP locus based on our data depicting suppression of BLCAP_v2a and NNAT in normal kidney due to methylated CpGs in the NNAT-CTCF site and NNAT promoter. Consequently, BLCAP_v2a and NNAT upregulation in Wilms tumors is caused by the demethylation of both regulatory regions. The BLCAP gene is expressed in both tumors and normal kidney due to unmethylted CpGs in its promoter. doi:10.1371/journal.pone.0067605.g005 [/fig]
|
Phosphoproteomic analysis of neoadjuvant breast cancer suggests that increased sensitivity to paclitaxel is driven by CDK4 and filamin A
Precision oncology research is challenging outside the contexts of oncogenic addiction and/or targeted therapies. We previously showed that phosphoproteomics is a powerful approach to reveal patient subsets of interest characterized by the activity of a few kinases where the underlying genomics is complex. Here, we conduct a phosphoproteomic screening of samples from HER2-negative female breast cancer receiving neoadjuvant paclitaxel (N = 130), aiming to find candidate biomarkers of paclitaxel sensitivity. Filtering 11 candidate biomarkers through 2 independent patient sets (N = 218) allowed the identification of a subgroup of patients characterized by high levels of CDK4 and filamin-A who had a 90% chance of achieving a pCR in response to paclitaxel. Mechanistically, CDK4 regulates filamin-A transcription, which in turn forms a complex with tubulin and CLIP-170, which elicits increased binding of paclitaxel to microtubules, microtubule acetylation and stabilization, and mitotic catastrophe. Thus, phosphoproteomics allows the identification of explainable factors for predicting response to paclitaxel. Patient heterogeneity represents a major problem in clinical oncology therapeutics. Regarding breast cancer, massive efforts by international consortia have revealed a virtually unique genomic landscape for each patient 1-6 . In addition, with the exception of the HER2 subtype, and compared to other tumour types, most patients lack a clear oncogenicaddiction driver, which suggests that most HER2-negative breast cancers are the result of the accumulation of several non-sufficient, non-necessary oncogenic alterations 1-9 . This fact complicates the precision medicine framework of the biomarker-driven administration of targeted (or nontargeted) therapies in breast cancer. Taking advantage of phosphoproteomics, however, we have recently demonstrated that considerably heterogeneous genomic landscapes converge into a discrete number of signalling aberrations, regardless of the presence or absence of mutations in the genes encoding for the aberrant signalling nodes 10 . A kinome-based taxonomy sorting ∼200 TNBCs according to the hyperactivation of 6 kinases added considerable parsimony into the classification of this disease compared to a genomic-centred classification and allowed the identification of powerful therapeutic targeted doublets 10 . This concept is currently being tested in a prospective clinical trial (NCT04494958), suggesting that phosphoproteomic profiling could solve some limitations of gene-centric approaches on which most precision oncology applications rely.
Precision oncology research is challenging outside the contexts of oncogenic addiction and/or targeted therapies. We previously showed that phosphoproteomics is a powerful approach to reveal patient subsets of interest characterized by the activity of a few kinases where the underlying genomics is complex. Here, we conduct a phosphoproteomic screening of samples from HER2-negative female breast cancer receiving neoadjuvant paclitaxel (N = 130), aiming to find candidate biomarkers of paclitaxel sensitivity. Filtering 11 candidate biomarkers through 2 independent patient sets (N = 218) allowed the identification of a subgroup of patients characterized by high levels of CDK4 and filamin-A who had a 90% chance of achieving a pCR in response to paclitaxel. Mechanistically, CDK4 regulates filamin-A transcription, which in turn forms a complex with tubulin and CLIP-170, which elicits increased binding of paclitaxel to microtubules, microtubule acetylation and stabilization, and mitotic catastrophe. Thus, phosphoproteomics allows the identification of explainable factors for predicting response to paclitaxel. Patient heterogeneity represents a major problem in clinical oncology therapeutics. Regarding breast cancer, massive efforts by international consortia have revealed a virtually unique genomic landscape for each patient [bib_ref] The clonal and mutational evolution spectrum of primary triple-negative breast cancers, Shah [/bib_ref] [bib_ref] The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups, Curtis [/bib_ref] [bib_ref] Mutational processes molding the genomes of 21 breast cancers, Nik-Zainal [/bib_ref] [bib_ref] The landscape of cancer genes and mutational processes in breast cancer, Stephens [/bib_ref] [bib_ref] Dynamics of breast-cancer relapse reveal laterecurring ER-positive genomic subgroups, Rueda [/bib_ref] [bib_ref] Genomic characterization of metastatic breast cancers, Bertucci [/bib_ref]. In addition, with the exception of the HER2 subtype, and compared to other tumour types, most patients lack a clear oncogenicaddiction driver, which suggests that most HER2-negative breast cancers are the result of the accumulation of several non-sufficient, non-necessary oncogenic alterations [bib_ref] The clonal and mutational evolution spectrum of primary triple-negative breast cancers, Shah [/bib_ref] [bib_ref] The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups, Curtis [/bib_ref] [bib_ref] Mutational processes molding the genomes of 21 breast cancers, Nik-Zainal [/bib_ref] [bib_ref] The landscape of cancer genes and mutational processes in breast cancer, Stephens [/bib_ref] [bib_ref] Dynamics of breast-cancer relapse reveal laterecurring ER-positive genomic subgroups, Rueda [/bib_ref] [bib_ref] Genomic characterization of metastatic breast cancers, Bertucci [/bib_ref] [bib_ref] Mutational landscape and significance across 12 major cancer types, Kandoth [/bib_ref] [bib_ref] The evolutionary history of 2658 cancers, Gerstung [/bib_ref] [bib_ref] Recurrent and functional regulatory mutations in breast cancer, Rheinbay [/bib_ref]. This fact complicates the precision medicine framework of the biomarker-driven administration of targeted (or nontargeted) therapies in breast cancer. Taking advantage of phosphoproteomics, however, we have recently demonstrated that considerably heterogeneous genomic landscapes converge into a discrete number of signalling aberrations, regardless of the presence or absence of mutations in the genes encoding for the aberrant signalling nodes [bib_ref] In vivo phosphoproteomics reveals kinase activity profiles that predict treatment outcome in..., Zagorac [/bib_ref]. A kinome-based taxonomy sorting ∼200 TNBCs according to the hyperactivation of 6 kinases added considerable parsimony into the classification of this disease compared to a genomic-centred classification and allowed the identification of powerful therapeutic targeted doublets [bib_ref] In vivo phosphoproteomics reveals kinase activity profiles that predict treatment outcome in..., Zagorac [/bib_ref]. This concept is currently being tested in a prospective clinical trial (NCT04494958), suggesting that phosphoproteomic profiling could solve some limitations of gene-centric approaches on which most precision oncology applications rely.
More than 90% of breast cancers are diagnosed in a localized stage. In this stage, achieving a pathologic complete response (pCR) to neoadjuvant therapy is the most accurate factor to predict disease-free survival in the long term [bib_ref] Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy, Symmans [/bib_ref] [bib_ref] Optimal management for residual disease following neoadjuvant systemic therapy, Foldi [/bib_ref]. Over the last 30 years, the agents that have led to the greatest increases in pCR rates in HER2-negative breast cancer are the chemotherapeutic agents paclitaxel [for hormone-positive and triple-negative breast cancer (TNBC)] [bib_ref] Randomized parallel study of doxorubicin plus paclitaxel and doxorubicin plus cyclophosphamide as..., Dieras [/bib_ref] and carboplatin (for TNBC). Compared to targeted therapies and immuno-oncology drugs, research on predictive factors or resistance mechanisms against traditional cytotoxics is less active. Increasing data suggest that interrogating the proteome represents an exceptional resource for tackling biological processes or undertaking diagnostic, biomarker or drug discovery efforts [bib_ref] Proteogenomic landscape of breast cancer tumorigenesis and targeted therapy, Krug [/bib_ref] [bib_ref] Proteogenomics connects somatic mutations to signalling in breast cancer, Mertins [/bib_ref] [bib_ref] Proteogenomic characterization reveals therapeutic vulnerabilities in lung adenocarcinoma, Gillette [/bib_ref] [bib_ref] Proteogenomic and metabolomic characterization of human glioblastoma, Wang [/bib_ref]. We hypothesized that an interrogation of the phosphoproteome might inform about mechanistic traits not capturable in genecentric layers [bib_ref] Landscape of somatic mutations in 560 breast cancer whole-genome sequences, Nik-Zainal [/bib_ref] [bib_ref] In vivo genetic screens of patient-derived tumors revealed unexpected frailty of the..., Bossi [/bib_ref] , aiming to understand the basic kinome landscapes underlying the response to the cytotoxic agent paclitaxel in HER2negative breast cancer. Paclitaxel is one of the most widely used agents against breast cancer, and, in contrast to the case for most targeted therapies, we currently lack specific predictive factors. We expected this approach to lead to the definition of predictive factors and an understanding of the mechanisms explaining paclitaxel sensitivity.
To that end, we conducted a phosphoproteomic screening of samples obtained from a clinical trial that compared paclitaxel against paclitaxel plus the antiangiogenic agent nintedanib in early HER2negative breast cancer [bib_ref] 18F-fluoromisonidazole P. E. T. and activity of neoadjuvant nintedanib in early HER2-negative..., Quintela-Fandino [/bib_ref]. The screening led us to 11 candidate biomarkers (kinases and phosphorylated proteins). To transfer these results to potential biomarkers assessable with routine clinical tools, the candidate biomarkers were then tested in two independent early breast cancer datasets by immunohistochemistry. For two of them (filamin A and CDK4), the predictive association with the response to paclitaxel in TNBC was preserved. Subsequent experimental work allowed us to understand the mechanism of sensitization, which was driven by a functional axis linking CDK4, filamin A and CLIP-170; this mechanism enhanced sensitivity to paclitaxel but not to other chemotherapeutics. Thus, our approach allowed us to find predictive factors specific for the response to paclitaxel in HER2-negative breast cancer.
# Results
## Discovery patient set for phosphoproteomic analysis
Samples were obtained from the clinical trial NCT01484080, which enrolled 130 early HER2-negative breast cancer patients (29 TNBC, 101 hormone-positive). Patients were treated with single-agent weekly paclitaxel (Standard Arm) or single-agent oral nintedanib for 14 days followed by nintedanib plus weekly paclitaxel (Experimental Arm). The endpoint of the trial was the rate of achieving a pathologic complete response (pCR), determined in surgery and assessed by the Symmans and Pusztai residual cancer burden (RCB) method, where a pCR is equivalent to RCB = 0 [this classification includes both patients with no evidence of residual tumour in the breast or axillary nodes and patients with only residual noninvasive in situ carcinoma in the breast and no tumour in the axillary nodes after neoadjuvant treatment [bib_ref] Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy, Symmans [/bib_ref] ]. [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref] shows the basic trial design and the number of harvested and valid samples (set at a minimum of 200 micrograms of purified protein isolated from a macrodissected tumour sample). Full trial results and patients' clinical characteristics are described elsewhere [bib_ref] 18F-fluoromisonidazole P. E. T. and activity of neoadjuvant nintedanib in early HER2-negative..., Quintela-Fandino [/bib_ref]. Although in the NCT01484080 trial patients were randomized 1:1, and thus the main clinical and pathologic characteristics of the patients were well balanced among both treatment arms, the cohort of patients with valid samples was slightly imbalanced (Supplementary : valid tumour samples from patients who received paclitaxel monotherapy (N = 39) were somehow smaller, with fewer involved axillary nodes, and of lower grade and replicative fraction than those from patients who received the combination treatment (N = 46); however, these differences did not reach statistical significance. The flow chart in [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref] depicts the steps followed and the samples used for biomarker discovery, confirmation, and experimental validation. In this study, we will refer exclusively to the baseline samples and their relationship with the response to paclitaxel; the effects of nintedanib on the phosphoproteome and their relationship with response or resistance against this agent will be reported elsewhere. These samples produced approximately 3 million spectra, of which 3834 unique phosphopeptides mapping to at least 1352 distinct phosphoproteins were identified (Supplementary lists the proteomic data from the whole trial and by treatment arm). Matching the predominance of serinephosphorylation sites in the proteome, we captured a predominance of serine over threonine phosphosites and a minority of tyrosine sites (Supplementary .
To search for potential differences in sample handling and/or preservation procedures across the different hospitals in which the samples were harvested that could result in signalling alterations (e.g., signalling changes in stress kinase pathways caused by tumour ischaemia [bib_ref] Ischemia in tumors induces early and sustained phosphorylation changes in stress kinase..., Mertins [/bib_ref] , hierarchical clustering was performed with the phosphopeptide intensity data matrix [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref]. Hierarchical clustering did not show that the samples seemed to be significantly grouped by the hospital in which they were harvested [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref]. Given the imbalance in certain patient characteristics when they were classified by treatment type , it was not surprising to observe a clustering of the samples by treatment arm [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref]. To further explore these clusters, we studied sample clustering by relying on other methodologies: consensus clustering 23 (Supplementary [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref] and the pvclust algorithm [bib_ref] Pvclust: an R package for assessing the uncertainty in hierarchical clustering, Suzuki [/bib_ref] [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref] ; however, we did not find a robustness of the association of the clusters with the trial arm [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref].
## Kinase and phosphopeptide enrichment among responders and nonresponders to paclitaxel
The functional impact caused by phosphorylation is unknown for most of the captured phosphorylation sites in high-throughput phosphoproteomic experiments. The advantage of phosphoproteomics over other -omic techniques, however, is that the phosphorylation status of all the encoded proteins is driven by a relatively low number of kinases (~500). We recently described an approach termed kinase set enrichment analysis (KSEA) that allows the mapping of phosphorylated peptides back to the kinase or kinases that can phosphorylate them: based on the affinity of the consensus sequences for each kinase, the main kinases that explain the phosphorylation status of a complex peptide mixture can be narrowed down [bib_ref] In vivo phosphoproteomics reveals kinase activity profiles that predict treatment outcome in..., Zagorac [/bib_ref]. By applying this technique and according to the amount of each detected phosphopeptide among two complex samples, the statistically significantly differentially functioning (hyper-or hypo-) kinases that account for each sample phosphoprofile can be determined. When applied to large sample sets (i.e., responders or nonresponders), we can infer which are the main kinases implicated in causing the differences between the sample sets' phosphoprofiles.
We compared the phosphoprofiles of the samples from patients who achieved a complete response to neoadjuvant therapy (pCR or RCB = 0; herein responders) versus those who did not (RCB > 0; herein nonresponders) [bib_ref] Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy, Symmans [/bib_ref] [bib_ref] Residual cancer burden after neoadjuvant chemotherapy and long-term survival outcomes in breast..., Yau [/bib_ref]. This comparison was run in a data matrix comprising 2151 phosphopeptides that mapped to 1027 unique proteins. The comparison in the Standard Arm (Arm B, paclitaxel only) suggested that the activity of 5 kinases was enriched in the samples from responder patients compared to the samples from nonresponders (P70S6K, CDK4, PKC, AMPK1/2 and CaMK-IV; the KSEA plots are shown in [fig_ref] Figure 2 |: Kinases driving the phosphoprofiles of responders to paclitaxel [/fig_ref]. When the same comparison was run considering all the trial samples together (i.e., mixing patients who received only paclitaxel and patients who received paclitaxel plus nintedanib; 2757/1252 phosphopeptides/unique proteins), P70S6K, CDK4 and AMPK1/2 activity enrichment-albeit present-lost statistical significance [fig_ref] Figure 2 |: Kinases driving the phosphoprofiles of responders to paclitaxel [/fig_ref]. A new kinase (HMG-CoA reductase kinase) was found to be enriched in responder samples [fig_ref] Figure 2 |: Kinases driving the phosphoprofiles of responders to paclitaxel [/fig_ref]. Finally, when responder and nonresponder samples from the paclitaxel plus nintedanib arm were compared, we did not find any significantly enriched kinase (2594/1212 phosphopeptides/unique proteins). The fact that the KSEAs lost significance and displayed reduced enrichment scores as we increased the proportion of samples from patients who received combination therapy relative to those who received monotherapy in the analysis suggests a certain specificity of the enrichment of the activity of P70S6K, CDK4, PKC, AMPK1/2 and CaMK-IV in responders to paclitaxel monotherapy (i.e., the enrichment score was lower when samples from both trial arms were mixed than when only samples from the monotherapy arm were evaluated; furthermore, when we included only samples from the combination arm, no enrichment was observed). Of note, in addition to the indirect detection of P70S6K, CDK4, PKC, AMPK1/2 and CaMK-IV activity enrichment by KSEA, more phosphorylated peptides mapping to the regulatory regions of these kinases were detected in samples from responders to paclitaxel than in those from nonresponders [fig_ref] Figure 2 |: Kinases driving the phosphoprofiles of responders to paclitaxel [/fig_ref]. Since the detected phosphorylation sites of these kinases are known to be involved in kinase activation, this finding further supports the involvement of those kinases in the responder phenotype.
Regarding potential resistance biomarkers (i.e., kinases enriched in samples from nonresponders compared to samples from responders), the data are shown in [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref]. Little overlap was observed when the analysis was run comparing responders and nonresponders in the whole trial (enrichment of SRC, JAK2, CDK5 and CK1; [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref] , in the experimental arm only (B-adrenergic receptor kinase, PIM1 and PDK1; [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref] , or in the standard arm only (RAF1, CK1 and B-adrenergic receptor kinase; Supplementary [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref].
Key information about potential predictive markers can also be obtained by examining the significantly up-/downregulated phosphoproteins in responders or nonresponders on top of the kinases predicted by KSEA, since phosphorylated proteins can mediate the observed effect from a mechanistic point of view. [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref] shows the volcano plots of the phosphopeptides enriched in responders or nonresponders. [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref] Sixty-two and sixty-one patients allocated to the Experimental and Standard arms, respectively, consented to and underwent a baseline biopsy (out of 130 patients). Patients allocated to the experimental arm underwent a 2-week course of singleagent nintedanib (150 mg orally twice a day), and then a second tumour sample was harvested (N = 58 patients consented to this second biopsy) prior to undergoing 12 weekly courses of paclitaxel combined with nintedanib. Those allocated to the standard arm immediately started weekly paclitaxel without the 14-day delay and did not have a second tumour sample harvested. The endpoint (tumour response according to the RCB score) was determined at the time of surgery, and patients then received standard treatment according to the referring physician's choice (radiation or hormonal therapy or further chemotherapy if indicated). B Flow chart depicting the study steps: biomarker discovery, biomarker confirmation and experimental validation. C Unsupervised hierarchical clustering. A phosphopeptide intensity data matrix was used for clustering analysis. Patient IDs are listed horizontally. The two following rows indicate, for each sample, whether they were allocated to the standard or experimental arm and the study site origin.
(Arm B), whereas Figs. 3B, C show the regulated peptides combining both treatment arms and Arm A only, respectively. In all volcano plots, the central dot cloud represents those peptides that were present both in responders and nonresponders and experienced some degree of regulation in one or another group of patients. The two additional lateral dot clouds represent phosphorylation sites that were undetectable in one group (responders or nonresponders) but were present in the other, suggesting that dramatic regulatory events were involved in the response or lack thereof. Phosphopeptide IDs with >2 4 (16-fold) regulation, together with their P value and FDR, are listed in Supplementary Data 1. Many peptide IDs map to proteins with little or uncharacterized functional significance; however, a considerable number of peptides upregulated in responders to paclitaxel mapped to proteins implicated in cytoskeletal polymerization and rearrangement, such as vimentin, laminin, plectin, tensin, filamin, and Rab7 (Group 1 and Group 2 phosphopeptide cloud in [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref] -C; Supplementary Data 1). As we show below, this fact is of key relevance, since paclitaxel exerts its antitumour effects by stabilizing and thereby blocking microtubule polymerization/depolymerization dynamics [bib_ref] Structure of tubulin at 6.5 A and location of the taxol-binding site, Nogales [/bib_ref] [bib_ref] Microtubules as a target for anticancer drugs, Jordan [/bib_ref] [bib_ref] How Taxol stabilises microtubule structure, Amos [/bib_ref]. No significantly regulated peptides were observed in tumours resistant to paclitaxel (no dots above the significance level in the far-left clouds are observed in [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref]. For subsequent experiments, only those phosphopeptides with FDR < 0.25 were studied.
Elevated CDK4 and filamin A levels narrow down a subset of patients with increased sensitivity to paclitaxel Mass spectrometry is not yet an over-the-counter technology in cancer hospitals. In addition, the phosphoscreening data (KSEAs or volcanoplot hits) are the result of a training set with relaxed FDR values to enhance the number of biomarker candidates, a strategy that was proven successful previously [bib_ref] In vivo phosphoproteomics reveals kinase activity profiles that predict treatment outcome in..., Zagorac [/bib_ref]. Because of these two reasons and aiming for the clinical applicability of our results, we sought to determine these potential hits by relying on immunohistochemistry (IHC) in additional patient series, a widely available technique in pathology diagnostic departments in hospitals for detecting protein and protein phosphorylation levels. IHC (as opposed to mass spectrometry) is optimized for formalin-fixed, paraffin-embedded samples, which in turn are the most common tissue vehicle in clinical routine.
To translate the phosphoscreening results (KSEAs and volcano plots of the proteins listed in Supplementary Data 1) to measurable data by immunohistochemistry, we followed our previously described mass spectrometry-to-immunohistochemistry approach [bib_ref] In vivo phosphoproteomics reveals kinase activity profiles that predict treatment outcome in..., Zagorac [/bib_ref]. Following the algorithm depicted in [fig_ref] Figure 4 |: Independent testing of biomarkers of response to paclitaxel [/fig_ref] , the approach yielded 11 potential antibodies for biomarkers of sensitivity to paclitaxel: p-P70S6K (Thr 389 ), CDK4, filamin-A, HMG-CoA reductase, p-vimentin (Ser 56 ), p-AMPK1/2 (Thr 172 ), p-Pan-PKC (Thr 497 ), p-CaMKIV (Thr 196/200 ), p-filamin A (Ser 2512 ), p-YAP1 (Ser 127 ) and plectin. Antibody setup and control stainings are shown in [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref] -K.
To account for the relaxed FDR boundaries of the training set, we applied a biomarker selection filter consisting of a two-step process in independent sample sets. The candidate biomarkers were first tested in a patient set of 117 high-risk early breast cancer patients of the three subtypes who were treated with paclitaxel-based neoadjuvant chemotherapy (Set 1; clinical and pathological characteristics shown in . In this first step, we aimed to detect which of the 11 potential biomarkers maintained an association with the response to paclitaxel after the mass-spectrometry-to-immunohistochemistry translation step, in a population of mixed breast cancer subtypes. shows that when the analysis was repeated combining the samples from the paclitaxel-only arm with those from the combination arm, virtually no significant enrichment was observed. For all KSEA plots, each vertical black line represents a phosphopeptide that can be phosphorylated by the depicted kinase that was detected in the samples from one of the compared conditions in the KSEA. High in responders and high in nonresponders refers to the increased abundance of phosphopeptides in the baseline samples of patients who achieved a pCR or patients who did not achieve a pCR, respectively. NES (normalized enrichment score) and FDR (false-discovery rate) values are depicted for each KSEA. A relaxed FDR boundary (up to 0.20) was allowed to ensure as little information loss as possible in the mass spectrometry-to-immunohistochemistry translation step, since the biomarker candidates underwent a subsequent 2-patient series filter.
We next filtered those that held significance through a second patient set of exclusively TNBC patients (N = 101; Set 2: clinical and pathologic characteristics shown in undergoing neoadjuvant paclitaxel-based treatment. By doing so, we expected to have available a small set of robust biomarkers for the breast cancer subtype where paclitaxel chemotherapy is more relevant. The 117 samples from Set 1 were stained with the 11 antibodies against the potential biomarkers, and an H-score was calculated. The H-score was divided into quartiles, and the association with response was studied by determining the probability of obtaining a pCR (RCB = 0) for the patients in the top quartile versus the remaining patients. A staining example of two patients from the upper and lower quartiles of filamin A is shown in [fig_ref] Figure 4 |: Independent testing of biomarkers of response to paclitaxel [/fig_ref]. The upper quartile cut-off value for each staining is depicted in Supplementary .
We found that the following biomarker candidates were associated with a pCR to paclitaxel-based neoadjuvant therapy in Set 1: p-P70S6K (Thr 389 ) (2.89-fold higher chance or achieving a pCR for patients in the upper H-score quartile versus patients with H-score in quartiles 2 to 4; P = 0.037), CDK4 (2.85-fold; P = 0.048), filamin A (3.28fold; P = 0.062), HMG-CoA reductase (4.00-fold; P = 0.064) and p-Vim (Ser 56 ) (2.93-fold; P = 0.047). The pCR rate in the whole Set 1 cohort, or divided by breast cancer subtype (luminal, HER2 or TNBC) according to the value of p-P70S6K (Thr 389 ), CDK4, filamin A, HMG-CoA reductase and p-Vim (Ser 56 ), is shown in [fig_ref] Figure 4 |: Independent testing of biomarkers of response to paclitaxel [/fig_ref]. Conversely, despite showing a potential association in the KSEA or volcano plot analysis, when translated to immunohistochemical staining, the following biomarker candidates did not show an association with a pCR in Set 1: p-AMPK (Thr 172 ) (2.07-fold higher chance or achieving a pCR for patients in the upper H-score quartile versus patients with H-score in quartiles 2 to 4; P = 0. [bib_ref] Residual cancer burden after neoadjuvant chemotherapy and long-term survival outcomes in breast..., Yau [/bib_ref] We next proceeded to the final biomarker filtering in Set 2. The upper quartile cut-off value for each staining in Set 2 is shown in Supplementary . Patients in the upper quartile of CDK4 displayed a 2.71-fold (P = 0.04) higher chance of achieving a pCR in response to neoadjuvant chemotherapy than patients in Q2-Q4. Similarly, patients with upper quartile filamin A staining exhibited a 3.36fold greater probability of achieving a pCR (P = 0.039). Conversely, no statistically significant associations were confirmed for p-Vim (Ser 56 ) (1.67-fold higher chance of achieving a pCR for patients in the upper H-score quartile versus patients with H-score in quartiles 2 to 4, P = 0.31), p-P70S6K (Thr 389 ) (1.1-fold; P = 0.90), or HMG-CoA reductase (0.4-fold, P = 0.11). The percentages of patients achieving a pCR according to their staining levels of CDK4, filamin A, p-Vim (Ser 56 ) and p-P70S6K (Thr 389 ) are shown in [fig_ref] Figure 4 |: Independent testing of biomarkers of response to paclitaxel [/fig_ref]. Ninety percent of patients with both CDK4 and filamin A staining levels in the upper quartile achieved a pCR [fig_ref] Figure 4 |: Independent testing of biomarkers of response to paclitaxel [/fig_ref]. Although this study is retrospective, its results compare favourably with commonly reported pCR rates in unselected TNBCs achieved with polychemotherapy regimens alone (40-50%) [bib_ref] Weekly carboplatin plus neoadjuvant anthracycline-taxane-based regimen in early triple-negative breast cancer: a..., Fontaine [/bib_ref] or in combination with immunotherapy (55-65%) [bib_ref] Neoadjuvant atezolizumab in combination with sequential nab-paclitaxel and anthracycline-based chemotherapy versus placebo..., Mittendorf [/bib_ref] [bib_ref] Pembrolizumab for early triple-negative breast cancer, Schmid [/bib_ref] [bib_ref] Pembrolizumab plus chemotherapy as neoadjuvant treatment of high-risk, early-stage triple-negative breast cancer:..., Schmid [/bib_ref]. Finally, it is worth mentioning that patients did not show significant staining heterogeneity regarding tumour cells/stromal or nuclear/cytoplasmic CDK4 or filamin A staining .
CDK4 and filamin A lead to increased microtubule stability and sensitivity to paclitaxel through CLIP-170
The notion that tumours with high replicative fractions are more sensitive to classic cytotoxics is commonly assumed in clinical oncology. However, a specific mechanistic explanation for each cytotoxic agent is lacking. Thus, we sought to understand how high levels of CDK4 and filamin A specifically sensitize tumour cells to paclitaxel. CDK4 is a well-known regulator of the mitotic process, as it phosphorylates and inactivates RB1, which in turn liberates the E2F transcription factor during the G1/S transition [bib_ref] Driving the cell cycle to cancer, Malumbres [/bib_ref]. Filamin A is a scaffolding protein that crosslinks actin filaments into networks and participates in anchoring membrane proteins to the actin cytoskeleton and in transducing signals from the tumour microenvironment [bib_ref] Filamins as integrators of cell mechanics and signalling, Stossel [/bib_ref]. How these functional features sensitize cells to paclitaxel treatment is currently unknown.
We generated stable variants of the TNBC cell line MDA-MB-231 that constitutively expressed elevated levels of CDK4 (MDA-MB-231 CDK4; [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref]. Compared to the parental MDA-MB-231 cells, sensitivity to paclitaxel was increased 2.7-fold (P = 0.0012), as evidenced by the colony assays and IC50 calculations shown in [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref] , in line with the data observed in patients. A previous study in the field of the regulation of cell motility and invasion described a physical interaction between cyclin D1 and filamin A; in addition, a positive correlation between CDK4 and filamin A (Ser 2152 ) phosphorylation levels was found, but direct phosphorylation was not demonstrated [bib_ref] Cyclin D1/cyclin-dependent kinase 4 interacts with filamin A and affects the migration..., Zhong [/bib_ref]. We were unable to show the coimmunoprecipitation of CDK4 and filamin A in either parental MDA-MB-231 or MDA-MB-231 CDK4 cells (Supplementary [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref]. Confocal imaging did not demonstrate the colocalization of CDK4 and filamin A in either cell line [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref]. In addition, an in vitro kinase assay did not show kinase activity of CDK4 over filamin A [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref]. The levels of CDK4 and filamin A, however, demonstrated a positive and statistically significant correlation in patients from Set 2 [fig_ref] 8 M: Urea, 100 mM Tris-HCl pH = 8 [/fig_ref]. Similarly, MDA-MB-231 CDK4 cells showed considerably higher levels of filamin A than parental MDA-MB-231 cells [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref]. Real-time PCR revealed a transcriptional mechanism of filamin A upregulation in MDA-MB-231 CDK4 cells [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref]. This observation is consistent with the fact that while active, RB1 sequesters and represses the E2F1 transcription factor, which is liberated when CDK4 phosphorylates and inactivates RB1. E2F1 has~3000 transcriptional targets, among which is filamin A, according to the last update of the ENCODE project (https://maayanlab.cloud/Harmonizome/gene_set/ E2F1/ENCODE+Transcription+Factor+Targets) [bib_ref] The harmonizome: a collection of processed datasets gathered to serve and mine..., Rouillard [/bib_ref]. We also generated stable MDA-MB-231 cells with elevated stable expression of filamin A (MDA-MB-231 FLNA, [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref]. The increased expression of filamin A did not change CDK4 levels [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref]. These cells showed a similar sensitization to that observed for MDA-MB-231 CDK4 (approximately 3-fold to 0.32 nM; P = 0.0004; [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref]. Together with the fact that transient filamin A knockdown in MDA-MB-231 CDK4 cells restored paclitaxel sensitivity to levels similar to those observed in MDA-MB-231 WT cells (IC50 = 0.92 nM; the comparison with the IC50 displayed by MDA-MB-231 CDK cells was statistically significant; P < 0.0001; [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref] , the transcriptional link between CDK4 and filamin A suggests that filamin A is a mediator of the increased sensitivity to paclitaxel in MDA-MB-231 CDK4 cells. High proliferation rates have been related to nonspecific sensitization to cytotoxic agents; interestingly, neither CDK4 nor filamin A overexpression sensitized tumour cells to the other cytotoxics received in the neoadjuvant TNBC setting (anthracyclines or platins; . Together with the limited correlation observed among CDK4, filamin A and the Ki67 replicative fraction in the TNBC patient series , these data suggest that CDK4 and filamin A are specifically involved in sensitization to paclitaxel only.
Paclitaxel exerts its cytotoxic effects by binding to beta-tubulin subunits of assembled microtubules, stabilizing them and interfering with the polymerization/depolymerization equilibrium that allows their extension and dynamics [bib_ref] Structure of tubulin at 6.5 A and location of the taxol-binding site, Nogales [/bib_ref] [bib_ref] Insights into the mechanism of microtubule stabilization by Taxol, Xiao [/bib_ref]. Microtubules in the mitotic spindle have rapid dynamics, and when these dynamics are interrupted, mitosis results in chromosomal mis-segregation, mitotic arrest or catastrophe and the formation of micronuclei, often leading to cell apoptosis [bib_ref] Microtubules as a target for anticancer drugs, Jordan [/bib_ref]. Although the link between CDK4 and elevated filamin A levels seemed clear and increased filamin A levels seemed to be a necessary link between CDK4 and increased sensitivity, the obvious remaining question was how to relate elevated filamin A levels with altered tubulin dynamics that could explain the increased sensitivity to paclitaxel. Filamin A and tubulin are reported to physically interact in response to forces in the process of the transduction of mechanotranscriptional signals by the microtubule network [bib_ref] Regulation of tension-induced mechanotranscriptional signals by the microtubule network in fibroblasts, D'addario [/bib_ref]. We analysed the spatial distribution of filamin A and tubulin and found a colocalization of these two proteins both in the cytoplasm and submembrane compartments . Accordingly, we hypothesized that changes in filamin A levels could lead to quantitative or qualitative changes in the protein complexes normally formed with tubulin or filamin A. In an attempt to identify regulated proteins within the filamin-tubulin complexes that could be implicated in microtubule dynamics, we performed a pull-down analysis of tubulin and filamin A, quantifying and comparing the complexes between MDA-MB-231 WT and MDA-MB-231 CDK4 cells. The plots shown in display the ratios of proteins bound to anti-filamin A antibody or isotype control in MDA-MB-231 WT (left panel) or MDA-MB-231 CDK4 cells (right panel). A migration in either the ratio or the intensity indicates potential regulation. shows the results of the tubulin pull-down comparison between MDA-MB-231 WT (left) and MDA-MB-231 CDK4 cells (right). The protein IDs bound to filamin A or tubulin (and their spectral intensity) in the two cell lines are gathered in Supplementary Data 2 and 3. Cytoplasmic linker protein 170, or CLIP-170, is a protein that belongs to the groups of plus-end tracking proteins, or +TIPs. Microtubules are nucleated out of the microtubule organizing centre (MTOC), where their minus end is anchored; their plus ends grow out of the MTOC [bib_ref] The minus end in sight, Dammermann [/bib_ref]. The plus ends are highly dynamic, and +TIPS are a class of microtubule-binding proteins that accumulate at and track with growing tubules, regulating their dynamics and interactions [bib_ref] Microtubule plus end: a hub of cellular activities, Lansbergen [/bib_ref] , of which CLIP-170 was the first discovered member [bib_ref] CLIP-170 links endocytic vesicles to microtubules, Pierre [/bib_ref]. It has been shown that enhanced CLIP-170 accumulation leads to reduced and slower tubulin polymerization [bib_ref] AMPK controls the speed of microtubule polymerization and directional cell migration through..., Nakano [/bib_ref]. In addition, others have found that CLIP-170 promotes paclitaxel binding to microtubules [bib_ref] Microtubule-binding protein CLIP-170 is a mediator of paclitaxel sensitivity, Sun [/bib_ref]. Quantitative mass spectrometry showed that the amount of CLIP-170 bound to tubulin and the amount of CLIP-170 bound to filamin A increased by >7-fold and >1000-fold, respectively, when CDK4 was overexpressed (Supplementary Data 2 and 3).
Whether CIP-170 and filamin A have a physical interaction or not is currently unknown. Thus, despite our mass spectrometry data, we sought to confirm this interaction through coimmunoprecipitation. proves that filamin A can bind CLIP-170. Taken together, these data led us to hypothesize that in the presence of increased filamin A, tubulin-filamin A complexes would be enriched in CLIP-170, which would lead to increased paclitaxel sensitivity through a twopronged mechanism: increased paclitaxel accumulation and increased microtubule stabilization .
To confirm this hypothesis, we experimentally tested the two potential mechanisms. shows the amount of paclitaxel bound to microtubules in MDA-MB-231 WT, CDK4 and FLNA variants. It can be appreciated that the CDK4 and FLNA cell lines accumulated more paclitaxel and did so earlier than their WT counterparts. In addition, when filamin A was knocked down in MDA-MB-231 CDK4 cells, the phenotype was reversed: these cell lines accumulated the same amount of paclitaxel and at a similar rate as that observed in WT cells.
Then, we studied the effects on microtubule stability. The status of microtubule dynamics can be assessed by determining several posttranslational modifications. Alpha-tubulin lysine acetylation is a key modification that stabilizes microtubules [bib_ref] Lysine acetylation of cytoskeletal proteins: emergence of an actin code, Latario [/bib_ref]. We measured the levels of [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref]. Both the baseline and paclitaxel-treated levels of acetylated tubulin were increased in the two cell variants, indicating increased microtubule stability as a result of elevated CDK4 and/or filamin A. When filamin A was knocked down in MDA-MB-231 CDK4 cells, we observed a decrease in tubulin acetylation levels [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref]. Interestingly, experimental replicates in which siRNA against filamin A knocked down filamin A with variable efficiency in MDA-MB-231 CDK4 cells showed that tubulin acetylation levels were modified accordingly with filamin A protein levels [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref].
Increased binding of paclitaxel to microtubules combined with a context prone to increased microtubule stabilization should enhance the consequences of tubulin polymerization/depolymerization equilibrium arrest in cell replication, explaining the observed increased sensitivity to paclitaxel. [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref] Taken together, our data suggest that CDK4 increases filamin A levels through a transcriptional mechanism. Increased filamin A, in turn, binds CLIP-170, increasing the amount of CLIP-170 in the tubulinfilamin complexes. This accumulation of CLIP-170 leads to increased microtubule acetylation and stabilization and increased paclitaxel binding to microtubules. These effects, combined, lead to mitotic catastrophe, explaining the increased sensitivity to this drug in tumours with elevated CDK4.
# Discussion
Precision oncology is highly based on gene-centric approaches. Although this perspective has led to considerable advances that have positively impacted clinical care, it falls short of answering certain translational research questions. One scenario of particular difficulty is that in which tumours are not driven by oncogene addiction mutations. In most tumours, each mutation confers only a small fitness advantage, but several of those low-penetrant mutations might cooperate and promote tumour progression [bib_ref] The mini-driver model of polygenic cancer evolution, Castro-Giner [/bib_ref] [bib_ref] Accumulation of driver and passenger mutations during tumor progression, Bozic [/bib_ref]. However, most of those mutations are nonrecurring (i.e., limited to one or a few patients, making the number of potential combinations of mutations potentially immense). In addition, the functional impact of each of those mutations still requires functional characterization. Together, these facts make it very complicated to issue accurate individual predictions in most cases. The most common solution for this situation relies on gene-expression panels, which have been successful in difficult tasks such as predicting benefit or not from hormonal treatment alone in non-HER2-positive breast cancers by grouping cancers on the basis of the expression of a number of genes [bib_ref] Adjuvant chemotherapy guided by a 21-gene expression assay in breast cancer, Sparano [/bib_ref]. However, the functional characterization of the majority of these genes or how they contribute functionally to the final tumour phenotype, depending on whether they are mutated or not, is not always available. Regardless of how accurate these approaches can be, they still lack functional specificity, hampering subsequent developments such as the rational design of therapies for adverse prognosis subgroups.
In the past, we have solved similarly complex questions in translational oncology by means of mass spectrometry-aided phosphoproteomics. The justification for relying on this technique is that different phenotypes (i.e., drug response or resistance) could be achieved by multiple different genomic landscapes, which ultimately collapse into the functional hyper-/hypoactivation of a discrete number of signalling axes, which in turn are the effectors of the tumour genotypes. This approach, for example, allowed us to solve a practical kinase-based taxonomy of TNBC 10 and to solve the problem of acquired resistance to antiangiogenics [bib_ref] Targeting tumor mitochondrial metabolism overcomes resistance to antiangiogenics, Navarro [/bib_ref].
This time, we aimed to understand what drives sensitivity to the most commonly used cytotoxic in breast cancer: paclitaxel. We took advantage of the samples obtained in a clinical trial that compared paclitaxel monotherapy in treatment-naive early breast cancer patients with paclitaxel plus the multikinase inhibitor nintedanib [bib_ref] 18F-fluoromisonidazole P. E. T. and activity of neoadjuvant nintedanib in early HER2-negative..., Quintela-Fandino [/bib_ref]. Our study has several features worth highlighting.
Our study provides a specific predictive factor of the efficacy of paclitaxel. Similar to our previous study [bib_ref] In vivo phosphoproteomics reveals kinase activity profiles that predict treatment outcome in..., Zagorac [/bib_ref] , from a relatively high number of candidate markers yielded from the mass spectrometry data [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref] , only a limited number of them were successfully confirmed by immunohistochemistry in independent patient sets [fig_ref] Figure 4 |: Independent testing of biomarkers of response to paclitaxel [/fig_ref]. Those markers that showed association in Set 1 were further filtered in a second set (Set 2) of a homogenous population of TNBC patients, leaving a final number of 2 markers: CDK4 and filamin A [fig_ref] Figure 4 |: Independent testing of biomarkers of response to paclitaxel [/fig_ref]. Filamin A and CDK4 were highly accurate in predicting a pCR, with patients with upper-quartile staining of both CDK4 and filamin A achieving a 90% pCR in response to paclitaxelbased chemotherapy [fig_ref] Figure 4 |: Independent testing of biomarkers of response to paclitaxel [/fig_ref] , which compares favourably with expected ratios in unselected patients receiving paclitaxel-based combinations [bib_ref] Weekly carboplatin plus neoadjuvant anthracycline-taxane-based regimen in early triple-negative breast cancer: a..., Fontaine [/bib_ref]. These markers could simplify complex treatment schedules now including up to 5 drugs, reserving immunotherapy and other combinations for patients with low levels of CDK4 and filamin A.
Interestingly, according to our data, elevated CDK4 and filamin A levels appear to be specifically linked to the mechanism of action of paclitaxel, adding robustness to their biomarker role. Filamin A upregulation by CDK4 [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref] is followed by increased binding of CLIP-170 to filamin A and tubulin, mediating an increased binding of paclitaxel to microtubules . This results in acetylated microtubules [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref] , mitotic arrest and mitotic catastrophe [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref]. In the absence of filamin A, paclitaxel binding to microtubules decreases along with tubulin acetylation [fig_ref] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization,... [/fig_ref] , reverting the increased sensitivity to paclitaxel [fig_ref] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models [/fig_ref] , which suggests a specific link between CDK4 and filamin A in regulating sensitivity to this drug. The experiments shown in further confirm drug specificity: although high replicative fractions have been nonspecifically associated with sensitization to cytotoxics, the role of CDK4 and filamin A seems specific for paclitaxel, since (1) these biomarkers were discovered in a series of patients who received singleagent paclitaxel [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref] ; (2) CDK4 and filamin A did not sensitize tumour cells to other cytotoxics used in the validation sets (Supplementary ; (3) the findings of mechanistic experiments suggested specific accumulation in their target (microtubules) elicited by filamin-A-mediated CLIP170 binding (Figs. 5-7); and (4) CDK4 and filamin A showed only low and no correlation, respectively, with Ki67 in | Tubulin and filamin A form a complex with CLIP170, which elicits increased binding of paclitaxel to microtubules. A Spatial colocalization of alpha-tubulin and filamin A in MDA-MB-231 cells (Pearson's colocalization coefficient = 0.5, n = 76 cells). Scale bars: 10 µm. B Filamin A pull-down in MDA-MB-231 WT or CDK4 cell lines. For each isolated protein, the X axis represents the Log 2 of the average ratio of the protein isolated in the anti-filamin A pull-down and the protein isolated in the IgG control antibody. The Y axis, conversely, represents the Log 10 of the sum of the average intensities found bound to the anti-filamin-A antibody and the IgG isotype control. C Same as in (B) for tubulin pull-downs. D Coimmunoprecipitation of CLIP-170 and filamin A. Filamin A was immunoprecipitated in whole-cell lysates from the three cell lines: MDA-MB-231 WT, CDK4 and FLNA. Three samples are shown for each cell line: total lysate, immunoprecipitated with anti-filamin A, and immunoprecipitated with isotype IgG control. Samples derive from the same experiment and blots were processed in parallel. Experiment was repeated 3 times with similar results. E Schematic representing the proposed mechanism. On the left-hand side, low CDK4 levels lead to average filamin A expression, which does not enhance the binding of CLIP-170 to microtubules. On the right-hand side, tumour cells with increased CDK4 levels would lead to the overexpression of filamin A. Filamin A would recruit an excess of CLIP-170 to tubulin, which ultimately leads to increased binding of paclitaxel and microtubule acetylation and hyperstabilization. F Paclitaxel-binding experiment. Fluorescently labelled paclitaxel was added to live cultures of MDA-MB-231 WT, CDK4 or FLNA cells. MDA-MB-231 CDK4 cells with filamin A knockdown were added to the experiment as well. The greater the green signal is, the higher the amount of paclitaxel bound to microtubules. It can be appreciated how both CDK4-and filamin A-overexpressing cell lines display both earlier and higher paclitaxel binding. Scale bar: 75 µm. The chart on the right-hand side depicts the signal (in fluorescent surface units) tracing paclitaxel accumulation over the 48-h time course, displaying a clear increase in the two overexpressing transfectants (CDK4 and FLNA) compared to the parental cell line and a reversion of the phenotype by filamin A knockdown in MDA-MB-231 CDK4 cells. Each dot represents mean signal intensity of six independent fields. Source data are provided as a Source data file.
the TNBC series , but both were strongly implicated in the sensitization mechanism. The previous state-of-theart knowledge in the field of predicting sensitivity to chemotherapy in breast cancer is limited to allocating chemotherapy to those patients with a high replicative fraction (i.e., high KI67 staining) or high-risk score in multigene expression panels. This knowledge, while correct in the practical sense [i.e., low-replication breast cancers derive low benefit from cytotoxics [bib_ref] Adjuvant chemotherapy guided by a 21-gene expression assay in breast cancer, Sparano [/bib_ref] [bib_ref] Ki67 index, HER2 status, and prognosis of patients with luminal B breast..., Cheang [/bib_ref] ] is ultimately nonspecific and conceptually incomplete. For example, one broad validation of a geneexpression score (the 21-gene assay) was performed in a patient cohort treated with single-agent tamoxifen. Patients with high scores experienced relapse [bib_ref] A multigene assay to predict recurrence of tamoxifentreated, node-negative breast cancer, Paik [/bib_ref] , and since then, they have been considered candidates for chemotherapy. The interpretation, however, is not that those patients obtained benefit specifically from one or another cytotoxic drug; it simply allows concluding that the benefit from tamoxifen was insufficient and thus they were offered multiagent chemotherapy regimens without a rationale choice of one cytotoxic over another. Understanding the mechanisms behind increased sensitivity in tumours with an increased replication fraction or increased CDK4/ filamin A levels allows, on the one hand, the making of rational treatment allocation decisions (for example, paclitaxel for patients with high CDK4 and filamin A levels and other cytotoxics for other patients), while on the other hand, it allows a more rational search for treatment approaches for resistant phenotypes (in this example, aiming for filamin A or CLIP-170 regulation in high-risk tumours that will require chemotherapy but do not show positive biomarkers of sensitivity).
Our study has several limitations. One potential limitation is the lack of a microdissection of the samples of the discovery set. The contamination of test samples by nontumor tissue is an important limitation in translational cancer studies (e.g., low variant allele frequency mutations can be overlooked if sequenced with low depth). However, phosphorylation events display manyfold regulation from one tissue to another; thus, even in cases where tumour and nontumor tissue are mixed, these post-translational modification events can be detected. Microdissection reagents decrease the amount of protein that can be obtained from tumour samples and interfere with the phosphopurification experimental protocols; since the amount of protein required for phosphoproteomic analyses is approximately 100-fold higher than what it is required for proteomic runs, we chose to proceed with macrodissection. All samples were macrodissected by an expert pathologist and were of >75% tumour purity. Thus, it is unlikely that the lack of microdissection impacted our results for two reasons: first, because highly regulated events such as phosphorylation would still be detected in the case of low tumour purity; second, because all the candidate biomarkers that were detected in the phosphoproteomic runs underwent confirmation in two external patient series. Thus, although contamination by nontumor tissue may have led us to overlook some additional potential biomarkers, it does not affect our conclusions about the involvement of CDK4 and filamin A in sensitizing tumour cells to paclitaxel.
The second limitation also concerns the discovery experiment: the lack of a normalization of phosphopeptide intensities by total protein intensity could be perceived as a shortcoming. Although the native protein can experience regulation from one sample to another, this is always of much smaller magnitude than the changes in phosphorylation. Thus, it is unlikely that many of the detected events with 10 5 −10 7 -fold regulation had not been detected if normalization techniques were used. Regardless, our strategy was to maximize the number of candidate biomarkers found in the proteomic experiment, applying low FDR boundaries for the KSEAs and, deliberately, not normalizing phosphoprotein intensity by total protein intensity (since this would have duplicated the cost and duration of the discovery step): no matter how reliable these candidate markers would have been because of the boundaries set in the discovery experiment, they still would have had to undergo confirmation in external sets and experimental mechanistic validation. Studies that base their conclusions exclusively on proteomic techniques (or other -omic techniques) obviously require more strict boundaries and different experimental processes [bib_ref] Proteogenomic landscape of breast cancer tumorigenesis and targeted therapy, Krug [/bib_ref] [bib_ref] Proteogenomics connects somatic mutations to signalling in breast cancer, Mertins [/bib_ref] [bib_ref] Proteogenomic characterization reveals therapeutic vulnerabilities in lung adenocarcinoma, Gillette [/bib_ref] [bib_ref] Proteogenomic insights into the biology and treatment of HPV-negative head and neck..., Huang [/bib_ref] [bib_ref] Proteogenomic integration reveals therapeutic targets in breast cancer xenografts, Huang [/bib_ref]. However, in a clinical oncology biomarker study, what matters most is whether the biomarkers can be confirmed and actually maintain a mechanistic relationship with the feature under study regardless of the origin of the candidate markers (literature, serendipitous experimental results, or discovery experiments, just to name a few), which we believe we have achieved. We have successfully applied the same strategy in the past [bib_ref] In vivo phosphoproteomics reveals kinase activity profiles that predict treatment outcome in..., Zagorac [/bib_ref] [bib_ref] Targeting tumor mitochondrial metabolism overcomes resistance to antiangiogenics, Navarro [/bib_ref] , and we think that it can be safely stated that this is an acceptable discovery biomarker strategy in clinical proteomics, particularly for difficult-to-solve translational oncology problems.
Regarding the discovery experiment, another point of criticism could be raised, which is the imbalance in certain clinical/pathologic characteristics between the patients who received paclitaxel monotherapy or combination treatment (Supplementary . Correlative studies are often performed with samples originating in large randomized trials, and usually, it is uncommon to observe 100% success in sample retrieval or sample validity. Thus, the question of whether the patient characteristics of the patient subcohort constituted by those with valid samples resemble those of the full cohort is normally addressed in such studies to be able to conclude whether the obtained results apply to the full cohort. We cannot confidently conclude that the nonsignificant differences in tumour size, grade, nodal status or Ki67 resulted in meaningful biases in the discovery set experiment, since although hierarchical clustering showed sample clustering by treatment arm [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref] , other clustering techniques [fig_ref] Figure 1 |: Clinical trial tumour samples [/fig_ref] did not. Clustering algorithms have limitations [bib_ref] Critical limitations of consensus clustering in class discovery, Senbabaoglu [/bib_ref] and are subjective since there are many choices of distances, linkages and numbers of groups that can affect the results. The differences obtained when the three clustering methods were applied may stem from the fact that the imbalances between treatment arms were not statistically significant or from the cluster methodologies per se. Nevertheless, potential biases in the discovery patient set and disagreement between clustering results when different methodologies are used would be relevant if we were aiming to establish conclusions just on the basis of the screening, attempting to extract conclusions about the biomarker role of filamin and CDK4 just for the NCT01484080 trial based only on this screening, or trying to classify patient subgroups on the basis of the discovery set. An imbalanced discovery set can impact the percentage of hits that are confirmed out of the screening candidates (introducing noise, i.e., yielding many hits that are not subsequently confirmed; in our study, only 2 out of 11 initial candidates were so). However, since our objectives were to determine potential biomarkers for the general TNBC population treated with paclitaxel and understand them from the biological point of view, filtering the candidate biomarkers through two external patient cohorts and preclinical experimentation can serve those purposes. Thus, we think it is unlikely that such an imbalance affected the results and conclusions of this study. However, the imbalances between the two arms, and above all, the fact that one arm studied the response to a single drug and the other arm studied the response to a combination of drugs with two different mechanisms of action (microtubule binderpaclitaxelor kinase inhibitornintedanib), may explain why there are differences between global appearance of the volcano plots shown in [fig_ref] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel [/fig_ref] and C (different lower bounds for nonadjusted P values in responders and nonresponders, or apparent flatlines in the P values, which are the results of changes only in the 3rd to 5th decimal place in the P values of the nonsignificantly regulated peptides, thus not being truly flat). Examples of apparent flatlines in the nonsignificant portion of the data are abundant in the literature reporting phosphoproteomic screenings [bib_ref] Metabolomic and proteomic analyses of persistent valvular atrial fibrillation and non-valvular atrial..., Hu [/bib_ref] [bib_ref] Proteomic analysis reveals dual requirement for Grb2 and PLCgamma1 interactions for BCR-FGFR1-Driven..., Peiris [/bib_ref] [bib_ref] Quantitative phosphoproteomics reveals extensive protein phosphorylation dysregulation in the cerebral cortex of..., Mees [/bib_ref] , which may be an inherent issue of this discovery technique but can also be seen in geneexpression volcano plots [bib_ref] TCF7L1 promotes skin tumorigenesis independently of beta-catenin through induction of LCN2, Ku [/bib_ref].
Taken together, our results suggest that although elevated CDK4 levels constitute a tumour progression factor, they also sensitize tumour cells to paclitaxel in a specific manner. This collateral effect, mediated by filamin A and CLIP-170, uncovers a potential liability with an easy-to-determine biomarker for TNBC. Our data ultimately showcase clinical phosphoproteomics as a tool to understand tumour biology in the context of cancer treatments, evidencing a great potential for complex contexts characterized by a lack of oncogene addiction hits.
# Methods
## Patient tumour samples
Patients enrolled in the CNIO-BR-2010-03 trial that consented for tumour sampling underwent image-guided (ultrasound or MRI) tumour biopsy. Three tissue cores obtained with a 14G needle were obtained and snap-frozen upon verification of presence of tumour tissue by the imprinting procedure, within less than 15 min after biopsy. The samples were stored in each study site and shipped to CNIO after trial completion.
Regarding the two patient sets samples used for immunohistochemical biomarker filtering and confirmation, an ad-hoc protocol was approved at three collaborating hospitals (Hospital Universitario Quiron, Hospital de Fuenlabrada and Hospital 12 de Octubre; protocol approval number CEI: 11/37). Formalin-fixed, paraffin-embedded samples from the pathology archive were retrieved for those patients that signed the informed consent form. The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board.
Both in the samples from the CNIO-BR-2010-03 trial, and in the other two sample sets used for biomarker confirmation, the definition of response to neoadjuvant paclitaxel-based treatment was achieving a pCR according to the Residual Cancer Burden method described by [bib_ref] Residual cancer burden after neoadjuvant chemotherapy and long-term survival outcomes in breast..., Yau [/bib_ref].
## Proteomics
Sample preparation. Protein extraction from the frozen tumour samples was performed with a urea lysis buffer (8 M urea, 50 mM Tris.HCl pH 8, 100 mM NaCl, 1X Roche Protease inhibitor, 1X Roche PhosSTOP phosphatase inhibitor) and using mechanical disruption with a Next Advance Bullet Blender with 1.0 mm glass beads and the following instrument settings: Speed 8, Time 3. Post disruption the buffer was removed and the beads discarded. The tissue extract was incubated at room temperature for 2 h with mixing at 2000 rpm in an Eppendorf ThermoMixer. The extract was centrifuged for 10 min at 5000 × g and 4°C. The protein concentration of the cleared extract was determined using a Qubit protein assay (Invitrogen). Protein was reduced with 10 mM dithiothreitol at 25°C for 30 min followed by alkylation with 15 mM iodoacetamide at 25°C for 45 min in the dark. Proteins were concentrated by addition of 4× volumes of −20°C acetone and overnight incubation at −20°C. Precipitated protein was collected by centrifuging for 10 min at 5000 × g and 4°C. Protein pellets were washed twice with −20°C acetone. Pellets were air dried to remove residual acetone. The washed pellets were reconstituted in 400 μL of urea lysis buffer the protein concentration was determined using a Qubit protein assay (Invitrogen). Protein digestion was performed by addition of 20 μg sequencing grade trypsin (Promega) to each sample and overnight incubation (16 h) at 37°C. The final digest volume was 2 mL adjusted with 25 mM ammonium bicarbonate. The digest was terminated with the addition of 10 μL TFA.
Mass spectrometry. Each digest sample was processed by solid phase extraction (SPE) using a Waters HLB PRiME 30 mg capacity C18 cartridge and gravity flow. Firstly, samples were loaded under vacuum at 5inHg and the cartridge was washed twice with 1 mL 0.1% TFA at 5inHg. Peptides were eluted with 2 × 500 μL of 90% acetonitrile, 0.1% TFA at 5inHg and peptide concentrations were determined by UV absorbance at 280 nm.
Phosphopeptides were enriched using Titansphere TiO 2 tips from GL sciences using the vendor protocol. Briefly, phosphopeptides were eluted from the tips using two eluents: 50 μL 5% NH4OH in water and 50 μL 5% Pyrrolidine in acetonitrile. The two elutions were pooled together and neutralized with acetic acid 50% and dried. Samples were reconstituted in 100 μL 0.1% trifluoroacetic (TFA) acid. Each enriched sample was desalted using a StageTip (ThermoFisher P/N SP301) per the vendor protocol. Peptides were dried and reconstituted in 70 μL of 0.1% TFA prior to analysis. Half of each enriched sample was analysed by nano LC-MS/MS with a Waters NanoAcquity HPLC system interfaced to a ThermoFisher Q Exactive mass spectrometer. Peptides were loaded on a trapping column and eluted over a 75 μm analytical column at 350 nL/min; both columns were packed with JupiterProteo resin (Phenomenex). The injection volume was 30 μL. The mass spectrometer was operated in data-dependent mode, with the Orbitrap operating at 60,000 FWHM and 17,500 FWHM for MS and MS/MS respectively. The fifteen most abundant ions were selected for MS/MS. Data processing. Data were processed with MaxQuant version 1.5.0.25 (Max Planck Institute for Biochemistry). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the Pride partner repository with the dataset identifier PXD034355. The fragmentation spectra were searched against the Homo sapiens Uniprot database (downloaded on 23-12-2013), using Andromeda as the search engine. The precursor mass tolerances were set to 20 ppm for the first search and 4.5 for the main search. Also, 0.05 and 0.5 Da were used for FT and IT detectors. Carbamidomethylation of cysteine was considered as fixed modifications, whereas oxidation of methionine (M); phosphorylation on serine (S), threonine (T) and tyrosine (Y); and protein N-terminal acetylation were chosen as a variable modification, and up to two tryptic missed cleavages were allowed. The match between run function was enabled. A target-decoy database searching strategy was used to evaluate the false-discovery rates (FDRs) at the peptide and protein level.
The identification of kinase-specific substrates was evaluated using linear sequence motifs analysis implemented in MaxQuant. For the identification of phosphorylated motifs Maxquant used the Phos-phoMotif Finder search tool at Human Protein Reference Database was used (http://www.hprd.org/PhosphoMotif_finder).
## Pull down and mass spectrometry analysis
For the immunoprecipitation studies, whole-cell lysates were prepared in RIPA lysis buffer (Sigma-Aldrich, #R0278) containing 1% Halt TM Protease & Phosphatase inhibitor cocktail, EDTA-free (Thermo Scientific #78441). Antibodies (anti-Filamin A -Abcam #ab254184-, anti-CDK4 -Invitrogen #MA5-12984-and anti-alpha Tubulin -Abca-m#ab7291-) and control isotypes IgGs were firstly incubated with protein lysates for an hour on rotation at 4°C (4 µg antibody/mg protein lysate). Then, protein A/G Plus agarose beads (Santa Cruz, #sc-2003) were added and the mix were incubated in rotation overnight at 4°C. For western blotting agarose beads were washed three times with lysis buffer and then boiled in presence of laemmli buffer (1X) (Sigma-Aldrich, #S3401). For mass spectrometry studies, proteins were eluted from the agarose beads in two consecutive steps by shaking for 10 min at 1250 rpm in an Eppendorf Thermomixer in 100 µL of elution buffer The mass spectrometer was operated in a data-dependent mode, with an automatic switch between MS and MS/MS scans using a top 12 method (minimum AGC target 1E3) and a dynamic exclusion of 20 s. MS (350-1400 m/z) and MS/MS spectra were acquired with a resolution of 60,000 and 30,000 FWHM (200 m/z), respectively. Peptides were isolated using a 1.4 Th window and fragmented using higherenergy collisional dissociation (HCD) at 27% normalized collision energy. The ion target values were 3E6 for MS (25 ms maximum injection time) and 1E5 for MS/MS (54 ms maximum injection time). Samples were analysed twice.
For data analysis, raw files were processed with MaxQuant (v 1.6.10.43) using the standard settings against a human protein database (UniProtKB/Swiss-Prot, 20,373 sequences) supplemented with contaminants. Label-free quantification was done with match between runs (match window of 0.7 min and alignment window of 20 min).
Carbamidomethylation of cysteines was set as a fixed modification whereas oxidation of methionines and protein N-term acetylation as variable modifications. Minimal peptide length was set to 7 amino acids and a maximum of two tryptic missed-cleavages were allowed. Results were filtered at 0.01 FDR (peptide and protein level).
For each pair bait/control, the data were normalized by the median of the ratio. For the purpose of calculating ratios, given the fact that proteins identified in the bait pull-downs are in the order of 10 7 to 10 9 psm, and the same proteins in the isotype pull-downs are in the order of 10 4 −10 5 psm, or simply non-detected (0 psm), non-detected proteins (0 psm) were transformed to 1 psm. This transformation allows performing ratio calculations (i.e., dividing a protein of high abundance in one condition by "0" in other condition turns into dividing it by 1, which still has the biological meaning of negligible protein abundance). Proteins with fold-change in Log2 > 2 or not identified in IgG control and with at least 3 psm in bait or 1 psm in both technical replicates, were considered as potential candidates.
## Antibody setup and immunohistochemistry staining
In order to confirm or reject the targets discovered by mass spectrometry analysis, immunohistochemistry stainings were performed when available antibodies were found. For this purpose, breast cancer tissues for routine histological analysis were fixed in 10% buffered formalin (Sigma-Aldrich, #HT501128) and embedded in paraffin. Tissue microarrays were mounted with two 1-mm cores per sample (Quick-Ray Instruments, UNITMA). An expert pathologist examined a template H&E slide from each sample to select the areas for core selection. Immunohistochemical staining was performed on 2.5-μm TMA sections. Immunohistochemistry was performed using an automated protocol developed for the Autostainer Link automated slide staining system (DAKO, Agilent). All steps were performed on this staining platform using validated reagents, including deparaffinization, antigen retrieval (cell conditioning), and antibody incubation and detection.
The following antibodies were used for IHC: phospho-P70S6K Corresponding TMA were acquired and digitalized using the AxioScan.Z1 system (Zeiss). Digitalized images were automatically analysed with the ZEN 2.3 lite software (Zeiss). For staining quartile determination, H-scores were calculated by formula: ((% of Area High Intensity × 3) + (% of Area Medium Intensity × 2) + (% of Area Low Intensity × 1))/100.
## Cell lines, cdna transfection and sirna knockdown
The human triple-negative breast cancer cell line MDA-MB-231 was acquired from the American Type Culture Collection (ATCC) (ATCC, #HTB-26). Cells were maintained following the ATCC recommendations and routinely tested for mycoplasma using the MycoalertTM Mycoplasma Detection Kit (Lonza, #LT07-318).
A vector for human CDK4 overexpression, pRc/CMV-CDK4 was kindly provided by M. Malumbres to generate stably MDA-MB-231 cells overexpressing CDK4. Cells were transfected at 60-70% density with Lipofectamine 3000 (Thermo Fisher Scientific, #L300008) according to the manufacturer's instructions. An empty vector was used as control; these cells were named MDA-MB-231 WT along the manuscript's main text. Transduced cells were selected in 1500 µg/mL neomycin for 2 weeks. For Filamin A knockdown, MDA-MB-231 CDK4 cells were seed in 6-well plates and transfected with Stealth siRNA Filamin A (250 pmol per well) (Thermo Fisher Scientific) using Lipofectamine 2000 (Thermo Fisher Scientific, #12566014) according to the manufacturer's instructions. The knockdown efficiency of Filamin A was tested by western blot after 48 h. A nontargeting siRNA (scramble) was used as a control.
To generate stably MDA-MB-231 cells overexpressing Filamin A, 1 × 10 6 cells were resuspended in 100 μL mixture of Opti-MEM (Thermo Fisher Scientific, #11524456) and 10 μg of pcDNA3-myc-FLNa vector (Addgene, #8982). Then, the mixture was transferred to a sterile Amaxa nucleofection cuvette (Clontech Laboratories) and cells were electroporated with a NEPA21 Super Electroporator (Nepagene) using the appropriate nucleofection programme. After nucleofection, the cells were immediately transferred into 3 mL of prewarmed medium in a 6-well plate. Transduced cells were selected in 1500 µg/mL neomycin for 2 weeks.
## Colony-formation assays
Colony-formation assays were conducted as follows: MDA-MB-231 WT, MDA-MB-231 CDK4 and MDA-MB-231 FLNA and MDA-MB-231 CDK4 siRNA FLNA cell lines were seeded at densities of 2000 cells per well in 12-well plates. After overnight incubation, medium was replaced with fresh medium with either vehicle (control) or drugs. Media and drugs were refreshed every 3-4 days. After 10 days of culture, cells were fixed and stained with 0.1% (w/v) crystal violet in 10% (v/v) ethanol. All experiments were performed at least in triplicate. The well area covered by colonies (colony area intensity) was quantified automatically from flatbed scanner-acquired images of colony assays conducted in multi-well plates using the ImageJ software.
To determine the inhibitory concentration of 50% (IC50) of paclitaxel, adriamycin or cisplatin in MDA-MB-231 cell lines, clonogenic survival assays were performed. Cells were exposed to increasing concentration range of each drug and the IC50 values were derived by a sigmoidal dose-response (variable slope) curve using GraphPad Prism software version 5.04. Values represent the mean of at least three independent experiments.
## Immunoblots
Cells were washed 2× with PBS and harvested in cold RIPA Buffer containing 1% protease and phosphatase inhibitor cocktail. Cell lysates were incubated at 4°C for 15 min, sonicated for 15 min and clarified by centrifugation at 14,000 × g at 4°C for 30 min. Protein concentration was estimated by BCA protein assay kit (Pierce TM BCA Protein Assay Kit, #23227) following the manufacture's instruction. 20 μg of proteins per sample were loaded on SDS-PAGE gel and transferred to nitrocellulose membranes for further processing. 5% BSA was used to block the membrane for 60 min at room temperature, followed by overnight incubation at 4°C with the primary antibodies.
The following primary antibodies were used: CDK4 (Cell Signaling, #12790, 1:1000), Filamin A (Abcam, #ab189183, 1:1000), acetylated alpha-Tubulin (K40) (Santa Cruz Biotechnology, #sc-23950, 1:5000), Vinculin (Sigma-Aldrich, #V913, 1:10,000), βActin (clone AC-15, 1:10,000) (Sigma-Aldrich, #A1978), CLIP-170 (Abcam #ab134907, 1:1000) and alpha-Tubulin (Abcam#ab7291, 1:10,000). Membranes were incubated with appropriate peroxidase-conjugate secondary antibodies (Sigma-Aldrich). Bands were visualized by the enhanced chemiluminescence (ECL) method (Clarity TM Western ECK substrate, Bio-Rad, #170-5060).
Regarding acetylated tubulin bands, band intensities were quantified using ImageJ software (NIH, Bethesda, MD, USA). For this purpose, developed films were scanned and band intensities representing acetylated alpha tubulin and total alpha tubulin expression were quantified. For each sample, the ratio between the amounts of acetylated and total alpha tubulin was calculated. Uncropped scans of blots are supplied in the Source data file.
## Rna extraction and quantitative rt-pcr
Total RNA was extracted from cells with TRIzol reagent (Life Technologies, #15596026) in accordance of the manufacturer's instructions. The same quantity of total RNA (1 µg) was retrotranscribed to cDNA using the Quantitect Reverse Transcription Kit (Qiagen, #205313) (2 min at 42°C, 15 min at 42°C, and 3 min at 95°C). One microliter of cDNA (dilution 1:3) was placed in a 384-well plate with 5 μL of Fast SYBR TM Green Master Mix (Applied Biosystems, #1129726) and 2 µL of the corresponding primers (Filamin A-Fw: TGCTGCCTACTCATGATGC; Filamin A-Rv: GGATG TGTGTCTTCTTCGGC) in a final volume of 10 µL. PCR amplification was performed using the QuantStudio TM 6 Flex Real_time PCR System (Applied Biosystems) under the following thermal cycler conditions: 2 min at 50°C, 10 min at 95°C, and 40 cycles (15 s at 95°C and 1 min at 59°C). To quantify transcription, the mRNA expression levels of the target genes were normalized to β-Actin. All samples were run in triplicates and relative quantification (RQ) was calculated following the ΔCt method: RQ = 2 − ΔCt, where ΔCt is the difference between the Ct of the gene of interest and the Ct of the endogenous gene control β-actin.
## Confocal studies
In order to evaluate the recovery of cells to paclitaxel treatment, we evaluate the percentage of abnormal divisions. The day after seeding cells on glass coverslip a concentration of 5 nM of paclitaxel was added for 4 h. After washing cells with PBS to eliminate remanent paclitaxel, normal medium was added and cells were grown for 48 h. Then, cells were fixed in 4% paraformaldehyde (PFA) for 10 min and permeabilized with 0.5% Triton X-100/PBS plus 0.05%SDS (20 min at room temperature). Coverslips were incubated in blocking solution (3% Bovine Serum Albumin in PBS-T (PBS + 0.05% Tween-20) for 1 h. Primary antibody solution was applied overnight at 4°C. The following primary antibodies were used: alpha-Tubulin (Abcam, #ab7291) (1:1000) and gamma-Tubulin (Invitrogen, #PA5-34815) (1:1000). Cells were then washed with PBS-T and incubated with Alexa Fluor 488-(Molecular Probes, #A11029) (1:200) or 555-conjugated secondary antibody (Molecular Probes, #A21429) (1:200). Cell nuclei were stained with DAPI (Sigma-Aldrich, #D9542) (1:2000). After washing, coverslips were mounted onto glass microscope slides with Prolong (Invitrogen, #P36930). Images were acquired in a Leica-TCS SP5X confocal microscope, with a HCX PL APO 63× 1.4 numerical aperture (NA) oilimmersion objective using LAS AF version 2.5.1 software. To estimate the abundance of abnormal divisions, all metaphases and anaphases found in each condition were acquired. For abnormal nuclei, 15random fields were acquired and around 1000 cells at interphase were scored in each assay. All experiments were done by triplicate.
In order to study the cellular localization of CDK4, alpha-Tubulin and Filamin A, cells grown on coverslips were subject to immunofluorescence studies as previously described. The following primary antibodies were used: Filamin A (Abcam, #ab254184) (1:500) and CDK4 (clone DSC5, Millipore, #MAB8879) (1:250) and alpha-Tubulin (Abcam, #ab7291; 1:1000).
For colocalization determination of CDK4 (green) and Filamin A (red), and α-Tubulin (green) and Filamin A (red) 30 fields were acquired and the correlation between pixels were estimated using Definiens Developer v2.5 software. Pearson index for colocalization was performed by mean of customized algorithms programmed in Definiens Developer using more than 250 cells per genotype. When Pearson index values are between 0.5 and 1 cells are considered as positive for colocalization.
## Paclitaxel accumulation and live cell imaging
In order to study the kinetics of paclitaxel accumulation in cells, MDA-MB-231 WT, MDA-MB-231 CDK4, MDA-MB-231 FLNA and MDA-MB-231 CDK4 siRNA FLNA cell lines were seeded at densities of 30,000 cells per well in a μ-slide 8 well chamber (Ibidi) and grew in 10% FBS DMEM medium overnight. After PBS rinsing, cells were incubated with 200 nM Flutax-2 (paclitaxel, Oregon Green 488 conjugate) (Invitrogen, #P22310) and 25 μM verapamil (Sigma-Aldrich, #152-11-4) in 10% FBS Hank's Balanced Salt Solution (HBBS) medium (Sigma-Aldrich, #H6648) for 48 h. Images were taken every 15 min during 48 h using a Leica-TCS SP5X confocal microscope, with a HCX PL APO 20X objective using LAS AF version 2.5.1 software. Quantitation of green fluorescence was performed in live cells using the Definiens Developer XD software with a customized script for detection and quantification of green area.
## Kinase assay
Kinase reactions were carried out at 30°C for 30 min using specific kinases buffer. Human recombinant CDK4/Cyclin D1 (Sigma-Aldrich, #C0620) was incubated with human recombinant Filamin A (Origene, #TP326488) in kinase assay buffer (20 mM MOPS, pH 7.2, 12.5 mM glycerol 2-phosphate, 25 mM MgCl 2 , 5 mM EGTA, 2 mM EDTA and DTT 0.25 mM) in presence of 50 μM cold ATP and 1.5 uCi [32 P] ATP. Recombinant RB (Santa Cruz Biotechnology, #sc-4112) was used as a positive phosphorylatable substrate for CDK4/Cyclin D1. Reactions were stopped by addition of Laemmli sample buffer. Radioactive samples were subject to acrylamide gel electrophoresis, followed by gel drying and autoradiography.
## Statistics
Clinical and demographic characteristics of the study patients by treatment arm were compared by the Mann-Whitney, Chi-Squared or Fisher's test as appropriate.
In order to study the relationship with response to paclitaxel of the candidate kinases or phosphoproteins resulting from the analysis of the discovery sets in Set 1 and Set 2, an immunohistochemical H-score was calculated for each candidate. The H-scores were then categorized as follows: for each candidate marker, patients with a H-score above the 75th percentile were encoded as High, whereas those with H-score below the 75th percentile were categorized as Low. In order to find the hazard ratio of obtaining or not a response in case of belonging to the High or Low categories, a univariate logistic regression model was run for each candidate marker in Sets 1 and 2. Then, for each marker with a positive association in the regression model (p-P70S6K, CDK4, p-vimentin, filamin A and HMG-Coa reductase), the proportion of patients achieving a pCR among those with high staining was compared with the proportion of patients achieving a pCR among those with low staining. The proportion of pCR for each category (i.e., High or Low) for each candidate marker was compared with a Chi-square test. For each breast cancer subtype (or the whole Set 1), 5 comparisons were run (one per marker) in Set 1. In Set 2, 5 comparisons were run as well; thus, the Bonferroni correction was applied to the type-1 error. P-values below 0.01 were considered significant, whereas P-values between 0.01 and 0.1 were considered borderline significant.
CDK4 and Filamin A correlation, or their correlation with Ki67 replicative fraction, were investigated with the Pearson´s coefficient in a pairwise manner. Drug sensitivity, mRNA levels and acetylated tubulin levels were compared with two-sided unpaired t-tests. All tests were performed with the SPSS Statistics V.19.0 software.
Unsupervised hierarchical biclustering was performed using Morpheus software (https://software.broadinstitute.org/morpheus). Consensus clustering and pvclust clustering methods were used to assess the robustness of the unsupervised clustering obtained by UPGMA. The R libraries ConsensusClusterPlus 59 and pclust [bib_ref] Pvclust: an R package for assessing the uncertainty in hierarchical clustering, Suzuki [/bib_ref] were used for this purpose.
A Gene Set Enrichment Analysis (GSEA) 60 was used to define sets of kinase substrate motifs that shows statistically significant, concordant differences between categories of interest, such as responders or nonresponders to paclitaxel, herein termed Kinase Set Enrichment Analysis (KSEA). To this end, KSEA was applied using annotations for motifs extracted from Perseus software. Leading proteins data matrix was ranked based on their t-test statistic. KSEA scoring scheme was classic. The kinase-set size limits were established at >5 and <1000. After Kolmogorov-Smirnoff testing with 1000 permutations, those kinase-sets showing FDR < 0.20 were selected as significant.
In order to compare phosphopeptides up-or downregulated in responders or nonresponders in the paclitaxel arm, the paclitaxel plus nintedanib arm, or in the whole trial, normality of the phosphopeptide intensity distribution was tested with the Shapiro Wilk normality test. H0 was that the distributions were normal, and H1 was that they were not. H0 was rejected in the three cases (P < 2.2 × 10 −16 in the three cases), and thus, the assumption of non-normality of the data distribution was adopted. According to this the median value for each phosphopeptide was calculated for each condition (i.e., responders or nonresponders, in Arm A, Arm B, or whole trial). The median values were compared with the non-parametric Mann-Whitney Wilcoxon test using 100,000 permutations; thus, P-values were calculated with 5 decimal places. The obtained P-values were adjusted by FDR using the Benjamini-Hochberg method to account for multiple testing.
Volcano-plots were depicted using the median log fold-change intensity (X-axis) values and raw P-values (Y-axis). The plots were generated with the GraphPad Prism software version 5.04. In order to avoid divisions by 0 in the phospho-peptide ratio calculation when two classes were compared (i.e., responders and nonresponders) and any given phospho-peptide was highly abundant in one class and undetectable in the other, the value "0" in spectral intensity was switched to "1" (a change in 1 spectral intensity when phospho-peptides are found in the range of 2 10 -2 20 intensity is biologically negligible, but allows the calculation of the ratios).
## Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
## Data availability
All the data supporting the findings of this study are available within the article and its supplementary information files. The mass spectrometry proteomics data are publicly available in the ProteomeXchage Consortium via the PRIDE partner repository with the dataset identifier PXD034355. Source data are provided with this paper.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2022 [bib_ref] The clonal and mutational evolution spectrum of primary triple-negative breast cancers, Shah [/bib_ref]
[fig] Figure 1 |: Clinical trial tumour samples: the phosphoprofiles were independent of the study site and/or treatment arm. A Clinical trial treatment and sampling schedule. After randomization, patients were scheduled for a fresh tumour biopsy. [/fig]
[fig] Figure 2 |: Kinases driving the phosphoprofiles of responders to paclitaxel. Panel A displays the significantly enriched kinases found in the baseline samples for patients who achieved a pCR in the paclitaxel monotherapy arm. However, Panel B [/fig]
[fig] Figure 3 |: Volcano plots with regulated phosphopeptides among responders and nonresponders to paclitaxel. Regulated phosphopeptides in patients who achieved a pCR versus those who did not in the standard arm (A), whole trial (B) and the experimental arm (C). Phosphopeptides with greater than 2 4 (16-fold) up-or downregulation in one or another condition with Mann-Whitney Wilcoxon P < 10 −2.5 are highlighted and colour-coded for each comparison; their IDs and FDRs are listed by colour group and comparison in Supplementary Data 1. [/fig]
[fig] Figure 4 |: Independent testing of biomarkers of response to paclitaxel. A Immunohistochemical staining of filamin A from patients with H-scores in the upper (Q1, left) and lower (Q4, right) quartiles. Scale bars: 100 µm. B Comparison of pCR rates for patients in the Q1 versus Q2-4 subgroups among the whole Set-1 series (N = 117) or divided by subtype (HR+: Hormone-positive, N = 62; HER2: HER2amplified, N = 35; TNBC: triple-negative breast cancer, N = 20), according to the staining of p-P70S6K (Thr 389 ), CDK4, p-Vimentin (Ser 56 ), filamin A or HMGCOA reductase. C Same as in (B) but for Set-2 patients (N = 101). D pCR rates of patients from Set-2 with combined elevated levels of CDK4 and filamin A compared to patients with low CDK4 and filamin A levels. Proportions were compared with a two-sided Xi-square test. P values are adjusted (Bonferroni). Error bars: standard error of the proportion. Source data are provided as a Source data file. [/fig]
[fig] Figure 5 |: CDK4-and filamin A-driven increased sensitivity to paclitaxel in TNBC models. A Left: Immunoblot showing different CDK4 levels between parental MDA-MB-231 and MDA-MB-231 CDK4 cells. Tubulin is shown for housekeeping purposes. Experiment was repeated 3 times with similar results. Middle: colony assay pictures (upper panel) and quantitation or relative plating efficiency chart (lower panel) of parental MDA-MB-231 and MDA-MB-231 CDK4 cells exposed to various concentrations of paclitaxel. In the relative plating efficiency chart, each condition is normalized to the untreated condition for each MDA-MB-231 variant. Data are presented as the mean ± SEM, n ≥ 6 independent experiments; two-sided unpaired t-test. Right: IC50 plots showing the~2.7-fold variation in paclitaxel sensitivity. Data are presented as the mean ± SEM, n ≥ 3 independent experiments; B Immunoblot showing the upregulation of filamin A in MDA-MB-231 CDK4 cells versus the parental cells; tubulin is shown for housekeeping purposes. Samples derived from the same experiment and blots were processed in parallel. Experiment was repeated 3 times with similar results. C Real-time PCR showed a 2.5-fold increase in filamin A mRNA levels in MDA-MB-231 CDK4 cells. Data are presented as the mean ± SEM, n = 4 independent experiments; two-sided unpaired t-test. D Left: Immunoblot showing filamin A levels in MDA-MB-231 FLNA and MDA-MB-231 WT cells. No changes were observed in CDK4 levels in the former. Vinculin: housekeeping. Samples derived from the same experiment and blots were processed in parallel. Experiment was repeated 3 times with similar results. Right: IC50 charts showing the sensitization to paclitaxel in MDA-MB-231 FLNA cells compared to the parental cells (Mean±SEM, n≥3 independent experiments) together with examples of colony assays. E Left: Transient filamin A knockdown with siRNA in MDA-MB-231 CDK4 cells. Experiment was repeated 3 times with similar results. Right: phenotype recovery: the increased sensitivity to paclitaxel in MDA-MB-231 CDK4 cells is restored back to normal (i.e., similar IC50 to that of MDA-MB-231 WT) when filamin A is downregulated. Data are represented as mean ± SEM, n ≥ 14 independent experiments. Examples of colony assays are shown. Source data are provided as a Source data file. [/fig]
[fig] Figure 7 |: Elevated CDK4 and filamin A lead to increased tubulin acetylation and stabilization, enhancing the effect of paclitaxel on mitotic aberrations. A Acetylated tubulin (i.e., stabilized microtubules) in parental, MDA-MB-231 CDK4 and MDA-MB-231 FLNA cells, untreated or in response to 5 nM paclitaxel. Samples derived from the same experiment and blots were processed in parallel. On the right-hand side, a bar chart showing the quantification of acetylated tubulin normalized to total tubulin in the different experimental conditions is shown. The quantification is the result of averaging three experimental replicates. Data are presented as the mean±SEM, n = 3 independent experiments; two-sided unpaired t-test. B Reversion of increased tubulin acetylation in MDA-MB-231 CDK4 cells by filamin A knockdown. Samples derived from the same experiment and blots were processed in parallel. C Evaluation of abnormal cell divisions, including metaphases and anaphases, in parental, CDK4 and FLNA MDA-MB-231 variants. Cells were exposed to 5 nM paclitaxel for 4 h, washed, and allowed to recover for 48 h. At this point, the percentages of normal metaphases/anaphases, multipolar metaphases, metaphases with misaligned chromosomes, metaphases with abnormal mitotic spindles, anaphases bridges, and anaphases with lagging chromosomes were quantified in each cell line. Data are presented as the mean ± SEM, n = 3 independent experiments. Supplementary Data 4 shows the percentage of cells displaying each aberration in either baseline or treated parental, CDK4 or FLNA MDA-MB-231 variants. The pictures on the right-hand side depict examples of each mitotic aberration type. Scale bars: 10 µm. D Same as in (C) regarding abnormal nuclei resulting from the previous aberrant processes. Scale bars: 25 µm. Source data are provided as a Source data file. [/fig]
[fig] 8 M: Urea, 100 mM Tris-HCl pH = 8.0). The supernatant obtained was digested by means of standard FASP protocol. Briefly, proteins were reduced (15 mM TCEP, 30 min, RT), alkylated (50 mM CAA, 20 min in the dark, RT) and sequentially digested with Lys-C (Wako) (protein:enzyme ratio 1:50, 4 h at RT) and trypsin (Promega, #V5071) (protein:enzyme ratio 1:50, o/n at 37°C). Resulting peptides were desalted using home-made C18 Stage-tips. For the proteomic analysis, LC-MS/ MS was carried out by coupling an Ultimate 3000 RSLCnano System (Dionex) with a Q-Exactive HF-X mass spectrometer (Thermo-Scientific). Peptides were loaded into a trap column (Acclaim Pep-MapTM 100, 100 µm × 2 cm, ThermoScientific) over 3 min at a flow rate of 10 µl/min in 0.1% FA. Then peptides were transferred to an analytical column (PepMapTM RSLC C18, 2 µm, 75 µm × 50 cm, ThermoScientific) and separated using a 60 min effective linear gradient (buffer A: 0.1% FA; buffer B: 100% ACN, 0.1% FA) at a flow rate of 250 nL/min. The gradient used was: 0-3 min 2% B, 3-5 min 6% B, 5-36 min 17.5% B, 36-60 min 25% B, 60-63 min 33% B, 63-65 min 45% B, 65-70 min 98% B, 70-80 min 2% B. The peptides were electrosprayed (1.5 kV) into the mass spectrometer through a heated capillary at 300°C and an Ionfunnel RF level of 40%. [/fig]
|
Candida Species Biofilms’ Antifungal Resistance
Candida infections (candidiasis) are the most prevalent opportunistic fungal infection on humans and, as such, a major public health problem. In recent decades, candidiasis has been associated to Candida species other than Candida albicans. Moreover, biofilms have been considered the most prevalent growth form of Candida cells and a strong causative agent of the intensification of antifungal resistance. As yet, no specific resistance factor has been identified as the sole responsible for the increased recalcitrance to antifungal agents exhibited by biofilms. Instead, biofilm antifungal resistance is a complex multifactorial phenomenon, which still remains to be fully elucidated and understood. The different mechanisms, which may be responsible for the intrinsic resistance of Candida species biofilms, include the high density of cells within the biofilm, the growth and nutrient limitation, the effects of the biofilm matrix, the presence of persister cells, the antifungal resistance gene expression and the increase of sterols on the membrane of biofilm cells. Thus, this review intends to provide information on the recent advances about Candida species biofilm antifungal resistance and its implication on intensification of the candidiasis.
# Introduction
During the last two decades, the occurrence of Candida species infections has been increasing and becoming more difficult to treat due to the growth of immunogenic diseases, the disproportionate use of immunosuppressive drugs, malnutrition, endocrine disorders, the widespread use of indwelling medical devices, broad spectrum antibiotics, aging and an increase in patient's population [bib_ref] Candida glabrata: An emerging oral opportunistic pathogen, Li [/bib_ref] [bib_ref] Current treatment of oral candidiasis: A literature review, Garcia-Cuesta [/bib_ref]. A fairly small number of Candida species are pathogenic for humans, causing superficial and deep-seated mycoses, disseminated worldwide [bib_ref] Candida biofilms and their role in infection, Douglas [/bib_ref]. Nonetheless, Candida is becoming a significant clinical problem that has taken the opportunity to create infections, called candidiasis [bib_ref] Candida glabrata: An emerging oral opportunistic pathogen, Li [/bib_ref] [bib_ref] Current treatment of oral candidiasis: A literature review, Garcia-Cuesta [/bib_ref] [bib_ref] Candida biofilms and their role in infection, Douglas [/bib_ref].
Candida albicans continues to be the most prevalent and problematic of all Candida species. However, with the development of molecular identification methods, the number of other Candida species, non-Candida albicans Candida (NCAC) species, identified as implicated in candidiasis, is now superior. This group includes, among others, Candida glabrata, Candida tropicalis, Candida parapsilosis and Candida krusei species [bib_ref] Biofilms: Survival Mechanisms of Clinically Relevant Microorganisms, Donlan [/bib_ref] [bib_ref] Resistance of Candida albicans biofilms to antifungal agents in vitro, Hawser [/bib_ref]. The pathogenicity of Candida species is attributed to certain virulence factors, such as the ability to evade host defences, adhesion and biofilm formation (on host tissues and or on medical devices), and the production of tissue-damaging hydrolytic enzymes, such as proteases, phospholipases and hemolysins [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref]. Biofilms are biological communities with an extraordinary degree of organization, in which Candida cells form structured, coordinated, and functional communities, embedded in a self-secreted extracellular matrix. Biofilm production is also related to a high level of antifungal resistance of the associated microorganisms. Moreover, the ability of Candida species to form drug-resistant biofilms is an important factor in their contribution to human disease [bib_ref] Candida glabrata: A review of its features and resistance, Rodrigues [/bib_ref]. In the Candida albicans +++++ Basal blastospore layer with a dense overlying matrix composed of exopolysaccharides and hyphae. [bib_ref] Dispersion as an important step in the Candida albicans biofilm developmental cycle, Uppuluri [/bib_ref] [bib_ref] Risk factors and outcomes of candidemia caused by biofilm-forming isolates in a..., Tumbarello [/bib_ref] Candida dubliniensis ++/+++ Chains of cells with thin extracellular matrix material. [bib_ref] Antifungal susceptibility of Candida biofilms: Unique efficacy of Amphotericin B lipid formulations..., Kuhn [/bib_ref] [bib_ref] Comparative analysis of Candida biofilm quantitation assays, Taff [/bib_ref] Hhigh variability among clinical isolates.
Candida glabrata ++/+++ Forms considerably less biofilm than C. albicans. [bib_ref] Resistance of Candida albicans biofilms to antifungal agents in vitro, Hawser [/bib_ref] [bib_ref] Risk factors and outcomes of candidemia caused by biofilm-forming isolates in a..., Tumbarello [/bib_ref] High in both protein and carbohydrate content.
## Candida biofilms and resistance mechanisms
The major classes of antifungal drugs used for treatment of Candida species infections are azoles, polyenes, and echinocandins [bib_ref] Recent advances on filamentous fungal biofilms for industrial uses, Gutiérrez-Correa [/bib_ref] [bib_ref] Epidemiology of invasive candidiasis, Arendrup [/bib_ref] [bib_ref] Antifungal treatment strategies in patients with haematological diseases or cancer: From prophylaxis..., Mikolajewska [/bib_ref] [bib_ref] Echinocandin antifungal drugs in fungal infections: A comparison, Chen [/bib_ref] [bib_ref] The evolution of fungal drug resistance: Modulating the trajectory from genotype to..., Cowen [/bib_ref]. Azoles (e.g., fluconazole, voriconazole and posaconazole) possess a fungistatic effect, blocking ergosterol synthesis, targeting the enzyme lanosterol 14α-demethylase (related to the ERG11 gene) and leading to an accumulation of toxic sterol pathway intermediates. Polyenes (e.g., amphotericin B and nystatin) are fungicidal, intercalating into membranes containing ergosterol, creating pores that destroy the proton gradient, which result in the outflow of the cytoplasm and other cell contents. Echinocandins (e.g., caspofungin, micafungin and anidulafungin) are also a fungicidal, targeting the synthesis of 1,3-β-glucan, a component of the Candida species cell wall. It is also important to address that preferably to azoles, the use of echinocandins and polyenes is recommended if the patient had prior azoles exposure and if the infection is markedly severe for patients infected with C. glabrata, which is consider as generally very azole-resistant. Echinocandins are, most frequently and according to the latest guidelines, the first antifungal drug choice in these severe cases of candidemia [bib_ref] Cellular, and Molecular Factors That Contribute to Antifungal Drug Resistance, White [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref]. There is some evidence that suggests that prophylactic use of fluconazole may be advantageous for preterm neonates, transplant recipients, intensive care unit patients, and other high-risk patient populations [bib_ref] Prophylactic antifungal therapy in the intensive care unit, Rex [/bib_ref] [bib_ref] Prophylactic fluconazole is effective in preventing fungal colonization and fungal systemic infections..., Manzoni [/bib_ref] [bib_ref] Clinical trials of antifungal prophylaxis among patients in surgical intensive care units:..., Lipsett [/bib_ref] [bib_ref] The use of prophylactic fluconazole in immunocompetent high-risk surgical patients: A meta-analysis, Ho [/bib_ref] [bib_ref] Risk factors for candidal bloodstream infections in surgical intensive care unit patients:..., Blumberg [/bib_ref]. Though, due to some controversies, this is not a standard for all hospitals.
Initial studies examined the impact of known mechanisms to play a role in drug resistance during planktonic Candida cells growth [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref] [bib_ref] Effects of fluconazole on Candida glabrata biofilms and its relationship with ABC..., Fonseca [/bib_ref] [bib_ref] Candida tropicalis biofilm's matrix-involvement on its resistance to amphotericin B, Fernandes [/bib_ref] [bib_ref] Molecular detection of resistance to azole components, Posteraro [/bib_ref]. As described, acquired planktonic cell resistance has been linked to increased efflux pump activity, mutations in genes encoding drug target enzymes and alterations in the composition of both the cell membrane and the cell wall [bib_ref] Effects of fluconazole on Candida glabrata biofilms and its relationship with ABC..., Fonseca [/bib_ref]. The Candida biofilm resistance phenomenon was for the first time demonstrated in 1995 for C. albicans by [bib_ref] Resistance of Candida albicans biofilms to antifungal agents in vitro, Hawser [/bib_ref] [bib_ref] Resistance of Candida albicans biofilms to antifungal agents in vitro, Hawser [/bib_ref]. After that, the ability of Candida species biofilms to survive extraordinarily to high antifungal concentrations has been the subject of numerous investigations for many researchers [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref] [bib_ref] Effects of fluconazole on Candida glabrata biofilms and its relationship with ABC..., Fonseca [/bib_ref] [bib_ref] Candida tropicalis biofilm's matrix-involvement on its resistance to amphotericin B, Fernandes [/bib_ref] [bib_ref] Candida glabrata's recurrent infections: Biofilm formation during Amphotericin B treatment, Rodrigues [/bib_ref] [bib_ref] Effect of voriconazole on Candida tropicalis biofilms: Relation with ERG genes expression, Fernandes [/bib_ref]. So, in the last decade, additional investigations began to focus on the role of biofilm-specific traits. These studies have examined the influence of high cell density, growth rate reduction, nutrient limitation, matrix extracellular production, presence of persister cells, gene expression alterations and sterols content increase on Candida membrane cells. The role of these mandatory factors is reviewed below and is schematized on [fig_ref] Figure 1: General scheme of the mechanisms described as involved on Candida species biofilm... [/fig_ref].
high antifungal concentrations has been the subject of numerous investigations for many researchers [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref] [bib_ref] Effects of fluconazole on Candida glabrata biofilms and its relationship with ABC..., Fonseca [/bib_ref] [bib_ref] Candida tropicalis biofilm's matrix-involvement on its resistance to amphotericin B, Fernandes [/bib_ref] [bib_ref] Candida glabrata's recurrent infections: Biofilm formation during Amphotericin B treatment, Rodrigues [/bib_ref] [bib_ref] Effect of voriconazole on Candida tropicalis biofilms: Relation with ERG genes expression, Fernandes [/bib_ref]. So, in the last decade, additional investigations began to focus on the role of biofilm-specific traits. These studies have examined the influence of high cell density, growth rate reduction, nutrient limitation, matrix extracellular production, presence of persister cells, gene expression alterations and sterols content increase on Candida membrane cells. The role of these mandatory factors is reviewed below and is schematized on [fig_ref] Figure 1: General scheme of the mechanisms described as involved on Candida species biofilm... [/fig_ref].
## Impact of candida cells density, nutrient and growth limitation
An important biofilm-specific trait suspected to influence antifungal resistance is the high relative concentration of Candida cells into biofilm communities comparatively to the great majority of planktonic conditions [bib_ref] Role for cell density in antifungal drug resistance in Candida albicans biofilms, Perumal [/bib_ref] [bib_ref] Biofilm lifestyle of Candida: A mini review, Seneviratne [/bib_ref]. [bib_ref] Role for cell density in antifungal drug resistance in Candida albicans biofilms, Perumal [/bib_ref] [bib_ref] Role for cell density in antifungal drug resistance in Candida albicans biofilms, Perumal [/bib_ref] after studying the cells density effect on antifungal treatment, observed that the azole antifungals' tolerance to planktonic cell cultures was effectively lower when compared to intact and/or disrupt biofilm communities. Candida
## Impact of candida cells density, nutrient and growth limitation
An important biofilm-specific trait suspected to influence antifungal resistance is the high relative concentration of Candida cells into biofilm communities comparatively to the great majority of planktonic conditions [bib_ref] Role for cell density in antifungal drug resistance in Candida albicans biofilms, Perumal [/bib_ref] [bib_ref] Biofilm lifestyle of Candida: A mini review, Seneviratne [/bib_ref]. [bib_ref] Role for cell density in antifungal drug resistance in Candida albicans biofilms, Perumal [/bib_ref] [bib_ref] Role for cell density in antifungal drug resistance in Candida albicans biofilms, Perumal [/bib_ref] after studying the cells density effect on antifungal treatment, observed that the azole antifungals' tolerance to planktonic cell cultures was effectively lower when compared to intact and/or disrupt biofilm communities. Candida albicans biofilm formation is associated with the dimorphic switch between yeast and hyphal growth, and biofilms of this species generically have two distinct layers: a thin, basal yeast layer and a thickener compact hyphal layer [bib_ref] Biofilms: Survival Mechanisms of Clinically Relevant Microorganisms, Donlan [/bib_ref]. In contrast, C. parapsilosis biofilms tend to be thinner, less structured, and consist almost exclusively of aggregates [bib_ref] Antifungal susceptibility of Candida biofilms: Unique efficacy of Amphotericin B lipid formulations..., Kuhn [/bib_ref] [bib_ref] Biofilms of non-Candida albicans Candida species: Quantification, structure and matrix composition, Silva [/bib_ref]. Candida tropicalis biofilms consist of a dense network of yeast cells with evident different filamentous morphologies and C. glabrata biofilms are structured on multilayers of blastospores with high cohesion among them [bib_ref] Biofilms of non-Candida albicans Candida species: Quantification, structure and matrix composition, Silva [/bib_ref]. In general, C. glabrata biofilms possessed higher density of cells comparatively to C. tropicalis and C. parapsilosis biofilms [bib_ref] In vitro biofilm activity of non-Candida albicans Candida species, Silva [/bib_ref] , which may be implicated on the usual highest resistance of C. glabrata biofilms to antifungal azoles and/or amphotericin B [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref] [bib_ref] Effects of fluconazole on Candida glabrata biofilms and its relationship with ABC..., Fonseca [/bib_ref] [bib_ref] Candida glabrata's recurrent infections: Biofilm formation during Amphotericin B treatment, Rodrigues [/bib_ref].
The well-structured biofilms layers open another hypothesis for Candida species antifungal resistance, that is, that cells placed in deeper layers of the biofilm grow slower owing to a lack of nutrients, and are subsequently more resistant to antifungal drugs. There is in fact a lack of work concerning this subject. However, by controlling nutrients in a perused biofilm fermentor, Baillie and Douglas [bib_ref] Effect of growth rate on resistance of Candida albicans biofilms to antifungal..., Baillie [/bib_ref] [bib_ref] Iron-limited biofilms of Candida albicans and their susceptibility to Amphotericin B, Baillie [/bib_ref] were able to compare the antifungal susceptibility of C. albicans biofilms growing at various rates. These authors, in opposition to what was expectable over a wide range of growth rates, verified that biofilm-associated cells exhibited similar levels of resistance to amphotericin B, suggesting that growth rate does not play a significant role in biofilm antifungal resistance. Similarly, C. albicans grown under glucose and iron limitation conditions were shown to both be highly resistant to amphotericin B [bib_ref] Iron-limited biofilms of Candida albicans and their susceptibility to Amphotericin B, Baillie [/bib_ref]. Nevertheless, factors including pH, temperature, and oxygen availability are described as possible inductors of biofilm architecture alteration and thus the antifungal sensibility [bib_ref] Defined anaerobic growth medium for studying Candida albicans basic biology and resistance..., Dumitru [/bib_ref] [bib_ref] Detailed comparison of Candida albicans and Candida glabrata biofilms under different conditions..., Kucharíková [/bib_ref] [bib_ref] Temperature affects the susceptibility of Cryptococcus neoformans biofilms to antifungal agents, Pettit [/bib_ref].
The general physiological state of sessile cells has also been reported as implicated in the susceptibility profiles of Candida biofilms. Metabolic activity confirms that cells within biofilms are undergoing mitochondrial respiration during development [bib_ref] Antifungal susceptibility of Candida biofilms: Unique efficacy of Amphotericin B lipid formulations..., Kuhn [/bib_ref] [bib_ref] Recent advances on filamentous fungal biofilms for industrial uses, Gutiérrez-Correa [/bib_ref] [bib_ref] Biofilm lifestyle of Candida: A mini review, Seneviratne [/bib_ref] [bib_ref] Temperature affects the susceptibility of Cryptococcus neoformans biofilms to antifungal agents, Pettit [/bib_ref] ].
## Contribution of the extracellular matrix production
Extracellular matrix (ECM) is a defining characteristic of all Candida species biofilms, providing the cells protection from hostile factors such as host immunity and antifungal agents [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref] [bib_ref] Standardized method for in vitro antifungal susceptibility testing of Candida albicans biofilms, Ramage [/bib_ref]. In some of the pioneer works, Candida species biofilm's matrices were shown to increase when biofilms were grown under dynamic flow conditions and their quantity is strongly species-and strains-dependent [bib_ref] Resistance of Candida albicans biofilms to antifungal agents in vitro, Hawser [/bib_ref] [bib_ref] Biofilms of non-Candida albicans Candida species: Quantification, structure and matrix composition, Silva [/bib_ref] [bib_ref] Penetration of Candida biofilms by antifungal agents, Al-Fattani [/bib_ref]. Subsequent work has shown that while ECM hampers diffusion, penetration of antifungal drugs is not thought to play an important role in biofilm resistance [bib_ref] Penetration of Candida biofilms by antifungal agents, Al-Fattani [/bib_ref]. However, more recent studies have provided new insights about the chemical composition of ECM and that it may play a central role in resistance by antifungal agents' neutralization.
It is important to address that the composition of the Candida biofilm matrices is species-variable. Little is known about matrix composition of NCAC species biofilms, but according to , C. albicans biofilm matrix is mainly composed of carbohydrates, proteins, phosphorus and hexosamines. Silva and colleagues 2009 [bib_ref] Biofilms of non-Candida albicans Candida species: Quantification, structure and matrix composition, Silva [/bib_ref] reported that the ECM of C. parapsilosis contained large amounts of carbohydrates and low levels of proteins. In the same study, C. glabrata biofilm matrices were found to have high levels of both proteins and carbohydrates, while C. tropicalis biofilm matrices had low levels of carbohydrates and proteins compared to the other NCAC species. Recently, [bib_ref] Candida glabrata's recurrent infections: Biofilm formation during Amphotericin B treatment, Rodrigues [/bib_ref] [bib_ref] Candida glabrata's recurrent infections: Biofilm formation during Amphotericin B treatment, Rodrigues [/bib_ref] revealed for the first time the presence of β-glucans in the C. glabrata matrices even when treated with amphotericin B. Furthermore, other authors [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref] showed that matrix material extracted from biofilms of C. tropicalis and C. albicans contained carbohydrates, proteins, hexosamine, phosphorus and uronic acid. However, the major component quantified in C. tropicalis biofilm matrices was hexosamine (27%). The same authors also reported that these biofilms partially detached after treatment with lipase type VII and chitinase, which is in contrast to biofilms of C. albicans that detached only after treatment with proteinase K, chitinase, DNase I or β-N-aceytyglucosamidase. In Candida species, there is scarce knowledge concerning the contribution of extracellular DNA to biofilm matrix and overall structure [bib_ref] Biofilms in chronic rhinosinusitis: Systematic review and suggestions for future research, Keir [/bib_ref].
In this sense, studies have been carried out to clarify the involvement of some of the matrix components in Candida biofilm resistance. [bib_ref] Addition of DNase improves the in vitro activity of antifungal drugs against..., Martins [/bib_ref] highlighted the importance of DNA in C. albicans biofilm formation, integrity and structure and that the addition of DNase improves the efficacy of polyenes and echinocandins, but not to azoles [bib_ref] Addition of DNase improves the in vitro activity of antifungal drugs against..., Martins [/bib_ref].
The major carbohydrate component is β-1,3 glucans, as treatment of C. albicans biofilms with β-1,3 glucanase helps detach biofilms from a substrate [bib_ref] Penetration of Candida biofilms by antifungal agents, Al-Fattani [/bib_ref]. Its contribution is confirmed, where it was shown to increase in C. albicans biofilm cell walls compared to planktonic organisms and was also detected in the surrounding biofilm milieu and as part of the ECM [bib_ref] Putative role of β-1,3 glucans in Candida albicans biofilm resistance, Nett [/bib_ref]. The involvement in the resistance was realized when it was also shown that biofilm cell walls bound four-to five-fold more azole than equivalent planktonic cells, and culture supernatant bound a quantifiable amount of this antifungal agent. Moreover, β-1,3 glucanase markedly improved the activity of both fluconazole and amphotericin B. The addition of exogenous biofilm ECM and commercial β-1,3 glucan also reduced the activity of fluconazole against planktonic C. albicans in vitro [bib_ref] Putative role of β-1,3 glucans in Candida albicans biofilm resistance, Nett [/bib_ref]. The group has recently shown that the ECM β-1,3 glucan is synthesized from Fks1 using a defined knockout and over-expressing strain [bib_ref] Genetic basis of Candida biofilm resistance due to drug-sequestering matrix glucan, Nett [/bib_ref]. This study demonstrated that β-1,3 glucan is responsible for sequestering azoles, acting as a sponge and conferring resistance on C. albicans biofilms [bib_ref] Genetic basis of Candida biofilm resistance due to drug-sequestering matrix glucan, Nett [/bib_ref]. Further studies have shown that they are also responsible for sequestering echinocandins, pyrimidines, and polyenes [bib_ref] Role of Fks1p and matrix glucan in Candida albicans biofilm resistance to..., Nett [/bib_ref]. Subsequent studies have identified a role for the SMI1 in C. albicans, a gene involved in cell-wall glucans, in biofilm ECM production and development of a drug-resistant phenotype, which appears to act through transcription factor Rlmp and glucan synthase Fks1. In addition to Fks1, a zinc-response transcription factor ZAP1 has been shown to be a negative regulator of ECM soluble β-1,3 glucan in both in vitro and in vivo C. albicans biofilm models [bib_ref] Interface of Candida albicans biofilm matrix-associated drug resistance and cell wall integrity..., Nett [/bib_ref]. Conversely, two glucoamylases, Gca1 and Gca2, are thought to have positive roles in matrix production. A group of alcohol dehydrogenases ADH5, CSH1, and LFD6 also have roles in matrix production, with ADH5 acting positively, and CSH1 and LFD6 acting negatively [bib_ref] Biofilm matrix regulation by Candida albicans Zap1, Nobile [/bib_ref]. It is also present on a number of other Candida species, including C. glabrata, C. parapsilosis and C. tropicalis [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref].
Recent studies revealed the involvement of the matrix on C. tropicalis strains on amphotericin B resistance [bib_ref] Candida tropicalis biofilm's matrix-involvement on its resistance to amphotericin B, Fernandes [/bib_ref]. These studies highlight the incapacity of this traditional antifungal to totally prevent biofilm formation and to eradicate C. tropicalis biofilms. Interestingly, it was observed that amphotericin B led to a significant increase of the biofilm production due to an augment of the total protein and carbohydrate contents of the matrix. [bib_ref] Effect of voriconazole on Candida tropicalis biofilms: Relation with ERG genes expression, Fernandes [/bib_ref] [bib_ref] Effect of voriconazole on Candida tropicalis biofilms: Relation with ERG genes expression, Fernandes [/bib_ref] revealed recently that voriconazole had no effect on pre-formed C. tropicalis biofilms. Remarkably, an increase in total biomass was observed when pre-formed biofilms were treated with this antifungal agent. This phenomenon is probably due to a response of C. tropicalis biofilm cells to the stress caused by the presence of the agent, which led to an expansion of the biofilm matrices' production. [bib_ref] Effects of fluconazole on Candida glabrata biofilms and its relationship with ABC..., Fonseca [/bib_ref] [bib_ref] Effects of fluconazole on Candida glabrata biofilms and its relationship with ABC..., Fonseca [/bib_ref] revealed a phenomenon similar for C. glabrata with an increase of proteins and carbohydrates in the matrices extracted from biofilms treated with fluconazole.
## Emergence of persister cells
An intriguing development in understanding Candida species biofilm resistance is the presence of persister cells. Persister cells are a subset of cells, dormant variants, which lie deep in a biofilm and exhibit tolerance to multiple antifungal drug classes [bib_ref] Candida albicans biofilms produce antifungal-tolerant persister cells, Lafleur [/bib_ref] [bib_ref] Multidrug tolerance of biofilms and persister cells, Lewis [/bib_ref]. LaFleur (2006) [bib_ref] Candida albicans biofilms produce antifungal-tolerant persister cells, Lafleur [/bib_ref] published the first study in fungi that demonstrated the presence of persister cells in C. albicans biofilms, but not in planktonic populations. In fact, a re-inoculation of cells, which survived from a biofilm treated with amphotericin B, was able to develop a new biofilm also with persister cells. This work suggested that these cells are not mutants but cells phenotype variants of the wild type. The presence of persisters in C. krusei, and C. parapsilosis biofilms treated with amphotericin B, were also described [bib_ref] Absence of amphotericin B-tolerant persister cells in biofilms of some Candida species, Al-Dhaheri [/bib_ref]. It was shown that the persister levels of the isolates varied from 0.2% to 9%, and strains isolated from patients with long-term carriage had high levels of persisters, whereas those from transient carriage did not [bib_ref] Candida albicans biofilms produce antifungal-tolerant persister cells, Lafleur [/bib_ref]. While the mechanisms of Candida persister cells transition remains unclear, transcriptional analysis of these cells shows differential regulation of genes involved in ergosterol (ERG1 and ERG25) and β-1,6 glucan (SKN1 and KRE1) pathways [bib_ref] A small subpopulation of blastospores in Candida albicans biofilms exhibit resistance to..., Knot [/bib_ref]. Moreover, superoxide dismutases (SOD) were found to be differentially expressed by miconazole-treated sessile C. albicans cells compared to untreated cells. Inhibition of SOD resulted in an 18-fold reduction of the miconazole-tolerant persister cells and increased endogenous reactive oxygen species (ROS) levels in these cells [bib_ref] Superoxide dismutases are involved in Candida albicans biofilm persistence against miconazole, Bink [/bib_ref]. [bib_ref] Superoxide dismutases are involved in Candida albicans biofilm persistence against miconazole, Bink [/bib_ref] [bib_ref] Superoxide dismutases are involved in Candida albicans biofilm persistence against miconazole, Bink [/bib_ref] also demonstrated that in biofilms from strains lacking sod4/sod5 at least three-fold less miconazole-tolerant persisters were observed, and ROS levels were also increased.
## Impact of sterols contents and its correlation with erg genes expression
Ergosterol is the most prevalent sterol in Candida cells plasma membrane. Moreover, antifungal agents (e.g., azoles and amphotericin B) act as ergosterol synthesis inhibitors by binding to lanosterol demthylase, a specific enzymatic in ergosterol biosynthesis. The observation that Candida mutants with altered ergosterol synthesis show enhanced resistance to azoles and amphotericin B led the investigators to question if Candida biofilm cells may employ similar mechanisms of resistance [bib_ref] Mechanism of fluconazole resistance in Candida albicans biofilms: Phase-specific role of efflux..., Mukherjee [/bib_ref]. Mukerjee and colleagues (2003) [bib_ref] Mechanism of fluconazole resistance in Candida albicans biofilms: Phase-specific role of efflux..., Mukherjee [/bib_ref] showed that, when comparing planktonic cells' membranes with the membranes of biofilm cells, the latter had a lower concentration of ergosterol, especially during the last steps of biofilm formation. This finding suggests that cells from mature biofilms rely less on ergosterol for maintaining its membrane fluidity and potentially limiting the efficacy of the ergosterol targeting drugs. In fact, several studies have demonstrated alterations in the transcriptional profile of sterol pathway genes in diverse Candida species [bib_ref] Genetic control of Candida albicans biofilm development, Finkel [/bib_ref]. Candida albicans microarray analysis demonstrated an increase of ERG25 and ERG11 in vitro biofilm growth when compared with its planktonic counterparts [bib_ref] Andes, D. β-1,3 Glucan as a test for central venous catheter biofilm..., Nett [/bib_ref]. Interestingly, transcriptional analysis of a rat venous catheter biofilm also found increased transcription of ERG25, but not ERG11 [bib_ref] Putative role of β-1,3 glucans in Candida albicans biofilm resistance, Nett [/bib_ref]. Moreover, the principal drug target, ERG11, can easily develop point mutations or even be over-expressed [bib_ref] Candida albicans drug resistance-Another way to cope with stress, Cannon [/bib_ref] [bib_ref] Increased mRNA levels of ERG16, CDR, and MDR1 correlate, with increases in..., White [/bib_ref].
These results confirm the involvement of the alterations on sterol content in membrane of C. tropicalis cells as in other Candida cells. Candida glabrata is assumed to be the most azole-resistant species of all Candida species [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref]. Besides, all the genes involved in the biosynthesis of ergosterol have been described as up-regulated in C. glabrata treated with azoles molecules in planktonic cells [bib_ref] Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal..., Silva [/bib_ref].
It is believed that the increase of C. glabrata infections is due to its intrinsically low susceptibility to azoles, including the imidazoles and the oral-parenteral triazoles (e.g., fluconazole, voriconazole) [bib_ref] Pathogenesis and antifungal drug resistance of the human fungal pathogen Candida glabrata, Tscherner [/bib_ref]. Additionally, it is known that the acquired resistance is resulted of rare mutations that are selected by drug pressure [bib_ref] An update on antifungal targets and mechanisms of resistance in Candida albicans, Akins [/bib_ref]. All the genes involved in the biosynthesis of ergosterol are expected to be up-regulated in the presence of azole molecules. Nevertheless, ERG genes are the ones more focused on in studies. Between them are ERG1, ERG3, ERG6, ERG7, ERG9, and especially ERG11. ERG11 is noticeably more referred as the central point on the increase of ergosterol production, in response to the azole attack to the C. glabrata cell membrane, which has great ease to acquire azole resistance [bib_ref] An update on antifungal targets and mechanisms of resistance in Candida albicans, Akins [/bib_ref] [bib_ref] Upregulation of ERG genes in Candida species by azoles and other sterol..., Henry [/bib_ref].
Induction of ergosterol genes has also been described in C. dubliniensis, where incubation with fluconazole and formation of biofilm was coupled with up-regulation of the ERG3 and ERG25 [bib_ref] The expression of genes involved in the ergosterol biosynthesis pathway in Candida..., Borecká-Melkusová [/bib_ref]. Moreover, up-regulation of genes involved with ergosterol biosynthesis has been described in C. parapsilosis biofilms [bib_ref] Correlation between biofilm formation and the hypoxic response in Candida parapsilosis, Rossignol [/bib_ref] , which are also resistant to azole antifungal therapy [bib_ref] Differential activities of newer antifungal agents against Candida albicans and Candida parapsilosis..., Katragkou [/bib_ref]. Regarding C. tropicalis, recently Fernandes and colleagues (2016) [bib_ref] Effect of voriconazole on Candida tropicalis biofilms: Relation with ERG genes expression, Fernandes [/bib_ref] demonstrated that, similar to C. albicans, voriconazole-resistant cells presented an increased on expression of ERG genes.
## Over-expression of other antifungal targets
Many cases of drug resistance are linked to the increase of efflux pumps in Candida cells membrane and the consequent reduction of the drug accumulation within the cells [bib_ref] The genetic basis of fluconazole resistance development in Candida albicans, Morschhäuser [/bib_ref]. In C. albicans, efflux pumps have been described as playing an important role in azole resistance, but not in resistance to amphotericin B and echinocandins [bib_ref] Regulatory circuitry governing fungal development, drug resistance, and disease. Microbiol, Shapiro [/bib_ref] [bib_ref] In vitro activity of caspofungin (MK-0991) against Candida albicans clinical isolates displaying..., Bachmann [/bib_ref] [bib_ref] Regulation of efflux pump expression and drug resistance by the transcription factors..., Schubert [/bib_ref] [bib_ref] Overexpression of Candida albicans CDR1, CDR2, or MDR1 Does Not Produce Significant..., Niimi [/bib_ref] [bib_ref] Characterization of caspofungin susceptibilities by broth and agar in Candida albicans clinical..., Silver [/bib_ref].
The ATP binding cassette transporters (CDR1 and CDR2) and major facilitator transporter (MDR1) are typically expressed at low levels in the absence of antifungal exposure [bib_ref] Overexpression of Candida albicans CDR1, CDR2, or MDR1 Does Not Produce Significant..., Niimi [/bib_ref]. The finding that azole-resistant clinical isolates often show constitutive over-expression of these pumps prompted investigators to postulate that the biofilm drug resistance phenotype may be related to increased efflux pump activity [bib_ref] Regulation of efflux pump expression and drug resistance by the transcription factors..., Schubert [/bib_ref] [bib_ref] Mechanisms of resistance to azole antifungal agents in Candida albicans isolates from..., Sanglard [/bib_ref] [bib_ref] Multiple efflux mechanisms are involved in Candida albicans fluconazole resistance, Albertson [/bib_ref] [bib_ref] A mutation in Tac1p, a transcription factor regulating CDR1 and CDR2, is..., Coste [/bib_ref] [bib_ref] Mutations in the multidrug resistance regulator MRR1, followed by loss of heterozygosity,..., Dunkel [/bib_ref]. [bib_ref] Fungal biofilm resistance, Ramage [/bib_ref] demonstrated that transcription of both MDR1 and CDR1 was more abundant in C. albicans biofilms than planktonic cultures of the same age. In opposition, some authors investigated the role of these efflux pumps by deletion of their genes and observed that, during the planktonic growth, these mutants displayed hypersensitivity to fluconazole. However, this phenotype was not observed when these same mutants were grown as biofilms, suggesting that the efflux pumps do not contribute significantly to drug resistance during the mature biofilm stage [bib_ref] Standardized method for in vitro antifungal susceptibility testing of Candida albicans biofilms, Ramage [/bib_ref]. [bib_ref] Mechanism of fluconazole resistance in Candida albicans biofilms: Phase-specific role of efflux..., Mukherjee [/bib_ref] examined the role of efflux pumps in antifungal resistance throughout the biofilm process. The researchers included early, intermediate and mature C. albicans biofilms with planktonic growth comparisons. Similar to the prior investigation, single, double and triple mutants of the three main efflux genes were no more susceptible to fluconazole treatment during mature biofilm growth than the parent strains; however, in the early phase, double and triple efflux pump mutants had significantly increased azole susceptibility when compared with the parent strains. This suggests that the efflux pumps contribute to resistance during the early biofilm developmental phase, and that the pumps may function in a cooperative manner. This theory of time-specific efflux pump functionality was further supported by transcriptional analysis, showing higher expression of efflux pump genes after 12 h biofilm formation when compared with mature, 48 h biofilm formation [bib_ref] Mechanism of fluconazole resistance in Candida albicans biofilms: Phase-specific role of efflux..., Mukherjee [/bib_ref]. This is collective evidence that Candida efflux pumps likely contribute to drug resistance during the early phase of biofilm growth, while their role in resistance in mature biofilms appears to be minimal at most. Investigations of C. glabrata and C. tropicalis biofilms have also shown up-regulation of efflux pumps [bib_ref] Fungal biofilm resistance, Ramage [/bib_ref] [bib_ref] Characteristics of biofilm formation by Candida tropicalis and antifungal resistance, Bizerra [/bib_ref]. [bib_ref] Effects of fluconazole on Candida glabrata biofilms and its relationship with ABC..., Fonseca [/bib_ref] evaluated the expression of C. glabrata ABC (ATP-binding cassette) transporters (CDR1, SNQ2 and PDR1) in presence of fluconazole, and observed that, in addition to high amounts, the matrix produced an over-expression of these efflux pumps.
This data supports the hypothesis that efflux pumps are an important-but not exclusive-determinant of fungal biofilm resistance to antifungal drugs. Their primary role may be for homeostasis within complex environments to protect themselves from acute toxicity, but within clinical environments, exposure to azoles drugs may enhance the levels of efflux pump expression, therefore either contributing towards or inducing clinical resistance.
# Conclusions
Reducing the incidence of biofilm-related candidiasis in hospitals is a requirement in the search for optimized patient care. However, the high degree of resistance of biofilm-associated Candida cells hinders rapid development toward highly efficacious therapies. Recent efforts of various excellent research groups tremendously broadened our knowledge on the complex mechanisms underlying biofilm resistance. According to the authors, the presence of matrix material is the most important biofilm resistance mechanism. However, several other important mechanisms such as cell density, differential regulation of drug targets, up-regulation of drug efflux pumps in developing biofilms, the presence of persisters into biofilms, up-regulation of different pathways and possibly yet-undefined mechanisms can further increase resistance to a maximum level. The elucidation of these resistance mechanisms provides a promising step toward the development of optimal therapies.
[fig] Figure 1: General scheme of the mechanisms described as involved on Candida species biofilm resistance. [/fig]
[fig] 36: Komshian, S.V.; Uwaydah, A.K.; Sobel, J.D.; Crane, L.R. Fungemia Caused by Candida Species and Torulopsis glabrata in the hospitalized patient: frequency, characteristics, and evaluation of factors influencing outcome. Clin. Infect. Dis. 1989, 11, 379-390. [CrossRef] 37. Walter, E.B.; Gingras, J.L.; McKinney, R.E. Systemic Torulopsis glabrata infection in a neonate. South. Med. J. 1990, 83, 837-838. [CrossRef] [PubMed] 38. Fortún, J.; Martín-Dávila, P.; Gómez-García de la Pedrosa, E.; Pintado, V.; Cobo, J.; Fresco, G.; Meije, Y.; Ros, L.; Alvarez, M.E.; Luengo, J.; et al. Emerging trends in candidemia: A higher incidence but a similar outcome. J. Infect. 2012, 65, 64-70. [CrossRef] [PubMed] 39. Ding, C.; Vidanes, G.M.; Maguire, S.L.; Guida, A.; Synnott, J.; Andes, D.; Butler, G. Conserved and divergent roles of Bcr1 and CFEM proteins in Candida parapsilosis and Candida albicans. PLoS ONE 2011, 6. [CrossRef] [PubMed] 40. Mohan, P.; Holland, L.; Butler, G.; Gacser, A.; Bliss, J.M. Candida parapsilosis is a significant neonatal pathogen: A systematic review and meta-analysis. Pediatr. Infect. Dis. J. 2013, 32, 206-216. 41. Wada, M.; Baba, H.; Imura, S. Prosthetic knee Candida parapsilosis infection. J. Arthroplast. 1998, 13, 479-482. [CrossRef] 42. Fox, P.M.; Lee, G.K. Tissue expander with acellular dermal matrix for breast reconstruction infected by an unusual pathogen: Candida parapsilosis. J. Plast. Reconstr. Aesthet. Surg. 2012, 65, e286-e289. [CrossRef] [PubMed] 43. Younkin, S.; Evarts, C.M.; Steigbigel, R.T. Candida parapsilosis infection of a total hip-joint replacement: Successful reimplantation after treatment with amphotericin B and 5-fluorocytosine. A case report. J. Bone Jt. Surg. Am. 1984, 66, 142-143. [CrossRef] [PubMed] 44. Mansur, A.J.; Safi, J.; Markus, M.R.P.; Aiello, V.D.; Grinberg, M.; Pomerantzeff, P. Late failure of surgical treatment for bioprosthetic valve endocarditis due to Candida tropicalis. Clin. Infect. Dis. 1996, 22, 380-381. [CrossRef] [PubMed] 45. Thomas, B.; Thomas, F.F. Oral biofilm-associated diseases: Trends and implications for quality of life, systemic health and expenditures. Periodontol 2000 2011, 55, 87-103. 46. Rautemaa, R.; Ramage, G. Oral candidosis-clinical challenges of a biofilm disease. Crit. Rev. Microbiol. 2011, 37, 328-336. [CrossRef] [PubMed] 47. Dongari-Bagtzoglou, A.; Kashleva, H.; Dwivedi, P.; Diaz, P.; Vasilakos, J. Characterization of mucosal Candida albicans biofilms. PLoS ONE 2009, 4. [CrossRef] [PubMed] 48. Ramage, G.; Tomsett, K.; Wickes, B.L.; López-Ribot, J.L.; Redding, S. Denture stomatitis: A role for Candida biofilms. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endodontol. 2004, 98, 53-59. [CrossRef] 49. Sardi, J.C.O.; Duque, C.; Mariano, F.S.; Peixoto, I.T.; Höfling, J.; Gonçalves, R. Candida spp. in periodontal disease: A brief review. J. Oral Sci. 2010, 52, 177-185. [CrossRef] [PubMed] 50. Carvalho, F.G.; Silva, D.S.; Hebling, J.; Spolidorio, L.C.; Spolidorio, D.M. Presence of mutans Streptococci and Candida spp. in dental plaque/dentine of carious teeth and early childhood caries. Arch. Oral Biol. 2006, 51, 1024-1028. [CrossRef] [PubMed] 51. Gündüz Arslan, S.; Akpolat, N.; Kama, J.D.; Özer, T.; Hamamci, O. One-year follow-up of the effect of fixed orthodontic treatment on colonization by oral Candida. J. Oral Pathol. Med. 2008, 37, 26-29. [CrossRef] [PubMed] 52. Canabarro, A.; Valle, C.; Farias, M.R.; Santos, F.B.; Lazera, M.; Wanke, B. Association of subgingival colonization of Candida albicans and other yeasts with severity of chronic periodontitis. J. Periodontal Res. 2012, 48, 428-432. [CrossRef] [PubMed] 53. Bamford, C.V.; D'Mello, A.; Nobbs, A.H.; Dutton, L.C.; Vickerman, M.M.; Jenkinson, H.F. Streptococcus gordonii modulates Candida albicans biofilm formation through intergeneric communication. Infect. Immun. 2009, 77, 3696-3704. [CrossRef] [PubMed] 54. Silverman, R.J.; Nobbs, A.H.; Vickerman, M.M.; Barbour, M.E.; Jenkinson, H.F. Interaction of Candida albicans cell wall Als3 protein with Streptococcus gordonii SspB adhesin promotes development of mixed-species communities. Infect. Immun. 2010, 78, 4644-4652. [CrossRef] [PubMed] 55. Coulthwaite, L.; Verran, J. Potential pathogenic aspects of denture plaque. Br. J. Biomed. Sci. 2007, 64, 180-189. [CrossRef] [PubMed] 56. Nett, J.E.; Marchillo, K.; Spiegel, C.A.; Andes, D.R. Development and validation of an in vivo Candida albicans biofilm denture model. Infect. Immun. 2010, 78, 3650-3659. [CrossRef] [PubMed] [/fig]
[table] Table 1: Characteristics of the most common Candida species biofilms. [/table]
|
Supramolecular chemistry: from aromatic foldamers to solution-phase supramolecular organic frameworks
This mini-review covers the growth, education, career, and research activities of the author. In particular, the developments of various folded, helical and extended secondary structures from aromatic backbones driven by different noncovalent forces (including hydrogen bonding, donor-acceptor, solvophobicity, and dimerization of conjugated radical cations) and solution-phase supramolecular organic frameworks driven by hydrophobically initiated aromatic stacking in the cavity of cucurbit[8]uril (CB[8]) are highlighted.2057
I was born on July 23rd, 1966 in the small, remote village of Fang-Liu (a combination of two common Chinese family names), which is located in Shangcai County in the Henan Province of central China. Living in the distant countryside, my childhood was simple and quiet. Later I learned that from 1966 to 1976 China had seriously suffered from the so-called Cultural Revolution. In 1972 I entered primary school. My class had about fifteen students. We were all children from the same village. For the first two years, we had only one teacher, Xi Liu, who taught us Chinese and arithmetic. Once per week he also taught music, mainly singing. Since we did not have a permanent classroom, Mr. Liu frequently moved the class to different empty rooms that he found in the village. As I remember it, I studied in at least four "classrooms". So every day we all needed to bring a small stool from home to "go to school", but I did not feel this was a burden. Actually this was possibly the happiest stage of my life because Mr. Liu was a neighbor and a family friend, and my scores in Chinese and arithmetic were always the best. In addition to going to school, I also spent a lot of time reading Chinese novels that I could find after I was able to read. Although there were no opportunities for modern sports such as basketball and football, I liked playing ground chess, a two-person game that was very popular during poor times but has now nearly completely disappeared. I enjoyed the chess game very much because I could often win even against adults. However, I never dreamed that I would become a chemistry professor in Shanghai many years later.
In 1979, I entered Caigou High School, which was located in Caigou Town one kilometer away from my home village. In 1977, China reinstated the long suspended national college entrance examination, thanks to the comeback of Mr. Xiaoping Deng, the greatest Chinese leader. For young people living in the rural areas, acceptance into college was the only route to change their fate at that time. Therefore, I studied very hard and again got the highest total score in the school in the national college entrance examination in 1981. In autumn of that year, I enrolled in the Department of Chemistry at Zhengzhou University. I chose chemistry as my major only because my chemistry score was 95 (out of 100), which was higher than my math or physics scores. I had no idea on how a university picked its students, but just instinctively believed choosing chemistry meant a better chance for being admitted. After more than 30 years, I still remember the days when I studied in Caigou High School. It was quite good at that time in the county, but it could not attract teachers and students in later years due to its remote location and was closed about ten years ago by the local government.
## Studying at zhengzhou university
Zhengzhou is the capital of the Henan Province. Being admitted to the University gave me the opportunity to leave Shangcai County and to take a train for the first time. The Chemistry Department enrolled a total of 120 undergraduates in that year. In the 1980s, the department had only the essential courses of inorganic, organic, physical and structural chemistry and chemical engineering, which were accompanied by a series of fundamental experiments in the same semester. This was the typical course system of the day in China and so all students received the same training. Many years later, China introduced the credit system and both mandatory and optional courses were offered. By the end of the third year, students were required to choose one of the above "secondary degree" disciplines. I chose organic chemistry because I felt it was the easiest course among the others. In particular, I enjoyed one course called "Organic Synthesis Skills" taught by Professor Zhixin Huang, which introduced multistep synthesis. I must say that since 1987 I have benefited greatly from this course as a graduate student at the Shanghai Institute of Organic Chemistry (SIOC). In the spring of 1985, the second semester of the fourth year, I entered Professor Zhendong Chen's lab to perform my dissertation research. I stayed in the lab for about three months and prepared several ferrocene-derived conjugated molecules to study the homologous linearity. I later realized that this was an important research project in physical organic chemistry in China in the 1980s. After graduation from Zhengzhou University, I was assigned to work at Henan Medical University as a teaching assistant. Two years later, I was enrolled as a graduate student at SIOC, the top research institution for organic chemistry in China.
Studying at SIOC: N···I interaction I joined SIOC in late August of 1987. For many years, graduate students in the first year studied physical organic chemistry, organic synthesis, and organic analytical chemistry. Impressively, all students performed 6-8 multistep synthesis experiments that involved the use of all standard organic synthesis techniques. This was a challenge to many students, but all received systematic training in organic synthesis upon completion of the experiments. In 1988, I entered Professor Ching-Sung Chi's group in the Laboratory of Organic Fluorine Chemistry for a master's degree. The lab is well-known for its longstanding research on reactions of fluorine-containing molecules. Professor Chi left as a visiting scholar at the University of Fribourg, Switzerland shortly after I joined the group and I was actually advised by Professor Yong-Da Lin. I finished the synthesis of several fluorine-containing macrocycles and published my first research paper in the journal Heterocycles. The starting materials for these macrocycles were initially designed for the preparation of biologically active molecules, which was the main project in this laboratory. Using them to prepare new macrocycles became a small independent dissertation project for me. In 1990, I joined Professor Qing-Yun Chen's group as a Ph.D. candidate. Professor Chen is a distinguished, esteemed Chinese chemist in organic fluorine chemistry. His group developed new trifluoromethylation reagents, which found many practical applications. I respect him for his persistent passion for science. Even at the age of 86 in 2015, he still goes to his office and advises his students and remains active in the field of fluorine chemistry. I received my Ph.D. degree in December of 1992. My dissertation research focused on photoinduced reactions of perfluoroalkyl iodides and pentafluoroiodobenzene with arenes, aromatic ethers and amines and heterocycles. These reactions led to the perfluoroalkylation or pentafluorophenylation of the aromatic compounds. One series of reactions involved liquid tetrafluoro-1,2-diiodoethane (1a) or dodecafluoro-1,6-diiodohexane (1b) and solid N,N,N',N'tetramethylphenylene-1,4-diamine. We found that when mixing in chloroform, the 1:1 mixtures readily gave a high yield of solid adducts, which melted at 85 and 65 °C, respectively. Elemental analyses supported a 1:1 stoichiometry for the solid adducts, andF NMR spectra showed downfield chemical shifting of the ICF 2 signal of 1a and 1b. Clearly, an important intermolecular interaction occurred which led to the solidification of the mixtures. However, we did not obtain the crystal structure of the mixtures, although I had been to Beijing for X-ray diffraction experiments at Peking University. We thus proposed that the two compounds formed 1:1 charge-transfer complexes (Scheme 1). Currently, this N···I interaction is termed as halogen bonding, which is widely used in supramolecular crystal engineering. My Ph.D. training in synthetic methodology and fluorine chemistry had an important influence on my research activity. When I initiated a project, the first thing I would think of and discuss with my students is the synthetic route for the target molecules. Fluorine-containing molecules have always been my favorite. When I performed my postdoctoral research in Denmark, I used fluorine-containing precursors to build catenanes, and many years later, I utilized fluorine as a hydrogen bonding acceptor to develop aromatic amide and triazole foldamers.
## Postdoctoral research at the university of south denmark: donor-acceptor interactiondriven catenanes
From October 1994 to December 1995, I performed postdoctoral research with Professor Jan Becher at Odense University (currently University of South Denmark) in Denmark. Through his research career, Professor Becher studied sulfur-containing molecules and since the early 1990s, tetrathiafulvalene (TTF) supramolecular chemistry. Before I went to Odense, the group had developed a very useful method of in situ generation of TTF thiolate anion from cyanoethylated precursors, which greatly simplifies the modification of the TTF core. Using this method, I could prepare bimacrocycle 3 in a short time. This macrocycle was used to template the formation of the so-called tetracationic "blue box"from 4 and 5 to give rise to the unique pseudocatenanes 6a and 6b (Scheme 2), together with a trace amount ofcatenane 6c with the tetracationic cyclophane holding one of the peripheral benzene rings. The TTF unit in 6a and 6b adopted a stable cis or trans configuration, although typically the two configurations easily isomerize into each other in solution. Macrocycle 3 is a brown solid due to the existence of the TTF unit. Catenanes 6a and 6b are blue as a result of the charge-transfer complex between the TTF and bipyridinium units, whereas catenane 6c is orange, which is attributed to the charge-transfer complex between the dioxybenzene and bipyridinium units. Impressively, the three catenanes could be separated from each other using a short silica gel column. Three colorful bands could be observed from the column, which made it easy to collect the respective solutions. Another impressive occurrence was the difference in the solubility of the catenanes with different counterions in water and organic solvents. With hexafluorophosphate as the counterion, the catenanes were well-dissolved in acetonitrile, while with chloride, their solubility was poor.
My stay in Odense initiated my research in supramolecular chemistry. For many years, I maintained interest in interlocked systems. TTF is also a favorite. Many years later at Fudan, I initiated a project to study the potential of its radical cation stacking in controlling the folded conformation of linear molecules and two-and three-dimensional supramolecular polymers and frameworks. My life in the small town of Odense was also memorable. Its calm is in sharp contrast to the bustle of Shanghai where I had lived since 1987. Odense is also the hometown of Hans C. Andersen, the great Danish writer of fairy tales. Since I had read his books before, I visited the small Hans Christian Andersen Museum, which is situated in a house in the old town where he was born.
## Early research at sioc
In January 1996, I returned to SIOC. Although I had hoped to continue research in fluorine chemistry, I was assigned to work for a big contract project in the laboratory of physical-organic chemistry. After finishing the contract research, I also did a small project: constructing calixarene-derived catenanes, such as 7a-c(Scheme 3). I chose this project because I believed that I could make progress in a short time. Most of the work was done by myself, but I also received assistance from several young technicians. We succeeded in investigating the effect of the calixarene moiety on the relative rotation of the two rings using 1 H NMR, and from 1998 to 2000, we published four papers from this project. However, this was generally a tough period for me. Due to some nonacademic reasons, I could not build an efficient research group after several years of effort. Therefore, I decided to make a change.
## Research at the university of illinois
In October 2000, I joined Professor Steven C. Zimmerman's group at the University of Illinois at Urbana. The Zimmerman group had been well-known for pioneering works in hydrogen bonding-related achievements. In 1998, the group reported the extremely stable quadruple hydrogen-bonded dimers 8·8(Scheme 4). By using an 1 H NMR dilution technique, a lower limit to the dimerization constant K dim (>10 7 M −1 ) was estimated. Professor Zimmerman hoped to make an accurate determination of the binding stability. Thus, I prepared compound 9, which bore a pyrene unit(Scheme 4), based on previously published research by Sijbesma and co-workers, who had taken advantage of the excimer signal of the pyrene dimer to evaluate the stability of their famous quadruple hydrogen bonded 2-ureido-4[1H]-pyrimidinone dimers. The excimer exhibits an emission band at 500-600 nm, which is separated Dimer 8·8 had another three tautomers. Thus, I also tried to obtain the single crystal structure of the binding motif. In total I prepared 28 derivatives by introducing different substituents and succeeded in growing the single crystal of compound 10(Scheme 4). The crystal structure showed the formation of a homodimer of the N(3H) protomer that was stabilized by four intermolecular N-H···O hydrogen bonds and the protomer itself was rigidified by an intramolecular six-membered N-H···O hydrogen bond [fig_ref] Scheme 1: Proposed structures of complexes between 1a and 1b with 2 [/fig_ref]. This motif was the one that had the highest proportion in chloroform.
## Research on foldamers: applications for molecular recognition and self-assembly
Donor-acceptor interaction and π-stacking for folding. In January 2001, I returned to SIOC again. When I left SIOC for Urbana in 2000, I had built a small research group with two graduates. Thanks to the persistence of Professor Xi-Kui Jiang, one of the greatest physical organic chemists in China, my small group survived until I returned to the institute. I was so impressed by the work of Professor Zimmerman on the folding of linear urea derivatives driven by intramolecular hydrogen bonding, that in 2001, our group started several projects searching for new folded frameworks. All the projects were done by Ph.D. students. Xin Zhao, who is currently a professor at SIOC, finished the first project in 2004. He prepared L-ornithine-derived δ-peptides 11a-g, which bore one to three electron-deficient pyromellitic diimide (PDI) and electron-rich 1,5-dioxynaphthalene (DAN) units on the two sides of the backbones(Scheme 5). An intramolecular donor-acceptor interaction between the DAN and PDI units, which is a well-established noncovalent force, induced the backbones to fold into zipper-featured foldamers in less polar chloroform and polar DMF [fig_ref] Scheme 2: The formation of catenanes 6a-c [/fig_ref]. This finding was supported by 1 H NMR, UV-vis, and fluorescence quenching studies. As expected, the folding state became more compact for longer sequences, which possess more donor-acceptor interacting sites. UV-vis experiments indicated that the folding state remained, even at 150 °C in DMF.
## Figure 2:
Zipper-featured folding motif of δ-peptides 11a-g driven by the cooperative donor-acceptor interaction.
Encouraged by the above work, Junli Hou, who is currently a professor at Fudan University, prepared naphthalene-incorporated oligo(ethylene glycols) 12a-g(Scheme 6). UV-vis, 1 H NMR, and fluorescence experiments in chloroform-acetonitrile binary solvents revealed that the naphthalene units in longer 12f-h stacked intramolecularly to induce the oligomeric chains to form a helical conformation at high acetonitrile content (Scheme 6). The compact helical conformation gave rise to a cavity similar to that of 18-crown-6 and thus could complex ammonium or ethane-1,2-diaminium in acetonitrile. The stability of the complexes increased with the elongation of the ethylene glycol chains.
## Aromatic amide oligomers: extended secondary structures.
Our students also tried to make use of hydrogen bonding to control the conformation of aromatic amide backbones. Previously, Hamilton, Gong, Huc and Lehn, and Huchad developed several series of elegant folded frameworks. We thus focused on the creation of extended frameworks, which we expected to be useful for the design of functional materials. In 2004, Zongquan Wu, who is currently a professor at Hefei University of Technology, reported the formation of straight conformations [fig_ref] Scheme 1: Proposed structures of complexes between 1a and 1b with 2 [/fig_ref] , which was driven by successive intramolecular hydrogen bonding(Scheme 7), whereas Jiang Zhu, who is currently an associate professor at North Sichuan Medical College, described the zigzag conformation of oligomers 15a-d, which is also stabilized by intramolecular hydrogen bonding(Scheme 7).
The aromatic units in oligomers 13-15 can be easily combined into one sequence. By changing their number and position, sequences of different length and shape or in a controlled conformation can be designed. Thus, oligomers 13-15 may be considered as structural prototypes for creating new modifiable backbones. For example, porphyrin-appended U-shaped molecular tweezers 16 and 17 have been produced (Scheme 8).
Compound 16 complexed C 60 or C 70 or their derivatives in chloroform or toluene through porphyrin-C 60 stacking, while compound 17 strongly complexed 18 in the chloroform-acetonitrile binary medium(Scheme 8). Because 18 could further form a threaded complex with 24-crown-8 19 driven by multiple O···H-N hydrogen bonds, the three components self-assembled into a unique dynamiccatenane. Under low temperature, this dynamiccatenane could be quantitatively generated. The intramolecular hydrogen bonds formed by the aromatic amide linkers remarkably Jiang further introduced amide subunits to the para-position of the ether groups of the benzamide rings of oligomers 15a and 15b to produce 20a and 20b, respectively (Scheme 9).
## H nmr dilution experiments in chloroform-d revealed that
both compounds formed stable homodimers 20a·20a and 20b·20b, with K dim being 3.0 × 10 3 and 2.3 × 10 5 M −1 , respectively. In contrast, even in nonpolar benzene-d 6 , the K dim for the dimerization of benzamide was only 40 M −1 . The result again shows that the intramolecular hydrogen bonding of the aromatic amide backbones promoted the appended amide subunits to bind in a cooperative manner by preorganizing the backbones.(Scheme 10). These two series of oligomers folded into helical secondary structures driven by intramolecular hydrogen bonding, which is typical for aromatic amide backbones . The diamine and diacyl chloride precursors for 21 had been used by Gong and co-workers to prepare the first family of hydrogen bonding-promoted aromatic amide macrocycles. The main aim for designing these folded structures was to explore their potential functions as acyclic receptors. Moore et al. had utilized this approach to investigate the binding of m-phenylene ethynylene foldamers for nonpolar organic molecules in polar media. All the C=O oxygen atoms of 21 point into the cavity of the helix, which has a diameter of approximately 0.8 nm. Thus, 21 complexed alkylated saccharide derivatives and a guest with three hydroxyl groups in chloroform. The binding also induced the backbone of 21 to produce helicity bias. The methoxy groups of 22a-c are all located inwards. These oxygen atoms are potential hydrogen bonding acceptors. 1 H NMR and fluorescence experiments in chloroform showed that this series of foldamers complexes primary and secondary alkyl ammonium products.
In most cases, we investigated the binding of foldamers for different guests in less polar chloroform. However, during the synthesis of the 22 series, which involved the hydrolysis of the methyl ester at one end with lithium hydroxide in heated dioxane-water solution, Huiping found that the nitro-bearing anisole unit at the other end was also hydrolyzed to afford a phenol derivative. In contrast, a short control did not exhibit similar reactivity. These observations suggested that the folded conformation, such as in 23, promoted the hydrolysis of the anisole, which we ascribed to the complexation of the foldamer to Li + cation (Scheme 11). We proposed that this complexation increased the efficient concentration of the OH − anion and also activated the nitro-bearing anisole. This result supports that the intramolecular MeO···H-N hydrogen bonding worked even in a highly polar solvent. The theoretical study by Pophristic and co-workers also shows that this hydrogen bonding exists, to some extent, in polar aqueous environments. The high stability of the folded conformation of the 22 series also make them useful frameworks for creating a variety of complicated supramolecular architectures. Although alkoxyl groups had been popularly used as acceptors for intramolecular hydrogen bonding , I was also interested in looking for other substitutes. For this purpose, I envisioned fluorine might be a good candidate, given its highest electronegativity. An investigation by Dunitz and Taylor led to the conclusion that "organic fluorine hardly ever accepts hydrogen bonds, that is, it does so only in the absence of a better acceptor". We conjectured that fluorine might work as an acceptor for intramolecular five-and six-membered F···H-N hydrogen bonding for aromatic amides because of their co-planarity. Chuang Li thus prepared oligomer 24 and shorter analogues. Systematic 1 H NMR experiments in chloroformd and crystal structure analysis of model molecules all supported that fluorine was engaged in the expected intramolecular F···H-N hydrogen bonding and 24 formed a helical conformation (Scheme 12). Moreover, Chuang found that intermolecular F···H-N hydrogen bonding could also be formed between 23 and aliphatic ammonium and, at high concentrations in chloroform, a chiral aliphatic ammonium induced 23 to produce helicity bias. One important difference between fluorine and alkoxy groups is that fluorine does not cause a steric effect on aromatic stacking. Zeng and co-workers demonstrated this difference by crystallizing a fluorine-bearing pentagon macrocycle . No stacking was observed for the methoxy-derived pentagon macrocycle. In contrast, their fluorine-engaged pentagon macrocycle stacked strongly. Moreover, the macrocycle stacked in the 2D space to give rise to the mathematically predicted, most densely packed lattice for a C 5 -symmetric macrocycle [49].
Jiang and co-workers also prepared fluorine-containing quinoline oligoamides, which they found self-assembled into unique double and quadruple helices. The fluorine atoms in the oligomers induced the backbone to fold by forming intramolecular F···H-N hydrogen bonding, whereas the small size of fluorine allowed the folded sequences to stack into double and quadruple helices. In an elegant study, they prepared fluorinecontaining hybrid sequences with different aromatic subunits, whose helical states could hold one linear molecule to give rise to foldamer-derived dynamic rotaxanes.
Aromatic hydrazide foldamers: Aromatic hydrazides have a high propensity towards co-planarity. Nowick introduced this unit into peptide backbones to increase the stability of artificial β-sheets. Junli hoped to extend the backbones of hydrogen-bonded foldamers and thus prepared oligomers 24a-c(Scheme 13). The octoxy groups provided solubility in common solvents such as chloroform. These and the methoxy groups formed successive hydrogen bonds to induce the backbones to form folded conformations, which have a cavity of about 1 nm in diameter. These folded structures could host alkylated saccharides in chloroform, which was stabilized by intermolecular hydrogen bonding formed between the carbonyl oxygen atoms of the foldamers and the hydroxy groups of the saccharides. By changing the position of the alkoxy groups, the shape of the backbones can be readily tuned. The large cavity tolerates the existence of the methoxy groups for the backbones to stack efficiently. Thus, the backbones can stack to form vesicles or gelate organic solvents depending on the appended side chains. Jiang and co-workers prepared methoxy-free backbones. It is expected that these backbones should exhibit increased conformational flexibility. However, they still displayed quite strong binding capacity to anions and saccharides.
## Scheme 13: the structures of foldamers 24a-c.
Studies on N-H····X (X = F, Cl, Br, I) hydrogen bonding. At SIOC, our group is the only one dedicated to physical-organic chemistry. Thus, we maintained a longstanding interest in the fundamental aspects of noncovalent forces. The establishment of intramolecular N-H····F hydrogen bonds raised the possibility of quantitative assessment of intermolecular N-H····F hydrogen bonds. Since this hydrogen bonding is generally weak, Yanhua Liu prepared compounds 25-27(Scheme 14), which have the identical backbones, in order to evaluate the stability of the heterodimers formed between them and 28. In chloroform, K dim for the three complexes was determined to be 11.2, 8.2 and 5.5 M −1 , respectively. In less polar binary chloroform-benzene (1:4, v/v), the K dim values of the first and third complexes were increased significantly to 25.8 and 19.3 M −1 , respectively. These results indicate that the intermolecular N-H····F hydrogen bond did exist for aromatic derivatives, but is quite weak.
After the establishment of the intramolecular N-H····F hydrogen bond motif, we further prepared various model molecules to exploit the possibility of forming similar N-H····X (X = Cl, Br, I) hydrogen bonding motifs by aromatic amide derivatives. We found that all these halogen atoms are able to form this hydrogen bonding, but the stability decreases successively, which is consistent with the decrease of their electronegativity, but may also reflect the increase of the van der Waals radius. These weak intramolecular hydrogen bonds cannot compete with the strong intermolecular N-H···O=C hydrogen bonds of the amide groups. Thus, for their formation, the latter has to be suppressed. For the same halogen atom, the five-membered N-H····X (X = Cl, Br, I) hydrogen bond is generally easier to form than the six-membered one. One straightforward explanation for this difference is that the formation of the former would confine the rotation of one single bond, while for the latter, it would confine two single bonds. Jiang and Huc and co-workers successfully utilized the five-membered N-H···Cl hydrogen bond to construct quinoline amide-derived double helices.
C-H····X (X = OR, F) hydrogen bonding-driven 1,2,3-triazole foldamers. In 2008, several groups independently described that the intermolecular C-H····Cl − hydrogen bonding could induce benzene-linked 1,2,3-triazole oligomers to form folded or helical conformations. Currently, this family of foldamers have found wide applications in anion binding and design of photo-active molecular devices. We were interested in developing inherently folded structural patterns for aromatic oligotriazole backbones. In 2009, Yuanyuan Zhu established that 1,5-diphenyl-1,2,3-triazole formed an intramolecular six-membered C-H····OMe hydrogen bonds. In 2012, Beye Lu further demonstrated that fluorine, chlorine or even bromine could form similar intramolecular C-H····X hydrogen bonds. Using the C-H····O hydrogen bonding motif, Liyan You created a family of oligotriazole foldamers, including 29 (Scheme 15), which had a cavity of approximately 1.8 nm in diameter. At first we expected that this series of triazole foldamers would be good hydrogen bonding acceptors. Intriguingly, the foldamers did not exhibit observable binding affinity to saccharides or amide derivatives even in less polar chloroform. Liyan designed different molecules to evaluate their binding to these triazole foldamers. He finally found that these foldamers were good halogen bonding receptors for tritopic and ditopic guests, including 30 in dichloromethane or its mixture with hydrocarbons. Yanhua further utilized the C-H····F hydrogen bonding to induce the folding of the same series of benzene/triazole oligomers. Recently, Jiang and co-workers reported that when fluorine or chlorine was introduced to the 2-position of the meta-substituted benzene linkers, the corresponding triazole oligomers folded to give new foldamers, which were stabilized by successive intramolecular C-H····F or C-H····Cl hydrogen bonds.
Conjugated radical cation dimerization-driven pleated foldamers. The stacking of the radical cations of viologen or TTF were observed in 1964 and 1979. This stacking is typically weak. Several approaches have been developed to enhance this stacking. As a result, strong stacking, which leads to the formation of stable homodimers, has been observed in many elegantly designed molecules and supramolecules. In recent years, this stacking has been utilized as a tunable, noncovalent force to induce the formation of interlocked systems and molecular switches. Our longstanding interest in foldamers prompted us to explore the application of this noncovalent force for the generation of new, folded patterns. Thus, Lan Chen prepared polymers P31a and P31b from the corresponding dialdehyde and di(acylhydrazine) precursors by forming dynamic hydrazone bonds(Scheme 16). This dynamic covalent chemistry approach allowed for quick synthesis of viologen/TTF-alternating polymers. Driven by the intramolecular donor-acceptor interaction between the TTF and viologen units, the polymers folded into pleated conformations in acetonitrile. Upon oxidation of the TTF units to radical cation TTF ·+ , the polymers adopted flexible conformations. When the viologen units were reduced to radical cations, the radical cations stacked intramolecularly to induce the backbone to form another kind of pleated secondary structure. Yunchang Zhang further illustrated that upon reducing the viologen units into radical cations, their intramolecular stacking also induced polymers P32a-d to form pleated conformations(Scheme 16). This occurred despite the fact that the folding of P32a needed the assistance of alkaline metal ions like Li + .
## Solution-phase supramolecular metal-organic frameworks (mofs)
Periodicity is the key feature of single crystals in which molecules arrange repeatedly in the three-dimensional space. Porous crystals such as metal-organic frameworks (MOFs) exhibit many unique properties mainly due to their large surface area. However, the achievement of periodicity is a challenge in solution for self-assembled architectures as a result of the impact of the solvent and the weakness of noncovalent forces that hold monomeric components together. In early 2012, I discussed the initiation of a project with Kangda Zhang to explore the possibility of generating honeycomb, supramolecular frameworks in water. In a short time, he prepared one triangular target molecule. However, the molecule was poorly soluble in water. He then prepared compound 33by introducing three hydrophilic amide chains, which provided good solubility in water (Scheme 17). Shortly after, Kangda revealed that the three phenyl-bipyridine units strongly stacked in the two-dimensional space when mixing with 1.5 equiv of cucurbituril (CB, which could encapsulate the stacked bipyridine dimers. In collaboration with Yi Liu at Lawrence Berkeley National Laboratory in the United States, we conducted solution-phase, small-angle X-ray scattering synchrotron experiments on the mixture of 33 and CBin water (1:1.5, 3.0 mg/mL), which supported the periodicity of the 2D honeycomb SOF structures in solution. Liang Zhang found that the stacking of viologen radical cations could also drive a similar triangular molecule to form 2D SOF, which was further stabilized with CBby encapsulating the stacking radical cation dimer. Very recently, Zhao and co-workers reported that a donor-acceptor interaction could drive the formation of square-shaped 2D MOFs from a porphyrin-derived bipyridinium precursor and 2,6-dioxynaphthalene-derived ditopic precursor. In this case, the 2,6-dioxynaphthalene-dipyridinium donor-acceptor complex was also stabilized by CB. These results showed that generation of 2D SOFs can be induced by discrete noncovalent forces.
In 2014, Jia Tian further extended the concept of solution-phase SOF to 3D space from the self-assembly of tetrahedral 34 and CB[86] (Scheme 18). The 1:2 mixture in water generated a 3D, homogeneous SOF of diamond topology. The pores of the 3D SOF could be observed by high-resolution TEM. As a supramolecular "ion sponge", the framework adsorbed various anionic guests, including drugs, peptides and DNA. These new 2D and 3D homogeneous SOFs all have defined cavities or pores, which are expected to display new, interesting properties.
## Future perspectives
My research on foldamers was heavily affected by the work of Professor Xi-Kui Jiang on the self-coiling of organic molecules in aqueous media driven by hydrophobicity. At an early stage, we developed quite a number of folding patterns to explore the so-called functions. Finally, I established the hydrogen-bonded aromatic amide and hydrazide sequences as a longstanding research area. With these sequences as a struc- tural basis, we were able to investigate many interesting phenomena or concepts in supramolecular chemistry such as stimuli responsiveness, "sergeant-soldier effect", and supramolecular devices. It is also noteworthy that the preorganization feature of the amide sequences can remarkably increase the selectivity of macrocyclization. By making use of the dynamic covalent chemistry approach, we have demonstrated that complicated macrocycles can be obtained in nearly quantitative yields.
My research on aromatic foldamers has lasted for more than ten years. In collaboration with Junli, we recently found that the tubular cavity of hydrazide foldamers could mediate the transmembrane transport of K +. One ongoing project is to explore the routes for improving the transport selectivity for proton, cations and anions by making use of helical foldamers as channels. Long helical foldamers can theoretically form deep spring-shaped tubes. However, the synthesis and characterization of long foldamer polymers has been a challenge. I hope that we can find solutions to address this issue in the future. Particularly, we are interested in obtaining stable, single macromolecular tubes, probably after suitable postmodification. Solution-phase SOFs are new, homogeneous, porous architectures. As ordered supramolecular polymeric electrolytes, this family of self-assembled systems may be developed as useful adsorbing materials. Currently we are investigating new behaviors of molecules or macromolecules adsorbed by SOFs in solution.
[fig] Scheme 2: The formation of catenanes 6a-c. [/fig]
[fig] Scheme 1: Proposed structures of complexes between 1a and 1b with 2. [/fig]
[fig] Scheme 3: The structures of cantenanes 7a-c. [/fig]
[fig] Scheme 4: The structures of dimer 8·8 and compounds 9 and 10. Scheme 5: The structures of compounds 11a-g. from that of the monomer. I thus determined the K dim of 9 as 3.0 × 10 7 M −1 in chloroform saturated with water ([water] ≈ 0.45 M) and 8.5 × 10 7 M −1 in freshly opened chloroform ([water] ≈ 17 mM). [/fig]
[fig] Figure 1: X-ray structure of 10 showing a quadruple hydrogen-bonded dimeric motif[17]. [/fig]
[fig] Scheme 6: The structures of compounds 12a-g and the formation of the helical conformation by the longer oligomers. [/fig]
[fig] Scheme 7: The structures of compounds 13a,b, 14, and 15a-d. Scheme 8: The structures of complex C 60 16 and dynamic [2]catenane formed by compounds 17-19. [/fig]
[fig] Scheme 9: The structure of homodimers 20a·20a and 20b·20b. Scheme 10: The structures of foldamers 21 and 22a-c. Aromatic amide oligomers: folded and helical secondary structures. In 2005, Huiping Yi finished the synthesis and characterization of aromatic amide oligomers 21 and 22a-c [/fig]
|
Loss of TaIRX9b gene function in wheat decreases chain length and amount of arabinoxylan in grain but increases cross‐linking
Wheat contains abundant xylan in cell walls of all tissues, but in endosperm, there is an unusual form of xylan substituted only by arabinose (arabinoxylan; AX) that has long chains and low levels of feruloylation, a fraction of which is extractable in water (WE-AX). WE-AX acts as soluble dietary fibre but also gives rise to viscous extracts from grain, a detrimental trait for some nonfood uses of wheat. Here, we show that a glycosyl transferase family 43 wheat gene abundantly expressed in endosperm complements the Arabidopsis irx9 mutant and so name the three homoeologous genes TaIRX9b. We generated wheat lines with a constitutive knockout of TaIRX9b by stacking loss-of-function alleles for these homeologues from a mutagenized hexaploid wheat population resulting in decreases in grain extract viscosity of 50%-80%. The amount and chain length of WE-AX molecules from grain of these triple-stack lines was decreased accounting for the changes in extract viscosity. Imaging of immature wheat grain sections of triple-stacks showed abolition of immunolabelling in endosperm with LM11 antibody that recognizes epitopes in AX, but also showed apparently normal cell size and shape in all cell types, including endosperm. We identified differentially expressed genes from endosperm of triple-stacks suggesting that compensatory changes occur to maintain this endosperm cell wall integrity. Consistent with this, we observed increased ferulate dimerization and increased crosslinking of WE-AX molecules in triple-stacks. These novel wheat lines lacking functional TaIRX9b therefore provide insight into control of wheat endosperm cell walls.2316
# Introduction
Xylan is a major component of primary and secondary cell walls in grasses, typically accounting for 25%-30% of cell wall by dry weight. It exists mostly in highly decorated forms called heteroxylans that have frequent arabinofuranosyl (Araf), glucuronosyl and acetyl substitutions. Some of these Araf are themselves linked to hydroxycinnamic acids ferulate (FA) and para-coumarate (pCA). FA is key to grass cell wall function, allowing cross-linking between chains of xylan and from xylan to lignin via radical coupling of the FA while the role of pCA on xylan is less clear as it does not seem to form cross-links, but it may be involved in facilitating the radical coupling of FA and of lignin [bib_ref] The transcriptional landscape of polyploid wheat, Ralph [/bib_ref]. In wheat endosperm, the tissue of major importance for wheat end uses, cell walls have an unusual composition that is dominated (60%-70% of the polysaccharide) by a simple xylan that has only Araf decorations (arabinoxylan; AX) some of which are feruloylated. A portion (typically 20%-30%) of this endosperm AX is extractable in water (WE-AX); compared to the rest of endosperm AX (water-unextractable AX; WU-AX), WE-AX has low feruloylation but long chains [bib_ref] Wheat arabinoxylans: exploiting variation in amount and composition to develop enhanced varieties, Saulnier [/bib_ref]. The greater solubility of WE-AX compared to WU-AX has been attributed to the much lower amounts of diferulate dimers that can act as cross-links between AX chains [bib_ref] Wheat arabinoxylans: exploiting variation in amount and composition to develop enhanced varieties, Saulnier [/bib_ref]. The abundance and long chains of WE-AX confer viscosity to aqueous extracts of wheat grain, a property that is detrimental for monogastric animal nutrition [bib_ref] Relationship between the levels of soluble nonstarch polysaccharides and the apparent metabolizable..., Annison [/bib_ref] and for alcohol production. Conversely, wheat grain WE-AX is an important source of soluble fibre in the human diet [bib_ref] Variation in the content of dietary fiber and components thereof in wheats..., Gebruers [/bib_ref] ; these conflicting needs may lead to an increasing separation between wheat varieties intended for human food and other uses.
The xylan backbone is considered to be synthesized by a complex comprising IRX9, IRX14 and IRX10 proteins [bib_ref] Asparagus IRX9, IRX10, and IRX14A are components of an active xylan backbone..., Zeng [/bib_ref] ; IRX9 and IRX14 are encoded by genes in the glycosyl transferase family (GT) 43 and IRX10, which is the catalytically active component [bib_ref] Arabidopsis thaliana IRX10 and two related proteins from psyllium and Physcomitrella patens..., Jensen [/bib_ref] [bib_ref] Two Arabidopsis proteins synthesize acetylated xylan in vitro, Urbanowicz [/bib_ref] , by GT47 genes. Loss-of-function mutations in any of these components in Arabidopsis result in the irregular xylem (irx) phenotype accompanied by severe dwarfing due to loss of mechanical strength in cell walls. The function(s) of IRX9 and IRX14 in the xylan synthase are not clear but are most likely accessory proteins. The most abundantly expressed genes in wheat endosperm that resemble IRX9 genes are three homeologues that we called TaGT43_2A, TaGT43_2B, TaGT43_2D [bib_ref] Cell walls of developing wheat starchy endosperm: comparison of composition and RNA-seq..., Pellny [/bib_ref]. We have previously shown that transgenic wheat expressing an RNAi construct targeting these genes resulted in decreases in the amount and chain length of WE-AX in endosperm [bib_ref] RNA interference suppression of genes in glycosyl transferase families 43 and 47..., Lovegrove [/bib_ref] and in decreased extract viscosity [bib_ref] Effect of suppression of arabinoxylan synthetic genes in wheat endosperm on chain..., Freeman [/bib_ref]. Suppression of TaGT43_2 genes did not only affect WE-AX; the amount of endosperm WU-AX was also decreased in amount and chain length [bib_ref] Effect of suppression of arabinoxylan synthetic genes in wheat endosperm on chain..., Freeman [/bib_ref] and while the wall thickness of endosperm cells was decreased, endosperm cell shape was unaltered [bib_ref] RNA interference suppression of genes in glycosyl transferase families 43 and 47..., Lovegrove [/bib_ref]. This indicates that endosperm cell wall strength was maintained, likely made possible by high levels of diferulate dimers in the remaining AX such that diferulate per unit endosperm tissue was not decreased in the RNAi lines [bib_ref] Feruloylation and structure of arabinoxylan in wheat endosperm cell walls from RNAi..., Freeman [/bib_ref]. Effects in these TaGT43_2 RNAi lines would be dictated by spatial and temporal expression of the transgene which was driven by the HMW1Dx5 endosperm-specific promoter; it was unclear whether a constitutive removal of TaGT43_2 function would have similar effect enabling us to eliminate residue viscosity derived from grain WE-AX without affecting cell wall function in endosperm and other tissues.
There are eight triads of homeologous genes resembling IRX9 in wheat [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref] , and the homeologues of each triad have similar expression patterns, as is the case for most wheat triads (Ram ırez-Gonz alez et al., 2018). While the TaGT43_2 genes are the most highly expressed of these genes in grain, in all other tissues they are less highly expressed and TaGT43_6 tends to be most expressed [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref]. TaGT43_2 transcripts account for 70% of total putative IRX9 transcripts in endosperm but~15% in non-grain tissues [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref]. We therefore postulated that constitutive knockout of TaGT43_2 function would have similar effects to the endosperm-specific silencing, a decrease in longchain WE-AX from grain, but with no consequences for cell wall function in endosperm or other tissues. Approaches for finding mutations in specific genes of wheat were therefore developed using variants of TILLING methodology on EMS-mutagenized populations of wheat, cv. Cadenza [bib_ref] High resolution melting analysis for the detection of EMS induced mutations in..., Botticella [/bib_ref] followed by crossing to achieve complete functional knockouts of gene function. Here, we report use of this approach to develop triple mutants lacking functional TaGT43_2 genes and consequent loss of long-chain WE-AX; we also examine pleiotropic effects on the transcriptome and AX molecule cross-linking.
# Results
## Gene identification
We assessed the likelihood that constitutive knockouts of TaGT43 genes would have effects on cell wall function in tissues other than grain using sequence and expression evidence. TaGT43_2 genes are orthologs of the rice gene OsGT43B which complements the irx9 Arabidopsis mutant [bib_ref] Functional roles of rice glycosyltransferase family GT43 in xylan biosynthesis, Lee [/bib_ref] , one of four rice genes that have been shown to complement irx9 [bib_ref] Functional roles of rice glycosyltransferase family GT43 in xylan biosynthesis, Lee [/bib_ref]. In wheat, there are 8 triads in the same clade as these genes and so seem likely to have IRX9 function, rather than the IRX14 which are also encoded by GT43 genes [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref]. If these genes are indeed functionally redundant, then the transcript abundance evidence [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref] suggests that a complete knockout of TaGT43_2 function should have little effect in tissues other than grain.
## Tagt43_2 functionally complements irx9 gene in arabidopsis
We expressed the coding region of the wheat gene TaGT43_2B driven by the Arabidopsis IRX3 promoter in the Arabidopsis irx9 mutant to test complementation of knockout of Arabidopsis IRX9. The IRX3 promoter was selected as IRX3 and IRX9 genes have similar expression patterns in Arabidopsis, predominantly in cells depositing secondary cell walls, but IRX3 is preferred over IRX9 due to higher absolute expression. We observed a rescued phenotype in these plants similar to wild-type and irx9 lines overexpressing Arabidopsis IRX9 [fig_ref] Figure 2: Visual phenotype [/fig_ref]. We also found that xylose (Xyl) content of alcohol insoluble residue (AIR) was restored to wild-type levels in irx9 lines overexpressing Arabidopsis IRX9 or TaGT43_2B [fig_ref] Figure 2: Visual phenotype [/fig_ref]. TaGT43_2 therefore complements irx9 and we rename this triad TaIRX9b; the b is to reflect the existing nomenclature for rice ortholog OsGT43B [bib_ref] Functional roles of rice glycosyltransferase family GT43 in xylan biosynthesis, Lee [/bib_ref] but using lowercase to avoid confusion with labelling of wheat A, B, D homeologs.
## Identification of mutants
We searched for loss-of-function alleles for the TaIRX9b genes in the mutagenized population of hexaploid wheat lines derived from cv. Cadenza [bib_ref] Diversity of agronomic and morphological traits in a mutant population of bread..., Rakszegi [/bib_ref]. We used highresolution melting to detect SNPs in amplicons produced by nested PCR as described by [bib_ref] High resolution melting analysis for the detection of EMS induced mutations in..., Botticella [/bib_ref] for starch branching enzyme IIa target wheat genes. We selected mutations that introduced premature stop codons in each of three homeologs of TaIRX9b that were predicted to truncate the encoded proteins by 306, 69, 62 residues for TaIRX9b A, B, D respectively [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref]. Co-dominant competitive allele-specific PCR (KASP) markers were designed for each of these alleles to facilitate genotyping of progeny from crosses.
## Triple mutants of tairx9b knockout homeoalleles confer greatly decreased we-ax
We compared grain WE-AX content in lines that were homozygous for these knockout (KO) alleles [fig_ref] Figure 3: Effect of different stacks of TaIRX9b KO alleles on [/fig_ref] ; lines carrying only one KO homeoallele did not differ significantly from a nullsegregant control line, but one out of the three double stacks (aabbDD) had significantly (F-test, 1-way ANOVA; P < 0.05) lower WE-AX and the aabbdd triple-stack had WE-AX content that was approximately 40% that of control (P < 0.001; . This dose effect on WE-AX of stacking KO alleles is similar to other examples for wheat where a big effect is only achieved when all KO homeoalleles are combined and has been termed the 'non-additive dose' case, which along with the 'fully redundant' case reflect hidden variation that would not be found by conventional wheat breeding [bib_ref] Applying the latest advances in genomics and phenomics for trait discovery in..., Borrill [/bib_ref]. Total AX content of grain was also decreased in triple mutants to about 65% of null-segregant control (P < 0.001; [fig_ref] Figure 3: Effect of different stacks of TaIRX9b KO alleles on [/fig_ref]. There were no significant differences between lines in grain weight ; F-test, 1-way ANOVA; P = 2.41).
Grain sections from triple mutants show loss of LM11 labelling but apparent normal cell shape
We made transverse grain sections at 10 and 17 days postanthesis (dpa) from three replicate plants of triple-stack and nullsegregant control lines and used the LM11 monoclonal antibody which was raised against an unsubstituted penta-b-1,4-xylanoside glycoprotein [bib_ref] Monoclonal antibodies to plant cell wall xylans and arabinoxylans, Mccartney [/bib_ref] and recognizes unsubstituted regions on AX but does not bind to highly substituted AX. We previously showed that immunolabelling with LM11 of grain sections was decreased in the TaIRX9b RNAi lines, explicable by the greater level of Araf substitution in the remaining AX [bib_ref] RNA interference suppression of genes in glycosyl transferase families 43 and 47..., Lovegrove [/bib_ref] , and we observed the same trends here [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref] ; further examples [fig_ref] Figure 3: Effect of different stacks of TaIRX9b KO alleles on [/fig_ref]. At 10 dpa LM11 labelling is mostly confined to the nucellar epidermis and crease region in controls and this was virtually abolished in triple-stack [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref] except for residual labelling occurring on the inner periclinal cell walls of nucellar epidermis [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref]. In the triple-stack we observed almost no labelling in endosperm and aleurone cells at 17 dpa [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref] with only the crushed nucellar epidermis still showing strong labelling . Images in [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref] are overlays of confocal fluorescence and transmitted light which show unlabelled cell walls in grey; from these representative images and others, we observed that cell size and shape in all grain tissues were normal in the triple mutants although there was evidence that cell wall width of endosperm cells was decreased (most clearly visible in noantibody control images; [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref] , as was also the case for TaIRX9b RNAi lines [bib_ref] RNA interference suppression of genes in glycosyl transferase families 43 and 47..., Lovegrove [/bib_ref].
## Transcriptome from triple-stack and control grain suggest pleiotropic effects to maintain cell wall integrity
We analysed the RNAseq transcriptome of endosperm and grain outer tissues from triple-stack, control null-segregant and wildtype Cadenza lines isolated at 21 dpa. This stage was selected to be within grain-fill period but still with high levels of transcription in the endosperm likely to influence properties of mature grain [bib_ref] Cell walls of developing wheat starchy endosperm: comparison of composition and RNA-seq..., Pellny [/bib_ref]. Viewing mapped reads from control and triple mutants for the TaIRX9b genes confirms the introduced SNPs in all the triple-stack samples [fig_ref] Figure 2: Visual phenotype [/fig_ref]. It also shows lower read count in the triple-stack [fig_ref] Figure 2: Visual phenotype [/fig_ref] ; [fig_ref] Figure 5: Transcript abundance of selected genes from RNAseq of endosperm tissue isolated at... [/fig_ref] ; this is expected when a mutation moves the stop codon before the last exon (as is the case here for all three homeoalleles) as this triggers nonsense-mediated decay [bib_ref] Current perspectives on mRNA stability in plants: multiple levels and mechanisms of..., Gutierrez [/bib_ref]. This phenomenon was presumably responsible for significantly lower TaIRX9b transcript abundance in endosperm [fig_ref] Figure 5: Transcript abundance of selected genes from RNAseq of endosperm tissue isolated at... [/fig_ref] in triplestack compared to control and Cadenza; in another study which targeted different genes from the same wheat population , rice (green) and wheat (black) genes. All wheat genes cluster together as triads from the A, B and D subgenomes and are named as previously [bib_ref] Cell walls of developing wheat starchy endosperm: comparison of composition and RNA-seq..., Pellny [/bib_ref]. Rice genes are as named in [bib_ref] Three novel rice genes closely related to the Arabidopsis IRX9, IRX9L, and..., Chiniquy [/bib_ref] and [bib_ref] Functional roles of rice glycosyltransferase family GT43 in xylan biosynthesis, Lee [/bib_ref]. Locus IDs for these genes are given in [fig_ref] Table 1: Bound hydroxycinnamate composition of whole grain and endosperm [/fig_ref]. The IRX9 clade in which genes are putatively redundant with Arabidopsis IRX9 is indicated; OsIRX9, OsIRX9L and OsGT43A and OsGT43E [bib_ref] Functional roles of rice glycosyltransferase family GT43 in xylan biosynthesis, Lee [/bib_ref] functionally complement irx9. (b, c) Public expression data for wheat genes in IRX9 clade taken from the expVIP platform [bib_ref] expVIP: a customizable RNA-seq data analysis and visualization platform, Borrill [/bib_ref] ; all hexaploid wheat studies were used, using means for each study, source tissue, stage, treatment combination as data points. (b) Transcript abundance (TPM; transcript reads per million mapped reads) of all putative wheat IRX9 genes shown as triads with homeologues ordered A, B, D. Means AE SD for samples from non-grain (n = 134) and grain (n = 34) tissues. (c) Proportion of total transcript counts of wheat IRX9 genes accounted for by sum of TaGT43_2A, B and D. Means AE SD for samples from non-grain (n = 134), whole grain (n = 13), endosperm (includes starchy endosperm and aleurone) (n = 16) and grain outer tissues (n = 4). [bib_ref] Combining mutations at genes encoding key enzymes involved in starch synthesis affects..., Botticella [/bib_ref] , expression of mutated target genes was also decreased.
We assessed other changes on the transcriptome by identifying all differentially expressed genes (DEGs) between the control and triple-stack (P < 0.05, adjusted for multiple testing). Both control and triple-stack lines would have many background mutations; discounting DEGs that also occurred as DEGs between Cadenza and control (likely to be non-specific responses) and TaIRX9b themselves, there are 355 and 394 DEGs in endosperm and grain outer tissues, respectively [fig_ref] Table 3: Bound hydroxycinnamate composition of water extracts from grain of control and triple-stack [/fig_ref] for endosperm and outer layers respectively) out of the 108 035 high confidence genes annotated in the IWGSC 1.1 wheat genome [bib_ref] Shifting the limits in wheat research and breeding using a fully annotated..., Appels [/bib_ref]. Contained within these sets are many transcripts associated with cell walls, suggestive of pleiotropic effects due to the loss of TaIRX9b function. For example, five expansin genes are down-regulated in endosperm [fig_ref] Table 3: Bound hydroxycinnamate composition of water extracts from grain of control and triple-stack [/fig_ref] , consistent with a cell wall integrity sensing mechanism responding to a loosening [bib_ref] To grow or not to grow that's the question, Voxeur [/bib_ref] caused by loss of AX chains. Since we had previously observed that diferulate (diFA) levels were maintained despite loss of AX in endosperm of RNAi lines [bib_ref] Feruloylation and structure of arabinoxylan in wheat endosperm cell walls from RNAi..., Freeman [/bib_ref] , we were interested to find a BAHD gene (TaBAHD2A) and a peroxidase gene (TaPRX19D) among the most up-regulated within endosperm DEGs. BAHD genes in this clade are putatively involved in addition of FA (de Souza et al., 2018) or p-coumarate (pCA) to AX [bib_ref] Overexpression of a BAHD acyltransferase, OsAt10, alters rice cell wall hydroxycinnamic acid..., Bartley [/bib_ref] ; we have recently demonstrated that the orthologs of these TaBAHD02 genes are dramatically up-regulated by methyl-jasmonate in Brachypodium accompanied by a rise in both FA and pCA linked to AX [bib_ref] Response of cell-wall composition and RNA-seq transcriptome to methyl-jasmonate in Brachypodium distachyon..., Hyde [/bib_ref]. Tandemly arranged TaBAHD02 paralogues are present on all subgenomes in wheat and were significantly up-regulated in the endosperm of triple mutants [fig_ref] Figure 5: Transcript abundance of selected genes from RNAseq of endosperm tissue isolated at... [/fig_ref]. TaPRX19D is from the family of class III peroxidases, members of which are involved in oxidative cross-linking of cell wall proteins in dicots [bib_ref] To grow or not to grow that's the question, Voxeur [/bib_ref] and FA in grasses; the homeologues of this gene are also upregulated in the endosperm of triple mutants [fig_ref] Figure 5: Transcript abundance of selected genes from RNAseq of endosperm tissue isolated at... [/fig_ref]. Some other DEGs in endosperm and outer layers are also potentially related to xylan synthesis and remodelling (annotated in Tables S3 and S4); the most up-regulated gene in endosperm being TraesCS6B02G187200 from the Trichome birefringencelike family [fig_ref] Table 3: Bound hydroxycinnamate composition of water extracts from grain of control and triple-stack [/fig_ref]. Members of this family are responsible for xylan acetylation [bib_ref] Xylan O-acetylation impacts xylem development and enzymatic recalcitrance as indicated by the..., Xiong [/bib_ref] and intense acetylation of wheat arabinoxylan has been reported in developing grain although AX in mature grain is not acetylated (Veli ckovi c et al., 2014).
## Cell wall ferulate composition of triple mutants shows increased dimerization
To decrease the number of background mutations, we undertook four rounds of backcrossing, selfed progeny and identified homozygous plants for the stacks and control (control = no TaIRX9b mutant alleles) lines. All results we present below are from these BC4F2 lines. We observed a decrease in cell wall ferulate monomer (FA) but an increase in diFA in mature grain from triple mutants compared to controls [fig_ref] Table 1: Bound hydroxycinnamate composition of whole grain and endosperm [/fig_ref]. This increased dimerization was also significant in the endosperm fraction [fig_ref] Table 1: Bound hydroxycinnamate composition of whole grain and endosperm [/fig_ref]. We previously observed similarly increased FA dimerization in endosperm cell walls in transgenic wheat where TaIRX9b genes were repressed by RNAi [bib_ref] Feruloylation and structure of arabinoxylan in wheat endosperm cell walls from RNAi..., Freeman [/bib_ref].
Viscosity of grain extracts from triple-stack and control lines is decreased due to loss of long-chain WE-AX Viscosity of water extracts from wholemeal flour is primarily determined by WE-AX with smaller contributions from soluble (1,3;1,4)-beta-glucan and low molecular weight molecules such as arabinogalactan peptide (AGP) and raffinose. These extract viscosities show considerable variation between samples of wildtype wheat [bib_ref] Effect of suppression of arabinoxylan synthetic genes in wheat endosperm on chain..., Freeman [/bib_ref] and the controls in this experiment were at the lower end of the range; despite this, we found that specific viscosity in triples was about 45% of that in controls [fig_ref] Table 2: Monosaccharide composition [/fig_ref]. As expected, we found decreased Xyl and Ara content in the extracts [fig_ref] Table 2: Monosaccharide composition [/fig_ref] ; Ara was decreased by less indicating increased Ara substitution in WE-AX as found previously in transgenics [bib_ref] RNA interference suppression of genes in glycosyl transferase families 43 and 47..., Lovegrove [/bib_ref] even allowing for Ara from AGP which will also be present. We examined the size-exclusion HPLC (SE-HPLC) profiles of extracts from grain of triple-stack and control lines [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref] and found that the contribution of high molecular weight molecules to viscosity was proportionately more decreased in triple mutants [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref]. The part of the profile between vertical lines in [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref] correspond to viscosity contributed only by large WE-AX and (1,3;1,4)-betaglucan molecules. The integrals of viscosity between these limits account for about 65% of bulk extract viscosities given in [fig_ref] Table 2: Monosaccharide composition [/fig_ref] , the remainder can be attributed to low molecular weight WE-AX and other small molecules with long retention times. We treated samples with lichenase to specifically remove high molecular weight (1,3;1,4)-beta-glucan [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref] which removed 0.01-0.02 contribution to integral of specific viscosity for all samples (average total integrals were 0.23 and 0.09 for controls and triples respectively). This suggests that amounts of high molecular weight (1,3;1,4)-beta-glucan were similar in extracts from control and triple grain. The lichenase-insensitive contribution to viscosity from molecules between the retention time limits is the product of the amount and intrinsic viscosity of WE-AX. Intrinsic viscosity of WE-AX is determined mostly by chain length [bib_ref] Experimental evidence for a semi-flexible conformation for arabinoxylans, Dervilly-Pinel [/bib_ref] , although some effect of diFA mediated cross-linking has been reported [bib_ref] Isolation of homogeneous fractions from wheat water-soluble arabinoxylans. Influence of the structure..., Dervilly [/bib_ref]. We found distributions of intrinsic viscosity to be not much affected by saponification and consequent loss of diFA in our samples [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref] ; average intrinsic viscosities (area under nonlog version of curves) for non-saponified and saponified were respectively 53 and 49 mL/g for control and 15 and 13 mL/g for triple-stack samples. We found much lower concentrations of all WE-AX molecules with log intrinsic viscosity >1 in triple mutants compared to controls and no WE-AX with log intrinsic viscosity >2.8 in triples. Using linear relationship between log chain length and log intrinsic viscosity derived for non-feruloylated WE-AX [bib_ref] Experimental evidence for a semi-flexible conformation for arabinoxylans, Dervilly-Pinel [/bib_ref] this corresponds to amounts of all WE-AX molecules >10 Xyl length being decreased in triple mutants and chains of >2000 being undetectable.
Grain WE-AX has more diFA and is more cross-linked in triple-stack lines
We found that WE-AX from triples had 2.0-fold higher FA content and 2.6-fold higher diFA content per unit xylose than controls [fig_ref] Table 3: Bound hydroxycinnamate composition of water extracts from grain of control and triple-stack [/fig_ref]. We investigated whether this increased amount of diFA was associated with increased inter-molecular cross-linking of WE-AX by examining effect of saponification (which removes FA and diFA) on WE-AX molecular weight (MWt) distribution estimated by multi-angle laser light scattering on SE-HPLC [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref]. We found that saponification decreased amounts of higher MWt WE-AX molecules in both controls and triples but the decrease was much greater in triples [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref]. The effect of saponification on average MWt (area under non-log version of curves in [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref] was a 6-fold decrease for triples and a 1.2-fold decrease for controls; this suggests most WE-AX molecules from triples are cross-linked via diFA bridges to multiple others whereas such cross-linking is comparatively rare in WE-AX from controls.
# Discussion
We have demonstrated that the three homeologous GT43_2 genes that are highly expressed in endosperm [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref] are functional orthologues of IRX9 in Arabidopsis [fig_ref] Figure 2: Visual phenotype [/fig_ref] so name them TaIRX9b. This is a noteworthy finding since the Arabidopsis IRX9 is involved in synthesis of secondary cell wall xylan which carries glucuronosyl and acetyl decorations whereas wheat endosperm AX is decorated only with arabinofuranosyl residues. By identifying and stacking loss-of-function alleles of TaIRX9b from a mutagenized wheat population, we were able to produce lines with decreased AX in the grain. The lack of a strong effect on AX in double mutants suggested that each homoelogue can support near-wild-type levels of AX biosynthesis; thus, only the triple-stack shows a substantial reduction [fig_ref] Figure 3: Effect of different stacks of TaIRX9b KO alleles on [/fig_ref]. Microscopy showed abolition of LM11 immunolabelling in triple mutants but no change in cell shape [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref] which is consistent with other observations that show large effects of abolition of TaIRX9b function on grain AX traits but no obvious effects on morphology.
We found a large decrease in viscosity of extracts from grain of triple mutants [fig_ref] Table 2: Monosaccharide composition [/fig_ref] ; low extract viscosity is a potentially valuable trait for use of wheat in alcohol production and animal feed. Using mutations avoids the problems of potential transgene silencing and regulatory restrictions of GM wheat lines. Furthermore, we observed no change in grain weight [fig_ref] Figure 3: Effect of different stacks of TaIRX9b KO alleles on [/fig_ref] , a key quality parameter for most uses. These lines therefore offer potential for development into wheat varieties intended for these applications.
The decrease in extract viscosity in triple mutants was entirely explained by a decrease in amount of long-chain WE-AX [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref]. A substantial decrease in amount of long-chain AX might be expected to weaken the starchy endosperm cell walls where it accounts for 65% of the polysaccharide and we found evidence that there were adaptive changes to compensate for such a weakening. In the transcriptome of the developing grain of controls and triple mutants, we found some of the most strongly differentially regulated genes are associated with cell wall synthesis or remodelling, particularly up-regulation of BAHD02 and peroxidase genes [fig_ref] Figure 5: Transcript abundance of selected genes from RNAseq of endosperm tissue isolated at... [/fig_ref] ; the BAHD02 genes are from a clade containing genes thought to be responsible for the addition of FA and pCA residues to AX [bib_ref] Overexpression of a BAHD acyltransferase, OsAt10, alters rice cell wall hydroxycinnamic acid..., Bartley [/bib_ref] [bib_ref] A novel bioinformatics approach identifies candidate genes for the synthesis and feruloylation..., Mitchell [/bib_ref] [bib_ref] Suppression of a single BAHD gene in Setaria viridis causes large, stable..., De Souza [/bib_ref] and peroxidases in the apoplast are believed to be required for oxidative dimerization of FA on AX. Therefore, the up-regulation of BAHD02 and peroxidase genes may be responsible for the increased FA per unit AX and FA dimerization that we observed here [fig_ref] Table 1: Bound hydroxycinnamate composition of whole grain and endosperm [/fig_ref] and previously [bib_ref] Feruloylation and structure of arabinoxylan in wheat endosperm cell walls from RNAi..., Freeman [/bib_ref] in response to TaIRX9b repression. Brachypodium orthologues of TaBAHD02 and multiple peroxidases were also substantially upregulated following methyl-jasmonate treatment which also induced increases in cell wall pCA, FA and FA dimerization [bib_ref] Response of cell-wall composition and RNA-seq transcriptome to methyl-jasmonate in Brachypodium distachyon..., Hyde [/bib_ref]. Here, we obtained evidence that the increased amount of diFA per unit AX resulted in greater cross-linking between WE-AX molecules from triples, shown by the much greater decrease in molecular weight induced by saponification [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref].
Such compensatory effects are consistent with knowledge of other primary cell walls, where extreme perturbations can be completely compensated for by changes in composition and structure that are induced by sophisticated sensing of cell wall integrity, although little is known about how this operates in grass primary cell walls [bib_ref] To grow or not to grow that's the question, Voxeur [/bib_ref]. Some of the other DEGs within the endosperm transcriptome from the triples [fig_ref] Table 3: Bound hydroxycinnamate composition of water extracts from grain of control and triple-stack [/fig_ref] may also reflect some of these cell wall adaptive changes, for example down-regulation of expansins and some aspects of the signalling steps involved (kinases and calciumbinding proteins). Our RNAseq transcriptomes therefore represent a resource that can be exploited to formulate hypotheses on control of wheat grain cell walls.
We have demonstrated that abolishing TaIRX9b function has a large effect on long-chain WE-AX from grain, but morphology and cell shape are unchanged. Assuming all xylan synthesis requires IRX9 components within xylan synthase complexes, the remaining xylan in the triples is presumably synthesized with IRX9 components encoded by some of the 7 other triads resembling IRX9 [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref]. We do not know whether the specific role of TaIRX9b genes in synthesizing long-chain WE-AX is due only to their expression pattern, which will determine the context of the encoded IRX9 protein (e.g. other proteins and metabolites present), or whether the sequence of the encoded IRX9 also confers specialist properties, for example leading to longer chains. Complementation experiments of these triple-stack lines with different IRX9 genes could be used to address this. The apparent lack of effect of knocking out TaIRX9b function apart from on grain AX raises questions why these genes were selected for during evolution of wheat ancestors. WE-AX acts as an antinutritive factor in wheat grain consumed by poultry, possibly due to its effect on viscosity [bib_ref] Cell walls and their components in cereral grain technology, Fincher [/bib_ref].
We speculate that WE-AX is present mostly as an antinutritive factor and that having separate IRX9 genes responsible for WE-AX allows for separate control of this (e.g. expression largely confined to grain during grain-filling) as opposed to xylans essential for cell wall function. Therefore, abolition of TaIRX9b function impacts greatly on viscosity of extracts from grain with little effect on cell wall integrity, although in the case of endosperm cell walls this appears to require compensatory mechanisms.
In conclusion, building on our previously reported effects of suppressing homeologous GT43 genes by RNAi, we have now shown that this wheat GT43 gene is a true IRX9 capable of restoring xylan synthesis to irx9 mutant, despite the major differences between wheat endosperm AX and xylan in secondary cell walls of Arabidopsis. We developed novel wheat lines lacking functional TaIRX9b by stacking three knockout homeoalleles. Using mutations avoids the problems of potential transgene silencing and regulatory restrictions of GM wheat lines. These lines therefore offer potential for development into wheat varieties intended for alcohol production and animal feed. They are also a resource for investigating the control of endosperm cell walls, as shown by the evidence of compensatory changes to maintain cell wall function, with mechanisms which may well apply more widely in primary cell walls of grasses.
# Materials and methods
# Phylogenetic analysis
Peptide sequences encoded by GT43 genes were taken from loci listed in [fig_ref] Table 1: Bound hydroxycinnamate composition of whole grain and endosperm [/fig_ref]. Alignments and generation of maximum likelihood tree were performed as previously described [bib_ref] Cell walls of developing wheat starchy endosperm: comparison of composition and RNA-seq..., Pellny [/bib_ref] [bib_ref] Suppression of a single BAHD gene in Setaria viridis causes large, stable..., De Souza [/bib_ref].
## Complementation
Expression vectors were obtained using Golden Gate Modular Cloning (MolClo) with the standard parts [bib_ref] Standards for plant synthetic biology: a common syntax for exchange of DNA..., Patron [/bib_ref]. AtIRX9, TaIRX9_2 coding sequences and IRX3 promoter domesticated DNA sequences where synthesized by GENEWIZâ. Expression vectors consisted in Level 2 construct containing in position R1 Oleosin-GFP selection marker [bib_ref] A rapid and nondestructive screenable marker, FAST, for identifying transformed seeds of..., Shimada [/bib_ref] and pIRX3-AtIRX9-myc or pIRX3-TaIRX9_2-myc in position R2. For complementation experiments, Arabidopsis irx9.1 mutant plants were transformed using the standard floral dip method [bib_ref] Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana, Clough [/bib_ref]. Seven-week-old transgenic lines and the controls were used to evaluate growth complementation. Neutral monosaccharide composition of non-cellulosic polysaccharides of basal stems was analysed by high-performance anion-exchange chromatography coupled to pulsed amperometric detector (HPAEC-PAD) as described in [bib_ref] Characterisation of FUT4 and FUT6 alpha-(1 -> 2)-fucosyltransferases reveals that absence of..., Tryfona [/bib_ref].
## Identification of mutant wheat lines
We screened lines from the mutagenized wheat population derived from cv. Cadenza [bib_ref] Diversity of agronomic and morphological traits in a mutant population of bread..., Rakszegi [/bib_ref] , searching for mutations in TaIRX9b genes. Identification of mutations in target genes was by high-resolution melt analysis [bib_ref] High resolution melting analysis for the detection of EMS induced mutations in..., Botticella [/bib_ref] followed by Sanger sequencing of amplicons. The high mutation rate in this population means that there is a good chance of finding loss-of-function mutations for any target in relatively few lines and we found a total of 7 premature stop codons over all three homeologs of TaIRX9b after screening 2150 lines; we selected one of these for each homeolog [fig_ref] Figure 1: GT43 gene family phylogeny and expression in wheat [/fig_ref]. These alleles were followed in progeny of crosses using co-dominant KASP markers (Limagrain genotyping facility, France) using primers given in [fig_ref] Table 2: Monosaccharide composition [/fig_ref] to identify homozygous plants. Lines carrying mutations in homeologs were crossed and selfed and F2 homozygotes identified to produce the aabbDD, aaBBdd and AAbbdd double stacks. The aabbDD stack was crossed with AABBdd, selfed and F2 homozygotes identified to produce the M5 triple-stack and null-segregant controls. The triple-stack line was subjected to 4 rounds of backcrossing to Cadenza, at each generation selecting lines carrying all three mutations. The BC4F1 line was selfed and homozygotes selected to give BC4F2 triple mutants and null-segregant controls.
## Plant sampling
Control and lines carrying mutant alleles were grown in small plots in the field at Rothamsted in 2016 [fig_ref] Figure 3: Effect of different stacks of TaIRX9b KO alleles on [/fig_ref] or in pots in a controlled temperature glasshouse with supplementary lighting [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref]. Results shown in Figures 3-5 are from M5 lines; results in Tables 1-3 and [fig_ref] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from... [/fig_ref] are from BC4F2 lines.
## Rnaseq
Endosperms were separated from grain outer layers by rolling them out at 21 days postanthesis (dpa) from three biological replicates of each line (null-segregant, triple mutant, non-mutagenized Cadenza); RNA was isolated from both endosperm and outer layers giving 18 samples. Total RNA isolation and quality control was as in [bib_ref] Cell walls of developing wheat starchy endosperm: comparison of composition and RNA-seq..., Pellny [/bib_ref]. Library preparation and sequencing were carried out by Oxford Genomics Centre, UK. Reads were trimmed to remove adapter and low-quality ends and then mapped to the IWGSC refseq1.1 genome [bib_ref] Shifting the limits in wheat research and breeding using a fully annotated..., Appels [/bib_ref] using the HISAT mapper with default settings [bib_ref] HISAT: a fast spliced aligner with low memory requirements, Kim [/bib_ref]. The 18 libraries all generated between 17 and 26 million uniquely mapped reads. Expression was estimated using Fea-tureCounts [bib_ref] featureCounts: an efficient general purpose program for assigning sequence reads to genomic..., Liao [/bib_ref] with parameter set for forward reads only (since stranded libraries were used) and other parameters set to default. Differentially expressed genes were identified using DEseq2 software [bib_ref] Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2, Love [/bib_ref]. All RNAseq read data and protocols are available under ArrayExpress accession E-MTAB-8237.
## Microscopy
Wheat grains were taken from the middle spikelets of the ear at 10 and 17 dpa. Wheat grains were fixed in 4% paraformaldehyde, 2.5% glutaraldehyde and dehydrated in a graded ethanol series. The samples were then processed through increasing concentrations of LR White Resin (Agar Scientific, Stansted, Essex, UK, R1281) and embedded at 58°C for 16-20 h in a nitrogen- rich environment. 1 µm sections of the resin blocks were cut with a Reichert-Jung ultramicrotome, and dried onto Polysine-coated slides (Agar Scientific, Stansted, Essex, UK, L4345) at 40°C. Sections were immunolabelled as described by [bib_ref] Distribution of gluten proteins in bread wheat (Triticum aestivum) grain, Tosi [/bib_ref] using the primary antibody LM11 and Alexa Fluor 488 goat antirabbit IgG (Invitrogen, Paisley, Renfrewshire, UK A11008). Images were acquired with a Zeiss, Cambourne, Cambridge, UK LSM 780 confocal microscope using Zeiss, Cambourne, Cambridge, UK ZEN 2010 software.
## Cell wall hydroxycinnamic acid content
Bound hydroxycinnamic acids were extracted from AIR prepared from whole grain or white flour, or from water extracts, and quantified by HPLC using a UPLC Kinetex Phenyl-Hexyl (150 mm 9 4.6 mm, 5 µm) column as previously described [bib_ref] Feruloylation and structure of arabinoxylan in wheat endosperm cell walls from RNAi..., Freeman [/bib_ref].
## Ax content, monosaccharides, viscometry and size-exclusion hplc
Total and WE-AX content were determined as detailed in [bib_ref] Influence of cultivar and environment on water-soluble and water-insoluble arabinoxylans in soft..., Finnie [/bib_ref]. Monosaccharide analysis following acid hydrolysis was as described by [bib_ref] GUX1 and GUX2 glucuronyltransferases decorate distinct domains of glucuronoxylan with different substitution..., Bromley [/bib_ref]. Relative viscosity of water extracts was determined using the method of [bib_ref] Effect of suppression of arabinoxylan synthetic genes in wheat endosperm on chain..., Freeman [/bib_ref]. WE extracts for use in SE-HPLC were prepared as for relative viscosity measurements except that following centrifugation the supernatant was aliquoted (to give equivalent of 200 mg/mL starting material), made up to 1 mL with water and digested with recombinant lichenase as described in [bib_ref] Feruloylation and structure of arabinoxylan in wheat endosperm cell walls from RNAi..., Freeman [/bib_ref]. Following boiling and centrifugation, aliquots of 950 µL of supernatants were removed and dried under vacuum. Samples for saponification were resuspended in 400 µL 2 M NaOH, vortexed for 30 s and incubated in the dark for 16 h at 40°C. Following incubation, samples were neutralized by addition of 415 µL of 2 M HCL and vortexed. Control samples were resuspended in 815 µL of 2 M NaCl. 500 µL of control or saponified samples were desalted using PD Minitrap G-25 columns (GE Healthcare, Amersham, Herts, UK) using the manufacturer's spin protocol. Samples were filtered through 0.45 µm PVDF filters and 100 µL injected onto the SE-HPLC system which was as in [bib_ref] Oat bran, but not its isolated bioactive beta-glucans or polyphenols, have a..., Kristek [/bib_ref] with the following modifications; the column temperature was set at 35°C and running buffer contained 0Á02% sodium azide and 0.1 M sodium nitrate. Size-exclusion columns were Shodex OH-Pak SB 806M HQ and SB 804 HQ columns in series, and a flow rate of 0.5 mL/ min. Data were analysed using Wyatt, Haverhill, Suffolk, UK ASTRA software.
## Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article.
## Figure s1
Alignment of TaGT43_2 (TaIRX9b) sequences showing mutations used. [fig_ref] Figure 2: Visual phenotype [/fig_ref] Confirmation of TaIRX9b genotype from RNAseq in controls and triple-stack lines. [fig_ref] Figure 3: Effect of different stacks of TaIRX9b KO alleles on [/fig_ref] Further example images as in [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref]. [fig_ref] Figure 4: Overlay of fluorescence [/fig_ref] No-antibody control images for immunolabelling of grain sections of control and triple-stack lines. [fig_ref] Table 1: Bound hydroxycinnamate composition of whole grain and endosperm [/fig_ref] GT43 gene loci from Arabidopsis, rice and wheat.
[fig] Figure 1: GT43 gene family phylogeny and expression in wheat. (a) Maximum likelihood tree of Arabidopsis (blue text) [/fig]
[fig] Figure 2: Visual phenotype (a) and (b) neutral monosaccharide composition of AIR from stems of Arabidopsis wild-type (Wt), irx9 mutant, irx9 mutant expressing AtIRX9 gene, irx9 mutant expressing TaGT43_2B gene under IRX3 promoter, two independent measurements are shown for WT and irx9 mutant, and three independent lines are measured for complemented irx9 transgenic plants. Ara, arabinose; Fuc, fucose; Gal, galactose; Glc, glucose; Man, mannose; Rha, rhamnose; Xyl, xylose. ª 2020 The Authors. Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd, 18, 2316-2327 [/fig]
[fig] Figure 3: Effect of different stacks of TaIRX9b KO alleles on (a) WE-AX, (b) total AX from grain and (c) grain weight. In mutant lines and null segregants, allele composition is shown with letter indicating subgenome and lowercase denoting recessive KO allele. Cadenza is non-mutagenized cultivar. Error bars are AESE, n = 3 independent biological reps.ª 2020 The Authors. Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd, 18, 2316-2327 [/fig]
[fig] Figure 4: Overlay of fluorescence (green) and transmitted light (grey) images of sections from control and triple-stack grain at 10 and 17 dpa. Fluorescence is due to LM11 immunolabelling of low substituted AX. Starchy endosperm (se), nucellar epidermis (ne) and outer pericarp (op) are indicated.ª 2020 The Authors. Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd, 18, 2316-2327 [/fig]
[fig] Figure 5: Transcript abundance of selected genes from RNAseq of endosperm tissue isolated at 21 dpa from Cadenza wild-type, null-segregant control and triple-stack of TaIRX9b KO alleles. (a) Example read coverage of TaIRX9b genes. SNP differences from IWGSC Chinese Spring reference are coloured with the SNP conferring premature stop codon in triple indicated by red arrow. (b, c) Mean transcript reads per million mapped reads (TPM) AE SD (n = 3). (b) TaIRX9b genes (The Authors. Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd, 18, 2316-2327 [/fig]
[fig] Figure 6: SE-HPLC analyses on extracts from control and triple-stack grain showing data from three independent biological reps of each. (a) Chromatograms showing specific viscosity of samples with and without lichenase treatment. Vertical dotted lines denote integration limits used. (b) Distributions of log intrinsic viscosity from lichenase-treated extracts with or without saponification treatment. (c) Distributions of log molecular weight from lichenase-treated extracts or without saponification treatment. ª 2020 The Authors. Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd, 18, 2316-2327 [/fig]
[table] Table 1: Bound hydroxycinnamate composition of whole grain and endosperm (white flour) from control and triple-stack lines. Averages AE SD, n = 3 [/table]
[table] Table 2: Monosaccharide composition (excluding Glc) and viscosity of water extracts from grain of control and triple mutant lines. Averages AE SD, n = 3 [/table]
[table] Table 3: Bound hydroxycinnamate composition of water extracts from grain of control and triple-stack. Two forms of diFA were detected; 8-O-4 0 diFA and 8-5 0 benzofuran (BF). The sum of these is given in diFA column. Averages AE SD, n = 3 The Authors. Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd, 18, 2316-2327 [/table]
[table] Table S2: Primers and conditions for competitive PCR assay marker. Table S3 DEGs identified from endosperm RNAseq. Table S4 DEGs identified from pericarp RNAseq. ª 2020 The Authors. Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd, 18, 2316-2327 [/table]
|
Fused in Sarcoma: Properties, Self-Assembly and Correlation with Neurodegenerative Diseases
Fused in sarcoma (FUS) is a DNA/RNA binding protein that is involved in RNA metabolism and DNA repair. Numerous reports have demonstrated by pathological and genetic analysis that FUS is associated with a variety of neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS), frontotemporal lobar degeneration (FTLD), and polyglutamine diseases. Traditionally, the fibrillar aggregation of FUS was considered to be the cause of those diseases, especially via its prion-like domains (PrLDs), which are rich in glutamine and asparagine residues. Lately, a nonfibrillar self-assembling phenomenon, liquid-liquid phase separation (LLPS), was observed in FUS, and studies of its functions, mechanism, and mutual transformation with pathogenic amyloid have been emerging. This review summarizes recent studies on FUS self-assembling, including both aggregation and LLPS as well as their relationship with the pathology of ALS, FTLD, and other neurodegenerative diseases.Molecules 2019, 24, 1622 2 of 17 specific transcription factors (e.g., Spi-1/PU.1, NF-κB, and Runx2) and transcription initiation factor TFIID[11][12][13][14]. One recognized mechanism of this regulation is that FUS binds to and recruits RNA polymerase II, subsequently altering its phosphorylation status[15]. RNA can promote the oligomerization of FUS, allowing it to form high-order components near TSS, which interact with CTD of RNA polymerase II. This allows more RNA polymerase II to be recruited to TSS and also prevents premature phosphorylation of Ser2 in CTD, stimulating the transition of polymerase from initiation to activity extension[5]. FUS also interacts with the DNA binding domain of certain nuclear hormone receptors to initiate their transcription[16]. Immunofluorescence studies have found that FUS preferentially localizes on active chromatin to execute its role in transcription regulation[17]. Specifically, during meiosis, FUS binds to autosomes other than transcriptionally silent sex chromosomes[18]. Similarly, it does not bind to transcriptionally silenced chromatin during mitosis[19].FUS modulates RNA splicing in conjunction with splicing regulators or precursor mRNAs[19]. FUS is a transporter of mRNA between the cytoplasm and the nucleus. For instance, the transfer of mRNA between neuronal dendrites and dendritic spines by FUS is essential for neuronal cell maturation, plasticity, and dendrite integrity[20]. Other relevant studies have indicated that FUS boosts the synthesis of microRNAs, as FUS is an integral part of the Drasha complex, which is required as a ribonuclease III enzyme for microRNA biosynthesis[21].Role of FUS in DNA Damage RepairDuring DNA damage repair, FUS is one of the proteins firstly recruited to the DNA damage site[22,23]. Loss of FUS results in impaired ATM/γH2AX (ataxia telangiectasia-mutated gene) signaling. The mechanism by which FUS regulates DNA damage repair also depends on interactions with histone deacetylase 1 (HADC1) and poly-ADP ribose, a by-product of DNA damage[22,24]. Knockdown of FUS during cell growth leads to defects in DNA damage recovery, and disease-associated mutations may limit FUS participation in DNA damage response (DDR)[25].In 1999, Baechtold demonstrated that FUS participates in DDR, which promotes complementary ssDNA annealing and single-stranded oligonucleotide uptake of homologous supercoiled DNA to form a D-loop. The D-loop formation is a principal step in DNA double-strand break repair via recombination, and the oncogenic fusion form FUS-CHOP does not promote DNA pairing[26]. Wild-type FUS can be phosphorylated by ATM in response to DNA double-strand breaks (DSBs), as phosphorylated FUS binds to dsDNA cleavage and Holliday junctions[27]. When DNA damage occurs, FUS is recruited by sense and antisense noncoding RNAs transcribed from 5 regulatory regions of the cyclin D1 (CCND1) gene[28]. FUS directly interacts with the CREB-binding protein/p300 histone acetyltransferase via its N-terminal domain, resulting in inhibition of CCND1 transcription[28]. Interestingly, FUS's ability to respond to DNA damage relies on allosteric interactions with single-stranded, low-copy-number long noncoding RNA transcripts[28]. Although the detailed processes of FUS in DDR have not been fully illustrated, its important function in DDR is without question.
# Introduction
Fused in sarcoma (FUS) is a DNA/RNA binding protein containing 526 amino acids. The FUS gene was initially identified as a fusion oncogene on chromosome 16 in human liposarcoma [bib_ref] Fusion of CHOP to a novel RNA-binding protein in human myxoid liposarcoma, Crozat [/bib_ref] , the translocation and fusion of which to transcription factors results in strong transcriptional activation of the proteins. FUS is one of the components of the heterogeneous nuclear ribonucleoprotein (hnRNP) complex. Increasing evidence suggests that FUS is involved in various cellular processes, including gene transcription and regulation, DNA repair, RNA shearing, RNA transport, translation, processing of microRNAs, and maintenance of genomic stability [bib_ref] TDP-43 and FUS/TLS: Emerging roles in RNA processing and neurodegeneration, Lagiertourenne [/bib_ref]. FUS can bind to RNA, ssDNA, and possibly to dsDNA [bib_ref] Converging mechanisms in ALS and FTD: Disrupted RNA and protein homeostasis, Ling [/bib_ref]. The most important binding sequence of FUS is GUGGU, which is rich in 5 untranslated regions (UTRs), and mutant FUS preferentially binds 3 -UTR and intron sequences [bib_ref] ALS-associated mutation FUS-R521C causes DNA damage and RNA splicing defects, Qiu [/bib_ref] [bib_ref] Biochemical Properties and Biological Functions of FET Proteins, Schwartz [/bib_ref] [bib_ref] The TET Family of Proteins: Functions and Roles in Disease, Tan [/bib_ref]. As an abundant nuclear protein, FUS forms stable complexes with numerous members of the hnRNP family, and some components of the FUS-hnRNP complex have RNA-binding activity [bib_ref] TLS (FUS) binds RNA in vivo and engages in nucleo-cytoplasmic shuttling, Zinszner [/bib_ref] [bib_ref] Identification of hnRNP P2 as TLS/FUS using electrospray mass spectrometry, Calvio [/bib_ref]. The physiological functions of FUS are apparently diverse. Here, we describe its role in RNA metabolism, DNA repair, and its relationship to disease.
## Role of fus in rna metabolism
FUS is well-studied for its transcriptional regulation function in cells. FUS binds to single-stranded DNA motifs in the promoter region of certain genes, which are transcribed by RNA polymerases II and III, and then accumulates near the transcription start site (TSS), resulting in changes of transcription levels [bib_ref] TLS/FUS (translocated in liposarcoma/fused in sarcoma) regulates target gene transcription via single-stranded..., Tan [/bib_ref] [bib_ref] TLS inhibits RNA polymerase III transcription, Tan [/bib_ref]. Additionally, FUS also regulates transcription levels partly via the interaction with coaggregates with other FET proteins in the pathologic inclusions of FTLD, while in all cases of ALS with FUS inclusions, the FUS mutant is dominant in the composition of ALS FUS inclusion [bib_ref] Role of FET proteins in neurodegenerative disorders, Svetoni [/bib_ref] [bib_ref] FET proteins TAF15 and EWS are selective markers that distinguish FTLD with..., Neumann [/bib_ref]. Further studies showed that normal FUS proteins are mainly located in the nucleus, whereas the mutants are primarily found in the cytoplasm. Most mutations alter the C-terminal nuclear localization signal, resulting in FUS redistribution from the nucleus to the cytoplasm. Excess FUS in the cytoplasm is involved in the formation of stress granules (SGs) during stress, which are membraneless granules composed of mRNAs, ribosome translation initiation factors, and other RNA binding proteins [bib_ref] ALS mutant FUS disrupts nuclear localization and sequesters wild-type FUS within cytoplasmic..., Caroline [/bib_ref] [bib_ref] Mutant FUS proteins that cause amyotrophic lateral sclerosis incorporate into stress granules, Bosco [/bib_ref] [bib_ref] FUS/TLS assembles into stress granules and is a prosurvival factor during hyperosmolar..., Sama [/bib_ref] [bib_ref] Stress granules as crucibles of ALS pathogenesis, Li [/bib_ref]. These granules can be induced by various cellular stresses such as oxidative stress, mitochondrial dysfunction, and viral infection that inhibit translation initiation [bib_ref] Dysregulated molecular pathways in amyotrophic lateral sclerosis-frontotemporal dementia spectrum disorder, Gao [/bib_ref].
FUS has two possible models for cytoplasmic aggregation leading to neurological diseases. One is the gain-of-function model, in which FUS gains a toxic function in the cytoplasm. FUS may sequester important regulators or trigger abnormal signaling pathways to alter cell physiology [bib_ref] Altered Ribostasis: RNA-Protein Granules in Degenerative Disorders, Ramaswami [/bib_ref]. The alternative model is that FUS aggregation depletes functional FUS, especially in the dendritic spines of neurons, which can cause instability of mRNA and trigger downstream immune responses such as inflammation, resulting in further neuronal damage [bib_ref] A new subtype of frontotemporal lobar degeneration with FUS pathology, Neumann [/bib_ref]. Loss of FUS in the nucleus affects transcription, alternative splicing, and also DNA repair [bib_ref] Mechanisms of FUS mutations in familial amyotrophic lateral sclerosis, Shang [/bib_ref]. FUS cytoplasmic aggregation may spread across the anatomical network in a prion-like manner, as observed by biophysical and histological analysis, ultimately leading to neurodegenerative disease [bib_ref] Neuronal cytoplasmic inclusions in tau, TDP-43, and FUS molecular subtypes of frontotemporal..., Richard [/bib_ref].
## Domains of the fus protein
FUS belongs to the FET/TET family, in which proteins share domains similar in both structure and function [bib_ref] Genomic structure of the human RBP56/hTAFII68 and FUS/TLS genes, Morohoshi [/bib_ref]. FUS contains an N-terminal transcriptional activation domain, which is a glutamine-glycine-serine-tyrosine-rich domain (QGSY-rich domain, amino acids 1-165); three arginine-glycine-glycine repetitive region (RGG, named RGG1-RGG3); an RNA recognition motif (RMM, amino acids 285-371); a Cys2-Cys2 zinc finger (ZnF) motif; a nuclear export signal (NES); and a C-terminus nonclassical nuclear positioning signal (NLS) [fig_ref] Figure 1: Schematic of fused in sarcoma [/fig_ref] [bib_ref] FUS-related proteinopathies: lessons from animal models, Lanson [/bib_ref] [bib_ref] TDP-43/FUS in motor neuron disease: Complexity and challenges, Guerrero [/bib_ref]. In addition, according to bioinformatics analysis, FUS has two prion-like domains (PrLDs) predicted to be amino acids 1-239 and 391-407 [bib_ref] Molecular Determinants and Genetic Modifiers of Aggregation and Toxicity for the ALS..., Sun [/bib_ref]. [bib_ref] Role of FET proteins in neurodegenerative disorders, Svetoni [/bib_ref] [bib_ref] FET proteins TAF15 and EWS are selective markers that distinguish FTLD with..., Neumann [/bib_ref]. Further studies showed that normal FUS proteins are mainly located in the nucleus, whereas the mutants are primarily found in the cytoplasm. Most mutations alter the Cterminal nuclear localization signal, resulting in FUS redistribution from the nucleus to the cytoplasm. Excess FUS in the cytoplasm is involved in the formation of stress granules (SGs) during stress, which are membraneless granules composed of mRNAs, ribosome translation initiation factors, and other RNA binding proteins [bib_ref] ALS mutant FUS disrupts nuclear localization and sequesters wild-type FUS within cytoplasmic..., Caroline [/bib_ref] [bib_ref] Mutant FUS proteins that cause amyotrophic lateral sclerosis incorporate into stress granules, Bosco [/bib_ref] [bib_ref] FUS/TLS assembles into stress granules and is a prosurvival factor during hyperosmolar..., Sama [/bib_ref] [bib_ref] Stress granules as crucibles of ALS pathogenesis, Li [/bib_ref]. These granules can be induced by various cellular stresses such as oxidative stress, mitochondrial dysfunction, and viral infection that inhibit translation initiation [bib_ref] Dysregulated molecular pathways in amyotrophic lateral sclerosis-frontotemporal dementia spectrum disorder, Gao [/bib_ref]. FUS has two possible models for cytoplasmic aggregation leading to neurological diseases. One is the gain-of-function model, in which FUS gains a toxic function in the cytoplasm. FUS may sequester important regulators or trigger abnormal signaling pathways to alter cell physiology [bib_ref] Altered Ribostasis: RNA-Protein Granules in Degenerative Disorders, Ramaswami [/bib_ref]. The alternative model is that FUS aggregation depletes functional FUS, especially in the dendritic spines of neurons, which can cause instability of mRNA and trigger downstream immune responses such as inflammation, resulting in further neuronal damage [bib_ref] A new subtype of frontotemporal lobar degeneration with FUS pathology, Neumann [/bib_ref]. Loss of FUS in the nucleus affects transcription, alternative splicing, and also DNA repair [bib_ref] Mechanisms of FUS mutations in familial amyotrophic lateral sclerosis, Shang [/bib_ref]. FUS cytoplasmic aggregation may spread across the anatomical network in a prion-like manner, as observed by biophysical and histological analysis, ultimately leading to neurodegenerative disease [bib_ref] Neuronal cytoplasmic inclusions in tau, TDP-43, and FUS molecular subtypes of frontotemporal..., Richard [/bib_ref].
## Domains of the fus protein
FUS belongs to the FET/TET family, in which proteins share domains similar in both structure and function [bib_ref] Genomic structure of the human RBP56/hTAFII68 and FUS/TLS genes, Morohoshi [/bib_ref]. FUS contains an N-terminal transcriptional activation domain, which is a glutamine-glycine-serine-tyrosine-rich domain (QGSY-rich domain, amino acids 1-165); three arginine-glycine-glycine repetitive region (RGG, named RGG1-RGG3); an RNA recognition motif (RMM, amino acids 285-371); a Cys2-Cys2 zinc finger (ZnF) motif; a nuclear export signal (NES); and a C-terminus nonclassical nuclear positioning signal (NLS) [fig_ref] Figure 1: Schematic of fused in sarcoma [/fig_ref] [bib_ref] FUS-related proteinopathies: lessons from animal models, Lanson [/bib_ref] [bib_ref] TDP-43/FUS in motor neuron disease: Complexity and challenges, Guerrero [/bib_ref]. In addition, according to bioinformatics analysis, FUS has two prion-like domains (PrLDs) predicted to be amino acids 1-239 and 391-407 [bib_ref] Molecular Determinants and Genetic Modifiers of Aggregation and Toxicity for the ALS..., Sun [/bib_ref]. The binding of FUS to RNA may be related to the ZnF motif. The RGG-ZnF-RGG domain is thought to be the major RNA-binding sequence [bib_ref] TLS (FUS) binds RNA in vivo and engages in nucleo-cytoplasmic shuttling, Zinszner [/bib_ref] [bib_ref] Domain Architectures and Characterization of an RNA-binding Protein, Iko [/bib_ref]. Furthermore, Lerga et al. identified that RNA oligoribonucleotides bind to the FUS protein via the GGUG motif [bib_ref] Identification of an RNA binding specificity for the potential splicing factor TLS, Lerga [/bib_ref]. Survival motor neuron (SMN) proteins are part of a large multiprotein complex that plays a vital role in the biogenesis of small nuclear ribonucleoprotein (snRNP) granules, which results in the fatal childhood motor neuron disease spinal muscular atrophy (SMA) [bib_ref] ALS-causative mutations in FUS/TLS confer gain and loss of function by altered..., Sun [/bib_ref]. further demonstrated that the RGG domain in FUS and the Tudor domain in SMN are necessary for the protein-protein interaction [bib_ref] The RNA-binding protein FUS/TLS is a common aggregate-interacting protein in polyglutamine diseases, Doi [/bib_ref]. The RGG2 domain recruits FUS to DNA damage sites, which is enhanced by PrLDs [bib_ref] The RNA-binding Protein Fused in Sarcoma (FUS) Functions Downstream of Poly(ADP-ribose) Polymerase..., Mastrocola [/bib_ref]. The C-terminal NLS of FUS is recognized by the nuclear input receptor, facilitating the transportation of the protein from the cytoplasm to the nucleus. The majority of mutations discovered in familial ALS is in the NLS. Loss-of-function mutations or deletion in this signaling motif prevents FUS from binding to transport proteins and being transported into the nucleus, and eventually, FUS accumulates in the cytoplasm [bib_ref] Rules for nuclear localization sequence recognition by karyopherin beta 2, Lee [/bib_ref] [bib_ref] Structural and energetic basis of ALS-causing mutations in the atypical proline-tyrosine nuclear..., Zhang [/bib_ref]. The binding of FUS to RNA may be related to the ZnF motif. The RGG-ZnF-RGG domain is thought to be the major RNA-binding sequence [bib_ref] TLS (FUS) binds RNA in vivo and engages in nucleo-cytoplasmic shuttling, Zinszner [/bib_ref] [bib_ref] Domain Architectures and Characterization of an RNA-binding Protein, Iko [/bib_ref]. Furthermore, Lerga et al. identified that RNA oligoribonucleotides bind to the FUS protein via the GGUG motif [bib_ref] Identification of an RNA binding specificity for the potential splicing factor TLS, Lerga [/bib_ref]. Survival motor neuron (SMN) proteins are part of a large multiprotein complex that plays a vital role in the biogenesis of small nuclear ribonucleoprotein (snRNP) granules, which results in the fatal childhood motor neuron disease spinal muscular atrophy (SMA) [bib_ref] ALS-causative mutations in FUS/TLS confer gain and loss of function by altered..., Sun [/bib_ref]. further demonstrated that the RGG domain in FUS and the Tudor domain in SMN are necessary for the protein-protein interaction [bib_ref] The RNA-binding protein FUS/TLS is a common aggregate-interacting protein in polyglutamine diseases, Doi [/bib_ref]. The RGG2 domain recruits FUS to DNA damage sites, which is enhanced by PrLDs [bib_ref] The RNA-binding Protein Fused in Sarcoma (FUS) Functions Downstream of Poly(ADP-ribose) Polymerase..., Mastrocola [/bib_ref]. The C-terminal NLS of FUS is recognized by the nuclear input receptor, facilitating the transportation of the protein from the cytoplasm to the nucleus. The majority of mutations discovered in familial ALS is in the NLS. Loss-of-function mutations or deletion in this signaling motif prevents FUS from binding to transport proteins and being transported into the nucleus, and eventually, FUS accumulates in the cytoplasm [bib_ref] Rules for nuclear localization sequence recognition by karyopherin beta 2, Lee [/bib_ref] [bib_ref] Structural and energetic basis of ALS-causing mutations in the atypical proline-tyrosine nuclear..., Zhang [/bib_ref].
## Fus self-assembly
## Fus self-assembly
## Prion-like domains and self-assembly of fus
Prions are self-replicating proteins with misfolded conformations and can lead to diseases, also known as infectious proteins. A single prion can act as a template to fold soluble proteins containing the same amino acid sequence into the prion conformation, resulting in self-replication [bib_ref] Prions as adaptive conduits of memory and inheritance, Shorter [/bib_ref]. Prions typically form stable amyloid fibers with a hallmark "cross-β" structure [bib_ref] Emergence and natural selection of drug-resistant prions, Shorter [/bib_ref] [bib_ref] Molecular mechanisms for protein-encoded inheritance, Wiltzius [/bib_ref]. Yeast prions such as Sup35, Ure2, and Rnq1 are rich in uncharged polar amino acids (glutamine, asparagine, tyrosine, and serine) and glycines, which are essential for prion-like propagation and amyloid formation, and those areas are referred to as the "prion domain". Adding this region to other innocuous proteins is sufficient to confer prion behavior [bib_ref] RNA-binding proteins with prion-like domains in health and disease, Harrison [/bib_ref]. Proteins containing prion-related Q/N-rich domains could induce conformational changes and finally lead to self-assembling, which alters protein functions such as SG formation and synaptic translation [bib_ref] Implications of the prion-related Q/N domains in TDP-43 and FUS, Udan [/bib_ref]. The most neurodegenerative diseases share similar cellular and biochemical mechanisms, which is the accumulation of misfolded proteins in brains. Thus, numerous reports suggest that aggregation-prone RNA-/DNA-binding proteins with PrLDs cause neurodegenerative disease by seed formation in a prion-like manner [bib_ref] Stress granules as crucibles of ALS pathogenesis, Li [/bib_ref] [bib_ref] Prion-like domains as epigenetic regulators, scaffolds for subcellular organization, and drivers of..., March [/bib_ref]. In other words, misfolded proteins act as seeds of aggregation, folding their primary subtypes and converting them into pathological aggregates, which ultimately generate neurological diseases through further recruitment and transformation [bib_ref] From nucleation to widespread propagation: A prion-like concept for ALS, Maniecka [/bib_ref]. A typical example of a pathogenic PrLD is that the trinucleotide is repeatedly amplified in the gene encoding ataxin 1 (ATXN1) in PrLDs and promotes aggregation of ATXN1, leading to polyglutamine protein production and, ultimately, to spinocerebellar ataxia type 1 [bib_ref] Prion-like domains as epigenetic regulators, scaffolds for subcellular organization, and drivers of..., March [/bib_ref] [bib_ref] Identification and characterization of the gene causing type 1 spinocerebellar ataxia, Banfi [/bib_ref] [bib_ref] Trinucleotide repeat disorders, Orr [/bib_ref] [bib_ref] Chaperone suppression of aggregation and altered subcellular proteasome localization imply protein misfolding..., Cummings [/bib_ref].
The experimental results from the interactions between FUS and polyglutamine inclusions and the predictions suggested that the FUS prion-like domains are located in the N-terminal QGSY domain and are part of the first RGG domain (amino acids 1-239), with an additional region in the RGG3 domain (amino acids 391-405) [bib_ref] Prion-like disorders: blurring the divide between transmissibility and infectivity, Cushman [/bib_ref]. By using a prediction algorithm based on a yeast prion domain to score 27,879 human proteins, FUS ranked 15th for its prion-like property and 1st in RNA binding proteins [bib_ref] Mechanisms of FUS mutations in familial amyotrophic lateral sclerosis, Shang [/bib_ref]. Prion-like domains in RNA-binding proteins are essential for neuronal proteins to enter ribonucleoprotein granules, forming a functional assembling state, which also drives pathological protein aggregation in neurons. It can therefore be speculated that the prion-like domain in FUS is essential for its aggregation in neurons [bib_ref] The tip of the iceberg: RNA-binding proteins with prion-like domains in neurodegenerative..., King [/bib_ref]. It was reported that the PrLD of FUS can be assembled into amyloid-like fibrils in a cell-free system and the morphology of aggregates is similar to the general pathogenic amyloid fibers. Reports showed that the process is reversible and highly regulated. Amyloid fibrils formed by FUS are susceptible to depolymerization due to a variety of factors, including FUS concentration, DNA or RNA levels, and SYGQ-rich domain phosphorylation [bib_ref] Molecular Determinants and Genetic Modifiers of Aggregation and Toxicity for the ALS..., Sun [/bib_ref] [bib_ref] Cell-free Formation of RNA Granules: Bound RNAs Identify Features and Components of..., Han [/bib_ref] [bib_ref] FUS/TLS forms cytoplasmic aggregates, inhibits cell growth and interacts with TDP-43 in..., Kryndushkin [/bib_ref] [bib_ref] A Yeast Model of FUS/TLS-Dependent Cytotoxicity, Ju [/bib_ref] [bib_ref] Phosphorylation-Regulated Binding of RNA Polymerase II to Fibrous Polymers of Low-Complexity Domains, Kwon [/bib_ref] [bib_ref] Cell-free Formation of RNA Granules: Low Complexity Sequence Domains Form Dynamic Fibers..., Kato [/bib_ref].
Phase separation is a physical process in which a supersaturated solution of components spontaneously separates into two phases with different densities, both stable and coexisting [bib_ref] Phase Separation of FUS Is Suppressed by Its Nuclear Import Receptor and..., Hofweber [/bib_ref]. This phenomenon is common in polymer chemistry and has recently been found in biomacromolecules [bib_ref] Binary-liquid phase separation of lens protein solutions, Broide [/bib_ref]. For example, FUS undergoes liquid-liquid phase separation (LLPS) both in vivo and in vitro at physiological concentrations [bib_ref] Residue-by-Residue View of In Vitro FUS Granules that Bind the C-Terminal Domain..., Burke [/bib_ref]. It is progressively recognized that phase separation controls the formation of membrane-free organelles, regulating biological functions and activities [bib_ref] Phase separation in biology; functional organization of a higher order, Mitrea [/bib_ref]. FUS is rapidly recruited to the site of DNA damage, then phase separation allows FUS to locally form droplet-like compartments, increasing the concentration of FUS and other substances and possibly promoting DNA repair [bib_ref] RNA-binding proteins protect from near and far, Dutertre [/bib_ref] [bib_ref] Liquids and Mesoscale Organization of Cytoplasm, Hyman [/bib_ref].
Studies have shown that FUS has three forms of existence-dispersion, droplets, and hydrogels-and the reversible dynamic phase transition is driven by the low-complexity (LC) domain at the N-terminus of FUS [bib_ref] Phase Transitions in the Assembly of Multi-Valent Signaling Proteins, Li [/bib_ref]. According to the results of Patel et al., FUS closed compartments have two states of droplets and hydrogels which are reversible and dynamic [bib_ref] A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease..., Patel [/bib_ref]. Their experiments showed that increasing the concentration of FUS in vivo or in vitro causes its transformation from liquid to an aggregated state. Under stress conditions, FUS rapidly passes through the nucleus to the cytoplasm, then forms SGs [bib_ref] TDP-43 and FUS: a nuclear affair, Dormann [/bib_ref]. SGs act as condensation sites, where the FUS concentration is locally increased and eventually may aggregate [bib_ref] Converging mechanisms in ALS and FTD: Disrupted RNA and protein homeostasis, Ling [/bib_ref].
The basis of the reversible phase separation of FUS is the self-assembly of the LC domain [bib_ref] Cell-free Formation of RNA Granules: Bound RNAs Identify Features and Components of..., Han [/bib_ref]. Once the self-assembly of the LC domain goes beyond normality, usually caused by mutations in ALS, FUS is driven to form a poorly soluble and stable hydrogel [fig_ref] Figure 2: Self-assembling model of FUS [/fig_ref] [bib_ref] Mechanisms of FUS mutations in familial amyotrophic lateral sclerosis, Shang [/bib_ref] [bib_ref] Decoding ALS: from genes to mechanism, Taylor [/bib_ref]. Murakami et al. demonstrated that pathogenic FUS mutations reduce the ability for reversible phase transitions. More specifically, the liquid-to-solid transition is accelerated by FUS mutants, indicating that FUS phase transitions are closely related to disease. In addition, the repeated cycles of gelation and degelation of FUS promote it to form an irreversible hydrogel [bib_ref] ALS/FTD Mutation-Induced Phase Transition of FUS Liquid Droplets and Reversible Hydrogels into..., Murakami [/bib_ref]. The droplets and hydrogels formed by the LC domains spontaneously mature into fibrous solid aggregates as the cell ages and the ability of the quality control mechanism diminishes [bib_ref] A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease..., Patel [/bib_ref]. However, the rest of the FUS domains will reduce the tendency of FUS assembly to hinder the formation of high-grade assemblies of the full-length protein.
Meanwhile, specific cellular proteins, such as the receptors transportin (TNPO1) and karyopherin-b2 (Kapb2), inhibit the solidification of FUS [bib_ref] Are aberrant phase transitions a driver of cellular aging?, Alberti [/bib_ref] [bib_ref] Reentrant Phase Transition Drives Dynamic Substructure Formation in Ribonucleoprotein Droplets, Banerjee [/bib_ref]. Structural analysis has shown that the LC domain noncovalently aggregates into morphologically uniform amyloid fibrils in FUS protein hydrogels [bib_ref] Phase Separation of FUS Is Suppressed by Its Nuclear Import Receptor and..., Hofweber [/bib_ref] , with a structure similar to FUS aggregates in the cytoplasm of spinal motor neurons of FUS mutant patients (with amyloid protein-like prominent filamentous cross-β structure) [bib_ref] Mutations in FUS, an RNA processing protein, cause familial amyotrophic lateral sclerosis..., Vance [/bib_ref]. The basis of the reversible phase separation of FUS is the self-assembly of the LC domain [bib_ref] Cell-free Formation of RNA Granules: Bound RNAs Identify Features and Components of..., Han [/bib_ref]. Once the self-assembly of the LC domain goes beyond normality, usually caused by mutations in ALS, FUS is driven to form a poorly soluble and stable hydrogel [fig_ref] Figure 2: Self-assembling model of FUS [/fig_ref] [bib_ref] Mechanisms of FUS mutations in familial amyotrophic lateral sclerosis, Shang [/bib_ref] [bib_ref] Decoding ALS: from genes to mechanism, Taylor [/bib_ref]. Murakami et al. demonstrated that pathogenic FUS mutations reduce the ability for reversible phase transitions. More specifically, the liquid-to-solid transition is accelerated by FUS mutants, indicating that FUS phase transitions are closely related to disease. In addition, the repeated cycles of gelation and degelation of FUS promote it to form an irreversible hydrogel [bib_ref] ALS/FTD Mutation-Induced Phase Transition of FUS Liquid Droplets and Reversible Hydrogels into..., Murakami [/bib_ref]. The droplets and hydrogels formed by the LC domains spontaneously mature into fibrous solid aggregates as the cell ages and the ability of the quality control mechanism diminishes [bib_ref] A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease..., Patel [/bib_ref]. However, the rest of the FUS domains will reduce the tendency of FUS assembly to hinder the formation of high-grade assemblies of the full-length protein.
Meanwhile, specific cellular proteins, such as the receptors transportin (TNPO1) and karyopherin-b2 (Kapb2), inhibit the solidification of FUS [bib_ref] Are aberrant phase transitions a driver of cellular aging?, Alberti [/bib_ref] [bib_ref] Reentrant Phase Transition Drives Dynamic Substructure Formation in Ribonucleoprotein Droplets, Banerjee [/bib_ref]. Structural analysis has shown that the LC domain noncovalently aggregates into morphologically uniform amyloid fibrils in FUS protein hydrogels [bib_ref] Phase Separation of FUS Is Suppressed by Its Nuclear Import Receptor and..., Hofweber [/bib_ref] , with a structure similar to FUS aggregates in the cytoplasm of spinal motor neurons of FUS mutant patients (with amyloid protein-like prominent filamentous cross-β structure) [bib_ref] Mutations in FUS, an RNA processing protein, cause familial amyotrophic lateral sclerosis..., Vance [/bib_ref].
## Amyloid core in lc domains of fus regulates its self-assembly
Recent solid-state nuclear magnetic resonance (NMR) studies have indicated a structurally ordered amyloid core in the LC domain of FUS, also known as the prion-like domain. At the same time, it was discovered that the residues 39-95 constitute the structurally ordered core of the prionlike domain fibers in combination with the computational assistance method based on the MCASSIGN algorithm [bib_ref] A general Monte Carlo/simulated annealing algorithm for resonance assignment in NMR of..., Hu [/bib_ref]. The droplets and hydrogels formed by FUS are stabilized by hydrogen bonds between the antiparallel β-sheet conformations. The hallmark structure of pathogenic amyloid fibers is an in-register parallel cross-β structure in which the β-strands are perpendicular to the fiber axis, generally formed from 2 to 9 repetitive residues [bib_ref] Structure of the cross-β spine of amyloid-like fibrils, Nelson [/bib_ref]. β-strands, irregular crimps, and loops constitute the backbone conformation of the fibers and the β-sheet layer interacts by van der Waals forces [bib_ref] The Amyloid State of Proteins in Human Diseases, Eisenberg [/bib_ref] [bib_ref] Amyloid fibrils Abnormal protein assembly, Rambaran [/bib_ref]. Sayawa et al. named this structure a "stereo zipper" that is highly stable and generally resistant to high temperatures, proteases, and detergents [bib_ref] Atomic structures of amyloid cross-β spines reveal varied steric zippers, Sawaya [/bib_ref]. The FUS-LC fibrils have a single cross-β unit, with all molecules having the same structural environments and conformations and a cross-β-core-aligned parallel intermolecular alignment.
## Amyloid core in lc domains of fus regulates its self-assembly
Recent solid-state nuclear magnetic resonance (NMR) studies have indicated a structurally ordered amyloid core in the LC domain of FUS, also known as the prion-like domain. At the same time, it was discovered that the residues 39-95 constitute the structurally ordered core of the prion-like domain fibers in combination with the computational assistance method based on the MCASSIGN algorithm [bib_ref] A general Monte Carlo/simulated annealing algorithm for resonance assignment in NMR of..., Hu [/bib_ref]. The droplets and hydrogels formed by FUS are stabilized by hydrogen bonds between the antiparallel β-sheet conformations. The hallmark structure of pathogenic amyloid fibers is an in-register parallel cross-β structure in which the β-strands are perpendicular to the fiber axis, generally formed from 2 to 9 repetitive residues [bib_ref] Structure of the cross-β spine of amyloid-like fibrils, Nelson [/bib_ref]. β-strands, irregular crimps, and loops constitute the backbone conformation of the fibers and the β-sheet layer interacts by van der Waals forces [bib_ref] The Amyloid State of Proteins in Human Diseases, Eisenberg [/bib_ref] [bib_ref] Amyloid fibrils Abnormal protein assembly, Rambaran [/bib_ref]. Sayawa et al. named this structure a "stereo zipper" that is highly stable and generally resistant to high temperatures, proteases, and detergents [bib_ref] Atomic structures of amyloid cross-β spines reveal varied steric zippers, Sawaya [/bib_ref]. The FUS-LC fibrils have a single cross-β unit, with all molecules having the same structural environments and conformations and a cross-β-core-aligned parallel intermolecular alignment.
The overall bone architecture of residues 43-95 and the orientation of the side chains of residues 44-52 and 65-95 have been determined, while residues 55-62 form a dynamic loop. The fragments of the β-strands forming the cross-β contain residues 44-46, 52-54, 62-64, 67-70, 85-90, and 93-95, determined from the characteristic arrangement of the backbone carbonyl and the fiber growth axis. This also indicates that many of the structurally ordered cores of the FUS-LC fibers do not participate in the formation of hydrogen bonds, which makes the cross-β structure more stable. Probably because of the more rational arrangements, residues 39-95 dynamically convert into stable cross-β structures, driven by entropy cost or desolvation energy [bib_ref] Role of water in Protein Aggregation and Amyloid Polymorphism, Thirumalai [/bib_ref]. From the structural model, it is known that S84 forms a hydrogen bond network with Y75 and T78, and the stability of the loop between 74 and 87 is related to this hydrogen bond network. Phosphorylation sites S84 and S87 are located between the loops 74 and 87; therefore, the disruption of the loop conformation may be critical to the formation of FUS-LC fibers. In addition, the folding of residues 44-50 against residues 64-80 is stabilized via the interaction of S48 with Q69 and T71 and the interaction of T45 and T47 with S77. The folding of residues 69-73 to 90-95 may be stabilized by the interaction of S70, Q73, S90, Q93, and S95 with each other. Y50 and Y66 may participate in π-stacking interactions (attractive, noncovalent interactions between aromatic rings) which help in stabilizing the core structure. In conclusion, hydrogen bonding and electrostatic dipole-dipole interactions between polar side chains may improve the stability of the PrLDs of the FUS structure. The prion-like domain fibers of FUS lack pure hydrophobic side chains, which are conducive to disassembly [bib_ref] Cell-free Formation of RNA Granules: Low Complexity Sequence Domains Form Dynamic Fibers..., Kato [/bib_ref].
Liu et al. found that 37SYSGYS42 and 54SYSSYG59 in the amyloid core region 39-95 can form reversible fibers that are sensitive to temperature and phosphorylation, termed the reversible amyloid nucleus [bib_ref] Atomic structures of FUS LC domain segments reveal bases for reversible amyloid..., Luo [/bib_ref]. Reversible amyloid cores (RACs) are assembled at low temperatures and disassembled as the temperature increases. The absence of either 37SYSGYS42 (RAC1) or 54SYSSYG59 (RAC2) significantly reduces the ability of self-assembling for FUS-LC; thus, it is considered the core area for FUS self-assembly [bib_ref] Atomic structures of FUS LC domain segments reveal bases for reversible amyloid..., Luo [/bib_ref]. Using microelectronics and X-ray diffraction techniques, Liu et al. determined that RAC1 is not a fiber spine of β-strands but a coil arranged along the fiber axis with Gly40 as a kink point. The hydroxyl group of the same layer of Tyr38 forms a hydrogen bond with the carbonyl group of Tyr41 to block the coil, and the adjacent sheets of Tyr38 and Ser42 also form hydrogen bonds, which facilitates the formation of a hydrogen bond network [bib_ref] Atomic structures of FUS LC domain segments reveal bases for reversible amyloid..., Luo [/bib_ref]. When the formation of these hydrogen bonds is disrupted, such as by Ser42 phosphorylation, the RAC1 fiber disassembles [bib_ref] FUS is phosphorylated by DNA-PK and accumulates in the cytoplasm after DNA..., Qiudong [/bib_ref]. The π-stacking of Tyr38 enhances the interaction of adjacent layers. The RAC1 fibril spine uses polar residues to form the interface between the mating layers, which is different from the hydrophobic interface composed of nonpolar residues of pathogenic fibers [bib_ref] Recent atomic models of amyloid fibril structure, Nelson [/bib_ref]. Additionally, the interaction between the pathogenic fibrous sheets also has hydrogen bonds and van der Waals forces, resulting in a relatively stable fiber structure [bib_ref] Atomic structures of FUS LC domain segments reveal bases for reversible amyloid..., Luo [/bib_ref] [bib_ref] The activities of amyloids from a structural perspective, Riek [/bib_ref]. Unlike RAC1, in the crystal structure, the RAC2 fiber contains a cross-β structure, the marker of the pathogenic fiber. However, in the RAC2 fibril spine, water molecules, together with the hydroxyl groups of Tyr58, form a hydrogen bond, so that the distance between the sheets increases to 13 Å, while the spacing of the layers of normal pathogenic fibers is~10 Å. The interlayer bonding is not tight and the fiber is remarkably unstable [bib_ref] Atomic structures of amyloid cross-β spines reveal varied steric zippers, Sawaya [/bib_ref] [bib_ref] Atomic structures of FUS LC domain segments reveal bases for reversible amyloid..., Luo [/bib_ref]. The unique structure of the fibers formed by RAC1 and RAC2 makes them reversible, facilitating FUS reversible self-assembly [bib_ref] Cell-free Formation of RNA Granules: Bound RNAs Identify Features and Components of..., Han [/bib_ref] [bib_ref] Atomic structures of FUS LC domain segments reveal bases for reversible amyloid..., Luo [/bib_ref].
## Regulation factors of fus self-assembly
The physiological function of FUS is performed while in a droplet state. In vitro, FUS droplets are amorphous, and it is speculated that, in vivo, FUS may have no ordered structure, and the oligomers formed may be gradually decomposed [bib_ref] Dynamic droplets: the role of cytoplasmic inclusions in stress, function, and disease, Amen [/bib_ref] [bib_ref] The Role of Post-Translational Modifications on Prion-Like Aggregation and Liquid-Phase Separation of..., Rhoads [/bib_ref]. Currently, the confirmed mechanism for driving droplet formation is through its prion-like LC domains [bib_ref] A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease..., Patel [/bib_ref]. It can therefore be hypothesized that increasing the interaction of the LC domains accelerates the formation of aggregates. Conversely, weakening the interaction of the LC domain can be detrimental to liquid-liquid phase separation [bib_ref] Residue-by-Residue View of In Vitro FUS Granules that Bind the C-Terminal Domain..., Burke [/bib_ref]. Irreversible FUS aggregates selectively capture ribonucleoproteins, thereby disrupting RNP granule function and impeding the synthesis of new proteins at the axon terminals of neuronal cells [bib_ref] ALS/FTD Mutation-Induced Phase Transition of FUS Liquid Droplets and Reversible Hydrogels into..., Murakami [/bib_ref] [bib_ref] Aggregation of ALS-linked FUS mutant sequesters RNA binding proteins and impairs RNA..., Takanashi [/bib_ref]. DNA-PK phosphorylates the LC domain of FUS [bib_ref] Identification and characterization of FUS/TLS as a new target of ATM, Mary [/bib_ref]. Tycko et al. identified 14 phosphorylation sites, including [bib_ref] Mutations in the FUS/TLS Gene on Chromosome 16 Cause Familial Amyotrophic Lateral..., Kwiatkowski [/bib_ref] [bib_ref] Dysregulated molecular pathways in amyotrophic lateral sclerosis-frontotemporal dementia spectrum disorder, Gao [/bib_ref] [bib_ref] Rules for nuclear localization sequence recognition by karyopherin beta 2, Lee [/bib_ref] [bib_ref] Prion-like domains as epigenetic regulators, scaffolds for subcellular organization, and drivers of..., March [/bib_ref] [bib_ref] Decoding ALS: from genes to mechanism, Taylor [/bib_ref] [bib_ref] Reentrant Phase Transition Drives Dynamic Substructure Formation in Ribonucleoprotein Droplets, Banerjee [/bib_ref] [bib_ref] Mutations in FUS cause FALS and SALS in French and French Canadian..., Belzil [/bib_ref] [bib_ref] Mutations of FUS gene in sporadic amyotrophic lateral sclerosis, Lucia [/bib_ref] [bib_ref] Novel missense and truncating mutations in FUS/TLS in familial ALS, Waibel [/bib_ref] , and 142 as well as Thr residues 7, 11, 19, and 68. In the fibril core formation segment, the phosphorylation of Ser and Thr residues disrupts the interaction of the cross-β, inhibiting hydrogel binding and droplet formation [bib_ref] Cell-free Formation of RNA Granules: Bound RNAs Identify Features and Components of..., Han [/bib_ref]. RAC1 does not form amyloid fibrils when Ser42 is phosphorylated, and S42D mutation significantly inhibits phase separation of FUS-LC [bib_ref] Atomic structures of FUS LC domain segments reveal bases for reversible amyloid..., Luo [/bib_ref].
Numerous studies have shown that the regions of FUS other than the LC domains, especially the C-terminal domain, are important in the phase transition behavior of FUS. Qamar et al. demonstrated that the synergistic cation-π interaction between the tyrosine of the LC domain and arginine of the structured C-terminal is related to phase separation [bib_ref] FUS Phase Separation Is Modulated by a Molecular Chaperone and Methylation of..., Qamar [/bib_ref]. The electrons of the tyrosine benzene ring form a cation-π interaction with the protons in the guanidino moiety of arginine, resulting in a local high concentration of the LC domains. Methylation of arginine is a common post-translational modification of RNA-binding proteins, and the C-terminal arginine rich in FUS is modified by methyltransferase with asymmetric dimethyl groups [bib_ref] Protein arginine methyltransferase 1 and 8 interact with FUS to modify its..., Chiara [/bib_ref]. This changes the local hydrophobicity and hydrogen bonding of arginine, thereby affecting the cation-π interaction [bib_ref] Chemical biology of protein arginine modifications in epigenetic regulation, Fuhrmann [/bib_ref]. The degree of methylation of arginine in the C-terminus regulates FUS assembly. Arginine hypomethylation at the C-terminus domain greatly accelerates phase separation and gelation, allowing FUS to form a stable hydrogel containing β-sheets. This is consistent with hypomethylation in FUS-related patients with frontotemporal degeneration [bib_ref] Phase Separation of FUS Is Suppressed by Its Nuclear Import Receptor and..., Hofweber [/bib_ref] [bib_ref] Arginine methylation next to the PY-NLS modulates Transportin binding and nuclear import..., Dormann [/bib_ref]. Experiments indicated that unmethylated FUS accounts for more than 1% of the total amount of FUS, which can induce the dispersed FUS to assemble into droplets and, in turn, become a stable fibrous gel [bib_ref] FUS Phase Separation Is Modulated by a Molecular Chaperone and Methylation of..., Qamar [/bib_ref].
The nuclear import receptor TNPO1 and Kapb2 regulate FUS phase separation by binding to FUS via its C-terminal domain [bib_ref] Rules for nuclear localization sequence recognition by karyopherin beta 2, Lee [/bib_ref]. They exist in the axonal terminal compartment of neurons, mediating nuclear import of FUS. TNPO1 acts as a molecular chaperone to reduce the tendency of FUS to form stress granules, suppressing FUS phase separation and liquid-solid phase transition. TNPO1 plays a central role in the FUS quality control mechanism. The highly stable combination of Kapb2 and FUS NLS weakens and dynamically interacts with the self-assembly-related regions, hindering the self-assembly of FUS. In vivo, Kapb2 dissociates FUS and RNA, regulating cytoplasmic RNA particles [bib_ref] Nuclear Import Receptor Inhibits Phase Separation of FUS through Binding to Multiple..., Yoshizawa [/bib_ref]. In vitro, Kapb2 dissolves the FUS phase separated liquid and aberrant fibrous hydrogels [bib_ref] Nuclear-Import Receptors Reverse Aberrant Phase Transitions of RNA-Binding Proteins with Prion-like Domains, Guo [/bib_ref].
It is currently known that over 50 mutations/deletions in FUS are associated with ALS, accounting for about 5% of familial cases (FALS) and 0.7-1.8% of sporadic cases (SALS) [bib_ref] Role of FET proteins in neurodegenerative disorders, Svetoni [/bib_ref]. Mutations in FUS mainly occur in exons 3-6 and exons 12-15. According to statistics, approximately two-thirds of FUS mutations are in exons 12-15, which encode ZnF motif, RGG2 and RGG3, and nuclear localization signaling domains. In particular, more than 50% of ALS-associated FUS mutations were found in the C-terminal NLS domain. Exons 3-6 encode the QGSY-rich and RGG1 regions, and about one-third of the mutations are in them [bib_ref] Mutations in the FUS/TLS Gene on Chromosome 16 Cause Familial Amyotrophic Lateral..., Kwiatkowski [/bib_ref] [bib_ref] Mutations in FUS, an RNA processing protein, cause familial amyotrophic lateral sclerosis..., Vance [/bib_ref] [bib_ref] The role of FUS gene variants in neurodegenerative diseases, Deng [/bib_ref]. The known mutations of FUS are shown in [fig_ref] Table 1: Summarizes of the mutations in the FUS gene [/fig_ref]. A series of mutations regulate the self-assembling of FUS. NLS mutations, on the one hand, can reduce the nuclear import, increasing the concentration of cytoplasmic FUS and promoting the formation of a more durable and mature fiber structure; on the other hand, they can change the kinetics of phase transition [bib_ref] A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease..., Patel [/bib_ref] [bib_ref] ALS/FTD Mutation-Induced Phase Transition of FUS Liquid Droplets and Reversible Hydrogels into..., Murakami [/bib_ref]. It has been shown that the NLS mutant of FUS reduces the binding of TNPO1, such as P525L and R495X, affecting the nuclear import of FUS [bib_ref] Phase Separation of FUS Is Suppressed by Its Nuclear Import Receptor and..., Hofweber [/bib_ref]. The NLS mutations of FUS cause the compulsive accumulation of FUS in the cytoplasm, resulting in aggregates that may produce toxic effects and trigger neuronal degeneration [bib_ref] The role of FUS gene variants in neurodegenerative diseases, Deng [/bib_ref]. In 2011, Suzuki and colleagues constructed a series of FUS-deleted domains and explored the SG formation of them. The results showed that C-terminal deletion formed a large number of stress granules, indicating that the C-terminus is important in stress granule formation. Later studies showed that mutation R514S or P525L formed SGs in HeLa cells, and double mutations of R514S and P525L or triple mutations of R514S, R521C, and P525L promoted the formation of stress granules [bib_ref] A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease..., Patel [/bib_ref]. Mutations in PrLDs accelerate the conversion of liquid to hydrogels. The G156E mutant tended to increase the aggregation of FUS in a manner of seed aggregation, causing rapid progression of ALS by using aggregated protein fibrils as a structural template for promoting nonaggregated protein fibrosis [bib_ref] ALS/FTD Mutation-Induced Phase Transition of FUS Liquid Droplets and Reversible Hydrogels into..., Murakami [/bib_ref] [bib_ref] Intranuclear Aggregation of Mutant FUS/TLS as a Molecular Pathomechanism of Amyotrophic Lateral..., Nomura [/bib_ref]. [bib_ref] Screening of the SOD1, FUS, TARDBP, ANG, and OPTN mutations in Korean..., Kwon [/bib_ref] Unlike the mechanism by which the mutated FUS locates in the cytoplasm, the FALS-associated FUS mutation enhances the interaction between FUS and SMN and reduces its levels at the normal expression sites, thereby impairing motor neurons [bib_ref] ALS-causative mutations in FUS/TLS confer gain and loss of function by altered..., Sun [/bib_ref]. Primarily mutant FUS impairs neuronal survival and causes defects in dendritic growth and synaptic connectivity by interfering with the ability of FUS in DNA damage repair and RNA splicing, leading to neurological disorders [bib_ref] Involvement of the pro-oncoprotein TLS (translocated in liposarcoma) in nuclear factor-kappa B..., Uranishi [/bib_ref]. Mutations of FUS in the glycine-rich domain (amino acids 156-262) and the C-terminal domain (amino acids 450-526) affect its interaction with the chromatin remodeling factor histone deacetylase 1, which plays a fundamental role in DNA repair and genomic stability maintenance via interaction with FUS. Experiments by showed that there is no wild-type FUS and interaction in FUS-R521C transgenic mice due to the mutated FUS-R521C protein in those mice, with HDAC1 forming a more stable complex with the wild-type FUS protein, implying that the mutated FUS acquired abnormal functional properties [bib_ref] ALS-associated mutation FUS-R521C causes DNA damage and RNA splicing defects, Qiu [/bib_ref]. The study of Nishimoto et al. indicated that FUS affects the integrity of paraspeckles (subnuclear structures that regulate gene expression by nuclear retention of RNA) by interacting with a long noncoding RNA nuclear-enriched abundant transcript (NEAT1), and mutations of FUS may disrupt integrity and maintenance of paraspeckles [bib_ref] Investigating FUS variation in Parkinson's disease, Labbé [/bib_ref] [bib_ref] The long non-coding RNA nuclear-enriched abundant transcript 1_2 induces paraspeckle formation in..., Nishimoto [/bib_ref]. In the same year, cytoplasmic aggregates of P45NRB/NONO, which are core proteins of paraspeckles, were found in spinal motoneurons in transgenic mice expressing truncated FUS mutant proteins (amino acids 1-359) [bib_ref] Fused in Sarcoma (FUS) Protein Lacking Nuclear Localization Signal (NLS) and Major..., Shelkovnikova [/bib_ref]. This verified the FUS-mutation-disrupting integrity of paraspeckles.
## Expectations
Most neurodegenerative diseases are associated with the misfolding of proteins. A growing number of studies have shown that FUS mutations in prion-like domains mediate the aggregation of FUS, which is an important reason for the development and progression of most neurodegenerative diseases. The mechanism of action of prion-associated regions of FUS needs to be further investigated, such as under what conditions the pathological aggregation of FUS occurs and whether they can play a role in disease progression by allowing the cell-cell metastasis of pathological protein aggregates. Although emergency particle assembly and prion proliferation has been confirmed, whether FUS mutations associated with neurodegenerative diseases cause normal function loss, a gain of toxic properties of FUS aggregates, or a combination of the both requires clarification and further studies. In addition, based on known mechanisms, the research of small-molecule-targeting inhibitors of FUS is also a topic of particular interest regarding finding a cure for related diseases. We are optimistic that with the deepening of FUS research, this field will make a major breakthrough.
Author Contributions: C.C. wrote the manuscript. S.X. contributed to discussions and revised the manuscript. X.D. and N.A. revised the manuscript. This review was written based on the design of S.-Z.L.
## Conflicts of interest:
The authors declare no conflict of interest.
[fig] Figure 1: Schematic of fused in sarcoma (FUS) protein domains. [/fig]
[fig] Figure 2: Self-assembling model of FUS. FUS can phase separate to form granules and reversible aggregates. Once the self-assembly of the low-complexity (LC) domain goes beyond normality, usually caused by mutations in neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS), FUS will form irreversible aggregation, possibly via liquid-liquid phase separation (LLPS). [/fig]
[fig] Funding: This work was supported by the National Natural Science Foundation of China (91853116, 21672019), the Fundamental Research Funds for the Central Universities and Research projects on biomedical transformation of China-Japan Friendship Hospital (No. PYBZ1812PYBZ1815). [/fig]
[table] Table 1: Summarizes of the mutations in the FUS gene. ALS: amyotrophic lateral sclerosis; FTLD, frontotemporal lobar degeneration; ET, essential tremor.Table 1. Cont. [/table]
|
Prostaglandin E1 in the medical management of erectile dysfunction in a genito-urinary medicine clinic.
Fifty consecutive patients with erectile failure who had previously proved refractory to papaverine and phentolamine intracavernosal therapy or were inappropriate candidates for such treatment were treated with intracavernosal prostaglandin E,. Forty patients (80%) achieved an erection sufficient forsexual intercourse and after a mean follow-up period of5.9 months, 32 patients were continuing to use treatment successfully. The average dose was 14 micrograms (range 2.5 to 30 micrograms). There were no cases of priapism or cavernosal fibrosis and no systemic side effects. A low incidence (8%) of local discomfort was reported. We conclude that prostaglandin is a safe and effective vasoactive agent for the treatment of erectile failure in a genito-urinary outpatient clinic.
# Introduction
The use of intracavernous vasoactive agents has been shown to be effective in the medical treatment of erectile dysfunction. Virag ' originally described the use of the smooth muscle relaxant papaverine while Brindley 2 demonstrated the effectiveness of the alpha adrenergic receptor blocker phenoxybenzamine. A combination of papaverine plus phentolamine (an alpha-blocker) was subsequently shown to be more effective than either agent alone 3'4. Ishii [bib_ref] Intracavernous injection of prostaglandin E, for the treatment of erectile impotence, Ishii [/bib_ref] has described the use of prostaglandin-E1, and studies have since demonstrated its efficacy in association with a reduced incidence of serious side effects6'7'8 We are unaware of reports of the use of this drug in a genito-urinary clinic, and describe our experience. PATIENTS AND METHODS Patients with erectile dysfunction were referred to the Genito-Clrinary Medicine Clinic, which provides a weekly clinic for erectile dysfunction, from their general practitioners or from consultant staff within the Northern Ireland hospital system. After a detailed sexual and medical history a full physical examination was performed and blood was taken for haemoglobin concentration, erythrocyte sedimentation rate, plasma glucose, liver function tests, testo-sterone and prolactin measurement. Intracavernous Prostaglandin El was first used in this clinic in November 1991, and this paper describes the first 50 patients. Those selected fulfilled one or more of the following criteria: (1) Previous failure to achieve satisfactory erectile response to papaverine 30 mg plus phentolamine 1 mg. (2) Previous priapism associated with papaverine plus phentolamine treatment or (3) Patients with a history of serious ischaemic heart disease but who were currently stable. The latter group of patients were assessed by a consultant cardiologist to ensure that treatment for their impotence was appropriate. Intracavernous injection of 10 ,ug of prostaglandin was given with a fine bore needle (26 gauge) into the lateral aspect of one of the corpora cavernosa at a site 1-2 cms proximal to the coronal sulcus. It was not our routine practice to use a constricting band at the base of the penis at the time of injection. The drug was supplied in 1 ml ampoules containing 10 jg prostaglandin El. These were individually prepared at the pharmacy of the Royal Victoria Hospital from 500,g ampoules (Upjohn). The initial dosage was 10 ,ug with appropriate titration according to response up to a maximum of 30 ,ug. Once an erection sufficient for sexual intercourse was achieved the patient was instructed in the technique of self-injection and supervised to ensure that their technique was correct. A supply of needles, syringes and prostaglandin El, was then provided for home use with advice that the injection should not be used more than twice per week; the patient attended on a regular basis to monitor effectiveness, side effects and to provide new supplies. Those who did not achieve an erection sufficient for sexual intercourse despite maximal dosage of 30 jig were assessed for a mechanical suction device. RESULTS Fifty patients satisfied the criteria for treatment between November 1991 and September 1992. This represented approximately one third of patients on intracavernosal treatment. Their average age was 52.5 years (range 25-74) with a mean duration of impotence of 5.7 years (range 6 months -30 years). The aetiology was organic in 44 cases, psychogenic in 3 cases and uncertain in three cases. Five patients had had previous episodes of priapism.
Alcohol-related 2
Forty patients (80%) achieved an erection sufficient for sexual intercourse; ten failed. In those treated successfully the average dose was 14 jg (range 2.5-30), the average duration of erection being 80 minutes (range 20 minutes -4 hours).
Of the original 40 patients with a satisfactory response 32 are still attending and regularly using injection therapy after a mean follow-up period of 5.9 months (range 1-11 months). Eight patients have failed to continue to attend for unknown reasons. Four patients complained of local pain in the penis, usually burning in nature and associated with the erection. The aetiology of erectile failure in these cases was vasculogenic in two, diabetic in one and psychogenic in one. In the diabetic case local pain necessitated reduction of the dose from 10 to 5 jig. His erectile ,response was still satisfactory at the lower dose. To date there have been no systemic side effects and no cases of priapism or corporeal fibrosis. DISCUSSION Prostaglandin El is an endogenous prostaglandin whose mode of action has yet to be fully elucidated. It acts as a powerful smooth muscle relaxant leading to dilatation of the cavernosal arteries and relaxation of the cavernosal sinusoidal smooth muscle. Its efficacy has been well documented, with erections sufficient for sexual intercourse reported in 68-86% of patients in unselected series5'9'10. The response rate is highest in those of psychogenic and neurogenic origin, 7,8 reaching almost 100% in some series 7, 8. A significantly lower rate of positive response occurs in cases of vasculogenic or diabetic impotence . Our own series was a selected group in terms of aetiology in that failure to respond to papaverine plus phentolamine tended to favour patients with a vasculogenic aetiology and this subgroup formed 60% of the total patient population. The aetiological diagnosis was presumptive on the basis of the clinical history, the presence of coexisting or previous significant medical conditions and thorough physical examination. Angiography was not available for further assessment of presumed vasculogenic cases and it is probable that there was some overlap in the categories, most notably in diabetic cases where both vasculogenic and neurogenic components are commonly present. The idiopathic group were patients in whom no overt organic risk factor for erectile dysfunction could be identified but who at the same time demonstrated no clear evidence of psychological problems. More intensive psychological and physical assessment might allow a more accurate classification of such patients. In this study the overall response rate of 80% demonstrates the superior efficacy of prostaglandin El in the treatment of vasculogenic impotence. In the ten cases where treatment was unsuccessful the presumed aetiology was vasculogenic in seven, idiopathic in two and diabetic in one. The use of papaverine has two principal complications, priapism and corporeal fibrosis. With the introduction of a combination of papaverine plus an alphablocking agent the dosage of papaverine required to induce an erectile response was significantly lowered, but despite this priapism and fibrosis remain important clinical side effects. Ten percent of our patients were commenced on prostaglandin El because of previous priapism associated with papaverine and phentolamine treatment. Prostaglandin El is a physiological agent partially metabolized locally within the cavernous tissue, and approximately 70% eliminated in a single passage C The Ulster Medical Society, 1994.
through the lungs 12. It is this rapid elimination which is believed to account for the absence of systemic side effects even in the presence of venous leakage. Many reports on its clinical use have shown a notable absence of priapism or fibrosis Some authors,, however, have documented priapism requiring treatment with metaraminol 10,' the most vulnerable group being those with a non-vasculogenic aetiology. Cases of sustained erections of up to 11 hours have also been described where spontaneous detumescence ultimately occurred 6 1O. In our own series there were no cases of priapism or sustained erection and this may be related to the selection of patients with vasculogenic disease. Corporeal fibrosis and fibrotic thickening of the tunica albuginea have been reported with papaverine injections and may be related to a number of factors including repetitive needle trauma, organization of haematomas associated with vascular puncture, or a local toxic reaction to papaverine solution which is of low pH 1316. We found no cases of corporeal fibrosis although the period of follow-up at present is relatively short.
In previous studies the most commonly described side effect of prostaglandin El is the occurrence of localized pain in the penis, varying in intensity from mild discomfort to a severe burning sensation prohibiting sexual intercourse. The reported incidence varies from 11% up to 75% 6-10, but was a complaint in only four patients (8%) in our own series. In only one was a dose reduction required to ameliorate this problem. Localized pain appears to occur more commonly in those individuals whose impotence is non-vasculogenic in origin 10, which may explain the low incidence of this series. Intracavernous prostaglandin E, has proved an effective and safe alternative agent for the treatment of erectile failure with particular application to those patients where treatment with papaverine and phentolamine had been ineffective or inappropriate. |
The Vitamin D Status of Prison Inmates
Introduction: There is no comprehensive, systematic analysis of the vitamin D status of prisoners in the scientific literature.Objective: To investigate the vitamin D status and its determinants in US prison inmates.Hypothesis: Given the uniformity of dietary intake amongst inmates, vitamin D status will be determined by non-dietary factors such as skin pigmentation, security level-, and the duration of incarceration.Subjects and Methods: A retrospective study of 526 inmates (males, n = 502, age 48.6612.5 years; females, n = 24, age 44.1612.2) in Massachusetts prisons. Vitamin D sufficiency, insufficiency, and deficiency were respectively defined as a 25(OH)D concentration 75 nmol/L; 50 to 75 nmol/L; and ,50 nmol/L. The Massachusetts Department of Correction Statement of Nutritional Adequacy stated that each inmate received the recommended daily allowance of vitamin D daily. Security level of incarceration was designated as minimum, medium, and maximum. Racial groups were categorized as Black, white, Asian, and Others.Results: Serum 25(OH)D levels peaked in summer and autumn, and decreased in winter and spring. Vitamin D deficiency occurred in 50.5% of blacks, 29.3% of whites, and 14.3% of Asian inmates (p = 0.007). Black inmates had significantly lower serum 25(OH)D level than white inmates at the maximum security level (p = 0.015), medium security level (p = 0.001), but not at the minimum security level (p = 0.40). After adjusting for covariates black inmates at a maximum security level had a four-fold higher risk for vitamin D deficiency than white inmates at the same security level (OR 3.9 [95% CI 1.3-11.7].Conclusions:The vitamin D status of prison inmates is determined by skin pigmentation, seasons, and the security level of incarceration.
# Introduction
The United States has the highest prisoner population per capita of any country in the world [bib_ref] The unmet medical needs of correctional populations in the United States, Cropsey [/bib_ref]. More than 8 million offenders are under the control of US correctional agencies, and these prisoners have poor nutritional status and poor health [bib_ref] Release from prison-a high risk of death for former inmates, Binswanger [/bib_ref] compared to the general population [bib_ref] Fatty acid analysis of blood from violent offenders, Corrigan [/bib_ref]. Recent studies of the United Kingdom and Unites States prison inmates showing poor intake of vitamin D [bib_ref] Food provision and the nutritional implications of food choices made by young..., Eves [/bib_ref] [bib_ref] What are we feeding our Inmates?, Collins [/bib_ref] and other nutrients have prompted a call for epidemiological studies devoted to criminology [bib_ref] The unmet medical needs of correctional populations in the United States, Cropsey [/bib_ref]. This call is timely as there is no systematic analysis of the vitamin D status of prisoners in global scientific literature despite recent studies linking vitamin D deficiency to poor skeletal health, lower muscle strength, increased body sway, falls, disability, low bone mineral density, osteoporosis, and fractures [bib_ref] The 2011 report on dietary reference intakes for calcium and vitamin D..., Ross [/bib_ref] [bib_ref] Vitamin D status, trunk muscle strength, body sway, falls, and fractures among..., Pfeifer [/bib_ref] [bib_ref] Neuromuscular and psychomotor function in elderly subjects who fall and the relationship..., Dhesi [/bib_ref] [bib_ref] Low vitamin D and high parathyroid hormone levels as determinants of loss..., Visser [/bib_ref] [bib_ref] Vitamin D Status Is Associated With Functional Limitations and Functional Decline in..., Sohl [/bib_ref].
The term vitamin D refers to a group of seco-steroid compounds, namely vitamin D 2 and D 3 , which are prohormones with skeletal and extra-skeletal functions [bib_ref] Vitamin D deficiency, Holick [/bib_ref]. Both vitamin D 2 and D 3 are hydroxylated by cytochrome P450 enzymes in the liver to form 25-hydroxyvitamin D [25(OH)D] which is the primary biomarker of vitamin D status. A further hydroxylation step in the kidneys yields the biologically active form of vitamin D called 1,25dihydroxyvitamin D or calcitriol, which serves to maintain plasma calcium and phosphate concentrations, thereby promoting bone health [bib_ref] Vitamin D, disease and therapeutic opportunities, Plum [/bib_ref].
A comprehensive analysis of the vitamin D status of prison inmates is the first step in the understanding of the impact of hypovitaminosis D on their bone health. Such an evaluation would form the basis for interventions to address the deleterious consequences of vitamin D deficiency-related bone diseases on inmates, their families, and the national healthcare budget. Our aim was to determine the baseline vitamin D profiles of prison inmates, and to investigate the non-dietary determinants of vitamin D status in this population. We hypothesized that (1) a high proportion of our cohort would be vitamin D deficient; and [bib_ref] Release from prison-a high risk of death for former inmates, Binswanger [/bib_ref] given the uniformity of dietary intake, vitamin D status in prison inmates would depend principally on non-dietary factors such as the degree of sun exposure, skin pigmentation, security level, and the duration of incarceration.
# Subjects and methods
# Ethics statement
The study protocol was jointly approved by the Massachusetts Department of Public Health Institutional Review Board, the Massachusetts Department of Correction, and the University of Massachusetts Institutional Review Board. All patient records and information were anonymized and de-identified prior to analysis.
## Subjects
All data were sourced from the Massachusetts Department of Correction's Inmate Management System (IMS) database. The medical records of 839 prison inmates in the 18 facilities in the Massachusetts Department of Correction system from June 1, 2010 through June 30, 2012 were reviewed [fig_ref] Figure 1: Study Flow Diagram [/fig_ref]. Of these, 744 inmates had complete data. Study subjects were included if they had serum 25(OH)D level obtained. The serum 25(OH)D estimations were conducted as part of routine biochemical laboratory testing in the inmates, and not due to any specific investigation for pathologies associated with vitamin D metabolism. Subjects were excluded if they were receiving pharmacological vitamin D or calcium supplementation prior to the date of 25(OH)D measurement. Seventy-eight inmates were excluded based on the above criteria. Subjects were further excluded if they were pregnant, lactating, receiving medications that affect vitamin D metabolism, or had a concurrent diagnosis of any disease that affects calcium or vitamin D metabolism such as diseases of bone, parathyroid gland, kidney, and liver. Specifically, subjects with serum creatinine of .1.5 mg/dL, or alanine aminotransferase of .2.5 times the upper limit of the normal range were excluded [bib_ref] Reduced growth hormone secretion in obesity is associated with smaller LDL and..., Makimura [/bib_ref]. An additional 140 inmates were excluded based on these latter exclusion criteria. Five hundred and twenty-six (526) [bib_ref] Dietary reference intakes for vitamin D: justification for a review of the..., Yetley [/bib_ref].
Anthropometry. Anthropometry was measured by correctional officers. Height and weight were measured in light clothing, with no shoes, using an upright scale. Body mass index (BMI) was derived using the formula weight/height 2 (kg/m 2 ). Age of study subjects was determined by the date of the 25(OH)D measure-ment. Subjects were categorized into normal weight, overweight, and obese groups, based on BMI criteria as BMI ,25 kg/m 2 , 25-29.9, and $30 kg/m 2 respectively.
Assay. Serum levels of 25(OH)D were analyzed using 25hydroxy chemiluminescent immunoassay (DiaSorin Liaison; Stillwater, Minnesota), which has a 100% cross-reactivity with both metabolites of 25(OH)D namely, 25(OH)D 2 and 25(OH)D 3 and thus measures total serum 25(OH)D content. Its functional sensitivity is 4 ng/mL (10 nmol/L), and its intra-and inter-assay coefficients of variation are 5% and 8.2% respectively.
Vitamin D Status. Vitamin D sufficiency was defined as a 25(OH)D concentration $75 nmol/L; vitamin D insufficiency as 25(OH)D of 50 to 75 nmol/L; and vitamin D deficiency as 25(OH)D level ,50 nmol/L according to the Endocrine Society criteria [bib_ref] Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical..., Holick [/bib_ref].
## Statistical analyses
Statistical analyses were performed using the SPSS Predictive Analytics SoftWare v.21 (IBM Corporation, Armonk, NY) and Microsoft Excel (2007). Means and standard deviations were calculated for descriptive summary statistics and 25(OH)D measurements. Anthropometrics and 25(OH)D levels were compared using Student's t test. Race, security level, and seasons of blood draw were compared using Fisher's exact test. A two-way ANOVA was first used to explore the differences in 25(OH)D between the four racial groups (Black, white, Asian, and Others) [F (3,522) = 7.836, p,0.001]; as well as the 3 security levels of incarceration [F (2,523) = 3.568, p = 0.029], followed by post-hoc comparisons. Data were expressed as mean 6 standard deviation The Vitamin D Status of Prison Inmates PLOS ONE | www.plosone.org (SD). Linear regression statistics was used to determine the relationship between the duration of incarceration and serum 25(OH)D level. The duration of incarceration was log transformed to achieve normality.
For the analyses, subjects were first stratified by gender to determine the effect of gender on serum 25(OH)D concentrations. They were then stratified by racial groups according to the official designation of inmates by the Massachusetts Department of Correction into white, black, Asians, and Others (consisting of Native Americans, Pacific Islanders, and Alaskan natives), to determine the relationship between race and 25(OH)D levels. We further investigated the relationship between 25(OH)D level and the duration of incarceration. Finally, the inmates were stratified by their level of incarceration into minimum-, medium-, and maximum security levels to investigate the effects of levels of incarceration and sun exposure on serum 25(OH)D concentrations. With respect to sun exposure, inmates at the minimum security level were allowed an average daily sun exposure of about 5-10 hours/day during which time they may engage in exercise, working outside the facilities as road crews, park maintenance crew, or in farms. Inmates at the medium security level spent an average of 1-5 hours/day in recreational activities under supervision. Inmates at the maximum security level spent only one hour in recreational activities per day under a heavy guard. [fig_ref] Table 1: Characteristics of Prison Inmates Stratified by Gender [/fig_ref] shows a comparative analysis of the inmates' characteristics stratified by gender. The male inmates were taller, and had significantly longer duration of incarceration than the female inmates. A significantly higher proportion of female inmates were incarcerated at the minimum security level compared to male inmates. In contrast, a higher proportion of male inmates were incarcerated at the medium security level. Only male inmates were incarcerated at the maximum security level.
# Results
## Patient characteristics stratified by gender
Proportion of subjects with vitamin D deficiency, insufficiency, and sufficiency [fig_ref] Table 2: Proportion of Inmates with Vitamin D Deficiency, Insufficiency, and Sufficiency within each... [/fig_ref] shows that the overall proportion of inmates in our cohort with vitamin D deficiency is 33%, vitamin D insufficiency 34%, and vitamin D sufficiency 31%. Vitamin D deficiency occurred in 50.5% of black inmates, 29.3% of white inmates, 14.3% of Asian inmates, and 35% of the Others (p,0.001). There was no significant difference in the proportion of inmates with vitamin D insufficiency among the four groups (p = 0.90). Vitamin D sufficiency occurred in 17.8% of black inmates, compared to 35.5% of white inmates, 42.9% of Asian inmates, and 25% of Others (p = 0.03) [fig_ref] Table 2: Proportion of Inmates with Vitamin D Deficiency, Insufficiency, and Sufficiency within each... [/fig_ref].
## The determinants of vitamin d status in prison inmates
Age. There was no relationship between age and 25(OH)D level in all inmates (b = 20.033, adjusted r 2 = 0.001, p = 0.45).
Race. There was a significant difference in 25(OH)D level across the racial groups (p,0.001). Black inmates had significantly lower 25(OH)D level compared to the white inmates Race and Gender. When the inmates were stratified by race and gender, the black male inmates had significantly lower 25(OH)D levels than the white male inmates (p,0.001). There was, however, no significant difference in 25(OH)D levels between black and Asian male inmates (p = 0.10), nor between the black male inmates and Others (p = 0.55). Both the white and Asian male inmates had similar vitamin D levels (p = 0.73). There was equally no difference in 25(OH)D level between the black and white female inmates (p = 0.58).
Seasons Duration of Incarceration. After adjusting for age, sex, season, and BMI, there was a weakly significant relationship between the duration of incarceration and serum 25(OH)D level in all inmates (b = 0.91, adjusted r 2 = 0.07, p = 0.042) [fig_ref] Figure 4: The relationship between serum concentrations of 25-hydroxyvitamin D [25 [/fig_ref]. However, a further investigation of the above relationship in
# Discussion
This study has 3 important findings: first, only 31% of prison inmates in our cohort were vitamin D sufficient, while the rest were either vitamin D deficient (33%), or vitamin D insufficient (34%). Second, a higher proportion of black inmates, regardless of their incarceration level, had lower 25(OH)D level compared to the non-black inmates. Third, after adjusting for age, sex, BMI, seasons, skin tone, eye and hair color, black inmates incarcerated at a maximum security level were 4 times more likely to be vitamin D deficient than white inmates at the same security level. Age, gender, and the duration of incarceration were not significant determinants of inmates' vitamin D status. The Statement of Nutritional Adequacy of the Massachusetts Department of Correction states that the inmates received the RDA for vitamin D in their diet. However, given that 67% of the inmates in our cohort had hypovitaminosis D, it is most likely that the amount of vitamin D in their diet was not sufficient to maintain a normal vitamin D status especially in the colder months, or that the process of food handling in prisons might have resulted in reduced amounts of vitamin D in their diet. This is supported by studies that either documented poor intake of vitamin D by inmates as a result of subnormal vitamin D contents of foods served in correctional facilities, or suggested that the method of food handling in correctional facilities could lead to vitamin D deficiency [bib_ref] Food provision and the nutritional implications of food choices made by young..., Eves [/bib_ref] [bib_ref] What are we feeding our Inmates?, Collins [/bib_ref]. Eves et al [bib_ref] Food provision and the nutritional implications of food choices made by young..., Eves [/bib_ref] reported that even though vitamin D intakes in correctional facilities could be in excess of recommended daily allowance, extensive warm-holding of foods prior to service could result in significant losses of some vitamins. Thus, the actual amounts of vitamin D consumed may be markedly less than the initial estimates by dietary analyses [bib_ref] Food provision and the nutritional implications of food choices made by young..., Eves [/bib_ref]. This scenario may be applicable to our study. Therefore, even though our study demonstrates that the independent determinants of vitamin D status in prison population are skin pigmentation, security level of incarceration, and seasons; dietary factors such as excessive warm holding of fold before service may also lead to subnormal amounts of vitamin D in diet, and consequent hypovitaminosis D.
This is the first in-depth examination of the unique determinants of vitamin D status of prison inmates in literature. First, this study confirms the effects of both isolation and insolation on vitamin D status. Insolation is defined as the incoming solar radiation that reaches the earth, its atmosphere, and surface objects [bib_ref] The Effect of Geography and Vitamin D on African American Stature in..., Carson [/bib_ref]. Insolation is the primary source of vitamin D [bib_ref] Vitamin D deficiency, Holick [/bib_ref]. This study demonstrates that individuals confined or restricted to similar environments such as prisons, where they are exposed to a similar pattern of nutritional intake, could have variable vitamin D status based on their level of insolation. Second, this study demonstrates the effect of skin pigmentation on vitamin D status in prison inmates. It shows that under conditions where individuals are exposed to a similar pattern of nutritional intake, darkly pigmented subjects would have lower serum concentrations of 25(OH)D compared to non-pigmented individuals. This is because melanin, which is found throughout the epidermis, is a natural sunscreen that absorbs ultraviolet B radiation thereby reducing the amount of vitamin D produced during sun exposure [bib_ref] Increased skin pigment reduces the capacity of skin to synthesise vitamin D3, Clemens [/bib_ref] [bib_ref] Melanin pigmentation in mammalian skin and its hormonal regulation, Slominski [/bib_ref]. Third, our study shows that age, gender, and the duration of incarceration have no biologically significant effect on inmates' vitamin D status.
These findings have implications at three concentric levels: the inmates [bib_ref] The unmet medical needs of correctional populations in the United States, Cropsey [/bib_ref] [bib_ref] Low vitamin D and high parathyroid hormone levels as determinants of loss..., Visser [/bib_ref] [bib_ref] Vitamin D Status Is Associated With Functional Limitations and Functional Decline in..., Sohl [/bib_ref] , the correctional system, and the national healthcare policy [bib_ref] The unmet medical needs of correctional populations in the United States, Cropsey [/bib_ref] [bib_ref] Food provision and the nutritional implications of food choices made by young..., Eves [/bib_ref] [bib_ref] What are we feeding our Inmates?, Collins [/bib_ref]. At the level of the inmates, prolonged vitamin D deficiency could lead to poor health, as strong associations between vitamin D deficiency and increased risk for several diseases such as type 1 and type 2 diabetes, cardiovascular diseases, rheumatoid arthritis, infectious diseases, depression, and cancers of the breast, prostate, colon, and pancreas, have been reported [bib_ref] Vitamin D deficiency, Holick [/bib_ref] [bib_ref] Relationship between vitamin D levels and depressive symptoms in older residents from..., Stewart [/bib_ref] [bib_ref] Nonclassic actions of vitamin D, Bikle [/bib_ref]. Though there is an ongoing debate on the significance of these extra-skeletal functions of vitamin D in humans [bib_ref] The 2011 report on dietary reference intakes for calcium and vitamin D..., Ross [/bib_ref] [bib_ref] Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical..., Holick [/bib_ref] , there is a universal consensus on its skeletal functions, as vitamin D has been demonstrated to be vital for bone mineralization, maintenance of bone strength, and the prevention of fractures and consequent immobilization [bib_ref] The 2011 report on dietary reference intakes for calcium and vitamin D..., Ross [/bib_ref] [bib_ref] Effectiveness and safety of vitamin D in relation to bone health, Cranney [/bib_ref].
At the level of the correctional facilities, this study shows that a large proportion of inmates have both vitamin D insufficiency and deficiency despite dietary analyses showing intakes above the RDA [bib_ref] Food provision and the nutritional implications of food choices made by young..., Eves [/bib_ref] [bib_ref] What are we feeding our Inmates?, Collins [/bib_ref]. Therefore, to ensure optimal bone and muscle health, and prevent functional limitations in inmates, adequate vitamin D supplementation protocol [bib_ref] The 2011 report on dietary reference intakes for calcium and vitamin D..., Ross [/bib_ref] [bib_ref] Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical..., Holick [/bib_ref] , coupled with regular monitoring of serum 25(OH)D levels occurring once or twice yearly will be necessary. The timing of the biannual vitamin D monitoring should be in the late fall and late winter seasons. In addition, inmates at the medium and maximum security levels, especially those with dark skin pigmentation, should be monitored closely, as they are at the greatest risk for vitamin D deficiency. Furthermore, the inmates in the high risk group may require a higher dose of vitamin D supplementation, of approximately 2000-4000 IU per day, to maintain 25(OH)D above 30 ng/mL which is the threshold recommended by the Endocrine Society [bib_ref] Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical..., Holick [/bib_ref]. Finally, correctional facilities or public health officials should institute periodic verification of the actual estimates of micronutrients and vitamins contained in inmates' meals by a process of direct measurement of the vitamin D content of homogenized samples of inmates' meals to ensure that the RDA is consistently met. At the national healthcare budgetary level, the implementation of a structured supplementation protocol is critical to the preservation of inmates' bone health [bib_ref] Effectiveness and safety of vitamin D in relation to bone health, Cranney [/bib_ref] , muscle strength [bib_ref] Low vitamin D and high parathyroid hormone levels as determinants of loss..., Visser [/bib_ref] , and the prevention of comorbidities derived from vitamin D deficiency-associated skeletal and muscular defects, such as falls, fractures, and immobilization [bib_ref] Vitamin D Status Is Associated With Functional Limitations and Functional Decline in..., Sohl [/bib_ref]. Such a supplementation protocol will help reduce the unnecessary healthcare expenditures on these preventable comorbidities.
This study has some limitations. First, the cross-sectional study design limits causal inference on the effects of seasons, race, and security level of incarceration on vitamin D status. Second, we did not study the entire Massachusetts prison population, but only those with laboratory values for 25(OH)D who met the inclusion criteria. This limited our ability to determine the prevalence of vitamin D deficiency or sufficiency in this population. We also did not have adequate power to detect a significant difference in 25(OH)D level between black and Asian male inmates because of the very few Asian male inmates in our cohort. Third, we neither administered a food-recall questionnaire, nor performed a biochemical quantification of the 25(OH)D content of the inmates' meals to accurately determine the vitamin D content of their diet. Our attempts to obtain samples of inmates' meals for quantification of the vitamin D content were unsuccessful. However, we reviewed both the Statement of Nutritional Adequacy and the daily logs of inmates' menus for the entire Massachusetts Correctional System for uniformity of RDAs for vitamin D. Fourth, we did not have data on parathyroid hormone levels which could be elevated in states of vitamin D deficiency and hypocalcemia. Fifth, our results were derived from a prison system in a single state in the United States, therefore, we are uncertain that our results are generalizable to other correctional systems, states, or countries. Finally, the study prison sites were located around latitude 42u N, thus, it is not certain whether similar results would be obtained in correctional facilities in other geographical locations.
The unique strength of this study is that it was conducted exclusively amongst inmates living in a uniform prison system with clear guidelines regarding nutrition, security level of incarceration and their associated hours of insolation, as well as the number of hours permitted for recreational activities. We had full access to the IMS database which enabled us to accurately confirm our inclusion and exclusion criteria. We had a fairly large study sample size to enable us detect differences between the groups of interest. Our sample contained a fair representation of the major racial groups, thus enabling us to analyze the effects of differential insolation on racial groups. We had a fair distribution of inmates in each of the three major security levels of incarceration, thus, enabling us to investigate the effect of security level on vitamin D status. The lack of a significant difference in 25(OH)D levels between the female inmates of different racial groups could be due to the fact that they were all housed in the single female correctional facility in Massachusetts, and none of the female inmates was being held at the maximum security level. The lack of a significant difference in 25(OH)D level between the black male inmates and the male inmates in the 'Others' category could be due to the fact that both groups possess variable degrees of dark skin pigmentation. Representative data on 25(OH)D levels were available for all the seasons of the year, thus allowing us to determine the variability of 25(OH)D with the seasons.
In conclusion, this study shows that in a prison system with uniform nutritional guidelines, three principal factors determine inmates' vitamin D status: skin pigmentation, seasons, and the security level of incarceration. The roles of diet and the method of food handling deserve further investigation as this study demonstrates a high occurrence of hypovitaminosis D amongst prisoners receiving the RDA of vitamin D in their diet. A comprehensive skeletal health policy that emphasizes the maintenance of normal serum 25(OH)D concentration is needed to ensure optimal skeletal and muscular health in inmates. Such a policy will ease the fiscal burden of these preventable, but costly, comorbidities on the national healthcare budget.
[fig] Figure 1: Study Flow Diagram. doi:10.1371/journal.pone.0090623.g001 [/fig]
[fig] .: Even though serum 25(OH)D levels peaked in Summer and Fall, and reached a nadir in Winter and Spring in all racial groups, black inmates had significantly lower mean 25(OH)D levels for all seasons of the year except for Fall, compared to white inmates (Figure 3) The seasonal pattern of change in the serum concentrations of 25(OH)D with seasons in blacks vs whites were as follows: Summer (5376231 nmol/L vs 71564, p = 0005), Fall (65565 nmol/L vs 7086265, p = 035), Winter (4506180 nmol/L vs 5996272, p = 0012), Spring (4036161 nmol/L vs 5886264, p = 0001) [/fig]
[fig] Figure 2: Box plots of the comparison of 25-hydroxyvitamin D values of inmates stratified by race. This figure shows that black inmates have a significantly lower 25(OH)D level compared to the white inmates (p,0.001), but similar levels to inmates categorized as Others (p = 0.54). There was inadequate power to detect a difference between black (n = 107) and Asian (n = 7) inmates (p = 0.1). Note: 50 nmol/L = 20 ng/mL. doi:10.1371/journal.pone.0090623.g002 [/fig]
[fig] Figure 3: Graph of the relationship between 25(OH)D and the seasons in black and white prisoners. Black inmates had significantly lower 25(OH)D levels in all of the seasons except for Fall. doi:10.1371/journal.pone.0090623.g003 [/fig]
[fig] Figure 4: The relationship between serum concentrations of 25-hydroxyvitamin D [25(OH)D] and years of incarceration. This figure shows a weakly significant relationship between the years of incarceration and serum concentration of 25(OH)D in all inmates after adjusting for age, race, season, and BMI (b = 0.09, r 2 = 0.07, p = 0.042). A similar investigation in inmates in the maximum security level showed no significant relationship between 25(OH)D and the duration of incarceration (b = 0.09, r 2 = 0.09, p = 0.49). doi:10.1371/journal.pone.0090623.g004 [/fig]
[fig] Figure 5: Complex box plots of the comparison of 25-hydroxyvitamin D values of inmates stratified by both race and the security level of incarceration. This figure shows a significant difference in 25(OH)D level between the black and white inmates at both the maximum security level (p = 0.015), and the medium security level (p = 0.001), but not at the minimum security level (p = 0.40). Note: 50 nmol/L = 20 ng/mL. doi:10.1371/journal.pone.0090623.g005 [/fig]
[table] Table 1: Characteristics of Prison Inmates Stratified by Gender. [/table]
[table] Table 2: Proportion of Inmates with Vitamin D Deficiency, Insufficiency, and Sufficiency within each Racial Category. [/table]
|
Self reported health status, and health service contact, of illicit drug users aged 50 and over: a qualitative interview study in Merseyside, United Kingdom
Background:The populations of industrialised countries are ageing; as this occurs, those who continue to use alcohol and illicit drugs age also. While alcohol use among older people is well documented, use of illicit drugs continues to be perceived as behaviour of young people and is a neglected area of research. This is the first published qualitative research on the experiences of older drug users in the United Kingdom.Methods: Semi-structured interviews were conducted in Merseyside, in 2008, with drug users aged 50 and over recruited through drug treatment services. Interviews were recorded and transcribed and analysed thematically. Only health status and health service contact are reported here.Results: Nine men and one woman were interviewed (age range: 54 to 61 years); all but one had been using drugs continuously or intermittently for at least 30 years. Interviewees exhibited high levels of physical and mental morbidity; hepatitis C was particularly prevalent. Injecting-related damage to arm veins resulted in interviewees switching to riskier injecting practices. Poor mental health was evident and interviewees described their lives as depressing. The death of drug-using friends was a common theme and social isolation was apparent. Interviewees also described a deterioration of memory. Generic healthcare was not always perceived as optimal, while issues relating to drug specific services were similar to those arising among younger cohorts of drug users, for example, complaints about inadequate doses of prescribed medication.Conclusion:The concurrent effects of drug use and ageing are not well understood but are thought to exacerbate, or accelerate the onset of, medical conditions which are more prevalent in older age. Here, interviewees had poor physical and mental health but low expectations of health services. Older drug users who are not in contact with services are likely to have greater unmet needs. The number of drug users aged 50 and over is increasing in Europe and America; this group represent a vulnerable, and in Europe, a largely hidden population. Further work to evaluate the impact of this change in demography is urgently needed.
# Background
The world's population is ageing. In 1900, the global population was estimated to have only 1% of people aged 65 years and over, by 2000 this figure was 7%, and by 2050, the estimated proportion will be 20% [1]. As our populations age, people who continue to use drugs and alcohol age also. Projections from the United States of America (USA), for example, suggest that the number of people aged 50 and over who need treatment for drug and alcohol problems will increase from 1.7 million in 2000 to 4.4 million in 2020, while estimates from Europe suggest that the number of people aged 65 and over needing treatment for such problems will more than double between 2001 and 2020. In the United Kingdom (UK), research using established monitoring systems has shown that the proportion of drug users aged 50 and over in contact with specialist drug services in Cheshire and Merseyside (the only area of the country to have collected data on drug treatment clients consistently since 1998 and therefore the only area able to monitor demographic changes) has increased between 1998 and 2004/05 from 1.5% to 3.6% (P < 0.001) and 1.9% to 3.2% (P < 0.001) for men and women respectively. Furthermore, the proportion aged 40 to 49, potential older drug users of the future, increased from 8.1% in 1998 to.6% in 2004/05 [bib_ref] Problematic drug use, ageing and older people: trends in the age of..., Beynon [/bib_ref].
A prospective cohort study undertaken in the USA demonstrated that older drug users have levels of morbidity considerably higher than the general population [bib_ref] Health conditions among aging narcotic addicts: medical examination results, Hser [/bib_ref] , while qualitative research also completed in the USA demonstrated that they experience loneliness, stress and fear of victimisation [bib_ref] The drug career of the older injector, Levy [/bib_ref]. The use of illicit drugs is, however, still largely perceived to be behaviour of the young and, consequently, continues to be a neglected area of research. In particular, there has to date, and to the best of our knowledge, been no qualitative research published on the experiences of older drug users in the UK. Information relating to what drugs older people in the UK use, the reasons for initiating or continuing drug use and the specific health and treatment needs of this group of people is, at present, lacking in the UK; our understanding can not be conceptualised using data from the USA due to cultural, legal and attitudinal differences between the two countries, and differences in the manner in which healthcare is delivered [bib_ref] Effects of program and patient characteristics on retention of drug treatment patients, Hser [/bib_ref].
The aim of this present study was to undertake an investigation into issues salient to older drug users. More specifically, the study aimed to identify substance (drug and alcohol) use, self-reported health status and contact with generic and specialist health services of a cohort of older drug users in contact with specialist drug treatment services. Here we define older drug users as those aged 50 and over.
# Methods
Qualitative research was undertaken using semi-structured interviews with prompts to capture older people's experiences across the life course. These interviews covered a range of topics but only results pertaining to health status and health service contact, plus brief contextual data, are presented here. While the interviewer had a number of pre-determined topics to cover, the interviewer was sensitive to the interviewee and their narrative, allowing for the exploration of topics that arose spontaneously, in order to enhance the qualitative information obtained. Interviews were conducted by the same researcher (LP) to ensure consistency and all were taped and transcribed. LP also provided additional information to enhance the quality of the transcriptions. Approval for this study was granted by the Liverpool John Moores University Ethical Committee. Each participant received a £10 shopping voucher in compensation for their time dedicated to the study.
## Recruitment
Staff working in voluntary sector, specialist drug treatment services in Merseyside were asked to identify clients within their care aged 50 and over and to introduce the study to them. People who expressed a willingness to participate were given an information sheet to take with them and the member of staff contacted the research team with the client's name and contact telephone number. A member of the research team (LP) made contact with the potential participant and organised a convenient time and place for the interview to occur. Prior to the interview, each potential participant was given another information sheet, and a consent form. Full informed consent was obtained, including specific consent to use anonymised quotes, and it was explained that participation would not influence their ongoing medical care. Interviews were conducted between 16 th January and 22 nd February 2008. Recruitment continued until no new major themes were identified.
# Analysis
Transcripts were scrutinised by all four authors, who analysed them using thematic analysis. Each member of the research team initially independently identified broad themes which were given codes; these were then discussed and agreed by the research team. Each transcript was then read again by all authors, were coded, and passages relating to each theme were grouped together. Themes included drug use, health, quality of life, lifestyle and employment, relationships and social networks and use of services.
## Reliability and validity
The first three interviews were used to establish the feasibility of recruitment, inclusion criteria and to test the methods of data collection. No changes to the interview schedule or prompts were made but we reduced the minimum age from 60 to 50 years because it proved difficult to recruit a sufficient number of people aged over 60. Thematic analysis showed no difference in the themes and sub-themes between these pilot interviews and the data from all other interviews and so the pilot interviews have been included in the results. Where interviewees provided contradictory or revisionist statements, the interviewer (LP) sought clarification in order to enhance the validity, accuracy and consistency of responses and our understanding of their discourse. Reliability of the analysis was assured by members of the team reading and re-reading transcripts, and by discussion and agreement of the themes and sub-themes. The measures described here are standard verification techniques for establishing quality, validity and rigour in qualitative research [bib_ref] Verification strategies for establishing reliability and validity in qualitative research, Morse [/bib_ref].
# Results and discussion
Context, demography and substance use Nine men and one woman were interviewed in a range of settings; within a doctor's surgery (n = 2), at home (n = 2), in a care home (n = 1) and at a drug treatment service (n = 5). The ages of interviewees ranged from 54 to 61 years. Six were single and four were divorced at the time of the interview, with two living with a male friend who acted as a carer, one more who lived next door to a friend/male carer and another who shared their accommodation with other drug users. Their accommodation varied; some lived in a hostel, others in their own home (council house, flat or housing association bed sit), one man lived in a care home and another lived in a caravan.
All but one interviewee had started substance use when in their adolescent, or early adult, years but there was no single pattern to drug initiation; drugs first used included alcohol, cannabis, lysergic acid (LSD), amphetamine, morphine hydrochloride, heroin, 'speedballs' (heroin in combination with a strong stimulant, usually cocaine) and (psilocybin-containing) mushrooms and people simply switched drugs according to availability. For these people, drug use across the life course was divided into three camps, with some interviewees ceasing drug use for many years at a time, some using drugs intermittently with short periods of abstinence, while others followed a course of near continual consumption. Conversely, one interviewee was a late onset user, having commenced drinking aged 30 (which progressed into problematic drinking) and drug use (heroin) aged 46. Every interviewee was, or had been, a problematic drug user (currently defined in the UK as a user of opiates and/or crack cocaine. Every interviewee had used heroin during their lives and some were still using this drug at the time of interview; many were in receipt of a methadone prescription as part of opiate substitution therapy. Six people also reported a history of cocaine and/or crack cocaine use. Some interviewees acknowledged that they had injected drugs or inferred that this was the case by discussing injection-related health consequences. A more detailed discussion of substance use and other themes arising from this study will be presented elsewhere.
## Physical health
The effects of the interaction between drug use and those processes that characterise ageing are poorly understood. However, there is evidence to show that drug use exacerbates, or accelerates the onset of, medical conditions which are more prevalent in older age. Stimulant use, for example, can lead to cardiovascular complications in younger drug users, while longer term use may increase the likelihood of premature atherosclerosis, ventricular hypertrophy and cardiomyopathy. Drug use has also been associated with an earlier onset on diabetes, while neurological disorders, respiratory disorders, cancer and other age associated diseases may also all be worsened by drug use [bib_ref] Drugs of abuse and the aging brain, Dowling [/bib_ref]. Furthermore, the disorientating or hallucinating effects of some psychoactive drugs increase the risk of accidents and falls, particularly in combination with prescribed or over-the-counter medications [bib_ref] Drugs of abuse and the aging brain, Dowling [/bib_ref] [bib_ref] Polypharmacy and falls in the middle age and elderly population, Ziere [/bib_ref]. Falls are a major cause of disability and mortality in older peopleand can be presumed to occur even more frequently among those who use psychoactive drugs including alcohol. Our interviewees reported a high level of physical morbidity; circulatory problems, respiratory problems, pneumonia, diabetes, hepatitis and liver cirrhosis. Overdose and injuries also featured in their narratives. Summarising, one man stated; "I'm... handicapped to the extent of can't go out, suffer with arthritis in the right leg and lower back, blind in one eye, suffer hepatitis C....". Hepatitis C infection, contracted by injecting drug users through sharing contaminated drug injecting and/or preparation equipment [bib_ref] Risk of hepatitis C virus transmission through drug preparation equipment: a systematic..., De [/bib_ref] , was a common theme, with half of those interviewed stating they were infected with the virus. Despite being a serious infection, progressing to cirrhosis, end stage liver disease and, in about one to five percent of those chronically infected, hepatocellular carcinoma, a synthesis of qualitative research demonstrated that drug users of all ages tend to see hepatitis C as somewhat of an inevitability of injecting, that being infected is normal and that it is physically benign, particularly relative to HIV [bib_ref] The social production of hepatitis C risk among injecting drug users: a..., Rhodes [/bib_ref]. This perception of its benignity is presumably even greater among those older drug users who contracted this infection many years previously and have experienced no adverse consequences of long term infection. Certainly our interviewees did not perceive their infection to be problematic;
## "i've never really had any serious illnesses.... i forgot the hep c y'know. i've had no symptoms. i haven't been sick or anything y'know, but i've got the virus".
There are many factors which influence the aetiology of hepatitis C infection and its impact on liver function and drug users of all ages are infected with the virus in the UK. However, those older drug users who contracted the virus many years previously may be at an advanced stage of disease progression. Furthermore, younger age at the commencement of treatment is one positive predictor of success in clearing the virusand healthcare services will need to be prepared to care for older drug users who may be disproportionately affected by treatment failure. Treatment services need also be aware of the wider determinants impacting on the likelihood of older drug users adhering to treatment and in particular the role of social support in coping with the relatively long and unpleasant nature of hepatitis C treatment [bib_ref] Resilient coping: applying adaptive responses to prior adversity during treatment for hepatitis..., Hopwood [/bib_ref]. Qualitative research from the USA reported that older drug users had severed links with non-drug using family and friends over the years of drug taking [bib_ref] The drug career of the older injector, Levy [/bib_ref] , a premise supported by the current research. For older drug users with little or no family support, the involvement of services to support a person while undergoing treatment will play an important role [bib_ref] Resilient coping: applying adaptive responses to prior adversity during treatment for hepatitis..., Hopwood [/bib_ref].
The high proportion of interviewees with hepatitis C reflects findings from the USA, where medical examinations of 108 older heroin-dependent men (mean age 58.4 years) showed that 95 percent of their subjects had been exposed to the hepatitis C virus, while more than half were found to have abnormal liver function; these rates were much higher than found in the general population [bib_ref] Health conditions among aging narcotic addicts: medical examination results, Hser [/bib_ref]. In recent years, the enormity of the UK's hepatitis C epidemic has been evidenced through a number of sources, with the Health Protection Agency consistently citing it as the most significant infectious disease affecting people who inject drugs. There are currently an estimated 330,000 problematic drug users (users of opiates and/or crack cocaine) in England and 120,000 current injectors of these substances [bib_ref] National and regional estimates of the prevalence of opiate use and/or crack..., Hay [/bib_ref] plus many others who are ex-drug injectors. According to Health Protection Agency estimates, approximately half of all current drug injectors are infected with hepatitis C, and both incidence and prevalence continue to rise. As the already ageing cohort of drug users [bib_ref] Problematic drug use, ageing and older people: trends in the age of..., Beynon [/bib_ref] age further, those infected with the virus will gradually develop symptoms of the infection and the full impact of hepatitis C infection among current and ex-drug injectors will become apparent. Currently, the burden of drug related chronic conditions such as hepatitis C, which disproportionately affect older drug users compared to their younger counterparts, is not being fully recognised because such deaths are not classified under UK or European definitions as drug related deaths and are therefore not counted in official figures [bib_ref] The role of substance use in non-drug related deaths: a cross-sectional study..., Beynon [/bib_ref] [bib_ref] When is a drug-related death not a drug-related death? Implications for current..., Beynon [/bib_ref].
Four interviewees discussed vein damage as a consequence of long term intravenous drug use and the fact that this made continuing to inject in their arms difficult. Consequently, interviewees had switched to more risky injecting practices (which included injecting into their groin and feet [bib_ref] Vascular complications of injecting drug misuse, Woodburn [/bib_ref] to overcome problems of gaining venous access, despite being aware of the elevated risk such practices convey;
"It's harder now to inject because your veins. Your veins sort of go y'know... I've been injecting in my foot, which is stupid really, y'know. You can catch an infection and you could end up losing your leg".
One man talked of having been hospitalised for deep vein thrombosis (a blood clot which forms in a deep vein), a recognised potential consequence of intravenous drug use for all ages [bib_ref] Vascular complications of injecting drug misuse, Woodburn [/bib_ref]. However, deep vein thrombosis becomes more likely among older injectors, particularly for those who have been heavy smokers, because in addition to the damage to veins caused by injecting, the ageing process is associated with changes in blood pressure, venous valve deterioration and reduced regenerative process; healthcare staff should ensure that older drug users, especially those who smoke, are aware of this increased risk.
Only three of the 10 interviewees complained of having respiratory problems despite research from the USA showing heroin-dependent males to have high rates of obstructive lung disease [bib_ref] Health conditions among aging narcotic addicts: medical examination results, Hser [/bib_ref]. This difference may be due to the men recruited in the USA study being predominantly smokers of heroin while the majority of those interviewed in this study discussed injecting as their mode of drug administration. One man interviewed in the present study was a heavy crack cocaine smoker and complained of respiratory problems and how this affected him; "I get out of breath easy. I absolutely panic then". A second crack cocaine smoker, who had experienced bronchial pneumonia twice, concurred; "it's tender inside and when I breathe I can't breathe hard cos it hurts like hell". The link between smoking any substance, including cigarettes, and respiratory complications is well known. Clinical guidelines recognise the high prevalence of smoking-related disorders among drug users, and the importance of offering smoking cessation interventions for smokers.
The people interviewed in the present study had used a variety of drugs across their life course, including alcohol, and had complex physical health needs. Drug use into older age can be seen as accelerating the physical health changes commonly associated with ageing. Furthermore, immunosenescence, the notion that there is an age related dysfunction of the immune system which leads to enhanced risk of infection, is an issue for all older people but will presumably have greater consequences for those older drug users who have experienced a lifetime of deprivation and poor environmental condition, making them particularly susceptible to opportunistic diseases such as pneumonia and other infections when compared to younger drug users or older people who do not use drugs [bib_ref] The role of substance use in non-drug related deaths: a cross-sectional study..., Beynon [/bib_ref]. Moreover, chronic poor nutrition will compound the physical problems arising from concomitant drug use and ageing; "I don't know but I think I'm seven stone now... I only have chips seven days a week". Consequently, for older drug users who have survived overdoses and other negative acute effects of drug use, it appears to be the longer term effects of drug use that become important in terms of their impact on physical health. Healthcare services, including drug specific services therefore need to recognise the impact of chronic multiple drug use on the physical health of older drug users and offer a package of care that addresses the person's general health in addition to the provision of drug treatment [bib_ref] The role of substance use in non-drug related deaths: a cross-sectional study..., Beynon [/bib_ref].
## Mental health
The links between mental health and substance use are complex but well documented and a high prevalence of co-morbidity exists. One study conducted in the UK, for example, reported that 44 percent of patients in community mental health services reported problem drug use and/or harmful drinking in the previous year, while 75 percent of people in contact with substance misuse services reported a psychiatric disorder during the same period of time [bib_ref] Comorbidity of substance misuse and mental illness in community mental health and..., Weaver [/bib_ref]. Psychological conditions, including depression, loneliness, anxiety, memory problems, cognitive impairment, dementia and confusion become more prevalent in older age. Ageing may therefore be associated with further elevation in the prevalence of mental health disorders in a population where the prevalence of such conditions is already considerably higher than that of the general population. Furthermore, prolonged use of some psychotropic drugs for example the benzodiazepines, is actually associated with depression and cognitive decline [bib_ref] Drugs of abuse and the aging brain, Dowling [/bib_ref]. Under-detection of such conditions including depression is relatively common among all older peoplebut may be a more significant issue for drug users because a considerable proportion do not seek treatment in services where screening for such disorders occurs [bib_ref] Hidden need for drug treatment services: measuring problematic drug use in the..., Beynon [/bib_ref].
Four interviewees commented that, during the course of their lives, their drug use had started, escalated or resumed following significant stressful events. For example, when discussing a relationship break up one man said; "You reach a point in our life, different points in your life where you've got a lot of stress and the easiest thing in the world is to say 'oh fuck it'. And once you've said that, if you're a drug user, a drug abuser, once you've said that, you're on the rocky road to destruction again". Across the life course, feeling sad or depressed due to stressful life events accounted for the reason why some of our interviewees recommenced drug use following long periods of abstinence and that using drugs was a means to forget about worries; That's all we do anyway, just take it [heroin] and just mong [be relaxed and docile] for the day like and forget about your worries and everything, got problems and you forget all that. Things like that; it's the reason why people take it like".
These findings highlight the need for healthcare services to be vigilant for drug and alcohol use by older adults, even among those who have ceased use for many years, while social support services for older people should also be aware of the need to screen people for drug and alcohol use, particularly those in contact during times of stress such as occurs following bereavement or relationship failure. However evidence from the USA suggests that healthcare professionals are currently poor at diagnosing substance use among older people [bib_ref] Drug abuse in older US adults worries experts, Boddiger [/bib_ref] possibly because of a lack of awareness that substance use among older adults occurs. Furthermore, the current screening tools for drug use, for example the Diagnostic and Statistical Manual of Mental Disorders IV for substance abuse, have not been validated for older populations [bib_ref] Drugs of abuse and the aging brain, Dowling [/bib_ref] , furthering problems of diagnosis.
Interviewees talked in terms of a life using drugs being depressing; "It's a depressing life y'know. It's no good" and how life not using drugs (where this was the case) had been more positive; "I was high on myself y'know [when not using drugs]. I'd look in the mirror and see somebody I loved. I don't any longer.... I can't see a road back to loving myself again". Conversely however, two interviewees were comparatively content with life at the time of the interview, with one recognising a link between reduced drug use and improvements in mood; "The ale was just a depressant wasn't it? And the heroin was just a depressant. Just making myself lower and lower all the time, wasn't I? Now I'm off the two of them I'm happy as a lark".
The loss of friends, predominantly through drug use, was a common theme through the interviews and was also associated with feelings of sadness and depression; this finding is unsurprising in light of extensive evidence to show that strong social networks have a positive effect on the health of older people, particularly on their mental health [bib_ref] The social capital of older people, Gray [/bib_ref]. Loss of friends, isolation and changes in social networks are issues experienced by many older people [bib_ref] Longitudinal studies of social networks and mortality in later life, Bowling [/bib_ref] but this is intensified among drug users and occurs earlier in life because rates of mortality are higher than in the general population. Furthermore, older drug users tend to associate exclusively with other drug users, having severed links with non-drug using family and friends over the years of drug taking [bib_ref] The drug career of the older injector, Levy [/bib_ref] , a premise supported by the findings of our interviews. Consequently, survivors are left alone or in contact with younger drug users from a different peer group. One particularly erudite man who had recently experienced the last of his friends die said; These results confirm findings from the USA, which highlighted isolation and loneliness as important issues for older drug users [bib_ref] The drug career of the older injector, Levy [/bib_ref]. For one interviewee, isolation was heightened by being morbidly obese and unable to walk more than a few steps; a physical state derived from cocaine-induced agoraphobia and inactivity;
[formula] " [/formula]
## "but i got so heavy with it [cocaine use] that i ended up getting agoraphobia cos i spent, must have been nearly four years in bed, just in bed...i think it was the four years what i ended up spending in bed that sort of got me into the state of putting on so much weight".
Furthermore, the isolation experienced by some older people may be exacerbated among older drug users by drug-related paranoia, experienced by three interviewees; "I don't like going to pubs. I always get paranoid... Everybody's looking at me... Drugs do it to you".
Younger drug users also experience friends dying (particularly through acute drug toxicity) but they may be somewhat cushioned from the effects by having larger social networks and more opportunities to make new friends. Younger drug users are also more likely to be in contact with their families. As older drug users die, the pool of peers grows ever smaller for those older drug users left behind. Of interest to this discussion, three interviewees appeared to have overcome this problem by forming new support structures, through developing friendships with much younger drug users who acted as carers. As the mobility, physical and mental capacity of older people fade they need to develop contacts with younger people to secure emotional and practical support. This is a considerable challenge for those without children [bib_ref] The social capital of older people, Gray [/bib_ref] , yet it appears that symbiotic alliances between older drug users and younger drug users develop out of this practical necessity. For example, one interviewee who had no family, had seen most of his friends die, who had recently had a stroke and who rarely left his house relied heavily on a much younger friend for assistance with daily tasks like collect-ing shopping, his benefits and his methadone prescription and doing the cooking. His younger friend, who was present during the interview comments "he helps me as I help him. We're good for each other". The younger drug user was not the focus of the interview and how exactly he benefited from this relationship was not explored although he did say that he enjoyed the company of the older drug user and suggested that he used his older friend's home as a preferable place to use drugs despite having his own place to live; "I'd rather do it [use drugs] here [older drug user's home] than some lavatory somewhere". Furthermore, the interviewee alluded to the fact that his younger friend might benefit financially be referring to his own social security payments in relation to "the money we get". These findings show a high level of adaptation and resilience among older drug users in relation to social support and we recommend further research should explore these relationships in more detail.
In general, ageing is associated with a decline in memoryand interviewees were specifically asked about this. Problems with memory, confirmed by six interviewees, were particularly related to what they termed 'short term' memory, for example, putting an object down and not remembering where it was placed a few minutes later. Of those who commented, two felt that age had contributed to their lack of memory, while most felt this was more associated with drug and alcohol use; "I can only put it down to just, down to drug abuse and drink abuse, really. I can't really put it down to anything else". The older interviewee detailed in the paragraph above had a significantly impaired memory and relied greatly upon his younger male friend/carer. Of particular importance, the interviewee relied upon this friend to remember to collect his methadone prescription, a drug which the older interviewee needed to consume daily. We know that high levels of alcohol consumption result in memory impairment due to thiamine deficiency, and the risk of even mild cognitive impairment in later life is significantly elevated among those who consume alcohol frequently compared to those who drink infrequently [bib_ref] Alcohol drinking in middle age and subsequent risk of mild cognitive impairment..., Anttila [/bib_ref]. Continuing drug use into older age may carry particular risks in terms of the concurrent effect on the brain; the brain changes in a number of ways across the life course, and the effect of these changes when combined with substance use are not well understood [bib_ref] Drugs of abuse and the aging brain, Dowling [/bib_ref].
## Health service contact
Interviewees discussed contact with both generic and specialist services. Care received in hospitals was not always optimal and some who had received care within this setting perceived hospital staff to treat them differently because they were drug users; "Like normally when you go into hospital and they find out that you're a heroin addict, they normally don't want to know you". One man found his expe-rience of hospital particularly humiliating with staff assuming he had hidden drugs within the toilet; "And they come in one night and says they couldn't find the toilet I put it [the drugs] in. In front of everybody and they put it in one of them big needles and squirted into my mouth in front of everybody. Makes you feel that big". In another instance, a confrontation caused by the misdiagnosis of a broken leg resulted in the police being called;
"They didn't diagnose it as a broken leg. They said there was nothing on the x-ray and I was in agony, y'know, with a walking stick and a broken leg. And I had to go three times to the hospital y'know... And I went back and they missed it again. And by then they're getting the police to throw me out the hospital".
The above negative accounts of experiences were countered by descriptions of instances of being cared for by healthcare staff with compassion and sensitivity; "the staff on that ward, they were good to me... I was happy in there. I could have stayed in there".
The de facto national service framework for drugs and alcohol highlights the importance of integrated care pathways and including non-drug specific services such as hospitals in the treatment of drug users; in particular in screening for drug use and referral into specialist services. While this integration is important for drug users of all ages, older drug users are more likely than their younger counterparts to be in contact with hospitals for a number of general chronic conditions, in addition to those specifically associated with their drug use. Hospitals therefore offer additional opportunities for screening and referral for older drug users and it is important that stigmatisation does not deter them from seeking hospital assistance.
Finally, with respect to hospitals, one man complained that doctors would not prescribe his dying friend, also a drug user, analgesia because he was using methadone as part of his drug treatment. The safe use of opioids to manage pain does present a challenge, but there are a number of reasons why drug users may have greater needs than non-drug users regarding pain management, and drug use per se should not preclude a practitioner prescribing an opioid. In particular, people with a history of drug use that need palliative care, should not be denied opioids to manage pain, and often these will need to be given in conjunction to drugs received as part of substitution therapy such as methadone. While mentioned by a single interviewee with only indirect experience, pain management and palliative care are likely to become important issues as a growing number of drug users reach older age and stigma and discrimination should not be allowed to affect care.
Other non-drug related care discussed by interviewees included generic rehabilitative treatment and this tended to be perceived as poor. The man who presented at hospital with a broken leg, for example, bemoaned; "I can't get a physiotherapist to come out. Or she did and she says 'oh that's great progress'. Just to tick me off the list y'know". A second man, who found the impaired verbal skills that followed his stroke very frustrating, had been offered, but had not accepted, speech therapy. He appeared to lack the understanding that such therapy could be advantageous; "I think my stroke will get better on its own if it's going to get better. I really do. I don't think anybody can do anything about it". It is a well accepted phenomenon that the least advantaged in society often make the least demands of specialist health services [bib_ref] Socioeconomic disparities in health care: Does universal coverage reduce inequalities in health, Veugelers [/bib_ref] despite their needs and our results confirm this; interviewees tended to have poor health but low expectations of specialist health services.
With respect to the views expressed by the interviewees about their care within drug services, no issues unique to older drug users emerged except for one man who said of his recent experience in residential rehabilitation that he was "a bit embarrassed, y'know, being the oldest one" and a second interviewee commented that there were many more services available to drug users now compared to when he had been younger. Interviewees were complimentary about outreach staff (who visit drug users in their homes and usually make few demands of their clients) and non-clinical drug service staff; "This place [drug service] is sound . They're all sound in here. Like the girl who I see, [member of staff's name], she's brilliant her... nothing bad to say about her". However, some complained about the doctors. One man, for example, was distressed and complained that a doctor would not increase his prescription of methadone ampoules, only his methadone 'linctus' which he complained made him vomit. Doctors in the UK are advised to prescribe methadone in an oral formulation and not injectable ampoulesbecause injecting carries a greater risk of dying from acute drug toxicity, and this situation demonstrates the difficulties clinicians face in terms of offering care in accordance with clinical guidelines, while still meeting the needs of their clients.
# Conclusion
Qualitative research makes no claims to generalisability; this study was carried out with white, predominantly male drug users in Merseyside, UK and included only ten people. Furthermore, these drug users were recruited through drug services and were all, or had been, problematic drug users (users of opiates and/or crack cocaine) and may not be representative of drug users not in contact with such services or users of different drugs. However, these findings add to knowledge and understanding of the physical and mental health of older drug users and their percep-tions of healthcare. We show that in general, drug users aged over 50 have poor levels of physical and mental health; our findings support the premise that substance use into older age exacerbates, or accelerates the onset of, medical conditions which are more prevalent in older age. Consequently, in relation to their needs, our interviewees appear to fit into the 'frail older people' category which is normally associated with people of a much older age.
We also highlight cases where drug users perceive their generic healthcare to be sub-optimal and conclude that older drug users, like other lower socio-economic groups, have relatively poor expectations of specialist healthcare. How drug services adapt to this ageing population of drug users and how geriatric services adapt to an increasing number of older people who use drugs need urgent consideration. People interviewed in the present study were recruited through drug services and those older people who are drug dependent and not in contact with services may be presumed to have even greater unmet needs.
Ageing drug users continue to be a neglected sub-group in the UK, largely because of societal attitudes about the behaviours deemed relevant to older people. The recently published national drug strategyfor example, failed to mention this change in demographic nor how services should respond to this challenge. This issue is not unique to the UK; the number of older drug users living in the UK and other developed nations will riseand further work is urgently needed to evaluate what challenges this change in demographic presents.
Publish with Bio Med Central and every scientist can read your work free of charge
[fig] ": BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime. Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours -you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv. [/fig]
|
Anesthesia for elective bilateral sagittal slip osteotomy of the mandible and genioplasty in a young man with Klippel-Feil syndrome, Sprengel deformity, and mandibular prognathism
Klippel-Feil syndrome is characterized by congenital fusion of two or more cervical vertebrae, a low hair line at the back of the head, restricted neck mobility, and other congenital anomalies. We report a 16-year-old young man with Klippel-Feil syndrome, Sprengel deformity of the right scapula, thoracic kyphoscoliosis, and mandibular prognathism with an anterior open bite. He was treated with orthodontic treatment and maxillofacial surgery. An anticipated difficult airway due to a short neck with restricted neck movements and extrinsic restrictive lung disease due to severe thoracic kyphoscoliosis increased his anesthesia risk. Due to his deviated nasal septum and contralateral inferior turbinate hypertrophy, we chose awake fiber optic orotracheal intubation followed by submental intubation. Considering the cervical vertebral fusion, he was carefully positioned during surgery to avoid potential spinal injury. He recovered well and his postoperative course was uneventful.
# Introduction
Klippel-Feil syndrome (KFS) is a congenital musculoskeletal condition characterized by the fusion of two or more cervical vertebrae. Common symptoms include a short neck, low hair line at the back of the head, and restricted neck mobility. Its incidence is one in 40,000-42,000 newborns worldwide with a slight predominance in females. KFS was initially reported in 1912 by Maurice Klippel and Andre Feil. The treatment of symptomatic patients includes several surgical procedures involving a multidisciplinary team of a neurologist, orthopedic surgeon, pediatrician, nurse practitioner, physical therapist, neurologist, neurosurgeon, and maxillofacial surgeon. While some patients have no symptoms, a significant number have myelopathy and neuropathy, which significantly decrease quality of life. KFS may be associated with Sprengel deformity, scoliosis, deafness, congenital heart disease, craniomaxillofacial deformities, pulmonary defects, facial nerve palsy, cleft palate, spina bifida, and genitourinary malformation. Mutations in the GDF6, GDF3, and MEOX1 genes can cause KFS. The GDF6 gene is involved in proper bone formation, while the GDF3 gene is involved in bone development. The MEOX1 gene creates the homeobox protein MOX1, which regulates separation of the vertebrae. GDF6 and GDF3 gene abnormalities are inherited in an autosomal dominant pattern, while MEOX1 mutations are autosomal recessive. Patients with KFS may be predisposed to congenital spinal stenosis; thus, a relatively low-impact injury may induce a significant neurological deficit. This syndrome may occur simultaneously with Sprengel deformity, Duane syndrome, renal agenesis, fetal alcohol syndrome, Goldenhar syndrome, Wildervanck syndrome, and other vascular and cardiac abnormalities. Approximately 50% of patients with KFS will present with concurrent scoliosis. Treatment is conservative and symptomatic only. In particular, with single or bilevel fusions below the third cervical vertebra are monitored and treated conservatively. We highlight this case to discuss anesthesia management during orthognathic surgery for patients with KFS.
## Case report
Here, patient-related factors like an anticipated difficult airway, thoracic kyphoscoliosis, cervical vertebral fusion with a potential of serious neurological injuries and other system involvement require consideration. Elective postoperative care is very important in these patients considering the potential for postoperative airway obstruction due to laryngeal edema and soft tissue swelling.
## Author orcids
Rathna Paramaswamy: https://orcid.org/0000-0002-0784-6952 |
Chronic Disease, Disability, Psychological Distress and Suicide Ideation among Rural Elderly: Results from a Population Survey in Shandong
Objective: Suicide is a major public health and social problem in contemporary societies. Previous studies showed that the older the seniors were, the more likely it was that they would experience disability, chronic disease, or both. The objective of this study was to examine the joint effects of chronic disease and physical disability on suicide ideation while controlling for psychological distress among the rural elderly living in Shandong Province, China. Method: A total of 5514 rural elderly individuals (60+) living in Shandong Province, China were included in this study. Suicidal ideation was assessed by using questions from the National Comorbidity Survey (NCS). Multiple logistic analyses were performed to examine the factors associated with suicide ideation. A path analysis was conducted to test the direct and indirect effects of chronic disease and of activity of daily living (ADL) limitation on suicide ideation while controlling for psychological distress. Results: The prevalence of suicide ideation among the rural elderly in Shandong, China was 11.0%. Psychological distress had the strongest direct (β = 0.392) and total effect (β = 0.392), chronic disease (β = −0.034; β = −0.063) had both direct and indirect impacts, and ADL (β = 0.091) had indirect impacts on suicide ideation. Psychological distress was a mediator between chronic disease, ADL limitation, and suicide ideation. Conclusions: Psychological distress was the greatest influencing factor of suicide ideation among the rural elderly, followed by chronic disease and disability. Effective intervention measures should be taken to facilitate the early detection of psychological distress in clinical practice among the rural elderly.
# Introduction
Suicide is a major public health and social problem in contemporary societies, and lays a huge economic, social, and psychological burden on individuals, families, communities, and countries. Every year, around 800,000 people worldwide die due to suicide, which corresponds to an adjusted global rate of 11.4/100,000 in 2012 . China, where suicide is ranked as the fifth leading cause of death, accounts for an estimated 22% of global suicides, or roughly 200,000 deaths per year [bib_ref] Women and suicide in rural China, Cui [/bib_ref].
Although the suicide rates in China have decreased in the past two decades [bib_ref] Economic growth and suicide rate changes: A case in China from 1982..., Zhang [/bib_ref] , the peak for the elderly has remained relatively unchanged and has even slightly increased in recent years, making 2 of 10 suicide among the elderly in China a prevalent public health problem which we should address [bib_ref] Suicide rates in China from 2002 to 2011: An update, Wang [/bib_ref]. Some previous studies conducted in China and other countries have indicated that the suicide rate was higher among the elderly [bib_ref] Analysis of suicide in the elderly in Italy. Risk factors and prevention..., Terranova [/bib_ref] [bib_ref] Suicide among the elderly in mainland China, Li [/bib_ref]. In addition, the rural elderly suicide rates are three to five times higher than those in urban areas [bib_ref] Suicide rates in China, Phillips [/bib_ref] [bib_ref] Suicide rates in China during a decade of rapid social changes, Yip [/bib_ref]. Suicide ideation is an important risk factor for attempted and completed suicide [bib_ref] The prediction of suicide, Sensitivity, specificity, and predictive value of a multivariate..., Goldstein [/bib_ref]. For suicide ideation, S. H. Kim found that chronic illnesses and functional limitations were associated with an increased risk of suicide ideation [bib_ref] Suicidal ideation and suicide attempts in older adults: Influences of chronic illness,..., Kim [/bib_ref]. H. Xu, L. Qin, and J. Wang et al. had similar results [bib_ref] A cross-sectional study on risk factors and their interactions with suicidal ideation..., Xu [/bib_ref]. Chronic disease and disability were also determinants of "high" psychological distress [bib_ref] Income-related inequalities in chronic conditions, physical functioning and psychological distress among older..., Korda [/bib_ref] [bib_ref] Mental Disorder, Psychological Distress, and Functional Status in Canadian Military Personnel, Sampasa-Kanyinga [/bib_ref]. Social demographics and other factors such as depression and insomnia have been associated with suicide ideation [bib_ref] Suicidal ideation and associated factors among community-dwelling elders in Taiwan, Yen [/bib_ref] [bib_ref] Suicidal ideation in elderly Korean population: A two-year longitudinal study, Kang [/bib_ref] [bib_ref] Insomnia as a predictor of high-lethality suicide attempts, Pompili [/bib_ref].
Previous studies have identified the relationship between chronic disease, disability, psychological distress, and suicide ideation. However, there have been few ideas concerning the direct and indirect effects of activity of daily living (ADL) limitation, chronic disease, and psychological distress on suicide ideation. The purpose of the current study is to examine the joint effects of chronic disease and physical disability on suicide ideation among the elderly while controlling for psychological distress. Specifically, chronic disease has a direct effect on suicide ideation, and both chronic disease and physical disability have an indirect effect on suicide ideation through psychological distress. In particular, the current study will address the following research questions and hypotheses. First, chronic disease and physical disability are associated with suicide ideation. Second, chronic disease has a partial direct effect and a partial indirect effect on suicide ideation through psychological distress. Third, physical disability has a full indirect effect on suicide ideation through psychological distress.
# Method
## Settings and participants
The design of the study consisted of face-to-face interviews using a structured questionnaire. This study was collected from the 2017 Survey of the Shandong Elderly Family Health Service, which was conducted by Shandong University. Stratified multi-stage random sampling was applied: in the first stage, six counties were selected from 137 counties as the primary sampling units (PSUs) throughout the eastern, central, and western regions of Shandong Province (which were divided into three districts and three counties that represented urban and rural areas, respectively). From each PSU, 18 villages in rural areas and 18 communities in suburban and urban areas were selected in the second stage as the secondary sampling units (SSUs). In the third stage, based on the roster of the residents by age and the total elderly population of each selected site provided by the local residential committee, an average of 66 individuals were stratified and randomly selected from each SSU, making up the total sample. The eligible participants for this survey were those aged 60 and above with local household registrations at the time of the interview. Initially, 7088 elderly individuals were selected and interviewed. Of these, 18 did not complete the survey because of an uncompleted questionnaire or an obvious logical error in the questionnaire. Finally, a total of 7070 individuals were included in the sample [fig_ref] Figure 1: Flow chart of participant enrolment [/fig_ref]. Our current study focused on respondents who answered the question "Have you ever seriously thought about committing suicide?" among the rural elderly (n = 5514).
## Variables and measures
## Dependent variable
Suicide ideation was assessed by using a question from the baseline National Comorbidity Survey (NCS) [bib_ref] Perceived abuse and neglect as risk factors for suicidal behavior in adolescent..., Lipschitz [/bib_ref] : "Have you ever seriously thought about committing suicide?" The question was designed to ask whether they had experienced this idea in their lifetime. For the purpose of the analysis, a positive response was considered to indicate the presence of suicide ideation.
## Variables and measures
## Dependent variable
Suicide ideation was assessed by using a question from the baseline National Comorbidity Survey (NCS) [bib_ref] Perceived abuse and neglect as risk factors for suicidal behavior in adolescent..., Lipschitz [/bib_ref] : "Have you ever seriously thought about committing suicide?" The question was designed to ask whether they had experienced this idea in their lifetime. For the purpose of the analysis, a positive response was considered to indicate the presence of suicide ideation.
## Independent variables
## Physical disability
The elderly individual's physical disability was measured using the ADL Scale, including ADL subscales [bib_ref] Studies of illness in the aged. The index of ADL: A standardized..., Katz [/bib_ref] and instrumental activity of daily living (IADL) subscales [bib_ref] Assessment of older people: Self-maintaining and instrumental activities of daily living, Lawton [/bib_ref]. The ADL subscale assesses six types of ability: feeding, dressing, bathing, toileting, grooming, and locomotion. The IADL subscale evaluates the ability to perform eight types of more complex activities, like using a telephone, using transportation, and shopping. Scores for performing activities range from 1 to 4 (1 point for each activity performed without help and 4 points for each activity that the individual is unable to perform). The maximum score is 56 (higher scores indicate greater dependence). Edwards confirmed the ADL scales in a Brazilian study, showing a Cronbach's α of 0.96-0.99 [bib_ref] The reliability and validity of self-report activities of daily living scales, Edwards [/bib_ref]. The Cronbach's α for the ADL was 0.93 in this sample.
## Psychological distress
Kessler 10 (K10) is a recognized and validated measure of psychological distress, in which respondents report how frequently they experienced specified symptoms of psychological distress [bib_ref] Evaluation of 2 measures of psychological distress as screeners for depression in..., Cairney [/bib_ref] [bib_ref] Short screening scales to monitor population prevalences and trends in non-specific psychological..., Kessler [/bib_ref]. An individual's sense of psychological distress is evaluated against ten items on a self-report questionnaire. Each item is rated on a five-point scale with numerical values assigned to the responses (none of the time, a little of the time, some of the time, most of the time, and all the time). The ratings are summed to yield a total score, with higher scores indicating higher levels of psychological distress. The scale's internal consistency, as assessed by Cronbach's α, was 0.87. The Cronbach's α for the K10 was 0.92 in this sample.
## Socio-demographic characteristics
The socio-demographic characteristics included chronic diseases in the past 6 months, gender, age, education level, residential status, self-rated health, and family relationship. Chronic disease was coded as yes and no in the past 6 months. Gender was coded as male and female. Education
## Independent variables
## Physical disability
The elderly individual's physical disability was measured using the ADL Scale, including ADL subscales [bib_ref] Studies of illness in the aged. The index of ADL: A standardized..., Katz [/bib_ref] and instrumental activity of daily living (IADL) subscales [bib_ref] Assessment of older people: Self-maintaining and instrumental activities of daily living, Lawton [/bib_ref]. The ADL subscale assesses six types of ability: feeding, dressing, bathing, toileting, grooming, and locomotion. The IADL subscale evaluates the ability to perform eight types of more complex activities, like using a telephone, using transportation, and shopping. Scores for performing activities range from 1 to 4 (1 point for each activity performed without help and 4 points for each activity that the individual is unable to perform). The maximum score is 56 (higher scores indicate greater dependence). Edwards confirmed the ADL scales in a Brazilian study, showing a Cronbach's α of 0.96-0.99 [bib_ref] The reliability and validity of self-report activities of daily living scales, Edwards [/bib_ref]. The Cronbach's α for the ADL was 0.93 in this sample.
## Psychological distress
Kessler 10 (K10) is a recognized and validated measure of psychological distress, in which respondents report how frequently they experienced specified symptoms of psychological distress [bib_ref] Evaluation of 2 measures of psychological distress as screeners for depression in..., Cairney [/bib_ref] [bib_ref] Short screening scales to monitor population prevalences and trends in non-specific psychological..., Kessler [/bib_ref]. An individual's sense of psychological distress is evaluated against ten items on a self-report questionnaire. Each item is rated on a five-point scale with numerical values assigned to the responses (none of the time, a little of the time, some of the time, most of the time, and all the time). The ratings are summed to yield a total score, with higher scores indicating higher levels of psychological distress. The scale's internal consistency, as assessed by Cronbach's α, was 0.87. The Cronbach's α for the K10 was 0.92 in this sample.
## Socio-demographic characteristics
The socio-demographic characteristics included chronic diseases in the past 6 months, gender, age, education level, residential status, self-rated health, and family relationship. Chronic disease was coded as yes and no in the past 6 months. Gender was coded as male and female. Education level was coded as illiteracy, primary, and middle or above. Residential status was coded as single and couple. Self-rated health was coded as good and bad. Family relationship was coded as good and bad.
# Statistical analysis
All statistical analyses were performed using SPSS 22.0 (IBM Corp, Armonk, NY, USA) and Amos 21.0 (IBM Corp, Armonk, NY, USA). t-test or χ 2 test was used to compare the difference in continuous or categorical variables. Logistic regression analysis was performed to examine the association of suicide ideation with some variables. Gender, ADL, chronic disease, and psychological distress were chosen as the independent variables in a path analysis model. The reported CI was calculated at the 95% level, and statistical significance was set at the 5% level. [fig_ref] Table 1: Socio-demographic factors associated with suicide ideation among the rural elderly in Shandong,... [/fig_ref] displays the socio-demographic details of the rural elderly about suicide ideation. A total of 5514 rural elderly individuals responded to the study. About 11.0% (n = 605) acknowledged suicide ideation (see [fig_ref] Table 1: Socio-demographic factors associated with suicide ideation among the rural elderly in Shandong,... [/fig_ref]. The age of the participants ranged from 60 to 101 (mean= 69.72; SD = 6.52), 57.1% were female, 80.8% lived with a spouse, and 39.2% was illiterate. Of the participants, 94.6% had a harmonious family. Furthermore, those people who had chronic disease accounted for 68.8%, whose self-rated health status was good accounted for 52.2%. The study compared participants in terms of the variables of age, gender, education level, residential status, self-rated health, family relationship, and chronic disease status in relation to suicide ideation. The participants differed significantly from each other in relation to suicide ideation.
# Results
## Socio-demographic characteristics of the sample
# Logistic regression analysis
The results of the logistic regression analysis of suicide ideation are shown in [fig_ref] Table 2: Logistic regression analysis for predictors of multiple suicide ideation [/fig_ref]. ADL and chronic disease were significantly associated with suicide ideation [fig_ref] Figure 1: Flow chart of participant enrolment [/fig_ref]. When we adjusted for psychological distress, the OR of ADL changed from 1.03 to 0.99 (Models 1 and 3) and the chronic disease status changed from 1.71 to 1.14 (Models 2 and 4). Age (OR = 0.97), education (primary OR = 0.73, middle or above OR = 0.49) and family relationship (OR = 1.39) were significantly associated with suicide ideation among the rural elderly in Shandong, China (Model 5). demonstrates the interaction between ADL limitation, chronic disease, psychological distress, and suicide ideation. Standardized regression coefficients are presented on each arrow. Using a path analysis, we found that psychological distress had a direct impact on suicide ideation. The influences of ADL on suicide ideation were mediated by psychological distress. Chronic disease had both direct and indirect effects on suicide ideation. The model had an absolute fit index (GFI) of 1.000, a simple fit index (P) of 0.866, a comparative fit index (CFI) of 1.000, and a root mean square error of approximation (RMSEA) of 0.000. Four fit indices indicated that the final path model adequately represented the data.
## The mediating effect of psychological distress between adl limitation and suicide ideation
(primary OR = 0.73, middle or above OR = 0.49) and family relationship (OR = 1.39) were significantly associated with suicide ideation among the rural elderly in Shandong, China (Model 5). demonstrates the interaction between ADL limitation, chronic disease, psychological distress, and suicide ideation. Standardized regression coefficients are presented on each arrow. Using a path analysis, we found that psychological distress had a direct impact on suicide ideation. The influences of ADL on suicide ideation were mediated by psychological distress. Chronic disease had both direct and indirect effects on suicide ideation. The model had an absolute fit index (GFI) of 1.000, a simple fit index (P) of 0.866, a comparative fit index (CFI) of 1.000, and a root mean square error of approximation (RMSEA) of 0.000. Four fit indices indicated that the final path model adequately represented the data. [fig_ref] Table 3: Standardized effects on suicide ideation from path analysis on psychological distress, chronic... [/fig_ref] presents the total, direct, and indirect effects of the variables on suicidal ideation. The total effects are the sum of the direct and indirect effects. Indirect effects represent the effects of one variable on another variable through indirect routes. We found that the value of standardized total effects of psychological distress was (0.392), while chronic disease was 0.096 and ADL limitation was 0.091.
## The mediating effect of psychological distress between adl limitation and suicide ideation
## Figure 2.
Direct and indirect effects of chronic disease, disability, and psychological distress on suicide ideation among the rural elderly in Shandong, China (OR and its 95% CI). ** p < 0.01, *** p < 0.001. . Direct and indirect effects of chronic disease, disability, and psychological distress on suicide ideation among the rural elderly in Shandong, China (OR and its 95% CI). ** p < 0.01, *** p < 0.001. [fig_ref] Table 3: Standardized effects on suicide ideation from path analysis on psychological distress, chronic... [/fig_ref] presents the total, direct, and indirect effects of the variables on suicidal ideation. The total effects are the sum of the direct and indirect effects. Indirect effects represent the effects of one variable on another variable through indirect routes. We found that the value of standardized total effects of psychological distress was (0.392), while chronic disease was 0.096 and ADL limitation was 0.091.
# Discussion
The lifetime prevalence of suicide ideation among the rural elderly was 11.0% (605 of 5514) in the current study. It was lower than the lifetime prevalence of 13.0% among the rural elderly living in Hunan Province, China [bib_ref] A case-control study of suicidal ideation among rural elderly in a city..., Tao [/bib_ref]. It was higher than the reported rates among the elderly in Shandong, China (4.2%) [bib_ref] Exploring the risk factors of suicidal ideation among the seniors in Shandong,..., Ge [/bib_ref] , Hong Kong (5.5%) [bib_ref] A prevalence study of suicide ideation among older adults in Hong Kong..., Yip [/bib_ref] , Taiwan (6.1%) [bib_ref] Prevalence and association of suicide ideation among Taiwanese elderly-A population-based cross-sectional study, Chan [/bib_ref] , and Canada (8.4%) [bib_ref] Gender differences in health service use for mental health reasons in community..., Vasiliadis [/bib_ref]. Because of the different levels of research backgrounds and the time point of definition, the prevalence of suicide ideation varied widely.
Psychological distress had the most powerful and direct relationship with suicide ideation in the current study. This is similar to previous studies. For example, one study found that psychological distress had a direct and mediating role in suicidal patients among adolescents [bib_ref] Suicidal ideation, suicide attempts, and psychological distress among intoxicated adolescents in the..., Puuskari [/bib_ref] , which was also a predictor for suicide ideation among the community-dwelling elderly [bib_ref] Psychological distress as a predictor of suicide ideation among the community-dwelling elderly, Fujita [/bib_ref]. This correlation was stronger in older people than in younger people [bib_ref] Gender differences in health service use for mental health reasons in community..., Vasiliadis [/bib_ref]. The K10 scale was used to represent psychological distress. Psychological distress was an effective screener to measure suicide ideation, which was composed of two primary factors (depression and anxiety) [bib_ref] The association between the Kessler 10 and suicidality: A cross-sectional analysis, O'connor [/bib_ref] [bib_ref] Suicidal ideation: The clinical utility of the K10, Chamberlain [/bib_ref].
In the current study, when psychological conditions were controlled, ADL limitation had little direct effect on suicide ideation. But suicide ideation was indirectly accompanied by psychological distress, which implies that psychological distress mediates the process of developing suicide ideation. Studies demonstrated that disability was significantly associated with suicidal ideation, particularly for the elderly [bib_ref] Factor structure and interpretation of the K10, Brooks [/bib_ref] [bib_ref] The influence of limitation in activity of daily living and physical health..., Dennis [/bib_ref]. Functional disability as a determinant of "high" psychological distress has been demonstrated by several studies [bib_ref] The Relationship Between Difficulties in Daily Living and Suicidal Ideation Among Older..., Zhang [/bib_ref]. One possible explanation for the mediating role of psychological distress was the lack of their daily housework and economic resources.
This study shows that chronic disease had both direct and indirect impacts on suicide ideation. Studies have found a positive association between chronic disease and suicide, such as coronary heart disease, diabetes and combined illness, etc. [bib_ref] Are lower levels of physical activity and self-rated fitness associated with higher..., Štefan [/bib_ref] [bib_ref] Association of chronic disease prevalence and quality of life with suicide-related ideation..., Joshi [/bib_ref] , which was similar to the findings of our study. Using a path analysis, we proposed that chronic disease would affect suicide ideation directly or through psychological distress, and we found that it had both direct and indirect effects on suicide ideation among the elderly. Some previous studies demonstrated that chronic disease was associated with psychological distress among the elderly [bib_ref] Combined effects of physical illness and comorbid psychiatric disorder on risk of..., Qin [/bib_ref] [bib_ref] Psychological distress and hypertension: Results from the National Health Interview Survey for..., Ojike [/bib_ref] [bib_ref] Psychological distress and type 2 diabetes mellitus: A 4-year policemen cohort study..., Li [/bib_ref]. Furthermore, psychological distress was found to be a predictor for suicide ideation among the elderly [bib_ref] Association between psychological distress and liver disease mortality: A meta-analysis of individual..., Russ [/bib_ref]. One possible explanation was that high levels of disease burden were more sensitive to psychological distress. In addition, the internet might also have positive effects on particularly older people who obtain information about their (chronic) diseases that contribute to their stress and access social support that is beneficial to their mental health [bib_ref] A cross-sectional study of family caregiver burden and psychological distress linked to..., Abreu [/bib_ref]. There is a considerable body of research on the effects, both positive and negative, of media on suicide among different age groups [bib_ref] First do no harm: Cross-sectional and longitudinal evidence for the impact of..., Scherr [/bib_ref]. The implication of this finding was that chronic disease may not only directly affect, but also prevent suicide ideation by reducing the psychological distress among elderly people.
Age, education, and family relationship were significantly associated with suicide ideation in our study, but we mainly studied the relationship between chronic disease, ADL limitation, psychological distress, and suicide ideation.
# Strength and limitations
The large sample size (n > 5000) provided strong statistical support for our study. However, the study still has several limitations, as indicated below. First, this study was based on a house-to-house interview, but village clinic interviews and house-to-house interviews were combined. The results may be biased because the individuals who arrived to the clinics usually had normal ADL, whereas the elderly with ADL limitation may not have been able to be interviewed. Second, information, including suicide ideation and health status, was self-reported, inevitably leading to the possibility of subjective bias. Third, we mainly focused on the risk factors for suicide ideation in the current study, and gave little consideration to the relationship between suicide ideation and attempted and completed suicide. Fourth, the dependent variable was lifetime suicide ideation. Chronic disease was diagnosed clearly by the medical staff in the first half of the investigation, but this study could not determine if suicide ideation occurred before, after, or both before and after experiencing psychological distress and health problems. Finally, the ORs for all of these variables were low. There were many statistically significant findings, but the effects noted in the data provided a limited ability to predict suicide ideation.
# Conclusions
Psychological distress was found to have the greatest total and direct effect on suicide ideation among rural elderly, followed by chronic disease and disability. Chronic disease had a partial direct effect and a partial indirect effect on suicide ideation through psychological distress, but disability only had an indirect effect on suicide ideation through psychological distress. These findings imply a
[fig] Figure 1: Flow chart of participant enrolment. [/fig]
[table] Table 1: Socio-demographic factors associated with suicide ideation among the rural elderly in Shandong, China. [/table]
[table] Table 2: Logistic regression analysis for predictors of multiple suicide ideation (OR and its 95% CI). [/table]
[table] Table 3: Standardized effects on suicide ideation from path analysis on psychological distress, chronic disease and activity of daily living (ADL) among the rural elderly in Shandong, China. [/table]
|
Screening and validation of lymph node metastasis risk-factor genes in papillary thyroid carcinoma
Background: Although most papillary thyroid carcinoma (PTC) cases have a good prognosis, some PTCs are more aggressive and are often accompanied by lymph node (LN) metastasis, a high recurrence rate, and poor prognosis. Distinguishing highly invasive metastatic PTC is an urgent problem that needs to be addressed clinically. We analyzed a microarray of metastasized PTC and validated it using quantitative reverse transcription PCR (RT-qPCR) and immunohistochemistry to identify biomarkers that can be used to assess the risk of PTC metastasis.Methods: The microarray of metastasized PTC was screened using the Gene Expression Omnibus (GEO) database. The differences between cancer and normal tissues were analyzed using the official GEO tool: GEO2R. Gene expression profile data (GEPIA) were used to verify the expression of differential genes in large samples and to analyze their correlation. The Kaplan-Meier plotter (KM-plotter) database was used for the analysis of genes potentially related to survival. RT-qPCR was used to check the expression of risk factor genes in pathological sections from PTC patients with clinical LN metastasis. Immunohistochemistry was used to verify the expression of core risk-associated genes.Results: Fourteen PTC metastasis-associated genes were identified. In metastasized PTC, CLDN1, LRP4, LRRK2, and TENM1 were highly expressed, whereas DIO1, DPP6, HGD, IPCEF1, MT1F, SLC26A4, SLC26A7, SPX, TFF3, and TPO were expressed at low levels, compared to expression in normal tissues. DIO1, HGD, SLC26A4, and TPO were found to be the core risk genes in the PTC metastatic risk set. Results based on clinical samples showed that the expression differences for metastasis risk-associated genes were consistent with the bioinformatics analysis results. Conclusions: Fourteen differentially expressed genes (CLDN1, LRP4, LRRK2, TENM1, DIO1, DPP6, HGD, IPCEF1, MT1F, SLC26A4, SLC26A7, SPX, TFF3, TPO) Frontiers in Endocrinology (2022) Screening and validation of lymph node metastasis risk-factor genes in papillary thyroid carcinoma. Front. Endocrinol. 13:991906.are associated with an increased risk of PTC metastasis, and DIO1, HGD, SLC26A4, and TPO are the key risk-associated genes in this set that might affect the occurrence and development of PTC through iodine metabolism. These genes could provide a reference for clinical metastatic PTC risk evaluation and treatment.
# Introduction
The current incidence of thyroid cancer (TC) accounts for 3.4% of all cancers, making it the tenth most prevalent cancer globally and seventh most prevalent in China . The reasons for the increased incidence of TC are complex. Overdiagnosis might be one of these, but an actual increase in TC cannot be ruled out [bib_ref] The thyroid cancer epidemic 2017 perspective, Roman [/bib_ref] [bib_ref] Incidence and mortality trends of thyroid cancer from 1980 to 2016, Wirth [/bib_ref]. TC can be distinguished into medullary thyroid carcinoma, anaplastic thyroid carcinoma (ATC), and differentiated thyroid carcinoma (DTC), based on the histological type. DTC can be further distinguished into follicular thyroid cancer (FTC) and papillary thyroid carcinoma (PTC). PTC accounts for approximately 95% of TC cases [bib_ref] New therapies for advanced thyroid cancer, Laha [/bib_ref]. Patients with early PTC have a 5-year survival rate >90% after conventional treatment, and these are generally considered "indolent tumors" [bib_ref] Overall survival of papillary thyroid carcinoma patients: A single-institution long-term follow-up of..., Ito [/bib_ref]. The recurrence rate of PTC is up to 30%, but the survival rate in patients with recurrence and metastasis is significantly reduced [bib_ref] A study of recurrence and death from papillary thyroid cancer with 27..., Grogan [/bib_ref]. Recurrence and metastasis are the key factors involved in PTC-associated mortality. Early detection and active intervention for PTC with a risk of metastasis is an effective strategy for improving prognosis. Unfortunately, a morphological evaluation of the primary tumor cannot be used to predict the likelihood of disease metastasis or recurrence.
Currently, fine-needle aspiration is the most reliable diagnostic tool for PTC; however, it cannot be used to cytologically distinguish between benign and malignant PTC and thus cannot be used to evaluate the risk of PTC metastasis [bib_ref] Advances in metabolomics of thyroid cancer diagnosis and metabolic regulation, Alexander [/bib_ref]. Previous studies have found characteristic genetic changes in PTC, such as RET-PTC, NTRK rearrangements, RAS and BRAF mutations, and PAX8/PPAR translocation [bib_ref] Update on fundamental mechanisms of thyroid cancer, Prete [/bib_ref] [bib_ref] Diagnostic approach to thyroid cancer based on amino acid metabolomics in saliva..., Zhang [/bib_ref] , which have been used for the diagnosis of PTC; however, the genetic changes currently found in PTC cannot be used to characterize tumors with different clinical behaviors, such as LN metastasis [bib_ref] Diagnostic approach to thyroid cancer based on amino acid metabolomics in saliva..., Zhang [/bib_ref] [bib_ref] Papillary thyroid carcinoma behavior: clues in the tumor microenvironment, Bergdorf [/bib_ref]. Identifying genetic changes that can help diagnose PTC and assess the risk of metastasis is crucial for early clinical intervention in high-risk PTC patients. Therefore, screening for genetic metastatic PTC-specific changes in patients with LN metastasis can provide a reference to assess the risk of PTC metastasis.
The present study aimed to identify prognostic biomarkers for clinical diagnosis and risk assessment with respect to PTC metastasis by screening risk-associated genes exhibiting differential expression. The overall goal of the study was to reduce metastasis and recurrence rates in PTC patients and improve their prognosis.
# Materials and methods
## Microarray data acquisition
The GEO (https://www.ncbi.nlm.nih.gov/geo/) database was searched using the keywords "Thyroid cancer, Human." Samples comprising metastatic PTC and matching paracancerous tissue were screened from the retrieved results.
## Data processing of degs
Using the GEO2R (https://www.ncbi.nlm.nih.gov/geo/info/ geo2r.html) online tool, we used a |logFC| > 2 to identify the DEGs between the metastatic PTC and normal thyroid gland tissue samples, and DEGs with P < 0.05 were screened. The raw data were then filtered online using the Venn diagram database to detect DEGs among the three datasets. DEGs with a logFC < −2 and logFC > 2 were considered downregulated and upregulated, respectively.
## Verification of degs
GEPIA (http://gepia.cancer-pku.cn/) was used to verify the expression of DEGs with statistically significant differences based on a large sample size.
## Analysis of degs associated with survival
The KM-plotter (https://kmplot.com/analysis/) was used to analyze the influence of genes on overall survival (OS) and relapse-free survival (RFS).
## Correlation analysis of degs
The obtained DEGs were filtered through the GEPIA database for correlation analysis. The correlation between each DEG in TC was obtained, and the core risk-associated genes in the set of metastasis-associated genes were screened based on these correlations. The R software (version 4.0.3) was used to visualize correlations.
## Validation of risk-associated gene sets
This study was reviewed by the Biomedical Research Ethics Committee of the Affiliated Hospital of Zunyi Medical University, batch no. KLL-2022-448. Preliminary data from major cancer databases represent a large sample size of real patient information. Therefore, in this study, eight clinical samples were collected for validation based on a small clinical sample set. The inclusion criteria were patients with confirmed PTC and metastatic lesions, between 18 and 70 years of age, who visited the Affiliated Hospital of Zunyi Medical University in 2020-2021. Patients with TC subtypes such as PTMC and ATC and other major diseases were excluded from the study. Pathological tissues of metastatic PTC patients meeting the inclusion criteria were collected and pathological sections and paraffin block were produced. The cancerous/paracancerous tissues were isolated from some sections under a microscope for RNA extraction. RT-qPCR was used to verify the 18 genes screened using bioinformatics associated with a risk of metastasis. Immunohistochemistry was performed to confirm the expression of the four core riskassociated genes that were highly associated with each risk gene.
## Rt-qpcr
RNA was extracted from tissues from patients with metastatic PTC using an RNeasy FFPE Kit (RNeasy FFPE Kit, QIAGEN, Hilden, Germany). cDNA was then synthesized using a reverse transcriptase system (SureScripTM First-Strand cDNA Synthesis Kit, GeneCopoeia, Guangzhou, China). RT-qPCR was used to validate gene expression in accordance with the manufacturer's instructions (BlazeTaq SYBR Green qPCR Mix 2.0, GeneCopoeia). The expression levels of all genes were normalized against ATCB gene expression, and relative gene expression was calculated using the 2 −DDCT method. Primers for RT-qPCR were purchased from Sangon Biotech (Shanghai, China).
## Immunohistochemistry
The tissues from patients with metastasized PTC were fixed with formalin, included in paraffin, and sections were prepared. For the immunohistochemical (IHC) analysis of deiodinase, iodothyronine type I (DIO1, 1:100, Affinity Biosciences, OH, USA), homogentisate 1,2-dioxygenase (HGD, 1:100, Santa Cruz Biotechnology, Dallas, TX, USA), solute carrier family 26 member 4 (SLC26A4, 1:100, Santa Cruz Biotechnology), and thyroid peroxidase (TPO, 1:100, Servicebio, Wuhan, China). Pathological changes were observed under an optical microscope (NI-U, Nikon, Tokyo, Japan), and images were captured with a digital camera (DS-RI2, Nikon).
## Institutional review board approval (or waiver) statement
In accordance with the Biosafety Law of the People's Republic of China and other laws, regulations, rules, normative documents, and international standards, this study was reviewed by the Biomedical Research Ethics Committee of Zunyi Medical University Hospital; the project was approved and the research conducted according to the scheduled plan (batch number, KLL-2022-448).
# Statistical analysis
The SPSS 22.0 software was used for all statistical analyses. All data are presented as the standard error of the mean and compared using Student's t-test. Statistical significance was set at P < 0.05.
# Results
## Microarray data information
The NCBI GEO database included three gene expression profiles that met the requirements (including metastasized PTC and normal paracancerous tissues) as follows: GSE151181 (15) from the GPL23159 platform ([Clariom_S_Human] Affymetrix Clariom S Assay, Human; includes Pico Assay), GSE6004 (16) and GSE60542 (17) from the GPL570 platform [(HG-U133_Plus_2) Affymetrix Human Genome U133 Plus 2.0 Array]. In total, 100 samples met these requirements, consisting of 53 metastatic PTC and 47 normal thyroid tissue samples.
## Identification of degs in metastasized ptc
The GEO2R online tool identified 462, 1392, and 2026 DEGs from GSE151181 (15), GSE6004 (16), and GSE60542 (17), respectively. A |logFC| > 2 and P < 0.05 were used as the condition to filter significant DEGs, and subsequently, Venn diagram online tools were used to determine the intersection of the three datasets to obtain 23 DEGs, including 8 upregulated and 15 downregulated genes.
## Verification of degs
The 23 identified DEGs were subjected to GEPIA for the verification of differential expression between TC and normal control samples, using a large sample size. Statistically significant differences in expression were obtained for 23 DEGs based on the large sample set (P < 0.05, [fig_ref] TABLE 1: Twenty-three differentially expressed genes between normal and thyroid cancer [/fig_ref] ; the 8 upregulated and 15 downregulated genes were confirmed.
## Analysis of degs associated with overall survival
The KM-plotter was used to analyze the association between OS and the 23 DEGs. Eighteen DEGs were associated with reduced OS in patients with TC (P < 0.05, and, whereas for five the association was not significant (P > 0.05, .
## Correlation analysis of degs
The 18 DEGs that affected the prognosis of patients might be risk factors for PTC metastasis. Correlation analysis was performed for this risk-associated set using GEPIA, and the results were visualized using the R software [fig_ref] FIGURE 3: Pairwise correlation of differential expressed genes [/fig_ref]. A Pearson correlation coefficient absolute value < 0.4 (very weak correlation) was scored as 0, 0.4-0.6 (weak correlation) was scored as 1 point, 0.6-0.8 (strong correlation) was scored as 2 points, 0.8-1.0 (very strong correlation) was scored as 3 points, and the highest cumulative score was considered the core risk-associated gene. The Pearson correlation coefficient cumulative scores of the 18 DEGs are ranked as follows: (2), TENM1 (1), TFF3 (1). The correlation between HGD, TPO, SLC26A4 and DIO1 is shown in [fig_ref] FIGURE 3: Pairwise correlation of differential expressed genes [/fig_ref].
## Rt-qpcr verification of risk gene set expression in clinical samples
Based on eight patients with metastatic PTC, cancerous and normal adjacent tissues were collected. Expression of the riskassociated gene set was verified using RT-qPCR, and expression of the top four core risk-associated genes was verified using immunohistochemistry. In the pathological tissues from patients, 14 genes were differentially expressed in metastatic PTC, with statistically significant differences (P < 0.05). Primer sequences are shown in [fig_ref] TABLE 1: Twenty-three differentially expressed genes between normal and thyroid cancer [/fig_ref] , and RT-qPCR results are shown in .
## Analysis of degs associated with relapse-free survival
Among the verified 14 genes with differential expression in the clinical samples, 9 DEGs significantly affected the RFS (P < 0.05,of patients with metastatic PTC, while the influence of 5 DEGs was not significant (P > 0.05).
## Ihc verification of core risk gene set expression in clinical samples
We analyzed DIO1, HGD, SLC26A4 and TPO protein expression with IHC staining in metastatic PTC and its adjacent tissues. DIO1, HGD, SLC26A4 and TPO were significantly overexpressed in paracancerous tissues . The judgment was based on the mean optical density (MOD) after Image-Pro Plus 6.0 analysis.
# Discussion
Although the 5-year survival rate of patients with PTC is more than 90%, approximately 30% of patients have recurrent PTC. The prognosis of patients with recurrent PTC is poor and often accompanied by LN metastasis. Without early intervention, the 5-year survival rate of patients with distal metastatic PTC is only 53.3% (19). Unfortunately, there currently are no reliable biomarkers for PTC risk [bib_ref] ARHGAP36 regulates proliferation and migration in papillary thyroid carcinoma cells, Yan [/bib_ref] , and the risk in patients cannot be assessed (21). However, a high recurrence rate and aggressive PTC are usually accompanied by severe thyroid epitaxy, LN metastasis, or distant metastasis (21). To identify PTC patients at a high risk of metastasis and recurrence at an early stage, this study focused on patients with metastatic PTC. The gene expression differences between metastatic PTC and normal tissues were analyzed using bioinformatics methods, and expression verification based on a large sample size was conducted using GEPIA, The Cancer Genome Atlas (TCGA), and other databases. Survival analysis was conducted to identify the genes associated with a risk of PTC metastasis. Clinical samples were collected for secondary validation of the screening results, to identify biomarkers that can be used to assess the risk of PTC metastasis. In our study, 14 PTC metastasis risk-associated gene sets were screened, including four upregulated genes (CLDN1, LRP4, LRRK2, TENM1) and 10 downregulated genes (DIO1, DPP6, HGD, IPCEF1, MT1F, SLC26A4, SLC26A7, SPX, TFF3, TPO). Among all the risk-related genes, DIO1, HGD, SLC26A4 and TPO have the highest correlation with each DEG and significantly affect the RFS of patients with metastatic PTC. Comprehensive multi-data analysis showed that these four genes Differentially expressed genes that significantly affect overall survival in thyroid cancer patients. may be the key risk genes leading to PTC metastasis. Through further analysis of the core metastasis risk-associated genes, we found that the four DEGs (DIO1, HGD, SLC26A4, TPO) that highly correlate with each risk-associated gene are closely related to the transport and activation of iodine.
## B a
Further analysis of the relationship between iodine and PTC through literature search showed that abnormal iodine intake is a risk factor for PTC [bib_ref] Review of factors related to the thyroid cancer epidemic, Liu [/bib_ref]. However, existing studies have not fully clarified the relationship between iodine and PTC, though some studies have shown that PTC patients have higher iodine exposure [bib_ref] Thyroid cancer in an iodide rich area: A histopathological study, Williams [/bib_ref] [bib_ref] Association between iodine intake, thyroid function, and papillary thyroid cancer: A case-control..., Kim [/bib_ref] and that high iodine intake is a risk factor for PTC. Other studies [bib_ref] High urinary iodine, thyroid autoantibodies, and thyroid-stimulating hormone for papillary thyroid cancer..., Zhao [/bib_ref] [bib_ref] High iodine intake and central LN metastasis risk of papillary thyroid cancer, Zhao [/bib_ref] have found that high iodine levels are not significantly correlated with the risk of PTC and only affect the growth and metastasis of PTC. Patients with PTC usually show high levels of urinary [bib_ref] Association between iodine intake, thyroid function, and papillary thyroid cancer: A case-control..., Kim [/bib_ref] and serum (28) iodine.
Concomitantly, the level of thyroid-stimulating hormone (TSH) in PTC patients is higher than that in normal controls [bib_ref] High serum TSH level is associated with progression of papillary thyroid microcarcinoma..., Kim [/bib_ref]. Interestingly, iodine deficiency can lead to decreased thyroid hormone synthesis and secretion and thus increased TSH levels. This suggests that PTC patients have two seemingly contradictory pathological features occurring concomitantly, high iodine and high TSH levels, which might thus be predictors of poor prognosis for PTC [bib_ref] High serum TSH level is associated with progression of papillary thyroid microcarcinoma..., Kim [/bib_ref] [bib_ref] The association of thyrotropin and autoimmune thyroid disease in developing papillary thyroid..., Lee [/bib_ref] [bib_ref] Association of iodine and iron with thyroid function, Luo [/bib_ref].
To explore the causes of both high iodine and high TSH levels in patients with PTC, this study further analyzed the relationship between the identified core risk-associated genes for metastasis and PTC. A further analysis of the functions of downregulated core risk-associated genes in metastasized PTC revealed that DIO1 is a type 1 iodothyronine deiodinase that can convert T4 into T3, and T3 into T2. Notably, high free T4 is a B A To investigate the expression of metastasis risk genes in papillary thyroid carcinoma (PTC) and adjacent clinical samples by RT-qPCR. 14 of which were significantly different. The Relapse-free survival of differentially expressed genes in thyroid cancer after validation by clinical samples. risk factor for PTC [bib_ref] Association between iodine intake, thyroid function, and papillary thyroid cancer: A case-control..., Kim [/bib_ref]. The possible mechanism through which low DIO1 expression affects the occurrence and development of PTC is as follows: low DIO1 expression leads to failed T4 deiodization and a lack of free of T4 in the body. HGD participates in tyrosine metabolism. Tyrosine metabolism is related to thyroid hormone synthesis, and HGD expression in turn is highly correlated with TPO. The low expression of HGD might be involved in the occurrence and development of PTC by affecting TPO expression and thyroid hormone synthesis. SLC26A4, which encodes the pendrin protein located on the apical membrane of thyroid follicular cells, is a chloride/iodide ion transporter. SLC26A4 is a key protein that transports iodine into the follicle to synthesize thyroid hormones, and its low expression is an early event in thyroid tumorigenesis [bib_ref] LncRNA SLC26A4-AS1 suppresses the MRN complex-mediated DNA repair signaling and thyroid cancer..., Yuan [/bib_ref]. TPO encodes thyroid peroxidase, which is synthesized by thyroid follicular cells and is present in the plasma membrane at the top of the follicular epithelial cells. Thyroid peroxidase can oxidize iodine in the follicular cells to activate it, providing activated iodine for thyroid hormone synthesis. TPO plays an important role in maintaining normal thyroid functions. With low expression of TPO, normal thyroid function cannot be maintained, and thyroid dysfunction is closely related to TC. In this study, we found that genes associated with iodine transport (SLC26A4), activated iodine (TPO), thyroid hormone synthesis (HGD), and disassembled free T4 (DIO1) are expressed at low levels in PTC, and their expression is further reduced in metastatic PTC . The expression levels in normal, PTC, and metastatic PTC tissues are shown in Supplementary Table 2. In conclusion, iodine utilization in patients with metastatic PTC is blocked in many ways, including thyroid hormone synthesis. This indicates that patients with PTC are unable to utilize iodine in their body. Although serum iodine and urine iodine levels are elevated, thyroid hormone synthesis is still not possible, leading to elevated TSH levels. This study could explain the possible reasons for high serum iodine and TSH levels in PTC patients.
Previous studies have found that the level of iodine intake can affect thyroid function, but the relationship between iodine intake and TC is not clear [bib_ref] Review of factors related to the thyroid cancer epidemic, Liu [/bib_ref] ; moreover, some studies [bib_ref] Association of high iodine intake with the T1799A BRAF mutation in papillary..., Guan [/bib_ref] suggest that high iodine intake is an important risk factor for BRAF mutations and the subsequent development of PTC. Some studies [bib_ref] High iodine blocks a Notch/miR-19 loop activated by the BRAF(V600E) oncoprotein and..., Fuziwara [/bib_ref] suggest that high iodine levels play a protective role B A Expression of DIO1, HGD, SLC26A4, and TPO in metastatic papillary thyroid carcinoma (PTC) and adjacent normal tissues, and in different stages in the database. in thyroid cells and attenuate acute BRAF oncogene-mediated miRNA dysregulation. Of concern, one studyfound that the 10, 15 and 20-year survival rate of patients receiving iodine-131 is 56%, 45% and 40%. However, the 10 and 15-year survival rate of patients without iodine-131 treatment was found to be 10% and 6%. Iodine intake is closely related to the prognosis of TC, and the expression of SLC26A4 and TPO in iodine-resistant PTC is lower [bib_ref] The molecular and gene/miRNA expression profiles of radioiodine resistant papillary thyroid cancer, Colombo [/bib_ref]. This further supports our speculation that serum iodine content and level of iodine intake might not fully reflect the effect of iodine on PTC and that the ability of the body to use iodine could be important for the occurrence and development of PTC. The found providing a new treatment idea for patients with metastatic PTC who do not respond to iodine radiotherapy. For example, iodine absorption and utilization should be first improved for iodine-refractory PTC to ensure the efficacy of radiotherapy.
Although this study comprised a large-sample size analysis using GEO, GEPIA, TCGA, and other databases, in addition to utilizing pathological wax blocks from patients with metastatic PTC for RT-qPCR and immunohistochemistry verification, this was only a preliminary screen of a risk-associated gene set that might lead to PTC metastasis. As such, this set has not been quantitatively studied, and an evaluation of the risk of PTC metastasis based on the expression level of this gene set requires more clinical samples and quantitative studies based on each stage-based subgroup. A possible mechanism by which DIO1, HGD, SLC26A4, and TPO participate in the development of TC is shown in .
Conclusions CLDN1, LRP4, LRRK2, TENM1, DIO1, DPP6, HGD, IPCEF1, MT1F, SLC26A4, SLC26A7, SPX, TFF3, and TPO affect the prognosis of patients with PTC and are riskassociated genes for PTC metastasis. Further, DIO1, HGD, SLC26A4, and TPO are potential core risk-associated genes. Our study will help assess the risk of metastasis in patients with PTC, so that patients at high risk of metastasis can be followed up early and closely.
# 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 in the article/Supplementary Material.
# Ethics statement
The studies involving human participants were reviewed and approved by Biomedical Research Ethics Committee of Zunyi Medical University Hospital. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. Author contributions QZ designed the study and edited the manuscript. JL reviewed and revised the manuscript, HS proofread the manuscript. XB and TZ systemically revised the manuscript for important content. PL proposed the concept and designed the structure of the study. All authors contributed to the article and approved the submitted version.
[fig] 23: differentially expressed genes (DEGs) in metastatic PTC and non-cancer tissues. (A) Common upregulated intersection DEGs obtained from Gene Expression Omnibus database: GSE15181, GSE6004 and GSE60542. (B) Common downregulated intersection DEGs obtained Gene Expression Omnibus database: GSE15181, GSE6004 and GSE60542. (C) Upregulated DEGs expression in thyroid cancer (TC) patients and normal controls. (D) Downregulated DEGs expression in TC patients and normal controls. Red: Cancer, black: Normal, *P< 0.05 vs. normal group. [/fig]
[fig] 2: Kaplan-Meier-Plotter analysis of the association between prognosis and 23 differentially expressed genes. Category Genes Significant impact on survival (P < 0.05) CLDN1, GABRB2, LRP4, LRRK2, SLC34A2, TENM1, DIO1, DPP6, HGD, IPCEF1, MT1F, PKHD1L1, SLC26A4, SLC26A7, SPX, TFCP2L1, TFF3, TPO Without significant impact on survival (P > 0.05) CRABP1, FAM20A, FN1, MPPED2, PLA2R1 [/fig]
[fig] FIGURE 3: Pairwise correlation of differential expressed genes (DEGs) significantly affecting overall survival of TC patients in the context of TC disease. (A) R Software visualization of pairwise correlations of DEGs that significantly affected overall survival. (B) The correlation figure of DIO1 and TPO, DIO1 and SLC26A4, DIO1 and HGD, HGD and TPO, HGD and SLC26A4, TPO and SLC26A4. [/fig]
[fig] Funding: This work was supported by the Program for Excellent Young Talents of Zunyi Medical University (No.18zy-006) and the Basic Research Program of Guizhou Provincial Department of Science and Technology (Natural Science) [Grant No. qiankehejichu-ZK [2022]yiban599]. [/fig]
[table] TABLE 1: Twenty-three differentially expressed genes between normal and thyroid cancer (TC) samples. [/table]
|
Identifying influenza-like illness presentation from unstructured general practice clinical narrative using a text classifier rule-based expert system versus a clinical expert
Background:We designed and validated a rule-based expert system to identify influenza like illness (ILI) from routinely recorded general practice clinical narrative to aid a larger retrospective research study into the impact of the 2009 influenza pandemic in New Zealand.
Results: We calculated a 98.2 % specificity and 90.2 % sensitivity across an ILI incidence of 12.4 % measured against clinical expert classification. Peak problem list identification of ILI by clinical coding in any month was 9.2 % of all detected ILI presentations. Our system addressed an unusual problem domain for clinical narrative classification; using notational, unstructured, clinician entered information in a community care setting. It performed well compared with other approaches and domains. It has potential applications in real-time surveillance of disease, and in assisted problem list coding for clinicians. Conclusions: Our system identified ILI presentation with sufficient accuracy for use at a population level in the wider research study. The peak coding of 9.2 % illustrated the need for automated coding of unstructured narrative in our study.
# Background
Influenza and influenza like illness (ILI) are very important clinical conditions responsible for a high burden of global mortality and morbidity. Identification of the emergence and surveillance of these flu like syndromes, was and is, important [bib_ref] Viral etiology and seasonality of influenza-like illness in Gabon, Lekana-Douki [/bib_ref] [bib_ref] Implementation of influenza-like illness sentinel surveillance in Togo, Maman [/bib_ref] [bib_ref] Pandemic influenza A(H1N1)v in New Zealand: the experience from, Baker [/bib_ref] [bib_ref] Interim Analysis of Pandemic Influenza (H1N1) 2009 In Australia: Surveillance trends, age..., Kelly [/bib_ref] in order to implement appropriate health service response at both general practice and secondary hospital levels. Pandemic H1N1 influenza was detected in New Zealand in April 2009, and within two months spread widely across the country. In New Zealand as in many other OECD countries General Practice and Primary care are responsible for the delivery of over 90 % of first line medical care, but are operated from small semi-autonomous units without a high degree of central control over clinical information recording and standardised coding systems. Community based services used a variety of response models. A research study was devised to acquire data on the pandemic and investigate the capacity for general practice to respond to the high demand for health care during the period.
Identification of patients with a particular condition is usually done using problem lists (disease classification data) or prescribed medicine lists [bib_ref] Validity of diagnostic coding within the General Practice Research Database: a systematic..., Khan [/bib_ref] [bib_ref] Clinical information for research; the use of general practice databases, Lawrenson [/bib_ref]. These lists are generated by clinicians during or shortly after a clinical encounter using the Read code classification system. Since the use of coded problem lists is highly variable between clinicians and between conditions [bib_ref] Quality of morbidity coding in general practice computerized medical records: a systematic..., Jordan [/bib_ref] , and we had no empirical evidence to suggest how comprehensively general practice had been classifying ILI, we could not rely solely on using this data set. We also suspected that prescribing of antiviral drugs (e.g., oseltamivir) would be an unreliable proxy, as it was not indicated in those solely presenting with ILI and some prescribing of it may have been prophylactic. Using software to classify clinical narrative has been postulated as a solution to such problems [bib_ref] Natural language processing and its future in medicine, Friedman [/bib_ref] [bib_ref] Emergency medical text classifier: new system improves processing and classification of triage..., Haas [/bib_ref] [bib_ref] Extracting information on pneumonia in infants using natrual language processing of radiology..., Mendonca [/bib_ref] [bib_ref] Automatic classification of diseases from free-text death certificates for realtime surveillance, Koopman [/bib_ref].
Our objective was to use automated software to identify ILI by classifying the unstructured clinical narrative written by physicians in community based care facilities. This approach has previously been successful in structured free text reports in an emergency department or hospital setting [bib_ref] Emergency medical text classifier: new system improves processing and classification of triage..., Haas [/bib_ref] [bib_ref] Extracting information on pneumonia in infants using natrual language processing of radiology..., Mendonca [/bib_ref] [bib_ref] Classifying free-text triage chief complaints into syndromic categories with natural language processing, Chapman [/bib_ref] [bib_ref] Identifying Respiratory Findings in Emergency Department Reports for Bio surveillance using MetaMap, Chapman [/bib_ref] [bib_ref] A comparison of classification algorithms to automatically identify chest X-ray reports that..., Chapman [/bib_ref] [bib_ref] ConText: an algorithm for identifying contextual features from clinical text, Chapman [/bib_ref] [bib_ref] A controlled trial of automated classification or negation from clinical notes, Elkin [/bib_ref] [bib_ref] Unlocking clinical data from narrative reports: a study of natural language processing, Hripcsak [/bib_ref] [bib_ref] A Study of Abbreviations in Clinical Notes, Xu [/bib_ref] [bib_ref] Defining Emergency Department Asthma Visits for Public Health Surveillance, Travers [/bib_ref] [bib_ref] Optimizing A syndromic surveillance text classifier for influenza-like illness: Does document source..., South [/bib_ref]. Little work has been done on information retrieval from community based care facilities and only a few papers have addressed searching narrative containing bespoke abbreviations or highly notational information [bib_ref] Limited Parsing of Notational Text Visit Notes: Ad-hoc vs. NLP Approaches, Barrows [/bib_ref]. Clinical narrative is traditionally difficult to process because it may often contain ambiguities, word compounds and a substantial body of synonyms to describe otherwise basic concepts. Because the clinical narrative in general practice is entered by a clinician using a keyboard during or shortly after a patient encounter it often contains typographical errors making the automation more difficult.
The aim of this paper is to describe the development and testing of a rule based expert-system to identify the presentation of ILI within general practice from routinely recorded clinical narrative. In doing so we outline a general approach towards algorithm development using text classifier processing that could be used in other clinical conditions where clinical coding is not universal.
# Methods
## System architecture
We developed a clinical algorithm in conjunction with clinical experts based on each encounter's narrative, problem lists and patient's age. Its purpose was to classify the encounter into one of two states; the patient had influenza like illness or not. To aid training and understanding of the problem in more detail the negative state was sub-classified into three groups; influenza prophylactic; respiratory; and other. The detail of this can be seen in . Influenza prophylactic presentations were those where procedures were undertaken (e.g., flu vaccine) or medicines prescribed (e.g., antiviral prescription for travelling abroad) relating to influenza, but the patient did not present with an ILI. The negative state subclassifications were not used in the final study analysis.
The software inference algorithm [fig_ref] Figure 2: Software ILI algorithm schematic [/fig_ref] is a more detailed implementation of the clinical algorithm Each lexical set was coded into a heuristic dictionary using regular expressions, to account for abbreviations, spelling errors, synonyms and negations. Regular expressions were used to modify full and correct forms of words to make portions of words optional to match abbreviations, or to substitute similar phonetic characters to improve detection of misspelled words. For example the term "fever" used a regular expression of "fe?ve?r(([iey]sh)|y|s?)" to detect words stemmed from "fever" including "fevers", "feverish" and "fevery" but to also account for abbreviations to a word stem of "fver", "fevr" or "fvr" as well as misspellings such as "feverysh". In this manner the algorithm sensitivity was increased by including more liberal character substitution expressions, while the specificity was decreased and vice versa. Matches to negation terms were controlled through the use of regular expression patterns, to both increase and decrease the appropriate word distances for particular terms which affected the measures of accuracy of the algorithm. The lexicon also included the concept of measurement-value pairs (e.g., "temp. 38.2" is analogous with a symptomatic description of pyrexia) used in other tools such as TOPAZ [bib_ref] Identifying Respiratory Findings in Emergency Department Reports for Bio surveillance using MetaMap, Chapman [/bib_ref]. Sentence detection was used to limit the effect of negations and temporal references to the scope of each text chunk (clinical narratives are not always formed into proper grammatical structures in this case a sentence is more accurately described as related narrative). In this manner we applied a pipeline of rudimentary operations which included word stemming, sentence detection and chunking. Each square box in the flow diagram in [fig_ref] Figure 2: Software ILI algorithm schematic [/fig_ref] represents an NLP processing step; each step used a separate lexical set and parameters to achieve its goal. [fig_ref] Table 1: Lexical sets [/fig_ref] shows how the lexical sets are used within the algorithm. The number of separate expressions is shown as the expression count. Some terms are codified into multiple regular expressions depending on the complexity of the way in which those terms may be expressed. The purpose of providing a count in this manner is to provide some proxy indicator to the relative complexity of detection of terms within each set. In general negation based expressions had more terms, were longer and contained more complexity. This occurred because of the increased need to account for the relationship between pertinent expressions and negations. Examples of terms used within the lexical sets are shown in [fig_ref] Table 2: Examples of terms used within lexical setsThe goal of each training round... [/fig_ref].
The tool applied two classes of logic for heuristic patterns. The first used logical disjunction to test for the existence of any appropriate expression. The disjunction was applied in a manner where the narrative was assessed against all terms in the lexical set with the presence of one or more terms resulted in a logical "true" output. The second used weights based on the proportion of pertinent expressions compared with negations. Each term present within the narrative for each lexical set was counted and contributed to the total weight of that lexical set for
## 30
Definitive Negatives Expressions that relate to procedures associated with influenza, particularly prophylaxis but do not indicate ILI is currently present.
10
## Negative pyrexia expressions
Expression that when either negated, or when present suggest a temperature that is not elevated.
## 16
## Negative respiratory expressions
Expressions that relate to signs or symptoms that when negated decrease likelihood of respiratory disease.
## 28
## Negative symptom expressions
Expressions that when present suggest an elevated temperature.
## 10
## Positive pyrexia expressions
Expressions that related to signs or symptoms that may be associated with respiratory disease but not necessarily ILI.
## 11
## Positive respiratory expressions
Expressions that related to signs or symptoms that when present, increase the likelihood that ILI is present.
## 36
## Positive symptom expressions
Expressions which related to signs or symptoms that when present increase the likelihood of ILI.
## Ili prophylactic expressions
Expressions that relate to influenza and contribute to workload in general practice but where there is no active disease process present in an individual. These expressions include those that related to the seeking of prophylactic measures for influenza.
comparison against other lexical sets. This allowed for decisions within the algorithm to be made on the balance of the evidence. The algorithm first identified if the patient had a respiratory condition by detecting an appropriate assigned Read code and then parsed the clinical narrative searching for expressions associated with respiratory illness and symptoms. If a patient was not coded with the more specific Influenza Read code (H27) the algorithm branched dependent on the patient's age. The lexicons used differed for children and adults even in identifying the same types of symptoms. Any phrases that explicitly negated the presence of influenza immediately caused the algorithm to code the presentation as being "other respiratory" with no further checks. We used asymmetric word distance to limit the scope of negative and temporal phrases to pertinent expressions. Negative and temporal phrase scopes were limited to sentence chunks where sentences were detected through punctuation or end-of-line characters. Negation phrases included "no", "not", "nil", "NAD" and permutations of these keywords. Regular expressions were formed to ensure that negation phrases were not detected within substrings of larger words. Temporal references included simple matches to keywords including "yesterday", "prior", "last time", as well as detection of phrases of word pairs indicating a time period ("days","months") matched with a reference to the past ("ago", "previously"). For those older than five, checks were made for expressions that indicated a diagnosis or suspected influenza, presence of pyrexia and general symptoms associated with viral infection. For those younger than five years only a check was made for expressions related directly to influenza and fewer assumptions were made on more general symptomatic presentation. Separate lexical sets were used for some detection of child symptoms. The final step of the algorithm was designed to eliminate those that present for advice or prophylactic medicine only.
How the algorithm parses a specific passage of text is illustrated in . It also provides an example of the typical format of a clinical narrative including descriptions of signs and symptoms and some of the abbreviations and typographical errors that can occur. It detects the presence of respiratory symptoms, does not find any statements explicitly ruling out ILI, finds observations indicating ILI, ascertains the patient has a high temperature, detects specific respiratory symptoms related to ILI and finally finds no evidence that the visit was only for discussion of ILI or for prophylactic treatment.
## Participants
All ninety-nine practices that were part of the primary health organisation (PHO) that collaborated on the research project were invited to participate in the study. Seventy-two practices agreed. The study data set included all patient encounters occurring from May to October, from 2007 to 2010 inclusive from participating practices. A total of 5.2 million routine encounter narrative records were included in the final study data set with associated problem lists, medications and disease screening activity information. Data was collected directly from patient management systems using standardised extraction scripts. No identifiable information was collected about any patient except the National Health Index unique identifier and all data used for the study was handled by the primary health organisation which routinely handles such data sets for contractual reporting purposes. The research team received no identifiable patient data. Proprietary software developed by the PHO was used to automate the extraction, packaging and secure transmission of the large data sets to the research team.
Demographics for patients funded in the participating practices over the study period were representative of all practices in the region. There appeared to be no substantial demographic factors that would be likely to contribute to any confounding within our results.
## Training
The algorithm was trained with a bias to promote specificity over sensitivity. Ten rounds of training were completed by a non-clinical analyst in conjunction with clinical experts. The training set consisted of six practices selected at random from all practices providing data for the research. One health care provider within each of these was selected at random to provide distinct sets of narrative. In total, 623 encounter narratives were used as the training set.
The training set was initially assessed by a clinical expert, following the agreed clinical algorithm. The assessor had access to the same data that the software algorithm could use, which included problem list data Software ILI algorithm applied to an example clinical narrative experience consistent with international evidence [bib_ref] Did you have flu last week?" A telephone survey to estimate a..., Payne [/bib_ref] [bib_ref] Incidence of 2009 pandemic influenza A H1N1 infection in England: a crosssectional..., Miller [/bib_ref] we wished to keep the total number of false positive results to a minimum. The analyst compared the results of the software algorithm with the clinical assessor for each round. Results from each training round were analysed to identify portions of the heuristics and domain lexicon that were performing poorly. Training consisted of modifying the domain lexicon in conjunction with the clinical expert and modifying the pattern dictionaries to account for errors involving contextual discrimination, including temporal discrimination, finding validation and contextual inference as described by Chapman et al. [bib_ref] Identifying Respiratory Findings in Emergency Department Reports for Bio surveillance using MetaMap, Chapman [/bib_ref].
## Testing
Eight practices were identified at random from the 72 providing data for the research. One provider in each of the eight practices was identified at random, and their clinical notes over a one week period were used as the test set. The six practices and providers whose notes were used as the training set were explicitly excluded from selection to eliminate any training bias. In total, 901 encounter narratives made up the test set.
The content of the test set was never used to inform the development or training of the clinical or software algorithm. It was used only once to determine final performance metrics and measures of accuracy.
## Gold standard
We considered the Gold standard to be the judgment of a clinical expert determining the presence of ILI on reading each clinical narrative. This was used to assess the performance of the algorithm during both training and test stages. A single clinical expert read each note, assessed the associated Read code information and classified each consult into an appropriate category using the algorithm as shown in . The expert was blinded from all results of the software algorithm.
## Ethics approval
This study was approved by the New Zealand Health & Disability Multi-region Ethics Committee under application number MEC/10/14/EXP.
# Results
The results of the ten rounds of training of the algorithm can be seen in [fig_ref] Figure 4: Algorithm sensitivity and specificity by training round The performance of the final... [/fig_ref]. While sensitivity was marginally higher during rounds 3 and 4 of training, the corresponding specificity at this point was not as high. The results of the tenth training round and final results of the test set can be seen in [fig_ref] Table 3: Software algorithm performance versus gold standard in test and tenth-round training set [/fig_ref]. The algorithm performed similarly against the test set as the training set.
Binomial confidence limits were calculated using the method described by Collett et al.. When applied across all practices, the system detected substantially more presentations of ILI than were coded by clinicians. The monthly clinical coding rate of ILI by provider ranged from 0.3 to 9.2 % of the ILI encounters detected by the system. This represents a range of 8.9 %, showing there is little consistency in the rate clinicians code such encounters.
# Discussion
The rate at which ILI was coded in the problem list by GPs in their routine consultations was low. The peak coding rate for ILI in any month by any individual clinician was 9.2 % of that detected by the algorithm and meant that had we relied on clinical coding, our principal study would have significantly under reported the presentation of ILI. The range of individual clinician coding rate compared to detected ILI was 8.9 % and was so great that it would have been difficult to build a model based on observed problem list coding alone. Manually reviewing all five million records would have been impractical. The approach to use automated software to identify appropriate presentations therefore seems appropriate.
# Strengths and weaknesses
The results we have achieved are favourable in comparison to other studies. The specificity of the software matches the best seen in literature for classifying medical text [bib_ref] Classifying free-text triage chief complaints into syndromic categories with natural language processing, Chapman [/bib_ref] [bib_ref] Classification Algorithms Applied to Narrative Reports, Wilcox [/bib_ref]. A recent rule-based approach to classifying influenza as a cause of death from death certificates reported higher PPV, substantially lower specificity and a slightly higher F-measure [bib_ref] Automatic classification of diseases from free-text death certificates for realtime surveillance, Koopman [/bib_ref]. As a comparison, our task was likely more difficult due to the nature of clinician entered clinical narrative containing abbreviations, acronyms and typographical errors without well-formed grammar or structure. We had to consider the presentation of signs and symptoms beyond the mention of diagnoses. The rules we applied extended to those that addressed clinical reasoning and placing signs and symptoms into the context of a diagnosis without the necessary mention of the diagnosis itself. For these reasons we believe achieving similar measures of accuracy is a favourable result for this specific problem domain.
Training of the algorithm was focused on maximising specificity to minimise type II errors in our calculations to be conservative in our analysis of disease presentation. There were two keys to keeping our specificity high; the detection of expressions that unambiguously discounted ILI and those that negated otherwise pertinent findings. There is an existing body of literature on the latter from which we could draw [bib_ref] ConText: an algorithm for identifying contextual features from clinical text, Chapman [/bib_ref] [bib_ref] A simple algorithm for identifying negated finds and disease in discharge summaries, Chapman [/bib_ref] [bib_ref] Evaluation of Negation Phrases in Narrative Clinical Reports, Chapman [/bib_ref]. Our system's sensitivity is not as high as seen in other literature [bib_ref] Classifying free-text triage chief complaints into syndromic categories with natural language processing, Chapman [/bib_ref] [bib_ref] Limited Parsing of Notational Text Visit Notes: Ad-hoc vs. NLP Approaches, Barrows [/bib_ref] , which suggests that more attention may need to be paid to identifying additional pertinent expressions or detecting the context in which they are used to include rather than negate them. Of course there is an inevitable trade-off between sensitivity and specificity and the balance within our algorithm may already be optimal. Our test method appears to be rigorous in the number of records we used, with narrow confidence intervals for the key parameters. A few prior studies use similarly large test datasets [bib_ref] Classifying free-text triage chief complaints into syndromic categories with natural language processing, Chapman [/bib_ref] [bib_ref] Evaluation of Negation Phrases in Narrative Clinical Reports, Chapman [/bib_ref] , with one using a significantly larger test set [bib_ref] Optimizing A syndromic surveillance text classifier for influenza-like illness: Does document source..., South [/bib_ref] although this was not entirely classified by clinicians.
The gold standard here of clinician diagnosed ILI has limitations. The ILI case definition has evolved over time, with the current WHO surveillance case definition being an acute respiratory illness with onset during the last 10 days with history of fever or measured fever of ≥ 38.0C and cough. Because this study was retrospective, it was not able to collect a complete set of data on each encounter sufficient to establish that the case did meet a standard ILI definition.
We used only one clinician to classify the test data. A previous study used overlapping sets across three clinicians, finding only 1 un-agreed result from 200 records [bib_ref] A simple algorithm for identifying negated finds and disease in discharge summaries, Chapman [/bib_ref]. Another using four clinicians disagreed on 12 from 200 records [bib_ref] Evaluation of Negation Phrases in Narrative Clinical Reports, Chapman [/bib_ref]. It is difficult to find clinicians with sufficient clinical experience, an appropriate analytical skill set and willingness to undertake such a tedious task. Given this and that little interrater disagreement has been previously found we believe that using a single clinician is appropriate as a 'gold standard' in this instance. We were conservative by using a high number of records for testing, and by minimising the introduction of bias by ensuring that the training sets were distinct from the test sets, and that test data were derived from a number of different clinicians.
Our results are satisfying considering the notational, varied and unstructured nature of the narrative we were attempting to classify. Based on our experience in this project of dealing with these narratives, the reported rate of 2 % spelling errors in medical notes [bib_ref] Evaluating and reducing the effect of data corruption when applying bag of..., Ruch [/bib_ref] seems to be an underestimation for general practice in New Zealand. In general practice, clinicians are responsible for entering notes themselves, and this relies not only on their spelling ability and keyboard skills but also on their propensity to identify and correct such errors. Using heuristic pattern matching has allowed us to compensate for this potential source of error. Traditional Natural Language Processing does not cope well with notational narrative and requires abbreviation and notational expansion as a pre-analysis operation [bib_ref] Limited Parsing of Notational Text Visit Notes: Ad-hoc vs. NLP Approaches, Barrows [/bib_ref]. Although a large body of literature uses this approach it would likely have required a much larger computational demand and may have not been as successful. Natural Language Processing tends to be used to classify a wider range of presentations or diagnoses and operates on more formal medical texts, such as x-ray reports or discharge summaries that are often dictated and transcribed by medical typists, which would increase sentence structures and decrease abbreviations and spelling errors.
The problem domain we have addressed is unusual in focusing on community care facility data that is highly notational. One study has approached the classification of clinical narrative using clinician entered, notational data in an ophthalmology clinic [bib_ref] Limited Parsing of Notational Text Visit Notes: Ad-hoc vs. NLP Approaches, Barrows [/bib_ref] , while another [bib_ref] Optimizing A syndromic surveillance text classifier for influenza-like illness: Does document source..., South [/bib_ref] used clinician entered data in an outpatient clinic but did not comment on the data structure or notational nature. Little literature uses routine general practice data for this purpose. The nature of general practice data is that there is a large variety in the presenting case-mix, with multiple problems arising in the course of one consultation making it difficult to use concepts such as primary reason for presenting. Although there are a multitude of systems used for academic purposes, very few are used in an operational medical environment and integrated with a clinical information system [bib_ref] What can natural language processing do for clinical decision support?, Demner-Fushman [/bib_ref]. Our system is integrated with existing practice management systems commonly used in New Zealand general practice.
The use of regular expressions to provide a lightweight method of accounting for abbreviations and typographical errors may explain some of our success. Using regular expressions to account for common errors in words, particularly in the use of vowels and repeated consonants is easy and highly efficient. Breaking the NLP problem down into discrete steps, with separate lexicons and parameters appears to have simplified the task and allowed us to alter discrete parts of the algorithm without influencing others.
# Error analysis
Although our approach performed well in general there were some situations that proved difficult for our algorithm to address. The use of temporal modifiers referring to signs, symptoms and diagnoses were not detected. In most cases a reference to a past sign or symptom was not relevant to the current presentation. This is particularly relevant for acute disease such as those that present as ILI. In longer term disease past signs and symptoms may be more useful in determining a final diagnosis. Clinical narratives at times contained references to signs, symptoms and diseases belonging to the subject's relatives, such a parent or sibling. This was more often associated with clinical narratives for children. The algorithm had difficulty distinguishing between signs or symptoms reported about the subject by a relative, or whether the signs and symptoms were those reported by the subject about a relative with the later contributing to type I errors.
The algorithm detected the negation of pertinent terms but was confused by particularly wordy or convoluted negation phrases e.g., "No obvious or real cough" or "No cough in the past few weeks but has started in the last day". It is likely that an improved part-of-speech tagging approach may address these limitations.
## Future work
Current influenza surveillance is carried out in general practice in New Zealand by paper based manual systems or on manual problem list coding by clinicians. Paper based systems can only be deployed to a limited number of practices, and suffer the same issue of relying on clinicians to remember to report cases of interest. Paper based systems suffer from a latency caused by the delays in completing paper work, transmission and collation. Existing electronic systems report on items recorded within problem lists. Our results show that at best, this type of reporting detects a small proportion of real cases, and the range of the rate makes it difficult to model and extrapolate such figures from manual recording alone.
The early detection of disease outbreaks can be an important factor in the ability of public health services to successfully intervene [bib_ref] The emerging science of very early detection of disease outbreaks, Wagner [/bib_ref] [bib_ref] Detection of epidemics in their early stage through inectious disease surveillance, Hashimoto [/bib_ref]. Text classifying expert systems have the potential to be used for large scale surveillance of presentation of ILI to general practice, or to other care settings where medical narrative is available electronically with a high resolution on a daily basis. Situating such an automated alerting or reporting system within general practice offers minimal latency of outbreak detection. The research team is currently developing this concept further.
This method could be scaled to be used across an even larger set of patient records in near-real time. Our approach is not computationally onerous and could be implemented within existing practice management systems, running on modest desktop computing equipment to aid in an increased rate of problem list coding. We ensured that the training and test records sets contained no cross-over between clinicians or facilities contributing to each which provides some demonstration that the algorithm is generally applicable beyond the records which were used to develop and train it. No single clinician or facility contributed more than 7 % of overall records for training or test purposes. Because of the small number of medical training facilities in New Zealand (two medical schools and a single vocational training organisation) we expect the variation between clinicians in the use of pertinent expressions and key semantics of note taking to be represented in the study group. Currently clinicians must manually find and select appropriate problem list codes to classify encounters within the mainstream patient management systems in general practice in New Zealand. This process contributes to the highly variable rate of coding across clinicians and conditions. Providing feedback, incentives and evidence based guidelines have been shown to improve data quality in clinical coding [bib_ref] The use of routinely collected computer data for research in primary care:..., De Lusignan [/bib_ref] and national and regional initiatives have done just this for long term conditions, but with little emphasis on the classification of acute presenting problems. A rule based approach could process narrative notes in real time, and suggest items of interest for coding with high accuracy. This approach has the potential to improve the coding of less significant or short-term problems [bib_ref] Submit your next manuscript to BioMed Central and take full advantage of:..., Meystre [/bib_ref]. Maintaining accurate and comprehensive problem lists can be useful to general practices who wish to keep a complete health record for their patients and manage them using this data. The limitation of this approach is that the algorithm development is based on specific target conditions each of which would have their own particular clinical presentations.
We could improve our approach by having the software output and store meta-data about the performance of each narrative and how it is branched through the algorithm to reach a final classification. This would likely help with training the algorithm and understanding how particular sub-components of the algorithm are performing which in turn would lead to better final performance.
# Conclusions
This paper describes the successful use of a heuristic rule-based expert system for identifying presentations of ILI in general practice from routine clinical narrative. The system performed to a standard sufficient to use for population based research similar to the best performing systems documented for other similar problems.
This research demonstrates that a text classifier can be applied successfully to a clinical narrative that is highly abbreviated and contains substantial spelling and typographical errors. General practice captures a large volume of medical information in narrative form on patients and the application of the techniques described can unlock this previously difficult to access information source. Furthermore this type of approach is computationally simple enough to apply to millions of records or in real time in general practice surgeries.
We have discovered that the clinical coding rate for ILI over the period of the study was very low, necessitating the use of such a tool for our specific research study. These results highlight an important finding for researchers that the routine coding of acute conditions such as ILI within general practice may be substantially lower than the actual presentation. The system also has potential to be used for automated on-going disease surveillance.
# Funding
The study was funded by grant 10/760 from the New Zealand Health Research Council.
[fig] Figure 2: Software ILI algorithm schematic [/fig]
[fig] Figure 4: Algorithm sensitivity and specificity by training round The performance of the final algorithm was pleasing with a mean classification rate of 7.3 clinical narratives per second. The algorithm was run on a physical server with a Dual Intel® Xeon® X5650 CPU running at 2.67 GHz with 24 Gb of RAM running Microsoft SQL Server 2008 R2. [/fig]
[table] Table 1: Lexical sets [/table]
[table] Table 2: Examples of terms used within lexical setsThe goal of each training round was to improve the algorithm specificity, while maintaining or improving sensitivity. Because we predicted a relatively low incidence of ILI presentations based on local clinical S) just back from [REDACTED], sore throat, temp., runny nose. strated Sunday, flew Staurday, no diarrhoea no rash. O) T 38.5 no specific glands, no allergies, no rash. Chin to chest easil adbo soft, dry cough A) sounds tracheal tracheitis not on oc/ not preg. [/table]
[table] Table 3: Software algorithm performance versus gold standard in test and tenth-round training set [/table]
|
Growth and Development in Chinese Pre-Schoolers with Picky Eating Behaviour: A Cross-Sectional Study
ObjectiveTo explore the associations between picky eating behaviour and pre-schoolers' growth and development. Corresponding potential mechanisms, such as nutrient and food subgroup intake, as well as micronutrients in the blood, will be considered.MethodsPicky eating behaviour was present if it was reported by parents. From various areas of China, 937 healthy children of 3-7 years old were recruited using a multi-stage stratified cluster sampling method. Children and their mothers' socio-demographic information and children's anthropometry, intelligence, blood samples, one 24-hour dietary intake record and food frequency questionnaire were collected. Z-scores and intelligence tests were used to evaluate growth and development (cognitive development). Multilevel models were employed to verify the associations between picky eating behaviour and growth and development.
# Results
The prevalence of picky eating as reported by parents was 54% in pre-schoolers. Compared with the non-picky eaters, weight for age in picky eaters was 0.14 z-score (95% CI: -0.25, -0.02; p = 0.017) lower while no significant difference was found in intelligence (p > 0.05). Picky eating behaviour lasting over two years was associated with lower weight for age, as was nit-picking meat (the prevalence from parents' perception was 23% in picky eaters) (p < 0.05). Picky eaters consumed fewer cereals, vegetables, and fish (p < 0.05), and had a lower dietary intake of protein, dietary fibre, iron, and zinc (p < 0.05). There were no differences in the concentrations of essential minerals in whole blood (p > 0.05).
# Introduction
During the time of scarcity in China, social resources and family life were devoted to farming, securing, procuring and preparing food, which was limited in variety, had poor palatability, and low contents of energy and nutrients. With the rapid increase of comparative wealth and general affordability of food, parents are paying more attention to the diversity and composition of foods in children's diets [bib_ref] Food neophobia and 'picky/fussy' eating in children: a review, Dovey [/bib_ref] [bib_ref] Does picky eating affect weight-for-length measurements in young children?, Ekstein [/bib_ref]. Picky eating is often characterised as eating limited amounts of food, rejecting some specific subgroups of food, being unwilling to try new foods, and having strong preferences for certain cooking methods regarding the smell and taste of food [bib_ref] Food neophobia and 'picky/fussy' eating in children: a review, Dovey [/bib_ref] [bib_ref] Problem eating behaviors related to social factors and body weight in preschool..., Dubois [/bib_ref]. Picky eating results in limited and imbalanced composition of the diet and lower diversity of food intake [bib_ref] Problem eating behaviors related to social factors and body weight in preschool..., Dubois [/bib_ref] [bib_ref] The phenomenon of "picky eater": a behavioral marker in eating patterns of..., Carruth [/bib_ref] [bib_ref] Prevalence of picky eaters among infants and toddlers and their caregivers' decisions..., Carruth [/bib_ref] [bib_ref] Associations of infant feeding practices and picky eating behaviors of preschool children, Shim [/bib_ref] [bib_ref] Picky eating during childhood: a longitudinal study to age 11 years, Mascola [/bib_ref]. Specifically in children [bib_ref] Fruit and vegetable intake in young children, Dennison [/bib_ref] [bib_ref] Parental influences on young girls' fruit and vegetable, micronutrient, and fat intakes, Fisher [/bib_ref] [bib_ref] Predictors and consequences of food neophobia and pickiness in young girls, Galloway [/bib_ref] with a lower intake of vegetables and fruit, and a higher consumption of unhealthy processed foods, picky eating behaviour may be of considerable concern to parents. This behaviour forms a common complaint at physician visits and creates conflicts in feeding practice [bib_ref] Picky eating during childhood: a longitudinal study to age 11 years, Mascola [/bib_ref].
Picky eaters have a lower intake of food groups rich in micronutrients [bib_ref] Associations of infant feeding practices and picky eating behaviors of preschool children, Shim [/bib_ref] , such as fruits, vegetables, and meats, and a lower intake of vitamin C, vitamin E, fibre, and folate [bib_ref] Food neophobia and 'picky/fussy' eating in children: a review, Dovey [/bib_ref] [bib_ref] Parental pressure, dietary patterns, and weight status among girls who are "picky..., Galloway [/bib_ref] [bib_ref] Protein-energy malnutrition and feeding refusal secondary to food allergies, Fortunato [/bib_ref] , which might result in a potential risk for nutritional deficits and inadequate weight. However, whether or not picky eating behaviour significantly stunts children's growth and development is still a controversial topic. Some studies [bib_ref] Does picky eating affect weight-for-length measurements in young children?, Ekstein [/bib_ref] [bib_ref] Early feeding problems in an affluent society. IV. Impact on growth up..., Dahl [/bib_ref] [bib_ref] Early feeding problems in an affluent society. III. Follow-up at two years:..., Dahl [/bib_ref] focusing on toddlers with eating problems found that problematic eaters had long-standing problems with associated failure to thrive. Other studies [bib_ref] Problem eating behaviors related to social factors and body weight in preschool..., Dubois [/bib_ref] [bib_ref] Parental pressure, dietary patterns, and weight status among girls who are "picky..., Galloway [/bib_ref] reported that pre-schoolers with picky eating behaviour were more likely to be underweight and less likely to be overweight. However, there were also some studies [bib_ref] Picky eating during childhood: a longitudinal study to age 11 years, Mascola [/bib_ref] [bib_ref] How do toddler eating problems relate to their eating behavior, food preferences,..., Wright [/bib_ref] showing no significant associations between picky eating behaviour and children's growth, even when evaluated with z-scores.
Nutrition during early life may have a long-lasting influence on children's cognitive development due to continuous development of frontal lobes throughout early childhood, especially for the first two years of life [bib_ref] The influence of dietary status on the cognitive performance of children, Benton [/bib_ref] [bib_ref] Intakes and adequacy of potentially important nutrients for cognitive development among 5-year-old..., Mcafee [/bib_ref]. Pre-schoolers with picky eating starting in early life might suffer poor cognitive development. However, there is no clear conclusion on whether long-term unhealthy dietary habits, such as picky eating, have negative effects on children's cognitive development.
Evidence [bib_ref] Predictors and consequences of food neophobia and pickiness in young girls, Galloway [/bib_ref] [bib_ref] Successful strategies to increase the consumption of fruits and vegetables: results from..., Lassen [/bib_ref] is well documented suggesting a need for increasing the consumption of fruits and vegetables, supporting the recommendation of consuming more servings of fruits and vegetables as part of healthy dietary patterns for children. Previous studies [bib_ref] Prevalence of picky eaters among infants and toddlers and their caregivers' decisions..., Carruth [/bib_ref] [bib_ref] Associations of infant feeding practices and picky eating behaviors of preschool children, Shim [/bib_ref] [bib_ref] Is picky eating an eating disorder?, Jacobi [/bib_ref] reported that picky eaters usually had limited dietary variety characterised by a lower intake of fruits, vegetables, and meats. Fruit and vegetable consumption might help children maintain a healthy weight by replacing excess energy derived from fatty foods [bib_ref] Food intakes of US children and adolescents compared with recommendations, Munoz [/bib_ref]. However, the association between the lower intake of fruits and vegetables and growth (especially for the underweight) is unclear. A previous study [bib_ref] Parental pressure, dietary patterns, and weight status among girls who are "picky..., Galloway [/bib_ref] found that picky eating was associated with a lower intake of micronutrients such as calcium and magnesium; however, it is still unclear whether picky eating behaviour results in a lower content of micronutrients in blood. Moreover, there is limited knowledge of the eating behaviours of Chinese pre-schoolers. Therefore, our research was conducted to estimate the proportion of pre-schoolers' picky eating behaviour in China, to investigate possible associations between picky eating behaviour and children's growth and development (including weight, height, BMI, intelligence, and zscores about weight, height, and BMI), and to identify corresponding potential mechanisms, such as nutrient and food subgroup intake, as well as micronutrient levels in the blood.
# Subjects and methods
## Study sample
The study protocol was reviewed and approved by the Ethical Committee of the Health Science Center at Peking University (NO.IRB00001052-11042). Written informed consent forms were signed by the legal guardians of all participants.
Data was collected between November 2011 and April 2012 from healthy pre-school children in China. By a using multi-stage stratified cluster sampling method, 937 children aged 3-7 years were enrolled in our study. In the first stage, according to geography and economic development level, seven cities (Beijing, Guangzhou, Chengdu, Shenyang, Suzhou, Lanzhou and Zhengzhou city), one village in the plains and one village in the mountainous area (both located in Xingtai, Hebei province), were selected via purposive sampling. In the second stage, in view of sample size and representativeness, one large kindergarten located in the semi-urban area of each city, and one large kindergarten in each village were selected. In the last stage, one junior-, one middle-and one senior-class in each kindergarten were selected randomly, and all children in the selected classes were surveyed. Criteria for eligibility were pre-schoolers aged 3-7 years, with no reported birth defects (including congenital heart disease, hydrocephalus or deformity at birth), no reported infantile paralysis and thalassemia, and no acute health problems (including common cold and diarrhoea) at present.
## Socio-demographics, anthropometry, intelligence, and blood measurement
The structured questionnaire was used to collect socio-demographic information through interviewing pre-schoolers' parents, including child's date of birth, gender, ethnicity, birth weight, dietary behaviour, and feeding patterns during the first four months after birth. Sociodemographic information about their parents was also collected, including date of birth, height, weight, educational level, and household income.
In every kindergarten, two well-trained investigators carried out the measurement of height and weight, respectively. Participants were asked to remove all heavy clothes and shoes (leaving only minimal clothing on in a comfortable preparation area), and weight was taken to the nearest 0.01 kg on a calibrated electronic weighing scale. Standing height was taken to the nearest to 0.1 cm using a height measuring tape suspended from the wall. The participants were required to stand erect with their shoulders level, hands by their sides, thighs together, and heels comfortably together, with the upper back, buttocks, and heels in contact with the wall during height measurement. All physical measurements were carried out three times, and the mean values were calculated. Intelligence was assessed in a quiet, isolated room using the Chinese version of the Wechsler Intelligence Scale for Children (C-WISC), which contains verbal and performance subtests. The Verbal (VIQ) and Performance (PIQ) scores from the Wechsler Intelligence Scale are composites of skills for performing tasks in the subtests, and all subtests are combined to produce a Full Scale IQ (FIQ). The C-WISC has shown good reliability and validity, and was used widely to measure pre-schoolers' general intelligence in China [bib_ref] Report on Shanghai norms for the Chinese translation of the Wechsler Intelligence..., Li [/bib_ref] [bib_ref] Neuropsychological and behavioral status of Chinese children with acyanotic congenital heart disease, Yang [/bib_ref] [bib_ref] Is there a difference in cognitive development between preschool singletons and twins..., Xing [/bib_ref]. Eight interviewers trained by a psychologist for children from Beijing Normal University administered the tests.
Fasting blood samples for the evaluation of microelements were collected in the morning. The blood samples were drawn via venepuncture, and were collected into 5-mL, metal free EDTA-treated tubes. During the process of blood collection, blood samples were kept on ice, occasionally shaken, then transferred to storage at 4°C. Calibration solution of the multi-minerals quality product in whole blood was obtained for quality assurances (BOHUI INNOVA-TION Ltd., Beijing). The concentration of calcium, copper, iron, magnesium, and zinc in blood were measured with a BOHUI 5100S atomic absorption spectrophotometer (BOHUI INNO-VATION Ltd., Beijing) using hollow cathode lamps (422.7, 285.2, 248.3, 213.9, and 324.7 nm for calcium, magnesium, iron, zinc, and copper, respectively). All analyses on whole blood were performed in the same laboratory.
## Variables for picky eating
In the current study, picky eaters were characterised as children who consumed an inadequate variety and amount of food through rejection of foods that are familiar (and unfamiliar) to them because of specific tastes, textures, smells or appearances. The definition was provided to caregivers before the picky eating assessment, and caregiver's perceptions were used to judge if their child was a "picky eater". At each assessment in the structured questionnaire, the primary caregiver (usually the mother) was asked "Do you consider your child as having picky eating behaviour at present?" by choosing from three possible answers: 1) never picky, 2) somewhat picky, and 3) always picky. This single question was chosen based on observations from previous studies [bib_ref] Prevalence of picky eaters among infants and toddlers and their caregivers' decisions..., Carruth [/bib_ref] [bib_ref] Associations of infant feeding practices and picky eating behaviors of preschool children, Shim [/bib_ref] [bib_ref] Perception of picky eating among children in Singapore and its impact on..., Goh [/bib_ref]. Children of those caregivers responding "somewhat" and "always" were categorised as picky eaters. Unless they answered "never", they were asked to describe perceptions of the age of onset and duration of picky eating behaviour, and the specific foods nit-picked by picky eaters with response choices of milk, soy, cereal, vegetables, fruits, meat, and eggs.
## Dietary measurement
The nutrient intake of the pre-schoolers was determined from the dietary data collected using a 24-hour food intake record/recall. One trained, experienced researcher interviewed the mothers and another trained researcher checked the quality of food record/recall during the dietary measurement. Dietary data was collected at both the kindergarten and individual levels. Food consumption in the kindergarten was determined by conducting a detailed examination of changes in inventory before and after each meal with a kitchen scale. The individual nutritional intake was determined from the dietary data collected using a 24-hour food intake record, which was finished by researchers and kindergarteners. In consideration of some pre-schoolers having extra meals, the parents were required to record the extra food intake at home for 24 hours. To identify and quantify the food more accurately, some food models and a series of pictures of standard bowls and spoons were shown and explained to the parents before the dietary survey. The parents were asked to report all of the food their children consumed at home and while away from home. The nutrient contents of the foods were estimated using the China Food Composition book (2004). If their child was taking dietary supplements, the parents were asked to provide specific information regarding the supplements including brand names, manufacturers, and daily dosages in order to estimate the composition and content of nutrients. The FFQ (Food Frequency Questionnaire) was used to obtain data on consumption of the following food subgroups: cereals, vegetables, fruits, meat, fish, eggs, milk, nuts and beans, and oils and fats. Parents were asked to define the frequency and quantity of each subgroup of food consumed by their children over the last six months.
# Statistical analysis
The database was established using Epi Data version 3.0, and a double data entry was carried out. We used mean value ± standard error or percentage to describe the baseline characteristics of participants. Comparison of categorical data was carried out by using Pearson's χ 2 test, and a Student's t-test was used for normally distributed data. According to the WHO Growth Standards, we calculated z-scores for height, weight, and BMI by using WHO AnthroPlus. To verify the functional relation between picky eating and growth and development of the pre-schoolers, to characterise the associations between the duration of picky eating and pre-schoolers' growth and development, and to estimate the associations between the subgroups of food nit-picked and growth and development we used the statistical software MLwiN 2.31 to construct multilevel (two levels) mixed-effects linear regression models with investigation areas (level 1) nested within individuals (level 2). The multilevel model accounts for the clustering of the data; it can correct biases in parameter estimates and can produce correct CIs and significance tests. The regression coefficients were expressed by β, SEM and the 95% CI in comparison with the non-picky eating group. All of the correlation analyses for height, weight, and BMI were adjusted for the following variables: age, gender, and birth weight of the children, education of the mothers, income per capita of the families, and the feeding patterns during the first four months after birth. All of the correlation analyses of z-scores were also adjusted for variables listed above (not including age and gender). As for nutrient intake, intake of food groups and mineral content in whole blood, mean value ± standard error was used to describe continuous variables, and Student's t-tests and covariance analyses adjusted for child's gender and age were used to compare the differences between picky eaters and non-picky eaters. All t-tests and covariance analyses were performed using SPSS 15.0. Statistical significance was determined at p < 0.05 (two-tailed tests).
# Results
At the initial interview, 937 healthy pre-schoolers participated in our survey. Only those preschoolers with their anthropometry and intelligence measured, and who completed their assessment of picky eating behaviour, were included in this study, resulting in the final analysis set of 911 eligible pre-schoolers including 423 (46%) non-picky eaters and 488 (54%) picky eaters. Socio-demographic information of the children and their mothers is provided in [fig_ref] Table 1: Socio-demographic characteristics of mother-pre-schooler dyads of the subjects [/fig_ref]. There were no significant differences in height, weight, and BMI of the mothers, or in the household income between the picky eating and non-picky eating groups. The mothers in the picky eating group were younger (p < 0.05) and had lower levels of education (p = 0.004). The group of picky eaters did not differ from the non-picky eating group in birth weight, ethnicity, gender, and feeding patterns during the first four months after birth. However, the pre-schoolers in the picky eating group were younger (p = 0.002).
The results of correlations between picky eating behaviour and the growth and development of children are provided in [fig_ref] Table 2: Parameters of growth and development of non-picky eating and picky eating groups [/fig_ref]. Crude and adjusted linear regression analyses indicated that children in the picky eating group had significantly lower weights and z-scores of weight for age compared with the non-picky eating group. Picky eating was associated with a 0.42 kg difference (lower) in weight (95% CI: -0.72, -0.12; p = 0.006) and a 0.14 unit lower in the zscore of weight for age (95% CI: -0.25, -0.02; p = 0.017), respectively, compared with non-picky eating. Though significant differences in height, z-score of height for age, and intelligence were found in picky eating and non-picky eating groups (p < 0.05), there were no significant differences between the two groups after adjusting for other influencing factors (p > 0.05). In the current study, a crude analysis indicated that children in the non-picky eating group did not have significantly lower BMIs and z-scores of BMIs for age, compared with the picky eating group (p < 0.05). Though significant differences in BMIs were found between the two groups according to the adjusted linear regression (p < 0.05), there was no significant difference in zscores of BMIs for age from the linear model (p > 0.05).
The picky eaters had different durations of picky eating behaviour: picky eating lasted 0-2 years in 203 (42%) subjects, 2-3 years in 124 (25%) subjects, and over 3 years in 104 (21%) subjects. To further verify the associations between picky eating behaviour and pre-schoolers' growth and development, comparisons between non-picky eating and picky eating groups with different durations were carried out [fig_ref] Table 3: Z-scores for height, weight, and BMI of pre-schoolers with picky eating habits... [/fig_ref]. The z-scores of height for age, weight for age, and BMI for age decreased with an increase in the duration of picky eating behaviour. Compared with the non-picky eating group, there were no significant differences in z-scores of height for age, weight for age and BMI for age during the first and second year of picky eating (p > 0.05). Compared with the corresponding values in the non-picky eating group, lower zscores of weight for age were found in pre-schoolers with picky eating lasting over two years, and lower z-scores of BMI for age were found in those with picky eating for over three years by crude and adjusted linear regression analyses (p < 0.05). Picky eating lasting 2-3 years and over three years were associated with 0.22 (95% CI: -0.39, -0.05; p = 0.011) and 0.25 (95% CI: -0.43, -0.06; p = 0.008) z-scores lower in weight for age, respectively. Being a picky eater for over three years was associated with a 0.10 (95% CI: -0.45, -0.06; p = 0.012) z-score lower in BMI for age. Compared with the non-picky eating group, there were no significant differences in intelligence of pre-schoolers with picky eating lasting any years (p > 0.05).
The food nit-picked by picky eaters were mostly vegetables (56% of the picky eaters), meat (23%), soy (21%), eggs (21%), cereal (19%), fruits (12%), and milk (8%). Children nit-picking of meat had a 0.19 (95% CI: -0.37, -0.02; p = 0.034) z-score lower in weight for age and a 0.23 (95% CI: -0.41, -0.04; p = 0.017) z-score lower in BMI for age. No significant associations were found between z-scores and nit-picking of vegetables, soy, eggs, cereal, fruits or milk (p >0.05) [fig_ref] Table 4: Influence of nit-picking of subgroups of food on growth and development of... [/fig_ref].
The results of dietary mean nutrient intake from the 24-hour dietary recall are shown in [fig_ref] Table 5: Dietary intake of energy, macronutrients, dietary fibre, minerals and vitamins of pre-schoolers... [/fig_ref]. Compared with the non-picky eating group, the intake of protein and dietary fibre was lower in the picky eating group according to the crude and adjusted linear regression analyses (p < 0.05). There was no significant difference in intake of energy, fat and carbohydrates or vitamins between the two groups (p > 0.05). The crude and adjusted linear regression analyses showed a lower intake of iron and zinc (p < 0.05), but no differences in the intake of magnesium, calcium, and copper (p > 0.05).
The mean intake of food groups (g/day) in the non-picky eating and picky eating groups are shown in . According to crude and adjusted linear regression analyses, picky eaters had a lower dietary intake of cereals (p < 0.05), vegetables (p < 0.001) and fish (p < 0.05). However, there were no significant differences in the intake of fruit, meat, eggs, milk, nuts and beans, or oils and fats (p > 0.05) between the two groups. a Values were given as Mean ± SE. † All of the models were constructed by using multilevel (two levels) mixed-effects liner regression with the iterative generalised least-squares estimation method. Results of the measurements from the regression models with adjustment for child's age, gender (female, male), birth weight, and feeding pattern during the first four months after birth (exclusive breastfeeding, mixed feeding, artificial feeding, and unclear), mother's education (middle school or below, high school, college or above, and unclear), and family's per capita monthly income (< 2000, 2000-4000, > 4000, and unclear, Yuan). Results of zscores from the regression models with adjustment for child's birth weight and feeding pattern during the first four months after birth, mother's education, and household income. ‡ β represents the difference in mean parameters of growth and development between non-picky eating and picky eating groups after adjusting for the covariates listed above.
doi:10.1371/journal.pone.0123664.t002
The micronutrients found in whole blood are summarised in . We found no difference between the two groups (p > 0.05).
# Discussion
We found that parents perceived 54% of their Chinese pre-schoolers to be picky eaters; the children consumed low quantities of limited diversity. Nit-picking, mostly vegetable, was reported in 56% of picky eaters. Picky eaters had younger, lower educated mothers. Growth was associated with picky eating: weight for age was lower and BMI as a consequence also, especially if the duration of picky eating was longer than 2 years; however, cognitive development was not. Dietary intake of picky eaters was lower in protein and fibre, as well as iron and zinc; consumption of vegetables, cereals and fish was lower. Whole blood values of minerals were unaffected. Nit-picking meat was associated with lower weight for age and BMI for age, but not 366 † All of the models were constructed by using multilevel (two levels) mixed-effects liner regression with the iterative generalised least-squares estimation method. Results of the z-scores from the regression models with adjustment for child's birth weight and feeding pattern during the first four months after birth, mother's education, and family's per capita monthly income. Results of intelligence from the regression models with adjustment for child's age, gender, birth weight, and feeding pattern during the first four months after birth, mother's education, and family's per capita monthly income. ‡ β represents the difference in mean z-scores between the non-picky eating group and the picky eating group after adjusting for the covariates listed above. a SE = standard error, b CI, confidence interval.
doi:10.1371/journal.pone.0123664.t003 Results of the z-scores from the regression models with adjustment for child's birth weight and feeding pattern during the first four months after birth, mother's education, family's per capita monthly income and other kinds of food related with children's growth listed above. ‡ β represents the difference in mean z-scores between the non-picky eating group and the picky eating group after adjusting for the covariates listed above. a CI, confidence interval.
doi:10.1371/journal.pone.0123664.t004 with height for age. Nit-pickers for meat consumed fewer meat, cereals, vegetables, and eggs (S3 , and had a lower dietary intake of magnesium (S2 , while whole blood values of minerals were unaffected (S4 . Previous studies [bib_ref] Associations of infant feeding practices and picky eating behaviors of preschool children, Shim [/bib_ref] [bib_ref] Children's preferences for high-fat foods, Birch [/bib_ref] [bib_ref] Using reinforcement and cueing to increase healthy snack food choices in preschoolers, Stark [/bib_ref] [bib_ref] Influence of parental attitudes in the development of children eating behaviour, Scaglioni [/bib_ref] [bib_ref] Variety is the spice of life: strategies for promoting fruit and vegetable..., Mennella [/bib_ref] have reported children's food preferences, which are developed from genetically determined predispositions and experiences, such as repeated exposure, feeding context, and social and physiological consequences, are the consistent determining factor of food acceptance. The first few years of life are considered a critical period for the development of food acceptance patterns [bib_ref] Development of food acceptance patterns in the first years of life, Birch [/bib_ref] [bib_ref] A sensitive period for learning about food, Cashdan [/bib_ref] due to early exposure to flavours through breastfeeding [bib_ref] Prenatal and postnatal flavor learning by human infants, Mennella [/bib_ref] [bib_ref] Demographic, familial and trait predictors of fruit and vegetable consumption by pre-school..., Cooke [/bib_ref] , artificial feeding [bib_ref] Sensitive period in flavor learning: effects of duration of exposure to formula..., Mennella [/bib_ref] , and complementary feeding [bib_ref] Longitudinal relationships between childhood, adolescent, and adult eating disorders, Kotler [/bib_ref] in babyhood and the toddler period. Food preference shaped by repeated experiences during the pre-school period may further lead to stability. Previous studies [bib_ref] Longitudinal correlates of the persistence of irregular eating from age 5 to..., Mcdermott [/bib_ref] [bib_ref] A prospective study of food variety seeking in childhood, adolescence and early..., Nicklaus [/bib_ref] reported that eating problems, including picky eating, were demonstrated continuously from early childhood into mid-adolescence, which was related to both later eating disorders, lasting fussy eating, and a limited variety of food preferred in adolescence and adulthood.
Picky eating is considered to be a complex behaviour [bib_ref] Picky eating during childhood: a longitudinal study to age 11 years, Mascola [/bib_ref] [bib_ref] Is picky eating an eating disorder?, Jacobi [/bib_ref] [bib_ref] Perception of picky eating among children in Singapore and its impact on..., Goh [/bib_ref] [bib_ref] Eating patterns in a populationbased sample of children aged 5 to 7..., Micali [/bib_ref] , which is difficult to quantify accurately in childhood. Picky eaters are usually defined as children who consume an inadequate amount and variety of food through rejection of some foods [bib_ref] Food neophobia and 'picky/fussy' eating in children: a review, Dovey [/bib_ref] [bib_ref] Characteristics of school children who are choosy eaters, Rydell [/bib_ref]. Our definition of picky eating behaviour was consistent with this. In our study, we found that picky eating . Comparison of intake (g/day) of various groups of food between pre-schoolers in non-picky and picky eating groups. behaviour was relatively common in Chinese pre-schoolers aged 3-7 years, and the proportion of picky eaters was up to 54%. Previous studies reported the frequency of picky eating behaviour with great variation, ranging from 8% to 50% of children, due to the different ages of children included, the varying areas/country of the study population, and the different definitions and assessment of picky eating [bib_ref] Food neophobia and 'picky/fussy' eating in children: a review, Dovey [/bib_ref]. In the present study, children who were picky eaters had younger mothers with lower levels of education than the non-picky eaters. Previous studies [bib_ref] Exploring the effects of maternal eating patterns on maternal feeding and child..., Morrison [/bib_ref] [bib_ref] Learning to overeat: maternal use of restrictive feeding practices promotes girls' eating..., Birch [/bib_ref] pointed out that parental control of feeding and eating patterns might contribute to children's eating behaviour. Older and more educated mothers might pay more attention to their children's eating behaviour and provide their children with a better selection of fruits and vegetables [bib_ref] Preschool children's eating behaviours are related to dietary adequacy and body weight, Dubois [/bib_ref] resulting in much better eating habits. Previous studies [bib_ref] Problem eating behaviors related to social factors and body weight in preschool..., Dubois [/bib_ref] [bib_ref] Picky eating during childhood: a longitudinal study to age 11 years, Mascola [/bib_ref] [bib_ref] How do toddler eating problems relate to their eating behavior, food preferences,..., Wright [/bib_ref] [bib_ref] Revisiting the picky eater phenomenon: neophobic behaviors of young children, Carruth [/bib_ref] showed inconsistent findings regarding the associations between picky eating behaviour and growth and development in children older than three years. Dubois et al [bib_ref] Problem eating behaviors related to social factors and body weight in preschool..., Dubois [/bib_ref] found that picky eaters were twice as likely as non-picky eaters to be underweight at 4.5 years old, but Mascola et al [bib_ref] Picky eating during childhood: a longitudinal study to age 11 years, Mascola [/bib_ref] did not observe similar results in their prospective study. In another study, Wright et al [bib_ref] How do toddler eating problems relate to their eating behavior, food preferences,..., Wright [/bib_ref] found weak associations between picky eating and poor growth, and simply eating a limited variety was found to be unrelated to growth. Carruth et al [bib_ref] Revisiting the picky eater phenomenon: neophobic behaviors of young children, Carruth [/bib_ref] followed 71 children with picky eating behaviour for 34-84 months and found that mean weights and heights were within normal values for age and genders. In the aforementioned studies, just a few results [bib_ref] Problem eating behaviors related to social factors and body weight in preschool..., Dubois [/bib_ref] were adjusted for confounding factors. Actually, children's characteristics, such as age, gender, birth weight [bib_ref] Longitudinal growth of very low birth weight neonates during first year of..., Mukhopadhyay [/bib_ref] , and feeding practices [bib_ref] Influence of Feeding Practices and Associated Factors on the Nutritional Status of..., Meshram [/bib_ref] during early life, and parents' characteristics, such as mother's education and household income, played an important role in children's growth and development. Other explanations for controversial results about picky eating may include inconsistent definitions and methods for the assessment of picky eating, lack of existing applicable measures, and the inappropriate comparison between the picky eating and non-picky eating groups.
In the present study, crude and adjusted linear regression analyses indicated significantly lower weights and z-scores of weight in the picky eating group than in the non-picky eating group. Previous studies [bib_ref] Problem eating behaviors related to social factors and body weight in preschool..., Dubois [/bib_ref] [bib_ref] The phenomenon of "picky eater": a behavioral marker in eating patterns of..., Carruth [/bib_ref] [bib_ref] Prevalence of picky eaters among infants and toddlers and their caregivers' decisions..., Carruth [/bib_ref] [bib_ref] Associations of infant feeding practices and picky eating behaviors of preschool children, Shim [/bib_ref] [bib_ref] Picky eating during childhood: a longitudinal study to age 11 years, Mascola [/bib_ref] reported that picky eaters had lower food diversity and ingested limited amounts of food, such as fruits, vegetables, and meat [bib_ref] Associations of infant feeding practices and picky eating behaviors of preschool children, Shim [/bib_ref] , and a lower intake of calories [bib_ref] Behavioral validation, precursors, and concomitants of picky eating in childhood, Jacobi [/bib_ref] , fat, calcium, magnesium, vitamin C, vitamin E [bib_ref] Parental pressure, dietary patterns, and weight status among girls who are "picky..., Galloway [/bib_ref] , and folate [bib_ref] Food neophobia and 'picky/fussy' eating in children: a review, Dovey [/bib_ref] [bib_ref] Protein-energy malnutrition and feeding refusal secondary to food allergies, Fortunato [/bib_ref] , which contribute to children's growth. At the same time, a lower intake of protein and iron by picky eaters was found in the present study, which predisposed the child to infection, being underweight, and a stunting of the growth [bib_ref] Protein Energy Malnutrition in India: The Plight of Our Under Five Children, Bhutia [/bib_ref]. In addition, lower intelligence in the picky eating group was not detected in adjusted linear regression analyses, though lower zinc intake was found in the picky eating group. Studies [bib_ref] Oral zinc supplementation may improve cognitive function in schoolchildren, De Moura [/bib_ref] in animal models and humans demonstrated the effects of zinc deficiency on decreased cognitive development activities; thus, it is necessary to identify whether long-term lower zinc intake among picky eaters may result in zinc deficiency and impairment in cognitive development in the future.
In this study, over 45% of picky eaters had this behaviour for over two years, and only 14% of the children were recognised as picky eaters during the past year, which suggests that picky eating starts mainly during early childhood and is often of longer duration. A previous study [bib_ref] Picky eating during childhood: a longitudinal study to age 11 years, Mascola [/bib_ref] reported picky eaters of longer durations were less likely to accept new foods and showed stronger liking or disliking of food than those of shorter durations. To further verify the association between picky eating and growth and development in pre-schoolers, we compared the non-picky eating group with the picking eating group separated into duration of pickiness. In agreement with our hypothesis, z-scores for weight and BMI were significantly lower in children with picky eating habits lasting longer than three years.
Although a lower intake of some subgroups of food and nutrients was found in picky eaters (and picky eating behaviour might have negative effects on children's growth), the associations between the intake of specific subgroups of food and the growth of children are still unclear. A previous study reported that replacing fresh produce, such as fruits and vegetables, with industrially processed foods might increase the risk of being overweight and having chronic diseases in the future [bib_ref] Food neophobia and 'picky/fussy' eating in children: a review, Dovey [/bib_ref]. On the other hand, nit-picking subgroups of food might result in children being underweight. In the present study, nit-picking meat was found to have lower z-scores for weight and BMIs. Meat, an important source of protein, fat, heme iron and fat-soluble vitamins, contributes to children's nutritional status and growth. Appropriate intake of meat is very important for the nutritional needs of children. The general recommendation should be provided for picky eaters (just like the intake of vegetables and fruits: at least five portions each day for children [bib_ref] Successful strategies to increase the consumption of fruits and vegetables: results from..., Lassen [/bib_ref].
Galloway [bib_ref] Parental pressure, dietary patterns, and weight status among girls who are "picky..., Galloway [/bib_ref] found that picky eaters consumed fewer foods containing vitamin E, vitamin C, folate and fibre, probably due to their lower consumption of fruits and vegetables compared to non-picky eaters. In the present study, a lower intake of protein was found in the picky eating group, probably due to lower consumption of fish. Lower dietary intake of fibre, iron, and zinc were also found, most likely due to a lower intake of cereals, vegetables and fish. Considering the long-term lower intake of proteins and micronutrients may result in nutritional deficiencies, which may have long-term developmental consequences [bib_ref] Stunting, Wasting, and Micronutrient Deficiency Disorders, Caulfield [/bib_ref] , considerable resources have been directed to the nutritional assessment [bib_ref] Micronutrient deficiencies are common in 6-to 8-year-old children of rural Nepal, with..., Schulze [/bib_ref] of the pre-school years when nutrient requirements are exacerbated by rapid growth. However, in our study, we did not find significant differences between picky eaters and non-picky eaters regarding the content of essential minerals in whole blood.
This study has several limitations. Firstly, it was not accurate enough to quantify picky eating behaviour in pre-schoolers by simply asking their parents whether or not they considered their child to be a picky eater. Differences in the judgment of picky eating exist in parents' perceptions, which might affect data interpretation [bib_ref] Food neophobia and 'picky/fussy' eating in children: a review, Dovey [/bib_ref]. Although a strict definition was provided in our study, some children still might be misclassified because of subjective reports and measurement. Therefore, it is essential that a universal and widely applicable psychometric tool be offered to measure picky eating behaviour with valid and reliable measures for future research. Secondly, although significant associations between picky eating behaviour and children's growth have been found, our cross-sectional study design did not provide direct evidence of the causality. A future prospective longitudinal study with a large sample size and consideration of potential confounders needs be carried out. Thirdly, the 24-hour dietary recall is widely used to assess children's nutrition intake [bib_ref] Prevalence of picky eaters among infants and toddlers and their caregivers' decisions..., Carruth [/bib_ref] [bib_ref] Predictors and consequences of food neophobia and pickiness in young girls, Galloway [/bib_ref] [bib_ref] Parental pressure, dietary patterns, and weight status among girls who are "picky..., Galloway [/bib_ref] , but its variability results in difficulty in accurately estimating individual's long-term diets [bib_ref] Food intake measurement: problems and approaches, Todd [/bib_ref]. In the present study, to reduce the influence of this limitation, a trained researcher interviewed parents and their child in a face-to-face interview with the aid of food models. Long-term dietary information from the FFQ was also collected, which contributed to the comprehensive understanding of children's nutrient intake.
# Conclusions
Picky eating in Chinese pre-schoolers is very common, reported more often by younger mothers with lower education. It is associated with lower weight for age especially when lasting over two years. Lower intake of protein, fiber, iron, and zinc is found in these pre-schoolers, probably due to lower vegetables, cereal and fish intake. Nit-picking meat is associated with lower weight for age and BMI for age, and this relationship requires further investigation to disentangle the possible nutrimental, environmental, and hereditary causal pathways of these mechanisms. We therefore recommend standardization of picky eating and propose the definition we used. Parents should pay attention to nit-picking meat and picky eating in order to prevent their lasting effects on the growth of pre-schoolers.
[fig] †: All of the models were constructed by using multilevel (two levels) mixed-effects liner regression with the iterative generalised least-squares estimation method. [/fig]
[table] Table 1: Socio-demographic characteristics of mother-pre-schooler dyads of the subjects. [/table]
[table] Table 2: Parameters of growth and development of non-picky eating and picky eating groups.Non-picky eating a (n = 423) Picky eating a (n = 488) p value Adjusted † β ‡ SEM 95% confidence interval [/table]
[table] Table 3: Z-scores for height, weight, and BMI of pre-schoolers with picky eating habits of different durations, and adjusted associations between z-scores and duration of picky eating behaviours. [/table]
[table] Table 4: Influence of nit-picking of subgroups of food on growth and development of pre-schoolers. [/table]
[table] Table 5: Dietary intake of energy, macronutrients, dietary fibre, minerals and vitamins of pre-schoolers of non-picky eating and picky eating groups. [/table]
|
Glycaemic, blood pressure and cholesterol control in 25 629 diabetics
Objective: To examine and compare the extent to which people with type 2 diabetes (T2DM) are achieving haemoglobin A 1c (HbA 1c ), blood pressure (BP) and LDL cholesterol (LDL-C) treatment targets. Methods: A review of databases (MEDLINE Ovid, Pubmed and Sabinet) was performed and limited to the following terms: type 2 diabetes mellitus AND guideline AND goal achievement for the years 2009 to 2014 (five years). Results: A total of 14 studies (25 629 patients) were selected across 19 different countries. An HbA 1c level of 7.0% (or less) was achieved by 44.5% of subjects (range 19.2-70.5%), while 35.2% (range 7.4-66.3%) achieved BP of 130/80 mmHg (or less), and 51.4% (range 20.0-82.9%) had an LDL-C level of either 2.5 or 2.6 mmol/l (100 mg/dl or less). Conclusion: Despite guideline recommendations that lowering of HbA 1c , BP and lipids to target levels in T2DM will lead to a reduction in morbidity and mortality rates, we found that control of these risk factors remains suboptimal, even across different settings.
International Diabetes Federation (IDF), there were 366 million people living with diabetes in 2011.By 2035, it is predicted that more than half a billion people will have the disease.
Trends in urbanisation and the adoption of unhealthy Western lifestyles have begun to affect low-and middle-income countries (LMICs). A prime example of this is South Africa, which previously had the dubious pleasure of infectious diseases being the primary source of mortality. Today, expansion of non-communicable diseases (NCD) is beginning to manifest and deplete the already strained health resources available. [bib_ref] Diabetes in sub-Saharan Africa 1999-2011: epidemiology and public health implications. A systematic..., Hall [/bib_ref] Rather than being limited to glycaemia alone, the management of type 2 diabetes mellitus (T2DM) includes multiple priorities, including identification and treatment of other modifiable risk factors. It is widely accepted that T2DM is associated with cardiovascular disease (CVD) and increased mortality rates. [bib_ref] Diabetic cardiomyopathy: mechanisms, diagnosis and treatment, Hayat [/bib_ref] In addition to lifestyle changes, the importance of reduction in levels of low-density lipoprotein cholesterol (LDL-C) and blood pressure (BP) has become an essential primary goal for the prevention of CVD in T2DM. [bib_ref] Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in..., Colhoun [/bib_ref] Furthermore, improved outcomes of diabetes-related chronic microvascular complications (retinopathy, neuropathy and nephropathy) are achieved through substantial reductions in incidence of both hyperglycaemia and hypertension. It is on the basis of this research that the Society for Endocrinology, Metabolism and Diabetes of South Africa (SEMDSA) recommends that most adults with diabetes should aim for an HbA 1c level of 7.0%, BP of 140/80 mmHg and LDL-C level of 2.5 mmol/l or less. There are many gaps in the management of T2DM that are proving difficult to close. Studies have revealed how clinical practice differs from clinical trials in that T2DM patients often cannot reach guideline-recommended targets. One of the ways to improve clinical outcomes is by comparing the performance of one clinical setting against another. In this study, our aim was to compare the achievement of the critical quality indicators: glycaemic, BP and lipid control in T2DM patients from different countries worldwide, in an attempt to benchmark which approach has been most successful.
# Methods
This study was a literature review using Ovid MEDLINE, Pubmed and Sabinet databases. Studies included were those conducted in the past five years (2009-2014) and limited to the following key terms: type 2 diabetes mellitus AND guideline AND goal achievement (HbA 1c , glycated haemoglobin, blood pressure, systolic, diastolic, lipids, cholesterol, LDL cholesterol). We also reviewed a selected number of reference lists of other department of pharmacy and pharmacology, school of therapeutic sciences, Faculty of health sciences, university of the witwatersrand, Johannesburg, south africa reviews and hand-searched several medical journals.
Studies that reported achievement of guideline-recommended targets of major risk factors for T2DM were included. The primary objective of this review was to provide an overview of achievement of major risk-factor targets (HbA 1c , BP and LDL-C) in the treatment of a sample of T2DM patients from different parts of the world. Specifically, the objectives would be addressed through comparison of the achievement of HbA 1c , BP and LDL-C targets, according to local or international guidelines, across different study samples.
The following data were extracted from the studies: author details, year of publication, study location, cohort size and achievement of major risk factors (combined systolic and diastolic BP, and HbA 1c and LDL-C levels). As different samples of study countries followed different guideline targets, flexibilities around these differences was needed. Studies selected for this article may have differed in the following parameters: recruitment and randomisation methods, total number of study participants recruited, study sites (e.g. single or multicentre), gender ratios, ethnicity ratios, timelines of results presented (e.g. single or longitudinal data) and periods of enrollment.
To compare results, we standardised (or converted or conformed) certain measurement units in order to maintain consistency (e.g. LDL-C in mmol/l instead of mg/dl). The control or baseline results of studies were reported instead of interventional group data. Only the latest data were selected from studies with multiple time periods.
Studies excluded from the review had one or more of the following characteristics: non-English language, studies conducted before 2009, participating patients younger than 18 years of age, participants reported to have had any diabetes other than T2DM (e.g. gestational, type 1 or steroid induced), studies that reported insufficient data or less than two of the three major risk factors being compared, and studies that consisted of large HMO claims databases. The latter was chosen as an exclusion criterion as larger-sized cohort studies would have biased the results of this review.
Data presented in this article were collected from the results of other studies and are limited to the authors' definitions of control. This review did not allow for the access of patient-level data of different studies included in the review to be accessed. It was assumed that all data extracted for this study were collected from the medical records of patients who willingly participated in the studies included in this review. The relevant data were captured into a secure database using Microsoft Excel 2010. Ethical approval was obtained from the University of the Witwatersrand Human Research Ethics Committee (Medical).
# Results
The authors of this study set out to determine how diabetes care compared across different settings, given the healthcare challenges faced especially by under-resourced areas. Of the 511 (154 from Ovid MEDLINE + 32 Pubmed + 325 Sabinet) titles initially identified between 2009 and 2014, 14 studies fulfilled the inclusion criteria. These 14 studies originated from 19 different countries (some studies included more than a single country) and we enrolled a total of 25 629 patients.
There were 17 high-income, one upper-middle-(South Africa) and one low-income (Uganda) country included in the review (grouped according to the United Nations' economies by per-capita country classification).Cohort sizes ranged from 50 to 4 926 patients. Twelve studies contained results for all major risk factors (HbA 1c , BP and LDL-C), while the rest included at least two-thirds of the measured risk factors. There were eight studies (57.1%) that defined treatment targets as per the American Diabetes Association.The characteristics of each study are outlined in [fig_ref] table 1: study characteristics [/fig_ref].
In 12 studies (25 354 patients) that used an HbA 1c level of 7.0% or less to define control, 44.5% (range 19.2-70.5%) of patients achieved target. [bib_ref] Evaluation of adherence to international guidelines for treating patients with type 2..., Dm [/bib_ref] [bib_ref] Poor achievement of guidelinesrecommended targets in type 2 diabetes: findings from a..., Braga [/bib_ref] [bib_ref] The prevalence of meeting A 1C , blood pressure, and LDL goals..., Stark Casagrande [/bib_ref] [bib_ref] Type 2 diabetes in primary care in Belgium: need for structured shared..., Goderis [/bib_ref] [bib_ref] Benchmarking is associated with improved quality of care in type 2 diabetes:..., Hermans [/bib_ref] [bib_ref] Suboptimal glycaemic and blood pressure control and screening for diabetic complications in..., Kibirige [/bib_ref] [bib_ref] Sub-optimal management of type 2 diabetes mellitus -a local audit, Klisiewicz [/bib_ref] [bib_ref] The current status of type 2 diabetes management at a university hospital, Lee [/bib_ref] [bib_ref] Pharmacist management of patients with diabetes mellitus enrolled in a rural free..., Sease [/bib_ref] [bib_ref] Diabetes care and complications in primary care in the Tshwane district of..., Webb [/bib_ref] [bib_ref] Quality of care of people with type 2 diabetes in eight European..., Stone [/bib_ref] In two studies (275 patients) where HbA 1c level was defined as < 6.5 and < 8.0%, respectively, 56.6 and 60.0% of patients reached their targets, respectively. [bib_ref] Assessment of treatment goals attained by patients according to guidelines for diabetes..., Morren [/bib_ref] [bib_ref] Impact of a quality improvement intervention on provider adherence to recommended standards..., Umar-Kamara [/bib_ref] In eight studies (18 089 patients), which had the definition of target BP of 130/80 mmHg or less (systolic and diastolic combined), 35.2% (range 7.4-66.3%) of patients achieved target. [bib_ref] Evaluation of adherence to international guidelines for treating patients with type 2..., Dm [/bib_ref] [bib_ref] Poor achievement of guidelinesrecommended targets in type 2 diabetes: findings from a..., Braga [/bib_ref] [bib_ref] The prevalence of meeting A 1C , blood pressure, and LDL goals..., Stark Casagrande [/bib_ref] [bib_ref] The current status of type 2 diabetes management at a university hospital, Lee [/bib_ref] [bib_ref] Pharmacist management of patients with diabetes mellitus enrolled in a rural free..., Sease [/bib_ref] [bib_ref] Diabetes care and complications in primary care in the Tshwane district of..., Webb [/bib_ref] [bib_ref] Assessment of treatment goals attained by patients according to guidelines for diabetes..., Morren [/bib_ref] In four studies (7 240 patients) where systolic BP targets of 130 mmHg or less (alone) defined control, 32.7% (range 21.3-50.0%) of the subjects achieved target. [bib_ref] Type 2 diabetes in primary care in Belgium: need for structured shared..., Goderis [/bib_ref] [bib_ref] Benchmarking is associated with improved quality of care in type 2 diabetes:..., Hermans [/bib_ref] [bib_ref] Sub-optimal management of type 2 diabetes mellitus -a local audit, Klisiewicz [/bib_ref] [bib_ref] Quality of care of people with type 2 diabetes in eight European..., Stone [/bib_ref] In two studies (300 patients) with a BP target of either < 140/90 or < 140/80 mmHg, 24.0 and 56.0% of patients achieved goal, respectively. 14,22
## Africa
In the 11 studies (24 702 patients) that used LDL-C levels of either 2.5 or 2.6 mmol/l (100 mg/dl or less) to define control, 51.4% (range 20.0-82.9%) of patients achieved goal. [bib_ref] Poor achievement of guidelinesrecommended targets in type 2 diabetes: findings from a..., Braga [/bib_ref] [bib_ref] The prevalence of meeting A 1C , blood pressure, and LDL goals..., Stark Casagrande [/bib_ref] [bib_ref] Type 2 diabetes in primary care in Belgium: need for structured shared..., Goderis [/bib_ref] [bib_ref] Benchmarking is associated with improved quality of care in type 2 diabetes:..., Hermans [/bib_ref] [bib_ref] Suboptimal glycaemic and blood pressure control and screening for diabetic complications in..., Kibirige [/bib_ref] [bib_ref] Sub-optimal management of type 2 diabetes mellitus -a local audit, Klisiewicz [/bib_ref] [bib_ref] The current status of type 2 diabetes management at a university hospital, Lee [/bib_ref] [bib_ref] Pharmacist management of patients with diabetes mellitus enrolled in a rural free..., Sease [/bib_ref] [bib_ref] Diabetes care and complications in primary care in the Tshwane district of..., Webb [/bib_ref] [bib_ref] Quality of care of people with type 2 diabetes in eight European..., Stone [/bib_ref] One study (225 patients) with a total cholesterol target of < 200 mg/ dl (5 mmol/l) had 49.3% of patients at goal. [bib_ref] Assessment of treatment goals attained by patients according to guidelines for diabetes..., Morren [/bib_ref] Two studies (702 patients) did not measure lipid levels. [bib_ref] Evaluation of adherence to international guidelines for treating patients with type 2..., Dm [/bib_ref] [bib_ref] Impact of a quality improvement intervention on provider adherence to recommended standards..., Umar-Kamara [/bib_ref] In general, more patients reached target for LDL-C than for HbA 1c levels, with the poorest achievement of targets being BP. The widest variability of target achievement was LDL-C (variation of 62.9%), followed by BP and then HbA 1c (least variability). The highest and lowest achieved targets were those by an American (LDL-C, 82.9%) and a German study (BP, 7.4%), respectively. [bib_ref] Pharmacist management of patients with diabetes mellitus enrolled in a rural free..., Sease [/bib_ref] [bib_ref] Diabetes care and complications in primary care in the Tshwane district of..., Webb [/bib_ref] discussion The quality of diabetes care cannot simply be measured across proportions of patients achieving guideline targets. However, a broad overview of the quality of care can be gauged when comparing target adherence across different countries, especially with adequately sized samples of patients. Hence the reason for this review, where countries from various economies were compared, according to achievement of modifiable risk factors against the guidelines. Based on the results of other studies, this review set out to establish the achievement of major risk-factor targets (HbA 1c , BP and LDL-C levels) in the treatment of DM patients in different parts of the world.
Given the increasing prevalence of T2DM, effective management of critical diabetes risk factors can significantly contribute towards improved outcomes. Attaining targets requires improved methods to increase adherence to lifestyle (exercise/diet) and pharmacological interventions. From this review, it was evident that certain studies appeared to be more successful at managing patients' risk factors than others.
Practitioners achieving better guideline adherence should be encouraged to share their management strategies for implementation with other healthcare facilities. Hermans et al. found that by benchmarking the level of care of 'three paramount cardiovascular risk factors' in a primary care setting has in itself led to a clinically significant improvement in T2DM care over time. [bib_ref] Benchmarking is associated with improved quality of care in type 2 diabetes:..., Hermans [/bib_ref] There is also evidence to suggest that performance with regard to management of a disease, when compared between a physician and his/her colleagues, has brought about an intellectual, emotional and competitive incentive for change. [bib_ref] OPtimal Type 2 dIabetes Management Including benchmarking and Standard trEatment) International Steering..., Nobels [/bib_ref] The most critical ways of reducing T2DM complications is by collectively managing HbA 1c , BP and LDL-C levels. More patients achieved LDL-C than HbA 1c targets in the studies reviewed, potentially owing to the progressive nature of the disease, where β-cell function gradually declines over time. BP control was the least-achieved risk factor across all the studies, and according to McLean et al., may have occurred due to the 'inadvertent under emphasis' of treating T2DM-associated risk factors (such as hypertension, when there is strong emphasis on glucose control). [bib_ref] Treatment and blood pressure control in 47,964 people with diabetes and hypertension:..., Mclean [/bib_ref] Perhaps it was due to inadequate dosages, poor adherence to medication, poor access to follow-up care or a combination of these. A well-designed, randomised, controlled trial may help address these questions.
Once considered rare in sub-Saharan Africa, the prevalence of T2DM is rapidly increasing. As many as four out of every five diabetics reside in LMICs, many of whom remain undiagnosed. 1 T2DM is a complex, resource-intensive disease requiring multifactorial yet individually tailoured, lifelong treatment.
Most of the studies found and included in this review were from higher-income countries. However patterns of poor control rates were common across all settings. For instance, less than 40% of patients from the USA, Europe (specifically Italy) and the UK studies (all high-income countries) achieved HbA 1c levels (< 7%) comparable with those of lower-to upper-middleincome countries (Uganda and South Africa, respectively). [bib_ref] Suboptimal glycaemic and blood pressure control and screening for diabetic complications in..., Kibirige [/bib_ref] [bib_ref] Pharmacist management of patients with diabetes mellitus enrolled in a rural free..., Sease [/bib_ref] [bib_ref] Diabetes care and complications in primary care in the Tshwane district of..., Webb [/bib_ref] [bib_ref] Quality of care of people with type 2 diabetes in eight European..., Stone [/bib_ref] Similarly, the combined results of six European countries, and other individual studies, had less than half of patients at LDL-C target, as seen in two separate non-high-income countries. [bib_ref] Benchmarking is associated with improved quality of care in type 2 diabetes:..., Hermans [/bib_ref] [bib_ref] Suboptimal glycaemic and blood pressure control and screening for diabetic complications in..., Kibirige [/bib_ref] [bib_ref] Quality of care of people with type 2 diabetes in eight European..., Stone [/bib_ref] Yet on the other hand, and possibly as expected from moredeveloped nations, two to three times more patients from separate European (specifically the Netherlands) and a USA study achieved HbA 1c (< 7%) and LDL-C (< 2.6mmol/l) targets in comparison with a lower-income country, respectively. [bib_ref] Suboptimal glycaemic and blood pressure control and screening for diabetic complications in..., Kibirige [/bib_ref] [bib_ref] Diabetes care and complications in primary care in the Tshwane district of..., Webb [/bib_ref] [bib_ref] Impact of a quality improvement intervention on provider adherence to recommended standards..., Umar-Kamara [/bib_ref] The differences across the sites in their abilities to achieve guideline targets may be attributed to socio-economic reasons. In resource-rich settings, where patients supposedly receive the extra time required for diabetes care through more regular physician interactions or appointments, appropriate reminder systems and adherence monitoring, this may improve the standards of diabetes care received. Lower-income countries face the realities of inadequate healthcare infrastructure, regular medication stock outs, few educational programmes and minimal healthcare facilities/professionals. [bib_ref] Diabetes trends in Africa, Motala [/bib_ref] This literature review covered the influences of multiple background factors occurring across healthcare systems in different countries, hence the differences in targets achieved across the environments studied.
As described above, Africa faces many healthcare challenges, both within and between countries. Despite resource constraints, by targeting the modifiable risk factors associated with DM, there is still the potential for improvement, and better patient outcomes. This review serves to highlight the proportion of patients achieving guideline targets across different settings. The aim of this review was to serve as a benchmark for those countries selected, in order to measure their performance against each other in terms of achieving guideline targets.
By recognising those healthcare settings with increased patient numbers achieving guideline targets, this could allow for future studies to identify the mechanisms and processes used to achieve their targets. Areas of interest for the improvement of diabetes care could include: organisational characteristics such as improved implementation of adherence to clinical guidelines (evidence-based), identification of individuals to act as guideline champions to deliver more performance measures, and feedback to healthcare providers on progress made. Perhaps, once identified, the settings achieving less-favourable control of modifiable risk factors may begin to explore approaches used in the more successful settings. In addition, given the chronic progressive nature of DM, it is hoped that attention will be prioritised not just on treatment but also on prevention strategies in those settings wishing to improve their level of diabetes care offered.
It has been predicted that the ageing populations of LMICs will face a significant increase in mortality rate due to NCDs over the next 25 years.Although not included in this review, a previous South African study revealed that only 30.4% of the 899 patients achieved HbA 1c levels < 7%, which is similar to the three studies included in this review from the same locale. [bib_ref] Diabetes guidelines and clinical practice: Is there a gap? The South African..., Amod [/bib_ref] The three South African studies included in this review had noticeably fewer patients at HbA 1c goal in comparison with other countries. One of the reasons for this may have been that all the South African studies included in this review were from the public sector, which has often been described as 'overburdened', and due to resource constraints, cannot always offer appropriate levels of healthcare or access to the most modern diabetes treatments currently available in South Africa's private sector or in those high-income countries included in this study. Furthermore, many of the patients serviced in South Africa's public sector settings originate from a lower socioeconomic background, which may indicate lower educational levels or limited access to healthier lifestyle choices. Perhaps HbA 1c level is still the most challenging of risk factors to control, especially in less-developed economies.
# Study limitations
Studies included in this review differed with regard to guidelines or targets, however, we tried to overcome this by consistently capturing and comparing similar risk factors using standardised units. Patients selected for the studies included in the review were treated to 'moving targets' and therefore studies conducted in the past five years only (2009-2014) were included, as recent guidelines have more similar targets.
Some eight out of the 14 studies provided information on the type of methods used for clinical and laboratory-based measurements, but with differing levels of detail. Notably, only two studies confirmed use of DCCT-standardised laboratory analysers for HbA 1c analysis. By converting to similar units (e.g. LDL-C from mg/dl to mmol/l), target values in this retrospective study [fig_ref] table 1: study characteristics [/fig_ref] were presented in a format that would allow for comparison. Ideally, a centralised laboratory should have been used for measurements included in this study, however, we relied on previously obtained measurements from other studies. We therefore cannot guarantee the accuracy or precision of measurements in the studies selected for this review, as methodologies may have differed. Studies from resourcerich settings may have implemented newer, more sophisticated and improved methods for A 1c and LDL-C measurements, influencing the results of those specific studies.
This review did not stratify the selected individual studies according to patient profiles, severity of disease, clinical settings (clinic or hospital) or involvement of specialists (factors affecting how individuals are managed and able to reach guideline targets). Although smoking is considered a critical risk factor in the prevention and management of CVD, it was not included as a crucial study parameter for this review (partly due to many studies not reporting this).
Although previously identified as a source of bias, we only included studies published in English, which according to an analysis, has little effect on summaries of treatment effect estimates. [bib_ref] Direction and impact of language bias in meta-analyses of controlled trials: empirical..., Jüni [/bib_ref] Publication bias may have occurred in our study in that a single reviewer (author) carried out the searches without the use of specific methodology (e.g. Cochrane data system).
# Conclusion
The results presented in this study demonstrate that T2DM patients remain inadequately controlled for their cardiovascular risk factors. Our review revealed that control of major risk factors did not differ significantly between countries or healthcare settings. There is substantial room for improvement in the way T2DM patients are being managed for their condition. Further efforts through multidisciplinary action to improve guideline adherence is critical for the prevention or delay of diabetesrelated complications.
[table] table 1: study characteristics [/table]
|
Direct purification of detergent-insoluble membranes from Medicago truncatula root microsomes: comparison between floatation and sedimentation
Background: Membrane microdomains are defined as highly dynamic, sterol-and sphingolipid-enriched domains that resist to solubilization by non-ionic detergents. In plants, these so-called Detergent Insoluble Membrane (DIM) fractions have been isolated from plasma membrane by using conventional ultracentrifugation on density gradient (G). In animals, a rapid (R) protocol, based on sedimentation at low speed, which avoids the time-consuming sucrose gradient, has also been developed to recover DIMs from microsomes as starting material. In the current study, we sought to compare the ability of the Rapid protocol versus the Gradient one for isolating DIMs directly from microsomes of M. truncatula roots. For that purpose, Triton X-100 detergent-insoluble fractions recovered with the two methods were analyzed and compared for their sterol/sphingolipid content and proteome profiles.Results: Inferred from sterol enrichment, presence of typical sphingolipid long-chain bases from plants and canonical DIM protein markers, the possibility to prepare DIMs from M. truncatula root microsomes was confirmed both for the Rapid and Gradient protocols. Contrary to sphingolipids, the sterol and protein profiles of DIMs were found to depend on the method used. Namely, DIM fractions were differentially enriched in spinasterol and only shared 39% of common proteins as assessed by GeLC-MS/MS profiling. Quantitative analysis of protein indicated that each purification procedure generated a specific subset of DIM-enriched proteins from Medicago root microsomes. Remarkably, these two proteomes were found to display specific cellular localizations and biological functions. In silico analysis of membrane-associative features within R-and G-enriched proteins, relative to microsomes, showed that the most noticeable difference between the two proteomes corresponded to an increase in the proportion of predicted signal peptide-containing proteins after sedimentation (R) compared to its decrease after floatation (G), suggesting that secreted proteins likely contribute to the specificity of the R-DIM proteome. Conclusions: Even though microsomes were used as initial material, we showed that the protein composition of the G-DIM fraction still mostly mirrored that of plasmalemma-originating DIMs conventionally retrieved by floatation. In parallel, the possibility to isolate by low speed sedimentation DIM fractions that seem to target the late secretory pathway supports the existence of plant microdomains in other organelles.
# Background
Biological membranes that compartmentalize cells into organelles or form a barrier to the outside environment are composed of lipids as well as a variety of trans-membrane, lipid-modified and lipid-associated proteins essentially involved in transport, signaling, differentiation and stress adaptation processes. Aside from the fluid mosaic model that refers to a homogenous distribution of lipids and proteins within the plasma membrane (PM) [bib_ref] Trehalose 6-phosphate is indispensable for carbohydrate utilization and growth in Arabidopsis thaliana, Schluepmann [/bib_ref] , a large body of evidence supports the microdomain hypothesis [bib_ref] Functional rafts in cell membranes, Simons [/bib_ref] , stating that membranes are also compartmentalized by uneven distributions of specific lipids and proteins into microdomains termed membrane rafts. Originally characterized in animal and yeast cells, membrane rafts are defined as plasma membrane [bib_ref] Trehalose 6-phosphate is indispensable for carbohydrate utilization and growth in Arabidopsis thaliana, Schluepmann [/bib_ref] nano-or microdomains enriched in sphingolipids and sterols, which act as platforms initiating signaling events in diverse physiological situations, including inflammation processes and apoptotic cell death [bib_ref] Lipid raft involvement in yeast cell growth and death, Mollinedo [/bib_ref]. The main hypothesis relative to the functional significance of these domains relies on the lateral segregation of membrane proteins that creates a dynamic scaffold to organize particular cellular processes [bib_ref] Plasticity of plasma membrane compartmentalization during plant immune responses, Urbanus [/bib_ref]. In plants, sphingolipid-and sterol-enriched membrane microdomains were also isolated from PM. Characterization of their protein content revealed their enrichment in proteins involved in signaling and response to biotic/abiotic stresses [bib_ref] Lipid rafts in higher plant cells: purification and characterization of Triton X-100-insoluble..., Mongrand [/bib_ref] [bib_ref] New insights into mathematical modeling of the immune system, Morel [/bib_ref] [bib_ref] Simon-Plas F: Quantitative proteomics reveals a dynamic association of proteins to detergent-resistant..., Stanislas [/bib_ref] [bib_ref] An update on plant membrane rafts, Simon-Plas [/bib_ref] , suggesting that plant membrane microdomains may exert similar signaling functions to their animal counterparts.
Due to their enrichment in sphingolipids and sterols, membrane rafts form tight packing liquid-ordered (Lo) phases that segregate from the rest of the PM. An increased resistance to solubilization by detergents of Lo versus liquid-disordered (Ld) phases has led researchers to consider that membrane fractions insoluble to nonionic detergents at low temperatures could contain the putative raft fractions. One caveat of this theory is that recovered detergent-insoluble membrane (DIM) fractions only exist after detergent treatment, and do not correspond to the native membrane structure [bib_ref] Detergent-resistant membranes should not be identified with membrane rafts, Lichtenberg [/bib_ref]. Nevertheless, their significant enrichment in sterols, sphingolipids and specific subsets of proteins, some of which displaying a clustered distribution within the PM [bib_ref] Revitalizing membrane rafts: new tools and insights, Simons [/bib_ref] , has encouraged their use as a biochemical counterpart of Lo microdomains existing in biological membranes. From an experimental perspective, upon detergent application to PM-enriched preparations, DIM fractions are usually purified by ultracentrifugation onto a sucrose gradient and appear as a ring floating at low density, which are structurally represented by vesicles and membranes sheets [bib_ref] Lipid rafts in higher plant cells: purification and characterization of Triton X-100-insoluble..., Mongrand [/bib_ref]. Initially, microdomains were thought to be exclusively present in PM and membranes belonging to the late secretory pathway [bib_ref] Contributions of quantitative proteomics to understanding membrane microdomains, Zheng [/bib_ref]. As indicated in [fig_ref] Table 1: Main literature background to microdomain preparations as related to initial fractions [/fig_ref] , most of DIM preparations were indeed carried out using PM-enriched fractions as starting material [bib_ref] Lipid rafts in higher plant cells: purification and characterization of Triton X-100-insoluble..., Mongrand [/bib_ref] [bib_ref] New insights into mathematical modeling of the immune system, Morel [/bib_ref] [bib_ref] Simon-Plas F: Quantitative proteomics reveals a dynamic association of proteins to detergent-resistant..., Stanislas [/bib_ref] [bib_ref] Identification of low-density Triton X-100-insoluble plasma membrane microdomains in higher plants, Peskan [/bib_ref] [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] [bib_ref] PAMP (pathogen-associated molecular pattern)-induced changes in plasma membrane compartmentalization reveal novel components..., Keinath [/bib_ref] [bib_ref] Detergent-resistant plasma membrane proteome in oat and rye: similarities and dissimilarities between..., Takahashi [/bib_ref] , thus hampering their identification within other cell membranes. The presence of raft-like regions within organelles was nonetheless further suggested to occur upon the characterization of DIMs extracted from membranes of Golgi complex [bib_ref] Insights into the role of specific lipids in the formation and delivery..., Laloi [/bib_ref] , mitochondrion [bib_ref] Proteomic analysis of lipid raft-enriched membranes isolated from internal organelles, Poston [/bib_ref] and vacuole [bib_ref] Tonoplast of Beta vulgaris L. contains detergent-resistant membrane microdomains, Ozolina [/bib_ref] [bib_ref] Studies on vacuolar membrane microdomains isolated from Arabidopsis suspension-cultured cells: local distribution..., Yoshida [/bib_ref]. To date, the widest investigation addressing the intracellular distribution of plant DIMs has been performed in Arabidopsis using whole cell membranes originating from liquid root callus cultures [bib_ref] Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts, Borner [/bib_ref]. Noteworthy, the results obtained strongly suggested that in A. thaliana roots, DIMs are predominantly derived [bib_ref] PAMP (pathogen-associated molecular pattern)-induced changes in plasma membrane compartmentalization reveal novel components..., Keinath [/bib_ref] Oat and Rye Leaves Microsomes PM Triton X-100 F [bib_ref] Detergent-resistant plasma membrane proteome in oat and rye: similarities and dissimilarities between..., Takahashi [/bib_ref] Red beet Roots Vacuoles Tonoplast Triton X-100 F [bib_ref] Tonoplast of Beta vulgaris L. contains detergent-resistant membrane microdomains, Ozolina [/bib_ref] Arabidopsis Cell cultures Vacuoles Tonoplast Triton X-100 F [bib_ref] Studies on vacuolar membrane microdomains isolated from Arabidopsis suspension-cultured cells: local distribution..., Yoshida [/bib_ref] Organelle versus microsomes and DIM recovery processes: floatation on sucrose gradient (F) versus sedimentation (S). GA, ER, Mmito and PM, and refer to Golgi apparatus, endoplamic reticulum, mitochondrial membrane and plasma membrane, respectively. Bold characters highlight the two protocols used in the current study.
from PM sphingolipid-and sterol-rich microdomains by virtue of their substantial depletion of intracellular organelle proteins. Whether this result also holds true for plants of agronomic has not been investigated yet, despite the recognized importance of membrane microdomains during plant-microbe interactions (reviewed in [bib_ref] Plasticity of plasma membrane compartmentalization during plant immune responses, Urbanus [/bib_ref] [bib_ref] An update on plant membrane rafts, Simon-Plas [/bib_ref]. Although Medicago truncatula has been retained more than ten years ago as the model for studying legumes and root symbiotic interactions with fungi and bacteria [bib_ref] Proteome analysis and identification of symbiosis-related proteins from Medicago truncatula Gaertn. by..., Bestel-Corre [/bib_ref] , only one report has been dedicated to the analysis of DIM fractions in barrel medic [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref]. The study showed that membrane raft domains corresponding to Triton X-100 insoluble membranes could be obtained from M. truncatula root PM. Additionally, evidence was given for their enrichment in proteins associated with signaling, cellular trafficking and redox processes. A raft protein termed Symbiotic REM (MtSYMREM1, or MtREM2.2) [bib_ref] Genome-wide annotation of remorins, a plant-specific protein family: evolutionary and functional perspectives, Raffaele [/bib_ref] was also found to control Sinorhizobium meliloti infection as well as rhizobial release into host cell cytoplasm within root symbiotic structures, the so-called nodules [bib_ref] A remorin protein interacts with symbiotic receptors and regulates bacterial infection, Lefebvre [/bib_ref]. Likewise, Haney and Long [bib_ref] Plant flotillins are required for infection by nitrogenfixing bacteria, Haney [/bib_ref] identified two microdomainassociated plant flotillins required for infection by nitrogen-fixing bacteria. These data raise the possibility that rafts may be involved in molecular events leading to successful nodule onset, and it is tempting to speculate that additional symbiotic associations like mycorrhiza may also require proper raft structures for their establishment and functioning. Elucidating microdomain function(s) in symbiosis and legume physiology thereby implies increasing knowledge about their cellular distribution coupled to fast and efficient methods dedicated to their isolation.
Although DIM fractions have been successfully prepared from M. truncatula root tissues using PM as starting material [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] , this protocol requires a huge amount of root tissues. Additionally, purifying PM fractions turns out to be somehow labor-intensive and timeconsuming. To overcome these technical limitations together with enlarging the coverage of DIM populations in legume roots, we investigated in the current study an alternative that relies on the possibility to skip the PM fractionation step, to isolate DIM fractions directly from microsomes, as previously described in other animal and plant model systems [fig_ref] Table 1: Main literature background to microdomain preparations as related to initial fractions [/fig_ref]. This work was thus intended to purify microdomains directly from M. truncatula root whole cell membranes by comparing two fast protocols previously described for DIM purification [bib_ref] Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts, Borner [/bib_ref] [bib_ref] Rapid preparation of nuclei-depleted detergent-resistant membrane fractions suitable for proteomics analysis, Adam [/bib_ref]. Using roots of soil-grown M. truncatula plants as starting material, we first analyzed the impact of detergent final concentration and detergent/protein ratio on lipid and protein patterns of DIM fractions. We then selected specific experimental conditions and used a GeLC-MS/MS proteomic approach, where biological samples are separated by SDS-PAGE, sliced, digested in-gel and analyzed by LC-MS/MS, on the DIM fractions retrieved from the two distinct protocols. Respective DIM protein populations were further contrasted with regard to their functional and cellular distributions.
# Results and discussion
Purification of DIMs from M. truncatula root microsomes
In the current study, whole root cell membranes from soil-grown plants were first extracted according to the differential centrifugation-based strategy initially developed for Nicotiana tabacum cell cultures [bib_ref] Simon-Plas F: Quantitative proteomics reveals a dynamic association of proteins to detergent-resistant..., Stanislas [/bib_ref]. DIMs were further isolated from the root microsomal fraction according to two distinct protocols. The former developed by Adam and collaborators [bib_ref] Rapid preparation of nuclei-depleted detergent-resistant membrane fractions suitable for proteomics analysis, Adam [/bib_ref] consists of a rapid method for purifying DIM fractions from human cells by low speed sedimentation that exploits the differential solubility of detergent-resistant microdomains in cold, non-ionic detergents. Briefly, upon cell mechanical disruption, the authors directly treated homogenates with cold Triton X-100 (TX-100) and centrifuged samples to recover detergent-insoluble material in the pellet. These DIMs were then solubilized using β-octylglucoside as detergent and the resulting supernatant recovered after centrifugation. This procedure referred to as Rapid or Rprotocol, was compared to that used by Borner et al. [bib_ref] Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts, Borner [/bib_ref] , which is classical floatation of cell extracts in a sucrose density Gradient (G), as illustrated in [fig_ref] Figure 1: Overview of the Rapid and Gradient protocols used for isolating DIM fractions... [/fig_ref]. The latter, initially carried out using Arabidopsis callus cultures, relies on the light buoyant density of TX-100insoluble microsomal membranes. R-and G-DIM subsets were thus prepared as explained in the section "Methods" and subsequently analyzed for their lipid and protein composition relative to the original microsomal fraction. Additionally, considering that R-and G-DIM extraction methods relied on the use of distinct TX-100/ protein ratios and TX-100 final concentrations (R3:1 and G3:2, respectively), the effects of detergent-to-protein ratio (w/w) and detergent final concentration (% v/v) on lipid and protein composition were also investigated.
Sterols, but not sphingolipids, are differentially enriched between R-and G-DIM fractions Three independent experiments were performed and both R-and G-DIM fractions were examined for their lipid content in order to assess enrichment in sterols and sphingolipids, a typical feature of membrane microdomains. Sterol composition was first determined by gas chromatography (GC) using epichoprostanol as internal standard. [fig_ref] Figure 2: Comparison of sterol and long-chain base [/fig_ref] shows a representative elution profile for R-and G-DIM fractions, relative to the microsomal (Mic) set used as starting material for DIM purification. In accordance with previous data [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] , spinasterol was recorded as the most abundant sterol in the two DIM fractions, but average enrichment-fold in spinasterol increased from 2.8 to 3.9 in R-and G-DIMs, respectively [fig_ref] Figure 2: Comparison of sterol and long-chain base [/fig_ref]. Due to the distinct TX-100/protein ratios and TX-100 final concentrations published for R-and G-DIM extractions, spinasterol content was thus quantified in relation to these two parameters. At identical TX-100/protein ratios and final TX-100 concentrations, significant differences in spinasterol enrichment were still registered between R-and G-DIM fractions (Additional file 1: [fig_ref] Figure 1: Overview of the Rapid and Gradient protocols used for isolating DIM fractions... [/fig_ref]. These results clearly indicated that respective purification steps of R-and G-methods, i.e. low speed centrifugation versus sucrose gradient, were responsible for differences in spinasterol concentration, irrespective of TX-100-related parameters.
As long-chain base (LCB) represent a common backbone to all sphingolipids, they were quantified by GC-MS [bib_ref] Rapid nanoscale quantitative analysis of plant sphingolipid long-chain bases by GC-MS, Cacas [/bib_ref] as a way to access the total enrichment in sphingolipids in R-and G-DIM fractions. Whatever the method used for DIM preparation, the resulting total LCB composition [fig_ref] Figure 2: Comparison of sterol and long-chain base [/fig_ref] was consistent with previously data reported for M. truncatula [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] , even though there was evidence for additional minor dihydroxylated LCB (d18:0, d18:1 and d18:2), the detection of which was previously ascribed to the high sensitivity of GC-MS [bib_ref] Rapid nanoscale quantitative analysis of plant sphingolipid long-chain bases by GC-MS, Cacas [/bib_ref]. Interestingly, both R-and G-fractions were highly enriched in trihydroxylated LCB (c.a. 6-fold increase in t18:0 and t18:1 when compared to Mic). These compounds are mainly found amidified in the sphingolipid class of glycosylinositolphosphoryl-ceramides [bib_ref] Lipids of plant membrane rafts, Cacas [/bib_ref]. Additionally, R-and G-samples also exhibited a very similar LCB profile with identical enrichment-folds whatever the TX-100 concentration used (Additional file 1: Additional A1B), strongly suggesting that sphingolipid content is not dependent on the method used for DIM isolation. Overall, the lipid composition of R-and G-DIMs confirmed their enrichment in sphingolipids and sterols relative to the microsomal fraction.
## Dim protein composition is impacted by the extraction method
Due to differences in the original setups for TX-100 concentrations between R and G protocols, the effects of detergent concentration and detergent/protein ratio (ratio detergent/protein = 3 to 6 and final concentrations 1 to 2) on DIM protein composition were also preliminary assessed on the basis of one-dimensional SDS-PAGE banding patterns visualized following Coomassie blue staining. As displayed in , the protein profiles obtained for R-DIM samples looked different from the microsomal fraction from which they originated, but roughly qualitatively similar in the conditions of interest (R3:1 and R3:2). Despite some minor differences, increasing the ratio d:p to 6:1 did not change drastically the protein pattern. These observations also hold true for G-DIM samples. By contrast, there were noticeable qualitative and quantitative differences in protein patterns between R-and G-fractions, indicating that in our experimental conditions DIM protein contents largely depends on the isolation process rather than detergent concentration.
To go further in analyzing and comparing the proteins co-extracted with sterol-enriched DIM fractions, a shotgun proteomic approach was performed using the original setup for TX-100 concentrations, namely R3:1 and G3:2 conditions after admitting the detergent-independence of DIM lipid and protein composition over this range. Due to the limitations of two-dimensional electrophoresis to resolve integral membrane proteins, 1D gel coupled to liquid chromatography-tandem mass spectrometry (GeLC-MS/MS) was chosen to investigate the protein composition of DIM fractions. This workflow that combines a size-based protein separation to an in-gel digestion of the resulting fractions proved to be successful in expanding the coverage of membrane proteins in M. truncatula roots [bib_ref] A mass spectrometric approach to identify arbuscular mycorrhiza-related proteins in root plasma..., Valot [/bib_ref] , and is also amenable to relative protein quantification methods such as spectral counting [bib_ref] Stable isotope-free quantitative shotgun proteomics combined with sample pattern recognition for rapid..., Wienkoop [/bib_ref].
GeLC-MS/MS was thus conducted on two independently-extracted sets of R-and G-DIMs and the initial root microsomal fraction. Using a probability of peptide misidentification inferior to 0.05, a total of 874 non redundant proteins were overall identified in the microsomal and DIM fractions when retaining only those co-identified in the two replicates of each DIM, as listed in additional data (Additional file 2: [fig_ref] Table 1: Main literature background to microdomain preparations as related to initial fractions [/fig_ref]. The Venn diagram distribution of microsomal, R-and G-DIM proteins, displayed in , indicated that relative to the 821 accessions initially identified in the microsomal fraction, R-and G-DIMs encompassed a rather similar number of proteins corresponding to 234 and 219 accessions, respectively. Although most of DIM-associated proteins (84%) were as expected also present in the original microsomal fraction, 53 accessions (16%) were uniquely identified in DIMs, indicating that the experimental procedure has enabled the identification of minor proteins that have escaped detection during mass analysis of whole membranes but are revealed upon fractionation. Noticeably, comparison of R-and G-DIMs showed that a common pool of 126 proteins was shared between both fractions, thereby defining a conserved core-set of DIM-associated proteins that overall represented 15% of the root microsomal proteome of M. truncatula.
To investigate whether there might be a difference in the quantitative distribution of these common proteins between R-and G-DIMs, protein abundance was estimated using spectral counting, which is based on the cumulative sum of recorded peptide spectra that can match to a given protein [bib_ref] Statistical analysis of membrane proteome expression changes in Saccharomyces cerevisiae, Zybailov [/bib_ref]. Following the calculation of a normalized spectral abundance factor (NSAF) value for each protein across the four replicates, only six proteins displayed a significant (p < 0.05) differential accumulation between R-and G-DIMs . Namely, a mitochondrial import receptor subunit TOM40 homolog, a fasciclin-like arabinogalactan protein and a hexokinase displayed a higher abundance in R-DIMs than in G-DIMs, whereas an elongation factor 1-alpha, a V-type proton ATPase subunit H and an asparagine synthetase over-accumulated in G-DIMs relative to R-DIMs. As a result, the 126 proteins shared between both fractions largely corresponded to a quantitatively conserved set of DIM-associated proteins irrespective of the extraction method, in which the top 10 major abundant proteins included transmembrane porins (aquaporins, OMP), respiratory chain related proteins (ATP synthases, flavoprotein), beta-glucosidase G1 and ubiquitin, as very often described in plant DIM fractions (Additional file 3: [bib_ref] Insights into the role of specific lipids in the formation and delivery..., Laloi [/bib_ref] [bib_ref] Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts, Borner [/bib_ref]. On the opposite, the Venn diagram also showed that out of the 234 and 219 proteins identified in R-and G-fractions, 108 (46%) and 93 (42%) proteins were unique to R-and G-DIM, respectively . This pointed out that 61% (201 proteins) of the 327 DIM-associated proteins in M. truncatula roots underwent a differential partition according to the purification procedure. Consequently, even though both approaches had equivalent protein extraction efficiencies, as inferred from the similar number of accessions identified in R-and G-fractions, they nonetheless displayed a differential selectivity toward microsomal proteins.
(See figure on previous page.) Comparison of the proteins identified in R-and G-DIMs relative the initial microsomal (Mic) fraction of M. truncatula roots. (A) One dimensional profile of the proteins (15 μg per lane) recovered in R (R) and G-DIMs (G) using variable Triton X-100 concentrations: 3:1, 3:2, 6:1 and 6:2 (detergent/protein ratio: final detergent concentration). (B) Venn diagram distribution of the 874 non redundant proteins overall identified using GeLC-MS/MS in the microsomal, R-and G-DIM fractions. (C) List of the proteins that display a differential accumulation (p < 0.05) between R-and G-DIMs. Comparison of protein abundance was performed using the Student's t-test on arsin square root-transformed normalized spectral abundance factors (NSAF). NSAF ratios of proteins between the two DIM fractions are provided in column 2. (D) Venn diagram distribution of the 227 proteins that reproducibly display at least a 2-fold higher abundance in DIM fractions than in microsomes. Subsets termed "R2xspecific" and "G2xspecific" refer to the proteins uniquely enriched in R-and G-DIMs, respectively, whereas "RG2xcore" designates the proteins enriched in both R-and G-DIMs, relative to microsomes. (E) Representation of previously published plant DIM-associated proteins within the proteins enriched in R-and G-DIMs relative to microsomes, by using identification mapping tools and homology search. Bold characters refer to canonical plant DIM markers.
## Dim-enriched proteins differ between r-and g-fractions
To further assess the extent to which protein composition of R-and G-DIMs quantitatively differed from that of initial microsomes, an abundance ratio between NSAF values of DIM and Mic fractions was calculated for each protein. On this basis, accessions that reproducibly displayed at least a 2-fold higher abundance in R-and G-DIMs than in microsomes were considered as DIMenriched proteins according to Borner's sensu. Among them, 65 were unique to R-DIMs (fraction termed "R2xspecific") and 46 were unique to G-DIMs (fraction termed "G2xspecific"), whereas 42 were shared between R-and G-DIMs (fraction termed "RG2xcore") . From these results, it was thus concluded that each extraction procedure generated a specific subset of DIMenriched proteins from Medicago root microsomes, which accounted for 7.4 and 5.3% of the initial 874 identifications, for R-and G-protocols, respectively. The rest of study was thus essentially dedicated to the comparison of these two specific proteomes and the core subset, relative to the microsomal fraction.
When investigating the representation of previously published plant DIM-associated proteins within our proteomic data by using identification mapping tools and homology search against the protein listed in [bib_ref] New insights into mathematical modeling of the immune system, Morel [/bib_ref] [bib_ref] Simon-Plas F: Quantitative proteomics reveals a dynamic association of proteins to detergent-resistant..., Stanislas [/bib_ref] [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] [bib_ref] Detergent-resistant plasma membrane proteome in oat and rye: similarities and dissimilarities between..., Takahashi [/bib_ref] [bib_ref] Recruitment and interaction dynamics of plant penetration resistance components in a plasma..., Bhat [/bib_ref] [bib_ref] Definition of Arabidopsis sterol-rich membrane microdomains by differential treatment with methyl-betacyclodextrin and..., Kierszniowska [/bib_ref] and [bib_ref] Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts, Borner [/bib_ref] , 152 proteins already described in plant microdomains were identified within the total 327 R-and G-proteins, including 33 proteins usually referred to as canonical plant DIM markers in the literature such as remorin (Additional file 3: . Noticeably, 14 DIM markers were overall identified within DIM-enriched Medicago proteins, which encompassed fasciclin-like arabinogalactan proteins, hedgehog-interacting protein, receptor-like kinases, 14-3-3 like protein, phospholipase D, dynamins, and flotillin. However, their distribution remarkably differed between R2xspecific and G2xspecific subsets , thereby comforting the view that R-and G-approaches displayed a differential selectivity toward certain classes of proteins.
Finally, to address whether known or putative nonmembrane proteins might be enriched in R-and G-DIM fractions, we used, as a point of reference for M. truncatula, the rationale described by Daher and co-workersthat favors similarity search on the basis of which homologous proteins share the same location in many organisms, a strategy recognized more confident than the use of in silico algorithmic predictors for protein localization [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref]. Consequently, DIM-enriched proteins obtained from Rand G-protocols were first compared with BLASTP to TAIR database accessions and were considered as membrane M. truncatula proteins when homologous sequences displaying at least 70% pair-wise identity and a cut-off expectation value of e −40 were experimentally demonstrated to have a membrane localization, including core integral or subunits of membrane complexes, on the basis of direct assays. In the absence of TAIR homologues, LegumIP annotations that overall agreed up to 80% with Arabidopsis-inferred cellular components, even though largely less detailed, were used to address protein localization (Additional file 3: . In the absence of confident membrane homologues, DIM-enriched proteins were retained as non-membrane proteins unless predicted to display at least one of the following criteria: to form an alpha helical TM domain or a beta barrel embedded in the membrane lipid bilayer, to be anchored to the membrane owing to hydrophobic tails, and/or to be targeted to the secretory pathway, as previously described [bib_ref] Contributions of quantitative proteomics to understanding membrane microdomains, Zheng [/bib_ref] [bib_ref] Greasing their way: lipid modifications determine protein association with membrane rafts, Levental [/bib_ref]. Using this design, 10 accessions mainly of cytosolic origin, out of the total 227 proteins previously recorded as DIMenriched were identified as potential contaminants of membrane fractions (Additional file 3: . However, when considering their known or putative functional relevance in microdomain formation with special regard to role in mediating hydrophobic interactions and/or responses to microbial ingress/accommodation at the interface of plant-microbe interactions that largely depend on exocytocis, endocytosis, or local secretion of defense compounds [bib_ref] Transport and secretion in plant-microbe interactions, Huckelhoven [/bib_ref] , we made the deliberate choice not to discard them from R-and G-DIM fractions. Namely, patellin-5 binds to hydrophobic molecules such as phosphoinositides and promotes their transfer between different cellular sites. The PLAT/LH2 family domain of lipase/lipoxygenase is found in a variety of membrane or lipid associated proteins, and dynein transports various cellular cargo by walking along cytoskeletal microtubules. Ubiquitin, linkage of which to PM proteins is known to induce endocytosis and/or proteasome-dependent degradation [bib_ref] Lipid rafts in higher plant cells: purification and characterization of Triton X-100-insoluble..., Mongrand [/bib_ref] , whereas caffeic acid 3-O-methyltransferase is involved in the reinforcement of the plant cell wall and in the responding to wounding or pathogen challenge by the increased formation of cell wall-bound ferulic acid polymers. Major latex proteins belong to cytokinin-specific binding proteins that also have role in pathogen defense responses. Sorting and assembly machinery component 50 (cell division protein FtsZ homolog) is part of a ring in the middle of the dividing cell that is required for constriction of cell membrane/cell envelope and localizes to very-long chain fatty acids-containing phospholipids that have an important role in stabilizing highly curved membrane domains [bib_ref] Detergent-resistant plasma membrane proteome in oat and rye: similarities and dissimilarities between..., Takahashi [/bib_ref] [bib_ref] Identification and biophysical characterization of a very-long-chain-fatty-acid-substituted phosphatidylinositol in yeast subcellular membranes, Schneiter [/bib_ref]. Finally, glycoprotein-binding proteins (lectins) have been suggested to contribute to stimulus-dependent microdomain assemblies via cross-linking of PM-resident proteins [bib_ref] Recruitment and interaction dynamics of plant penetration resistance components in a plasma..., Bhat [/bib_ref] [bib_ref] Lipid rafts and apical membrane traffic, Fullekrug [/bib_ref].
Taken together, the above data confirmed that both the Rapid (R) and Gradient (G) protocols enabled the isolation of microdomain fractions directly from M. truncatula root microsomes, as inferred from sterol enrichment, presence of typical sphingolipid long-chain bases from plants, enrichment in membrane proteins including well-known plant DIM reporters, but also showed that the method used for DIM extraction, namely low-speed centrifugation versus floatation, qualitatively impacted the composition of the proteome enriched in DIM fractions relative to initial microsomes. Consequently, to get a deeper insight regarding the processes by which DIM-enriched proteins may preferentially partition to either R-or G-DIM fraction, the corresponding M. truncatula proteins were further characterized with regard to their subcellular localization, functional relevance and features known to drive membrane association.
## R-and g-dim-enriched proteins differ in their cellular location
To analyze the subcellular localization of DIM-associated proteins of M. truncatula roots relative to the microsomal fraction, we used the above-described workflow that favors similarity searches over in silico predictions. Using these criteria, 23 different localizations were recorded for the 227 proteins enriched in the current DIM fractions, as detailed in Additional file 2: [fig_ref] Table 1: Main literature background to microdomain preparations as related to initial fractions [/fig_ref]. In this respect, because chloroplast-located proteins in roots refer to those belonging to non-photosynthetic plastids, they were further termed non-green plastid proteins. To minimize misinterpretation, these 23 localizations were restricted to 17 after combining when possible each membrane fraction to its counterpart organelle, as for example plastids with plastidial membranes. Actually, although each cellular compartment was experimentally checked, reference to whole organelle localization can also include its membrane residents when not specifically addressed in the corresponding study. To take into account the multiple cell locations that a protein very often inhabits according to TAIR and LegumeIP annotations (Additional file 3: , a subcellular profile was thereby drawn for each of the R, G and microsomal fractions of interest by plotting the rate of occurrence of each cellular component within the corresponding proteomic data sets, as displayed in [fig_ref] Figure 4: Cellular [/fig_ref].
Keeping in mind that each frequency does not refer to an exclusive subcellular component and that frequencies may be biased toward the most studied Arabidopsis and legume organelles, it nonetheless appeared from [fig_ref] Figure 4: Cellular [/fig_ref] that plasma membrane had the highest rate of occurrence within the RG2xcore proteome, a result that substantiates the view according to which the PM largely contributes to microdomain-enriched proteins [bib_ref] Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts, Borner [/bib_ref]. However, although proteins ascribed to mitochondrion were largely depleted in this core fraction relative to initial microsomes, as previously observed by Zheng et al. [bib_ref] Contributions of quantitative proteomics to understanding membrane microdomains, Zheng [/bib_ref] , those located to other cellular components such as cell wall and non-green plastids happened to be enriched in Medicago root DIMs. Consequently, the subcellular profile obtained for this core fraction agreed with the idea that besides the plasma membrane, DIMs can be extracted from several other cellular compartments, as essentially demonstrated before for endomembrane systems when analyzing organelleenriched fractions (Additional file 2: [fig_ref] Table 1: Main literature background to microdomain preparations as related to initial fractions [/fig_ref]. In this respect, whereas the presence of plant cell wall-related proteins within DIM fractions has been widely reported in the literature [bib_ref] Detergent-resistant plasma membrane proteome in oat and rye: similarities and dissimilarities between..., Takahashi [/bib_ref] [bib_ref] Definition of Arabidopsis sterol-rich membrane microdomains by differential treatment with methyl-betacyclodextrin and..., Kierszniowska [/bib_ref] , the retrieval of plastidial component in microdomains is far less documented. Nonetheless, Arabidopsis TOC75 protein, a component of the plastid outer membrane, was found in a fraction of detergentinsoluble membranes [bib_ref] Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts, Borner [/bib_ref] , supporting the idea that specific proteins might be included in microdomains of plastid membranes. In the current study, a beta-hydroxyacyl-(acyl-carrier-protein) dehydratase FabZ (Medtr2g008620), experimentally ascribed to the chloroplast envelope and the cell wall and reminiscent of the beta-hydroxyacyl-(acyl-carrier-protein) dehydratase precursor previously identified in M. truncatula DIMs [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] , was enriched more than 50 fold in both R-and G-DIM fractions, relative to microsomes, Because this enzyme displayed no chloroplast transit peptide (cTP), but may be plastid-encoded according to HAMAP prediction (data not shown), it is likely that this protein that has role in lipid biosynthesis may serve specific function(s) at the plastid membrane. Likewise, phospholipase D alpha, a noticeable plant DIM marker that participates to the metabolism of phosphatidylcholines, which are important constituents of cell membranes, lipase/lipoxygenase, and patellin-5 (see above), also belonged to those lipid-related proteins coenriched in R-and G-DIMs that can localize to non-green plastids (Additional file 3: . Regarding plastids, it is worth noting that these organelles are specialized, among other features, for the synthesis of fatty acid precursors that are either directly assembled within their own membranes, exported to the ER for extraplastidial lipid assembly, or reimported for the synthesis of plastidial lipids [bib_ref] Chloroplast lipid synthesis and lipid trafficking through ER-plastid membrane contact sites, Wang [/bib_ref].
With special interest in those proteins specifically enriched in R-and G-DIMs relative to microsomes, the most remarkable differences recorded between the subcellular patterns of these two fractions included enrichment in proteins ascribed to cytosol/cell wall/undefined membrane components and a depletion of nuclear proteins in the R-specific subset, whereas plasmodesmaand nucleus-associated proteins were enriched in the G-specific fraction [fig_ref] Figure 4: Cellular [/fig_ref]. Among the 22 cytosolic proteins recorded as specifically enriched in R-DIMs, only 6 didn't display any feature driving association to membranes and were exclusively assigned to cytosol according to experimental annotations, indicating that association of cytosolic proteins to R-DIM was not driven in the majority by non-membrane proteins. Likewise, all the 13 proteins located to the cell wall that were exclusively enriched in R-DIMs were predicted to have a membrane signature, as illustrated by germin-like protein, alpha-D-xylosidase, alpha-1,4-glucan-protein synthase, cysteine proteinase inhibitor 5, pectinesterase, beta-D-glucosidase, beta xylosidase, xylan 1,4-beta-xylosidase (Additional file 3: . It was also noticeable that R-procedure generated a DIM fraction largely depleted in nuclear proteins opposite to what observed for the gradient-based method, as previously depicted by Adam et al. [bib_ref] Rapid preparation of nuclei-depleted detergent-resistant membrane fractions suitable for proteomics analysis, Adam [/bib_ref]. Although mainly consisting of ribosomal proteins, most of the nucleus-ascribed proteins specifically enriched during G-DIM isolation displayed at least a membrane-related feature, which overall minimizes the likelihood that free ribosomes may have stricken to the lipid fraction during our extraction procedures [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref]. Finally, plasmodesma-located proteins happened to be selectively enriched in G-DIMs, as inferred from the presence of 22 accessions ascribed to this compartment, although not exclusively, among which the DIM-marker flotillin belongs to (Additional file 3: . In plants, plasmodesmata correspond to membranous channels that allow intercellular communication. Embedded in the cell wall, they are defined by specialized domains of the endoplasmic reticulum and the plasma membrane, which may explain the large representation (Additional file 3: of transporters and receptorlike kinase within the plasmodesma proteins enriched in G-DIMs, similar to what observed in the proteomic studies recently dedicated to plasmodesmata [bib_ref] Dissecting plasmodesmata molecular composition by mass spectrometry-based proteomics, Salmon [/bib_ref] [bib_ref] Receptor-mediated signaling at plasmodesmata, Faulkner [/bib_ref]. Due to the relative specialization of each organelle toward protein sorting and/or particular metabolic pathways, we thereby anticipated that the differential distribution of R-and G-DIM-enriched proteins over distinct cellular compartments might be of functional relevance.
## R-and g-dim-enriched proteins differ in their functional relevance
To obtain an overview of the functionality of R-and G-DIM-enriched proteins relative to the microsomal proteome, the corresponding 874 identifications were classified according the FunCat annotation scheme [bib_ref] The FunCat, a functional annotation scheme for systematic classification of proteins from..., Ruepp [/bib_ref] that assigned them to seven known biological processes. When regarding the fraction of the proteins co-enriched in Rand G DIMs, [fig_ref] Figure 4: Cellular [/fig_ref] showed a noticeable increase in the "transport and vesicular traffic" category, which is consistent with anterior repertoires showing that DIMassociated proteins are largely described in the context of membrane transport [bib_ref] Roles of lipid rafts in membrane transport, Ikonen [/bib_ref] [bib_ref] Membrane microdomains, rafts, and detergent-resistant membranes in plants and fungi, Malinsky [/bib_ref]. Signal transduction relatedproteins were also enriched in this core proteome, but to a lesser extent than expected from what usually observed in animal systems. This feature previously reported for DIMs extracted from the PMs of oat and rye, but also within sterol-dependent proteins of Arabidopsis, thereby comforts the opinion that not all signaling proteins are necessarily enriched in microdomains [bib_ref] Detergent-resistant plasma membrane proteome in oat and rye: similarities and dissimilarities between..., Takahashi [/bib_ref] [bib_ref] Definition of Arabidopsis sterol-rich membrane microdomains by differential treatment with methyl-betacyclodextrin and..., Kierszniowska [/bib_ref]. Remarkably, [fig_ref] Figure 4: Cellular [/fig_ref] also indicated that different functional categories were preferentially represented depending on the protocol used. Namely, proteins playing roles in "proteins synthesis/fate" were depleted in R-DIMs, relative to microsomes. Not observed for G-DIMs, this later pattern was consistent with the depletion of nuclear ribosomal proteins noticed in the current study and by Adam et al. [bib_ref] Rapid preparation of nuclei-depleted detergent-resistant membrane fractions suitable for proteomics analysis, Adam [/bib_ref]. Additionally, opposite to the functional partitioning obtained with the gradient method, proteins ascribed to "defense/cell rescue" and "energy/metabolism" processes were specifically enriched in R-DIMs. Finally, the category related to "cell structure/membrane shaping" displayed a larger increase within those proteins enriched in G-DIMs than after R-mediated sedimentation. To check whether the distribution of these functional categories could mirror the differential partitioning of cellular components in each DIM fraction, we then plotted the rate of occurrence of the Funcat-ascribed functions within the proteins specifically-enriched in each organelle, relative to microsomes. showed that cytosolic and cell wall/membrane proteins typical of R-DIMS were essentially involved in defense/cell rescue and energy/metabolism, respectively. By contrast, the PM and plasmodesmalocated proteins that were preferentially recruited in G-DIMs mostly played role in cell structure/membrane shaping processes. In a previous work [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] , proteins sustaining cellular trafficking and cell wall functioning were also found well-represented in DIM fractions prepared from M. truncatula root plasma membrane. The latter two functional categories, together with that corresponding to biotic/abiotic stress responses, were highlighted in DIM recovered from tobacco PM [bib_ref] New insights into mathematical modeling of the immune system, Morel [/bib_ref].
Taken together, the above-data showed that despite the existence of a conserved core of proteins, R-and Gprotocols each resulted in the enrichment of a particular DIM-proteome displaying specific cellular localizations and biological functions. Differences in protein and/or lipid associated with DIM fractions have been reported earlier, but essentially as dependent upon extraction parameters such as temperature, concentration and type of detergent [bib_ref] Detergent resistance as a tool in membrane research, Lingwood [/bib_ref] [bib_ref] Resistance of cell membranes to different detergents, Schuck [/bib_ref]. However, in the current study, DIM extraction procedures (temperature, duration, and detergent) were identical for both R and G protocols, which consequently only differed at DIM isolation process, namely sedimentation versus floatation on sucrose gradient. In this context, it was reasonable to assume that the selectivity displayed by each method could have arisen from some particular membrane compatible characteristics displayed by these differentially-enriched proteins.
## Membrane-associative features as related to r/g protein partitioning
Among the features favoring protein embedment in the hydrophobic lipid bilayer are membrane spanning protein domains that include typically alpha-helices or betasheets with hydrophobic surfaces serving as the interface to the hydrocarbon core of the lipid bilayer. In addition to, signal peptides of nascent proteins can also mediate protein translocation across or integration to membranes along the secretory pathway [bib_ref] Targeting proteins to membranes: structure of the signal recognition particle, Egea [/bib_ref]. Protein association to membranes can also be driven by lipidic anchors, which can either be permanent co-translational additions or posttranslational modifications. These lipid modifications include glycophosphatidylinositol (GPI) anchors, N-terminal myristic acid tails (myristoylation), and cysteine acylation (palmitoylation) [bib_ref] Greasing their way: lipid modifications determine protein association with membrane rafts, Levental [/bib_ref]. Consequently, we further monitored and compared the putative presence of trans-membrane (TM) spanning domains, signal peptide (SP) sequences, and lipid modifications and within R-and G-enriched proteins, relative to microsomes, as inferred from the corresponding online predictor tools (see Methods).
With regard to the distribution of lipid modifications within each of the cellular components specifically enriched in R (cytosol, CW, undetermined membranes) and G-DIMs (PM, plasmodesmata), indicated that the proportion of putative palmitoylated proteins was not affected by either of the method used. This result also holds true when considering the total subsets of the proteins enriched in R-and G-fractions . Overall, S-palmitoylation largely prevailed as the most abundant putative lipid modification associating proteins to membrane compartments in M. truncatula, as predicted for 487 (59%) candidates within the total microsomal fraction. This result is consistent with recent proteomic data derived from root callus culture of Arabidopsis, in which the number of proposed Sacylated proteins has increased from 30 to over 500 [bib_ref] A proteomic approach identifies many novel palmitoylated proteins in Arabidopsis, Hemsley [/bib_ref]. By contrast, GPI anchors that seemed more frequently predicted for R-enriched proteins than for G, relative to microsomes, only encompassed a few number of proteins . Likewise, the presence of putative N-terminal myristic acid tails was anecdotic within the proteins enriched in R-and G-DIMs. It appeared from that the most noticeable difference between the two proteomes corresponded to an increase in the proportion of predicted SP-containing proteins after sedimentation (R) compared to its decrease after floatation (G). This behavior, which was also observed at a larger scale when looking at the total subsets of the proteins enriched in R-and G-fractions , indicated that part of the late secretory pathway may likely contribute to the specificity of the R-DIM proteome. The classical eukaryotic pathway for secretion includes translocation of nascent proteins into the ER lumen and then through the secretory pathway, which comprises the Golgi apparatus (GA) and trans-Golgi network, protein packaging into vesicles that migrate to, and fuse with the PM, releasing the protein cargo into the cell wall, or are targeted to the vacuole or other post-Golgi compartments [bib_ref] Straying off the highway: trafficking of secreted plant proteins and complexity in..., Rose [/bib_ref]. The default pathway for proteins with signal peptides and with no additional targeting information is to proceed through the ER, Golgi, and PM where they are secreted into the cell wall. It is noteworthy that for targeting to the PM, proteins do not necessary require a signal peptide in so far as additional sequences, including membrane spanning regions, also allow them to flow to the PM [bib_ref] Insertion of Membrane Proteins into the ER Membrane, Lodish [/bib_ref]. Consequently, the preferential partitioning of SP-predicted proteins within the cell wall and endomembrane components specifically enriched in the R-fraction suggested that DIMs containing proteins targeted to secretion might not have a buoyancy similar to those uniquely retrieved by floatation (e.g., mainly PM and plasmodesmata DIMs). Consistent with this possibility, when investigating the distribution of microdomains in leek plant cells, Laloi et al. [bib_ref] Insights into the role of specific lipids in the formation and delivery..., Laloi [/bib_ref] showed that replacement of usual Δ 5 -sterols induced a preferential formation of DIMs in the GA compared to the PM, indicating that sterols didn't contribute equally to DIM formation along the secretory pathway. There were additional evidences for a preferential accumulation of sterols in the PM, with a progressive increase through the secretory pathway both in Zea mays and Allium porrum [bib_ref] Plant membrane sterols: isolation, identification and biosynthesis, Hartmann [/bib_ref] [bib_ref] A method to study the in vivo intermembrane transfer of lipids and..., Moreau [/bib_ref]. Remarkably, inhibition of sterol biosynthesis also led to a decrease in DIM protein markers within the PM without affecting vesicular transport to the cell surface, suggesting a likely requirement of particular sterols for protein targeting to PM microdomains [bib_ref] Insights into the role of specific lipids in the formation and delivery..., Laloi [/bib_ref] [bib_ref] Cell polarity and PIN protein positioning in Arabidopsis require STEROL METHYLTRANSFERASE1 function, Willemsen [/bib_ref]. In M. truncatula roots, spinasterol was reported as the dominant compound in DIMs prepared from the PM [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] , and, as described above, we also demonstrated that this sterol was enriched to a larger extent in G-than R-fractions. Inferred from these results and from the preferential retrieval of PM-related proteins in G-DIMs relative to R-fractions in which protein secretion dominates, we thus assume that spinasterol enrichment participates to the buoyancy of PM and plasmodesma DIMs, opposite to the DIMs located along the late secretory pathway in which sterols are suspected to be less concentrated [bib_ref] Insights into the role of specific lipids in the formation and delivery..., Laloi [/bib_ref].
In a similar line of reasoning, it has been shown that sterol enrichment within the PM can increase its hydrophobic thickness relative to ER and Golgi endomembranes, thereby mediating the selective targeting of proteins that have corresponding TM hydrophobic length [bib_ref] A comprehensive comparison of transmembrane domains reveals organelle-specific properties, Sharpe [/bib_ref] [bib_ref] Organellar lipidomics-background and perspectives, Klose [/bib_ref]. This process, referred to as the bilayer-mediated mechanism [bib_ref] Insights into the role of specific lipids in the formation and delivery..., Laloi [/bib_ref] , mirrors the view that due to their high (See figure on previous page.) Comparison of the distribution of (A) protein functional categories and (B) membrane-associative features between the cellular components enriched in R-DIMs (Cytosol, Cell wall, Membranes) and those enriched in G-DIMs (Plasma membrane, Plasmodesmata). Subsets termed "R2xspecific" and "G2xspecific" refer to the proteins uniquely enriched in R-and G-DIMs, respectively, relative to microsomes (Mic). The number (n) of proteins assigned to a given cellular component is indicated into brackets for Mic, "R2xspecific" and "G2xspecific" fractions (A) Functional distribution of the proteins ascribed to a cellular compartment enriched in R-DIMs (blue color) or G-DIMs (red color) relative to microsomes (yellow color), as inferred from the FunCat scheme. (B) Venn diagram distribution of membrane-associative features (%) within the proteins ascribed to a given cellular compartment, as predicted for protein sequence motifs (alpha-helices (blue), beta-strands (orange) and signal peptide (SP) (black) and lipid modifications (palmitoylation (yellow), GPI anchor (green), myristoylation (pink)).
concentrations of sterols and sphingolipids with long, saturated hydrocarbon chains, microdomains may have thicker bilayers than the surrounding lipid matrix containing unsaturated phospholipids [bib_ref] Cholesterol and the Golgi apparatus, Bretscher [/bib_ref]. Consequently, proteins with relatively long TM hydrophobic regions would be expected to localize in the thick bilayers, whereas shorter TM proteins should localize in the thinner nonraft regions [bib_ref] Structure, composition, and peptide binding properties of detergent soluble bilayers and detergent..., Gandhavadi [/bib_ref] [bib_ref] Altering hydrophobic sequence lengths shows that hydrophobic mismatch controls affinity for ordered..., Lin [/bib_ref]. In the current study, analysis of TM domain distribution, as displayed in , indicated a substantial depletion of predicted alpha-helices in the proteins enriched within microdomain fractions relative to microsomes, which dropped from 63 to 18 and 32% in Rand G-DIMs, respectively. This pattern, which was also observed when considering DIM-enriched cellular compartments with the exception of cell wall-related proteins , is consistent with the view that the presence of TM domains per se is not a prerequisite to drive protein association to microdomains [bib_ref] Greasing their way: lipid modifications determine protein association with membrane rafts, Levental [/bib_ref]. When refining the partitioning of TM domains with special attention to their number and length that may contribute to both organelle localization and microdomain affinity, it turned out that relative to the microsomal fraction, proteins predicted to be anchored by a single membrane-spanning helix were largely enriched in both DIM fractions, and to a larger extent in R-(83%) than in G-(60%) DIMs . One rationale for the recruitment of single-span TM proteins Comparison of Venn diagram distribution of predicted membrane-associative features (%) between the "R2xspecific", "G2xspecific", "RG2xcore", and Mic protein subsets, as related to protein sequence motifs (alpha-helices (blue), beta-barrels (orange) and signal peptide (SP) (black) and lipid modifications (palmitoylation (yellow), GPI anchor (green), myristoylation (pink)). The number (n) of proteins contained within each diagram is indicated into brackets. Subsets termed "R2xspecific" and "G2xspecific" refer to the proteins uniquely enriched in R-and G-DIMs, respectively, relative to microsomes (Mic), whereas "RG2xcore" designates the proteins enriched in both R-and G-DIMs, relative to microsomes.
within DIMs, as reported earlier in the PM microdomains of tobacco and barrel medic [bib_ref] New insights into mathematical modeling of the immune system, Morel [/bib_ref] [bib_ref] Characterization of lipid rafts from Medicago truncatula root plasma membranes: a proteomic..., Lefebvre [/bib_ref] , may be that oligomerization of monomeric TM proteins would increase their affinity for microdomains, which otherwise would only display a short residency time within rafts [bib_ref] Lipid domain structure of the plasma membrane revealed by patching of membrane..., Harder [/bib_ref]. When examining the length of these single-span TM domains, showed that although helices comprising 19 amino acids tended to be more frequently enriched in R-DIMs relative to G-DIMs in which stretches of 24 and 26 amino acids dominated, these differences may not be of significant relevance as they only encompassed a very few number of proteins. In silico analysis of the number of TM domains also indicated that proteins predicted to contain eight to twelve alpha-helices were enriched in G-DIMs but not in R-DIMs, relative to microsomes . This feature that was previously highlighted in the PM microdomains of tobacco [bib_ref] New insights into mathematical modeling of the immune system, Morel [/bib_ref] , therefore agrees with the preferential recruitment of plasmalemma-located proteins within G-DIMs [fig_ref] Figure 4: Cellular [/fig_ref]. Finally, we remarked that dynamin-2B, flotillin, stomatin and leucine-rich repeat (LRR) proteins belonged to those membrane-localized proteins predicted to contain 2 beta-strands, which were enriched in G-but not in R-DIMs relative to microsomes bottom panel). In this regard, dynamin-2B, flotillin and stomatin can display a hairpin-like topology, susceptible to influence membrane curvature and scaffolding processes in microdomains [bib_ref] Characterization of the stomatin domain involved in homo-oligomerization and lipid raft association, Umlauf [/bib_ref]. Likewise, LRR have curved horseshoe structures that are known to drive protein-proteininteractions [bib_ref] The leucine-rich repeat as a protein recognition motif, Kobe [/bib_ref] and can also act as a raft nanodomain targeting signal [bib_ref] Identification of a lysine-rich region of Fas as a raft nanodomain targeting..., Rossin [/bib_ref]. Overall, in silico predictions showed that R-and G-DIM-enriched proteins differ in the distribution of the three protein motifs currently investigated that can drive association to membrane compartments, namely alpha-helices, beta-strands, and signal peptides. The preferential partitioning of predicted SP-containing proteins within the cell wall and endomembrane components specifically enriched in R-DIMs, suggested that part of the late secretory pathway may contribute to the specificity of the R-DIM proteome.
# Conclusions
In the current study, DIMs were prepared for the first time directly from M. truncatula root microsomes that consist of a complex membrane mix relative to the PM conventionally used as starting material for microdomain isolation. We clearly established that both long-lasting sucrose gradient centrifugation (G protocol) and rapid microfuge sedimentation at low speed (R protocol) enable the recovery of membrane fractions that meet the criteria of DIMs, as inferred from sterol enrichment, presence of typical sphingolipid long-chain bases from plants, and enrichment in membrane proteins including canonical DIM markers. Proteomic analysis of the corresponding fractions also show that, despite the existence of a conserved core of proteins, R-and G-protocols result in the enrichment of a particular DIM-proteome displaying specific cellular localizations and biological functions. Collectively, even though microsomes were used as initial material, we show that the composition of the G-DIM fraction still mostly mirrored that of PM microdomains conventionally retrieved by floatation. In parallel, the possibility to isolate by rapid differential centrifugation a DIM fraction that seems to target the late secretory pathway opens new avenues to study plant microdomains. Finally, with regard to our initial questioning addressing the intracellular distribution of plant DIMs, current results obtained in M. truncatula roots clearly support the existence of microdomains not only in PM and the late secretory pathway, but also in additional membrane organelles, including non-green plastids.
# Methods
# Plant material
Medicago truncatula cv. Jemalong 5 seeds were surfacesterilized and germinated at 27°C in the dark onto 0.7% sterile agar [bib_ref] Proteome analysis and identification of symbiosis-related proteins from Medicago truncatula Gaertn. by..., Bestel-Corre [/bib_ref]. Two-day old seedlings were then transferred on soil and grown into 400 ml plastic pots containing a mix of sterile soil of Epoisses (neutral clay loam from Domaine d'Epoisses, INRA Dijon France) and sand (1:2, v/v) supplemented twice a week with a nitrogenenriched nutrient solution (Long Ashtonunder controlled conditions (16 h photoperiod, 220 μE.m −2 .s −1 light irradiance). After 4 weeks, roots were collected, gently rinsed with deionized water to get rid of soil, deep frozen in liquid nitrogen and stored at −80°C until further use.
## Microsomal protein purification
All steps of microsome preparation were carried out at 4°C according to [bib_ref] The membrane proteome of Medicago truncatula roots displays qualitative and quantitative changes..., Abdallah [/bib_ref]. Microsomes of M. truncatula roots were obtained by differential centrifugation as previously described for tobacco cells [bib_ref] Simon-Plas F: Quantitative proteomics reveals a dynamic association of proteins to detergent-resistant..., Stanislas [/bib_ref]. Briefly, frozen roots (about 100 g fresh weight) were homogenized using a Waring Blender in grinding buffer (50 mM Tris-MES, (See figure on previous page.) Comparison of the characteristics of the predicted trans-membrane domains between the "R2xspecific", "G2xspecific", "RG2xcore", and Mic protein subsets. (A) Distribution (%) of the number of alpha-helices (top panel) and beta-strands (bottom panel). (B) Distribution (%) of the length of single alpha-helicoidal spanning domain. Subsets termed "R2xspecific" and "G2xspecific" refer to the proteins uniquely enriched in R-and G-DIMs, respectively, relative to microsomes (Mic), whereas "RG2xcore" designates the proteins enriched in both R-and G-DIMs, relative to microsomes. The number (n) of involved proteins is indicated into brackets for each fraction. pH 8.0, 500 mM sucrose, 20 mM EDTA, 10 mM DTT and 1 mM PMSF). The homogenate was successively centrifuged at 12.000×g and 16.000×g for 20 min. After centrifugation, supernatants were collected, filtered through two successive meshes , and centrifuged at 100.000×g for 1 h. Microsomal pellets were resuspended in buffers according to the DIM fraction isolation procedure used (see below), homogenized with a glass pestle. Protein contents were quantified using the RCDC (bicinchoninic acid) Protein Assay Kit (BioRad) to avoid TX-100 interference, using BSA as standard.
## Detergent-insoluble-membrane fraction isolation
For both the Rapid (R) and the Gradient (G) protocol [fig_ref] Figure 1: Overview of the Rapid and Gradient protocols used for isolating DIM fractions... [/fig_ref] , three independent extractions of DIMs were performed, each from a 6 mg protein equivalent of microsomes. Anotations R/G x:y were used to refer to the R or G protocol, the x detergent/protein ratio (w/w), and the y (%, v/v) final detergent concentration, respectively.
The rapid protocol was performed as described in [bib_ref] Rapid preparation of nuclei-depleted detergent-resistant membrane fractions suitable for proteomics analysis, Adam [/bib_ref]. The microsomal fraction was first resuspended in 10 mM Tris-MES buffer pH 7.3, containing 250 mM sucrose, 1 mM EDTA, 10 mM DTT, 1 mM PMSF, 10 μg/ml aprotinin and 10 μg/ml leupeptin. To increase solubilisation, resuspended microsomal proteins were aliquoted into 1 mg fractions that were further diluted (270 μl final volume) with 25 mM Tris-MES pH 6.5, containing 150 mM NaCl. Then, 30 μl of 10% (v/v) Triton X-100 was added to reach a final detergent-to-protein ratio of 3:1 (w/w) and a 1% final detergent concentration (v/v). Incubation was performed on ice under gentle shaking for 30 min. Samples were centrifuged at 16.000xg for 20 min. The resulting Triton X-100insoluble fraction (DIM) was homogenized (using a micropestle) in 200 μl of beta-octylglucoside-containing buffer (60 mM beta-octylglucoside, 10 mM Tris HCl pH7.5, 150 mM NaCl) and incubated on ice under gentle shaking for 30 min. A last centrifugation step (16.000×g for 20 min) was performed to separate soluble DIM fraction (R-DIM) from both Triton X-100 and betaoctylglucoside-insoluble pellet. Aliquoted DIM fractions were pooled and protein amount was measured using the RCDC Protein Assay Kit (BioRad).
For the gradient protocol, the microsomal pellet was resuspended in TNE buffer (25 mM TrisHCl pH 7.5, 150 mM NaCl, 5 mM EDTA) according to Borner et al. [bib_ref] Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts, Borner [/bib_ref]. The microsomal fraction (6 mg protein equivalent) was treated with 1% (w/w) Triton X-100 forinterface, washed with an excess of TNE buffer and centrifuged at 100.000×g for 1 hour to remove residual sucrose. The DIM pellet (G-DIMs) was homogenized in 1 ml of beta-octylglucoside containing buffer (60 mM beta-octylglucoside, 10 mM Tris HCl pH7.5, 150 mM NaCl). Protein concentration was determined with the RCDC Bio-Rad protein assay.
## Free sterol extraction and analysis
Total lipids were extracted from microsomal and DIM fractions (R and G; 100 μg equivalent proteins per sample, previously diluted in 0.37 M KCl to reach a final volume of 500 μl) according to Folch et al. [bib_ref] A simple method for the isolation and purification of total lipides from..., Folch [/bib_ref] with chloroform/ methanol (2:1, v/v). The extraction of lipids was carried out in 25 ml glass tubes with Teflon lined screw caps. Sterol internal standard (10 μg epichoprostanol) was added to Mic, R and G samples but not into one Mic sample (Mic-Std). All solutions were then mixed with MetOH/chloroform 2:1 (v/v), shaken and left overnight at 4°C. The next day, chloroform was added to reach a final ratio MetOH/chloroform 2:4 (v/v) and the mixture was centrifuged at 320×g for 8 min. The sterol-containing lower phase was collected, dried under nitrogen flux, washed once with EtOH and dried again. Lipids were saponified using ethanolic KOH (1 ml EtOH, 100 μl KOH (11 N)). Upon warming at 80°C for 1 h, 2 ml H 2 0 and 2 ml hexane were added and the lipids were recovered by centrifugation (320×g, 8 min). The upper phase was evaporated to dryness under nitrogen flux. Then, the residue was derivatized by adding 100 μl of the silylating agents (BSTFA/TMCS mixture (5:1 v/v)) and warmed at 80°C for 1 hour. Samples were finally analyzed for their sterol content by gas chromatography after addition of 400 μl hexane.
Gas chromatography analyses were carried out on Agilent 7890 GC instrument equipped with a with a flame ionization detector, on a Varian Factor Four VF-5 ms capillary column (15 m, 0.32 mm i.d. × 0.25 μm film thickness). Sample manual injection (1 μl) was performed in splitless mode (split vent at 30 seconds; injector temperature 240°C). Helium was the carrier gas at 1.5 ml/min in constant flow mode. Temperature program was programmed from 120°C to 240°C at 9°C/min. Data were processed using the Agilent EZ Chrom Elite software providing retention time and area for each compound of interest.
## Quantification of sphingolipid long-chain bases
LCB content of DIM and microsomal fractions were determined as previously described Cacas et al. [bib_ref] Rapid nanoscale quantitative analysis of plant sphingolipid long-chain bases by GC-MS, Cacas [/bib_ref] from three individual experiments. Briefly, LCB were released from fractions by direct overnight incubation at 110°C in 1 ml dioxane and 1 ml 10% (w/v) Ba(OH) 2 solution prepared in water. Before incubation, standard LCB (d14:1, d17:1 and d20:0) used for quantification were directly added to the dioaxane/barium mixture (10 μg for each standard LCB/sample). Upon cooling and addition of 6 ml distilled water, LCB were extracted twice with 4 ml diethylether. Pooled diethylether phases were dried under nitrogen flux. Dry residues were dissolved in 1 ml methanol containing 100 μl of a freshly prepared 0.2 M metaperiodate (NaIO 4 ) solution. Oxidation of extracted LCB into aldehydes was then carried out in the dark for 1 hour at room temperature and under mild shaking, as described by Kojima et al. [bib_ref] Composition and molecular-species of ceramide and cerebroside in scarlet runner beans (Phaseolus-Coccineus..., Kojima [/bib_ref]. LCB-derived long-chain aldehydes were extracted into 1 ml hexane following addition of 1 ml water. To concentrate samples, the aldehyde-containing hexane phase was dried under nitrogen flux, and aldehydes were finally resuspended in 100 μl hexane to be injected into GC-MS.
For the separation of LCB-derived fatty aldehydes, a 30 m × 250 μm HP-5MS capillary column (5% phenylmethyl-siloxane, 0.25 μm film thickness, Agilent) was used with helium carrier gas at 2 ml/min; injection was in splitless mode; injector and MS detector temperatures were set to 250°C (Agilent 6850 coupled to a mass analyzer Agilent 6975); the oven temperature was held at 50°C for 1 min, then programmed with a 25°C/min ramp to 150°C (2 min hold), a 10°C/min ramp to 210°C, and 75°C/min ramp to 320°C (5 min hold). Upon separation by GC and detection by MS, fatty aldehydes were identified based on their retention time and fragmentation [bib_ref] Rapid nanoscale quantitative analysis of plant sphingolipid long-chain bases by GC-MS, Cacas [/bib_ref]. The ion current of each molecular species of interest was determined and further used for calculating the amount of molecules by comparison with the appropriate internal standards. These results were expressed in micrograms. Taking into account the molecular weight of individual LCB-derived aldehydes, the quantity in moles for each molecular species was calculated and expressed as mole%.
## One-dimensional sds-page and nano-lc-ms/ms analysis
Microsomal and DIMs samples (20 μg protein equivalent) were mixed at a ratio of 1 to 1 with Laemmli buffer [bib_ref] Cleavage of structural proteins during the assembly of the head of bacteriophage..., Laemmli [/bib_ref] without any heating denaturation step. Samples were separated onto small 12% polyacrylamide gels with 4.5% stacking gel and proteins were stained with colloidal blue (G250). Proteins were separated along a short (1 cm)-migration. Individual gel lanes were sliced in 7 pieces for in-gel digestion and LC-MS/MS analysis. Each section was washed in water and completely destained using 100 mM NH 4 CO 3 in 50% acetonitrile (ACN). A reduction step was performed by addition of 100 μl of 50 mM NH 4 CO 3 , pH 8.9, and 10 μl of 10 μM TCEP (Tris(2-carboxyethyl) phosphine HCl) at 37°C for 30 min. The proteins were alkylated by adding 100 μl of 50 mM iodoacetamide and allowed to react in the dark at 20°C for 40 min. Gel sections were first washed in water, then ACN, and finally dried for 30 min. In-gel digestions were performed with trypsin in the Progest system (Genomic Solution, East Lyme, CT, USA) according to a standard protocol. Gel pieces were washed twice by successive baths of 10% (v/v) acetic acid, 40% (v/v) ethanol and ACN. They were then washed twice with successive baths of 25 mM NH 4 CO 3 and ACN. Digestion was subsequently performed for 6 h at 37°C with 125 ng of modified trypsin (Promega) dissolved in 20% (v/v) methanol and 20 mM NH 4 CO 3 . Peptides were extracted successively with 2% (v/v) TFA and 50% (v/v) ACN and then with pure ACN. Peptide extracts were dried and suspended in 20 μl of 0.05% (v/v) TFA, 0.05% (v/v) HCOOH, and 2% (v/v) ACN.
Mass spectrometry analysis was carried out on 2 independent replicates for each DIM fraction (R and G). Peptide separation was performed using an Eksigent 2D-ultra-nanoLC (Eksigent Technologies, Livermore, CA, USA) equipped with a C18 column (5 μm, 15 cm × 75 μm, PepMap, LC packing). The mobile phase consisted of a gradient of solvents A 0.1% HCOOH (v/v) in water and B 99.9% ACN (v/v), 0.1% HCOOH (v/v) in water. Peptides were separated at a flow rate of 0.3 μl/min using a linear gradient of solvent B from 5 to 30% in 60 min, followed by an increase to 95% in 10 min. Eluted peptides were online analysed with a LTQ XL ion trap (Thermo Electron) using a nanoelectrospray interface. Ionization (1.5 kV ionization potential) was performed with a liquid junction and a non-coated capillary probe (10 μm i.d.; New Objective). Peptide ions were analyzed using Xcalibur 2.0.7, with the following data-dependent acquisition stepsfull MS scan (mass to charge ratio (m/z) 300-2000, centroid mode), (2) MS/MS (qz = 0.25, activation time = 30 ms, and collision energy = 35%, centroid mode).
Step 2 was repeated for the three major ions detected in step 1. Dynamic exclusion was set to 45 s.
## Protein identification and quantification
Searches were performed using the Mascot search engine (http://www.matrixscience.com) on the Medicago truncatula pseudomolecule database (http://www.jcvi.org/ cgi-bin/medicago/annotation.cgi) version 3.5v3 (47529 entries). Trypsin digest was set to enzymatic cleavage, carboxyamidomethylation of, and oxidation of methionines were defined as fixed and variable modifications, respectively. Precursor mass precision was set to 2.0 Da with a fragment mass tolerance of 0.5 Da. Sequences corresponding to keratins or trypsin were removed by querying a homemade contaminant database as a first step of filtration. Identified proteins were validated according to the presence of at least two peptides with an E value smaller than 0.05. To take redundancy into account, proteins were grouped according to the Legoo server (http://www.legoo.org/).
[fig] Figure 1: Overview of the Rapid and Gradient protocols used for isolating DIM fractions from M. truncatula roots microsomes. [/fig]
[fig] Figure 2: Comparison of sterol and long-chain base (LCB) contents in R-DIMs(R) and G-DIMs (G) relative to the initial microsomal (Mic) fraction of M. truncatula roots. All experiments were performed with 100 μg protein equivalents (A) Representative GC profile of total extracted sterols in the three fractions. Epichoprostanol (O) was used as an internal standard (10 μg) to quantify major sterol peaks (*) and added to the Mic, R and G fractions but not to the Mic minus Std (Mic-Std) sample. (B) Sterol enrichment from 100 μg protein equivalents. Results are expressed as the means ± SE (vertical bars) of at least three independent preparations. (C) Representative distribution of LCB in the three fractions; Abbreviations used: t18:1(8Z): 4-hydroxysphing-8(Z)-enine, t18:1(8E): 4-hydroxysphing-8(E)-enine, t18:0: 4-hydroxysphinganine (phytosphingosine), d18:1(8Z): sphing-8(Z)-enine, d18:1(8E): sphing-8(E)-enine, d18:0: sphinganine (dihydrosphingosine), d18:2(4E,8E,Z): sphinga-4 (E),8(E,Z)-dienine. [/fig]
[fig] Figure 4: Cellular (A) and functional (B) distribution of the 227 proteins recorded as enriched above 2-fold in DIM-fractions relative to microsomes of M. truncatula roots. Subsets termed "R2xspecific" and "G2xspecific" refer to the proteins uniquely enriched in R-and G-DIMs, respectively, whereas "RG2xcore" designates the proteins enriched in both R-and G-DIMs, relative to microsomes. (A) Localization was inferred from TAIR and LegumeIP homologous proteins having experimentally checked cellular components. (B) Functional classification was performed using the FunCat scheme. [/fig]
[table] Table 1: Main literature background to microdomain preparations as related to initial fractions [/table]
|
Salmonella enterica Serovar Enteritidis Liver Abscess after Blunt Abdominal Trauma
[bib_ref] Intrahepatic hematoma with secondary Salmonella infection via biliary fistula, Cerwenka [/bib_ref]
## Case presentation
A 27-year-old patient, who lived in the Adriatic coastal area, was admitted to the regional hospital after being injured in a motorbike accident, in which he acquired a blunt abdominal trauma to the right hemiabdomen, a clavicular fracture, serial rib fractures and fractures of the metacarpal bones. His medical history was unremarkable. He did not have any chronic viral infections, and the serology for HIV was negative. Upon admission, the patient was haemodynamically stable. The initial laboratory results revealed anaemia (haemoglobin 98 g/L), hepatic injury (aspartate transaminase 968 U/L, alanine transaminase 1520 U/L) and kidney injury (creatinine 147 µmol/L). The computed tomography (CT) scan revealed a right kidney and right liver lobe rupture, with a subcapsular liver haematoma . Two days later his physical state deteriorated, with dyspnoea, tachycardia and hypotension. Laboratory tests represented a sepsis: a leukocyte count of 5.200/mm 3 with 30% of nonsegmented polymorphonuclear leukocytes, haemoglobin 89 g/L, platelet count 123/mm 3 , C-reactive protein 184 mg/L, aspartate transaminase 1165 U/L, alanine transaminase 1809 U/L, creatinine 155 µmol/L, serum procalcitonine 26.24 ng/mL and lactate 4.1 mmol/L. There were no other serum abnormalities. Intravenous meropenem 3x1 g (Meronem, AstraZeneca Limited; Cheshire, United Kingdom) and vancomycin 2x1 g (Edicin, Lek farmacevtska druzba d.d.; Ljubljana, Slovenia) were administered. A control CT scan revealed gas inside the huge lesion of the right liver lobe, i.e. it revealed a newly formed liver abscess. Because of the physical deterioration, the patient was sent to a tertiary-care clinic for further medical care. On admission to our clinic for surgery, the patient was haemodynamically stable with a noradrenaline (Arterenol, Sanofi-Aventis; Frankfurt/Main, Germany) support. On physical examination there was a diminished breath sound in the basal part of the right lung lobe. A contrast-enhanced CT scan of the thorax, abdomen and pelvic region was obtained. It confirmed a huge eight-centimetre abscess [fig_ref] FIG. 2: Contrast-enhanced computed tomography axial scan at followup showing a huge gas-forming abscess... [/fig_ref]. Despite the antibiotic treatment, the patient was continuously febrile. C-reactive protein was repeatedly elevated, up to 299 mg/L [fig_ref] TABLE 1: Ladic et al [/fig_ref] (6). The isolate was sensitive to amoxicillin, ceftriaxone, ciprofloxacine and trimethoprim/sulfamethoxazole, so the antibiotic therapy was switched to ciprofloxacine, 2x400 mg (Ciprinol, Krka d.d.; Novo Mesto, Slovenia) intravenously daily. Ultrasound-guided percutaneous abscess drainage was performed with only a small quantity of the material obtained. In light of the constant elevation of inflammatory markers and not entirely successful percutaneous liver drainage, an operation was indicated. It began as a laparoscopic procedure, but following CO 2 insufflation, the patient became tachycardic, and the operation was converted to midline laparotomy extended to the right subcostal area. A liver bisegmentectomy of the sixth and seventh segment was performed, in addition to cholecystectomy and appendectomy. An intraoperative swab from the liver abscess was positive for S. enterica serovar Enteritidis. The post-operative course was uneventful. The patient continued with intravenous ciprofloxacine treatment (Ciprinol, Krka d.d.; Novo Mesto, Slovenia), 2x400 mg daily. Two weeks later, the patient was febrile again, with a body temperature of 39 °C, with chills and cold sweats. After a thorough examination, we noticed a skin wound and took a swab -it was positive for a Coagulasenegative Staphylococcus, sensitive to vancomycin. The ciprofloxacine treatment was changed after 16 days to intravenous vancomycin 2x1 g (Edicin, Lek farmacevtska druzba d.d.; Ljubljana, Slovenia) and piperacillin/tazobactam 3x4.5 g (Tazocin, Wyeth Lederle S.p.A.; Catania, Italy) daily, according to the recommendation of the clinical pharmacologist. The rest of the post-operative course was uneventful and the patient was discharged home four weeks after the operation. Written informed consent was obtained from the patient.
# Discussion
Salmonella infection is a very common bacterial infection, obtained through contaminated water or food. More than 2.500 Salmonella serovars exist; the serovar Enteritidis is among the most commonly reported serovars of human salmonellosis in industrialized countries [bib_ref] Global monitoring of Salmonella serovar distribution from the World Health Organization Global..., Hendriksen [/bib_ref]. It causes diverse clinical syndromes -from unnoticed infection or gastroenteritis with febrile episodes in otherwise healthy subjects to complicated intraabdominal infections in immune-compromised patients [bib_ref] Salmonella chronic carriage: epidemiology, diagnosis and gallbladder persistence, Gunn [/bib_ref]. Chronic carriage of non-typhoidal Salmonella serovars (NTS), such as serovar Enteritidis, occurs in 0.5% of cases, compared to 3% of typhoidal serovars [bib_ref] Salmonellosis and the GI Tract: More than Just Peanut Butter, Crum Cianflone [/bib_ref]. A growing number of clinical cases report on chronic NTS carriage in immunocompetent patients. According to the available data from our country (Croatia), the prevalence of a chronic Salmonella carrier state is 3% [bib_ref] Epidemiological characteristics of salmonellosis in New Zagreb during the 1990-2009 period, Ban [/bib_ref]. Both NTS and typhoid serovars initially adhere to the small intestine epithelium, where they encounter gastrointestinal microbiome -a huge body of intestinal bacteria that acts as a defence mechanism against multiple pathogens. In salmonellosis, it protects the gut by producing toxic metabolites that repress Salmonella virulence gene expression and mediate bacterial gut clearance [bib_ref] The microbiota mediates pathogen clearance from the gut lumen after nontyphoidal Salmonella..., Endt [/bib_ref]. Recent reports have suggested, though, that certain NTS serovars, by inducing intestinal inflammation, develop the ability to consume non-fermentable substrates, in contrast to the host microbiota, which relies on fermentative metabolism. This gives Salmonella an advantage in terms of growth over the host microbiome [bib_ref] The microbiota mediates pathogen clearance from the gut lumen after nontyphoidal Salmonella..., Endt [/bib_ref]. After the initial intestinal infection, typhoid serovars inhabit the gall bladder, where they live in a dormant-like state (3). There is no strong evidence that serovar Enteritidis inhabits the gall bladder as well, but it is known that it can be cultivated from bile and faeces. It can reach the gall bladder through blood drainage but also from the bowel along the bile ducts (3). Hepatic abscesses are uncommon, according to a North American study that found an incidence of 2.3 cases per 100.000 inhabitants [bib_ref] Population-Based Study of the Epidemiology of and the Risk Factors for Pyogenic..., Kaplan [/bib_ref]. Predisposing factors for the formation of abscesses are cholelythiasis, intrahepatic haematomas and hepatocellular carcinoma, as well as blunt and penetrating trauma to the abdomen. Hepatic abscesses are mostly treated conservatively -with percutaneous or surgical drainage, in conjunction with antibiotic therapy. The duration of antibiotic therapy depends on the underlying symptoms and possible complications and should be accompanied by careful imaging and monitoring for complications. S. enterica serovar Enteritidis is rarely described as a liver abscess pathogen [bib_ref] Intrahepatic hematoma with secondary Salmonella infection via biliary fistula, Cerwenka [/bib_ref]. In contrast to all other Salmonella serovars, S. enterica serovar Enteritidis has the potential to form a gas from carbohydrates. The gas accumulation results in an impaired transport of nutrients in adjacent tissues and promotes tissue destruction [bib_ref] Pyogenic liver abscess caused by Salmonella Enteritidis: A rare case report, Mahajan [/bib_ref]. This makes S. enterica serovar Enteritidis an extremely noxious pathogen. Cerwenka et al. (5) presented a case 20 years ago in which their patient had a blunt abdominal trauma, after which he developed a small biliary fistula that filled the abscess with Salmonella-infected biliary content. Our patient was young and previously healthy, but most likely a chronic Salmonella carrier, who probably went through the same pathogenic process -his abscess was filled with bacteria through injured small bile ducts. Although the current recommendations are for conservative treatment of a liver abscess for as long as is possible, we decided on surgery very early on. This decision was based upon: a) extremely rapid deterioration with dysregulated host response and organ dysfunction, b) unsatisfying results of percutaneous liver drainage. The surgical approach was chosen due to the fact that the abscess was filled not only with pus, but also with necrotic and haemorrhagic content.
In conclusion, Salmonella may carry an unpredictable disease course, especially in immune-compromised patients. A Salmonella liver abscess should be considered in regions/ countries with a higher incidence of chronic carriers. The most appropriate approach to this problem lies in the primary prevention of foodborne illnesses.
Financial Disclosure: No financial disclosure was declared by the authors.
[fig] FIG. 2: Contrast-enhanced computed tomography axial scan at followup showing a huge gas-forming abscess in the right liver lobe (arrow). [/fig]
[table] TABLE 1: Ladic et al. S. enterica Serovar Enteritidis Liver Abscess Preoperative laboratory parameters of the patient with post-traumatic liver abscess (S. enterica serovar Enteritidis) [/table]
|
A rare case of multiple giant colonic diverticula successfully treated with laparoscopic sigmoidectomy
Colonic diverticulosis is pervasive in Western society, with over half of individuals over the age of 60 carrying the diagnosis. A Giant Colonic Diverticulum (GCD) is a rare presentation of diverticulosis, involving one or more colonic diverticula that measure 4 cm or greater. Less than 200 reports of GCD have been published in the literature. Almost all GCD patients present with symptoms, with abdominal pain being the most common. Diagnosis is usually made with CT imaging and recommended treatment is segmental colectomy. We present an atypical case of GCD with an asymptomatic presentation, initial diagnosis made during endoscopy and a minimally invasive resection of multiple GCD within the same patient.
# Introduction
Colonic diverticulosis is pervasive in the Western world. Nearly 80% of individuals over the age of 80 have diverticulosis of the colon. Diverticula form at anatomic 'weak' points where arteries enter the circular smooth muscle of the colon wall, allowing herniation of mucosal and submucosal layers. In contrast to 'true' diverticula, which involve all layers of the bowel wall and are most often congenital, these outpouchings signify 'false' or 'pseudo' diverticula. Most are small, measuring less than a few centimeters (cm) in greatest dimension, but a diverticulum measuring 4 cm or greater is defined as a giant colonic diverticulum (GCD), with reports of GCD over 20 cm. In a review of the literature, Nigri et al.found less than 200 reported cases of GCD. The pathogenesis of GCD is largely unknown but the ball-valve theory is most common, hypothesizing that chronic high pressure from gas entering the diverticulum but unable to exit due to occlusive debris or inflammation at the diverticular neck stretches the tissues over time. The rising pressure is relieved either via diverticular decompression intraluminally or free perforation intraperitoneally.
McNutt et al.defined three types of GCD. Type one is the result of a normal, smaller, pseudo-diverticulum slowly increasing in size with intact mucosa in the diverticular wall. Type two describes a pseudo-diverticulum that has perforated and formed a chronic abscess cavity, as might occur after treatment of diverticulitis by laparoscopic washout without resection. Type three is a true, congenital, diverticulum with all layers of the colon wall present. Most patients are symptomatic (93%), with abdominal pain the most common (69%) presenting symptom. Diagnosis is usually made with imaging and computed Diverticulum contains all layers of the colon wall: serosa, muscularis, submucosa and mucosa tomography (CT) scan is the study of choice. Endoscopy can also serve as a means of diagnosis, but increased risk of perforation limits its use. Although diverticulectomy has been reported, most recommend surgical treatment with segmental colectomy. Many authors have reported successful laparoscopic resections for GCD; however, the majority of resections reported to date have been performed open-likely reflecting the common high acuity presentation that may include peritonitis or perforation. We present an atypical case of multiple GCD with an asymptomatic presentation and an endoscopic mode of diagnosis, treated by successful minimally invasive segmental resection.
## Case report
A 68-year-old woman with history of hypertension, breast cancer, gastroesophageal reflux disease, osteoarthritis, hypothyroidism and prior GI bleeding with frequent NSAID use underwent diagnostic colonoscopy to evaluate anemia to 9.6g/dL refractory to iron supplementation. Prior abdominal surgery included umbilical hernia repair and caesarian section. She denied abdominal symptoms including pain. Colonoscopy identified widespread left-sided diverticulosis and two large, cavernous areas partially filled with dark debris in the sigmoid colon without active bleeding. CT imaging confirmed severe diverticulosis with two GCD arising from the posterior wall of the proximal and mid-sigmoid colon measuring 6.8 × 4.4 × 4.3 cm and 6.1 × 5.2 × 3.9 cm, respectively. Mild wall thickening and trace, low-attenuation fluid in the adjacent sigmoid mesentery were suggestive of an inflammatory state without evidence of perforation. The patient was appropriately counseled and offered laparoscopic sigmoidectomy. Intra-operatively, the two GCD were seen extending well into the pelvic cavity with minimal scarring and no evidence of perforation. Sigmoid colectomy was performed to an area of pliable, descending colon and specimen extraction was performed via a Pfannenstiel incisionbefore performing an end-toend, circular stapled coloproctostomy anastomosis. The patient recovered well, with return of bowel function on post-operative day 4 and discharge home on post-operative day 6. Her anemia corrected with ongoing iron supplementation, and she experienced no complications within 10 months of follow-up.
Macroscopic pathologic examination confirmed two giant diverticula with narrow ostia connecting them to the colonic lumen adjacent to more typical, smaller diverticula. Microscopic examination confirmed the GCD to be 'false' or 'pseudodiverticula' with only mucosa and submucosa present in the diverticular wall. Significant fibrosis and marked submucosal plasma cell infiltration were present, features of a chronic inflammatory state without perforation, consistent with a type I GCD.
# Discussion
Less than 200 cases of GCD have been reported since the first recorded report in 1946, making it a rare variant of an exceedingly common condition. Most patients are symptomatic at presentation, primarily reporting abdominal pain. Despite having two GCD, our patient was asymptomatic, placing her in the overwhelming minority of GCD patients. Some authors have noted endoscopy to be ineffective in diagnosing GCD with limited visibility from narrow ostia or unsafe due to increased perforation risk. Our patient, similar to a case described by Mehta et al., safely underwent colonoscopy to rule out malignancy as the cause of her anemia and was discovered to have two GCD. Notably, the patient had history of frequent NSAID use which has been linked to nearly a threefold increased risk of diverticular bleeding.
A review by Steenvoorde et al.found that only 9% of patients present with multiple GCD, which can make surgical approaches more complex and highlights the importance of CT imaging for surgical planning. To the best of our knowledge, this is the first reported laparoscopic colectomy involving multiple GCD. Of the three types, this patient's GCD is compatible with type 1 as mucosa was present in the walls of the non-perforated diverticula. The inherent perforation present in patients with type II GCD may prove more technically challenging and has historically mandated an open surgical approach, especially for fulminant diverticulitis. Very large GCD over 10 cm may be difficult to resect using a minimally invasive approach given concerns of iatrogenic injury with specimen handling and limitations for extraction of such a large specimen. Ultimately, surgical approach must be tailored to each individual patient's clinical presentation and unique anatomy. Our technique did not differ from the standard laparoscopic sigmoidectomy technique. This case confirms that GCD may be asymptomatic and demonstrates the feasibility and safety of minimally invasive surgical techniques for patients with multiple GCD. |
A novel long noncoding RNA HOXC-AS3 mediates tumorigenesis of gastric cancer by binding to YBX1
Background: Recently, increasing evidence shows that long noncoding RNAs (lncRNAs) play a significant role in human tumorigenesis. However, the function of lncRNAs in human gastric cancer remains largely unknown. Results: By using publicly available expression profiling data from gastric cancer and integrating bioinformatics analyses, we screen and identify a novel lncRNA, HOXC-AS3. HOXC-AS3 is significantly increased in gastric cancer tissues and is correlated with clinical outcomes of gastric cancer. In addition, HOXC-AS3 regulates cell proliferation and migration both in vitro and in vivo. RNA-seq analysis reveals that HOXC-AS3 knockdown preferentially affects genes that are linked to proliferation and migration. Mechanistically, we find that HOXC-AS3 is obviously activated by gain of H3K4me3 and H3K27ac, both in cells and in tissues. RNA pull-down mass spectrometry analysis identifies that YBX1 interacts with HOXC-AS3, and RNA-seq analysis finds a marked overlap in genes differentially expressed after YBX1 knockdown and those transcriptionally regulated by HOXC-AS3, suggesting that YBX1 participates in HOXC-AS3-mediated gene transcriptional regulation in the tumorigenesis of gastric cancer. Conclusions: Together, our data demonstrate that abnormal histone modification-activated HOXC-AS3 may play important roles in gastric cancer oncogenesis and may serve as a target for gastric cancer diagnosis and therapy.
# Background
Gastric cancer is one of the leading causes of cancerrelated deaths worldwide and the most common gastrointestinal malignancy in East Asia [bib_ref] Cancer statistics, Siegel [/bib_ref] [bib_ref] Screening of gastric cancer in Asia, Sugano [/bib_ref]. Gastric cancer is diagnosed at an advanced stage accompanied by malignant proliferation in most patients, and the prognosis for advanced stage patients remains very poor. Therefore, additional research is needed to discover and develop effective biomarkers and targets for gastric cancer diagnosis and treatment.
To date, gastric cancer research has mainly focused on the deregulation of protein-coding genes to identify oncogenes and tumor suppressors that could serve as diagnostic and therapeutic targets. However, protein-coding sequences occupy less than 2% of the human genome [bib_ref] Chromatin signature reveals over a thousand highly conserved large non-coding RNAs in..., Guttman [/bib_ref] [bib_ref] A long noncoding RNA controls muscle differentiation by functioning as a competing..., Cesana [/bib_ref]. LncRNAs are operationally defined as RNA transcripts that are > 200 nt with limited protein coding potential [bib_ref] No-nonsense functions for long noncoding RNAs, Nagano [/bib_ref] , which have been shown to play a key role in tumorigenesis, including GC [bib_ref] The emergence of lncRNAs in cancer biology, Prensner [/bib_ref] [bib_ref] Long noncoding RNA GAPLINC regulates CD44-dependent cell invasiveness and associates with poor..., Hu [/bib_ref] [bib_ref] LncRNA GClnc1 promotes gastric carcinogenesis and may act as a modular scaffold..., Sun [/bib_ref]. Many studies found that lncRNAs could play an important role in regulating gene expression by different mechanisms, including chromatin modification, and transcriptional and posttranscriptional processing [bib_ref] Long non-coding RNAs: insights into functions, Mercer [/bib_ref] [bib_ref] Long noncoding RNAs: functional surprises from the RNA world, Wilusz [/bib_ref] [bib_ref] A functional genomic approach identifies FAL1 as an oncogenic long noncoding RNA..., Hu [/bib_ref]. For example, HOTAIR is involved in the transcriptional repression of HOX loci and promotes breast metastasis by binding to PRC2 (Polycomb Repressive Complex) [bib_ref] Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis, Gupta [/bib_ref]. However, the biological functions of lncRNAs in the control of GC tumorigenesis are not well characterized. Therefore, a better understanding of the role of lncRNAs underlying GC progression will enrich the understanding of the molecular mechanisms of GC carcinogenesis and provide information for improving the diagnosis and treatment of GC.
In our present study, we identified the full sequence of HOXC-AS3 and found that gain of H3K4me3 and H3K27acetylation could activate the expression of HOXC-AS3, both in cells and in tissues. HOXC-AS3 was also significantly upregulated in GC tissues compared with the corresponding nontumor tissues and may serve as an independent predictor for the overall survival in GC. In addition, HOXC-AS3 regulated cell proliferation and migration both in vitro and in vivo. RNA-seq analysis for whole transcriptome studies indicates an important role for HOXC-AS3 in the tumorigenesis of GC, and the activated function of HOXC-AS3 was mediated, in part, by interaction with YBX1. These results suggest that further studies to identify nonprotein-coding genes that contribute to oncogenesis are necessary for elucidating the complex genetic rewiring that is driven by HOXC-AS3 in GC.
# Results
Identification of HOXC-AS3 by analyzing gastric cancer RNA-expression profiling data Raw microarray data was downloaded from GEO, including GSE50710 (n = 20) and GSE58828 (n = 6). To obtain differentially expressed lncRNAs, signal data were normalized and z-score-transformed (paired or group t-test according to the experimental design was used to validate statistical significance). The top 20 dysregulated lncRNAs in these GEO datasets were shown in , b. For example, the lncRNA that showed clear upregulation was Screening HOXC-AS3 by bioinformatics analysis. a, b Raw microarray data were downloaded from GEO, including GSE50710 (n = 20) and GSE58828 (n = 6). Then, the raw data were normalized and z-score-transformed using RMAExpress. The top 20 dysregulated lncRNAs in these GEO datasets are shown in a and b HOXA11-AS (the largest increase in GSE50710), and our previous study showed that HOXA11-AS could regulate cell proliferation and invasion of gastric cancer by scaffolding chromatin modification factors and patients with high HOXA11-AS expression had a shorter survival and poorer prognosis [bib_ref] LncRNA HOXA11-AS promotes proliferation and invasion of gastric cancer by scaffolding the..., Sun [/bib_ref]. The most obvious increase in GSE58828, LINC00473, could mediate tumor growth and elevated LINC00473 expression correlated with poor prognosis of lung cancer [bib_ref] cAMP/CREB-regulated LINC00473 marks LKB1-inactivated lung cancer and mediates tumor growth, Chen [/bib_ref]. Our previous study showed that FEZF1-AS1 could promote gastric cancer proliferation and higher expression of FEZF1-AS1 predicted poor prognosis [bib_ref] LincRNAFEZF1-AS1 represses p21 expression to promote gastric cancer proliferation through LSD1-mediated H3K4me2..., Liu [/bib_ref]. Qin et al. found that LINC01296 was upregulated in GC tissue and correlated with poor prognosis [bib_ref] Long intergenic noncoding RNA 01296 aggravates gastric cancer cells progress through miR-122/MMP-9, Qin [/bib_ref]. CCAT1 is an oncogenic long noncoding RNA in human cancers including GC [bib_ref] CCAT1: an oncogenic long noncoding RNA in human cancers, Guo [/bib_ref]. Our previous work found that CCAT1 could regulate cell proliferation and migration in esophageal squamous cell carcinoma and higher expression of CCAT1 is correlated with poorer outcomes [bib_ref] H3K27 acetylation activated-long non-coding RNA CCAT1 affects cell proliferation and migration by..., Zhang [/bib_ref]. In addition, CCAT1 could serve as important prognostic biomarkers in colorectal cancer [bib_ref] CCAT1 and CCAT2 long noncoding RNAs, located within the 8q.24.21 'gene desert',..., Ozawa [/bib_ref]. We focused on overexpressed lncRNAs because these lncRNAs may be more readily used as early diagnosis markers or therapeutic targets. Therefore, although a small portion of lncRNAs have been functionally characterized, many members in the class remain uncharacterized. In our analysis of the results, HOXC-AS3 was the lncRNA clearly upregulated in both GSE50710 and GSE58828 (the rank of HOXC-AS3 is in the exceeding front position, both in GSE50710 and GSE58828), and the role in GC tumorigenesis including other cancer types remained unclear. This prompted us to explore the role of HOXC-AS3 in human GC. To further validate this result, we analyzed that RNA-Seq data (from The Cancer Genome Atlas (TCGA)) of lncRNAs of gastric cancer were from TANRIC [bib_ref] TANRIC: an interactive open platform to explore the function of lncRNAs in..., Li [/bib_ref] (http://ibl.mdanderson.org/tanric/_design/basic/ index.html). As shown in , HOXC-AS3 was significantly upregulated in GC tissues from the TCGA data.
Gain of H3K4me3 and H3K27 acetylation-activated HOXC-AS3 is upregulated in human GC tissues and correlates with poor prognosis For further study, we first performed rapid amplification of cDNA ends (RACE) to identify the full sequence of HOXC-AS3 in BGC-823 cells according the sequence archived in the RefSeq database of NCBI (440 bp, with poly (A) tail, . To validate the expression results from high-throughput data, as shown in , the HOXC-AS3 expression level in tumor tissues was significantly higher in 112 pairs of GC tumor tissues compared with adjacent normal tissues (P < 0.001). One tumor tissue showed an upregulation of HOXC-AS3 greater than 521-fold relative to normal tissue. Next, we explored the correlation between HOXC-AS3 expression and the clinicopathological factors of patients with GC.
The result showed that the HOXC-AS3 level was associated with TNM stage. Patients with advanced TNM stage (III/IV) were associated with higher HOXC-AS3 expression, whereas patients with local TNM stage (I/ II) were associated with a lower HOXC-AS3 level (27.1317 ± 80.26254 vs 4.7483 ± 3.02402, P = 0.046) . Furthermore, we divided the samples into relatively high (above the mean, n = 56) and relatively low (below the mean, n = 56) HOXC-AS3 expression groups according to the median value of HOXC-AS3 levels. A chi-square test was then performed to evaluate clinicopathological factors between the two groups. As shown in Additional file 1: , the relative HOXC-AS3 level was also correlated with histological grade (P = 0.002), tumor invasion depth (P = 0.008), lymph node metastasis (P = 0.035), and TNM stage (P = 0.002). No relationship between HOXC-AS3 expression and other clinical factors, such as sex (male, female) and patient age (≤ 60, > 60), was found in our study.
To determine the relationship between HOXC-AS3 expression and GC patient prognosis, we evaluated the correlation between HOXC-AS3 expression and clinical outcomes. Kaplan-Meier analysis and log-rank test were used to evaluate the effects of HOXC-AS3 expression and the clinicopathological characteristics on overall survival (OS). The median survival time for low HOXC-AS3 expression groups was 34 months, whereas for high HOXC-AS3 expression groups, it was only 20 ± 1.357 months. As shown in , overexpression of HOXC-AS3 predicted a poor prognosis in patients with GC (P = 0.004). Similarly, the correlation between HOXC-AS3 expression levels and the survival of GC patients was also supported by Kaplan-Meier Plotter analysis (http://kmplot.com/analysis/, detailed steps were described in Additional file 8: Supplementary Methods), which indicated that higher HOXC-AS3 expression correlated with worse OS, using publicly available chip data from 631 GC patients .
Then, univariate and multivariate survival analyses (Cox proportional hazards regression model) were performed. Univariate analysis identified two prognostic factors: TNM stage and HOXC-AS3 expression. Multivariate analysis further revealed that HOXC-AS3 expression was an independent predictor for overall survival in patients with GC (P < 0.001), as well as TNM stage (P = 0.023) (Additional file 2: .
To explore the mechanism of high expression of HOXC-AS3 in GC, firstly, by using the UCSC Genome Bioinformatics Site (http://genome.ucsc.edu/), we found high enrichment and overlaps of H3K4me3 and H3K27Ac peaks at the promoter region of HOXC-AS3 [fig_ref] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro [/fig_ref] and H3K27Ac, two markers of active promoters). Using ChIP assays, we found gain of H3K4me3 and H3K27Ac in cancer tissues compared with normal tissues (n = 4) at the promoter of HOXC-AS3. We also observed the gain of H3K4me3 and H3K27Ac in GC cells (BGC-823) compared with normal human esophageal epithelial cells Identification of HOXC-AS3 and expression of HOXC-AS3 in GC tissues and its clinical parameters and gain of H3K4me3 and H3K27 acetylation could activate HOXC-AS3 in GC. a HOXC-AS3 expression in GC tissues (n = 285) compared with noncancerous tissues (n = 33) analyzed using the TCGA database (data from TANRIC). b The full sequence of HOXC-AS3 was confirmed by RACE. c HOXC-AS3 was detected in 112 pairs of GC tissues by qRT-PCR. The levels of HOXC-AS3 in GC tissues are significantly higher than in nontumor tissues. The ΔCt value was determined by subtracting the GAPDH Ct value from the HOXC-AS3 Ct value. A smaller ΔCt value indicates higher expression. d HOXC-AS3 expression was significantly higher in patients with an advanced TNM stage. e Patients with high levels of HOXC-AS3 expression showed reduced survival times compared with patients with low levels of HOXC-AS3 expression (P < 0.01, log-rank test). f Kaplan-Meier survival plots demonstrating that high HOXC-AS3 expression levels correlated with worse OS in GC patients (n = 631). g The UCSC Genome Bioinformatics Site (http://genome.ucsc.edu/) showed high enrichment of H3K4me3 and H3K27Ac at the promoter of HOXC-AS3. ChIP assays detected the level of H3K4me3 and H3K27Ac at the promoter of HOXC-AS3 in GC tissues and cells. *P < 0.05, **P < 0.01 (GES-1) at the promoter of HOXC-AS3 . Taken together, these data confirm that HOXC-AS3 is frequently increased in GC. Abnormal histone modification, gain of histone sites H3K4me3 and H3K27Ac of the promoter may partially account for the significant activation of HOXC-AS3.
## Hoxc-as3 regulates gc cell proliferation and migration in vitro
To explore the role of HOXC-AS3 in GC, we examined the HOXC-AS3 expression levels in gastric cancer cell lines. As shown in [fig_ref] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro [/fig_ref] , gastric cancer cell lines expressed significantly higher levels of HOXC-AS3 than a normal gastric epithelium cell line (GES-1). To further confirm the function of HOXC-AS3, we used an LNA-ASO (Locked Nucleic Acid, antisense oligonucleotide) targeting HOXC-AS3 and control LNA-ASO. As shown in (Additional file 3: , the ASOmediated knockdown and plasmid-mediated overexpression were used for exogenously manipulating the expression of HOXC-AS3 in BGC-823 and SGC-7901 cells. Then, MTT assays showed that knockdown of HOXC-AS3 expression significantly inhibited cell proliferation compared with control cells. By contrast, overexpressed HOXC-AS3 promoted cell proliferation [fig_ref] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro [/fig_ref]. EdU assays demonstrated that HOXC-AS3 had a significant impact on GC cell proliferation [fig_ref] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro [/fig_ref]. Similarly, the results of the colony-formation assay revealed that clonogenic survival was significantly decreased following knockdown of HOXC-AS3. Furthermore, overexpression of HOXC-AS3 could increase the number of clones [fig_ref] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro [/fig_ref]. Next, transwell assays revealed that knockdown of HOXC-AS3 significantly repressed cell migration compared with the control in BGC-823 cells. By contrast, overexpression of HOXC-AS3 promoted cell migration [fig_ref] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro [/fig_ref]. Flow cytometric analysis was performed to further examine whether HOXC-AS3 affected the proliferation of GC cells by altering cell-cycle progression. The results revealed that cell-cycle progression of knockdown-HOXC-AS3 cells was significantly stalled at the G1-G0 phase compared with controls both in BGC-823 and in SGC-7901 cells. Overexpressed HOXC-AS3 promoted S phase progression [fig_ref] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro [/fig_ref]. Additionally, knockdown of HOXC-AS3 significantly induced BGC-823 and SGC-7901 cell apoptosis [fig_ref] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro [/fig_ref].
## Hoxc-as3 regulates gc cell proliferation and migration in vivo
To further determine whether HOXC-AS3 affects cell proliferation of GC in vivo, BGC-823 cells and SGC-7901 stably transfected with sh-HOXC-AS3 or control vector were inoculated into nude mice. Sixteen days after the injection, the tumors formed in the sh-HOXC-AS3 group were substantially smaller than those in the control group [fig_ref] Figure 4: HOXC-AS3 regulates GC cell proliferation and migration in vivo [/fig_ref]. Moreover, the mean tumor weight at the end of the experiment was markedly lower in the sh-HOXC-AS3 group compared with the control vector group [fig_ref] Figure 4: HOXC-AS3 regulates GC cell proliferation and migration in vivo [/fig_ref]. Tumors formed from stably sh-HOXC-AS3-transfected BGC-823 and SGC-7901 cells exhibited decreased positivity for Ki-67 than tumors from the control cells [fig_ref] Figure 4: HOXC-AS3 regulates GC cell proliferation and migration in vivo [/fig_ref]. These findings indicate that knockdown of HOXC-AS3 inhibits tumor growth in vivo.
To validate the effects of HOXC-AS3 on cell metastasis in vivo, BGC-823 cells stably transfected with sh-HOXC-AS3 or control vector were injected into the tail veins of nine mice. Metastatic nodules on the surface of the lungs were counted after 7 weeks. Ectopic knockdown of HOXC-AS3 reduced the number of metastatic nodules compared with the control group [fig_ref] Figure 4: HOXC-AS3 regulates GC cell proliferation and migration in vivo [/fig_ref]. This difference was further confirmed following examination of the entire lungs and by hematoxylin and eosin (HE) staining of lung sections [fig_ref] Figure 4: HOXC-AS3 regulates GC cell proliferation and migration in vivo [/fig_ref]. Our in vivo data, therefore, complemented the results of the functional in vitro studies involving HOXC-AS3.
## Hoxc-as3 interacts with ybx1
To explore the mechanism of HOXC-AS3, firstly, we examined whether HOXC-AS3 acts in cis, affecting nearby gene expression (HOXC10; HOXC-AS3 was an antisense transcript of HOXC10). We found that knockdown of HOXC-AS3 did not affect expression of HOXC10 (data not shown), indicating it may act in trans. Then, subcellular fractionation location assays showed a considerable increase in HOXC-AS3 expression in the nucleus versus the cytosol , thus suggesting that HOXC-AS3 may play a major regulatory function at the transcriptional level, indicating it might interact with nucleus molecules or proteins. Then, we performed an RNA pull-down assay followed by a proteomic analysis of the HOXC-AS3-associated protein complex in BGC-823 cells . We incubated the in vitro transcribed HOXC-AS3 bound to beads with BGC-823 nuclear extracts to purify the HOXC-AS3 RNA-protein complex, and protein identity was revealed by mass spectrometry. Among the highly enriched proteins (Additional file 4: , only YBX1 was detected by western blotting from three independent RNA pull-down assays. YBX1 attracted our attention because of its established role in tumorigenesis and this notable protein was a known transcription factor [bib_ref] Shrimp miR-S8 suppresses the stemness of human melanoma stem-like cells by targeting..., Yang [/bib_ref] [bib_ref] Endogenous tRNA-derived fragments suppress breast cancer progression via YBX1 displacement, Goodarzi [/bib_ref] , with 12 unique peptides detected in this MS analysis. To further validate the physical interaction between HOXC-AS3 and YBX1, we performed RNA pull-down followed by western blotting with YBX1 antibodies. Our results showed that labeled HOXC-AS3 RNA, but not empty vector or an antisense HOXC-AS3, specifically retrieved YBX1 from BGC-823 cell extracts . We also performed RNA immunoprecipitation (RIP) for the RNA-YBX1 complex using different YBX1 antibodies and measured the amount of HOXC-AS3 associated with YBX1 immunoprecipitates but not MEG3. LncRNA MEG3 was a negative control . Using a series of HOXC-AS3 deletion mapping, the YBX1-binding activity mapped to nucleotides 115 nt of HOXC-AS3 . Additionally, the qRT-PCR and western blots showed that knockdown of HOXC-AS3 could have had no effect on the expression of YBX1 . Our results proved that HOXC-AS3 could bind to YBX1 but have no effect on YBX1 expression. These results indicated that HOXC-AS3 may participate in the tumorigenesis of GC through transcriptional regulation of other genes via binding to YBX1.
A large set of genes that are linked to cell proliferation and cell migration were coregulated by interaction of HOXC-AS3 and YBX1
To ascertain the mechanism of action of HOXC-AS3/ YBX1-driven oncogenesis of GC and to determine the gene expression changes downstream of HOXC-AS3, we evaluated the global effects of HOXC-AS3 knockdown compared with those of YBX1 knockdown using RNA transcriptome sequencing , Additional file 5: and Additional file 6: . A total of 3122 genes were significantly upregulated or downregulated in BGC-823 cells as a consequence of HOXC-AS3 knockdown (fold change ≥ 1.5) . For RNA-seq results of HOXC-AS3 knockdown, gene ontology (GO) analysis showed that the most significantly overrepresented biological processes included pathways involved in cell division, cell proliferation, and cell death . Gene set enrichment analysis (GSEA) revealed that the gene sets were significantly related to cell proliferation and metastasis . We also built a coexpression network based on the differentially expressed genes that are functionally related. The co-expression network was constructed according to the gene normalized expression (RPKM), and the k-core was calculated based on Pearson's relationship to represent the core status of each gene in different groups. The k-core factor was then applied to identify key regulatory genes, which likely play pivotal roles in gene interactions and regulation , including for key cancer genes, such as p21, FAS, CCND1, and CDK2; real-time PCR confirmed these genes after knockdown of HOXC-AS3 (Additional file 3: . Additionally, a total of 1480 genes were significantly upregulated or downregulated in BGC-823 cells as a consequence of YBX1 knockdown (fold change ≥ 1.5) and Additional file 3: . Intriguingly, the expression of 805 of the 3122 HOXC-AS3-regulated genes was also regulated by knockdown of YBX1 . The high overlap between HOXC-AS3 and YBX1-mediated regulation strongly indicates a functional interaction between HOXC-AS3 and YBX1. YBX1 functions as an RNA-binding protein and has been implicated in numerous cellular processes, including the regulation of transcription and translation [bib_ref] Y-box binding protein 1-a prognostic marker and target in tumour therapy, Kosnopfel [/bib_ref]. YBX1 was shown to be a protein with a nucleic acid-binding common domain in gene promoter, CCAA T-box, which is a high consensus sequence in eukaryotes [bib_ref] Eukaryotic cold shock domain proteins: highly versatile regulators of gene expression, Mihailovich [/bib_ref]. Moreover, YBX1 serves as a transcriptional activator and YBX1 activation was associated with cancer progression, including gastric cancer [bib_ref] Endogenous tRNA-derived fragments suppress breast cancer progression via YBX1 displacement, Goodarzi [/bib_ref] [bib_ref] Identification of Y-box binding protein 1 as a core regulator of MEK/ERK..., Jurchott [/bib_ref] [bib_ref] Human helicase RECQL4 drives cisplatin resistance in gastric cancer by activating an..., Mo [/bib_ref]. Our results showed that YBX1 was significantly upregulated in 60 pairs of GC tissues by qRT-PCR analysis. Knockdown of YBX1 could significantly inhibit cell proliferation and migration (Additional file 3: . Previous research supports a role for lncRNA-YBX1 interactions as mediators of gene transcriptional regulation [bib_ref] Epigenetic inactivation of the p53-induced long noncoding RNA TP53 target 1 in..., Diaz-Lagares [/bib_ref]. In particular, 476 genes were corepressed by knockdown of HOXC-AS3 and YBX1. These results indicated that a common set of target genes linked to cell proliferation and cell migration were shared by HOXC-AS3 and YBX1.
Genes were corepressed by knockdown of HOXC-AS3/ YBX1. They included many well-known proliferation and migration-related genes (e.g., CDK2, HOXB13, IGFBP4, ATF5, MAPK4, MMP7, MMP24, BIRC2, WNT10B, and HDAC5). These genes were selectively confirmed by qRT-PCR after knockdown of HOXC-AS3 and YBX1 in BGC-823 and SGC-7901 cells (all these genes containing CCAAT-box, YBX1-binding site in their promoters) . Given that YBX1 regulates gene transcription by binding to the promoter regions of target genes, we examined whether HOXC-AS3 knockdown affected YBX1 occupancy of the promoter regions in these target genes. Then ChIP assays followed by qPCR demonstrated that knockdown of HOXC-AS3 decreased the binding of YBX1 levels across the promoters of the most coregulated genes among the ten target genes tested, YBX1 occupancy were indeed affected in the promoter regions of seven genes following knockdown of HOXC-AS3). These findings demonstrate that HOXC-AS3 is important in regulating gene transcription expression, presumably in part by regulating the association between YBX1 and the promoter regions of its target genes. HOXC-AS3 Interacts with YBX1. a After nuclear and cytosolic separation, RNA expression levels were measured by qRT-PCR. GAPDH was used as a cytosol marker and U6 was used as a nucleus marker. b For the in vitro transcribed, pull-down assays-MS and WB assays showed that desthiobiotinylation-HOXC-AS3 could bind YBX1 in BGC-823 cells. c RIP experiments for YBX1 (two different antibodies from Abcam) were performed and the coprecipitated RNA was subjected to qRT-PCR for HOXC-AS3. The fold enrichment of HOXC-AS3 in RIPs is relative to its matching IgG control RIP. LncRNA MEG3 was a negative control. d Using a series of HOXC-AS3 deletion mapping constructs and RNA pull-down-WB assays, the YBX1-binding activity mapped to nucleotides 115 nt of HOXC-AS3. The profiles are established by RNA pulldown of the BGC-823 extract, and the retrieved proteins are detected by immunoblotting. e The qPCR and western blot assays detected the expression of YBX1 after knockdown of HOXC-AS3. *P < 0.05, n.s., not significant
## Hoxc-as3 regulates the expression of hdac5
Among the common targets of HOXC-AS3 and YBX1, HDAC5 is of particular interest because of its remarkable expression fold change upon HOXC-AS3 knockdown and its significant contribution to tumorigenesis [bib_ref] Histone deacetylases and their inhibitors in cancer, neurological diseases and immune disorders, Falkenberg [/bib_ref]. Aberrant activation of HDAC5 in tumor cells leads to dysregulation of a diverse set of genes mainly involved in the regulation of proliferation, migration, and angiogenesis [bib_ref] Histone deacetylase inhibitors and the promise of epigenetic (and more) treatments for..., Minucci [/bib_ref]. Our results showed that HDAC5 was significantly upregulated in 60 pairs of GC tissues. Further analysis demonstrated that HOXC-AS3 was positively correlated with HDAC5 expression in GC tissues [fig_ref] Figure 8: HOXC-AS3 regulates expression of HDAC5 [/fig_ref]. Then, we verified the results at the protein level for HDAC5 after knockdown of HOXC-AS3 [fig_ref] Figure 8: HOXC-AS3 regulates expression of HDAC5 [/fig_ref].
Furthermore, the increase of HDAC5 by HOXC-AS3 was reversed with treatment by YBX1 siRNAs at both the mRNA and protein levels [fig_ref] Figure 8: HOXC-AS3 regulates expression of HDAC5 [/fig_ref]. Moreover, knockdown of HDAC5 induced the suppression of proliferation and migration in BGC-823 cells [fig_ref] Figure 8: HOXC-AS3 regulates expression of HDAC5 [/fig_ref] and Additional file 3: . In addition, knockdown of HDAC5 could reverse HOXC-AS3-mediated growth and migration promotion [fig_ref] Figure 8: HOXC-AS3 regulates expression of HDAC5 [/fig_ref]. Together, these results demonstrate that HOXC-AS3 exerts its function, at least in part by regulating HDAC5 expression.
# Discussion
Before the discovery of noncoding RNAs, explorations for cancer drivers focused on protein-coding genes that Downstream gene network of HOXC-AS3 and a common set of target genes shared by HOXC-AS3 and YBX1. a Mean-centered, hierarchical clustering of gene transcripts altered (≥ 1.5-fold change) after knockdown of HOXC-AS3 and YBX1 in BGC-823 cells, with three replicates. b Gene ontology analysis for all genes with altered expression after knockdown of HOXC-AS3. c GSEA showed that genes in response to HOXC-AS3 knockdown are enriched for the gene sets significantly related to cell proliferation and metastasis. d Coexpression-network analysis of differentially expressed genes for knockdown of HOXC-AS3. The solid lines represent positive Pearson's correlation, whereas the dotted lines represent negative Pearson's correlation. The node size and node color represent the coexpression ability k-core of gene, the greater the node size, the greater the k-core value resided in recurrent alterations in cancer genomes. However, the newly discovered lncRNAs have emerged as important players in cellular development and human diseases, especially in cancer. In the present study, utilizing publicly available lncRNA expression profiling data of gastric cancer and integrating analyses of TCGA data, we screened and identified a novel lncRNAs HOXC-AS3. The high expression of HOXC-AS3 in GC patients was positively correlated with advanced TNM stage. Moreover, high HOXC-AS3 expression in GC tissues was associated with a poor prognosis and could be an independent prognostic indicator. In addition, GTEx data (https://www.gtexportal.org/) showed that HOXC-AS3 had lower basal expression in normal GC tissues (Additional file 3: . This further demonstrates an important role of HOXC-AS3 in the carcinogenesis of GC. These results and our functional evidence for HOXC-AS3 suggested that HOXC-AS3 might exhibit an important role in GC progression.
HOXC-AS3, which is located at chromosome 12q13.13, was an antisense transcript of HOXC10. HOX genes are essential for morphogenesis and development [bib_ref] Modulating Hox gene functions during animal body patterning, Pearson [/bib_ref] [bib_ref] Functional demarcation of active and silent chromatin domains in human HOX loci..., Rinn [/bib_ref]. The dysregulation of HOX gene expression has been shown in many diverse cancers [bib_ref] The Hox genes and their roles in oncogenesis, Shah [/bib_ref] [bib_ref] Evidence for a functional role of epigenetically regulated midcluster HOXB genes in..., Di Pietro [/bib_ref] , and lncRNA generation in HOX genes may play important roles in tumorigenesis. For example, as a well-known lncRNA, HOTAIR, also located at chromosome 12q13.13, is an antisense transcript of HOXC11 and was an oncogenic lncRNA in many different types of cancer [bib_ref] LncRNA HOTAIR as prognostic circulating marker and potential therapeutic target in patients..., Botti [/bib_ref]. Our previous study also found that a lncRNA of the HOX gene family, HOXA11-AS, could play an important role in GC tumorigenesis [bib_ref] LncRNA HOXA11-AS promotes proliferation and invasion of gastric cancer by scaffolding the..., Sun [/bib_ref]. In our present study, we found a novel lncRNA in the HOX genes family. Our results revealed that gain of H3K4me3 and H3K27Ac activation of the promoter also partly contributed to activation of HOXC-AS3 in GC, both in cells and tissues. Similar to protein coding transcripts, the transcription of lncRNAs is subject to typical epigenetics-mediated and transcription factormediated regulation. For example, the lncRNA MEG3 was lost in tumors due to an increase in CpG methylation within the promoter [bib_ref] MEG3 noncoding RNA: a tumor suppressor, Zhou [/bib_ref]. Histone deacetylase3-suppressed the lncRNA LET in hepatocellular carcinoma by reducing the histone acetylation-mediated modulation of the promoter region [bib_ref] Repression of the long noncoding RNA-LET by histone deacetylase 3 contributes to..., Yang [/bib_ref].
In our study, we found that inhibition of HOXC-AS3 repressed GC proliferation and migration both in vitro and in vivo. RNA-seq found that knockdown of HOXC-AS3 affected key cancer-related genes, such as p21, FAS, and CCND1. Mechanistic investigations found that HOXC-AS3 could bind to YBX1, but not affect YBX1 expression. These results indicated that HOXC-AS3 may participate in the tumorigenesis of GC through the HOXC-AS3 regulates the transcription of a large set of genes through an interaction with YBX1. a The altered mRNA levels of genes were selectively confirmed by qRT-PCR in knockdown HOXC-AS3 and YBX1. b ChIP-qPCR of YBX1 of the promoter region of these gene loci after knockdown HOXC-AS3 in BGC-823 cells. Antibody enrichment was quantified relative to the amount of input DNA. Antibody directed against IgG was used as a negative control. *P < 0.05, **P < 0.01. n.s., not significant transcriptional regulation of other genes via binding to YBX1 in trans. To probe the HOXC-AS3-associated pathway on an unbiased basis in the tumorigenesis of GC, RNA-Seq assays were used after simultaneous knockdown HOXC-AS3 and YBX1. Notably, loss of HOXC-AS3 in GC cells recapitulated the phenotype observed after YBX1 knockdown. In addition, a significant fraction of the genes regulated by HOXC-AS3 loss were similarly regulated by loss of YBX1. Thus, HOXC-AS3 may act, in part, by regulating the interaction between a Based on qRT-PCR assays, the level of HDAC5 was upregulated in 60 pairs GC tissues. The ΔCt value was determined by subtracting the GAPDH Ct value from the HDAC5 Ct value. A smaller ΔCt value indicates higher expression. b The level of HOXC-AS3 in GC tissues showed a significant positive correlation with the relative level of HDAC5 expression (n = 60). c Western blot assays detected the expression of HDAC5 after knockdown of HOXC-AS3 in BGC-823 cells. The promotion of HDAC5 (mRNA and protein) by HOXC-AS3 was significantly reversed by knockdown of YBX1, based on qPCR and western blot assays. d MTT assays showed that knockdown of HDAC5 could suppress cell proliferation, and knockdown of HDAC5 could also reverse HOXC-AS3-mediated growth promotion. e Proposed model in which HOXC-AS3 mediates the proliferation and migration progression of GC YBX1 and the promoter of target genes, although HOXC-AS3 likely interacts with other RNA-binding proteins that will need to be identified to fully understand its molecular function. YBX1 was reported to play a role in regulating cell signaling, transcription, and tumorigenesis [bib_ref] Endogenous tRNA-derived fragments suppress breast cancer progression via YBX1 displacement, Goodarzi [/bib_ref]. Furthermore, YBX1 serves as a transcriptional activator and regulates much gene transcription [bib_ref] RNAbinding specificity of Y-box protein 1, Dong [/bib_ref]. YBX1 was a protein with a nucleic acid-binding common domain in the gene promoter, CCAAT-box, which is a high consensus sequence in eukaryotes. We found that YBX1 is overactive in GC and knockdown of YBX1 inhibits the proliferation of GC cells. Our results showed that HOXC-AS3 could bind to YBX1, thus transcriptionally regulating a large set of genes that are linked to cell proliferation and cell migration in gastric cancer cells, such as MMP7, WNT10B, and HDAC5, thus promoting GC cell proliferation and migration.
The HDAC5 gene is a member of the histone deacetylase (HDAC) from a family of enzymes. Histone acetylation and deacetylation play important roles in chromatin remodeling and gene expression. An imbalance of these reactions leads to the growth, migration, and apoptosis of cancer cells. Histone deacetylase (HDAC) inhibitors were shown to have antitumor effects in clinical trials [bib_ref] Inhibition of histone deacetylase activity down-regulates urokinase plasminogen activator and matrix metalloproteinase-9..., Lee [/bib_ref] [bib_ref] HDAC5 is required for maintenance of pericentric heterochromatin, and controls cell-cycle progression..., Peixoto [/bib_ref]. Over-activation of HDAC5 was found in many different types of cancer [bib_ref] Functional interaction of histone deacetylase 5 (HDAC5) and lysine-specific demethylase 1 (LSD1)..., Cao [/bib_ref] [bib_ref] HDAC up-regulation in early colon field carcinogenesis is involved in cell tumorigenicity..., Stypula-Cyrus [/bib_ref] [bib_ref] HDAC5 promotes cell proliferation in human hepatocellular carcinoma by upregulating Six1 expression, Feng [/bib_ref]. A previous study showed that HDAC5 was induced in gastric cancer cells [bib_ref] Inhibition of histone deacetylase activity down-regulates urokinase plasminogen activator and matrix metalloproteinase-9..., Lee [/bib_ref]. Here, we also provide evidence for high expression of HDAC5 in gastric cancer, and knockdown of HDAC5 inhibits the proliferation of GC cells. We also found that the transcriptional activation of HDAC5 is partly mediated by HOXC-AS3 in the tumor progression of GC through binding to YBX1, thus facilitating GC cell proliferation and migration. In addition to HDAC5, there are many other important genes related to tumorigenesis, and they are also regulated by HOXC-AS3 in a similar manner.
# Conclusions
In summary, abnormal histone modification-mediated activation of a novel lncRNA HOXC-AS3 promotes GC cell proliferation and migration through transcriptional activation of a large set of genes through an interaction with YBX1. Our data reveal a role for HOXC-AS3 in GC tumorigenesis and may provide a strategy for using HOXC-AS3 as a potential biomarker and a therapeutic target for patients with GC [fig_ref] Figure 8: HOXC-AS3 regulates expression of HDAC5 [/fig_ref].
# Methods
## Tissue collection and ethics statement
A total of 112 patients in this study underwent resection of the GC at The Affiliated Jiangyin Hospital of Southeast University Medical College, affiliated Xuzhou Central Hospital of Southeast University Medical College.
The study was approved by the Medical Ethical Committee of Southeast University Medical College (Nanjing, Jiangsu, PR China), and it was performed in compliance with the Helsinki Declaration. All patients have given written informed consent for publication. The clinicopathological characteristics of the GC patients are summarized in Additional file 1: .
Gastric cancer RNA-expression data retrieval and analysis Microarray data analysis Microarray datasets from the GEO database were used to test HOXC-AS3 differential expression. Raw microarray data was downloaded from GEO including GSE50710 and GSE58828. Then, the raw microarray data were normalized and z-score-transformed using RMAExpress (http://www.rmaexpress.bmbolstad.com/). RNA-Seq data (from TCGA) of lncRNAs of gastric cancer were from TANRIC (http://ibl.mdanderson.org/tanric/_design/basic/index.html) [bib_ref] TANRIC: an interactive open platform to explore the function of lncRNAs in..., Li [/bib_ref].
## Race (rapid amplification of cdna ends)
5′-RACE, 3′-RACE, and full-length amplification of HOXC-AS3 were performed using the SMART RACE cDNA Amplification Kit (Cat. 634858, Clontech, Palo Alto, CA, USA) according to the manufacturer's instructions. The gene-specific primers used for RACE analysis were presented in Additional file 7: .
## Transfection of cell lines
LNA-ASO (Locked Nucleic Acid, antisense oligonucleotide) targeting HOXC-AS3 and negative control LNA-ASO were designed and synthesized by Exiqon (Exiqon, Vedbaek, Denmark). GC cells were transfected with the LNA-ASOs using Oligofectamine transfection reagent (RNAi MAX, Invitrogen) according to the manufacturer's instructions. Cells were harvested for analyses 48 h after transfections. The sequences of ASO were listed in Additional file 7: . The sequences for siRNAs were listed in Additional file 7: . Scrambled negative control siRNA was purchased from Invitrogen (Invitrogen, CA, USA). The interference target sequence of YBX1 was acquired according to a previous study [bib_ref] YB-1 evokes susceptibility to cancer through cytokinesis failure, mitotic dysfunction and HER2..., Davies [/bib_ref]. The HDAC5 siRNA was from Santa Cruz (sc-35542). The plasmid was transfected into GC cells using the X-tremeGENE™ HP DNA Transfection Reagent (Roche) according to the manufacturer's instructions.
## Subcellular fractionation location
Separation of the nuclear and cytosolic fractions was performed using the PARIS Kit (Cat. AM1921, Invitrogen, CA, USA) according to the manufacturer's instructions.
[fig] Figure 3: HOXC-AS3 regulates GC cell proliferation and migration in vitro. a Analysis of HOXC-AS3 expression levels in GC cell lines (SGC-7901, BGC-823, MGC-803, and SNU-601) compared with a normal gastric epithelium cell line (GES-1) by qRT-PCR. b MTT assays were performed to determine cell proliferation of BGC-823 and SGC-7901 cells after transfection of overexpression plasmid and ASO of HOXC-AS3. c EdU assays of BGC-823 and SGC-7901 cells transfected with an overexpression plasmid and knockdown of HOXC-AS3. d Colony formation assays of BGC-823 and SGC-7901 cells transfected with an overexpression plasmid and knockdown of HOXC-AS3. e Transwell assays were used to investigate the changes in migratory abilities of BGC-823 cells after transfection, respectively. f At 48 h after transfection, cell cycle was analyzed by flow cytometry. The bar chart represents the percentage of cells in G1-G0, S, or G2-M phase, as indicated. g At 48 h after transfection, BGC-823 and SGC-7901 cells were stained and analyzed by flow cytometry. LR, early apoptotic cells. UR, terminal apoptotic cells. *P < 0.05, **P < 0.01 [/fig]
[fig] Figure 4: HOXC-AS3 regulates GC cell proliferation and migration in vivo. a Scramble or shHOXC-AS3 was stably transfected into BGC-823 and SGC-7901 cells, which were injected into nude mice, respectively. Tumor volumes were calculated after injection every 2 days. Tumor weights are represented as the means of tumor weights ± S.D. (standard deviation). The qRT-PCR was performed to detect the average expression of HOXC-AS3. Immunohistochemistry analysis revealed that the tumors developed from shHOXC-AS3 cells displayed lower Ki-67 staining than the control group. b Analysis of an experimental metastatic animal model was performed by injecting BGC-823 cells stably transfected with HOXC-AS3 knockdown into nude mice. Lungs from the mice in each experimental group, together with the number of tumor nodules on lung surfaces are shown. Visualization of the entire lung and HE-stained lung sections. *P < 0.05, **P < 0.01 [/fig]
[fig] Figure 5: HOXC-AS3 Interacts with YBX1. a After nuclear and cytosolic separation, RNA expression levels were measured by qRT-PCR. GAPDH was used as a cytosol marker and U6 was used as a nucleus marker. b For the in vitro transcribed, pull-down assays-MS and WB assays showed that desthiobiotinylation-HOXC-AS3 could bind YBX1 in BGC-823 cells. c RIP experiments for YBX1 (two different antibodies from Abcam) were performed and the coprecipitated RNA was subjected to qRT-PCR for HOXC-AS3. The fold enrichment of HOXC-AS3 in RIPs is relative to its matching IgG control RIP. LncRNA MEG3 was a negative control. d Using a series of HOXC-AS3 deletion mapping constructs and RNA pull-down-WB assays, the YBX1-binding activity mapped to nucleotides 115 nt of HOXC-AS3. The profiles are established by RNA pulldown of the BGC-823 extract, and the retrieved proteins are detected by immunoblotting. e The qPCR and western blot assays detected the expression of YBX1 after knockdown of HOXC-AS3. *P < 0.05, n.s., not significant [/fig]
[fig] Figure 8: HOXC-AS3 regulates expression of HDAC5. [/fig]
|
Coronary sinus as a site for stable temporary atrial pacing to tide over premature ventricular complex–triggered recurrent ventricular fibrillation in a patient with severe left ventricular dysfunction after coronary bypass surgery
Coronary sinus as a site for stable temporary atrial pacing to tide over premature ventricular complexetriggered recurrent ventricular fibrillation in a patient with severe left ventricular dysfunction after coronary bypass surgery a b s t r a c t Use of atrial pacing has been known, yet underutilized tool for effective temporary pacing whenever needed early after cardiac surgery. The reasons may be frequent failures of epicardial wires (fixed over atria intra-operative) leading to loss of capture. Endocardial atrial pacing sites for temporary pacing are unstable and hence continuous pacing with acceptable thresholds is impossible. We describe a case of ischemic cardiomyopathy and severe left ventricle systolic dysfunction who required atrial pacing post coronary artery bypass grafting (CABG) surgery for around 48e72 hours starting 4th day post operation when he had multiple episodes of ventricular fibrillation (VF) needing many defibrillation shocks. VF episodes were triggered by premature ventricular complexes (PVC) falling on as R on T. Episodes were refractory to anti-arrhythmic drugs and general anesthesia. These PVC were successfully overdriven by atrial pacing by fluoroscopically placing pacing lead in coronary sinus. This led to complete suppression of PVC induced VF for next 48e72 hours while the antiarrhythmic drugs were continued. Subsequently the patient was discharged in stable state. Atrial pacing, though quite valuable during the post-operative period after cardiac surgery, is underutilized, especially when pacing through surgically placed epicardial wire fails. This report is helpful in drawing attention towards coronary sinus (CS) as an alternate site for achieving stable, temporary atrial pacing during the post-operative period. This site can also be utilized for short term dual chamber pacing if required in post-operative state using transvenous CS site for atrial pacing and intraoperatively placed epicardial wire for ventricular pacing.
A 45-year-old gentleman presented with recent anterior wall myocardial infarction (AWMI). The baseline electrocardiogram showed changes suggestive of evolved AWMI, and the echocardiogram showed severe left ventricle (LV) dysfunction with an ejection fraction of 0.3. The coronary angiogram revealed critical triple-vessel disease for which he underwent off-pump coronary artery bypass graft (CABG) surgery.
After an initial uneventful recovery, he had recurrent episodes of ventricular fibrillation (VF) requiring multiple defibrillation shocks on the fourth day after CABG. He was given intravenous amiodarone, lignocaine, metoprolol, and magnesium sulfate in usual protocols and doses. He was put on mechanical ventilation with general anesthesia. Despite this, he continued to have recurrent VF requiring multiple defibrillation shocks (27 in a span of 3 hours!). All episodes were triggered by short-coupled premature ventricular complexes (PVCs) with R-on-T phenomenon [fig_ref] Figure 1: Monitor images of rhythm recordings showing multiple episodes of ventricular fibrillation [/fig_ref]. Ventricular overdrive pacing by the epicardial pacing wire placed during CABG failed to suppress the PVCs and subsequent VF. Epicardial wire for atrial pacing was not capturing even at maximum outputs. Transesophageal echo did not show any new regional wall motion abnormality to suggest new-onset ischemic insult. The patient was shifted to catheterization laboratory for temporary atrial pacing as an attempt to suppress the PVCs. Patient already had a right internal jugular (triple lumen) venous line in situ, so femoral access was sought. Continuous pacing through a lead placed in the right atrium was not possible due to unstable position of the lead in the right atrial (RA) cavity with resultant intermittent loss of capture; this occurred despite positioning it in the RA appendage. Hence, the pacing lead was positioned in coronary sinus (CS). The nondeflectable temporary pacing lead could not be negotiated directly, and hence, it was advanced after first placing a deflectable decapolar electrophysiology (EP) catheter in CS and then advancing the temporary lead as a "buddy technique" [fig_ref] Figure 2: Left upper panel [/fig_ref] , left upper panel). Then, the EP catheter was withdrawn [fig_ref] Figure 2: Left upper panel [/fig_ref] , left lower panel). Stable continuous left atrial capture was confirmed with acceptable pacing threshold. The patient was put on overdrive atrial pacing at the rate of 110/min. This led to complete suppression of PVC and subsequent VF [fig_ref] Figure 2: Left upper panel [/fig_ref] , right panel). The patient was kept on mechanical ventilation and general anesthesia for 24 hours and then extubated after weaning over next 36 hours. The atrial pacing rate was reduced slowly at every 4-hour intervals till spontaneous sinus rhythm persisted, without PVCs. The temporary pacing lead was withdrawn after 72 hours. He was anticoagulated with subcutaneous enoxaparin for deep vein thrombosis prophylaxis during the period of temporary pacing. His remaining hospital stay was uneventful. He was discharged in a stable state.
The PVCs that precede VF initiation have been believed to be the mechanism that triggers VF episodes in arrhythmic storm, and VF recurrence has been avoided by suppressing these PVC by either drugs/pacing/ablation. [bib_ref] Premature ventricular complexes as a trigger for ventricular fibrillation, Muñoz [/bib_ref] Analysis of Multicenter Automatic Defibrillator Implantation Trial II-type patients revealed that PVCs triggered 77% of VF episodes, in contrast to a short-long-short sequence found in 23% of patients. [bib_ref] Prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced..., Moss [/bib_ref] Usually surviving Purkinje fibers with shortened action potential duration or reduced amplitude, depolarized membrane potentials, and reduced V max are thought to be the source of automatic foci for PVC triggering VF in such scenarios. Polymorphic ventricular tachycardia (VT)/VF in these patients is triggered and possibly maintained by activity originating from the distal Purkinje arborization localized to the border zone of the myocardial infarction (MI) in which tissue heterogeneity is particularly marked. Interestingly, the Purkinje tissue cells have been demonstrated to be able to survive transmural infarction in experimental models, leading to speculation that their proximity to the endocardium allows exposure to cavitary blood. These surviving Purkinje fibers crossing the border zone of the infarct demonstrate heightened automaticity, triggered activity, and supernormal excitability, which, when coupled with prolongation of the action potential duration in this region may result in the necessary milieu for polymorphic VT. At times, reperfusion injury after revascularization is responsible for such arrhythmia. Electrolyte imbalance (especially hypokalemia), hypothermia, and inotropic support may be other causes responsible for ventricular arrhythmia in postoperative state. [bib_ref] Prognosis of patients with frequent premature ventricular complexes and nonsustained ventricular tachycardia..., Pinto [/bib_ref] If antiarrhythmic drugs and sedation fail to suppress Purkinjerelated PVC triggering VF in the early post-MI period, temporary overdrive pacing may be a reasonable simpler alternative to emergency ablation strategies targeting PVC. By overdrive atrial pacing, an automatic focus of PVC may be captured and suppressed or exit block achieved by making the surrounding myocardium refractory. However, the obvious prerequisite for using atrial pacing to suppress ventricular arrhythmias is 1:1 atrioventricular (AV) conduction at fast rates, and once this is met, it can be extremely useful in refractory VF triggered by PVCs to reduce the need for recurrent defibrillation while waiting for drug therapy to take effect. Even in patients for whom 1:1 AV conduction is not possible, dual chamber AV sequential overdrive pacing using transvenous CS site and surgically placed epicardial ventricular wire can be tried. Overdrive atrial or a dual-chamber pacing appears superior in this respect to isolated ventricular pacing for two reasons: (i) overdrive ventricular pacing may itself become a source for triggering recurrent ventricular arrhythmia and (ii) ventricular pacing leads to suboptimal cardiac outputs compared with atrial pacing which is more physiological.
Atrial pacing in the early postoperative period may also be helpful in overdrive suppression of atrial flutter/tachycardia or using it as hemodynamically superior AV sequential dual-chamber pacing over isolated ventricular pacing whenever required for bradycardia/heart blocks. Hence, its routine use in almost all (but especially in those with LV dysfunction, high risk of AV block, atrial or ventricular arrhythmias) patients with cardiac surgery is relevant.
The Achilles heel in temporary RA transvenous pacing is unstable position and failure of continuous capture; latter may lead to intermittent PVC triggering VF. Positioning the pacing lead inside the CS has to be careful, and the advancement should be rather passive with beat-to-beat movement of lead, even slight forceful maneuver may lead to CS dissection, especially at the site of valve. For continuous atrial capture, the lead should be fluoroscopically ensured to be in main body of CS; sometimes, a venogram with an Amplatz Left (AL1) catheter may be helpful for delineating CS anatomy. For the same reason, lead placements into a venous branch and hence inadvertent ventricular capture or oversensing of ventricular signals have to be avoided. The CS can be a safe nontraumatic site for stable temporary atrial transvenous (jugular/ femoral) pacing, ensuring continuous capture for as long as 72 hours as has been shown in this case with complete suppression of PVCs and resultant VF.
[fig] Figure 1: Monitor images of rhythm recordings showing multiple episodes of ventricular fibrillation (VF) requiring defibrillation shocks. All the episodes were triggered by similar morphology short-coupled PVC (white arrow) triggering VF as R-on-T phenomenon. [/fig]
[fig] Figure 2: Left upper panel (left anterior oblique fluoroscopic view) shows placement of temporary pacing lead with help of deflectable decapolar electrophysiology catheter as a buddy technique. Left lower panel (left anterior oblique fluoroscopic view) shows after withdrawing decapolar and position of pacing lead in CS. Right panel shows continuous stable atrial pacing and complete suppression of PVC-triggered VF. CS, coronary sinus. [/fig]
|
A novel gingipain regulatory gene in Porphyromonas gingivalis mediates host cell detachment and inhibition of wound closure
The black pigmentation-related genes in Porphyromonas gingivalis are primarily involved in regulating gingipain functions. In this study, we identified a pigmentation-related gene, designated as pgn_0361. To characterize the role of pgn_0361 in regulating P. gingivalis-mediated epithelial cell detachment and inhibition of wound closure, PgΔ0361, an isogenic pgn_0361-defective mutant strain, and PgΔ0361C, a complementation strain, were constructed using P. gingivalis ATCC 33277. The gingipain and hemagglutination activities, as well as biofilm formation, were examined in all three strains. The effect of P. gingivalis strains on epithelial cell detachment was investigated using the HO-1-N-1 and Ca9-22 epithelial cell lines. The inhibition of wound closure by heat-killed P. gingivalis cells and culture supernatant was analyzed using an in vitro wound closure assay. Compared to the wild-type strain, the PgΔ0361 strain did not exhibit gingipain or hemagglutination activity but exhibited enhanced biofilm formation. Additionally, the PgΔ0361 strain exhibited attenuated ability to detach the epithelial cells and to inhibit wound closure in vitro. Contrastingly, the culture supernatant of PgΔ0361 exhibited high gingipain activity and strong inhibition of wound closure. The characteristics of PgΔ0361C and wild-type strains were comparable. In conclusion, the pgn_0361 gene is involved in regulating gingipains. The PGN_0361defective strain exhibited reduced virulence in terms of epithelial cell detachment and inhibition of wound closure. The culture supernatant of the mutant strain highly inhibited wound closure, which may be due to high gingipain activity.K E Y W O R D Sbiofilm, epithelial cell detachment, gingipain, pgn_0361, Porphyromonas gingivalis, wound closure
## | introduc ti on
The colonies of several oral bacterial species, including Porphyromonas gingivalis and Prevotella spp., exhibit black pigmentation on the blood agar medium. The black pigmentation phenotype is associated with the initiation and development of various infections, including periodontitis and endodontic infections. In a mouse infection model, the administration of black-pigmented P. gingivalis was lethal, whereas the administration of a spontaneously pigment-less P. gingivalis variant did not cause lethality. This indicated that the pigment-less phenotype of P. gingivalis is associated with reduced virulence.
Various studies on P. gingivalis have revealed that the black colony pigmentation is caused due to heme accumulation on the cell surface. P. gingivalis acquires iron through gingipain-mediated hemoglobin digestion in the microenvironment. The hemagglutinins (such as HagB and HagC) facilitate the acquisition of heme through erythrocyte binding, allow the gingipains to release the heme moiety from the hemoglobin molecule, and uptake by the bacterial. Meanwhile, the heme-binding domains (HA2) of gingipain and the adhesion domains of hemagglutinins are both able to bind heme and hemoglobin, facilitating the heme deposition. Heme and peptides that are released after digestion are used by the cells as nutrients or stored on the cell surface, which leads to a black pigmentation phenotype. Based on their specificity, gingipains can be classified as arginine-specific proteases (RgpA and RgpB) and lysine-specific proteinases (Kgp). Interestingly, all known pigmentation-related genes are functionally linked to gingipains, especially kgp, or genes involved in the regulation, transportation, and attachment of gingipains.
Porphyromonas gingivalis is reported to have a complex gingipain regulation system, which involves several genes or gene clusters. Several studies have focused on the genes and gene clusters related to gingipain function. However, there are limited studies that have evaluated the involvement of gingipain in the interaction between P. gingivalis and host cells. Gingipains are reported to be key mediators in P. gingivalis-induced inflammation and tissue damage. Previous studies have demonstrated that gingipains can degrade various human cellular proteins, including cytokines, integrins, and collagen, or can alter cellular signal transduction. Gingipains are also involved in detaching the epithelial cells from the connective tissues of the gingivaand inhibiting the wound healing process. However, it is not clear whether the gingipain-regulating genes affect the interaction between P. gingivalis and host solely through regulation of gingipain function or in combination with other functions of the individual gene.
The screening of the non-pigmented P. gingivalis strains in a transposon mutagenesis library revealed several pigmentation-related genes, such as kgp and gtfB. However, the two transposon mutagenesis systems that were screened (Tn4351 and Tn4400) have several disadvantages. In both systems, the elements are inserted preferentially into the "hotspots" in the genome, which limits the distribution of interrupted genes. Additionally, introducing Tn4351 into P. gingivalis may result in duplication and transposition of the endogenous insertion element. Therefore, the complexity of P. gingivalis genomic rearrangement after Tn4351 transposition restricts its use in transposition mutagenesis. The Himar 1 Mariner mini-transposon system was developed in 2012, which addressed the limitations of previous systems. This system allows a single, stable transposition event. This system has been successfully applied to identify the genes that are essential for the growth of P. gingivalisand to identify pigmentation-related genes.
In this study, we generated a transposon library of the P. gingivalis ATCC 33277 strain using the Mariner system independently to isolate several non-pigmented mutants. Interestingly, we also identified an insertion within the pgn_0361 gene, which was previously characterized as the pigmentation-related gene byusing the same Mariner system. This study aimed to investigate the involvement of pgn_0361 in the interaction between P. gingivalis and host cells. Specifically, the role of pgn_0361 in regulating the P. gingivalis-mediated epithelial cell detachment and inhibition of wound healing was examined.
## | material s and me thods
## | bacterial strains and growth conditions
Porphyromonas gingivalis ATCC 33277 was used as the parental strain and all mutants were constructed based on this genetic background. The P. gingivalis strains were routinely cultured on 5% sheep blood tryptic soy agar plates (Becton Dickinson B.V., Breda, the Netherlands) supplemented with hemin (5 μg/mL) and menadione (1 μg/mL) (BA plates) under anaerobic conditions (80% N 2 , 10% H 2 , and 10% CO 2 ) at 37°C. The mutant strains (transposon mutants) were selected in the presence of 50 μg/mL gentamicin and 5 µg/mL erythromycin. The complement mutants were selected in the presence of 50 μg/mL gentamicin and 1 µg/mL tetracycline.
Escherichia coli S17-1 was used for maintenance, construction of plasmids, and conjugation. The pSAM_Bt plasmid containing the bla, ermG, and himar1c9a genes was obtained from Dr. Brian Klein. S17-1 was cultured in Luria-Bertani (LB) broth or LB agar. Ampicillin (100 µg/ml) was added to the medium when necessary.
## | transposon mutagenesis
The Mariner-based transposon mutants were constructed by conjugation, following the protocol described by. Briefly, the P. gingivalis cells (OD 600 = 0.5-1) were mixed with the culture of E. coli S17-1 (OD 600 =0.5) harboring the pSAM_Bt plasmid. The mixture was plated on the BA plates containing 50 μg/ml gentamicin and 5 µg/ml erythromycin. The erythromycin-resistant transposon mutants were selected.
## | construction of the p. gingivalis mutant strain
To generate the mutant strain, the pgn_0361 coding region was replaced with the ermF gene using the PCR ligation mutagenesis method as described previously. The genomic DNA (gDNA) of the P. gingivalis ATCC 33277 strain was isolated using the GeneJet genomic DNA purification kit (ThermoFisher, Waltham, MA, USA) and was used as a template for PCR. The flanking regions of the PGN_0361 gene were amplified from gDNA using the following two PCR primer pairs: 361uf/361ur (361uf 5′-TCTGCAAGGCTACGACTATCT-3′; 361ur 5′-ACTAGTGAGAAGGCACTGTTCCAAGA-3′) and 361df/361dr (361df 5′-GGCGCGCCTGATACATCTTCCCTTAGAAAG-3′; 361dr 5′-GCAGC CATACACCTGCTCATT-3′). The erythromycin resistance gene (Em gene) was amplified from the pEP4351 plasmid using the following primer pair: Emf/Emr (Emf 5′-ACTAGTGTTTCCGCTCCATCGCCAATTTGC-3′;
Emr 5′-GGCGCGCCCGATAGCTTCCGCTATTGC-3′). These three PCR fragments were digested with restriction enzymes SpeI and AscI (indicated by the underlined sequences) and ligated for 16 h at 4 °C. The ligated fragment was used as the template for the amplification of the mutant construct using the 361uf/361dr primer pair. The resulting PCR product was purified and transformed directly into the P. gingivalis cells by electroporation. The mutant strain was obtained via a double crossover. The insertion of the Em gene was verified by PCR and Sanger sequencing. The resulting strain was designated as PgΔ0361.
## | complementation of the pgδ0361 mutant strain
To complement the function of pgn_0361, a region containing the complete PGN_0361 sequence (including a 105 bp upstream region) was amplified from the gDNA of the ATCC 33277 strain by PCR. The complete fragment was cloned into the shuttle vector, pT-COW. The primer sequences used were as follows: 361F_com, 5′-GCATGCCGGTAGGAGAGGACGACCTCTTGT-3′; 361R_com, 5′-GT CGACCATGGCGATCCGAATACGATCA-3′. The resulting plasmid was transformed into E. coli S17-1 and the recombinant colonies were selected based on ampicillin resistance. The plasmid was introduced into the corresponding PgΔ0361 strain by conjugation. The presence of the newly constructed plasmid was confirmed by PCR. The PgΔ0361 strain containing the pT-COW-derived plasmid was designated as PgΔ0361C.
## | characterization of the mutant pgδ0361 strain
The mutant PgΔ0361 strain was characterized along with the corresponding wild-type ATCC 33277 strain and PgΔ0361C strain. The hemagglutination and gingipain activities, as well as the ability to form biofilms, were investigated in all three strains.
## | gingipain activity
The late exponential culture (2 ml; OD 600 = 1) of P. gingivalis strains was subjected to centrifugation at 21,200 g for 2 min. The supernatant was collected, and the pellet was resuspended in Tris-buffered saline (TBS, 50 mM Tris, 150 mM NaCl, pH 7.6). The supernatant was filter-sterilized using a 0.2-μm filter before further analysis. The
## | hemagglutination activity
The hemagglutination activities of all three strains were examined following previously described protocols. Briefly, the OD 600 of the late exponential P. gingivalis cells was adjusted to 1 in phosphate-buffered saline (PBS) and serially diluted in a round-bottom 96-well plate (100 μl/well). The P. gingivalis cells were incubated with an equal volume of 1% PBS-washed sheep erythrocytes (Biotrading, Mijdrecht, the Netherlands) at 22°C for 3 h.
Hemagglutination was visually assessed. The last dilution that exhibited complete hemagglutination was considered as the titer.
## | biofilm formation
Biofilms were grown in a 96-well active attachment model. The 24-h pre-culture of each P. gingivalis strain was diluted
(1:20) in fresh brain-heart infusion (BHI) broth supplemented with hemin (5 μg/ml) and menadione (1 μg/ml) (BHIHM) and pipetted into a 96-well plate (200 μl/well). The lid with 96 pegs (NuncTM, Roskilde, Denmark) was placed into the microtiter plate to allow biofilm formation. The pegs were transferred to fresh BHIHM after 24 h. The biomass of 48-h P. gingivalis biofilms was quantified by crystal violet assay. Three experiments were performed with quadruplicate samples from each group per experiment.
## | eukaryotic cell culture
The human oral buccal epithelial cell line (HO-1-N-1)and gingival epithelial cell line (Ca 9-22) (Japanese Collection of Research Bioresources) were cultured in tissue culture flasks containing Dulbecco's modified Eagle's medium/Nutrient Mixture F-12 (DMEM/F-12; Life Technologies, Waltham, MA, USA) supplemented with 10% fetal calf serum (Hyclone, Logan, UT, USA) and 1% antibiotics (penicillin, streptomycin, and fungizone, Sigma, St. Louis, MO, USA) in a humidified atmosphere at 37°C and 5% CO 2 .
## | detachment assay
The ability of the three P. gingivalis strains to detach the HO-1-N-1 and Ca9-22 cells was evaluated using a detachment assay. The confluent (80%) cell cultures were trypsinized and resuspended in the DMEM/F-12 medium supplemented with 10% fetal calf serum. The resuspended cells were seeded in a 96-well plate at a cell density of 3.5 × 10 4 cells/well for 24 h. The cells were washed with PBS twice, followed by the addition of 100 μl P. gingivalis cultures at a multiplicity of infection (MOI, the ratio of bacteria: cells) of 50,000 or only 100 μl medium (control). The P. gingivalis cultures were prepared by centrifuging the late exponential cultures at 5,000 g for 15 min. The cell pellet was resuspended with the culture medium. The co-culture was incubated at
## | in vitro wound closure assay
The effects of the three P. gingivalis strains on the wound healing process were evaluated using an in vitro wound closure assay. Both heatinactivated P. gingivalis cells and supernatants collected from the late exponential P. gingivalis cultures (OD 600 = 1) were used to challenge the host cells. The supernatants were filter-sterilized before use.
The heat-inactivated P. gingivalis cells were prepared according to the protocol described by. The late exponential P. gingivalis cells were washed twice in PBS (pH 7.4) and resuspended in keratinocyte serum-free medium (SFM; ThermoFisher, Waltham, MA, USA). The resuspended cells were incubated at 60°C in a water bath for 60 min. The killing of bacterial cells was confirmed by the absence of P. gingivalis growth on the BA plates.
The wound closure assays were performed as described previously. The HO-1-N-1 cells were seeded in a 24well plate at a concentration of 3 × 10 5 cells/well for 6 h. An artificial wound was created by scraping the confluent cell monolayer using a sterile 1-ml micropipette tip. The epithelial cells were treated with P. gingivalis culture supernatant or heat-inactivated P. gingivalis cells.
The supernatant was diluted three times with SFM before addition to the wells. The P. gingivalis growth medium (BHIHM) was diluted similarly to serve as a negative control for the supernatant groups.
The heat-inactivated P. gingivalis cells at MOIs of 10, 100, and 1,000 were tested. The SFM was used as a negative control for the heat-inactivated groups. The co-cultures were incubated in a humidified atmosphere at 37°C and 5% CO 2 in a Bioflux system (Fluxion).
The cells in the middle of the well were simultaneously imaged at a magnification of 50X after the addition of P. gingivalis challenges.
The images were captured at 2-h intervals for 20 h. The area of the scratch in each image was calculated using the ImageJ software (ver-
# | statistical analysis
The data were analyzed using the Statistical Package for Social Science (SPSS, version 23.0). One-way analysis of variance (ANOVA), followed by Bonferroni post hoc test was used to compare the gingipain activity and host cell detachment data of the wild-type, PgΔ0361, and PgΔ0361C P. gingivalis strains. The Kruskal-Wallis test, followed by Mann-Whitney U post hoc test, was used to compare the biomass of three P. gingivalis strains as not all data met the prerequisites of normal distribution and equality of variance. The α value was set at 0.0167 (0.05/3). To evaluate the influence of the PGN_0361 deletion on wound healing ability, the relative cell coverage area was analyzed using a twoway ANOVA, followed by the Bonferroni test. The difference was considered statistically significant when the p-value was less than 0.05.
## | re sults
## | identification of genes associated with nonpigmentation
In total, 400 transposon mutants were screened on the BA agar plates for colony pigmentation. Among these mutants, 11 colonies displayed altered pigmentation. To locate the transposon insertion in the genome, the gDNA of these 11 mutants with altered pigmentation was extracted and used as a template in a nested semi-random PCR. The PCR products were subjected to sequencing analysis, which revealed one insertion site in the pgn_0361 gene that encodes a putative glycosyltransferase (family 2). Another mutant contained an insertion site in the pgn_1240 gene which encodes a putative glycosyltransferase (group 1). The rest mutants contained insertions sites in either the rgpA/B gene or kgp gene.
## | characterization of the mutant pgδ0361 strain
We constructed a PgΔ0361 strain by gene-directed mutagenesis and a PgΔ0361C strain by introducing the wild-type pgn_0361 gene into the PgΔ0361 strain. The PgΔ0361, PgΔ0361C, and wild-type strains were characterized . The colonies of PgΔ0361 did not exhibit black pigmentation on the BA plate, whereas those of PgΔ0361C exhibited the same black pigmentation as the wild-type strain . The Rgp and Kgp activities of the cell pellets and culture supernatants of the three P. gingivalis strains are presented in . The Rgp and Kgp activities of the PgΔ0361 cell pellets were 17 ± 15% and 5 ± 3%, respectively, which were significantly lower than those of the wild-type and PgΔ0361C strains (p < 0.05).
The Rgp and Kgp activities of the PgΔ0361 cell culture supernatant were 155 ± 7% and 158 ± 26%, respectively, which were significantly higher than those of the wild-type and PgΔ0361C strains (p < 0.05).
The hemagglutination activity and biofilm formation of the strains are shown in and 1D, respectively. The PgΔ0361 strain did not exhibit hemagglutination activity. The wild-type and PgΔ0361C strains exhibited similar hemagglutination activities. The biomasses of the PgΔ0361 biofilms were 2-fold higher than those of the wildtype biofilms. The biofilm biomass of the PgΔ0361C strain was lower than that of the PgΔ0361 strain and was significantly higher than that of the wild-type strain (p < 0.0167).presents the cell detachment results obtained from all MOIs. Generally, increasing cell detachment could be observed as the MOIs of P. gingivalis wild-type and complement strains increased. However, the mutant strain was unable to detach any tested epithelial cells irrespective of MOI values.
## | role of pgn_0361 in epithelial cell detachment and wound closure
In the wound closure assay, only the HO-1-N-1 cells were tested as they exhibit strong adhesion to the surfaces of the polypropylene 24-well plate. As shown in , the heat-killed P. gingivalis strains dose-dependently inhibited the epithelial cell migration after co-incubation for 20 h. However, the levels of epithelial cell migration inhibition varied among the tested strains. For example, the PgΔ0361 strain exhibited significant inhibition of wound closure at the highest MOI tested (MOI 1,000), whereas the other two strains exhibited strong inhibition at MOI of 100. Additionally, the PgΔ0361 cells at the highest MOI exhibited non-significantly lesser inhibition of wound closure than the wild-type and PgΔ0361 strains (p > 0.05).
The culture supernatant of PgΔ0361 exhibited almost complete inhibition of wound closure after 20 h . The supernatants of the other two strains exhibited limited inhibition of wound closure. Additionally, treatment with the supernatant negative control, which was prepared by diluting the BHIHM 3 times with the SFM medium, delayed the wound closure. After 20 h, the wound closure of this group was only 50% of that in the SFM control group (data not shown).
## The dynamic effects of p. gingivalis heat-killed cells (moi 100)
and supernatants on wound closure are shown in , respectively. The percentage of wound closure was plotted against time at 2-h intervals. In the medium control group, the scratch closure rate increased with time following a linear pattern.
The heat-killed wild-type and PgΔ0361C strains
## | discuss ion
In this study, we identified the black pigmentation-related gene (pgn_0361) in a transposon mutant library of P. gingivalis ATCC 33277, which was generated using the Mariner transposon system.
The pgn_0361 insertion mutant was isolated by other researchers using the same method. Klein et al. reported that this gene is involved in pigmentation. The pgn_0361 gene deletion strain generated in this study exhibited altered phenotypes, such as non-pigmented colonies on the BA plate and impaired gingipain activities. Additionally, the mutant strain exhibited decreased hemagglutinin activity, which was consistent with the results of a previous study.
So far, fifteen genes have been reported to be associated with black pigmentation in P. gingivalis. The deletion of most of these genes, except ugdA, results in impaired hemolytic and hemagglutinin activities.
The reduction in hemagglutinin activity of the pgn_0361 mutant strain indicated that the pgn_0361 gene shared general functions of the pigmentation-related genes. Of the 15 genes, two are putative glycosyltransferase genes (gtfB and vimF). Interestingly, the pgn_0361 gene identified in this study was also annotated as a putative glycosyltransferase 2-3 domain in a previous study. Similar to the gtfB and vimF mutant strains, the pgn_0361 mutant strain also exhibited In our study, the activities of Rgp and Kgp in the culture supernatant of the pgn_0361 mutant strain were significantly higher than those in the culture supernatant of the wild-type strain, whereas these activities in the cell pellets were significantly less than those of the wild-type.have also shown that the cell-associated gingipain activities of gtfC (PGN_0361) mutant strain were reduced whereas most of its gingipain activities were observed in the supernatant. Similar observations have been reported for other gingipain regulatory genes, such as porR and gppX. These findings indicate that these mutant strains were able to produce Rgp and Kgp but could not retain them on the cell surfaces. It is known that both Rgp and Kgp belong to a group of C-terminal domain (CTD) proteins, which are attached to the cell surfaces through the modification by anionic lipopolysaccharide (A-LPS). Since both porR and gtfC mutant strains were shown to be deficient in A-LPS, it is likely that these strains, including the pgn_0361 mutant strain, failed to anchor the produced gingipains to the outer membrane of the cells due to the lack of A-LPS; thus, the majority of gingipains was detected in the culture supernatant instead of the cellular fraction. However, our results did not concur with the report of, where the culture supernatant of the pg_0264 (a pgn_0361 homolog) mutant strain, constructed in strain W83, did not exhibit any Rgp and Kgp activities. In ours and the study of, the mutant strains were constructed in strain ATCC33277. Therefore, the discrepancy might have resulted from strain variation. Previously, it has been shown that the distribution of Rgp and Kgp in various cell fractions of P. gingivalis was different in different strains.
In addition to characterizing the pgn_0361 deletion strain, we also investigated the role of pgn_0361 in the interaction between P. gingivalis and epithelial cells in vitro. The pgn_0361 deletion mutant could not detach the epithelial cells and exhibited attenuated ability to inhibit wound closure. However, the culture medium of the mutant strain exhibited the strongest inhibition of cell migration. Arg-specific and Lys-specific gingipains are reported to mediate the inhibition of wound healing caused by P. gingivalis. This may be due to the high concentration of gingipains detected in the medium of the pgn_0361 deletion strain. Our results further demonstrated that the released gingipains in the culture medium of strain PgΔ0361 could actively inhibit the wound healing process during the entire test period (20 h).
The PgΔ0361 and wild-type strains exhibited a significant difference in the inhibition of wound closure only at MOI of 100. At higher MOI, the levels of wound closure inhibition were similar in the PgΔ0361 and wild-type strains.reported that the kgp and rgpArgpB mutant strains at MOI of 1,000 exhibited a significantly lower inhibition of wound closure than the wild-type strain.
The PgΔ0361 cells may have residual gingipains, whose activity was too low to be detected using the fluorogenic probes. When the bacterial counts were high enough, these residual gingipains could be sufficient to inhibit the wound closure. In addition to gingipains, other factors, such as capsular polysaccharidein the PgΔ0361 cells, may also be involved in the inhibition of wound closure. Previously, the non-encapsulated P. gingivalis strain was found to inhibit the wound closure more strongly than the encapsulated strain. The deletion of a glycosyltransferase may result in a loss of capsule, thereby inhibit wound closure even when the gingipains were absent.
Oral diseases, such as periodontitis or mucositis, are infectious diseases involved host tissue destruction and regeneration, where periodontal pathogens played an important role in modulating tissue homeostasis. Cell detachment assay and wound healing assay have been used in vitro to investigate the influence of risk factors on tissue homeostasis and the intervention strategies. Hence, this study chose these two assays to investigate the role of pgn_0361 in the interaction between P. gingivalis and the host. The host inflammatory response to microbes is also a critical host factor for the development of the infectious disease. The role of pigmentation-related genes in this inflammatory process is certainly worth further investigation.
## | con clus ions
This study confirmed that the pgn_0361 gene of P. gingivalis is involved in pigmentation and gingipain regulation. Additionally, this study demonstrated that the deletion strain exhibited reduced hemagglutinin activity and enhanced biofilm formation capacity.
## F i g u r e 3
Wound closure in the oral buccal epithelial cells challenged with heat-inactivated Porphyromonas gingivalis strains. (A) Mean relative closure (±SD) from all biological replicates of scratches in the oral buccal epithelial cells challenged with heat-inactivated P. gingivalis strains after 20 h. (B) Time curve of percentage closure of a scratch in the oral buccal epithelial cells challenged with heat-inactivated P. gingivalis strains at MOI 100. The cells treated with the medium were used as a control. *p < 0.05. Moreover, the pgn_0361 deletion strain also exhibited reduced virulence via its reduced ability to detach the adherent epithelial cells and inhibit wound closure. However, the culture medium of the mutant strain exhibited enhanced inhibition of wound closure, which may be due to high gingipain activity in the medium.
## Ack n owled g em ents
This work was supported by the National Natural Science Foundation of China (81800960) and Guangdong Basic and Applied Basic Research Foundation (2020A1515010315).
## Co n fli c t o f i nte r e s t
None declared.
## Auth o r co ntr i b uti o n s
Hongyan Liu: Data curation (equal); formal analysis (equal); funding acquisition (equal); investigation (equal); methodology (equal); software (equal); visualization (equal); writing -original draft (equal). Lijia Huang:
Data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); software (equal); validation (equal); visualization (equal); writing -original draft (equal); writing -review and editing (equal). Yanling Cai: Data curation (supporting); formal analysis (supporting); funding acquisition (equal); investigation (supporting); software (supporting); writing -review and editing (supporting). Floris J. Bikker:
Formal analysis (supporting); methodology (equal); validation (equal); writing -review and editing (supporting). Xi Wei: Conceptualization (lead); funding acquisition (lead); project administration (lead); resources (lead); supervision (lead); writing -review and editing (equal). Dong Mei Deng: Conceptualization (lead); formal analysis (lead); funding acquisition (equal); investigation (lead); resources (lead); software (equal); visualization (equal); writing -review and editing (equal).
## E th i c s s tatem ent
None required.
## Data ava i l a b i l i t y s tat e m e n t
All data are provided in full in the results section of this paper.
## O rci d
## Hongyan
|
Clinical and Disease Characteristics From REVEAL at Time of Enrollment (Baseline): Prospective Observational Study of Patients With Polycythemia Vera in the United States
REVEAL is the first large, prospective observational study aimed at examining the contemporary demographics, burden of disease, clinical management, patient-reported outcomes, and health care resource utilization in patients with polycythemia vera in the United States.Background: Polycythemia vera (PV) has a prevalence of 44 to 57 per 100,000 people in the United States. Prospective data concerning the demographics, clinical characteristics, and treatment patterns of patients with PV in the United States are lacking.Patients and Methods:The ongoing, prospective, observational REVEAL study evaluates demographics, disease burden, clinical management, patient-reported outcomes, and health care resource utilization of adult patients with PV in the United States. This report summarizes the demographics and clinical characteristics of patients at enrollment (baseline).Results: Patients (n = 2510) were a median age of 67.0 years, 54.2% were male, and 89.1% were white. The median time from PV diagnosis to study enrollment was 4.0 (range, 0-56.3) years. Most patients (89.7%) were diagnosed after an abnormal blood test. Less than half (49.2%) underwent JAK2 mutation analysis, of whom 95.8% were JAK2 V617F mutation positive; < 1% were positive for JAK2 exon 12 mutations. At enrollment, 47.7% of patients had elevated hematocrit (> 45%), 35.8% had elevated platelets (> 400 × predominantly therapeutic phlebotomy alone (33.6%), hydroxyurea monotherapy (29.0%), or hydroxyurea plus phlebotomy (23.7%). Thrombotic events occurred in 11.9% of patients before PV diagnosis (venous, 6.7%; arterial, 5.7%), and 8.3% between diagnosis and enrollment. Hypertension (70.6%) was the most common previous medical condition.Conclusion: REVEAL enrollment data inform our understanding of the baseline demographics, diagnostic approach, disease characteristics, and treatment patterns of patients with PV in the United States. Longitudinal real-world data collected in this study will complement information collected during randomized controlled clinical trials.
# Introduction
Polycythemia vera (PV) is a myeloproliferative neoplasm in which clonal proliferation of a hematopoietic progenitor cell results in erythrocytosis, leukocytosis, and thrombocytosis. [bib_ref] Myeloproliferative neoplasms and thrombosis, Barbui [/bib_ref] With respect to the epidemiology of PV in the United States, analyses of administrative claims from 2008 to 2010 reported a prevalence of approximately 44 to 57 per people. [bib_ref] Epidemiology of myeloproliferative neoplasms in the United States, Mehta [/bib_ref] Most patients with PV have a Janus kinase 2 (JAK2) mutation that results in constitutive activation of hematopoietic signal transduction, thus explaining the exaggerated hematopoiesis. More than 95% of patients with PV carry the JAK2 V617F mutation; ~4% have mutations in JAK2 exon 12. [bib_ref] Response criteria for essential thrombocythemia and polycythemia vera: result of a European..., Barosi [/bib_ref] [bib_ref] Acquired mutation of the tyrosine kinase JAK2 in human myeloproliferative disorders, Baxter [/bib_ref] [bib_ref] A prospective study of 338 patients with polycythemia vera: the impact of..., Pmssamonti [/bib_ref] [bib_ref] The JAK2 exon 12 mutations: a comprehensive review, Scott [/bib_ref] [bib_ref] Molecular and clinical features of the myeloproliferative neoplasm associated with JAK2 exon..., Pmssamonti [/bib_ref] Patients with PV have an increased mortality risk relative to the general population, primarily due to thrombotic events, transformation to leukemia or myelofibrosis, and solid tumors. [bib_ref] Treatment related changes in anti-fibrinolytic activity in patients with polycythemia vera, Sonmez [/bib_ref] [bib_ref] Patterns of survival among patients with myeloproliferative neoplasms diagnosed in Sweden from..., Hultcrantz [/bib_ref] [bib_ref] Life expectancy and prognostic factors for survival in patients with polycythemia vera..., Passamonti [/bib_ref] [bib_ref] Vascular and neoplastic risk in a large cohort of patients with polycythemia..., Marchioli [/bib_ref] [bib_ref] Acute leukemia in polycythemia vera: an analysis of 1638 patients enrolled in..., Finazzi [/bib_ref] Patients also experience a broad range of symptoms that may lead to decreased quality of life and productivity. [bib_ref] Myeloproliferative Neoplasm (MPN) Symptom Assessment Form Total Symptom Score: prospective international assessment..., Emanuel [/bib_ref] [bib_ref] Association between quality of life and clinical parameters in patients with myeloproliferative..., Johansson [/bib_ref] [bib_ref] Myeloproliferative neoplmsm.s (MPNs) have a significant impact on patients' overall health and..., Mesa [/bib_ref] [bib_ref] The Myeloproliferative Neoplasm Symptom Assessment Form (MPN-SAF): international prospective validation and reliability..., Scherber [/bib_ref] Large multicenter observational studies of PV have focused primarily on European patients, [bib_ref] Myeloproliferative Neoplasm (MPN) Symptom Assessment Form Total Symptom Score: prospective international assessment..., Emanuel [/bib_ref] [bib_ref] Association between quality of life and clinical parameters in patients with myeloproliferative..., Johansson [/bib_ref] [bib_ref] The Myeloproliferative Neoplasm Symptom Assessment Form (MPN-SAF): international prospective validation and reliability..., Scherber [/bib_ref] [bib_ref] Patients with polycythemia vera have the worst impairment of quality of life..., Abelsson [/bib_ref] [bib_ref] Polycythemia versa: the natural history Of 1213 patients followed for 20 years, Policitemia [/bib_ref] [bib_ref] Risk and cause Of death in patients diagnosed with myeloproliferative neoplasms in..., Hultcrantz [/bib_ref] [bib_ref] Bleeding, thrombosis, and anticoagulauon In myeloproliferative neoplasms (MPN): analysis from the German..., Kaifie [/bib_ref] [bib_ref] Survival and prognosis among 1545 patients with contemporary polycythemia vera: an international..., Tefferi [/bib_ref] whereas US studies have been limited to single-center settings. [bib_ref] Sex differences in the JAK2 V617F allele burden in chronic myeloproliferative disorders, Stein [/bib_ref] [bib_ref] Age-related differences in disease characteristics and clinical outcomes in polycythemia vera, Stein [/bib_ref] Patterns of care can be characterized with administrative claims databases and electronic medical record audits. However, these data sources frequently lack important clinical information and do not capture patient-reported outcomes (PROs).
## The prospective observational study of patients with polycythemia vera in us clinical
Practices (REVEAL; ClinicalTrials.gov, NCT02252159) was designed to collect contemporary demographics and data on the burden of disease, clinical management, PROs, and health care resource utilization of US patients with PV being seen at community and academic medical centers. This report describes the demographics as well as the clinical and disease characteristics of patients in the REVEAL study at the time of study enrollment.
# Patients and methods
## Study design
REVEAL is a multicenter, noninterventional, prospective observational study of patients with PV conducted in accordance with the Declaration of Helsinki. Approval of all study materials by central (Sterling IRB; Atlanta, GA) and local institutional review boards and provision of informed consent by all patients was required, including a separate informed consent form documenting willingness to participate in an optional biological substudy to examine the molecular features of PV. Patients with a clinical diagnosis of PV were enrolled from community practices and academic centers in the United States over a 24-month enrollment period. Physician assessments and PRO outcomes (previously published 24 ) were recorded during routine care visits (approximately every 6 months). Blood samples were collected at enrollment and annually for patients agreeing to participate in an optional molecular substudy. All treatment decisions during the study were made by patients' treating physicians.
## Eligibility requirements
At enrollment, patients were ≥ 18 years old; had a clinical diagnosis of PV per physician judgment; were willing and able to provide informed consent and complete patient assessments and questionnaires, either alone or with minimal assistance from a caregiver or trained site personnel; and were currently under physician supervision for management of PV. Patients were excluded if they were participating in an active, blinded clinical trial; had a life expectancy < 6 months; had a diagnosis of myelofibrosis, acute myeloid leukemia, or myelodysplastic syndrome; had a history of or active plan (within 3 months of enrollment) to proceed to allogeneic hematopoietic stem-cell transplantation; or had undergone splenectomy.
## Assessments
This analysis describes patient demographics as well as clinical and disease characteristics at time of diagnosis and around the time of enrollment. Historic data pertaining to the evaluations and disease characteristics at time of diagnosis and physician-reported data from usual care visits, including information from the 6 months before enrollment, were collected and abstracted from electronic medical records into an electronic data capture system. For the purpose of this analysis, physician-reported data in the following categories were examined: demographics and PV disease characteristics; education, employment, insurance, and residency characteristics; laboratory tests and procedures used to determine a diagnosis of PV; signs and symptoms at time of PV diagnosis; laboratory values at time of enrollment; PV-directed therapy; comorbid conditions; thrombotic event history; concomitant medications; and health care resource utilization.
The Charlson Comorbidity Index (CCI) was used to assess morbidity. [bib_ref] A new method of classifying prognostic comorbidity in longitudinal studies: development and..., Charlson [/bib_ref] The CCI uses a 6point scale to rate diagnosis codes from the International Classification of Diseases for each comorbid condition, which are summed to generate a total index score for each patient; higher scores correspond with more severe morbidity.
# Statistical analysis
Descriptive statistics were used. Missing data were not imputed; however, missing dates for temporal end points were imputed when applicable. If the entire date (ie, day, month, and year) was missing, all corresponding data were excluded from analyses involving dates. If the day was missing but the month and year were available, the date was imputed as the first day of the month. If the day and month were missing but the year was available, the date was imputed as January 1.
# Results
## Patient demographics and pv disease characteristics
In this ongoing study, 2510 patients were enrolled from July 22, 2014, to August 3, 2016, from 227 US sites (39 academic and 188 community [self-identified]; [fig_ref] Figure 1: IRB-Approved Study SitesParticipating in REVEAL [/fig_ref]. Patients had a mean (SD) age of 66.3 (12.3) years , and 2.9% were young (aged < 40 years). Most patients were male (54.2%) and white (89.1 and 48.1% had a history of smoking; 11.1% were active smokers at time of enrollment. Family history of PV was documented in 132 patients (5.3%).
Median (range) disease duration was 4.0 (0-56.3) years [fig_ref] Figure 2: Duration Myeloma Leuk [/fig_ref] , 77.3% of patients were classified as having high-risk disease (aged ≥ 60 years and/or history of thrombotic events), and 72.6% were obese or overweight. Spleen assessment by physical examination was performed in 1601 patients (63.8%) at time of enrollment. Of these patients, 282 (17.6%) had palpable splenomegaly. Spleen length measurements were available for 122 patients, 51.6% of whom had spleen length ≥ 5 cm from the left costal margin to the point of greatest splenic extension.
## Patient education, employment, insurance, and residency characteristics
Most patients (62.7%) had at least some college education, 33.5% were employed full time or part time, 51.1% were retired, and 4.4% were disabled or unable to work (Supplemental . The vast majority of patients (99.1%) had health insurance (public only, 47.8%; private only, 40.4%; both public and private, 10.9%). Most patients (94.9%) were living independently at home, 3.7% were living as a dependent at home, and 0.8% were living in assisted housing; 0.1% and 0.4% were living in long-term care facilities or other residency settings, respectively.
## Pv diagnosis
The majority of patients were reported to have abnormal hemoglobin (59.0%) or hematocrit (57.5%) results at time of diagnosis [fig_ref] Table 2: Abnormal Laboratory Tests and Polycythemia Vera-Related Signs and Symptoms at Time of... [/fig_ref]. An abnormal platelet count or erythropoietin level was reported in only 28.7% and 20.1% of patients at time of diagnosis, respectively. In almost half (49.2%) of patients, JAK2 mutational testing results were reported; most of these patients were positive for JAK2 V617F (95.8%), with only 0.8% positive for JAK2 exon 12 mutations. A JAK2 test was reported for 19.9% of patients diagnosed before January 1, 2006, compared to 54.3% of patients diagnosed after January 1, 2006. Bone marrow biopsy/ aspiration results were reported in 24.4% of patients. Palpable splenomegaly was reported as a factor in the diagnosis of 7.0% of patients. With respect to PV-specific symptoms, the most commonly reported symptoms noted at diagnosis were tiredness (22.0%), itching (13.9%), muscle aches and/or bone pain (8.8%), and difficulty sleeping (7.8%; [fig_ref] Table 2: Abnormal Laboratory Tests and Polycythemia Vera-Related Signs and Symptoms at Time of... [/fig_ref].
## Laboratory values
At time of enrollment, 5.3% of patients had an elevated hemoglobin level (> 18.5 g/dL in men or > 16.5 g/dL in women), 47.7% had elevated hematocrit (> 45%), 35.8% had elevated platelets (> 400 × 10 9 /L), and 37.0% had elevated leukocytes (> 10 × 10 9 /L). Mean hemoglobin, hematocrit, platelet, and leukocyte values at enrollment were 14.5 g/dL, 45%, 367.5 × 10 9 /L, and 10.4 × 10 9 /L, respectively .
## Pv-directed therapy at time of enrollment
At time of enrollment, the majority (54.6%) of patients with low-risk disease were being managed with therapeutic phlebotomy. However, 18.1% and 15.8% of low-risk patients were being managed with hydroxyurea alone and phlebotomy plus hydroxyurea, respectively. Only 4.9% of patients with low-risk PV were managed with watchful waiting. In contrast, hydroxyurea alone (32.2%) and phlebotomy alone (27.4%) were the most commonly reported PV-directed therapies in patients with high-risk PV. Management with watchful waiting was reported for 5.7% of patients with high-risk disease [fig_ref] Figure 3: Figure 3. [/fig_ref]. The most common hydroxyurea daily doses were 500 mg (35.4%) and 1000 mg (30.2%). The mean hydroxyurea daily doses were 769.5 and 834.9 mg for patients undergoing and not undergoing phlebotomy, respectively. Among patients who were treated with therapeutic phlebotomy, mean number of phlebotomies was 2.5 (median, 2) during the 6 months before enrollment. A small proportion of patients were treated with other agents (8.2%), including ruxolitinib (3.8%), anagrelide (2.0%), and interferon (1.4%).
## Comorbid conditions
The most common comorbid conditions were hypertension (70.6%), hyperlipidemia (30.4%), obesity (17.5%), and diabetes mellitus (15.7%). Mean (SD) CCI was 3.4 (0.83). A history of skin cancer (not otherwise specified) and nonmelanoma skin cancer was documented in 3.5% and 3.3% of patients, respectively. Malignant melanoma was noted in 1.1% of patients. A history of other malignancies included prostate (2.9%), breast (1.9%), and colon (1.0%) cancers (Supplemental [fig_ref] Table 2: Abnormal Laboratory Tests and Polycythemia Vera-Related Signs and Symptoms at Time of... [/fig_ref].
## History of thrombosis
At time of study enrollment, for patients with a known PV diagnosis date (n = 2485), 468 patients (18.8%) had a documented thrombotic event [fig_ref] Table 4: History of Thrombotic Events in 2485 Patients With Evaluable Data [/fig_ref] , and 4.6% had at least 2 thrombotic events. Venous thrombotic events had occurred in 279 patients (11.2%); the most common were deep vein thrombosis (6.1 and pulmonary embolism (2.6%). Arterial thrombotic events had occurred in 218 patients (8.8%); the most common were cerebrovascular arterial thrombosis (5.1%) and acute myocardial infarction (1.8%). A total of 29 patients (1.2%) experienced both a venous and arterial thrombotic event. Before PV diagnosis, 296 patients (11.9%) experienced thrombotic events (venous, 6.7%; arterial, 5.7%). Between diagnosis and enrollment, 207 patients (8.3%) experienced a thrombotic event (venous, 5.2%; arterial, 3.5%). Of the patients who experienced thrombotic events before PV diagnosis, 35 (11.8%) had a subsequent thrombotic event after PV diagnosis. The exposure-adjusted thrombotic event rate between diagnosis and enrollment was 1.55 per 100 patient-years. Among patients who did not have a history of thrombotic events at time of diagnosis, 7.0% had a thrombotic event after diagnosis.
## Concomitant medications
The 3 most common categories of concomitant medications were statins (26.3%), selective β-blocking agents (18.9%), and angiotensin-converting enzyme inhibitors (16.6%). Aspirin use was reported in 62.9% of patients.
## Health care resource utilization
There were 667 patients (26.6%) who reported at least one hospital, urgent care, emergency room, or outpatient visit within the 6 months before enrollment. Most of these patients (82.6%) reported having outpatient medical care visits. PV was a causal factor for 39.9% of all hospital, urgent care, emergency room, and outpatient visits.
# Discussion
The REVEAL study is the first large, prospective observational study to examine the contemporary demographics, burden of disease, clinical management, PROS, and health care resource utilization among US patients with PV. Enrolled patients represent a broad, real-world segment of the PV population who are actively being managed at community or academic centers. In this analysis of clinical and disease characteristics at time of enrollment to REVEAL, patients were mostly male (54.2%) and had a mean (SD) age of 66.3 (12.3) years, with a median (range) PV disease duration of 4.0 (0-56.3) years. In the Cytoreductive Therapy in Polycythemia Vera (CYTO-PV) study, which had similar eligibility criteria, 62.2% were male, mean (SD) age at recruitment/enrollment was 65 (12) years, and mean (SD) time from diagnosis to enrollment was 4.3 (5.4) years. [bib_ref] Cardiovascular ewents and intensity of treatment in polycythemia vera, Marchioli [/bib_ref] Baseline data reported from the 1638 patients included in the database collected during the observational study of the European Collaboration on Low-Dose Aspirin in Polycythemia Vera (ECLAP) project revealed that 57.5% were male, mean (SD) age at recruitment/enrollment was 60 (13) years, and mean (SD) time from diagnosis to enrollment was 5.0 (5.0) years. [bib_ref] Leukocytosis as a major thrombotic risk factor in patients with polycythemia vera, Landolfi [/bib_ref] Demographic statistics from REVEAL are similar to those reported by other highly regarded studies.
Bone marrow biopsy and JAK2 mutation testing are important to differentiate between a PV diagnosis and myelofibrosis or essential thrombocythemia; confirmation of diagnosis by bone marrow biopsy may affect treatment selection and potentially outcomes. At the time of diagnosis, it was reported that < 25% of patients underwent bone marrow biopsy/mspiration, and < 50% underwent JAK2 mutation testing. Of the patients who did undergo JAK2 mutation testing, approximately 96% had a JAK2 V617F mutation, and approximately 1% had other JAK2 mutations (eg, exon 12). Because the majority of patients did not have JAK2 mutation testing to verify diagnosis, it is possible that some do not truly have PV. Importantly, JAK2 mutation testing is increasing over time. According to data reported in this study, of the patients diagnosed in the 3-year intervals of , and 2014-2016, 38%, 57%, 57%, and 61% underwent JAK2 mutation testing, respectively. However, even when taking into consideration that the JAK2 mutation was discovered in 2005 and the time needed for this mutation testing to become widely available to all practice settings, the rates of JAK2 mutation testing reported in this study seem low.
Similarly, the percentage of patients who reportedly underwent bone marrow biopsy and/or erythropoietin testing seems low. These low rates may be related to the various ways in which these data are acquired, recorded, and abstracted from electronic medical records. [bib_ref] Commonly used data-collection approaches in clinical research, Saczynski [/bib_ref] Regarding the management of patients with low-risk disease, in line with National Comprehensive Cancer Network clinical practice guidelines,the majority (54.6%) of patients were managed with phlebotomy with or without aspirin. However, > 40% of lowrisk patients received hydroxyurea or some other cytoreductive therapy with or without aspirin. Conversely, approximately one third of high-risk patients were managed without cytoreductive medications with or without aspirin. These data suggest variance from guideline-defined risk-adapted treatment of PV, which was published after many of these patients were diagnosed and enrolled. [bib_ref] Philadelphia-negative classical myeloproliferative neoplasms: critical concepts and management recommendations from European LeukemiaNet, Barbui [/bib_ref] It is important to note that treatment with aspirin was reported in 62.9% of patients, but it is likely that, in some patients, over-the-counter use of aspirin was not reported or recorded.
With respect to thrombotic events, 11.9% of patients in REVEAL experienced thrombotic events before PV diagnosis (venous, 6.7%; arterial, 5.7%), and 8.3% experienced a thrombotic event between diagnosis and enrollment. In comparison, 14% of patients with PV had thrombotic events before diagnosis, and 19% had a thrombotic event during a 20year follow-up in a retrospective study of 1213 patients in Italian hematology institutions conducted in 1995. 18
# Conclusion
As the first prospective observational study of its size to assess US patients with PV, REVEAL is expected to provide information that will improve our understanding of the demographics, diagnosis, management, and outcomes of these patients. This analysis describes patient demographics as well as clinical and disease characteristics at the time of diagnosis and around the time of enrollment. REVEAL baseline data indicated that the JAK2 V617F mutation rate among patients who underwent molecular testing was consistent with previous reports. [bib_ref] Acquired mutation of the tyrosine kinase JAK2 in human myeloproliferative disorders, Baxter [/bib_ref] [bib_ref] The JAK2-V617F mutation is frequently present at diagnosis in patients with essential..., Lippert [/bib_ref] However, the reported percentage of patients who underwent testing was much lower than expected, which could be related to the challenges of capturing retrospective data from medical records. Many patients had poorly controlled hematocrit, platelet, and leukocyte counts at enrollment, with elevated hematocrit values observed in nearly half of all evaluated patients. Additional prospective data from REVEAL (eg, treatment patterns, PROS, and cardiovascular risk assessments) will be important for understanding how the disease progresses and how management changes with time.
## Clinical practice points
## -
PV is a Philadelphia chromosome-negative myeloproliferative neoplasm that can be diagnosed on the basis of absolute erythrocytosis (hemoglobin > 16.5 g/dL in men, hemoglobin > 16 g/dL in women), bone marrow biopsy, and the presence of JAK2 V617F or JAK2 exon 12 mutation. According to World Health Organization PV diagnostic criteria, bone marrow biopsy may not be required in cases of sustained absolute erythrocytosis (hemoglobin > 18.5 g/dL in men or > 16.5 g/dL in women) in the presence of one of the JAK2 mutations and a subnormal serum erythropoietin level.
## -
In this observational study, the retrospective review and abstraction of data from electronic medical records limits definitive conclusions regarding diagnostic approach.
## -
The primary goal of treatment is prevention of arterial and venous thrombotic events. In general, risk-adapted treatment consists of phlebotomy and aspirin in low-risk patients (< 60 years and no history of thrombosis). In high-risk patients (≥ 60 years and/or with a history of thrombosis), cytoreductive therapy is usually added to phlebotomy and aspirin. Data collected from REVEAL suggest variance from guideline-defined risk-adapted treatment of PV.
- During treatment, patients with PV should be monitored every 3 to 6 months for new thrombohemorrhagic events, increased need/intolerance of phlebotomy, progressive/symptomatic splenomegaly, symptomatic thrombocytosis, progressive leukocytosis, and progressive disease-related symptoms. This close clinical surveillance is necessary to determine whether a cytoreductive therapy is indicated for low-risk patients or whether a different medication should be selected for high-risk patients.
- Baseline characteristics of patients with PV enrolled in REVEAL are similar to those reported by other highly regarded studies.
# Supplementary material
Refer to Web version on PubMed Central for supplementary material. Data are presented as n (%) unless otherwise indicated.
# Pv-related
Abbreviations: EEC = endogenous erythroid colony; JAK2 = Janus kinase 2; LDH = lactate dehydrogenase.
a As reported by survey respondents; definitions of "abnormal" were not available. A patient could be counted in > 1 category. Patients without known PV diagnosis date were excluded from evaluable patient population.
a Thrombotic events without event dates were excluded.
b Patient groups reporting events "before diagnosis" and "between diagnosis and enrollment" are not mutually exclusive.
Clin Lymphoma Myeloma Leuk. Author manuscript; available in PMC 2021 May 25.
[fig] Figure 1: IRB-Approved Study SitesParticipating in REVEAL. A Total of 2510 Patients Were Enrolled From 227 Sites Across the United States (39 Academic and 188 Community [Selfidentified]) Abbreviations: IRB = institutional review board; REVEAL = Prospective Observational Study of Patients With Polycythemia Vera in US Clinical Practices. Myeloma Leuk. Author manuscript; available in PMC 2021 May 25. [/fig]
[fig] Figure 2: Duration Myeloma Leuk. Author manuscript; available in PMC 2021 May 25. [/fig]
[fig] Figure 3: Figure 3. [/fig]
[table] Table 2: Abnormal Laboratory Tests and Polycythemia Vera-Related Signs and Symptoms at Time of Diagnosis in 2510 Patients [/table]
[table] Table 4: History of Thrombotic Events in 2485 Patients With Evaluable Data [/table]
|
Diagnosis-linked antibiotic prescribing in Swedish primary care - a comparison between in-hours and out-of-hours
Background: The rise in antibiotic resistance is a global public health concern, and antibiotic overuse needs to be reduced. Earlier studies of out-of-hours care have indicated that antibiotic prescribing is less appropriate than that of in-hours care. However, no study has compared the out-of-hours treatment of infections to in-hours treatment within the same population.Methods: This retrospective, descriptive study was based on data retrieved from the Kronoberg Infection Database in Primary Care (KIDPC), which consists of all visits to primary care with an infection diagnosis or prescription of antibiotics during 2006-2014. The purpose was to study the trends in antibiotic prescribing and to compare consultations and prescriptions between in-hours and out-of-hours. Results: The visit rate for all infections was 434 visits per 1000 inhabitants per year. The visit rate was stable during the study period, but the antibiotic prescribing rate decreased from 266 prescriptions per 1000 inhabitants in 2006 to 194 prescriptions in 2014 (mean annual change − 8.5 [95% CI − 11.9 to − 5.2]). For the out-of-hours visits (12% of the total visits), a similar reduction in antibiotic prescribing was seen. The decrease was most apparent among children and in respiratory tract infections. When antibiotic prescribing during out-of-hours was compared to in-hours, the unadjusted relative risk of antibiotic prescribing was 1.37 (95% CI 1.36 to 1.38), but when adjusted for age, sex, and diagnosis, the relative risk of antibiotic prescribing was 1.09 (95% CI 1.08 to 1.10). The reduction after adjustment was largely explained by a higher visit rate during out-of-hours for infections requiring antibiotics (acute otitis media, pharyngotonsillitis, and lower urinary tract infection). The choices of antibiotics used for common diagnoses were similar.Conclusions: Although the infection visit rate was unchanged over the study period, there was a significant reduction in antibiotic prescribing, especially to children and for respiratory tract infections. The higher antibiotic prescribing rate during out-of-hours was small when adjusted for age, sex, and diagnosis. No excess prescription of broad-spectrum antibiotics was seen. Therefore, interventions selectively aiming at out-of-hours centres seem to be unmotivated in a low-prescribing context.
# Background
The rise of antibiotic resistance is a global public health threat according to the World Health Organization, and antibiotic overuse is common and results in medicalization, unnecessary costs, and increased antibiotic resistance. However, studies on antibiotic prescribing in primary care regardless of indication show a high level of variability between physicians in different countries.
In primary care in-hours (IH) are usually office hours (in Sweden 08:00 to 17:00) during business days, and out-of-hours (OOH) are the remaining hours. Earlier studies of OOH care have suggested that compared to IH care there are lower adherence to antibiotic guidelines, a higher antibiotic prescribing rate, a higher rate of prescriptions for broad-spectrum antibiotics, and more antibiotic prescriptions during weekends than weekday evenings. In a qualitative study from Belgium, the physicians reported that the threshold for prescribing antibiotics was lower during OOH, but the choice of antibiotics was the same. A more recent Belgian OOH study showed a high antibiotic prescribing rate for all indications, a high rate of not using recommended antibiotics, and an overuse of quinolones. However, a Dutch study found the prescribing quality to be appropriate, and the higher rates of prescribing in OOH were explained by a different population of presenting patients. No previous study has compared the OOH treatment of infections to IH within the same population.
Although Sweden belongs to the European countries with low levels of antibiotic prescriptions, there is still room for improvement. Previous registry-based studies in Sweden have shown a significant reduction in antibiotic prescriptions over the last decade, but these studies have not included OOH. Several Swedish national guidelines concerning the evaluation and treatment of infectious diseases have been published, and generally these guidelines aim at better diagnostics, fewer antibiotics, and more targeted treatments.
Because visits for infectious diseases are common at OOH centres, it is important to evaluate whether OOH visits are associated with increased antibiotic prescribing rates because this would warrant interventions in OOH settings.
The purpose of the study was to describe the trends in antibiotic prescribing over time and to compare diagnosis-linked prescribing in general and in detail between IH and OOH in the same population.
# Methods
## Description of the study population
In 2014, Kronoberg County in southern Sweden had 189,128 inhabitants, which was equal to 2% of the Swedish population. During 2014, there were a total of 243,502 physician visits for all causes and 238,164 other visits (nurses, physiotherapists, behavioural therapists) in primary care, thus there were 1300 physician visits and 1300 other visits per 1000 inhabitants.
During the study period, the number of primary healthcare centres (PHCCs) varied between 28 and 35, with 1-8 family physicians each. There were approximately 100 family physician positions and 50 junior physician positions. At the study start, all PHCCs were publicly run, but since March 2009 a third of the PHCCs have been privately run due to new legislation allowing publicly funded private PHCCs.
At the PHCCs, the patient normally booked an appointment through a telephone call with an office nurse who assessed if the patient needed a physician visit. IH were business days 08:00 to 17:00. In the region there were two OOH centres (OOHCs), and the PHCCs staffed the OOHCs with physicians. Patients were supposed to call a nurse triage first, but could also walk in. The visit fees were the same as for IH visits. Home visits were rare, and usually only performed for urgent cases at elderly care homes. Nurses at the OOHCs were responsible for phone advice, and there was also a national phone advice number for patients where nurses provided medical advice. At the time of the study, no Internet services were available.
OOHC 1 served approximately 125,000 inhabitants and was situated in the neighbourhood of the hospital in city 1. During 2006-2007 the centre was open from 17: 00 to 24:00 on weekdays and from 08:00 to 24:00 on weekends and holidays. From 2008 the centre closed at 21:00. Walk-in patients met a nurse who assessed whether a meeting with a physician was warranted.
OOHC 2 served approximately 63,000 inhabitants and was situated at the emergency department of the hospital in city 2. During 2006-2007 the centre was open from 17:00 to 08:00 on weekdays and around the clock on weekends and holidays. From 2008 the centre closed at 21:00. Walk-in patients generally got to see a physician.
## The kronoberg infection database in primary care (kidpc)
This retrospective, descriptive study was based on data from the KIDPC database, which contains information on all visits with an infection diagnosis and all antibiotic prescriptions with or without a visit in primary care in Kronoberg County in 2006-2014. Annually, there were on average 86,000 visits for infections and 43,000 antibiotic prescriptions reported in the database.
The data in the KIDPC were extracted from the electronic medical records (EMR) used in Kronoberg County (Cambio Cosmic software, Cambio Healthcare Systems AB, Linköping, Sweden) at one instance in 2015 using BusinessObjects (SAP AG, Walldorf, Germany). These data contain detailed information about the patients (age, sex, anonymous ID), the visits (PHCC, geography, IH or OOH), the care providers (physicians, nurses), the investigations (diagnostic tests, x-rays, cultures), and the prescriptions (drugs, dosages, durations). The data were linked together using the anonymous patient ID and visit date. For all physician visits, at least one diagnosis was registered according to the simplified Swedish primary care edition of the International Classification of Disease and Related Health Problems -Tenth Revision (ICD-10). The diagnoses were validated and grouped into four main groups and several subgroups by one of the authors (OC) according to recommendations by Public Health Agency of Sweden. The main groups are respiratory tract infections (RTIs), urinary tract infections (UTIs), skin and soft tissue infections (SSIs), and other infections. The RTI group includes ear infections, and the UTI group includes urogenital infections. The other infections group includes eye infections, gastrointestinal infections, and rare infections (See Additional file 1). Because at least one diagnosis had to be recorded for each physician visit, the data set is considered to be complete. However, no diagnoses were recorded for phone, mail or e-mail consultations, and in these cases, the prescriptions could not be linked to a diagnosis.
Antibiotic prescriptions were identified according to Anatomical Therapeutic Chemical Classification (ATC) code group J01, which includes all oral and parenteral antibiotics, but not antibiotics in ointments or eye drops. Antibiotic prescriptions were linked to diagnoses if within a week after a visit. Antibiotic treatment without a diagnosis of an infection could also result from consultation with a care provider other than a physician or from a non-infection diagnosis at a visit. Information on whether the patients collected the medication at the pharmacies was not available in the present study.
## Data set
All physician visits with an infection diagnosis and all antibiotic prescriptions were extracted from the KIDPC database, resulting in a data set with 702,048 physician visits and 389,263 prescriptions over 9 years. For each visit, data on the patient's age and sex, infection diagnoses, antibiotic treatments, and PHCC were extracted.
A visit was defined as a physical visit to a physician, and a consultation was defined as a phone, mail, e-mail or nurse contact. It was compulsory for the physician to code the diagnosis when documenting the visit. Only physician coded diagnoses were used in this study for consistency. In 3% of the visits more than one infection diagnosis was recorded, and in these cases the main diagnosis was selected based on the severity and the likelihood of the diagnosis resulting in an antibiotic prescription. Consultations were not coded for diagnoses, but could in some instances result in antibiotic prescriptions, for example treatment for UTI or repeat prescriptions.
This study presents descriptive annual data and mean annual change for infections and antibiotic prescribing per 1000 inhabitants divided per main infection group, age group, sex, and per IH and OOH (Tables 1, 2, 3, 4). The data are presented as numbers per 1000 inhabitants per year based on the population of the region as of December 31 of each year. Because the population of Kronoberg County is only 2% of the population of Sweden and the antibiotic prescription rate was lower than the average in Sweden, the numbers reported cannot be extrapolated to the national level. However, the trends are likely to be generalisable.
The IH and the OOH cohorts were compared. The relative risk of receiving antibiotics during OOH was calculated. The proportions of the choice of antibiotics for common infections were reported.
# Statistical methods
All analyses were performed using Excel 2013 (Microsoft, Redmond, WA, USA) and SPSS Version 23 (IBM Corp, Armonk, NY, USA). For descriptive statistics, means, and proportions were used. For annual trends, linear regressions were calculated and presented as mean annual change with 95% confidence interval. Comparisons between groups after adjusting for sex, age, and diagnosis were presented as relative risks with 95% confidence interval. Comparisons between proportions of categorical variables in two independent groups were performed with the chi-square test. P-values ≤ .05 were considered statistically significant.
# Results
The physician visit rate for infections varied during the study and reached a maximum of 469 visits per 1000 inhabitants per year in 2011 and a minimum of 398 visits in 2014. Female patients have more infection visits than male patients, 502 and 366 visits per 1000 inhabitants per year respectively. Children 0-4 years and adults over 80 years had the highest visit rates, 995 and 576 visits per 1000 inhabitants per year respectively. No significant trends were observed in total visit rate nor in visit rate by sex, but the mean annual change in visit rate per 1000 inhabitants per year decreased in children 0-4 years (− 33.7 (95% CI − 56.0 to − 11.5)), increased in adults 65-79 years (7.7 (95% CI 1.1 to 14.3) and in adults over 80 years (13.9 (95% CI 7.6 to 20.2)).
The antibiotic prescriptions per 1000 inhabitants per year decreased significantly from 266 prescriptions in 2006 to 194 prescriptions in 2014 (mean annual change − 8.5 (95% CI − 11.9 to − 5.2)). There was no sex difference, but the decrease in antibiotic prescriptions was more pronounced in children 0-4 years (mean annual change − 35.2 (95% CI − 46.9 to − 23.5)) and in children 5-19 years (mean annual change − 11.7 (95% CI − 17.0 to − 6.5). The antibiotic prescribing frequency decreased mainly for RTIs (mean annual change − 6.5 (95% CI − 9.0 to − 3.9)), explaining 76% of the total reduction. Antibiotic prescriptions without an infection diagnosis and prescriptions for UTIs also decreased, explaining a further 11 and 8% of the total reduction, respectively.
Of all antibiotic prescriptions, 75% were linked to an infection visit on the same day, another 3% were linked to an infection visit within a week before the prescription day, and finally 22% were not possible to link to an infection visit. These proportions were stable during the study period. Of all antibiotics prescribed at visits, 66% were antibiotics commonly used for RTIs, 12% were commonly used for SSIs, 16% were commonly used for UTIs, and 6% were other antibiotics. Of the antibiotics prescribed without an infection diagnosis, 38% were antibiotics commonly used for RTIs, 25% were commonly used for SSIs, 29% were commonly used for UTIs, and 8% were other antibiotics. Of the UTI antibiotics, 36% were prescribed without an infection diagnosis.
During the study period, the OOH infection visits decreased from 65 visits per 1000 inhabitants in 2006 to 43 visits in 2014 (mean annual change − 3.0 visits (95% CI − 4.2 to − 1.7)). Also, the antibiotic prescribing decreased from 43 prescriptions per 1000 inhabitants in 2006 to 26 prescriptions in 2014 (mean annual change − 2.2 prescriptions (95% CI − 3.3 to − 1.2)).
The diagnoses and antibiotic prescription rates between IH and OOH are shown in. During IH, there were 382 infection visits per 1000 inhabitants per year compared to 51.4 during OOH. Thus 12% of all visits were during OOH. RTIs were the most common diagnoses during both IH and OOH. However, acute otitis media, pharyngotonsillitis, and lower UTIs were more common during OOH. A total of 15% of all antibiotics were prescribed during OOH. The likelihood of receiving an antibiotic prescription was 55% during OOH visits compared to 41% during IH visits. The unadjusted relative risk of antibiotic prescribing in OOH was 1.37 (95% CI 1.36 to 1.38) compared to IH. The difference remained unchanged when only adjusted for age and sex 1.37 (95% CI 1.37 to 1.38) and 1.37 (95% CI 1.37 to 1.38), respectively. However, when adjusted for age, sex, For the six most common diagnoses treated with antibiotics, a comparison of treatment choice per diagnosis with IH and OOH visits was made. The prescription rate was higher during OOH for pneumonia, acute otitis media, and pharyngotonsillitis. Although the difference was statistically significant, the choices of treatment for each diagnosis were comparable between IH and OOH prescriptions.
# Discussion
During the study period, the level of infection visits was constant, but the antibiotic prescription rate decreased. Fewer prescriptions in children and for RTIs were the main reasons for the reduction. During OOH, there was a reduction both in infection visits and in antibiotic prescribing. The antibiotic prescription rate was higher during OOH than during IH, and when adjusting for age, sex, and diagnosis the difference was significant but small. The choices of treatments were similar. This study showed that women visited primary care for infections more often than men and also received antibiotic treatment more often than men. The same pattern has been seen in other studies from Denmark, the Netherlands, and the United Kingdom. The sex difference in the incidence of lower UTI was an important reason.
Our data on visit rates per 1000 inhabitants per years for infections were similar to the Primary Care Record of Infections in Sweden (PRIS) database, which consists of data since 2007 on visits with an infectious diagnosis and all antibiotic prescriptions from voluntarily participating PHCCs on an annual basis. Antibiotic prescriptions are in most cases linked to diagnoses and also includes information about age, sex, and laboratory results. The database has a larger dataset than in this study covering PHCCs in other regions but lacks OOH data. In the PRIS database, the visit rates per 1000 persons per year for infections during IH were 457 (in 2008), 441 (in 2010), and 406 (in 2013).
The total antibiotic prescribing in primary care decreased by 27% in this study. However, in the PRIS databasethe reduction of IH antibiotic prescribing was 36%, as the IH antibiotic prescription per 1000 persons per year decreased from 245 (in 2008) to 201 (in 2010) to 157 (in 2013). For the corresponding years in our study, the IH antibiotic prescriptions per 1000 inhabitants were 212, 217, and 186, respectively. It is possible that participation in the PRIS database could have triggered a more restrictive antibiotic prescribing behaviour compared to our real-life study. A Finnish studyreported a 47% reduction in antibiotic prescriptions to children in primary and other out-patient care between 2010 and 2016, whereas our present study showed a 38% reduction in children in primary care between 2010 and 2014.
Several explanations are possible for the reduction in antibiotics prescriptions. For example, there might be increasing awareness among the general public that the use of antibiotics should be avoided when they are not needed. Also, physicians might have become more restrictive in prescribing. Another reason might be due to the antibiotic stewardship work performed by the Strama group, the Swedish strategic programme against antibiotic resistance. In 2005, Strama together with the government launched a national strategy to prevent antibiotic resistance and healthcare-associated infections. Several actions have been performed in relation to this strategy. Diagnosis-specific guidelines for optimal antibiotic use have been published and promoted, and the use of antibiotics has been reported at the local, regional, and national level. During 2011-2014, the Swedish government ran a patient safety campaign aiming to decrease antibiotic use with the goal of fewer than 250 annual prescriptions in out-patient care per 1000 inhabitants for all prescribers together (primary and secondary care, dental care) resulting in a decrease from 385 prescriptions (2011) to 328 prescriptions (2014). Furthermore, a pneumococcal conjugate vaccine was introduced in the Swedish national vaccination programme for children in 2009. Finally, a national economic bonus system was introduced for regions achieving a reduction in the antibiotic prescription levels, and incentive for quality outcome with the same goal was introduced in 2011 at the PHCC level in Kronoberg County.
During the period studied here, the number of OOH infection visits decreased by a third. Factors contributing to the decrease were shorter opening hours at the end of the study, a penalty fee (100 euros) introduced in 2008 for the PHCC for each patient attending the OOHC, and the introduction of a nurse triage system for walk-in patients at OOHC1. The OOH antibiotic prescription rate per 1000 inhabitants per year was at the same level in the Netherlands, Sweden, and England (20, 28, and 31 prescriptions, respectively), but higher in Denmark (80 prescriptions). Two English studies have shown stable or increased OOH antibiotic prescription rates from 2010 to 2014. In contrast, our study showed a decrease in antibiotic prescription rates.
The main explanation for excess prescribing during OOH is that infections that are often treated with antibiotics were more common during OOH visits such as acute media otitis, pharyngotonsillitis, and lower UTIs. The relative risk of antibiotic prescribing was decreased when adjusting for diagnoses. For SSI, the relative risk of receiving antibiotics during OOH remained elevated 1.20 (95% CI 1.18-1.23). It was uncommon to prescribe UTI antibiotics without a visit with infection diagnosis during OOH service (9% of UTI antibiotic prescriptions were without a visit during OOH compared to 39% during IH) although it was in line with current guidelines. This fully explained the higher UTI visit rate during OOH.
These results are similar to other European studies when comparing OOH and IH. A Norwegian comparison of tonsillitis and acute media otitis showed no difference in the prescription rate at OOHCs, and a Dutch study showed higher prescription levels during OOH for common infections and argued that the patients were sicker in the sense that they had more urgent problems that could not wait until the next day based on a revision of the EMR.
The remaining excess prescriptions during OOH after adjusting for diagnosis were estimated, leading to 2.2 more prescriptions per 1000 inhabitants per year compared to IH, which corresponds to 7.9% of the prescriptions during OOH and to 1.2% of all prescriptions during IH and OOH together. These prescriptions could partly be explained by sicker patients in need for urgent evaluation and an absence of control visits in the OOH setting. On the other hand, a reason could be a lower threshold to prescribe during OOH for example due to high workload or due to limited possibility to arrange for follow-ups.
Apart from the high relative risk of receiving antibiotics for SSI during OOH, there are no apparent areas to intervene. But because the total decrease of antibiotic prescriptions during the study period is 27% and the excess prescriptions during OOH are just above 1% of all antibiotic prescriptions, there There were no differences in treatment choice, which corresponds with other quantitative studies from Norway and the Netherlandsand with a Belgian qualitative study where physicians reported the treatment choice to be the same as during IH, although the threshold to prescribe was lower at OOHCs. In contrast, an English study noted a higher proportion of broad-spectrum antibiotics during OOH.
# Strengths
The data set was complete for infection visits and antibiotic prescriptions in primary care in a region in Sweden. Because the whole region was included, the data were real-life data without any selection due to study participation. Also, the same EMR system was used during the study period thus decreasing the risk for information errors. Because writing a diagnosis was compulsory for all visit records, very few diagnoses were missing. All OOH infection visits and prescriptions were included, which enabled comparisons between IH and
# Limitations
Limitations of the study include that no validation of diagnoses by examining the EMR was done. Also, the reason why some antibiotics are prescribed without a coded infection diagnosis has not been explored. A lower threshold to diagnose infections and to prescribe antibiotics in the OOH setting cannot be ruled out but would also be hard to verify in the EMR. Other antibiotics than oral and parenteral antibiotics (ATC code J01) are missing in the dataset, such as antibiotics in topical skin and eye preparations. The antibiotic rate for the elderly (> 80 years) might be underestimated due to partly missing data for patients with medication administered through a dispensing system. Furthermore, we could not measure the rate of delayed prescribing because we did not have access to pharmacy dispensing data. The common way of delayed prescribing in Sweden is that the patient receives an electronic prescription but is recommended to wait a few days before collecting the prescription.
# Conclusions
Although the infection visit rate was unchanged, there was a significant reduction in antibiotic prescribing, especially to children and for RTIs. The increased antibiotic prescribing rate during OOH was small when adjusted for age, sex, and diagnosis, and no excess prescribing of broad-spectrum antibiotics was seen. Therefore, interventions selectively aiming at OOHCs seem to be unmotivated in a low-prescribing context.
## Supplementary information
Supplementary information accompanies this paper at https://doi.org/10. 1186/s12879-020-05334-7. |
Structural transitions in the RNA 7SK 5′ hairpin and their effect on HEXIM binding
S1 Molecular Dynamics simulations of HP1 24-871 In our first simulation study of HP1 1 we considered the system with all bases neutrally charged.2The hairpin was observed to be rather flexible. However, a careful analysis of the crystal 3 structure, also confirmed by the recent NMR structure (Exp4), highlighted the presence of three 4 protonated bases (C71, C75 and A77). We report here results of molecular dynamics (MD) 5 simulations for the experimental structure Exp2 in a system with all neutral bases 1 and for a 6 system where the bases C71, C74 and A77 are positively charged, as suggested by the crystal 7 structures. MD simulations were performed for the fully solvated system and run for 200 ns. 8 We observe that the protonated molecule adopts a straighter configuration with respect to the 9 neutral model, and it fluctuates less (seeFigure S1). This increased rigidity of the protonated 10 model is also reflected by the RMSD and by the overall bend. The RMSD for the protonated 11 molecule remains closer to the initial (experimental) structure, with values ranging from 2 to 12 4Å . The RMSD for the neutral model has significantly higher values and larger fluctuations, 13 ranging from 6 to 10Å with respect to the crystal structure. We also analyse the overall bend of 14 the molecule, as given by the angle formed by atoms P28-P48-P57. We find that the protonated 15 model remains straighter (values closer to 180 degrees) and exhibits fewer fluctuations than 16 the unprotonated model, for which the bend increases as the simulation progresses, with values 17 ranging from the initial 160 degrees to as low as 100 degrees. 18 * [email protected] † [email protected] S1 The output of H-REX simulations are trajectories for a fixed value of λ (λ-trajectories), there-98 fore subject to the exchanges of replicas. Using PLUMED, we reconstructed the 24 temporal 99 trajectories, which are continuous trajectories for a given molecule, subject to a Hamiltonian of 100 varying intensity. In our analysis we focused on the behavior of the variable ξ, discriminating 101 between the three families of configurations, M1, M2 and E. 102 From the temporal trajectories we computed the autocorrelation time for ξ for the simulations 103 starting from M1 and from M2 (Table S2). The overall simulation time of 200 ns is large com-104 pared to the characteristic times of the simulations, which average between 8 and 14 ns. We 105 observe in some examples that in a given temporal trajectory ξ exhibits several transitions 106 between values characteristic of M1, M2 and E (Fig. S10) 107 S15 Figure S10: Temporal trajectories: evolution of ξ for two trajectories starting from M1 (top) and two trajectories starting from M2 (bottom). A block analysis for the λ=1 trajectory (full Hamiltonian) exhibits a shift in population from 108 M1 to M2 for simulations starting from M1, and a more stable M2 population for simulations 109 starting from M2 (Fig. S11). From the λ-trajectories we computed the populations of M1, M2 110 and E for each value of λ(Table S3)focusing on the second half of the trajectories (100 ns to 111 200 ns) to avoid the initial transient behavior of the simulations. As expected, as the value of 112 λ decreases, local structures are lost, ξ assumes a broad distribution, and results are similar for 113 the simulations starting from Exp1 and from Exp2. 114 S16 : Comparison of the fluctuations of the experimental structure Exp2 for a fully atomistic MD simulation. Left: Superposition of simulated structures for the model with protonated C71, C74 and A77; Right: superposition of simulated structures for the model with all neutral bases.
## S2 molecular dynamics setup
## 19
S2.1 Equilibration 20
The following protocol was used for both standard MD simulations and for H-REX simulations. All the runs were performed using Gromacs 5.1.2. 2 Each RNA system was first neutralized with 22 potassium cations, and then solvated in a 6 nm side cubic box containing a 0.2 M KCl solution. All bonds involving hydrogen atoms were maintained rigid using the LINCS algorithm. 3 Tem-24 perature was maintained at 300 K using the stochastic velocity rescaling algorithm with a time 25 constant of 0.1 ps, 4 and pressure was fixed to 1 bar using the Parinello-Rahman barostat 5 with a 26 time constant of 0.5 ps. A cut-off of 1.1 nm was employed for van der Waals and real-space elec-27 trostatics interactions, while long-range electrostatics was treated using PME 6 with a Fourier 28 grid spacing of 1Å −1 . All simulations were then equilibrated as follows: first, the potential en-29 ergy was minimized until convergence; the solvent molecules were then relaxed for 80 ps at the 30 target temperature, while the macromolecule was maintained fixed. The whole system was then 31 heated to 300 K for 400 ps, and later equilibrated for another 400 ps at 300 K before subsequent 32 production runs were performed. : Comparison of the RMSD (top) and the overall bending (bottom) of the experimental structure Exp2 for a fully atomistic MD simulation of the protonated and the neutral system.
## 33
## S2
## S2.2 hamiltonian replica exchange simulations
## 34
For selected systems, enhanced sampling of the RNA conformational space was achieved using 35 a Hamiltonian replica exchange scheme. The REST2 7-9 algorithm was employed, as it can 36 be successfully applied to biomolecules in explicit solvent. Instead of performing simulations 37 of several replicas with different physical temperatures, as in the more standard temperature 38 S3 replica exchange approach, in the REST2 approach only the Hamiltonian of the biomolecule is 39 rescaled, and the number of replicas thus scales with the biomolecule size, and not that of the 40 entire simulation box. Details about this algorithm and its application to biomolecules can be 41 found elsewhere. 7-9 Briefly, in a given replica n, all RNA-RNA interactions are rescaled by λ n 42 and all RNA-solvent interactions by √ λ n , with 0 < λ n ≤ 1. In our REST2 setup, we employed 43 24 replicas and a distribution of λ n such that for the n-th replica (from 1 to 24) λ n = (1/2) n/24 .
## 44
REST2 simulations were propagated for 200 ns using Gromacs 5. : Projection of the path sampling database of native sequences onto the variables ξ and the angle U44-A65-U63, describing the inward or outward position of U63. Structures included in heat-plots are the minima accessible from M1 or M2 with a threshold of 15 kcal/mol from the bottom of the respective free energy funnels. This threshold means we only include minima that can be reached from the respective structure on a time of milliseconds. The projection only represents the density of minima, and no occupation probabilities or energies are used to weight the result. : Hydrogen bonds formed between peptide residues and RNA. We measure the presence of a bond as a percentage over the full MD trajectory (dark blue 100%, white 0%). Dashed boxes highlight the interactions of U40 (green) and of the GAUC/GAUC motif (red).
## S8
## S5 mutations: data and figures
Here, we present more detailed results for the mutants. and S8 show the cor-94 responding energy landscapes. In a summary of structural descriptors is given, with 95 labels indicating their position in the disconnectivity graphs provided.
[formula] 96 S9 A B C D 4 kcal mol −1 ξ 0 - 90 - 180 - A B 4 kcal mol −1 ξ 0 - 90 - 180 - A B C D E F G 4 kcal mol −1 ξ 0 - 90 - 180 - U40C [/formula]
U40C-U41C U41C : Potential energy disconnectivity graphs for the U40C, U40C+U41C and U41C mutants of the RNA 7SK HP1 hairpin. The labels A to G correspond to distinct structural ensembles, which are characterised in detail in . This figure otherwise corresponds to in the main text. : Potential energy disconnectivity graphs for the delU63, the doubleCG and doubleUA mutants of the RNA 7SK HP1 hairpin. While the deletion of U63 still allows for the formation of some M2 structures, the changes in the 7SK motif in the doubleUA and doubleCG mutants, which destabilise the T2 triplet, lead to an increase in the number of extended structures. The labels A to D correspond to distinct structural ensembles, which are characterised in detail in .
[formula] S10 A B C 4 kcal mol −1 ξ 0 - 90 - 180 - A B C 4 kcal mol −1 ξ 0 - 90 - 180 - A B C D 4 kcal mol −1 ξ 0 - 90 - 180 - delU63 doubleCG doubleUAS11 A B C D 4 kcal mol −1 ξ 0 - 90 - 180 - A B C D E F 4 kcal mol −1 ξ 0 - 90 - 180 - A39G A39G-U68C A B C D E 4 kcal mol −1 ξ 0 - 90 - 180 - A B C D 4 kcal mol −1 ξ 0 - 90 - 180 - A39U [/formula]
A39U-U68A : Disconnectivity graphs for the potential energy landscapes of the A39G, A39G-U68C, A39U and A39U-U68A mutants. The labels A to F correspond to distinct structural ensembles, which are characterised in detail in . S12 : Summary of discrete path sampling results for all mutants. For distinct funnels on the energy landscapes shown in -S8 we report the potential energy difference between the global minimum and funnel bottom, ∆V gmin , the depth of the funnel (potential energy from the lowest minimum in the funnel to the lowest energy transition state connecting it to the global minimum), ∆E funnel , the classification as M1, M2 or E structures, indicating whether we detect the artificial kink of the lowest stem folding back on the rest of the structure (Ek), the position of U63, the average number of hydrogen bonds made by the nucleotides 40 and 41, including bonds made by the base as well as by the phosphate group, and the percentage of structures exhibiting a U40-X68 base pairing. S13 : Projection of the path sampling database of native and mutated sequences onto the variables describing the behaviour of U40 and U41 with respect to the triplets T1 and T2 (T2 angle U40-U41-G42 ≤ 40, T1: angle U40-U41-G42 ≥ 160) and the behavior our U63 (inward: angle U44-A65-U63 ≤ 90, angle outward: U44-A65-U63 ≥ 90). These projections simply present densities of minima with respect to the two chosen parameters. These plots are not weighted by occupation probabilities or energies. S14 M1 (ns) : Temporal trajectories: autocorrelation times of the variable ξ from the 24 reconstructed continuous trajectories of the H-REX for the wild system, starting from either M1 or M2 as initial configurations. The last line gives the average and the standard deviation over all replicas. : Block analysis of the distribution of ξ for the full Hamiltonian trajectories starting from M1 (top) and M2 (bottom). The trajectory has been divided into 5 blocks of 40 ns each.
## S6 statistical analysis of h-rex
## 97
|
The Impacts of Climate Change Mitigation Strategies on Animal Welfare
Simple Summary: Climate change is probably the most important environmental issue of our time. Raising animals for food contributes to the production of greenhouse gases implicated in the global warming that is causing climate change. To combat this ecological disaster, a number of mitigation strategies involving changes to agricultural practices have been proposed. However, some of these changes will impact the welfare of farmed animals. This paper reviews selected climate change mitigation strategies and explains how different approaches could have negative or positive effects.Abstract: The objective of this review is to point out that the global dialog on reducing greenhouse gas emissions in animal agriculture has, thus far, not adequately considered animal welfare in proposed climate change mitigation strategies. Many suggested approaches for reducing emissions, most of which could generally be described as calls for the intensification of production, can have substantial effects on the animals. Given the growing world-wide awareness and concern for animal welfare, many of these approaches are not socially sustainable. This review identifies the main emission abatement strategies in the climate change literature that would negatively affect animal welfare and details the associated problems. Alternative strategies are also identified as possible solutions for animal welfare and climate change, and it is suggested that more attention be focused on these types of options when allocating resources, researching mitigation strategies, and making policy decisions on reducing emissions from animal agriculture. Animals 2015, 5 362
# Introduction
Climate change is arguably the most important environmental issue of our time. With severe and widespread destructive effects, warming of the planet threatens ecological systems, peoples' livelihoods, and species survival [bib_ref] Contribution of Working Group II to the Fifth Assessment Report of the..., Ipcc ; Field [/bib_ref]. Animal agriculture is an important source of greenhouse gas (GHG) emissions and has been implicated as a serious contributor to climate change. To reduce the contribution of emissions attributable to animal agriculture, a number of mitigation strategies involving changes to farming practices have been proposed. Although this is an important and timely goal, and many of the proposed solutions seem reasonable on the surface, mitigation strategies can have complex effects on people and animals in practice. While there has been occasional mention [bib_ref] The nexus of environmental quality and livestock welfare, Place [/bib_ref] , in the global discussion on climate change there has generally been a dearth of attention paid to the animal welfare impacts of the proposed abatement options, and some of the suggested livestock management approaches would have severe and wide-ranging impacts on the animals.
At the same time, in other arenas there is a growing international social movement afoot aimed at addressing animal welfare, the physical and psychological state of animals [bib_ref] Science-based assessment of animal welfare: Farm animals, Duncan [/bib_ref]. There are many animal welfare issues associated with commercial farming practices, especially in industrial agricultural production systems, where animals are often confined indoors at high stocking densities. The conditions in which the animals are kept are a matter of serious deliberation in legislative, corporate, investment and trade organizations, among many others around the world. Overarching calls to increase animal productivity, by, for example, pushing animals (either genetically through selective breeding, with hormone drugs, or through feed changes) to produce more milk or meat, or to produce more offspring faster, can also be considered in light of potential health and welfare consequences. This paper argues that the potential impact of climate change mitigation strategies on animal welfare should be acknowledged and considered in the ongoing debate and discussion centered on climate change approaches and climate change research, with the aim of maximizing co-benefits of sustainability while avoiding negative tradeoffs. Specific mitigation strategies are discussed and their potential impacts on the animals are assessed. As examples of alternative pathways forward, some mitigation approaches that have no effect on animal welfare or even a positive effect are also reviewed. Given the scope and severity of the impacts on animals, and the increasing global support for more welfare-friendly practices, animal welfare concerns should become a more prominent part of the dialog and policy decisions on climate change mitigation practices in animal agriculture.
## Background on climate change and mitigation strategies
The Earth's surface has been successively warmer over each of the last three decades compared to any preceding decade since 1850. Human activities are responsible (with a 95%-100% probability) for the recent global warming and the marked increase in global atmospheric concentrations of carbon dioxide (CO 2 ), methane (CH 4 ) and nitrous oxide (N 2 O) to above pre-industrial values [bib_ref] The Physical Science Basis. Contribution of Working Group I to the Fifth..., Ipcc [/bib_ref]. By sector, greenhouse gas pollution originates primarily from industry; agriculture, forestry and other land use (AFOLU); buildings; and transport [bib_ref] Mitigation of Climate Change. Contribution of Working Group III to the Fifth..., Ipcc ; Edenhofer [/bib_ref]. Separated out, animal agriculture emits a significant portion, estimated from 7% to 18% of anthropogenic greenhouse gas emissions.
Livestock contribute both directly and indirectly to climate change. Enteric fermentation and manure associated emissions are direct, while production and transport of feed (including the fossil fuels used in manufacturing chemical fertilizers) and land use changes (such as conversion of forest to pasture and crop land) contribute indirectly. About 44% of the emissions generated by livestock are CH 4 , which is released during enteric fermentation (eructation in ruminants) and emitted from manure decomposition; 27% are in the form of CO 2 emitted during the production and transport of animal products and feed, and 29% are N 2 O attributable to manure and fertilizer. There is a clear need to address the contribution that farming animals makes to global warming.
To combat climate change caused by animal agriculture, a number of mitigation strategies have been proposed, as well as policy options to implement such strategies [bib_ref] Policy options in addressing livestock's contribution to climate change, Gerber [/bib_ref]. A theme that runs through this literature is greater efficiency of production, with the aim to achieve more output for a given level of resources used. As explained in the pivotal 2006 United Nations, Food and Agriculture (FAO) report, Livestock's Long Shadow:
The most promising approach for reducing methane emissions from livestock is by improving the productivity and efficiency of livestock production, through better nutrition and genetics. Greater efficiency means that a larger portion of the energy in the animals' feed is directed toward the creation of useful products (milk, meat, draught power), so that methane emissions per unit product are reduced. The trend towards high performing animals and towards monogastrics and poultry in particular, are valuable in this context as they reduce methane per unit of product. The increase in production efficiency also leads to a reduction in the size of the herd required to produce a given level of product. Because many developing countries are striving to increase production from ruminant animals (primarily milk and meat), improvements in production efficiency are urgently needed for these goals to be realized without increasing herd sizes and corresponding methane emissions(pp. [bib_ref] Behavioral effects of "step-up" ractopamine feeding program on finishing pigs, Poletto [/bib_ref] [bib_ref] Effects of a "step-up" ractopamine feeding program, sex, and social rank on..., Poletto [/bib_ref] However, by far, the most substantial emissions reductions can come through adaptations in current systems rather than requiring a shift to industrialized systems. This is important from an animal welfare perspective. While there can be negative animal welfare impacts in all production systems, industrialized production has inherently problematic effects on billions of animals globally. Thus, it is important that there is significant alignment in both animal welfare and climate goals. Some ways in which they are less compatible are discussed in Section 2.
## Mitigation strategies and possible negative animal welfare effects
## Feed changes
Given that feed production accounts for about 47% of livestock emissions(p. 18), it is a key target for mitigation. However, there are a number of welfare concerns with some of the current proposals to reduce emissions using feed alterations.
## Grain finishing feedlot cattle
One strategy for mitigating methane emissions with feed changes is by including more concentrates in addition to, or in place of, a proportion of dietary forage. Although emissions per animal may rise as productivity improves, there is generally a reduction in the greenhouse gases released on a per kg-product produced basis. In intensive production systems, grain feeding of livestock often entails removing cattle raised for beef from pastures to provide the finishing ration in a feedlot. This is a common system in the United States and is being emulated in other parts of the world, such as Latin America. In Brazil, the practice of finishing cattle in feedlots increased 50% (from two to three million animals) from 2003 to 2010 [bib_ref] Current outlook and future perspectives of beef production in Brazil, Millen [/bib_ref].
The digestive system of cattle is best adapted to roughage provided by grass-based diets, and the natural behavior of cattle is to graze throughout the day. However, the typical routine in feedlots is to gradually reduce the proportion of forage in the ration over time, eventually reaching diets that can be as high as 90% grain [bib_ref] A review of bloat in feedlot cattle, Cheng [/bib_ref]. In enlarging Brazilian feedlots, the expected trend is a decrease in the amount of roughage in feedlot diets, with an already observed reduction from 29% in 2009 to 21% in 2011 [bib_ref] Survey of nutritional recommendations and management practices used by Brazilian feedlot consulting..., Millen [/bib_ref].
In cattle feeding systems with high ratios of concentrate to forages, digestive problems are common. The starches in this concentrated diet are quickly digested and fermentation acids can build up disrupting the normal function of the rumen [bib_ref] A review of bloat in feedlot cattle, Cheng [/bib_ref]. Abnormal rumen function and digestive disorders can lead to acidosis, bloat, and if persistent, liver abscesses, and even the foot disorder laminitis [bib_ref] Laminitis-like changes in the claws of feedlot cattle, Greenough [/bib_ref] [bib_ref] Sudden death of feedlot cattle, Glock [/bib_ref]. Maintaining normal rumen function on these diets is a constant challenge for the feedlot industry; one quarter of cattle mortality in feedlots can be attributed to digestive disorders [bib_ref] Impact of disease on feedlot performance: A review, Smith [/bib_ref].
For both acidosis and bloat, increasing the amount of forage in the diet can stimulate chewing and promote saliva production, which have a buffering effect, and thus reduce the likelihood of a sinking pH level and destabilization of the microbial population in the rumen [bib_ref] A review of bloat in feedlot cattle, Cheng [/bib_ref] [bib_ref] Acidosis in cattle: A review, Owens [/bib_ref].
Finishing calves in a feedlot usually involves transporting them from their natal pastures. In the United States, nearly two-thirds (63%) of cattle and calves raised for beef are sold through an auction barn after leaving the farm, thus, many calves are moved twice before arrival at a feedlot.
Transport takes a physical and psychological toll on animals. Stressors include unfamiliar surroundings, novelty, noise, vibration, social regrouping, loading and unloading, and feed and water deprivation [bib_ref] The welfare of calves during handling and transport, Trunkfield [/bib_ref] [bib_ref] Cattle transport: Historical, research, and future perspectives, Swanson [/bib_ref] , although many factors, such as preconditioning [bib_ref] Effects of pre-haul management and transport duration on beef calf performance and..., Schwartzkopf-Genswein [/bib_ref] and location in the vehicle during the journey [bib_ref] Associations of beef calf wellness and body weight gain with internal location..., White [/bib_ref] can alter the effect. The stress of transport can lead to immunosuppression and an increase in disease susceptibility, including the ubiquitous Bovine Respiratory Disease (RBD) problem [bib_ref] Effect of transportation and weaning on humoral immune responses of calves, Mackenzie [/bib_ref] [bib_ref] Influence of stress and nutrition on cattle immunity, Carroll [/bib_ref] ].
## Concentrated diets for pigs
In cases where intensification of animal agriculture is accompanied by better access to resources, then there may be benefits for welfare in terms of providing more nutritionally complete rations. However, pigs are also often fed highly concentrated grain-based diets in intensive production. Concentrated diets may be fed to pigs in a finely ground or pelleted ration and are expected to reduce GHG emissions due to their high digestibility. For example, in a case study of intensive pig production in East and South East Asia, the FAO explored changes to swine diets as a potential mitigation strategy, including increased digestibility of feed. In intermediate (not fully industrialized) systems, by changing the feed along with other management changes, emissions could be reduced by 32% to 38%.
Pigs' stomachs are biologically designed for numerous, small meals throughout the day of high fiber feedstuffs [bib_ref] Feeding behaviour in pigs, Nielsen [/bib_ref]. However, in intensive production, they have little access to roughage. Feed processing improves feed-to-gain efficiency, but it is also associated with intestinal problems. These low fiber diets can impact mucosa integrity leaving pigs prone to gastric ulcers in the pars esophageal region of the stomach [bib_ref] Effect of feed processing on growth performance and gastric mucosa integrity in..., Millet [/bib_ref]. This is a problem worldwide. Reports vary, with incidence of pigs showing distinct signs of ulceration between farming operations ranging from zero to nearly 90% [bib_ref] Factors in the housing environment of finisher pigs associated with the development..., Amory [/bib_ref] [bib_ref] Risk factors for gastric ulcers in australian pigs at slaughter, Robertson [/bib_ref] [bib_ref] The incidence of gastric lesions in slaughtered pigs in Ibadan, Majekodunmi [/bib_ref] [bib_ref] Gastric ulcers in fattening pigs: Isolation of Arcobacter spp. From stomachs with..., De Oliveira [/bib_ref] [bib_ref] A note on the frequency of gastric ulcers detected during post-mortem examination..., Swaby [/bib_ref]. For pigs in the growing/finishing stages of production, ulceration is a common cause of mortality [bib_ref] Gastric ulcers: An under-recognized cause of mortality and morbidity, Friendship [/bib_ref]. The production systems in which most pigs are kept appear to have a large impact on the incidence of these ulcers, as pigs with access to straw, sawdust, or outdoor paddocks have fewer ulcers than those confined on bare, solid or slatted concrete floors [bib_ref] Factors in the housing environment of finisher pigs associated with the development..., Amory [/bib_ref] [bib_ref] Effects of cereal disintegration method, feeding method and straw as bedding on..., Nielsen [/bib_ref] [bib_ref] Comparison of the severity of esophagogastric, lung and limb lesions at slaughter..., Ramis [/bib_ref] ].
## Feed additives
Feed additives can be used to inhibit CH 4 production in a variety of species. However, some dietary additives that are being tested have potential animal health concerns. For example, in vitro studies have shown that fumarate can work as an alternative to CH 4 as a sink for H 2 in the rumen, but feeding fumarate in its free acid form can lower rumen pH and inhibit fiber digestion, while feeding it in salt form can induce toxicity. Effective doses may well exceed current dietary recommendations [bib_ref] Proprionate precursors and other metabolic intermediates as possible alternative electron acceptors to..., Newbold [/bib_ref]. Especially in low protein diets, nitrates are also potential CH 4 mitigation agents. However, the intermediate product from nitrate metabolism, nitrite, is toxic, and so feeding it requires a gradual acclimation to a high dietary intake in order to avoid nitrite toxicity [bib_ref] Effects of acclimation to high nitrate intake on some rumen fermentation parameters..., Alaboudi [/bib_ref]. Another proposed supplement, sulfate, has been linked to sulphur-induced polioencephalomalacia, a condition in US feedlot cattle associated with excessive production of H 2 S in the rumen from diets high in distillers grains-co-products from the ethanol industry, which are high in sulfate [bib_ref] Hydrogen Sulphide: Synthesis, Physiology, Roles and Pathology Associated with Feeding Cattle Maize..., Schoonmaker [/bib_ref]. The FAO cautions that more research is needed. Mitigation strategies should steer clear of feeding substances that might endanger the animals, either outright or in the absence of highly skilled management.
## Genetic selection
In general, as animals become more productive, greater feed intake is needed to support the energy requirements associated with growth or lactation, but the proportion of total energy requirements required for maintenance decreases [bib_ref] Developing breeding schemes to assist mitigation of greenhouse gas emissions, Wall [/bib_ref]. Thus the feed input per unit of product produced is reduced. Further, with increases in productivity, fewer animals are needed to reach a given level of output [bib_ref] Developing breeding schemes to assist mitigation of greenhouse gas emissions, Wall [/bib_ref]. In the climate change mitigation literature, this concept is often exemplified with dairy cows. For example, in 1944, the diverse breeds comprising the US dairy cow population produced an annual milk yield of 2074 kg per cow, but by 2007 the genetic pool had shrunk, and Holsteins were the dominant breed, while milk production more than quadrupled to 9193 kg per cow. In 1944 there were 25.6 million dairy cows in the United States with an annual production of 53.0 billion kg of milk, while in 2007 the total US dairy herd had declined to 9.2 million cows, but they produced 84.2 billion kg of milk. Because of this per cow increase in production, the total carbon footprint for the US dairy industry in 1944 was 194 million tons of CO 2 -equivalent (CO 2 -eq.), while in 2007 the industry's carbon footprint was only 114 million tons [bib_ref] The environmental impact of dairy production: 1944 compared with, Capper [/bib_ref]. Much of this individual increase in productivity can be attributed to genetic selection, although dietary changes also contributed.
Swine and poultry produce relatively small amounts of enteric CH 4 , but their manure can be a significant source of GHG production. Therefore, maintaining or improving feed conversion efficiency in these species, and subsequently reducing the volume of manure produced for a given level of output, is considered a major strategy for mitigating CH 4 and N 2 O [bib_ref] Mitigation of methane and nitrous oxide emissions from animal operations: III. A..., Hristov [/bib_ref]. However, there is evidence across species that overemphasis on productivity in genetic selection can lead to unintended negative effects on animal welfare.
## Effects of breeding for yield on dairy cow welfare
While selective breeding has been successful in producing cows that yield more milk with less feed, this change is not without costs. Detrimental, unintended side-effects on the health of dairy cows have accompanied the physiological demands of greater productivity [bib_ref] Undesirable side effects of selection for high production efficiency in farm animals:..., Rauw [/bib_ref].
Following the birth of her calf, a dairy cow enters a state of negative energy balance. Genetic selection for increased milk yield increases feed intake, but high yielding cows can experience greater negative energy balance [bib_ref] Genetic correlation between days until start of luteal activity and milk yield,..., Veerkamp [/bib_ref] than cows producing less, so they are vulnerable to health problems. There is a strong correlation between maximum metabolic load and laminitis [bib_ref] Relationships between energy balance and health traits of dairy cattle in early..., Collard [/bib_ref] , lameness and ketosis [bib_ref] Genetic parameters of health disorders, and relationships with 305-day milk yield and..., Van Dorp [/bib_ref] [bib_ref] The relationship between milk yield and the incidence of some diseases in..., Fleischer [/bib_ref]. Failure to meet the nutritional demands of highly productive dairy cows can also lead to an array of additional metabolic diseases including fatty liver disease, ketosis and periparturient hypocalcemic paresis (milk fever) [bib_ref] Major advances in our understanding of nutritional influences on bovine health, Goff [/bib_ref].
Mastitis, a bacterial infection of the udder, is the most commonly reported health problem in the US dairy industry, affecting 16.5% of cows. The udder and teats of mastitic cows are sensitive to the touch, and in severe cases, clinical signs include swelling and pain sometimes accompanied by fever and depression. It is one of the leading reasons for culling of cows. Although there are many risk factors for mastitis, such as lack of cleanliness in cow housing areas, research has consistently shown an association between mastitis and the amount of milk produced. For example, a study of over 3000 cows in 85 trials in eight different countries found that even with similar management conditions, higher yielding cows contracted more mastitis than lower yielding cows [bib_ref] Clinical mastitis in cows treated with somatribove (recombinant bovine somatotropin) and its..., White [/bib_ref]. Many studies have reported that milk yield and mastitis are genetically correlated [bib_ref] Genetic correlations between production and disease traits during first lactation in Holstein..., Hagiya [/bib_ref] [bib_ref] Genetic parameters for production, health, fertility and longevity traits in dairy cows, Pritchard [/bib_ref] [bib_ref] Genetic analyses of mastitis data using animal threshold models and genetic correlation..., Hinrichs [/bib_ref] [bib_ref] Genetic parameters for clinical mastitis, somatic cell score, and production in the..., Carlén [/bib_ref] [bib_ref] Linear and threshold model genetic parameters for disease, fertility and milk production..., Kadarmideen [/bib_ref] [bib_ref] Genetic parameters of udder, mastitis and milk traits in two different climactic..., Amin [/bib_ref].
Further, high genetic merit dairy cows require high energy feeds to meet the substantial energy demands of early lactation. They often receive a high proportion of cereal grains, and one model found that cows selected for greater milk fat and protein fed a 50% grain diet had lower emissions (1.05 kg CO 2 -eq. per kilogram of energy-corrected milk), than cows with a 25% grain diet and summer grazing (1.19 kg per kilogram of energy-corrected milk) [bib_ref] The effect of improving cow productivity, fertility, and longevity on the global..., Bell [/bib_ref]. However, dairy cows are prone to digestive disorders such as sub-acute ruminal acidosis (SARA) [bib_ref] Diet-induced bacterial immunogens in the gastrointestinal tract of dairy cows: Impacts on..., Dong [/bib_ref] [bib_ref] Subacute ruminal acidosis (SARA): A review, Kleen [/bib_ref]. Diets that are rich in grains alter rumen metabolic balance and can induce an explosive proliferation of Streptococcus bovis and Lactobacillus bacteria species. [bib_ref] A metabolomics approach to uncover the effects of grain diets on rumen..., Saleem [/bib_ref]. The resulting rumen acidosis eventually leads to erosion and ulceration of the ruminal epithelium and chronic symptoms including lethargy, reduced feed intake, diarrhea, and discomfort among others.
## Effects of breeding for productivity on swine welfare
Increases in the productivity of swine, largely achieved through genetic selection, have been cited as climate change mitigation progress(p. [bib_ref] Breeding for better welfare: Genetic goals for broiler chickens and their parents, Dawkins [/bib_ref] [bib_ref] Lowering ruminant methane emissions through improved feed conversion efficiency, Waghorn [/bib_ref]. For example, a 2013 FAO report cited data provided by the US National Pork Board. Between 1959 and 2009, there was a 29% increase in the number of hogs marketed in the United States, yet the size of the breeding herd has been reduced by 39%. This is due in part to the steady increase in litter size from 7.10 in 1974 to 9.97 piglets in 2011 and the concomitant change in the amount of pork produced from a single breeding animal from 775 to 1828 kg. Feed efficiency also increased by 33%. This has reduced the carbon footprint per 454 kg of hot dressed carcass weight produced by 35% [bib_ref] Mitigation of methane and nitrous oxide emissions from animal operations: III. A..., Hristov [/bib_ref]. However, the original source for this calculation [bib_ref] A 50-Year Comparison of the Carbon Footprint of the US Swine Herd, Boyd [/bib_ref] does not appear to be peer-reviewed, and total emissions in the US pork industry have increased overall with greater pork production. Despite this, the suggestion is that the type of throughput achieved in the United States, if adopted more widely in the developing world, could be an approach to climate change mitigation. However, some of the key unintended animal welfare consequences that may accompany such a change include low piglet viability and severe feed deprivation of breeding sows in order to prevent them from becoming obese.
Selectively breeding for increased litter size reduces the proportion of surviving piglets [bib_ref] Selection for litter size at day five to improve litter size at..., Su [/bib_ref] [bib_ref] Perinatal mortality in the pig: Environmental or physiological solutions?, Edwards [/bib_ref]. [bib_ref] Effect of selection for lean tissue growth on body composition and physiological..., Herpin [/bib_ref] postulated that genetic changes have altered body fat metabolism, body composition, and hormonal state, resulting in lean tissue growth that makes piglets heavier but less mature at birth compared to lines that have not been so intensely selected for production traits. This reduces piglet survival rate [bib_ref] Effect of selection for lean tissue growth on body composition and physiological..., Herpin [/bib_ref] , which is an intractable problem in the pork industry. Selection for leanness may have also inadvertently decreased the nutritional quality of sows' milk [bib_ref] Breeding pigs for improved welfare, Rydhmer [/bib_ref] , in turn further affecting survival of the piglets.
Selectively bred for productivity, growing/finishing pigs in the United States are fed to satiety to meet their full genetic potential, but breeding animals must often be feed restricted to prevent excessive weight gain and fat deposition. Confined sows are typically fed just half the amount they would choose to consume if fed ad libitum. This leads to a chronic hunger problem including persistent, unfulfilled feeding motivation [bib_ref] Stereotypic oral behaviour in captive ungulates: Foraging, diet and gastrointestinal function, Bergeron [/bib_ref] [bib_ref] A comparison of operant responding by farrowing sows for food and nest-building..., Hutson [/bib_ref] [bib_ref] Methodology for measuring hunger and food needs using operant conditioning in the..., Lawrence [/bib_ref]. Short feeding times and lack of an outlet for natural foraging behavior are underlying causes of abnormal stereotypic behavior, discussed below in Section 2.4.1.
## Effects of breeding for productivity on the welfare of chickens
As with swine, genetically linked productivity improvements in poultry have also been tied to climate change benefits through feed conversion efficiencyand the extremely rapid growth rate of modern broiler chickens (raised for meat). Broilers now grow from hatch to market weight in five weeks or less. Annual growth rates have increased 3.3% over the past 50 years, and broiler chickens now grow more than four times faster than they did in 1957. Using a life-cycle assessment model, one study estimated that genetic changes in average U.K. broiler chickens have achieved a 1% reduction per year in greenhouse gases over the last 20 years [bib_ref] The effect of genetic improvement on emission from livestock systems, Jones [/bib_ref].
These animals now grow so fast that muscle outpaces bone development, leading to metabolic bone disease. As a result, broiler chickens often suffer from leg deformities and lameness [bib_ref] Welfare of broilers: A review, Bessei [/bib_ref] [bib_ref] Monitoring leg problems in broilers: A survey of commercial broiler production in..., Sanotra [/bib_ref] [bib_ref] Production and growth related disorders and other metabolic diseases of poultry-A review, Julian [/bib_ref] , including tibial dyschondroplasia, angular bone deformities, spondylolisthesis, epiphyseitis, and in severe cases, rupture of the gastrocnemius tendon. Studies are consistent in showing that approximately 26%-30% of broiler chickens by 40-42 days of age suffer from gait defects severe enough to impair walking ability [bib_ref] Production and growth related disorders and other metabolic diseases of poultry-A review, Julian [/bib_ref] [bib_ref] Leg disorders in broiler chickens: Prevalence, risk factors and prevention, Knowles [/bib_ref] [bib_ref] Prevalence of leg weakness in broiler chickens and its relationship with genotype, Kestin [/bib_ref]. Additional research strongly suggests that while conformational differences account for some gait differences [bib_ref] An examination of anatomic, physiologic, and metabolic factors associated with well-being of..., Skinner-Noble [/bib_ref] , birds at this level of lameness are probably in pain [bib_ref] Self-selection of the analgesic drug carprofen by lame broiler chickens, Danbury [/bib_ref] [bib_ref] Impact of lameness on broiler well-being, Nääs [/bib_ref] [bib_ref] Effect of carprofen on lameness in broiler chickens, Mcgeown [/bib_ref].
Not only do rapidly growing broiler chickens suffer from musculoskeletal diseases, but they are also afflicted by metabolic disorders. Ascites is a condition in which rapidly growing broiler chickens do not have the heart and lung capacity needed to distribute oxygen throughout the body and is a leading cause of mortality as the birds reach market weight [bib_ref] Managing rapid growth rate in broilers, Boersma [/bib_ref]. Sudden Death Syndrome (SDS) is associated with acute heart failure caused by dysrhythmias. Young birds die from SDS after sudden convulsions and wing-beating, and are frequently found lying on their backs [bib_ref] Evaluating the impact of metabolic disorders on the welfare of broilers, Julian [/bib_ref]. The condition has been recognized since the 1950s as more broiler chickens were grown in large numbers for commercial production [bib_ref] Production and growth related disorders and other metabolic diseases of poultry-A review, Julian [/bib_ref]. Sudden death syndrome (SDS) and ascites together can account for 50% of the mortality of highly productive broiler chicken strains [bib_ref] Metabolic disturbances in male broilers of different strains. 1: Performance, mortality, and..., Gonzales [/bib_ref] [bib_ref] Relationships between age, body weight, and season of the year and the..., Gardiner [/bib_ref].
The parent generation of broiler chickens grown for meat shares the same genetics as their progeny. Like breeding sows, these broiler breeders must be feed restricted, or they will suffer from health and reproductive disorders due to weight gain. Parent birds are usually feed-restricted starting when they are as young as one week old. In many parts of the world broiler breeders may be fed on a "skip-a-day" regimen, in which the animals are fed as infrequently as every other day-though this practice has been outlawed in several European countries. Experimental studies suggest that artificial selection for increased body weight may have altered the brain mechanism controlling satiety and appetite [bib_ref] Genetic selection for growth rate alters hypothalamic satiety mechanisms in chickens, Burkhart [/bib_ref] , and evidence from behavioral studies suggests that feed restriction interferes with learning [bib_ref] Too hungry to learn? Hungry broiler breeders fail to learn a Y-maze..., Buckley [/bib_ref] and causes stress [bib_ref] Relationships between the degree of food restriction and welfare indices in broiler..., Hocking [/bib_ref] [bib_ref] Major welfare issues in broiler breeders, Jong [/bib_ref] , boredom, and chronic hunger [bib_ref] Assessment of hunger in growing broiler breeders in relation to a commercial..., Savory [/bib_ref] [bib_ref] Broiler breeders: Feed restriction and welfare, Mench [/bib_ref] [bib_ref] Feed restriction and welfare in domestic birds, Jong [/bib_ref]. Breeding birds receive only 25%-50% of the amount of feed they would otherwise eat if given free access [bib_ref] Assessment of hunger in growing broiler breeders in relation to a commercial..., Savory [/bib_ref]. While free-range chickens normally devote about 50% of their daily time budget to foraging [bib_ref] Feeding behaviour in a population of domestic fowls in the wild, Savory [/bib_ref] [bib_ref] Time budgets in red junglefowl as a baseline for the assessment of..., Dawkins [/bib_ref] , feed restricted parent birds can consume their daily feed allotment in as little as 15 min. Feed restriction is believed to cause undernourishment, nutritional deficiency, and abnormal behavior including increased pecking at non-feed objects, pacing, and heightened aggression [bib_ref] Influence of degree of food restriction, age and time of day on..., Savory [/bib_ref]. Given that target body weights for broiler breeders have changed little in the past 30 years, but broiler body weight continues to increase, the welfare of parent birds may become more serious in the future as they will be feed deprived to an even greater extent [bib_ref] Feed restriction, Hocking [/bib_ref].
## Genetic selection with regard to welfare
While genetic selection for productivity can have negative side-effects for animal welfare, this is not necessarily always the case. Although progress may be slow, it is possible to select for both welfare associated traits and productivity simultaneously. For example, while there is an overall negative correlation between walking ability and growth rate of broiler chickens, some individuals do have both good gait scores and high productivity. Given enough time, it might be possible to successfully select for both traits [bib_ref] Breeding for better welfare: Genetic goals for broiler chickens and their parents, Dawkins [/bib_ref]. Where animals can be bred to produce the same amount of meat or milk, but with less feed, GHG emission intensity (CH 4 /unit animal product) is reduced. The most efficient animals produce less CH 4 and N 2 O than the population average [bib_ref] Lowering ruminant methane emissions through improved feed conversion efficiency, Waghorn [/bib_ref] , and there might not be a further welfare impact if selection focused on these animals. Direct selection on reduced CH 4 production could also be a possibility. For example, in one study of sheep the heritability of g CH 4 /day was 0.29 with no genetic correlation with live weight, and a slightly economically favorable correlation with fleece weight [bib_ref] Heritability estimates of methane emissions from sheep, Pinares-Patino [/bib_ref].
## Mismatching genetics and environment
According to the FAO, the single most effective GHG mitigation strategy in many regions of the world is to increase animal productivity(p. 100). Raising imported animals selected for production traits, instead of traditional breeds is part of that strategy. However, the FAO also cautions that while achieving the genetic potential of highly productive breeds is important, these animals should not be imported into environments where climate or management limitations would prevent the animals from achieving this potential. In new regions or countries, animals from genetic lines with high production potential may actually perform worse than native breeds or crossbreeds, especially compared to those well-adapted to local disease and climactic conditions(p. 104). Some regions have seasonal fluctuations in feed quality and nutrient management, and, as high genetic merit breeds have exacting nutritional requirements, they do not always thrive as expected.
In southern India, for example, efforts to replace the native Vechur cow with imported Holsteins have been disappointing, because Holsteins were not bred for performance under the local conditions. While the small Vechur cows produce only one-tenth of the milk hoped for from Holsteins, they are hardy, heat and humidity tolerant and resistant to local diseases and ticks. Under the these environmental conditions in India, Holsteins do not produce the same milk yield that they do in more temperate climates, and the costs of keeping the animals alive and healthy may be greater than any added productivity [bib_ref] Agriculture: Steps to sustainable livestock, Eisler [/bib_ref].
In Africa, resistance to trypanosomiasis is a factor that limits the geographical distribution of cattle breeds. Indigenous taurine (humpless) breeds of cattle including the N'Dama and West African Shorthorn are genetically trypanotolerant [bib_ref] Genetic resistance to African trypanosomiasis, Murray [/bib_ref] , but European taurine breeds have no resistance, and even Zebu cattle imported from other regions of the continent do not thrive in the humid tropics of West Africa occupied by tsetse flies, vectors of trypanosomiasis [bib_ref] Agriculture: Steps to sustainable livestock, Eisler [/bib_ref]. Clearly there are limits to importing more highly productive breeds in this large swath of the world, and such circumstances call in to question the notion of "genetic improvement"-indeed the improved breeds here would appear to be the ones that can best survive the local conditions.
In Brazil, more fertile and adapted breeds, such as the Brazilian Gir, are increasing in popularity [bib_ref] How sustainable are the breeding programs of the global main stream dairy..., Madalena [/bib_ref]. Indigenous breeds are better adapted to local conditions, and more readily able to avoid predators, tolerate heat or other weather extremes, resist parasites and endemic diseases, and scavenge for feed or consume local feed resources. For example, in the subtropical climate zone, with seasonal rainfall patterns that do not permit continuous vegetation growth, Zebu and Sanga cattle are more resistant to ticks and tick-transmitted diseases as well as feed scarcity and seasonal high temperatures [bib_ref] Invited review: Are adaptions present to support dairy cattle productivity in warm..., Berman [/bib_ref].
While promoting highly productive genetics to address climate change will reduce GHG emissions in some ways, maintaining biodiversity and the genetic resources of locally adapted breeds will be needed to complement other climate change mitigation strategies and to ensure the animals thrive. A potential solution may be more research and genetic selection of local breeds, but only with due consideration to welfare. Cradle-to-grave analyses that include the full range of trade-offs will become increasingly necessary [bib_ref] Livestock production: Recent trends, future prospects, Thornton [/bib_ref]. Genetic suitability for survival in less hospitable regions with more variable climates and diverse, changing feed resources is currently, and may become even more so, an essential requirement for both human and animal welfare.
## Growth and yield promoting biotechnologies
Some authors are calling attention to the wide application of growth promoting technologies, including β-adrenergic agonists (βAAs) and recombinant bovine somatotrophin (rbST), for their climate change mitigation potential [bib_ref] The environmental impact of recombinant bovine somatotropin (rbST) use in dairy production, Capper [/bib_ref] [bib_ref] Growth-promoting technologies decrease the carbon footprint, ammonia emissions, and costs of california..., Stackhouse [/bib_ref]. While the research to date has focused on US production systems, the call is being made in the context of meeting increasing global demand for animal protein. Zilpaterol hydorchloride and ractopamine hydrochloride are βAAs that are used as feed additives in grain rations for cattle [bib_ref] Increased mortality in groups of cattle administered the b-adrenergic agonists ractopamine hydrochloride..., Loneragan [/bib_ref]. βAAs are fed in a 20-40 day period prior to slaughter to improve feed efficiency and weight gain [bib_ref] A meta-analysis of zilpaterol and ractopamine effects on feedlot performance, carcass traits..., Lean [/bib_ref]. The growth hormone rbST is used in the dairy industry to increase milk yield and feed conversion efficiency [bib_ref] The environmental impact of recombinant bovine somatotropin (rbST) use in dairy production, Capper [/bib_ref]. However, both rbST and βAAs are cause for substantial animal welfare concern [bib_ref] Increased mortality in groups of cattle administered the b-adrenergic agonists ractopamine hydrochloride..., Loneragan [/bib_ref] [bib_ref] A meta-analysis review of the effects of recombinant bovine somatotropin: 1. Methodology..., Dohoo [/bib_ref] ].
## Β-adrenergic agonists
Using Life Cycle Analysis (LCA) modeling, [bib_ref] Growth-promoting technologies decrease the carbon footprint, ammonia emissions, and costs of california..., Stackhouse [/bib_ref] found that the use of βAAs, in combination with antibiotics, ionophores, and hormone implants, as is common in US feedlots, decreased the carbon footprint of Angus beef production by 9% [bib_ref] Growth-promoting technologies decrease the carbon footprint, ammonia emissions, and costs of california..., Stackhouse [/bib_ref] , and therefore advocated their use as a potential mitigation option. Cooprider et al. (2011) calculated that feedlot cattle fed ractopamine hydrochloride along with hormone implants and antibiotics had a 31% decrease in emissions per finished steer given the improved feed efficiency [bib_ref] Feedlot efficiency implications on greenhouse gas emissions and sustainability, Cooprider [/bib_ref]. However, anecdotal reports of changes in behavior and more cattle deaths when βAAs were being fed has led researchers to investigate [bib_ref] Growth-promoting technologies decrease the carbon footprint, ammonia emissions, and costs of california..., Stackhouse [/bib_ref]. One recent study found that cattle fed βAAs along with a growth implant (trenbolone acetate and estradiol) engaged in more agonistic behavior including bulling and pushing [bib_ref] Effects of growth-promoting technology on feedlot cattlebehavior in the 21 days before..., Stackhouse-Lawson [/bib_ref]. A large-scale investigation by two major US agricultural universities examined datasets provided by feedlot operators also confirmed the field observations, finding that the cumulative death rate is as much as 90% greater in animals administered a βAA, and that 40%-50% of deaths among groups administered a βAA could be explained by administration of the drug [bib_ref] Increased mortality in groups of cattle administered the b-adrenergic agonists ractopamine hydrochloride..., Loneragan [/bib_ref]. Imports of animal products containing beta agonists is banned in many countries, including the EU, Russia, and China.
The βAA ractopamine hydrochloride is also fed to swine, where the side-effects that have been elucidated to date include elevated heart rate, difficulty walking [bib_ref] The effects of ractopamine on the behavior and physiology of finishing pigs, Marchant-Forde [/bib_ref] , increased aggression [bib_ref] Effects of ractopamine feeding, gender and social rank on aggressiveness and monoamine..., Poletto [/bib_ref] , abnormal behavior [bib_ref] Behavioral effects of "step-up" ractopamine feeding program on finishing pigs, Poletto [/bib_ref] , and hoof lesions [bib_ref] Effects of a "step-up" ractopamine feeding program, sex, and social rank on..., Poletto [/bib_ref]. Observations at slaughter plants additionally find that difficulty walking due to ractopamine may contribute to a greater incidence of non-ambulatory (or "downed") pigs, those too weak to stand and walk on their own accord [bib_ref] The effects of ractopamine on the behavior and physiology of finishing pigs, Marchant-Forde [/bib_ref]. The number of studies investigating the animal welfare effects of βAAs is, thus far, small, but given the problems identified already, calls for their increased use clearly should be made more cautiously.
## Recombinant bovine somatotrophin
Also called Bovine Growth Hormone, rbST is a protein hormone that redirects nutrient partitioning during lactation and can increase milk yield by 8%-36% [bib_ref] Somatotropin treatment for a second consecutive lactation, Annexstad [/bib_ref]. On a per unit of milk produced basis, it has been show that rbST reduces the maintenance energy and protein requirements of cows by 11.8% and 7.5%, respectively, and total feed requirements by 8.1% [bib_ref] Somatotropin treatment for a second consecutive lactation, Annexstad [/bib_ref]. It also reduces N excretion on a per liter of milk produced basis. Although overall N excretion increases in rbST treated cows, because they consume more feed, each cow produces more milk, so estimates calculated on a per liter basis report a reduction in nitrogen excretion between approximately 9% [bib_ref] The environmental impact of recombinant bovine somatotropin (rbST) use in dairy production, Capper [/bib_ref] and 15% [bib_ref] Animal production and european pollution problems, Williams [/bib_ref]. Per unit of milk, the dilution of maintenance conferred by the use of rbST can result in a reduction in manure production by 6.8% and CH 4 output by 7.3% [bib_ref] The environmental impact of recombinant bovine somatotropin (rbST) use in dairy production, Capper [/bib_ref].
While rbST is banned in the European Union, Canada, Japan, Australia, and New Zealand, the drug is approved in the United States, despite substantial controversy. One of the primary welfare concerns associated with the increased milk production caused by rbST is the potential risk for mastitis. Some of the initial studies on rbST indicated that its use was linked to an increase (for example, and , but see for counter examples).
A 1999 report of the Scientific Committee on Animal Health and Animal Welfare (SCAHAW), a working party of independent scientists appointed by the European Commission to assess rbST, reviewed approximately two dozen studies and concluded that rbST usage increases the risk of clinical mastitis and the length of infections. This finding was supported by a review published by the Canadian Veterinary Medical Association (CVMA) [bib_ref] A meta-analysis review of the effects of recombinant bovine somatotropin: 1. Methodology..., Dohoo [/bib_ref]. The CVMA performed a series of meta-analyses to combine data from all randomized clinical trials that had been published in peer-reviewed journals or which were provided by Health Canada for the registration of rbST in the country. The CVMA found that rbST increased the risk of clinical mastitis by approximately 25%, closely agreeing with the SCAHAW report. Subsequently, rbST was not approved for use in Canada, due to animal health concerns.
The debate between researchers in different countries largely centers on the question of whether increases in mastitis are due to rbST treatment per se or to the increase in productivity caused by rbST, the same as would be expected when genetic selection leads to greater milk production. [bib_ref] Clinical mastitis in cows treated with somatribove (recombinant bovine somatotropin) and its..., White [/bib_ref] argue that treatment differences in mastitis are within the range predicted using estimates from genetic studies. And using this logic they conclude that rbST treatment had no effect on mastitis [bib_ref] Clinical mastitis in cows treated with somatribove (recombinant bovine somatotropin) and its..., White [/bib_ref]. In contrast,argue that what matters is the result on the welfare of the animals, which is the same whether it is the drug itself or the increase in productivity that causes the mastitis.
In addition to problems with mastitis, there is also concern that rbST treatment may increase the risk of lameness and foot and leg disorders. While many early studies failed to find a detectable difference between rbST treated and control animals [bib_ref] Somatotropin treatment for a second consecutive lactation, Annexstad [/bib_ref] [bib_ref] Health and reproductive performance of dairy cows treated for up to two..., Burton [/bib_ref] [bib_ref] Use of prolonged-release bovine somatotropin for milk production in british friesian dairy..., Weller [/bib_ref] [bib_ref] Effects of recombinant methionyl bovine somatotropin (sometribove) in high producing cows milked..., Jordan [/bib_ref] [bib_ref] Effect of long-term administration of a prolonged release formulation of bovine somatotropin..., Wells [/bib_ref] [bib_ref] Responses of dairy cows supplemented with somatotropin during weeks 5 through 43..., Chalupa [/bib_ref] , others found a slight increase in the number of cows culled due to lameness [bib_ref] Effects of treatment of dairy cows with recombinant bovine somatortropin over three..., Oldenbroek [/bib_ref] and slightly more feet and leg problems [bib_ref] Effects of treatment of dairy cows with recombinant bovine somatortropin over three..., Oldenbroek [/bib_ref] [bib_ref] Lactation, health, and reproduction of dairy cows receiving daily injectable or sustained-release..., Zhao Burton [/bib_ref]. As with the early work on mastitis and rbST, in studies with small cow numbers, low statistical power may have prevented detection of effects, and slight increases would be hard to find.
However, both the SCAHAW committee and the CVMA review panel found an increased incidence of foot and leg disorders associated with the long-term administration of rbST [bib_ref] A meta-analysis review of the effects of recombinant bovine somatotropin: 1. Methodology..., Dohoo [/bib_ref]. The CVMA review panel found that cows treated with rbST were at an estimated 55% increase in the risk of developing clinical signs of lameness, and this is now reflected on the Posilac drug label, which states "Cows injected with Posilac may have more enlarged hocks and disorders of the foot region".
Recombinant bovine Somatotropin has also been shown to reduce the ability of cows to cope with high environmental temperatures, which can lead to heat stress [bib_ref] Interactions of heat stress and bovine somatotropin affecting physiology and immunology of..., Elvinger [/bib_ref]. It is, thus, not advisable in geographic regions where severe heat can be expected, especially if management options to alleviate heat stress are insufficient.
Clearly the use of rbST carries unintended risks to the welfare of animals, and these should be seriously considered before such growth-promoting compounds are touted as a possible mitigation strategy for climate change.
## Species shifts
In addition to increasing productivity within species, the current global trend toward intensively producing greater numbers of non-ruminant animals-pigs and chickens-is often viewed as a more efficient form of meat production compared to traditional ruminant cattle, sheep and goat rearing systems [bib_ref] Livestock production and the global environment: Consume less or produce better?, Steinfeld [/bib_ref] [bib_ref] Responses on environmental issues, Steinfeld [/bib_ref] [bib_ref] Biomass use, production, feed efficiencies, and greenhouse gas emissions from global livestock..., Herrero [/bib_ref]. To meet the growing demand for animal protein in emerging economies, the expansion in supply is largely coming from increased local production, as opposed to from imports, and in these regions, livestock systems are rapidly shifting toward monogastric species. Pigs and poultry account for 77% of the expansion in production, a trend that is often accompanied by intensification, vertical integration, geographic concentration, up-scaling of production, and feeding more grain-or concentrate-based diets-in short, industrialized farming.
Monogastrics have improved GHG efficiencies compared to cattle. Cattle accounted for 77% of total non-CO 2 GHG emissions from the livestock sector in 2000, while monogastrics accounted for just 10% [bib_ref] Biomass use, production, feed efficiencies, and greenhouse gas emissions from global livestock..., Herrero [/bib_ref]. It takes at least three times as much land to produce one kilogram of beef as it does to produce one kilogram of pork or chicken [bib_ref] Comparing environmental impacts for livestock products: A review of lifecycle assessments, De Vries [/bib_ref]. Cattle contribute about seven times the CO 2 -eq. emissions of pigs and chickens. This biological difference can be explained by three main factors: greater feed efficiency, lower enteric CH 4 emissions, and faster rates of reproduction. CH 4 emissions from ruminants originate from both manure and enteric fermentation, while CH 4 emissions from monogastrics are primarily associated with manure alone. Because sows and hens have a relatively large number of progeny and earlier sexual maturity, the number of breeding adults per animal slaughtered for human food is reduced in these production systems [bib_ref] Comparing environmental impacts for livestock products: A review of lifecycle assessments, De Vries [/bib_ref]. Together, these factors account for the reduction in inputs and lower emissions of pig and poultry production worldwide.
If the increasing consumption of animal protein in developing countries will be supplied in a substantial way by expansion of pork and chicken industries, and if the expansion will be in the form of specialized, intensive confinement facilities that typify the bulk of production in North America and other developed regions, then there is a serious potential for a concomitant increase in the animal welfare problems that characterize those systems.
As covered in Section 2.2.3, the intensive production of modern strains of poultry is the cause of one of the most severe and wide-spread animal welfare problems in the world. Globally, nearly 60 billion chickens are slaughtered for meat each year , and given the wide prevalence of metabolic problems and painful leg disorders, the intensive production of meat chickens already constitutes one of the most problematic aspects of intensive agriculture. Sweeping calls to use more of these compromised animals to meet efficiency goals is alarming.
In an ethical analysis, overall welfare would be a function of both the state of the individual animal and the number of animals subjected to that level of welfare. It takes far more animals to produce the same amount of meat when beef is replaced in the diet with chicken, compounding the overall magnitude of any negative animal welfare impacts.
Another of the most salient issues with intensive confinement of animals in agriculture is the way those animals are housed. Intensive systems for pork and poultry production remove animals from the land and instead raise them in indoor, total confinement facilities. While well-managed, large-scale production can have some animal welfare advantages, such as access to resources that might be economically out of reach to smaller producers, they often greatly limit the space allotment per animal. In the case of breeding sows in the pork production industry the animals are further partitioned into enclosures that prevent normal movement and severely restrict the animals' natural behavior, which is a serious welfare problem.
## Industrial production of pigs
Swine are highly intelligent, social animals with complex behavioral repertoires. Yet, in many countries, intensive production facilities house female breeding pigs in narrow gestation crates, metal stalls that measure approximately 0.61 m by 2.31 m, which is only slightly larger than the sow's own body. Each sow can take a step forward and backward, but she cannot turn around for the length of her pregnancy, which is approximately 114 days. Due to decreased fitness from chronic lack of exercise, crated sows have lower bone strength and muscle weight [bib_ref] Effects of dry sow housing conditions on muscle weight and bone strength, Marchant [/bib_ref] , and have higher resting heart rates [bib_ref] The effects of housing on heart rate of gestating sows during specific..., Marchant [/bib_ref] , compared to those permitted more freedom of movement and exercise.
Fed concentrated grain-based diets that can be consumed quickly, and without opportunity to express species-typical foraging behavior, sows in intensive confinement operations often begin to display abnormal, stereotypic behavior including repetitive bar-biting, head-weaving, sham-chewing and drinker pressing. Rooted in stress and lack of stimulation [bib_ref] A review of the welfare issues for sows and piglets in relation..., Barnett [/bib_ref] [bib_ref] Tethered versus loose sows: Ethological observations and measures of productivity. I. Ethological..., Vestergaard [/bib_ref] , these "stereotypies", are common in captive animals kept in barren, restrictive environments [bib_ref] Can't stop, won't stop: Is stereotypy a reliable animal welfare indicator?, Mason [/bib_ref]. Stereotypic behavior is often reduced when sows are fed a high fiber, bulky diet instead of, or in addition to, concentrates [bib_ref] High-fiber diets in pregnant sows: Digestive utilization and effects on the behavior..., Ramonet [/bib_ref] [bib_ref] Feeding motivation and stereotypies in pregnant sows fed increasing levels of fibre..., Bergeron [/bib_ref] [bib_ref] Effect of dietary fibre and feeding system on activity and oral behaviour..., Brouns [/bib_ref] [bib_ref] Does the number of daily meals affect feeding motivation and behaviour of..., Robert [/bib_ref] [bib_ref] Influence of the nature of dietary fibre on digestive utilization, some metabolite..., Ramonet [/bib_ref] , although some fiber sources, such as soybean hulls, may be less effective [bib_ref] Effects of a high-fiber diet and frequent feeding on behavior, reproductive performance,..., Holt [/bib_ref]. Ethologists attribute this stereotypic behavior to boredom and frustration resulting from an impoverished environment, confinement, restraint, and unfulfilled behavioral needs [bib_ref] The Effects of Alternative Forms of Intensive Pig Husbandry on Measures of..., Mendl [/bib_ref].
Most large, industrialized confinement facilities for pigs utilize a liquid manure handling system with slatted floors. While a 2013 FAO review noted that slatted floor manure systems tend to decrease GHG and NH 3 emissions compared with deep litter systems, these bare, concrete floors are hard on the animals' feet. One survey found that pigs allowed outdoor access had a lower prevalence of foot and limb injuries, while those confined indoors on hard, slatted flooring had more bruising, calluses, locomotion problems, and adventitious bursae. The problem is especially acute for piglets; foot and limb injuries affect 60% to 90% of young piglets housed indoors, while only about 9% of piglets housed outdoors suffer these types of injuries. While outdoor production does have its own suite of welfare concerns, including severe weather and predators, in farrowing huts common to pasture-based systems, soil covered with straw provides a soft, non-abrasive, protective surface. Behavioral testing has shown that pigs prefer earthen floors over concrete [bib_ref] Operant preference tests with pigs, Van Rooijen [/bib_ref] [bib_ref] Preference testing of substrates by growing pigs, Beattie [/bib_ref]. While bedding, such as straw provides comfort, warmth and outlets for natural rooting and nesting behavior, it is not usually provided in indoor operations due to cost, difficulty of cleaning, and incompatibility with slatted floors.
There are however, more welfare-friendly intensive systems. For example, sows can successfully be housed in group pens as opposed to individual gestation crates, as long as they are managed well to prevent aggression. Group housing systems work best if they deliver the feed in a non-competitive way, such as by using trickle feeding, feeding stalls, or Electronic Sow Feeding systems (which use a computer to precisely deliver each sow's daily allotment) [bib_ref] Group housing of sows: The potential for the future, Gonyou [/bib_ref]. Straw bedding can be incorporated into these systems in purpose-built un-slatted sections of the barn. However, N 2 O emissions in a straw-based, deep litter group housing systems are usually greater compared to concrete slatted floors [bib_ref] Ammonia and greenhouse gas emission from group-housed gestating sows depends on floor..., Philippe [/bib_ref]. Further research on group housing systems design could focus on lowering emissions while simultaneously fulfilling the behavioral needs of the animals.
## Further disadvantages of shifting production systems
Another problem with large, industrialized animal production sites is the ratio of animals to human caretakers. With increased farm size, new technologies and increased mechanization, such as automated feed and water delivery and manure removal, coupled with economic pressure to reduce labor, the result is that fewer workers now tend to more animals. As such, individualized attention to each animal is generally lacking. With the use of efficiencies in pen and barn design, one person may now be responsible for the care of 8000 pigs per day(p. 291). One of the main daily tasks on a commercial-scale broiler grow-out facility is picking up the dead birds. While large farms can be well managed to reduce death losses, the health of the animals on large operations is often addressed at the herd or flock level, and individual veterinary care is not generally provided. An individual animal who becomes injured or ill is likely to go unnoticed, and could experience prolonged pain and suffering before their eventual death.
On top of the animal welfare concerns, the industrialization of agriculture carries with it numerous other societal implications. Switching to large-scale use of monogastrics threatens genetic diversity, and subsequent genetic adaptation of domestic species to climate change [bib_ref] Climate change and the characterization, breeding and conservation of animal genetic resources, Hoffmann [/bib_ref]. It could also mean the proliferation of more and more serious environmental issues, including air, soil, and water pollution. For example, the rapid expansion of pig production in the coastal areas of China, Vietnam, and Thailand is thought to be a major source of nutrient pollution in the South China Sea. Runoff is causing red tides and threatening fragile mangroves and coral reefs(p. 71). Concerns such as these must be properly weighed in the discussion of intensification and shifts to monogastric production as proposed climate change mitigation strategies.
System changes, however, (e.g., from extensive to industrialized production) are not required for some of the most substantial emissions reductions, which can be achieved within production systems, across all species and regions. The FAO found that applying the techniques of 10% of the most efficient producers to current systems, by for example, improved pasture management and feed quality, and better animal health, among others could cut nearly a third of emissions from the farm animal sector. Production system changes further reduced emissions a mere 2%(pp. [bib_ref] Mitigation of methane and nitrous oxide emissions from animal operations: III. A..., Hristov [/bib_ref] [bib_ref] Undesirable side effects of selection for high production efficiency in farm animals:..., Rauw [/bib_ref]. Clearly, there is a large opportunity to improve emissions by picking the low-hanging fruit, i.e., the least efficient producers [bib_ref] Productivity gains and greenhouse gas emissions intensity in dairy systems, Gerber [/bib_ref] , while achieving multiple benefits, such as animal welfare, or at least avoiding additional negative impacts. Such improvements should also aim at improved food security (not just increased production) and fit well within the concept of "sustainable intensification" [bib_ref] Sustainable intensification in agriculture: Premises and policies, Garnett [/bib_ref].
## Potential solutions: mitigation strategies that both improve animal welfare and reduce environmental impacts
While it is clear that animal welfare concerns for many different climate change mitigation strategies exist, fortunately a great many more potential solutions have benign or even beneficial welfare outcomes. Many of these options are strategies for extensive production, systems that, if improved, have very high welfare potential. Basic improvements to these systems would also have the co-benefit of improving the livelihoods of smallholder farmers. With more attention paid to the health, physiology, and behavioral effects on animals at the outset, triple-win strategies can be championed that mitigate climate change, improve food security, and improve animal welfare.
## Improved animal health and longevity
Where climate change solutions and improved animal welfare may dovetail best begins with a focus on the animals. Improving the efficiency of animal production systems by improving animal health and reducing mortality could reduce CH 4 and N 2 O emissions considerably by reducing losses due to health related production declines and emissions attributed to animals that die before they can reproduce or produce consumable products. In locations where basic veterinary care, vaccinations, clean drinking water, shade, husbandry skills, technical knowledge, financial aid and other vital resources are lacking, there is great potential for improvement. For example, in many developing African and Asian countries, approximately 80% of poultry is still located in village flocks [bib_ref] The Relative Contribution of Indigenous Chicken Breeds to Poultry Meat and Egg..., Pym [/bib_ref]. Globally, backyard flocks kept for meat and eggs contribute 42.5 million tons of CO 2 -eq. in greenhouse gas emissions and have higher emission intensity than intensively confined chickens, due in part to the low feed quality of scavenging birds and the toll that disease and predation take on productivity. Vaccinations, providing simple, low-cost protection for newly hatched chicks from predators, supplementary feeding of balanced rations, and basic biosecurity education [bib_ref] Technology and programmes for sustainable improvement of village poultry production. In Alternative..., Besbes [/bib_ref] could simultaneously improve the welfare of the flocks, alleviate poverty, and potentially reduce greenhouse gas emissions. The climate change mitigation potential of these improvements has not specifically been studied, to our knowledge.
Extending dairy cow lifetime is another potential improvement. Cows can easily live to 15 years of age or longer, but on most intensive, indoor dairy production facilities that typify developed countries the lifespan of a cow is typically closer to six years [bib_ref] The Importance of Improving Cow Longevity, Rushen [/bib_ref]. A USDA summary reports the proportion of dairy cows permanently removed from the herd for reproductive problems was 26.3%; 23.0% were removed for udder or mastitis problems, 16.0% for lameness or injury, and 16.1% for poor production. Similar factors for culling are reported for Canada and European countries [bib_ref] The Importance of Improving Cow Longevity, Rushen [/bib_ref]. In one study of 14 Irish dairy herds, the rate at which cows were replaced in the herd increased from 16% in 1990 to 27% in 2003, an increase of 0.8% per year, and the proportion of cows culled for infertility increased 2.2% per year, from 27% in 1990 to 40% in 2002 [bib_ref] Financial implications of recent declines in reproduction and survival of Holstein-friesian cows..., Evans [/bib_ref]. Improved longevity would reduce the total lifetime emissions of dairy cows when accounting for the resources needed for rearing replacement heifers. The proportion of total methane emissions produced by replacement heifers has been estimated to be up to 27% for U.K. dairy herds [bib_ref] The environmental impact of fertility in dairy cows: A modelling approach to..., Garnsworthy [/bib_ref]. Relative to the period from birth to first calving (a period that usually lasts from 25 months to about 3 years, depending on the region), a greater number of lactation periods per lifetime will reduce overall CH 4 loss per unit of milk yield [bib_ref] Greenhouse gas abatement strategies for animal husbandry, Monteny [/bib_ref]. By improving the lifespan of a dairy cow from 3.02 to 3.5 lactations, methane emissions would be reduced by 3% [bib_ref] Developing breeding schemes to assist mitigation, Wall [/bib_ref]. Selection for increased productive life could also lead to a decrease in disease incidence of dairy cows [bib_ref] Genetic correlations among protein yield, productive live, and type traits from the..., Rogers [/bib_ref].
In the future, one possibility for addressing the lack of individual attention on industrialized farms could be "Precision Livestock farming" (PLF). In PLF, sensors continuously monitor key indicators, such as animal movement, feed intake, or temperature and alert caretakers when a problem arises. Examples include using cameras to monitor the behavior of hens, pedometers to monitor estrus behavior of dairy cows, and automated scales for measuring weight gain of chickens [bib_ref] Is precision livestock farming an engineer's daydream or nightmare, an animal's friend..., Wathes [/bib_ref]. PLF has good potential for automated detection of animal health problems. For example, auditory sensors can be used to detect disease in fattening pigs by analysis of coughing sounds, which change qualitatively when the respiratory system is infected [bib_ref] Automatic On-line Monitoring of Animals by Precision Livestock Farming, Berckmans [/bib_ref] [bib_ref] Recognition system for pig cough based on probabilistic neural networks, Chedad [/bib_ref]. However, it is uncertain how applicable these highly technical solutions would be in developing countries, and there is a danger of losing genuine husbandry skills and knowledge when relying even more heavily on automation.
## Improved animal nutrition
As discussed previously, changes to animal diets can reduce greenhouse gas emissions and increase animal productivity, but they can also risk harming animal welfare. However, bolstering nutrition by supplementing poor diets or using improved forages can both reduce greenhouse gas emissions and improve welfare at the same time. In regions of the world with a seasonally dry tropical climate, such as Africa, the low nutritional value of most animal feeds during the dry season is a major constraint on animal productivity [bib_ref] The Role of Legumes in Farming Systems of Sub-Saharan Africa. Potentials of..., Tothill [/bib_ref]. In developing countries, ruminant animals are often fed on low quality grasses and crop residues (stover). Inadequate nutrition is a key factor limiting fertility. A large potential to mitigate emissions exists in these low-yield systems. For example, modeling of small ruminant production in West Africa suggests that improvements in forage digestibility, along with other animal health and husbandry measures, can potentially reduce emissions by 27% to 41% of baseline emissions, amounting to 7.7 to 12 million tons of CO 2 -eq..
Improved forages can be used as a climate change mitigation strategy in both developing and developed regions. Legumes available for grazing ruminants in the United States include clover, alfalfa, crown vetch, and birdsfoot trefoil (lotus). These nutritionally valuable, high protein forages can be incorporated into established pasture or rangeland with interseeding. Not only can improved pastures reduce CH 4 emissions from animals, but carbon sequestration in grassland soils can also be increased by planting nitrogen-fixing legumes [bib_ref] Grassland management and conversion into grassland: Effects on soil carbon, Conant [/bib_ref] [bib_ref] Response of soil carbon and nitrogen to transplanted alfalfa in North Dakota..., Liebig [/bib_ref].
The mitigation potential of feeding legumes has been the subject of a substantial number of studies. Simulation results in one study showed that CH 4 production is 28% lower with legume forage (alfalfa hay) as compared to grass (timothy hay) when expressed as a proportion of gross energy (GE) intake, and 21% lower when expressed relative to digestible energy (DE) [bib_ref] Evaluation of dietary strategies to reduce methane production in ruminants: A modelling..., Benchaar [/bib_ref]. These results are corroborated by multiple field studies, which also find lower CH 4 with pasture legumes for sheep [bib_ref] Mathanogenesis from forages fed to sheep, Waghorn [/bib_ref] and cattle [bib_ref] Impact of pasture type on methane production by lactating beef cows. Can, Mccaughey [/bib_ref] [bib_ref] Methane emissions by dairy cows fed increasing proportions of white clover (Trifolium..., Lee [/bib_ref]. For example, [bib_ref] Mathanogenesis from forages fed to sheep, Waghorn [/bib_ref] found that when fed to sheep, a sulla/lucerne feed mixture produced just 19.0 g CH 4 per kg DMI compared to fresh pasture, which produced 25.7 g CH 4 per kg DMI. In a second trial, lotus produced just 11.5 g CH 4 per kg DMI, and by using polyethylene glycol binding to remove condensed tannins (CT), the researchers demonstrated that CT played an important role in this reduction [bib_ref] Mathanogenesis from forages fed to sheep, Waghorn [/bib_ref]. Condensed tannins not only lower CH 4 production in sheep, but may also have other beneficial effects including improving wool growth and reducing intestinal worm burdens [bib_ref] The implications of condensed tannins on the nutritive value of temperate forages..., Barry [/bib_ref]. However, at least one study failed to find reduced methane production feeding white clover to sheep [bib_ref] Effects of feeding fresh white clover (Trifolium repens) or perennial ryegrass (Lolium..., Hammond [/bib_ref] [bib_ref] Effects of feeding fresh white clover (Trifolium repens) or perennial ryegrass (Lolium..., Hammond [/bib_ref] , and so more study is required.
One caution: legume consumption in some cases is associated with animal health issues, such as fetal hydrops in periparturient ewes [bib_ref] Theriogenology of sheep and goats, Edmondson [/bib_ref] , and bloat with some tropical legumes [bib_ref] Chemical constraints to the use of tropical legumes in animal nutrition, D'mello [/bib_ref]. Feeding legumes must, therefore, be carefully managed to avoid animal health problems.
Many other feeding management strategies that can reduce greenhouse gas emissions without compromising animal welfare could be better emphasized and further explored. These include improving silage quality with more effective preservation methods; adding dietary fats, such as coconut, linseed, soybean, or sunflower seed oils [bib_ref] Methane mitigation in ruminants: From microbe to the farm scale, Martin [/bib_ref] [bib_ref] The effect of varying levels of coconut oil on intake, digestibility and..., Jordan [/bib_ref] [bib_ref] Comparative evaluation of the effects of coconut oil, oilseeds and crystalline fat..., Machmüller [/bib_ref] ; changes to pasture management such as matching grazing intensity and pasture quality [bib_ref] Grassland management and conversion into grassland: Effects on soil carbon, Conant [/bib_ref] [bib_ref] Methane emissions of beef cattle on forages: Efficiency of grazing management systems, Deramus [/bib_ref] supplementation of the diet with enzymes [bib_ref] Animal production and european pollution problems, Williams [/bib_ref] [bib_ref] Use of exogenous fibrolytic enzymes to improve feed utilization by ruminants, Beauchemin [/bib_ref] or yeasts [bib_ref] In vitro utilization by a ruminal acetogenic bacterium cultivated alone or in..., Chaucheyras [/bib_ref] , and precision feeding (better agreement between the supply of N in the diet and requirements of the animal, in order to reduce N excretion) [bib_ref] Impact of nutrition on nitrogen, phosphorus, Cu and Zn in pig manure,..., Dourmad [/bib_ref] [bib_ref] Minimizing environmental impacts of livestock production using diet optimization models, Moraes [/bib_ref] , to name just a few.
## Improved manure and land management
In addition to improving animal health, nutrition or longevity, a great many more climate change mitigation options remain that could be used without negatively affecting the welfare of animals. Ruminants may be key parts to well-managed, mixed crop-livestock systems, helping to reduce a number of environmental impacts [bib_ref] Integrated crop-livestock systems: Strategies to achieve synergy between agricultural production and environmental..., Lemaire [/bib_ref]. Well-managed grazing can improve soil organic carbon and nitrogen content [bib_ref] Surface soil changes during twelve years of pasture management in the Southern..., Franzluebbers [/bib_ref] and partially offset net GHG emissions [bib_ref] Mitigating the greenhouse gas balance of ruminant production systems through carbon sequestration..., Soussana [/bib_ref]. Most manure management options are probably welfare-neutral, unless collection systems require the intensive confinement of animals. Hristov et al. (2013) provide a thorough summary of such options including reducing manure storage time, composting, and carefully timed land application. Several techniques for manure application also hold mitigation promise, for example subsurface application, application during non-rainy periods, and controlling soil pH. Perennial crops can be used to improve soil structure, water retention capacity and, in turn, carbon sequestration capacity [bib_ref] Perennial Crops for Food Security, Cox [/bib_ref]. Converting from conventional tilling to conservation or no-till cropping practices, especially with the simultaneous use of cover crops, can reduce soil disturbance and improve soil biodiversity and is another way to sequester carbon [bib_ref] Soil carbon sequestration to mitigate climate change, Lal [/bib_ref]. Other options include restoration of degraded range land (re-vegetation, water conservation) [bib_ref] Potential for reduced methane and carbon dioxide emissions from livestock and pasture..., Thornton [/bib_ref] , and silvopastoral and agroforestry systems [bib_ref] Potential for reduced methane and carbon dioxide emissions from livestock and pasture..., Thornton [/bib_ref] [bib_ref] Successes and failures with animal nutrition practices and technologies in developing countries:..., Owen [/bib_ref] among many others.
## Reduced animal numbers
The demand for meat and dairy products is rapidly growing. This demand is attributable to the growing world population, urbanization and increasing affluence in emerging economies [bib_ref] Worldwide transformation of diets, burdens of meat production and opportunities for novel..., Smil [/bib_ref]. The projected global growth in annual consumption of meat and milk is 5.41% and 3.43%, respectively. In Asia this growth is projected to be even greater, at 7.99% and 11.85% [bib_ref] Impact of increased demand for animal protein products in Asian countries: Implications..., Cao [/bib_ref]. Meat consumption is expected to continue to grow in developing countries until they meet the levels of developed countries, with poultry accounting for around 50% of the increase. If this demand is met, the growing number of animals will substantially increase the problem of climate change [bib_ref] Forecasting potential global environmental costs of livestock production 2000-2050, Pelletier [/bib_ref] , even if production efficiencies per unit of product produced are realized.
The growing demand for animal-based protein is usually presented as inevitable. While the general trajectory seems certain, the call to fill this demand in every case is not necessarily prudent. Reducing animal numbers could come, in part, with changes in production efficiency, as mentioned previously. However, these production-side efficiency approaches seem, in some ways, to avoid the heart of the problem.
Feeding animals grain to produce animal-based food is inherently inefficient. For every kg of chicken live weight gain, 1.7 kg is required in feed, and chicken is the most highly feed efficient species; the ratio of feed to gain for pigs is 2.4 to 1 and for cattle it is as high as 10 to 1 [bib_ref] Livestock-related greenhouse gas emissions: Impacts and options for policy makers, Garnett [/bib_ref]. The production of feed for animals can have a larger environmental impact in terms of energy use and global warming, than the management of those same animals and their manure on the farm [bib_ref] Environmental system analysis of pig production, Strid Eriksson [/bib_ref] [bib_ref] A carbon footprint analysis of egg production and processing supply chains in..., Pelletier [/bib_ref]. Using IPCC figures,estimates that the reductions achievable for agriculture mitigation options amount to only about 30% of the total impact; following this she argues: "[h]ence the conclusion that we need also to moderate our consumption of livestock products seems inescapable."
One approach to this is to address the immense amount of food that is wasted. This category is broad, including waste at harvest, storage, processing, and a number of post-consumer pathways [bib_ref] Food waste within food supply chains: Quantification and potential for change to..., Parfitt [/bib_ref]. The inefficiencies above suggest that waste of animal-based foods carries a particularly high environmental cost. The FAO has estimated global food waste at about 1/3, with waste of meat products of about 20% [bib_ref] Global Food Losses and Food Waste, Gustavsson [/bib_ref]. [bib_ref] How much land-based greenhouse gas mitigation can be achieved without compromising food..., Smith [/bib_ref] found savings of about 6% in food crop area, and 13%-47% GHG savings in 2050 through waste reduction compared to the reference case [bib_ref] How much land-based greenhouse gas mitigation can be achieved without compromising food..., Smith [/bib_ref].
Another approach focuses specifically on consumption, particularly in developed countries and in populations that face chronic disease due to overconsumption. Many studies show that reducing consumption of animal products would have a substantial effect on reducing GHG emissions, and the IPCC recognized this potential in their Fifth Assessment Report. [bib_ref] How much land-based greenhouse gas mitigation can be achieved without compromising food..., Smith [/bib_ref] found greater mitigation potential in the Agriculture, Forestry, and Other Land Use (AFOLU) sector on the consumption rather than production side, when incorporating associated benefits in spared land use (e.g., afforestation and bioenergy) [bib_ref] How much land-based greenhouse gas mitigation can be achieved without compromising food..., Smith [/bib_ref].
While variables such as energy used in transport, processing and refrigeration do not permit blanket statements about animal products over alternatives, meat, seafood, and some dairy products tend to have large GHG emissions per kg of product at the checkout as compared to most plant-based foods [bib_ref] Food and life cycle energy inputs: Consequences of diet and ways to..., Carlsson-Kanyama [/bib_ref]. Some analyses show that vegetarian and vegan diets have potential GHG savings of 22% and 29% in the UK and US, respectively [bib_ref] The relative greenhouse gas impacts of realistic dietary choices, Berners-Lee [/bib_ref] [bib_ref] Climate change mitigation and health effects of varied dietary patterns in real-life..., Soret [/bib_ref]. However, a 2014 study conducted in the UK found that dietary GHG emissions for meat-eaters were approximately twice as high as those of vegans [bib_ref] Dietary greenhouse gas emissions of meat-eaters, fish-eaters, vegetarians and vegans in the..., Scarborough [/bib_ref].
Vegan and vegetarian diets also have lower environmental impact including less land use, water [bib_ref] A global assessment of the water footprint of farm animal products, Mekonnen [/bib_ref] , and fossil fuel requirements [bib_ref] Quantification of the environmental impact of different dietary protein choices, Reijnders [/bib_ref]. Reducing climate change mitigation costs, as well as potential to improve public health [bib_ref] Food, livestock production, energy, climate change, and health, Mcmichael [/bib_ref] , have also been cited as a co-beneficial effects. Substantial effects could be realized even by simply reducing meat consumption, such as through efforts like the Meatless Monday campaign, or partial substitution of plant protein in ground and processed meats [bib_ref] Worldwide transformation of diets, burdens of meat production and opportunities for novel..., Smil [/bib_ref]. Another potential future alternative is cultured meat production. Cultured meat is laboratory produced using in vitro tissue engineering techniques, and initial LCA estimates show that it could lead to 78%-96% lower GHG emissions than conventionally produced European meat [bib_ref] Environmental impacts of cultured meat production, Tuomisto [/bib_ref] , while simultaneously avoiding all of the animal welfare impacts of intensive production.
# Conclusions
Proposals for climate change mitigation strategies often use the term, "sustainable intensification." Yet, societal norms related to the way we treat animals and use them in agriculture are changing and some past practices are no longer as widely accepted. Intensification of animal farming may not ever truly be sustainable, unless, among other things, there is concomitant attention to the health and behavioral needs of the animals, a meaningful effort to provide them with a life worth living. There is an urgent need to integrate these other sustainability measures into GHG mitigation assessments [bib_ref] Greenhouse gas mitigation in animal production: Towards an integrated life cycle sustainability..., De Boer [/bib_ref].
While multiple approaches are necessary to help stop anthropogenic climate change, those that offer the largest potential climate and sustainability benefits, while avoiding negative tradeoffs, should be the focus. This could mean different approaches (and levels of focus) for pre-market (production side) versus the market and beyond (demand side). For example, with production-side measures, there can be a rapid marginalization of further climate benefits beyond initial productivity gains [bib_ref] Productivity gains and greenhouse gas emissions intensity in dairy systems, Gerber [/bib_ref]. Demand-side approaches, on the other hand, seem to offer a largely untapped potential with multiple co-benefits.
More research is needed to explore the combination of animal welfare and environmental improvements. There are some promising examples in pigs [bib_ref] Energy and carbon inventory of iowa swine production facilities, Lammers [/bib_ref] , chickens [bib_ref] The effects of welfare-enhancing system changes on the environmental impacts of broiler..., Leinonen [/bib_ref] , and dairy cows [bib_ref] Comparing intensive, extensified and organic grassland farming in Southern Germany by process..., Haas [/bib_ref] [bib_ref] Greenhouse gas emissions from selected austrian dairy production systems-Model calculations considering the..., Hörtenhuber [/bib_ref]. These studies exhibit potential and the factors dictating these results, as well as ways to improve upon them, should be the focus of future research agendas.
Calls to globalize the production systems that were developed in the West are concerning when the impacts on the animals are taken into account, and there is worldwide momentum toward abandoning the worst practices. For example, multinational corporations such as McDonalds, Compass Group, and Nestle are asking producers for plans to eliminate gestation crates from their supply chains. Legislative and voluntary phase-outs of intensive animal confinement systems are increasingly taking hold across the globe, for example in the European Union, South Africa, Australia, Canada, New Zealand and several US states, with active campaigns in many additional regions of the world. The global movement toward farm animal welfare reform must be an important consideration going forward in the discussion of climate change mitigation strategies, and research and policy should increasingly reflect this.
© 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).
AcknowledgmentsWe would like to thank Mike Appleby for initial edits and comments.Author ContributionsShields and Orme-Evans researched and wrote the manuscript.Conflicts of InterestThe authors work in the Farm Animals section of Humane Society International.Animals 2015, 5 |
Predictive and Incremental Validity of Parental Representations During Pregnancy on Child Attachment
Parental pre-natal representations predict the interactive patterns that parents will put in place after childbirth. Early interactions defined by high parental emotional availability (EA) influence the development of security in children. To date, research on the predictive role of parental pre-natal representations on child attachment is still poor. Moreover, investigations on pre-natal representations have mainly focused on mothers. This study aimed at: investigating the criterion validity of the Interview of Maternal Representations During Pregnancy-Revised (IRMAG-R) and of the Interview of Paternal Representations During Pregnancy (IRPAG), using EA, parental attachment, and child attachment toward both parents, as criteria; testing the incremental validity of the IRMAG-R and IRPAG in the prediction of child attachment, controlling for other covariates, such as depressive and anxious levels during pregnancy, EA, and parental attachment; evaluating the possible mediation role of EA on the relationship between parental representations during pregnancy and child attachment. Fifty couples of primiparous parents were recruited during pregnancy, when the IRMAG-R and IRPAG were administered to mothers and fathers. At 6-9 months after childbirth, the mother-child and father-child interactions were coded by means of the EA Scales (EAS). At 14-18 after childbirth, the Adult Attachment Interview (AAI) was administered to parents, and the Strange Situation Procedure (SSP) was carried out to assess children's attachment toward mothers and fathers, respectively. The results showed significant correlations between parental pre-natal representations and EA, parental attachment and child attachment. As regards the prediction of child attachment, the IRMAG-R/IRPAG categories showed: a significant and large unique contribution for maternal representations; a close to be significant contribution for paternal representations (with a higher effect size for mothers than fathers). Moreover, while the indirect effect of pre-natal representations in the prediction of child attachment was not significant for mothers, it was instead significant for fathers. The results of this study confirmed the criterion validity of the IRMAG-R and IRPAG, and supported the incremental validity of the IRMAG-R and IRPAG in the prediction of children's attachment categories. Finally, the mediation models revealed that EA did not mediate the relationship between maternal pre-natal representations and child attachment, while it totally mediated the relationship between paternal pre-natal representations and child attachment.
# Introduction
The parent-child relationship begins during pregnancy, when both women and men face dramatic psychological reorganizations related to the new task of becoming parents . In women, these transformative processes are sustained by the development of mental representations, concerning themselves as mothers, the unborn infant, and the future relationship with him (5-9). Maternal representations become particularly clear and rich by the third trimester of pregnancy, when the mothers can fully perceive their infant's vitality thorough intrauterine movements and ultrasound images . These perceptual experiences allow the mothers to experience a shift from focusing on themselves to their infant as a separate object.
Although pre-natal representations have been poorly investigated in men, the existing literature shows that expectant fathers also create an emotional bond with the infant, and this bond increases as the pregnancy progresses.
Pre-natal mental representations include projections, dreams, attributions, and fantasies, which are strictly related to parents' childhood attachment experiences. During pregnancy, such early experiences become closer to conscious awareness, allowing both women and men to identify with their parents and-at the same time-develop their own parental identity. As George and Solomonhave underlined, a relevant change takes place in the representational world of expectant parents, whose goals switch from being cared to being caregivers.
A central task for expectant parents is developing a feeling of intimate connection to the infant and, progressively, recognizing him/her as a separate individual who has the need for both care and autonomy. Consistent with these considerations, Tambelli et al.have underlined that, when pre-natal representations are flexible and open to change, parents can "have an unconditional acceptance of the infant and a realistic consideration of the baby's individual characteristics and of any difficulties emerging in the relationship with him or her" (pp. 378-379).
Parental pre-natal representations tend to be relatively stable after their formation, serving as a sensitive indicator of the caregiving behaviorsand interactive patterns that parents will put in place after childbirth.
## Emotional availability and child attachment
After childbirth, the parent-child relationship takes the form of a "reciprocal interchange, " that occurs between the innate propensity of infants to engage with and share the subjective states of othersand the ability of parents to interpret and respond appropriately to the emotional underpinnings of their infant's overt behavior.
From such a perspective, the theoretical construct of Emotional Availability (EA)provides a relevant description of the parent-infant relationship. Such a constructwhich is theoretically grounded on attachment theoryand integrated with Emde's conceptualization of emotions (36)-refers to the "capacity of a dyad to share an emotional connection and to enjoy a mutually fulfilling and healthy relation" [(37), p. 1]. Scientific literature has documented that EA predicts a wide range of child outcomes, including attachment security [for a review, see].
During the first year of life, repeated interactions with the parents are internalized as an internal working models (IWMs) of attachment. IWMs can be regarded as generalized representations of "lived experiences" with primary caregivers, that remain fairly stable across the lifespan.
Early positive emotional interactions allow the children to consider the parents as secure base, that is as someone who is emotionally available to them in times of distress. The internalization of these positive interactive experiences (and of their related affects) promote the development of attachment security in children. In these cases, IWMs include positive expectations about others' EA and willingness to provide support, along with positive representations of the self as competent and valued. Conversely, when parents are not emotionally available and responsive, children develop doubts about their self-worth and others' goodwill, and use defensive strategies other than confident proximity seeking, to face distress. As a result of these negative emotional experiences, two insecure IWMs-avoidant or resistant/ambivalent-are likely to develop. Avoidance reflects a tendency to use deactivating strategies, in response to parents that children perceive as insensitive or rejecting to their attachment needs of reassurance. In these cases, children tend to hide or suppress negative emotions (such as anxiety, fear, anger or need for consolation) and deal with distress autonomously to avoid the frustration caused by the potential rejection from the parents. On the contrary, resistant/ambivalent attachment reflects the use of hyperactivating strategies, in response to parents who show inconsistent, hesitant or unpredictable EA toward the emotional needs of their children. In these cases, children tend to amplify proximity seeking behaviors to demand or force the parents to be accessible and to pay more attention to them.
## Aims of the study
Whereas, a large body of research has explored the influence of parental post-natal representations on the quality of attachment in children, studies on the predictive role of parents' prenatal representations on child attachment are still very scarce. Given that pre-natal representations are found to be related to both post-natal representations and post-natal parent-infant interaction, it would be important that the influence of parental pre-natal representations on parent-child attachment may also be investigated. At the same time, it is worth noting that, even though research has come to document the relevant influence of fathers on children's development, over the past decades, research on parental prenatal representations have mainly focused on mothers. We believe that the lack of studies on fathers represents a further relevant gap within scientific literature, which should be filled by greater attention toward paternal contribution to child socioemotional development.
Considering the importance of exploring, both in mothers and in fathers, the complex constellations of mental representations during pregnancy as well as their influence on child attachment, this study aimed at:
-investigating the criterion validity of the IRMAG-R and IRPAG, using EA in mother-child and father-child interactions (hereafter referred to as maternal EA and paternal EA), parental attachment, and child attachment toward parents as criteria; -testing the incremental validity of the IRMAG-R and IRPAG in the prediction of child attachment, with respect to parental depressive and anxious levels during pregnancy, EA scales, and parental attachment; -evaluating the possible mediation role of EA on the relationship between parental representations during pregnancy and child attachment.
We expected that, both in mothers and in fathers, mental representations during pregnancy will be positively correlated with EA as well as parental and children's attachment categories.
We also expected that, both in mothers and fathers, the categories of pre-natal mental representations will provide a unique incremental contribution in the prediction of children's attachment categories, even when parental depressive and anxious level during pregnancy, EA, and attachment were included as covariates.
Finally, we expected that, both in mothers and in fathers, EA will mediate the effect of pre-natal parental representations on children's attachment.
# Materials and methods
## Participants
Initially, 189 couples of primiparous parents were recruited at seventh/eighth month of pregnancy, while they were attending childbirth preparation courses at maternity and child health services. These parents had been enrolled in a larger extensive research, aimed at assessing the effects of early interventions on parents at risk for psychopathological symptoms and on their children's socio-emotional development during the first year of life. The screening of parental anxiety and depression revealed 78 couples in which both parents were within the normal range. These parental couples were excluded from the general study and were enrolled in the present investigation. Of these 78 couples, 28 were ruled out because they did not complete all the longitudinal observations. Thus, the final samples consisted of 50 couples of primiparous mothers (mean age = 33.88 years; SD = 4.58) and primiparous fathers (mean age = 36.90 years; SD = 6.69).
The gestation period did not reveal complications for 66% of mothers, and only 8% of them reported that they needed at least one hospitalization. In addition, 20% of mothers reported having had abortions previously. At the time of the study neither mothers nor fathers showed the presence of anxious or depressive symptoms.
This study was carried out in accordance with the recommendations of the Ethics Committee of the Department of Dynamic and Clinical Psychology, "Sapienza" University of Rome. Prior to data collection, the parents received complete information concerning the rationale of the study procedures and provided their written informed consent to participate to the research study, as stated in the Declaration of Helsinki.
## Procedure
The longitudinal study included three measurement occasions, in which different types of instruments were administrated at the Department of Dynamic and Clinical Psychology: semistructured interviews, self-report scales, and rating scales applied to videotaped materials. (61)]. Even though this self-report scale was originally developed to measure depressive symptomatology in mothers during the post-natal period, its validity has also been successively demonstrated during pregnancy as well as in its application with fathers. The EPDS includes 10 items that explore the presence of the following depressive symptoms during the past week: inability to laugh, inability to enjoy, unmotivated feelings of guilt, state of anxiety or worry, moments of fear or panic, feeling of being overwhelmed by things, difficulty in sleep due to sadness and unhappiness, feeling of sadness, presence of excessive crying, thinking of getting hurt. The internal consistency of the Italian version of the EPDS was evaluated both with a Cronbach's estimation (alpha = 0.79) and Guttman split-half index (r tt = 0.82). A Global Score for depressive symptomatology was computed summing up all items. The cut-off value for the Italian version of the scale is 12/13 for clinical depression and 9/10 for screening purposes.
Interview of Maternal Representations During Pregnancy-Revised (IRMAG-R)and Interview of Paternal Representations During Pregnancy (IRPAG). These semistructured interviews consist of 47 open questions, designed to assess maternal and paternal representations during the third trimester of pregnancy, by examining parental narratives regarding the future child and the unfolding of the relationship with him/her. Parental narratives are coded as a function of seven different dimensions (i.e., richness of perceptions, openness to change, intensity of investment, coherence, differentiation, social dependency, and dominance of fantasies), that allow the mothers' and fathers' transcripts to be classified into one of three categories: Integrated/Balanced, Restricted/Disengaged and Not Integrated/Ambivalent. The Integrated/Balanced category is characterized by the ability of parents to provide a consistent picture of their experience in the context of their personal history; they give rich, affectively involved and flexible representations of their children, even though still unborn, and of their future with him/her. Parents consider pregnancy as an important step of personal development and the fulfillment of their personal identity. Restricted/Disengaged category is characterized by rigid representations, impersonality, poor fantasies, and high emotional control and inhibition. Moreover, restricted/disengaged parents usually show difficulty imagining and managing the relationship with their children, and recognizing the experience of pregnancy. Finally, in Not Integrated/Ambivalent category, parents tend to report not organized and poorly coherent narratives, in which different tendencies toward parenthood and the child coexist (defined by excessive involvement and the struggle to impose distances), as they are strongly absorbed by their conflicts with their original family or partners. The degree of inter-rater reliability for all dimensions as estimated in terms of agreement between judges was: 0.86 for Richness of Perceptions; 0.89 for Openness to Change; 0.90 for Intensity of Investment; 0.84 for Coherence; 0.93 for Differentiation; 0.97 for Social Dependency, and 0.86 for Dominance of Fantasies, confirming the high level of reliability of the instrument.
Emotional Availability Scales (EAS). The EAS coding system [EAS 4th Edn;] was applied to 15/20 min of video-recorded free-play home-interactions. The instrument was composed of six scales designed to assess different dimensions of parent-child emotional regulation. Four scales concern parental EA toward children (Sensitivity, Structuring, Non-Intrusiveness, and Non-Hostility), and two concern children's EA toward parents (Responsiveness and Involvement), with a range from one (highly emotional unavailable) to seven (highly emotional available) points. Sensitivity refers to parental affectivity, acceptance, flexibility, clarity of perceptions, affect regulation, and variety and creativity that was shown during play toward children. Structuring refers to parental capacity to give rules, regulations and a supportive framework for interaction.
Non-Intrusiveness refers to parental capacity to interact with the child without being over-directive, over-stimulating or overprotective. Non-Hostility concerns covert and overt parental hostility. Responsiveness refers to children's availability toward their parents' requests of interaction, along with children's enjoyment of the interaction. Involvement regards children's willingness to interact with their parents. Inter-rater reliability, assessed with mean absolute agreement intraclass correlation coefficients (ICC), ranged from 0.81 to 0.93.
Adult Attachment Interview (AAI). The AAI is a semi-structured interview formed by 20 questions requesting respondents to describe their relationship with main attachment figures during childhood, specific positive or negative memories, traumas, and current attachment relationships. Some questions specifically concern crucial events related to attachment relationships, such as illnesses, separations and rejections. Adult participants are asked to recall autobiographical memories from early childhood in order to evaluate the narratives produced, by considering the structural dimension of the transcript rather than its content. The AAI coding system was applied to categorize participants into one of five categories corresponding to different states of mind with respect to attachment: Secure/Autonomous (F); Dismissing (Ds); Preoccupied (E); Unresolved/Disorganized (U); Cannot Classify (CC). The F classification includes individuals who value attachment relationships, describe their attachment experiences (whether positive or negative) coherently and consider them important for their own personality. In the DS classification, adults tend to minimize the importance of attachment for their own lives or to idealize their childhood experiences. Adults classified as E tend to maximize the importance of attachment, are still very much involved and preoccupied with their past experiences, and are unable to describe them coherently and reflectively. Anger or passivity characterizes the discourse style of these adults. The additional classification U is applied to interviewees who show signs of unresolved experiences of trauma usually involving the loss of attachment figures. Finally, the CC classification is applied when a transcript has strong characteristics of both the dismissing and preoccupied categories. Inter-rater reliability with respect to the main category was 89% with a k = 0.74, p = 0.001.
Strange Situation Procedure (SSP). The SSP is a standardized laboratory observational procedure, commonly carried out between 12 and 18 months after childbirth, during which the child's attachment behavior toward his/her parent is activated and intensified by the child's exposure to a moderately, yet increasingly stressful situation (i.e., the presence of a strange person and two short separations from the mother). The SSP originally classified infants into three categories: Secure (B); Insecure Avoidant (A); and Insecure Resistant/Ambivalent Attachment (C). The B classification characterizes children who use parents as a secure basis when they are present, show distress when were separated from them and actively seek contact when they return with a certain predisposition to be easily consoled. The A classification characterizes children who do not seek contact and play with parents, and do not show distress when are separated from them. During the reunion with the parents, these children are not interested in seeking proximity to them, manifest a tendency to avoid contact with them and continue to play or to explore the environment. Moreover, they are not disturbed in the presence of the unfamiliar adult (the stranger) and during the entire procedure. Finally, The C classification characterizes children who are strongly focused on parents during SSP, show reluctance to explore the environment, and express high levels of distress during the separations from the parents as well as inconsolability during the reunions with them. Main and Solomon (53) later added a fourth category, Disorganized/Disoriented (D) defined by odd, awkward behavior and unusual fluctuations between anxiety and avoidance. As reported by George and Solomon (65), when coders are trained to categorize attachment styles using all categories the percentage of agreement between judges was from 80 to 88%.
# Data analysis
To investigate the criterion validity of the IRMAG-R and IRPAG categorizations, point-biserial correlations (for relationships between dichotomous and scale variables) and phi correlations (for relationships between dichotomous variables) were used.
To investigate the incremental validity of the IRMAG-R and IRPAG categorizations in the prediction of children's attachment, with respect to levels of anxiety and depression during pregnancy, parent-child EA, and adult attachment), two logistic regressions were conducted.
Finally, to investigate the mediation role of parent-child EA on the relationship between parental representations during pregnancy and children's attachment patterns, two mediation models were analyzed (one for mothers and one for fathers) through Mplus software, Version 8 (66), using the weighted least square mean and variance adjusted estimator (WLSMV), which allows the computation of the indirect effects also with dichotomous outcomes. Moreover, a biased corrected estimation of confidence intervals of parameters through a bootstrap procedure was used. As regards statistical power, it is worth noting that medium/large effect sizes are expected for the relationships among the main factors investigated (i.e., parental representations during pregnancy, EA and children's attachment patterns). In this view, to reach a statistical power of 0.80 with medium/large effects and a level of alpha to 0.05, the number of subjects needed is about 50 subjects for both logistic regressions [G * Power software;] and mediation models with biased corrected confidence intervals for indirect effects' parameters.illustrates the frequency and percentages for each category of the IRMAG-R and IRPAG, maternal AAI and paternal AAI, and SSP, as carried out with the mothers and fathers, respectively (hereafter referred as maternal SSP and paternal SSP).
# Results
## Descriptive statistics
As regards the narratives about parenthood during pregnancy, more than 50% of representations were Integrated/Balanced for both mothers and fathers. Restricted/Disengaged representations were slightly higher for fathers, while Not Integrated/Ambivalent were slightly higher for mothers.
Regarding parental AAI, more than 50% of both mothers and fathers showed a Secure attachment. The number for Dismissing attachment was higher than that for Preoccupied for both mothers and fathers.
As regards the SSP procedure, more than 50% of the children showed a Secure attachment both toward mothers and fathers. However, a higher percentage of Secure attachment emerged toward mothers rather than toward fathers. The number of children's Insecure Avoidant attachment was two times higher toward fathers rather than toward mothers, while a similar number for Insecure Resistant/Ambivalent attachment was found toward mothers and fathers.shows the descriptive statistics of all scales considered in the present study, with mean scores, standard deviation as well as skewness and kurtosis values for both maternal and paternal scales. Interestingly, mean values are higher for maternal than paternal scores in all scales, especially for trait anxiety (STAI-TRAIT) and depressive symptoms. All scales showed close to normal distribution, except for the kurtosis value of maternal STAI-STATE and scores on EAS related to father-child interactions, which revealed values higher than one.
## Criterion validity of the irmag-r and irpag
To evaluate the criterion validity of the IRMAG-R and IRPAG, the correlations of maternal and paternal representations during pregnancy with parental EA, and with parental and children's attachment were estimated. Since some of the IRMAG-R, IRPAG, AAI and SSP categories showed a too low frequency to conduct appropriate statistical analysis (e.g., seeAs illustrated in, the IRMAG-R and IRPAG categories (Integrated vs. Not Integrated representations) were positively and significantly correlated (phi correlation) with maternal and paternal AAI and SSP categories, with a high effect size, providing support for their criterion validity. Inpointbiserial correlations between the IRMAG-R/IRPAG categories and EAS scores are also reported. Significant and positive correlations were found between parental representations during pregnancy and parental EAS with a high effect size, with a further support for the criterion validity of IRMAG/ IRPAG interviews.
## Incremental validity of the irmag-r and irpag in the prediction of child attachment
In order to investigate the incremental validity of the IRMAG-R and IRPAG, two logistic regressions (one for mothers and one for fathers) were conducted, including parental representations during pregnancy (Integrated vs. Not Integrated), EA and the parental attachment model (Secure vs. Insecure) as predictors, children's attachment (Secure vs. Insecure) as a criterion, and age and children's gender as covariates. Parental anxious and depressive scores were excluded from these analyses, as they did not show significant correlations neither with IRMAG-R/IRPAG, nor with parental and children's attachment and with EA scales. As illustrated in, overall maternal predictors accounted for a considerable portion (R 2 Nagelkerke = 0.54) of maternal SSP variability. Moreover, the Hosmer and Lemeshow test indicated an adequate fit for the model [χ 2 (8) = 5.09, p = 0.75].
As regards the single predictors, a significant unique contribution emerged for the IRMAG-R with a high effect size in terms of Odds Ratio, while not significant unique contributions were found for maternal AAI, EAS, as well as for children's gender and maternal age.
Similarly, paternal predictors accounted for a high portion (R 2 Nagelkerke = 0.64) of paternal SSP variability (See, and the Hosmer and Lemeshow test again indicated an adequate fit for the model [χ 2 (8) = 8.48, p = 0.39]. As regards single predictor effects, a significant unique contribution emerged for paternal EAS, with a high effect size (as evaluated in terms of Odds Ratio), as well as a close to be significant contribution (with a large Odds Ratio) of the IRPAG categories. Not significant contributions were found for AAI categories, children's gender and paternal age.
## Direct and indirect effects of parental representations during pregnancy on child attachment: the mediating role of ea
In order to further investigate the effect of parental representations during pregnancy on children's attachment, two mediation models were tested (one for mothers and one for fathers), including the IRMAG-R/IRPAG categories (Integrated vs. Not Integrated) as independent variables, maternal and paternal EAS scores as mediators, SSP categories (Secure vs. Insecure) toward mothers and fathers as dependent variables, and maternal and paternal AAI categories (Secure vs. Insecure) as covariates. These mediation analyses were conducted with Mplus Version 8 software using the WLSMV estimator and bootstrapping procedure to estimate confidence interval of indirect effects. The first mediation model, which included maternal variables, showed an adequate fit with the data [χ 2
(1) = 0.17, p = 0.68; RMSEA = 0.00 (0.00-0.28); CFI = 1.00; TLI = 1.12; WRMR = 0.14]. The model accounted for 53% of children's attachment variability, and for 46% of maternal EA. As illustrated in, results showed significant direct effects of IRMAG-R categorizations on both EAS scores (Biased Corrected Bootstrap 99% CI: from 0.47 to 0.82) and SSP categories (Biased Corrected Bootstrap 99% CI: from 0.07 to 1.10), and not significant effects of EAS scores (Biased Corrected Bootstrap 99% CI: from −0.37 to 0.62) and maternal AAI categories on SSP ones (Biased Corrected Bootstrap 99% CI: from −0.61 to 0.24).
Moreover, a not significant indirect effect (Estimate = 0.02, Estimate/SE = 0.18, p < 0.86) of the IRMAG-R categorizations, via maternal EAS, on children's attachment toward mothers was found (Biased Corrected Bootstrap 99% CI: from −0.25 to 0.45), indicating that the impact of maternal representations during pregnancy on children's attachment was exclusively direct.
Similar to the first model, the second one, which included paternal variables, also showed an adequate fit with the data [χ 2= 0.55, p = 0.46; RMSEA = 0.00 (0.00-0.34); CFI = 1.00; TLI = 1.05; WRMR = 0.26]. The model accounted for 76% of children's attachment variability, and for 28% of paternal EA. As illustrated in, a significant direct effect of the IRPAG categories on EAS scores was found (Biased Corrected Bootstrap 99% CI: from 0.22 to 0.71), along with a significant effect of the EAS scores on children's attachment (Biased Corrected Bootstrap 99% CI: from 0.47 to 1.03). Conversely, not significant direct effects of both IRPAG (Biased Corrected Bootstrap 99% CI: from −0.22 to 0.55) and paternal AAI categories on children' attachment (Biased Corrected Bootstrap 99% CI: from −0.57 to 0.24) were found.
Different from the first model, a significant indirect effect (Estimate = 0.41, Estimate/SE = 4.39, p < 0.001) of the IRPAG categories, via paternal EA, on children's attachment toward fathers was found (Biased Corrected Bootstrap 99% CI: from 0.19 to 0.68), indicating that the impact of paternal representations during pregnancy on children's attachment is totally indirect.
# Discussion
Whereas the influence of parental post-natal representations on child attachment has been extensively investigated, much less is known about the predictive role of parental prenatal representations on child attachment. Moreover, the existing literature on pre-natal mental representations have been mainly focused on mothers, while paternal mental representations during pregnancy have received scarce attention from the research.
Beginning from these premises, we firstly investigated the criterion validity of the IRMAG-R and IRPAG, using maternal EA, paternal EA, parental attachment, and child attachment toward parents as criteria.
Consistent with our expectations, both in mothers and in fathers, the results showed a strong relationship between pre-natal representations and EA, and between pre-natal representations and child attachment categories. These associations are supported by the evidence that parental prenatal expectations, thoughts, and fantasies shape an anticipatory working model (6, 69), that sustains women and men in the achievement of a parental identity and in the development of an early attachment bond with their infants. Consistent with the results of previous investigations, small size correlations (in terms of Cohen's standards) were also found between pre-natal representations and parental attachment, with a further support for the criterion validity of the IRMAG-R and IRPAG. These results may be interpreted taking into account the role played by the reworking of early attachment relationship in enabling expectant mothers and fathers to achieve their own parental identity and develop the capacity to recognize the unborn infant as a separate individual with specific needs. Parents with positive and stable childhood experiences are more likely to develop and maintain flexible and coherent representations about attachment and caregiving during the transition to parenthood.
Having explored the criterion validity of parental prenatal representations, we tested the incremental validity of the IRMAG-R and IRPAG in the prediction of child attachment, with respect to children's gender, parental age, EA scales, and parental attachment.
Contrary to our expectations, even though maternal and paternal predictors accounted for a considerable portion of variability of attachment categories in children, relevant differences between mothers and fathers were found regarding the effects of the single predictors. As regards the mothers, the categories of mental representation during pregnancy provided a significant unique incremental contribution in the prediction of children's attachment categories, while all the other considered predictors did not provide a unique contribution. These results evidenced that, compared to children whose mothers have not integrated pre-natal representations, children whose mothers report integrated representations during pregnancy have a higher possibility to develop a sense of security in the attachment relationship. Thus, in our study, maternal pre-natal representations have a specific role in the construction of the attachment bond with the child. Different from that observed in mothers, in the case of fathers, child attachment categories were accounted for by the significant unique contribution of EA during dyadic interactive exchanges. Moreover, fathers' prenatal representations resulted in providing a weaker (close to be significant) unique contribution to child attachment if compared to maternal representations. These results show that children's attachment security toward fathers is more likely to be associated with high EA during dyadic interactive exchanges. Undoubtedly, the results concerning fathers confirmed those of previous investigations that have documented the role of parentchild EA in predicting secure attachment in children. At the same time, it is worth noting that, even though pre-natal representations resulted in predicting child attachment both in mothers and (with a lower effect size) in fathers, only in mothers the other predictors did not provide any unique contribution.
Finally, taking into account the peculiarities between maternal and paternal contributions on child attachment, we evaluated the possible mediation role of EA on the relationship between parental representations during pregnancy and child attachment.
Even in this case, the analyses produced unexpected findings. As regards the mothers, results showed significant direct contributions of mental representations in predicting both EA and child attachment categorizations, while no direct effects were found for EA and maternal attachment on child attachment categories. Moreover, no indirect effect of maternal pre-natal representations, via EA, on children's attachment categories was found. Different from that observed in mothers, in fathers, a significant direct effect of pre-natal representations on EA was found, along with a significant effect of EA on children's attachment category. Conversely, neither paternal pre-natal representations nor paternal attachment category resulted in having a direct effect on children's attachment categories.
The results of our study may be explained by taking into account the well-known condition of primary maternal preoccupation, that has been conceptualized as "almost an illness" that a mother must experience and recover from, in order to provide the infant with an environment that can meet his/her physical and psychological needs. As Leckman et al.have evidenced, such preoccupations develop during the last months of pregnancy, affecting both mothers and (to a lesser extent) fathers, with the aim of heightening parental ability to anticipate the infant's needs, learn his/her emotional signals, and gradually recognize him/her as an individual. It may be assumed that, because of more intense preoccupations, mothers may be more prone than fathers to develop vivid mental representations of their infants and an early sensitive attitude toward them. The results of our study seem to prove that these maternal inclinations are so consolidated during pregnancy as to shape the ground in which the child's sense of security will be rooted.
In mothers, the experience of a somatic gestation (26) contributes consistently in increasing the richness and specificity of mental representations about their unborn infants. During pregnancy, maternal mental representations are sustained by the perception of the baby, whose vitality is manifested through intrauterine movements and ultrasound images . The father's emotional relationship with the unborn infant is instead indirect, as it is experienced via the mother's willingness to share with them the affective and somatic experience of pregnancy. In this view, it may be assumed that fathers' contribution to their child's attachment security may fully emerge only when they will have the possibility to really interact with their real child (77).
## Limitations and strengths of the study
The main constraint of this study is the small number of the recruited parents, as it produced an increase of parameters' standard errors and a decrease of statistical power that limited the possibility to detect low size effects. The small number of participants did not even permit to test the effect of some potentially relevant variables, such as previous abortions (which was reported only by 10 mothers).
As regards the mediation analysis, it is worth noting that all variables included in the model were assessed only on one occasion of measurement. As a consequence, the analyzed models did not include residual change estimations of both mediator and outcome (as computed using autoregression-based statistical procedures), with a possible distortion of parameter estimations (for an extensive explanation, see. Further longitudinal studies with multiple measurements of all variables are needed to address this potential source of distortion.
In this study, we adopted a dyadic perspective to evaluate (separately for mothers and fathers) the predictive role played by pre-natal representations on child attachment. This did not allow examination whether pre-natal triadic family relations might predict mother-and father-child attachment relationship. As regard this issue, a recent investigation has shown that children's attachment toward fathers (but not toward mothers) is predicted by pre-natal triadic family alliance, that is by the ability of the mother and father to cooperate and support each other in their parental roles.
Notwithstanding these limitations, no previous study has ever investigated the predictive and incremental validity of maternal and paternal pre-natal representations on child attachment. We believe that our results (albeit preliminary) may provide the starting point for future researches, aimed at shedding further light on the distinct (even though complementary) paths, that mothers and fathers follow to contribute to their children' attachment security.
These reflections lead us to consider the inclusion of fathers (who have been long overlooked in scientific literature on parenting) as a further strength of our study.
Finally, we believe that, in this study, the combined use of clinical semi-structured interviews and observational procedures may have provided an articulated description of the complexity underlying the construction of mother-and fatherchild attachment relationship.
# Conclusions
The results of this study may have relevant implications for prevention, clinical practice, and future researches, as they indicate pregnancy as a privileged time for the intervention programs that may be designed to support the parents in creating that intersubjective matrix (80), within which the child's sense of security develops.
The assessment of mental representations during pregnancy provides the opportunity to recognize parents who will have nonoptimal interactions with their infants, after childbirth. Indeed, the IRMAG-R and IRPAG, beyond assessing the emotional valence of parental representations, also allow to identify the presence of defensive strategies (toward pregnancy and the unborn infant) that are sensitive predictors of early impairments in parental EA. These aspects are particularly evident among parents with psychopathological symptomsand with whom this study needs to be replicated.
# Data availability statement
The datasets generated for this study are available on request to the corresponding author.
# Ethics statement
Prior to data collection, the participants received complete information concerning the rationale of the study procedures and provided their written informed consent to participate to the research study, as stated in the Declaration of Helsinki.
# Author contributions
RT: conceived the work, monitored data acquisition, and provided a substantial contribution to the interpretation of the data. As first author, she was primarily accountable for all aspects of the work. CT: wrote the Introductions, Discussion, and Conclusions sections, revised the paper for intellectual content, and approved its final version to be published. FD: analyzed data, wrote the Methods and Results sections, revised the paper for intellectual content, and approved its final version to be published. All authors agreed to be accountable for all aspects of the work and to ensure that questions related to the accuracy or integrity of any part of the work were appropriately investigated and resolved. |
Deconvolution and phylogeny inference of structural variations in tumor genomic samples
Motivation: Phylogenetic reconstruction of tumor evolution has emerged as a crucial tool for making sense of the complexity of emerging cancer genomic datasets. Despite the growing use of phylogenetics in cancer studies, though, the field has only slowly adapted to many ways that tumor evolution differs from classic species evolution. One crucial question in that regard is how to handle inference of structural variations (SVs), which are a major mechanism of evolution in cancers but have been largely neglected in tumor phylogenetics to date, in part due to the challenges of reliably detecting and typing SVs and interpreting them phylogenetically. Results: We present a novel method for reconstructing evolutionary trajectories of SVs from bulk whole-genome sequence data via joint deconvolution and phylogenetics, to infer clonal sub-populations and reconstruct their ancestry. We establish a novel likelihood model for joint deconvolution and phylogenetic inference on bulk SV data and formulate an associated optimization algorithm. We demonstrate the approach to be efficient and accurate for realistic scenarios of SV mutation on simulated data.
# Introduction
Genomic methods have provided a wealth of information about mutational landscapes of developing cancers, but have also created a great need for sophisticated computational models to make sense of the resulting data. They have revealed extensive variation patient-topatient (intertumor heterogeneity) as well as cell-to-cell within single patients (intratumor heterogeneity) [bib_ref] Tumor heterogeneity: causes and consequences, Marusyk [/bib_ref] and suggested a far more complex landscape of somatic variations in cancer development than earlier mutational models [bib_ref] A genetic model for colorectal tumorigenesis, Fearon [/bib_ref] [bib_ref] The clonal evolution of tumor cell populations, Nowell [/bib_ref] had anticipated. Extracting meaningful biological insight from such data nonetheless remains challenging. Much effort has focused on the difficulty of identifying those variants relevant to tumorigenesis and progression, known as the drivers, from the background noise of the many more chance mutations carried along with a developing tumor despite being functionally irrelevant, known as the passengers [bib_ref] Clonal status of actionable driver events and the timing of mutational processes..., Mcgranahan [/bib_ref]. More recently, attention has shifted to understanding what one can learn even from passengers regarding how a particular tumor's mutational spectrum [bib_ref] Mutational signatures: the patterns of somatic mutations hidden in cancer genomes, Alexandrov [/bib_ref] shapes its genome across stages of progression and how that knowledge can predict its future progression and help improve prognosis. These remain substantively unsolved problems that must be better tackled if cancer researchers are to make sense of enormous and ever-growing libraries of genetic variations in cancers.
One key advance in understanding tumor genomic data was the advent of tumor phylogenetics, i.e. the use of phylogenetic inference to reconstruct tumor progression. This field arose from the observation that cancer progression is fundamentally the evolution of clonal cell populations and thus in principle interpretable via algorithms for reconstructing evolutionary trees, i.e. phylogenetics. Tumor phylogenetics itself has greatly evolved, from its initial use in making sense of intertumor heterogeneity via oncogenetic tree models [bib_ref] Inferring tree models of oncogenesis from comparative genomic hybridization data, Desper [/bib_ref] , through the advent of methods for interpreting variation between distinct tumor regions [bib_ref] The clonal evolution of metastases from primary serous epithelial ovarian cancers, Khalique [/bib_ref] [bib_ref] Genetic clonal diversity predicts progression to esophageal adenocarcinoma, Maley [/bib_ref] , between distinct cells in single tumors [bib_ref] Reconstructing tumor phylogenies from heterogeneous single-cell data, Pennington [/bib_ref] and ultimately to recent variants that seek to explain whole-genome evolution of numerous single-cells per tumor [bib_ref] Assessing intratumor heterogeneity and tracking longitudinal and spatial clonal evolutionary history by..., Jahn [/bib_ref] [bib_ref] Onconem: inferring tumor evolution from single-cell sequencing data, Ross [/bib_ref] [bib_ref] Sifit: inferring tumor trees from single-cell sequencing data under finite-sites models, Zafar [/bib_ref].
Single-cell genomic data is beginning to become available in quantity, though most studies of non-trivial patient populations are still limited to bulk sequence data, providing at best variant frequencies averaged across many single cells. Modern methods for working with such data combine phylogenetic inference with a deconvolution step, in which one infers clonal sub-populations from mixed genomic samples prior to or concurrent with inferring phylogenetic relationships between those sub-populations [bib_ref] Applying unmixing to gene expression data for tumor phylogeny inference, Schwartz [/bib_ref]. Numerous tumor phylogeny methods now work on this basic model of joint deconvolution and phylogenetics, with prominent examples including THeTA [bib_ref] Quantifying tumor heterogeneity in whole-genome and whole-exome sequencing data, Oesper [/bib_ref] , Pyclone [bib_ref] Pyclone: statistical inference of clonal population structure in cancer, Roth [/bib_ref] , PhyloWGS [bib_ref] PhyloWGS: reconstructing subclonal composition and evolution from whole-genome sequencing of tumors, Deshwar [/bib_ref] , SPRUCE [bib_ref] Inferring the mutational history of a tumor using multi-state perfect phylogeny mixtures, El-Kebir [/bib_ref] and CITUP [bib_ref] Clonality inference in multiple tumor samples using phylogeny, Malikic [/bib_ref]. See [bib_ref] Cancer evolution: mathematical models and computational inference, Beerenwinkel [/bib_ref] for recent reviews.
Despite many advances, though, key aspects of the problem of reconstructing tumor evolution from variation data remain unresolved, an important one being the interpretation of structural variations (SVs). SVs, along with the copy number aberrations (CNAs) they frequently induce, are the primary mechanism of phenotypic adaptation in developing cancers [bib_ref] Pan-cancer patterns of somatic copy number alteration, Zack [/bib_ref]. Most tumor phylogeny methods until recently focused primarily on single nucleotide variations (SNVs) [e.g. [bib_ref] Reconstruction of clonal trees and tumor composition from multi-sample sequencing data, El-Kebir [/bib_ref] [bib_ref] Fast and scalable inference of multi-sample cancer lineages, Popic [/bib_ref] ]. SNVs are generally abundant and make for computationally simpler analyses than other marker types but omit much of the functional mutation that we often seek to understand with tumor phylogenetics. Some early methods did focus primarily on CNAs for deconvolution [bib_ref] Robust unmixing of tumor states in array comparative genomic hybridization data, Tolliver [/bib_ref] and phylogenetics [bib_ref] Phylogenetic analysis of multiprobe fluorescence in situ hybridization data from tumor cell..., Chowdhury [/bib_ref] [bib_ref] Reconstructing tumor phylogenies from heterogeneous single-cell data, Pennington [/bib_ref] [bib_ref] The Cancer Genome Atlas Network. (2012) Comprehensive molecular portraits of human breast..., Schwarz [/bib_ref] and several tools are now available for joint inference of SNVs and CNAs [e.g. [bib_ref] PhyloWGS: reconstructing subclonal composition and evolution from whole-genome sequencing of tumors, Deshwar [/bib_ref] [bib_ref] Inferring the mutational history of a tumor using multi-state perfect phylogeny mixtures, El-Kebir [/bib_ref] [bib_ref] Assessing intratumor heterogeneity and tracking longitudinal and spatial clonal evolutionary history by..., Jahn [/bib_ref] ]. There is, to our knowledge, however, no method that handles phylogenetics of SVs more comprehensively. Despite their importance, SVs introduce a number of technical challenges, including difficulty of reliable detection leading to a high expected missing data rate, of reconstructing variants that by their nature are associated with copy number variant regions of the genome, and of interpreting these more complicated event types phylogenetically.
The goal of this paper is to address the lack of methods for tumor deconvolution and phylogenetics of diverse classes of SVs at nucleotide resolution. Specifically, we develop a new method for simultaneously deconvolving inferences of SVs, derived from the Weaver variant caller [bib_ref] Allele-specific quantification of structural variations in cancer genomes, Li [/bib_ref] and reconstructing the likely evolution of clonal populations via these SV events. The method relies on a novel model extending prior literature on SNV and CNA phylogenetics [bib_ref] Inferring the mutational history of a tumor using multi-state perfect phylogeny mixtures, El-Kebir [/bib_ref] to handle SVs. It depends on a model of joint likelihood of genomic sequence data and clonal phylogenies, which we pose and solve through a combinatorial coordinate descent inference strategy. We demonstrate, on simulated and the Cancer Genome Atlas [TCGA (The Cancer Genome Atlas Network, 2012)] samples that these methods are practical and effective in inferring progression of major clones from bulk whole genome sequence (WGS) data.
# Materials and methods
## Breakpoint and structural variant definitions
Let chrm: pos denote the position and chromosome for each base pair in a reference genome. For example, 7:501 represents the base pair at position 501 on chromosome 7. We define a breakpoint as any base pair c: i that is found non-adjacent to either base pair c: i-1 or base pair c: i þ 1. If base pair c: i was found nonadjacent to base pair c: i-1 we denote the breakpoint as [c: i[as the intact chromatin extends to the right, while if base pair c: i was found non-adjacent to base pair c: i þ 1 we denote the breakpoint as]c: i]. We define a structural variant (SV) as a pair of breakpoints found adjacent to one another in the cancer genome but at nonadjacent positions in a reference genome. We call each such pair of breakpoints a mated pair, or mates for short. For example, SV]2:30],[5:10[means that the segment on the reference genome on chromosome 2 at position 30 extending to the left was found next to the segment on the reference genome on chromosome 5 at position 10 in the cancer genome. This is specifically an example of a translocation SV, as the re-arrangement involves different chromosomes.
To relate SVs to CNAs, we assume the reference genome is partitioned into r segments, with breakpoints positioned on the ends of segments excluding ends of chromosomes. (In practice, edges of segments are not always supported by breakpoints as mated breakpoints cannot always be supported with a sufficient number of reads). Each breakpoint is found in exactly one segment. Because of this, we can define both the number of times a mated breakpoint appears in a genome (denoted c b for the copy number of breakpoint b) and the copy number of the segments containing each breakpoint (denoted c b for the copy number of the segment containing breakpoint b). A more in depth example for the appearance and copy number of breakpoints is given in [fig_ref] Figure 1: Example genomes before and after segmental deletion and duplication [/fig_ref].
# Problem statement
Our method takes as input variant calls. We currently assume these calls are of the form produced by Weaver [bib_ref] Allele-specific quantification of structural variations in cancer genomes, Li [/bib_ref] , which calls SVs and CNAs from bulk genomic read data and estimates copy numbers for copy number segments and breakpoints supporting the SVs. Weaver partitions the genome into r segments and infers the mixed copy number of these segments. Weaver reports the copy number of ' phased breakpoints with sufficient number of reads supporting them, as well as a mapping of mated breakpoints to form SVs. Although Weaver provides additional phase information, we combine homologous chromosomes by summing copy number segments of sister chromatids and assuming SVs initially appear on only one of the chromatids. We use the Weaver output to construct an m  ' þ r ð Þmixed copy number matrix F, the m rows of which represent tumor samples and columns of which represent mutations. The first ' columns correspond to breakpoints and the next r to mixed segmented copy numbers. The variant calls also provide a mapping of breakpoint positions to segments, which we code as an '  r binary matrix Q. We also use information mapping breakpoints to structural variants, encoded as '  ' binary matrix G. From these inputs, we seek simultaneously to infer an integer copy number matrix C, which describes copy numbers across the genome regions profiled for each inferred clonal cell population; a mixture fraction matrix U, which describes how clonal populations are distributed among tumor samples and a phylogeny T, describing ancestral relationships among the clones. We assume the number of leaves n in the phylogenetic tree containing N ¼ 2n À 1 total nodes (clones) is known. More formally, given
[formula] F 2 R mÂð'þrÞ !0 f p;s is mixed copy number of variant s in sample p Q 2 f0; 1g 'Âr q b;s is 1 iff breakpoint b is in segment s G 2 f0; 1g 'Â' [/formula]
g s;t is 1 iff breakpoints s and t are mated pairs n 2 N number of leaves in the phylogenetic tree c max 2 N >2 maximum allowed sub-clonal copy number for breakpoints and segments k 1 2 R !0 regularization term to weight total tree cost k 2 2 R !0 regularization term to weight breakpoint consistency i358 J.Eaton et al.
where
[formula] P r s¼1 q b;s ¼ 1 8 b 2 f1; . . . ; 'g; P ' b 0 ¼1 g b;b 0 ¼ 1 8 b 2 f1; [/formula]
. . . ; 'g, we seek to determine and minimize the objective function
[formula] min U;C jF À UCj þ k 1 R þ k 2 S ð Þ(1) [/formula]
where jF À UCj describes the deviation between true and inferred mixed copy numbers, R is a phylogenetic cost, S is a cost capturing consistency between SVs and copy number segments, and k 1 and k 2 are regularization terms (constants). An overview of of the inputs and outputs to this problem including a toy example is given in .
## Coordinate descent algorithm overview
We solve for U, C and T given F, Q and G using coordinate descent [bib_ref] The copy number tree mixture deconvolution problem and applications to multi-sample bulk..., Zaccaria [/bib_ref]. We write two linear programs: one solving for U given F and C and the other solving for C given U and F. We then iteratively alternate between solving for U and for C while holding the other constant, either until convergence where U and C remain unchanged between iterations, or until a maximum number of iterations is reached. To avoid local minima, we run coordinate descent on multiple random initializations of U. Each row in U is independently randomly uniformly initialized so P N k¼1 u p;k ¼ 1 8 p 2 f1; . . . ; mg and samples independently distributed.
## Estimating u
In solving for Equation (1), we define the L1 distance jF À UCj as
[formula] f D;p;s ! f p;s À X N k¼1 u p;k Á c k;s 8 p 2 f1; . . . ; mg; s 2 f1; . . . ; ' þ rg (2) f D;p;s ! Àf p;s þ X N k¼1 u p;k Á c k;s 8 p 2 f1; . . . ; mg; s 2 f1; . . . ; ' þ rg (3) jF À UCj ¼ X m p¼1 X 'þr s¼1 f D;p;s(4) [/formula]
Assume then that F and C are given. To ensure each element u p;k 2 U is a percentage of cell type k in sample p and that percentage for a single sample sum to 1, we constrain u p;k so 0
[formula] u p;k 1 8 p 2 f1; . . . ; mg; k 2 f1; . . . ; Ng (5) X N k¼1 u p;k ¼ 1 8 p 2 f1; . . . ; mg(6) [/formula]
Since the regularization terms in our minimization [Equation (1)] do not depend on U, we can then simply find U to minimize jF À UCj [Equation (4)] given F and C subject to constraints Equations (2), (3), (5) and (6).
## Estimate c
We then estimate C and T given F, U, Q and G.
## Binary indicator variables
Any variable x has an associated indicator variable
x defined as bin
[formula] x ð Þ ¼ x ¼ 1 x > 0 0 x ¼ 0 ((7) [/formula]
This is used throughout the following sections. To linearly define x, we introduce temporary variable y b 2 f0; 1g as the bit representation of x over q bits [bib_ref] The copy number tree mixture deconvolution problem and applications to multi-sample bulk..., Zaccaria [/bib_ref]. The values of temporary variable y b only apply to Equations (8) and (9). y b is then defined by
[formula] X b log 2 xmaxcþ1 b¼0 2 b Á y b ¼ x(8) [/formula]
and constrains x as
[formula] 0 y b x X b log 2 xmaxcþ1 a¼0 y a 8 b 2 f0; . . . ; b log 2 x max c þ 1g (9) so [/formula]
x is 0 if all bits b are 0 and 1 if any bit of x is 1. In this way, any integer variable x with a maximum value x max can be represented in binary form
x. Binary indicator variables are noted with a bar on top
x or by bin (x). C 2 Z NÂð'þrÞ !0 c k;s is the integer copy number of segment or breakpoint s in clone k U 2 ½0; 1 mÂN u p;k is the cell type k that makes up sample p E 2 f0; 1g NÂN e i;j is 1 iff directed edge (v i , v j ) exists in the inferred phylogeny T
[formula] A B C D E F A D E F Seg [/formula]
## Phylogenetic constraints
Since the individual rows of C are not independent but instead share a phylogenetic history, we create a tree structure T representing the inferred relationships between rows in C. We define a binary tree T using a N Â N directed adjacency matrix E. To impose a tree structure on E, assume the first n clones are leaf nodes and clones n þ 1 through 2n À 1 ¼ N are internal nodes, with node N as the root. We constrain element e i;j as follows: root, incoming edges
[formula] e i;N ¼ 0 8 i 2 f1; . . . ; Ng(10) [/formula]
non-root, incoming edges
[formula] X NÀ1 i¼1 e i;j ¼ 1 8 j 2 f1; . . . ; N À 1g(11) [/formula]
leaves, outgoing edges
[formula] e i;j ¼ 0 8 i 2 f1; . . . ; ng; j 2 f1; . . . ; Ng(12) [/formula]
internal nodes, outgoing edges
[formula] X N j¼1 e i;j ¼ 2 8 i 2 fn þ 1; . . . ; Ng(13) [/formula]
Equations (10) and (11) ensure the root has no in-edges and all other nodes have exactly one in-edge. Equations (12) and (13) force leaves to have no out-edges and all internal nodes to have exactly two out-edges.
## Phylogenetic cost
We next ensure all copy numbers are below some input maximum c max and force the normal (non-tumor) root node to be diploid
We next model a phylogenetic tree cost, using CNAs to estimate evolutionary distance q i;j across each tree edge v i ; v j À Á 2 E. We approximate evolutionary distance by the L1 distance between the copy number profiles of an edge's endpoints P r s¼1 jc i;sþ' À c j;sþ' j. While there are more sophisticated models of copy number distance in the literature [bib_ref] Algorithms to model single gene, single chromosome, and whole genome copy number..., Chowdhury [/bib_ref] [bib_ref] Inferring models of multiscale copy number evolution for single-tumor phylogenetics, Chowdhury [/bib_ref] [bib_ref] Complexity and algorithms for copy-number evolution problems, El-Kebir [/bib_ref] [bib_ref] The Cancer Genome Atlas Network. (2012) Comprehensive molecular portraits of human breast..., Schwarz [/bib_ref] , we use L1 distance as an approximation as it can be coded and computed efficiently within the ILP framework. To linearly define q i;j we use temporary variable x i;j;s 2 N NÂNÂr , defined as the absolute change in copy number of segment s on edge (v i , v j ). Here, the values of temporary variable x i;j;s only apply to Equation (17)
x i;j;s ! c i;sþ' À c j;sþ' À c max Á 1 À e i;j À Á 8 i; j 2 f1; . . . ; Ng; s 2 f1; . . . ; ' þ rg
x i;j;s ! Àc i;sþ' þ c j;sþ' À c max Á 1 À e i;j À Á 8 i; j 2 f1; . . . ; Ng; s 2 f1; . . . ; ' þ rg
Equation (17) sets the cost to zero for any pair of nodes (v i , v j ) where v i is not the parent of v j , while Equations (18) and (19) set the cost . Illustrative example of the TUSV algorithm. This figure provides an overview of the method and its inputs and outputs using a small artificial example. m ¼ 2 samples are run through Weaver, which produces ' ¼ 4 breakpoints (2 SVs) and r ¼ 5 segments. Each breakpoint and segment has an average copy number represented by matrix F. Simultaneous phylogenetic inference and deconvolution yields matrices C, U and E which are visually depicted above as a simple n ¼ 2 leaf phylogeny. Each node is represented by the inferred vector of segment copy numbers and by the inferred copy number for each breakpoint. Segments are represented by adjacent boxes while breakpoints are shown by lines between boxes with inferred copy numbers above those lines. The appearance of breakpoints 1 and 3 along the edge to the left leaf corresponds to the regional duplication of segments 2 and 3 while the appearance of breakpoints 2 and 4 along the edge to the right leaf corresponds to a deletion across segments 2 and 3. In this ideal scenario, U Á C exactly equals F to be the absolute difference between copy number for of end nodes for any edge (v i , v j ). We then define the cost across edge (v i , v j ) and total cost of tree as
[formula] q i;j ¼ X r s¼1 x i;j;s 8 i; j 2 f1; . . . ; Ng (20) R ¼ X N i¼1 X N j¼1 q i;j(21) [/formula]
2.5.4 Perfect phylogeny on appearance of breakpoints We next impose a perfect phylogeny on breakpoints. While the perfect phylogeny assumption is problematic for other variant types, we argue that it is sufficiently unlikely for a base-resolution breakpoint to recur that it can be neglected. Note that violations of the infinite sites model due to allelic loss are handled separately by treating a lost allele as having copy number zero. We therefore impose constraints to force each breakpoint to appear across exactly one edge in T and for mated breakpoints to appear together. Define W 2 f0; 1g NÂNÂ' , where each element w i;j;b is 1 if the copy number of breakpoint b goes from 0 to a positive integer across edge (v i , v j ) and 0 otherwise. To linearly define w i;j;b we define temporary variable x i;j;b 2 f0; 1; 2; 3g to be
[formula] x i;j;b ¼ 2 þ c i;b À c j;b À e i;j 8 i; j 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g(22) [/formula]
so x i;j;b is 0 iff the copy number of breakpoint b increases from 0 across edge (v i , v j ). The value of temporary variable x i;j;b only applies to Equations (22) and (23). Using the binary representation x i;j;b of x i;j;b , define w i;j;b and ensure w i;j;b is 1 for a single edge in the tree. w i;j;b ¼ 1 À x i;j;b 8 i; j 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g (23)
[formula] X N i¼1 X N j¼1 w i;j;b ¼ 1 8 b 2 f1; . . . ; 'g(24) [/formula]
Using breakpoint mate indicator g s;t 2 f0; 1g, where g s;t is 1 iff breakpoints s and t are mates, we force breakpoint indicators to be equal for mates. w i;j;s À w i;j;t 1 À g s;t 8 i; j 2 f1; . . . ; Ng; s; t 2 f1; . . . ; 'g (25) Àw i;j;s þ w i;j;t 1 À g s;t 8 i; j 2 f1; . . . ; Ng; s; t 2 f1; . . . ; 'g (26) Note we extend the notation of breakpoint appearance indicator w i;j;b to have w j;b ¼ P N i¼1 w i;j;b be 1 if breakpoint b appears at node v j and 0 otherwise.
## Ancestry condition for non-disappearing svs
We next impose the two-state perfect phylogeny ancestry condition as described in [bib_ref] Reconstruction of clonal trees and tumor composition from multi-sample sequencing data, El-Kebir [/bib_ref] for the appearance of breakpoints. For any breakpoint s that appears as an ancestor to breakpoint t, the total fraction of cells with breakpoint s must be larger than the fraction with breakpoint t so long as breakpoint s never subsequently disappears. To enforce this, the fraction of cells / p;b containing breakpoint b in sample p is defined as
[formula] / p;b ¼ X N k¼1 u p;k Á c k;b 8 p 2 f1; . . . ; mg; b 2 f1; . . . ; sg(27) [/formula]
We then must define a few variables to force / p;s ! / p;t if breakpoint s appears before breakpoint t and is never subsequently lost. Let v i be the ith node in the phylogeny and v i 0 v j denote that node v i is an ancestor of v j . We first define ancestor variables a i;j 2 f0; 1g as 1 if v i 0 v j and 0 otherwise for all i; j 2 f1; . . . ; Ng. Linearly define a i;j by root v N is ancestor to all nodes a N;j ¼ 1 8 j 2 f1; . . . ; N À 1g (28) root v N has no ancestors
[formula] a i;N ¼ 0 8 i 2 f1; . . . ; Ng(29) [/formula]
any parent is an ancestor a i;j ! e i;j 8 i; j 2 f1; . . . ; Ng
child gets parent's ancestor profile a g;j ! e i;j þ a g;i À 1 8 i; j 2 f1; . . . ; Ng; g 2 f1; . . . ; i À 1; i þ 1; . . . ; Ng (31) a g;j 1 À e i;j þ a g;i 8 i; j 2 f1; . . . ; Ng; g 2 f1; . . . ; i À 1; i þ 1; . . . ; Ng (32)
Next, define the number of descendants to node v i with at least one copy of breakpoint b as d i;b for all i 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g. To linearly define d i;b , define temporary binary variables x i;j;b 2 f0; 1g for Equation (33) through Equation (36) for all i; j 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g to be 1 if a i;j and c j;b and zero otherwise.
[formula] x i;j;b ! a i;j þ c j;b À 1 8 i; j 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g(33) [/formula]
x i;j;b a i;j 8 i; j 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g (34)
[formula] x i;j;b c j;b 8 i; j 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g (35) d i;b ¼ X N j¼1 x i;j;b 8 i 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g(36) [/formula]
Define temporary binary variables y i;b 2 f0; 1g for Equation (37) through Equation (39) to be 0 to be zero if all descendants of node v i contain at least one copy of breakpoint b and 1 otherwise.
[formula] y i;b ¼ X N j¼1 a i;j À d i;b 8 i 2 f1; [/formula]
. . . ; Ng; b 2 f1; . . . ; 'g (37) y i;b ¼ bin y i;b À Á 8 i 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g
Define temporary binary variable z i;j;s;t 2 f0; 1g for Equations (39) and (40) to be 0 only if breakpoint s appears at node v i , breakpoint t appears at node v j , node v i is an ancestor to node v j and breakpoint s never disappears. z i;j;s;t ¼ 3 À w i;s À w j;t À a i;j þ y i;s 8 i; j 2 f1; . . . ; Ng; s; t 2 f1; . . . ; 'g
Finally, apply the condition that the fraction of cells / p;s containing breakpoint s in sample p must be larger than the fraction of cells / p;t containing breakpoint t in sample p if breakpoint s appears in an ancestor to the node where breakpoint t appears and breakpoint s is never lost in any descendant (no descendant has copy number 0 for breakpoint s).
[formula] / p;s ! / p;t À 1 þ X N i¼1 X N j¼1 [/formula]
1 À z i;j;s;t À Á 8 p 2 f1; . . . ; mg; s; t 2 f1; . . . ; 'g
Note that P N i¼1 P N j¼1 1 À z i;j;s;t À Á can only take on values 0 or 1 since breakpoint appearance indicator w i;s and w j;s can only be both 1 at most once across all i, j. This means the condition / p;s ! / p;t only holds when breakpoint s appears before breakpoint t and never subsequently disappears. Note the ancestry condition is implied by but weaker than the sum condition described in [bib_ref] Reconstruction of clonal trees and tumor composition from multi-sample sequencing data, El-Kebir [/bib_ref] , but can similarly be enforced by linear constraints.
## Structural variant and segment consistency
Since each breakpoint belongs to exactly one segment, we define the copy number of each segment containing a breakpoint b and constrain it so a breakpoint's copy number never exceeds that of its containing segment:
[formula] c k;b c k;b ¼ X r s¼1 q b;s Á c k;sþ' 8 k 2 f1; . . . ; Ng; b 2 f1; . . . ; 'g(41) [/formula]
where input q b;s 2 f0; 1g is 1 if segment s contains breakpoint b. P r s¼1 q b;s ¼ 1 as each breakpoint belongs to a single segment. We similarly define w p;b directly from the input to be the mixed copy number of the segment containing breakpoint b.
[formula] w p;b ¼ X r s¼1 q b;s Á f p;sþ' 8 p 2 f1; . . . ; mg; b 2 f1; . . . ; 'g p p;b ¼ f p;b w p;b 8 p 2 f1; . . . ; mg; b 2 f1; . . . ; 'g [/formula]
Intuitively, the ratio p p;b of the mixed copy number of a breakpoint to the mixed copy number of the segment containing that breakpoint should be maintained in the integer output as this preserves the difference in mutation types (duplication, deletion). To penalize for discrepencies between the inferred ratio of breakpoint and its segment copy number given p p;b , we incorporate the following quantity into our objective function:
[formula] p p;b À P N k¼1 u p;k Á c k;b À Á P N k¼1 u p;k Á c k;b À Á [/formula]
To convert this from a ratio to units of copy numbers, we re-arrange the expression and define S for the final term in the objective function Equation (1) to be
[formula] z p;b ! p p;b Á X N k¼1 u p;k Á c k;b À Á À X N k¼1 u p;k Á c k;b À Á 8 p 2 f1; . . . ; mg; b 2 f1; . . . ; 'g(42)z p;b ! Àp p;b Á X N k¼1 u p;k Á c k;b À Á þ X N k¼1 u p;k Á c k;b À Á 8 p 2 f1; . . . ; mg; b 2 f1; . . . ; 'g(43) [/formula]
S
[formula] ¼ X m p¼1 X ' b¼1 z p;b(44) [/formula]
In this way, increased emphasis is placed on the relationship between segments and breakpoints. The solution for C and T is found by minimizing Equation (1) subject to constraints Equation (2) through Equation (44).
# Results
## Simulated data
To validate accuracy of the method on data of known ground truth, we assess accuracy in inference of copy number profiles across clones. For each such test, we generate a copy number matrix C tru containing breakpoints and segments, mix this matrix with a mixture fraction matrix U tru to get the mixed copy number matrix (C tru  U tru ! F), run our deconvolution algorithm and compare the inferred copy number matrix C inf with the original true copy number matrix C tru . We score our result as the L1 distance (jC tru À C inf j) between copy number matrices after a maximum matching between copy number profiles (for clones).
To generate C tru , we simulated mutation data varying the expected number of mutations l, number of samples m and number cell types n. For each triplet (l, m, n), five synthetic patients were generated. Reported scores are averaged across those five patients. For each run of the simulation, we generated a binary tree T with n leaves and a random topology. Mutations were assigned so that the expected numbers of mutations across all edges in each tree are equal. We start with a genomic profile for the root (assumed to be a normal diploid cell containing no structural variants) and progressively added a Poisson-distributed number of mutations across each edge down to the leaves. Initially, the root node contains three pairs of homologous chromosomes of the same lengths as human chromosomes 1-3. To generate mutations, a central location is uniformly chosen across all chromosomes, then a mutation size is sampled from an exponential distribution, with expectation equal to the mean structural variant size found across 59 TCGA samples (approximately 5 745 000 base pairs). The mutation type is uniformly randomly selected to be either a tandem duplication, deletion, or inversion. From the generated tree, we obtain a copy number matrix C tru . We then create a cell type mixture matrix U tru by uniformly randomly assigning cell type fractions such that the fraction of all cell types in each sample sums to 1. U tru and C tru are subsequently multiplied to generate mixed copy number matrix F.
Since there is no method for validating how accurate the choice of regularization terms k 1 and k 2 are on real data, we define empirical values for these terms based on each sample and show they perform well on simulated data. We choose regularization terms k 1 and k 2 empirically from the data to be k 1 ¼ 'þr ' Á m N and k 2 ¼ 'þr ' . This allows the maximum error in the jF À UCj term in the minimization, which is m Á ' þ r ð ÞÁc max , to equal the maximum errors in k 1 R and k 2 S terms, which are ' Á N Á c max and ' Á m Á c max , respectively. To show these empirical definitions do as well as iteratively choosing the hyperparameters, we test on simulated data generated for n ¼ 3 leaves, m ¼ 3 samples and l ¼ 50 mutations as this produces approximately 100 breakpoints, a value comparable to the average number of breakpoints found in real, TCGA samples. To ensure consistency in scoring, we generate five simulated patients with exactly 99 segments (not 100 since we have an odd number of chromosomes) and report the mean L1 distance between copy number segment matrices across the n ¼ 3 leaves. shows that automatically selecting hyperparameters k 1 and k 2 (solid green curve) leads to very good performance relative to that seen across a scan of possible parameter values (dotted blue curve), suggesting the automated parameter inference is effective. Both outperform the algorithm when excluding the regularization terms (dashed red curve), indicating the usefulness of including phylogenetic cost and breakpoint-segment consistency into the model. To further assess the novel value of including the SV phylogeny constraints in our model, we removed all phylogenetic constraints as well as structural variants from our model and found the results to be nearly the same as those when excluding both regularization terms (mean score of 480.6 and 487.0, respectively). This result further demonstrates the value of simultaneous SV phylogenetic inference and deconvolution even when the method is judged solely on deconvolution quality.
We further evaluated the effectiveness of the methods by their ability to identify the correct phylogenetic trees. We assessed accuracy using Robinson Foulds (RF) distance, which measures the number of bi-partitions differing between two trees on a common set of nodes, between the true and inferred trees for each of the simulated test cases. We found that three of the five inferred trees had identical topology to the true trees (RF distance 0). The remaining two trees differed solely by swapping the root node with one leaf neighbor of the root (RF distance 2). While the trees are too simple and few in number to attach any significance to this result, it does demonstrate that the method is generally accurate at inferring correct or nearcorrect phylogenies despite some error in deconvolution of the nodes of the trees.
## Tcga data
We next apply the methods to a selection of TCGA breast cancer (BRCA) samples (The Cancer Genome Atlas Network, 2012), restricting analysis to a sub-set of 59 samples for which WGS data was available. Of these, 31 ran successfully within a prescribed run time limit of 2 days, while 28 with the highest SV counts timed out before completion or required more memory than was available to us (128 Gb of RAM). Since there is no known ground truth for these samples, we cannot assess their individual accuracies. Nonetheless, they provide some basis for analysis of trends across samples. Space does not permit us to display all observed trees, so for purposes of illustration we classify them into seven observed topologies (A-G), shown in , with frequencies of occurrence shown in . None of the inferred trees are purely linear, consistent with a model of significant sub-clonal heterogeneity rather than a simple sequential model of clonal progression. Quantitation by several measures of heterogeneity, as shown in [fig_ref] Figure 6: Violin plots quantifying heterogeneity of the 31 TCGA trees [/fig_ref] , likewise suggests a wide diversity among samples. The data is suggestive of a possible clustering into distinct low-diversity and high-diversity sub-clusters, but with substantial overlap between clusters.
# Discussion
We have developed a new method for automated joint deconvolution and phylogeny inference of tumor genomic data designed to address the important unsolved problem of describing progression via SVs. We specifically learn a model encompassing CNAs and SVs of major clones, mixture fractions of these clones across samples and a phylogenetic tree relating the clones. We pose the model inference problem to balance the likelihood of sequence read data with respect to copy numbers and observed breakpoints against the evolutionary cost of the phylogenetic tree. We solve the resulting model via a coordinate descent algorithm posed as a pair of MILPs. We demonstrate that the method can accurately and efficiently reconstruct . Deconvolution quality on simulated data for varying hyperparameters k 1 and k 2 . Accuracy is scored by the sum of L1 distances between the true C tru and inferred C inf copy number matrices for all segments in each leaf after maximum matching (lower means better performance). The reported score is an average across five simulated datasets each containing 99 segments with standard errors shown as error bars. The dotted blue line is the score when k2 ¼ 0 is held constant and k 1 varies from 0.01, 0.05, 0.25, 1.25, 6.25 across the x-axis. The solid green line shows the score when hyperparameters are automatically chosen based on the number of SVs and CNAs in the dataset, while the dashed red line shows scores when only the first term in the objective function, corresponding to accuracy of copy number deconvolution, is used clonal populations and phylogenetic histories from simulated tumor data. Application to WGS data from the TCGA shows the method to be effective on real data supportive of a range of tree topologies and complexities. This work provides a proof-of-concept demonstration of the feasibility of more comprehensively modeling the important role of SVs in tumor evolution, but also suggests a number of avenues for future work. Our methods currently rely on a sometimes costly and potentially sub-optimal model fitting algorithm, and further algorithmic advances might plausibly lead both to greater efficiency and improved solution quality. In particular, there are currently practical limits on the total SV counts the method can handle without excessive run time and memory usage. While most of the TCGA BRCAs considered fell within those limits, a significant minority did not. The method also makes some assumptions about its input data that may not always be satisfied, particularly that base-pair resolution SV breakpoints can be inferred accurately and will form sufficiently rarely that we can assume a perfect phylogeny of SVs. While we argue that this is a sounder assumption for SVs than for SNVs, one might nonetheless anticipate some violations either due to truly recurrent mutation or to errors in breakpoint assignment that might lead to conflation of distinct breakpoints. Extending the model to allow for tolerance of such violations of the SV perfect phylogeny assumption would thus be a good avenue for future work. Furthermore, our work focuses only on the sub-problem of handling SVs (and associated CNAs), and will likely benefit from incorporating other variant types, most notably SNVs but also potentially expression, methylation, or other markers of cell state. In addition, biotechnology for data generation is continuing to advance, with growing numbers of computational methods taking advantage of single-cell sequence or long read technologies that can provide direct single-cell readouts of SNV or CNA data. SV detection is problematic for all current single-cell technologies, and we can anticipate value in combining single-cell methods with bulk deconvolution methods such as ours for SVs. Finally, the present work has focused only on the development of the new technology and its validation. The ultimate value of the work will lie in bringing SV-aware phylogenetics to diverse patient cohorts, to begin to develop a comprehensive understanding of the landscape of SV variation in tumor progression and its implications for patient prognosis and treatment.
[fig] Figure 1: Example genomes before and after segmental deletion and duplication. Top images are the reference genome while bottom images are the genome after deletion/duplication. Each colored box represents a single base pair and base pairs between two vertical orange lines represent segments. The letters below a base pair identify the position of that base pair in the reference genome. Assume this is example holds for any single chromosome labeled z.(a) Shows a deletion of segment 2 (base pairs B through C) producing structural variant ]z: A], [z: D[. The copy number of each of the mated breakpoints (c z:A and c ½z:D½ ) and the copy number of each of the segments containing these breakpoints (c z:A and c ½z:D½ ) are all 1 (c z:A ¼ c ½z:D½ ¼ c z:A ¼ c ½z:D½ ¼ 1). (b) Shows a duplication of segment 2 producing structural variant [z: B[,]z: C]. The copy number of each of the mated breakpoints is 1 (c ½z:B½ ¼ c z:C ¼ 1) while the copy number of each of the segments containing breakpoints is 2 (c ½z:B½ ¼ c z:C ¼ 2) [/fig]
[fig] Figure 4, Figure 5: Tree topologies observed across 31 TCGA BRCA samples, grouped into seven categories (A-G) Histogram of occurrences of tree topologies across 31 TCGA BRCA samples [/fig]
[fig] Figure 6: Violin plots quantifying heterogeneity of the 31 TCGA trees. (a) Number of unique clones. (b) Diversity, defined in terms of the clonal frequency vector u p;k as 1 À P N k¼1 u p;k . (c) Number of breakpoints [/fig]
|
Pros and Cons of Skeletal Medications in the COVID-19 Era
Purpose of Review This review provides an overview regarding osteoporosis therapies during the COVID-19 pandemic. Recent Findings The COVID-19 pandemic has disrupted treatments for osteoporosis and resulted in decreased adherence particularly for parenteral regimens. Osteoporosis medications are safe and effective during the pandemic and should be continued whenever possible. Bisphosphonates have long-lasting effects on bone turnover such that delays in their administration are unlikely to be harmful to skeletal health. In contrast, interruption of denosumab treatment is strongly discouraged because of rapid loss of bone mass and an associated increased risk for rebound vertebral fractures. When osteoanabolic treatments cannot be continued during the pandemic, change to an oral bisphosphonate is advised. Preclinical data suggest possible beneficial effects of some therapies against COVID-19, but require validation in clinical studies. Vitamin D deficiency is associated with a more severe COVID-19 clinical course but data supporting improvements in outcomes with vitamin D supplementation are lacking. Summary The impact of the COVID-19 pandemic on long-term bone health remains unknown but focused interventions to ensure osteoporosis treatment initiation/maintenance should Curr Treat Options in Rheum (2022) 8:56 -69 Pros and Cons of Skeletal Medications in the COVID-19 Era, E. Tsourdi and M. T. Drake be implemented. Future studies are needed to determine whether osteoporosis medications have an impact on SARS-CoV-2 pathophysiology and COVID-19 clinical outcomes.
# Introduction
At the time of this writing, SARS-CoV-2, the causative virus for the COVID-19 pandemic, has infected more than 554 million people and resulted in over 6.3 million deaths worldwide . The COVID-19 pandemic has disrupted routine medical care globally, with the reallocation of personnel and infrastructure for the acute management of patients suffering from COVID-19 having taken its toll on the treatment of many chronic conditions including musculoskeletal diseases. In addition, social distancing and isolation have led to reduced physical activity and increased sedentary lifestyles, both of which are associated with sarcopenia, osteoporosis, and an increased incidence of fragility fractures [bib_ref] Sarcopenia during COVID-19 lockdown restrictions: long-term health effects of short-term muscle loss, Kirwan [/bib_ref] [bib_ref] Osteoporotic fractures in the time of COVID-19, Bhattacharyya [/bib_ref]. In view of the impact of the pandemic on musculoskeletal health, both professional societies and bone specialists have issued recommendations for the optimization of osteoporosis treatment in this era - , 6- , . In this review, we examine the current literature on pharmacologic and non-pharmacologic regimens for patients suffering from osteoporosis in the COVID-19 era and discuss putative "pros" and "cons" of such agents and approaches in this setting.
## Vitamin d and calcium supplementation
The importance of vitamin D for the regulation of calcium and phosphate homeostasis is well established. Vitamin D insufficiency leads to impaired calcium absorption, secondary hyperparathyroidism, osteoporosis, and fragility fractures [bib_ref] Skeletal and extraskeletal actions of vitamin D: current evidence and outstanding questions, Bouillon [/bib_ref] , while daily supplementation of 800-2000 IU cholecalciferol co-administered with calcium reduces hip fracture rates by 15-30% and other non-vertebral fractures by 20% [bib_ref] Impact of vitamin D supplementation on falls and fractures-a critical appraisal of..., Chakhtoura [/bib_ref]. Whether the beneficial effects of vitamin D extend beyond musculoskeletal health has been the subject of research from the earliest stages of the pandemic. Theoretically, additional non-musculoskeletal benefits seemed possible, since vitamin D is known to modulate both innate and adaptive immunity, is involved in inflammatory processes, and may decrease the risk of infection [bib_ref] The trinity of COVID-19: immunity, inflammation and intervention, Tay [/bib_ref] [bib_ref] Vitamin D supplementation to prevent acute respiratory tract infections: systematic review and..., Martineau [/bib_ref]. Numerous observational studies have investigated the association of circulating vitamin D levels with a number of outcomes including the severity of COVID-19 disease, need for hospitalization, duration of hospital stay, intensive care unit (ICU) admission and duration of stay, time to symptomatic recovery and time to seronegative conversion, risk of complications, and mortality rate following COVID-19 infection. Two recent narrative reviews [bib_ref] Vitamin D and coronavirus disease 2019 (COVID-19): rapid evidence review, Raisi-Estabragh [/bib_ref] , as well as a systematic review and meta-analysis , have presented these findings. To summarize the evidence, observational studies have shown that serum 25(OH)D concentrations < 20 ng/ mL are linked to a higher probability of SARS-CoV-2 infection and are associated with increased risk for ICU admission, mechanical ventilation, and COVID-19 mortality [6, . It is noteworthy, however, that observational studies have a high risk of bias and are limited by confounding, i.e., the relationship between vitamin D status and other comorbidities also linked to adverse outcomes following COVID-19 infection such as diabetes mellitus, obesity, cardiovascular and respiratory diseases, and malignancy [bib_ref] Clinical features of patients infected with 2019 novel coronavirus in Wuhan, Huang [/bib_ref]. In addition, since vitamin D is an inverse acute phase reactant with lower levels during times of increased physiologic stress [bib_ref] Vitamin D: a negative acute phase reactant, Waldron [/bib_ref] , observational studies in this setting are potentially prone to reverse causation [bib_ref] Vitamin D and coronavirus disease 2019 (COVID-19): rapid evidence review, Raisi-Estabragh [/bib_ref]. Given such caveats associated with association studies, randomized-controlled studies (RCTs) of vitamin D supplementation in the context of COVID-19 infection are more likely to provide evidence of potential benefits [bib_ref] COVID-19 and high-dose VITamin D supplementation TRIAL in high-risk older patients (COVIT-TRIAL):..., Annweiler [/bib_ref]. Two RCTs which used single-bolus high-dose vitamin D versus placebo (540,000 IU and 200,000 IU respectively), however, reported no survival benefit or differences in secondary outcomes related to hospitalization [bib_ref] Early high-dose vitamin D 3 for critically Ill, vitamin D-deficient patients, Ginde [/bib_ref] [bib_ref] Effect of a single high dose of vitamin D3 on hospital length..., Murai [/bib_ref] , findings which were confirmed in a meta-analysis of RCTs ]. Moreover, a Mendelian randomization study based on participants from eight distinct genome-wide association studies (GWAS) showed no effect of alleles associated with 25(OH)D levels on COVID-19 susceptibility, indicating no genetic evidence to support vitamin D supplementation . In conclusion, since existing data do not support a clear and strong cause-effect relationship between vitamin D status and COVID-19 outcomes, no recommendations for supraphysiologic doses of vitamin D supplementation to treat or prevent COVID-19-related complications can be issued. In fact, supplementation with high-dose bolus vitamin D may even have a negative impact on fall and fracture outcomes [bib_ref] Annual high-dose oral vitamin D and falls and fractures in older women:..., Sanders [/bib_ref] [bib_ref] Monthly high-dose vitamin D treatment for the prevention of functional decline: a..., Bischoff-Ferrari [/bib_ref]. It is notable that there are currently > 40 RCTs in ClinicalTrials.gov examining the role of vitamin D in COVID-19. However, until more information becomes available, it seems reasonable to tailor vitamin D supplementation according to osteological society recommendations, i.e., daily supplementation with 800-2000 IU to maintain 25(OH) D concentrations ≥ 30 mg/mL [bib_ref] Commentary: myths and facts on vitamin D amidst the COVID-19 pandemic, Chakhtoura [/bib_ref]. Finally, patients infected with COVID-19 commonly present with hypocalcemia. Although hypocalcemia has been shown to correlate with higher rates of adverse outcomes [bib_ref] Hypocalcemia in COVID-19: prevalence, clinical significance and therapeutic implications, Di Filippo [/bib_ref] [bib_ref] Hypocalcemia is associated with severe COVID-19: a systematic review and meta-analysis, Martha [/bib_ref] , it is unclear whether low serum calcium levels constitute a COVID-19-specific effect or rather serve as a general marker of severe illness. No RCTs on the effects of supplementation with calcium alone on the course of COVID-19 have been published to date, although a retrospective cross-sectional study of 490 patients treated with calcium supplements for osteoporosis revealed a decreased risk of COVID-19 infection (RR = 0.64, 95% CI 0.37, 1.12) in this population [bib_ref] The ability of a single BMD and fracture history assessment to predict..., Blanch-Rubió [/bib_ref]. A phase I/II study to investigate the tolerability and safety of calcium carbonate versus placebo concomitant with the best available treatment (BAT) is currently in the recruitment phase (NCT04900337).
## Pharmacological treatment of osteoporosis
There is a well-recognized gap in the treatment of osteoporosis that extends worldwide, with just one-fifth of patients provided with treatment after a hip fracture, the time when patients are most at risk for another fracture . The discrepancy between a diagnosis of osteoporosis and the provision of optimal care has been further exacerbated during the COVID-19 pandemic, with the treatment of osteoporosis appearing to rank comparatively lower when compared to other competing clinical priorities. As evidence of this, a number of national and international surveys have documented considerable delays and poor adherence to treatment, especially with regard to parenteral osteoporosis treatments, both as a result of reallocation of personnel to acute services, and because affected patients have been unwilling or unable to attend care facilities for scheduled treatments [bib_ref] Osteoporosis care during the COVID-19 pandemic in the Netherlands: a national survey, Peeters [/bib_ref] [bib_ref] Poor adherence to parenteral osteoporosis therapies during COVID-19 pandemic, Kocijan [/bib_ref] [bib_ref] Diagnosis and management of osteoporosis during COVID-19: systematic review and practical guidance, Hampson [/bib_ref] [bib_ref] How has COVID-19 affected the treatment of osteoporosis? An IOF-NOF-ESCEO global survey, Fuggle [/bib_ref]. Importantly, osteoporosis treatments are not known to increase the risk for adverse events from COVID-19, and preclinical data suggest that some therapies may have favorable effects. Below, we discuss what is known regarding commonly used medications for the treatment of osteoporosis with respect to COVID-19 susceptibility, pathogenesis, and clinical course as well as the impact of the COVID-19 pandemic on individual osteoporosis therapeutic regimens.
## Estrogen and selective estrogen receptor modulators (serms)
Hormone replacement therapy (HRT) improves bone mineral density (BMD) and reduces fracture risk in women with osteoporosis. Low-dose transdermal HRT is less likely to cause breast cancer, endometrial hyperplasia, coronary artery disease (CAD), and venous thromboembolism (VTE), which have been linked to standard-dose oral HRT regimens [bib_ref] Estrogen therapy for osteoporosis in the modern era, Levin [/bib_ref]. HRT is therefore used for the primary prevention and treatment of osteoporosis in appropriate candidates [bib_ref] Estrogen therapy for osteoporosis in the modern era, Levin [/bib_ref]. Interestingly, males are more susceptible than females to SARS-CoV-2 infection, and older men with comorbidities have a greater risk of developing severe COVID-19 disease [bib_ref] Gender differences in patients with COVID-19: focus on severity and mortality, Jin [/bib_ref]. Although environmental factors are undoubtedly important, it is possible that such differences may be at least partially attributable to sex-specific genetic and hormonal factors [bib_ref] Bone metabolism in SARS-CoV-2 disease: possible osteoimmunology and gender implications, Salvio [/bib_ref]. Innate and adaptive immune responses to many viral infections differ between males and females, with evidence pointing to the more favorable clearance of viral pathogens in women [bib_ref] Coronavirus COV-19/SARS-CoV-2 affects women less than men: clinical response to viral infection, Conti [/bib_ref] , as well as reduced expression of ACE2, a receptor facilitating the entry of SARS-CoV-2 into target cells [bib_ref] Sex-specific SARS-CoV-2 mortality: among hormone-modulated ACE2 expression, risk of venous thromboembolism and..., Vignera [/bib_ref] , and negative correlations between estradiol and IL-6, IL-2R, and interferon γ-inducible protein 10 [bib_ref] Association of circulating sex hormones with inflammation and disease severity in patients..., Dhindsa [/bib_ref]. In this vein, it is noteworthy that there are currently two RCTs (NCT04539626, NCT04865029) underway designed to investigate the effects of a short course of systemic treatment with estradiol and progesterone in addition to BAT on COVID-19 symptoms in patients with mild/moderate disease.
Selective estrogen receptor modulators (SERMs) are characterized by mixed agonism/antagonism effects at the estrogen receptor (ER) of specific tissues, with beneficial estrogenic effects in some organs and avoidance of adverse or off-target effects in other tissues. Raloxifene, a widely used SERM, has been approved for the prevention and treatment of osteoporosis in postmenopausal women, as well as to reduce the risk of invasive breast cancer in postmenopausal women [bib_ref] Effects of raloxifene on fracture severity in postmenopausal women with osteoporosis: results..., Siris [/bib_ref] [bib_ref] Continuing outcomes relevant to Evista: breast cancer incidence in postmenopausal osteoporotic women..., Martino [/bib_ref]. The most common adverse effects associated with raloxifene therapy are hot flushes, nausea, and vomiting, while the most significant adverse effects are venous thromboembolic events [bib_ref] Risk-benefit profile for raloxifene: 4-year data from the Multiple Outcomes of Raloxifene..., Barrett-Connor [/bib_ref]. Although raloxifene is not approved for use in men, existing data demonstrate tolerability as well as efficacy in men in terms of bone turnover markers and lipid metabolism [bib_ref] Effects of raloxifene, a selective estrogen receptor modulator, on bone turnover markers..., Doran [/bib_ref]. Of note, an older study in humans had shown that raloxifene decreases levels of both IL-6 and tumor necrosis factor α (TNFα), thus highlighting its anti-inflammatory effects [bib_ref] Raloxifene modulates interleukin-6 and tumor necrosis factor-alpha synthesis in vivo: results from..., Gianni [/bib_ref]. In addition, in vitro studies have confirmed inhibition of viral replication and/or infection against HCV [bib_ref] Selective estrogen receptor modulators inhibit hepatitis C virus infection at multiple steps..., Murakami [/bib_ref] , HBV [bib_ref] Screening and identification of compounds with antiviral activity against hepatitis B virus..., Lamontagne [/bib_ref] , and the Zika virus [bib_ref] Identification of estrogen receptor modulators as inhibitors of flavivirus infection, Eyre [/bib_ref] , while the addition of raloxifene to BAT of interferon 2α/ribavirin proved beneficial for the treatment of chronic hepatitis C infection in humans [bib_ref] Raloxifene hydrochloride is an adjuvant antiviral treatment of postmenopausal women with chronic..., Furusyo [/bib_ref].
Recently, raloxifene was identified in silico among 400,000 candidate molecules and was preselected to proceed with in vitro testing among 7000 molecules via the supercomputing platform 'ExaSCale smart pLatform Against paThogEns' (EnsEXSCALATE) as a promising molecule with antiviral activity to treat oligosymptomatic COVID-19 disease . According to data generated via this platform, raloxifene was predicted to bind to relevant SARS-CoV-2 proteins, while also having a higher pulmonary distribution relative to other SERMs. Notably, the relatively low pharmacologic concentrations in the lung have been a major limitation of currently used antiviral medications used to treat COVID-19 respiratory complications [bib_ref] Repurposing the estrogen receptor modulator raloxifene to treat SARS-CoV-2 infection, •• Allegretti [/bib_ref]. It is also noteworthy that no clinically relevant interactions have been described between raloxifene and frequently co-administered drugs . A multicenter placebocontrolled phase 2 study to investigate the efficacy and safety of two different doses of oral raloxifene in patients with early diagnosis of pauci-symptomatic COVID-19 was recently completed, with the results pending (NCT05172050).
In conclusion, both HRT and raloxifene have pleiotropic effects which may affect COVID-19 severity, including modulating ACE2 expression which might impact both infection risk and clinical course, and via inhibition of IL-6 signaling which could potentially mitigate cytokine storm. However, as both HRT and raloxifene are associated with a modest increase in thrombotic risk, the results of ongoing RCTs investigating their efficacy and safety in the setting of COVID-19 will be particularly important. It will also be of interest to ascertain whether the effects of HRT and raloxifene differ between males and females.
## Bisphosphonates
Bisphosphonates are the most commonly used medications for the treatment of primary and secondary osteoporosis in both women and men. After bisphosphonate discontinuation, bone turnover markers which had been suppressed eventually revert to baseline levels, and bone mineral density (BMD) remains stable or gradually diminishes over years [bib_ref] The effect of 3 versus 6 years of zoledronic acid treatment of..., Black [/bib_ref]. Bisphosphonates' antiresorptive effects last after treatment is discontinued as a result of the strong affinity of bisphosphonates for binding to hydroxyapatite. This property holds especially true for the aminobisphosphonates alendronate and zoledronate. There is also some evidence for long-term antifracture effects following bisphosphonate cessation, although these findings are less well supported [bib_ref] Fracture risk remains reduced one year after discontinuation of risedronate, Watts [/bib_ref] [bib_ref] Efficacy of continued alendronate for fractures in women with and without prevalent..., Schwartz [/bib_ref].
A number of studies have investigated whether bisphosphonate use for the treatment of osteoporosis is associated with COVID-19 outcomes. A nationwide, multicenter, retrospective study from Turkey during the first wave of the COVID-19 pandemic compared patients diagnosed with COVID-19 who were treated with anti-osteoporosis medications with COVID-19 positive patients who did not receive bone active drugs. Of the 1997 women included, 89.5% were treated with bisphosphonates. This study revealed that hospitalization risk, ICU admissions, and mortality rates were not influenced by bisphosphonates or any other drug category [bib_ref] No association of anti-osteoporosis drugs with COVID-19-related outcomes in women: a nationwide..., Atmaca [/bib_ref]. In comparison, a separate study showed that treatment with zoledronate was associated with a 40% decreased risk of COVID-19 infection, whereas oral bisphosphonates had no effect on COVID-19 incidence . A third population-based retrospective observational cohort study comprising 9% of the Italian population also showed no benefit of oral bisphosphonate treatment with regard to the incidence of COVID-19 hospitalization, need for ICU care, and mortality rates . Although these findings need to be verified in further studies, the potent aminobisphosphonate zoledronate might differentially modulate aspects of COVID-19 susceptibility and clinical course.
It is notable that earlier studies not performed in patients with COVID-19 infection had reported lower mortality with zoledronate treatment, with postulated mechanisms including a reduction in cardiovascular risk and fracture prevention, but also enhanced regulation of the immune system and a lower incidence of pneumonia [bib_ref] Zoledronic acid and clinical fractures and mortality after hip fracture, Lyles [/bib_ref] [bib_ref] Potential mediators of the mortality reduction with zoledronic acid after hip fracture, Colón-Emeric [/bib_ref] [bib_ref] Bisphosphonates and mortality: confounding in observational studies?, Bergman [/bib_ref]. These immunomodulatory effects have been ascribed to an increase in natural killer cells as well as a stimulatory effect on γδ cells T cells in response to zoledronate [bib_ref] Zoledronic acid-induced expansion of γδ T cells from early-stage breast cancer patients:..., Sugie [/bib_ref] [bib_ref] DC-like cell-dependent activation of human natural killer cells by the bisphosphonate zoledronic..., Nussbaumer [/bib_ref]. In this context, it is worth noting the distinct profile involving selective expansion of γδ T cell populations was described in the aftermath of the 2003 SARS outbreak [bib_ref] Anti-severe acute respiratory syndrome coronavirus immune responses: the role played by V..., Poccia [/bib_ref] , while T-cells, including γδ cells T cells, are depleted in patients suffering from severe COVID-19 . In this context, zoledronate could plausibly hinder endosomal homeostasis, a process which appears to be pivotal for SARS-CoV-2 survival [62- ]. In terms of mechanism of action, zoledronate inhibits prenylation of small GTPases, proteins necessary for endosomal trafficking in osteoclasts [bib_ref] Protein geranylgeranylation is required for osteoclast formation, function, and survival: inhibition by..., Coxon [/bib_ref]. Given that osteoclasts and dendritic cells arise from a common precursor and share a number of functions [bib_ref] Identification of the common origins of osteoclasts, macrophages, and dendritic cells in..., Xiao [/bib_ref] [bib_ref] Bone marrow microenvironment controls the in vivo differentiation of murine dendritic cells..., Wakkach [/bib_ref] , zoledronate may have similar effects on endosomal exocytosis of SARS-CoV-2 infected dendritic cells [bib_ref] This review highlights the distinct mechanism through which NBPs such as zoledronate..., Brufsky [/bib_ref]. It remains to be seen whether clinical trials will be able to support these putative effects.
To summarize, while bisphosphonates (particularly zoledronate) are theoretically capable of ameliorating the immune host status to provide protection against SARS-CoV-2 infections, clinical studies to this effect are lacking. Current data do not support the hypothesis that oral bisphosphonates can prevent COVID-19 or mitigate its severity; however, oral bisphosphonates do not appear to increase infection risk. With regard to osteoporosis treatment, in patients in whom intravenous bisphosphonate treatment cannot be administered, delays of even several months are acceptable because of their long-lasting residence and maintenance of activity within the bone.
## Denosumab
Denosumab, a monoclonal antibody against the receptor activator of nuclear factor B ligand (RANKL), is a potent antiresorptive agent that suppresses bone turnover markers (BTMs), increases BMD, and significantly lowers fracture risk, with evidence for a good safety profile when provided every 6 months for up to 10 years [bib_ref] 10 years of denosumab treatment in postmenopausal women with osteoporosis: results from..., Bone [/bib_ref]. When denosumab is discontinued, however, its effects on BMD and BTMs are rapidly reversed. After 1-2 years after treatment discontinuation, BMD returns to pre-treatment baseline values, with BTMs exceeding baseline values within 3 months and thereafter remaining persistently elevated before slowly returning to baseline levels [bib_ref] Effects of denosumab treatment and discontinuation on bone mineral density and bone..., Bone [/bib_ref]. There is now good evidence that denosumab discontinuation may be associated with multiple vertebral fractures (VFx), with current expert recommendations counseling against denosumab discontinuation unless an alternative treatment is subsequently initiated . Given these considerations, it is evident that disruption of the standard denosumab every 6-month dosing administration schedule, as has increasingly been documented during the pandemic [bib_ref] Osteoporosis care during the COVID-19 pandemic in the Netherlands: a national survey, Peeters [/bib_ref] [bib_ref] Poor adherence to parenteral osteoporosis therapies during COVID-19 pandemic, Kocijan [/bib_ref] [bib_ref] Diagnosis and management of osteoporosis during COVID-19: systematic review and practical guidance, Hampson [/bib_ref] [bib_ref] How has COVID-19 affected the treatment of osteoporosis? An IOF-NOF-ESCEO global survey, Fuggle [/bib_ref] , can have devastating skeletal effects. As an example, a retrospective analysis of 768 patients receiving denosumab at a large hospital in Singapore during the COVID-19-first wave period revealed that adherence to treatment decreased significantly compared to the pre-COVID-19 period, with the odds of adherence increased if the treating physician was an endocrinologist [bib_ref] Adherence to dosing schedule of denosumab therapy for osteoporosis during COVID-19 lockdown:..., Chandran [/bib_ref].
A preliminary analysis of the same study showed that fractures on follow-up occurred less frequently in patients adherent to a 6-month treatment dosing interval [bib_ref] Adherence to dosing schedule of denosumab therapy for osteoporosis during COVID-19 lockdown:..., Chandran [/bib_ref]. Moreover, a retrospective analysis of 638 patients in China demonstrated that postponing denosumab treatment for 3 months (i.e., administration 9 months after the most recent previous dose) resulted in significant BMD decreases at the lumbar spine [bib_ref] Retrospective analysis of the effects of non-compliance with denosumab on changes in..., Huang [/bib_ref].
Regarding a possible influence of denosumab treatment on COVID-19 incidence, the study of Blanch-Rubió et al. showed a 40% decreased risk of COVID-19 infection in patients treated with denosumab, a similar finding to that found with zoledronate. Immune responses involving the RANK/ RANKL system include lymph node formation, lymphocyte differentiation, dendritic cell survival, and T-cell activation [bib_ref] Effects of RANKL-targeted therapy in immunity and cancer, Cheng [/bib_ref]. Furthermore, RANKL inhibition modifies immune cell profiles and decreases the release of pro-inflammatory cytokines [bib_ref] Denosumab: targeting the RANKL pathway to treat rheumatoid arthritis, Chiu [/bib_ref]. This attenuated inflammatory response may be beneficial during viral infections, as has previously been shown for osteoprotegerin, a decoy receptor for RANKL [bib_ref] Beyond bone remodeling-emerging functions of osteoprotegerin in host defense and microbial infection, Kobayashi-Sakamoto [/bib_ref]. On the other hand, a recent meta-analysis demonstrated an increased risk for ear, nose, and throat and gastrointestinal infections in patients treated with denosumab, albeit without a higher overall risk for any infection or for mortality . Moreover, survey studies have not identified an association between denosumab therapy and increased risk of COVID-19 infection .
In summary, patients should be strongly dissuaded from stopping denosumab due to concerns related to rapid loss of bone mass and increased risk for multiple vertebral compression fractures. When denosumab treatment continuation cannot be guaranteed within 7 months of the most recent prior injection, a temporary transition to an oral bisphosphonate should be recommended [bib_ref] Osteoporosis management in the era of COVID-19, Yu [/bib_ref]. Although data with regard to COVID-19 infection in patients treated with denosumab are scarce, it seems unlikely that denosumab therapy will aggravate the clinical course of COVID-19.
## Romosozumab
Romosozumab is a humanized monoclonal antibody directed against sclerostin. It has a dual mode of action, i.e., enhancement of bone formation with simultaneous suppression of bone resorption, and proven efficacy with regard to BMD increase and fracture reduction [bib_ref] A systematic review and metaanalysis of efficacy and safety of Romosozumab in..., Singh [/bib_ref]. However, like other bone anabolic agents, romosozumab discontinuation is associated with rapid bone loss within 1 year, with BTMs rising as early as 3 months from the time of the most recent previous dose if subsequent antiresorptive therapy is not provided [bib_ref] Effects of 24 months of treatment with romosozumab followed by 12 months..., Mcclung [/bib_ref]. However, transitioning from romosozumab to alendronate results in sustained BMD gains [bib_ref] Romosozumab or alendronate for fracture prevention in women with osteoporosis, Saag [/bib_ref]. Thus, if romosozumab is discontinued in the setting of the pandemic, it is prudent to transition to an oral bisphosphonate [bib_ref] Osteoporosis management in the era of COVID-19, Yu [/bib_ref] [bib_ref] Diagnosis and management of osteoporosis during COVID-19: systematic review and practical guidance, Hampson [/bib_ref]. To date, no studies have investigated the putative effects of treatment with romosozumab on SARS-CoV-2 infection. Nevertheless, recent evidence has shown that upregulation of the canonical Wnt/β-catenin pathway is associated with inflammation and cytokine storm in patients infected with COVID-19 [bib_ref] Wnt5a and Wnt11 as acute respiratory distress syndrome biomarkers for severe acute..., Choi [/bib_ref]. In these patients, the Wnt/β-catenin pathway is stimulated via transforming growth factor-β (TGF-β) and can cause pulmonary fibrosis and pulmonary infarctions . The impact of sclerostin, one of the major inhibitors of the Wnt/β-catenin pathway, on SARS-CoV-2 infections is unknown. Accordingly, it is currently difficult to predict whether romosozumab may influence the clinical course of COVID-19 infection.
## Teriparatide/abaloparatide
Teriparatide and the parathyroid hormone-related protein (PTHrP) analogue abaloparatide are osteoanabolic treatments with significant antifracture benefits. Teriparatide cessation is associated with progressive BMD loss over the course of 1 year [bib_ref] Effects of teriparatide treatment and discontinuation in postmenopausal women and eugonadal men..., Leder [/bib_ref] , with the antiresorptive treatment required following teriparatide discontinuation to avoid loss of the osteoanabolic effect [bib_ref] Enhancement of bone mass in osteoporotic women with parathyroid hormone followed by..., Rittmaster [/bib_ref] [bib_ref] One year of alendronate after one year of parathyroid hormone (1-84) for..., Black [/bib_ref]. Given that abaloparatide shares a similar biological profile to teriparatide, sequential treatment is equally applicable, with beneficial skeletal effects having been demonstrated for abaloparatide followed by alendronate [bib_ref] Fracture and bone mineral density response by baseline risk in patients treated..., Leder [/bib_ref]. Based on the above evidence, temporary transition to an oral bisphosphonate should be offered when either of these treatments must be discontinued during the pandemic [bib_ref] Osteoporosis management in the era of COVID-19, Yu [/bib_ref] [bib_ref] Diagnosis and management of osteoporosis during COVID-19: systematic review and practical guidance, Hampson [/bib_ref]. Although PTH receptors are located on various immune cells (neutrophils, B and T cells), clinical studies of the immunological effects of teriparatide have been limited by different PTH formulations (rat, bovine, and human) and have primarily been conducted in patients with renal impairment [bib_ref] Effects of parathyroid hormone on immune function, Geara [/bib_ref]. To date, there are insufficient data to predict whether teriparatide and abaloparatide may have an impact on SARS-CoV-2 infections.
# Conclusion
The COVID-19 pandemic has created many challenges for patients with musculoskeletal diseases and the long-lasting impact of SARS-CoV-2 on bone health remains unknown. The disruption of access to healthcare and medications is expected to result in poor disease control and an increase in fracture risk and fracture outcomes. Osteoporosis drugs are safe and effective and should be continued during the pandemic. Although limited preclinical data exist for the beneficial effects of raloxifene and zoledronate in COVID-19 infection, any salutary effects must be confirmed in robust clinical studies. While osteoporosis drugs have not been proven to prevent SARS-CoV-2 infections or alleviate COVID-19 severity, they do not appear to increase any risks associated with COVID-19 infection [fig_ref] Table 1: Pros and cons of skeletal medications in the COVID-19 eraAbbreviations [/fig_ref].
## Key points
- Although vitamin D deficiency is associated with more severe clinical outcomes in patients infected by COVID-19, clear evidence to support the improvement of clinical outcomes through vitamin D supplementation is lacking.
- HRT and raloxifene may have potential positive effects on infection risk and the clinical course of COVID-19 infection, but confirmatory studies are pending.
- Zoledronate is theoretically capable of modulating the immune host status and thereby protecting against SARS-CoV-2 infections.
- Denosumab discontinuation should be strongly discouraged, with a temporary transition to oral bisphosphonate therapy recommended when denosumab cannot continue to be provided per standard clinical dosing guidelines
# Declarations
## Conflict of interest
ET has received research funding from MSD, honoraria for lectures from Amgen, UCB, Shire, Kyowa Kirin, and educational grants from Shire and UCB. Matthew T. Drake declares that he has no conflict of interest. This ASBMR perspective provides evidence, where available, and clinical recommendations based on expert opinion, when data are lacking, for the general care of patients with osteoporosis in the COVID-19 era. Particular emphasis is placed on the transition from parenteral osteoporosis therapies. 5.- Napoli N, Elderkin AL, Kiel DP, Khosla S. Managing fragility fractures during the COVID-19 pandemic. Nat Rev Endocrinol. 2020;16(9):467-8. This Expert Opinion article focuses on measures of acute and long-term fragility fracture care during the COVID-19 pandemic. 6.- Cromer SJ, Yu EW. Challenges and opportunities for osteoporosis care during the COVID-19 pandemic. J Clin Endocrinol Metab. 2021;106(12):e4795-808. This mini-review summarizes the available evidence regarding the effects of COVID-19, its treatment, and the consequences of the pandemic itself on bone health. Additionally, evidence and expert recommendations regarding the putative effects of osteoporosis medications on COVID-19 outcomes and vaccine therapy are reviewed.
[table] Table 1: Pros and cons of skeletal medications in the COVID-19 eraAbbreviations: HRT, hormone replacement therapy; RCT , randomized-controlled trialOsteoporosis medication AdvantagesDisadvantages Vitamin D Theoretical non-musculoskeletal benefits as modulator of innate and adaptive immunity Association of deficiency with more worsened COVID-19 clinical outcomes (observational studies) Supplementation as advised by osteological societies for musculoskeletal health No clear evidence of COVID-19 course improvement through supplementation (RCTs) Negative impact of high-dose bolus on both falls and fracture risk HRT/raloxifene Potential beneficial effects on infection risk and COVID-19 clinical course (experimental & observational data) Raloxifene identified in drug repurposing program as an eligible candidate against oligosymptomatic COVID-19 RCT results confirming COVID-19 course improvement pending Associated with modest increase in thrombotic risk Bisphosphonates Long-lasting effects on suppression of bone turnover such that interruption during the COVID-19 pandemic is unlikely to be harmful Zoledronate theoretically capable of modulating the immune host status to protect against SARS-CoV-2 infections (experimental and observational data) Acute phase reaction following zoledronate infusion may resemble flu-like symptoms of COVID-19 Logistical difficulties of providing zoledronate infusion during the COVID-19 pandemic Denosumab Theoretical benefit of attenuation of inflammatory response to SARS-CoV-2 infections (experimental and observational data) No increased risk of COVID-19 illness in patients treated with denosumab Discontinuation causes rapid bone loss and an overshoot in bone turnover in all patients and multiple vertebral fractures in some patients Logistical difficulties in providing denosumab during the COVID-19 pandemic Romosozumab Self-administration logistically beneficial for skeletal health during the COVID-19 pandemic Upregulation of the Wnt/β-catenin pathway associated with cytokine storm; impact of sclerostin on SARS-CoV-2 infections unknownTeriparatide/abaloparatide Self-administration logistically beneficial for skeletal health during the COVID-19 pandemic PTH receptor expression on immune cells, but insufficient data regarding effects on SARS-CoV-2 infections [/table]
|
Neural underpinnings of morality judgment and moral aesthetic judgment
Morality judgment usually refers to the evaluation of moral behavior`s ability to affect others` interests and welfare, while moral aesthetic judgment often implies the appraisal of moral behavior's capability to provide aesthetic pleasure. Both are based on the behavioral understanding. To our knowledge, no study has directly compared the brain activity of these two types of judgments. The present study recorded and analyzed brain activity involved in the morality and moral aesthetic judgments to reveal whether these two types of judgments differ in their neural underpinnings. Results reveled that morality judgment activated the frontal, parietal and occipital cortex previously reported for motor representations of behavior. Evaluation of goodness and badness showed similar patterns of activation in these brain regions. In contrast, moral aesthetic judgment elicited specific activations in the frontal, parietal and temporal cortex proved to be involved in the behavioral intentions and emotions. Evaluation of beauty and ugliness showed similar patterns of activation in these brain regions. Our findings indicate that morality judgment and moral aesthetic judgment recruit different cortical networks that might decode others' behaviors at different levels. These results contribute to further understanding of the essence of the relationship between morality judgment and aesthetic judgment.OPENSubjects. Twenty-five female college students (Mean age = 21 years, SD = 1.7) from South China Normal University were recruited to participate in this study. All subjects were right-handed, had normal vision or corrected vision, had no history of neurological or psychiatric problems. Some studies have found that women rated heterosexual faces more consistently than men 22,23 , and women are more emotional than men and more easily moved by moral events 5,27-30 . Therefore, the experimental materials designed in this study were all male moral behaviors and the recruited subjects were all female, in order to eliminate the possible confusion between the gender of the subjects and the gender of the protagonist in the experimental stimulus 23,31 . In addition, the results of our pilot experiment, which asked subjects to make subjective judgment about the degree of moral behavior, also confirmed the above findings, indicating that women's evaluations of men's moral behavior were more consistent than men's evaluations of women's moral behavior, and were able to distinguish between different degrees of moral behavior. All the subjects signed the informed consent before the start of the experiment and were paid a certain amount of money after completing the experiment. The purpose and procedure of the experiment have been approved by the Academic Committee and the Ethical Review Board of the School of Psychology, South China Normal University. All methods used in the current study were performed in accordance with the relevant guidelines and regulations of the ethical review board. The data from three subjects were excluded from analyses because of large head movements (> 3 mm maximum displacement or 3° rotation). Thus, our fMRI analyses included data from 22 subjects with an average age of 21 years (s.d. = 1.7).
When you see a man rescuing the drowning child, you might be impressed by the kindness of this man's behavior. Besides that, you might also think that this man is a beautiful person on his inside. Similarly, when you see a man vandalizing flowers and trees in public, you may feel angry with this person for his bad behavior; at the same time, you may also think that this man is an ugly person on his inside. In the process of evaluating the actions in the above scenarios, people actually make two kinds of judgments on the same behavior. One kind is morality judgment, that is, people make a judgment about an individual's social behavior as morally good/bad decision-making process 1 , which usually refers to the evaluation of the moral behavior's ability to affect others' interests and welfare. Goodness and badness are a very common dichotomy in ethics, philosophy, and psychology. In the general context, badness is the absence or antithesis of goodness. It is driven by fear and manifests itself through violence and division.
Another is moral aesthetic judgment, that is, the decision-making process of making a judgment about whether an individual performing good or bad behavior is intrinsically beautiful or ugly. Moral beauty is also called spiritual beauty or inner beauty; moral ugliness is also called spiritual ugliness or inner ugliness. From a theological point of view, spiritual beauty exists whenever the attributes of God (love, justice, mercy, truth, generosity, grace) are manifested. From a philosophical point of view, human virtue is a sign of the spiritual beauty. For example, in Aristotle's Ethics, he describes human virtues as moral beauty, and the purpose of all virtues is beauty. For Aristotle, a person with good character is one who often uses spiritual beauty to guide his or her behavior, and beauty is the highest goodness in human behavior. Although ugliness is a topic that has been largely neglected by aestheticists because of the intense uneasiness associated with saying ugliness, a long intellectual tradition still uses ugliness as a mark of bad character, such as the ugly stepmothers and stepsisters of Grimm's fairytales. The aesthetic judgment is emotional. According to Kant, aesthetic judgment starts from the individual object to reflect whether its form can cause some universal pleasure, in which the intermediary between the object and the subject is pleasant and unpleasant emotion. Moral aesthetic judgment often implies an appraisal of the moral behavior's capability to provide aesthetic pleasure.
Haidt and Keltner define moral beauty as "the ability to discover, recognize, and enjoy the presence of goodness in society". He says, "When we see or imagine acts of charity or gratitude, we are impressed by its beauty and we also have a strong desire to do acts of charity and gratitude". This strong desire is called moral elevation, of which neural mechanism was found to be in line with the neural mechanism of mentalizing, both activate the mPFC/SFG and bilateral TPJ implicated in theory of mind (ToM), as well as the ACC, PCC/precuneus and anterior insula involved in empathic process. According to Haidt and Keltner, moral beauty is a beauty (ugliness) that people experienced in the face of virtuous (evil) behavior or the expression of moral goodness (badness) in society. Some studies have shown that aesthetic processing usually distinctly activate the cortical network involved in the theory of mind (ToM) and empathy, and better understanding of another's intentions and emotions is related to greater aesthetic appreciation.
Beauty and goodness, as well as ugliness and badness, are intrinsically linked. Traits such as honesty and kindness, or selfishness and cowardice, are imperceptible psychological traits, the goodness or badness of which stems from adherence to or violation of rational principles. If a person is morally good then, to this extent, they are beautiful; or, conversely, if a person is morally bad then, to this extent, they are ugly. In recent years, more and more researchers have begun to use brain imaging technology to investigate the relationship between the neural mechanisms of morality judgment and aesthetic judgment and the brain regions. Tsukiura and Cabeza used fMRI technology to scan the responses of subjects when they judged the beauty of faces and judged the goodness of hypothetical moral behaviors, in order to investigate the neural mechanisms of aesthetic judgment and morality judgment. They found that the processing of morality and aesthetic judgments was related to the medial orbitofrontal cortex (mOFC) and insula activity. The better the moral behavior described and the more beautiful the face, the stronger the mOFC activity but the weaker the insula activity; on the contrary, the worse the moral behavior described and the uglier the face, the stronger the activity in the insula but the weaker the activity in the mOFC.observed the relationship between the neural correlates of moral perception and facial aesthetic judgment and the brain activity in the ITG, mOFC and SFG. They suggested that two types of aesthetic judgments both involved the joint participation of perceptual, emotional and cognitive components 6 . Luo et al. asked subjects to make an integrated aesthetic judgment of facial portrait and moral description to investigate the neural underpinnings of the integration of facial beauty and moral beauty, and found that the appreciation of facial beauty and moral beauty recruited a common network involving the mOFC and middle occipital gyrus (MOG). They suggested that brain regions associated with sensory perception and reward might be recruited in the integrated aesthetic judgments 23 . Heinzelmann et al. suggested that all studies of morality judgment and aesthetic judgment involved different stimulus material and differences in visual processing of assessing the two judgment types, which make them vulnerable to confound, and as a result, any difference in neural activity between two types of judgments may be due to differences between stimulus materials, rather than the difference between morality and aesthetic judgments 24 . They measured brain activity of subjects judging the beauty of artistic images and the moral goodness of the actions depicted in the same images, and found that both common and distinct representations for morality and aesthetic judgments in temporoparietal and prefrontal regions were activated. Our recent study required subjects to rate the same positive moral behavior in terms of the good and beautiful extent to investigate the relationship between moral goodness and moral beauty from the neural perspective. We found that lateral OFC was commonly involved in the processing of both goodness and beauty. Furthermore, compared with moral goodness, moral beauty specifically activated cortical network which was implicated in the processing of mentalizing. Moral aesthetics seems to be related to the internal processes that underlie a perceiver's aesthetic experience.
The above results have important enlightening significance for us to understand the relationship between morality judgment and moral aesthetic judgment through the study of neural mechanism. However, there are still some limitations in previous studies. Firstly, researchers have only studied the neural mechanism of moral process and aesthetic process respectively, but have not directly compared these two processes of morality judgment and moral aesthetic judgment under the same condition. Secondly, previous studies on the relationship between the two processes were usually local comparisons and lack of integration. For example, we have recently investigated the neural mechanisms of moral goodness and moral beauty and found that they relied on both common processes as well as distinct cognitive components. Goodness and beauty belong to one aspect of morality judgment and moral aesthetic judgment respectively, their results did not represent the neural underpinnings of morality judgment and moral aesthetic judgment. In addition to the above limitations, the research paradigms and technology used are also different, and the results of different studies differ greatly, which is difficult to integrate. Therefore, the major question whether the processes of morality judgment and moral aesthetics are the same has not been truly answered. To solve this question, the present study used the method of parameter design to record and analyze the brain activity of moral brain in the process of making morality judgment about the goodness or badness of moral behavior and the brain activity of aesthetic judgment about the beauty or ugliness of the same moral behavior, aiming to reveal the neural underpinnings of morality judgment and moral aesthetic judgment. Morality judgment will be compared with moral aesthetic judgment on the same stimulus material. Parameter design has proved valuable in exploring the relationship between experimental parameters and physiological responses to system changes. Based on previous research results and theoretical discussion, we hypothesized that, on one hand, morality judgment and moral aesthetic judgment might share the common process, for example, in addition to what we've shown is the shared mechanism of moral beauty and moral goodness, moral ugliness might rely on similar neural mechanisms with moral badness; on the other hand, but that compared with morality judgment, moral aesthetic judgment might involve more complex and Stimuli. The experimental stimuli consisted of a series of scenes depicting moral behaviors of the male protagonists in the daily life. We first collected short sentences describing positive (he helped an old man cross the road, he escorted the injured to the hospital, etc.) and negative (he beat an old man, he cheated in the examination, etc.) moral behaviors. Then, scene drawings were drawn based on the sentence content. The protagonist in the scene drawing is marked with a red capital letter A to distinguish the other character(s) in the scene (See.
Twenty-four female college students were recruited to evaluate the effectiveness of the experimental materials. The sample size was determined with a power analysis using the G*Power tool. This analysis indicated that a sample size of 24 participants would provide 80% power, reaching a minimum recommended power value. The rating experiment asked subjects to rate the degree of moral behavior of the protagonist in the scene drawing on a 5-point scale (1 = extremely good, 5 = extremely bad). Based on the results, we selected 256 scenarios depicting two different moral degrees of positive behaviors (extremely good and somewhat good) and 256 scenarios depicting two different moral degrees of negative behaviors (extremely bad and somewhat bad), each containing 128 scenarios. A total of 512 scene drawings were used as stimulus materials for the following fMRI experiment. The scene drawing was 589 × 500 pixels.
Scenarios of positive and negative moral behaviors were pseudorandomly divided into two groups respectively. Half of positive behavior and half of negative behavior scenes were used as stimulus material for morality judgments, and another half of positive behavior and another half of negative behavior scene were used as stimulus material for moral aesthetic judgments, in order to avoid using the same stimulus material for different www.nature.com/scientificreports/ judgment tasks. The order of the two groups was counterbalanced across subjects. Each group contains 128 scenarios. The data from two subjects were excluded from analyses because of two much missing values. A 2 (group 1 and group 2) × 2 (behavior nature: positive behavior and negative behavior) × 2 (moral degree: extremely and somewhat) repeated measure ANOVA was used to test whether the two groups of material were equivalent. Evaluation results show that the main effect of group was not significant, F (1, 21) = 0.008, p > 0.05, the main effect of behavior nature was significant, F (1, 21) = 1202.34, p < 0.001, η 2 p = 0.98, the main effect of moral degree was significant, F (1, 21) = 85.54, p < 0.001, η 2 p = 0.8. The interaction between behavior nature and moral degree was significant, F (1, 21) = 1265.01, p < 0.001, η 2 p = 0.98, which was driven by the significant difference of the moral degrees of positive moral behaviors, t (21) = 17.85, ps < 0.001, and also the significant difference of the moral degrees of negative moral behaviors, t (21) = 31.24, ps < 0.001. The two-way interactions between behavior nature and group (F (1, 21) = 0.058, p > 0.05), between group and moral degree (F (1, 21) = 0.003, p > 0.05), and the three-way interaction between group, behavior nature and moral degree (F (1, 21) = 1.19, p > 0.05) were all not significant. This result indicates that positive and negative moral behaviors can be distinguished well, and there is no significant difference between the two groups of materials, which ensures the equivalence of the two groups of materials. This result is important because it suggests that differences in brain activity involved in morality processing and moral aesthetic processing are not caused by differences in experimental stimuli. Therefore, each judgment task contains scenarios of positive and negative behaviors with two different moral degrees. The allocation of these materials is balanced among the different judgment tasks. The moral degrees of the behaviors involved in each task were used as a parameter to determine the brain activation changes that varied with the degree of morality in the judgment tasks.
Procedure. The experiment consists of two tasks: morality judgment (MJ) and moral aesthetic judgment (MA). In the morality judgment, the subjects were asked to rate the degree of goodness or badness of the moral behavior in the scenarios on a 4-point scale (1 = extremely good, 4 = extremely bad); In the moral aesthetic judgment, the subjects were asked to rate the degree of beauty or ugliness of the moral behavior in the scenarios on a 4-point scale (1 = extremely beautiful, 4 = extremely ugly), while their brain activity was scanned. Scan was an event-related design.
During the MRI scan, the task instructions of 4 s were presented at the center of the computer screen, which told the subjects about the upcoming task. Next, one scenario was presented with a 4-point scale under it in which the number indicated the corresponding moral degree. The subjects were instructed to assess the moral degree of behavior of the protagonist (red A marked) in the scene drawing as quickly as possible within 4000 ms. Using a response box with buttons labeled 1-4, half of the subjects made a left index finger response (i.e., 1 or 2) for positive behaviors and a right index finger response (i.e., 3 or 4) for negative behaviors. Reverse responses were used for the other half of subjects to balance any influence of movement on data. Once the response was recorded, the rating bar disappeared and was followed by a blank display for a random period ranging between 500 and 5000 ms. If the scenario slide was not presented for 4000 ms after subjects pressed the button, it disappeared and a blank display will appear for the remaining time (See. The maximum duration of the presentation of the
## Fmri data acquisition. a siemens 3 t trio scanner with a 12-channel head coil at the brain imaging
Center of South China Normal University was used to collect the MRI image data, and a special pad was used to fix the head and prevent the head from moving. The T2* weighted functional images were obtained by using the echo plane imaging sequence. The specific scanning parameters were as follows: 33 layers of the whole brain image were scanned, each layer thickness was 3.5 mm, layer spacing was 0.8 mm, TR = 2000 ms, TE = 30 ms, Flip Angle = 90°, FOV = 204 × 204 mm 2 , Acquisition matrix = 64 × 64. High resolution T1-weighted structural images were obtained by using the MP-RAGE sequence, with specific parameters as follows: TR = 1900 ms, TE = 2.52 ms, Flip Angle = 9°, FOV = 256 × 256mm 2 , Acquisition matrix = 256 × 256, thickness = 1.0 mm, number of layers = 176, voxel size = 1 × 1 × 1 mm 3 . All subjects were asked to keep their heads still and stay awake during the MRI scan.
fMRI data analysis. Data preprocessing and statistical analyses of all images were conducted using SPM12
(Welcome Trust Center for Imaging, London, UK; http:// www. fil. ion. ucl. ac. uk/ spm). For each run of each participant, the preprocessing analysis consists of the following steps: (1) interscan slice-time correction, (2) spatial realignment and motion correction, (3) spatial normalization of aligned functional images to the Montreal Neurological Institute (MNI) template and resampling with a voxel size of 2 × 2 × 2 mm 3 , and (4) spatial smoothing with an isotropic Gaussian kernel of 6 mm FWHM. The three-dimensional translation of all the subjects' heads was less than 3.0 mm, and the three-dimensional rotation was less than 3.0°. Statistical analysis was performed after preprocessing. We used parametric designto study the BOLD response pattern of the whole brain to MJ and MA judgments. In this study, we manipulated the moral degree of the protagonist's behavior in the scene drawing to explore the neural underpinnings related to the processing of morality judgment and moral aesthetic judgment.
At the individual level, the trial-related activity observed in each subject was modeled by convolving a vector of trial onsets with a canonical hemodynamic response function (HRF) within the context of the General Linear Model (GLM). We input the rating scores for moral degree of each behavior into the SPM12 parametric modulator to identify brain regions displaying altered activity simply as a function of MJ and MA ratings. The missed trials were not modeled as predictors. On a given trial, the duration of the predictor of interest is equal to 4 s as shown in the scenario. Six head movement parameters calculated during the realignment were added to the model as non-interest predictors to offset the effects of head movement. Trials that did not respond to were not added to the model and were therefore included in the baseline. Low frequency noise was removed by 1/128 Hz high pass filter (HPF). For each subject, the brain regions associated with the changes in the degree of morality were identified by comparing the parameter-adjusted predictors of interest to baseline.
In the first-order analysis, each subject was analyzed separately. Then the first-level individual contrast images were fed to the second-level analysis employing a whole-brain random-effects model. Firstly, the cortical networks processed by MJ and MA were determined by parameter analysis, separately. Secondly, we calculated the conjunction analysis of MJ and MA contrasts to identify the common brain regions involved in the processing of two types of judgments. Third, we calculated a direct comparison between MJ and MA to obtain their unique brain activity. In addition, the parameter estimates in the regions of interest (ROI) was obtained through leaveone-subject-out-cross-validation (LOOCV) analysis. In this analysis, in each iteration, the ROI of the excluded subjects was defined using data from all subjects except one, which served as the test set. Each ROI was defined using a sphere with 6 mm radius centered on the peak voxel. This procedure was repeated 22 times by omitting a different subject each time, and the parameter estimates were then averaged across subjects and plotted. The LOOCV analysis method shows that the data used to define the ROI and the data extracted from that ROI are independent. We report the neural results at a voxel-level threshold of p < 0.05 (FDR-corrected) and cluster-level threshold of p < 0.05 (FWE-corrected) to correct for multiple comparisons. Statistical maps were labeled based on the MRIcro atlas (http:// www. mricro. com), and results were visualized with the BrainNetViewer (http:// www. nitrc. org/ proje cts/ bnv/).
# Results
Behavior results. The mean rating scores and the mean response times (RTs) of ratings for each moral degree in MJ and MA tasks were calculated and subjected to two 2 (task: MJ and MA) × 4 (moral degree: (1) extremely good/beautiful, (2) somewhat good/beautiful, (3) somewhat bad/ugly, and (4) extremely bad/ugly) repeated measures ANOVAs. In this analysis, half of the subjects' ratings were converted to the same criteria as the other half (i.e. reverse coding), with a larger number representing particularly bad or particularly ugly. Greenhouse-Geisser correction was performed on the non-sphericity of the data as needed, and the uncorrected degrees of freedom, the corrected p value, and the effect size (η 2 p ) were reported. Bonferroni correction was used to adjust the p-value of the paired comparison. In all analyses, the significance level was set at 0.5. The results of rating scores showed that the main effect of task was significant, F (1, 21) = 6.2, p = 0.02, η 2 p = 0.23, with significantly higher score of MJ than that of MA. The main effect of moral degree was significant, F (1, 21) = 274.14, p < 0.001, η 2 p = 0.93, with significant differences between degree 1 and degree 2, t
# Rois analysis.
In order to further investigate whether the above significant activated clusters selectively respond to the four categories of goodness, badness, beauty and ugliness during the process of MJ and MA, we performed a LOOCV analysis. ROIs were defined as a 6 mm sphere centered on the peak voxel in areas identified by direct task contrasts MJ-MA and MA-MJ. Our recent study on the neural mechanisms of moral goodness and moral beauty 17 found the common involvement of left lateral OFC (MNI peak coordinates: − 36, 24, − 15) in the processing of moral goodness and moral beauty. In the present study, we found that both morality judgment and moral aesthetic judgment activated the right lateral OFC (MNI peak coordinates: 30, 27, 3), although the result of the conjunction analysis showed that there was no shared brain region between these two types of judgments. It is necessary to further investigate the response pattern of bilateral lateral OFC in the four categories of goodness, badness, beauty and ugliness. Therefore, ROI analysis was also conducted on these two brain regions. Each ROI was iteratively defined on n-1 subjects, and then β value (parameter estimates) of excluded subjects were extracted from above defined ROIs. One-way repeated measures of ANOVA showed that there were significant www.nature.com/scientificreports/ response as compared to good response, t(21) < 3.91, ps < 0.05, a higher signal for ugly response as compared to bad response, t(21) < 6.18, ps < 0.01. And the difference between beauty and ugly was not significant (See. These results demonstrated that the brain regions involved in the process of MJ were only sensitive to good and bad responses, while the brain regions involved in the process of MA were only sensitive to beautiful and ugly responses.
In addition, differences in the four response categories were also significant in the bilateral lateral OFC (See an Image inserted in the upper left corner of(lOFC, F(3, 72) = 16.82, p < 0.001, η 2 p = 0.29; rOFC, F(3, 72) = 5.21, p = 0.002, η 2 p = 0.11). In the left lateral OFC, the differences were mainly driven by a higher signal for good response as compared to bad response, t(21) = 6, p < 0.001, a higher signal for beautiful responses as compared to ugly, t(21) = 2.75, p = 0.04, and bad responses, t(21) = 6.23, p < 0.001; in the right lateral OFC, the differences were mainly driven by a higher signal for good response as compared to bad response, t(21) = 2.13, p = 0.04, a higher signal for beauty response as compared to ugly response, t(21) = 2.73, p = 0.009. These results revealed that the bilateral lateral OFC had stronger activation for goodness and beauty than badness and ugliness, which was consistent with a recent finding on the neural mechanism of moral goodness and moral beauty judgments 17 , namely, the lateral OFC is commonly engaged in the processing of both moral goodness and moral beauty.
# Discussion
In the present fMRI study, we required the subjects to assess the degree of goodness or badness, as well as the degree of beauty or ugliness of moral behaviors in daily life, to determine whether morality judgment and moral aesthetic judgment differ in their neural underpinnings or not. Behavioral results revealed that the rating of morality judgment was significantly higher than those of moral aesthetic judgment for the same moral behavior, indicating that people may be more inclined to make morality judgment. The fMRI results revealed that when the same experimental material was used but varying task demands were systematically considered, the distinct nervous systems of morality judgment and moral aesthetic judgment emerge with more clarity. Specifically, direct contrasts showed specific activations for morality judgment; these were mainly located in the frontal, parietal and occipital cortex, including the bilateral SMA, IPL/SPL, MFG/pars opercularis of IFG extending onto the ventral premotor cortex (PMv), and IOG/MOG: i.e., brain regions which have been previously reported for motor representations of behavior during understanding others` behaviors. These areas are only sensitive to goodness and badness, but not to beauty and ugliness. In contrast, moral aesthetic judgment significantly elicited specific activations in the frontal, parietal and temporal cortex, including the bilateral ACC/mPFC/mOFC, PCC/ Precuneus, TPJ and anterior insula: i.e., brain regions which have been proved to be involved in understanding other people's behavioral intentions and emotions. These areas are only sensitive to beauty and ugliness, but not to goodness and badness. Interestingly and more importantly, when subjects judged a behavior to be good (as in contrast to bad) or to be beauty (as in contrast to ugly), the lateral OFC showed enhanced BOLD signals, which was in line with what was found in a recent study that lateral OFC was commonly involved in the processing of goodness and beauty. But no common brain region was found to be involved in the processing of badness and ugliness. Against the background of the literature, present findings indicate that morality judgment and moral aesthetic judgment seemly activate completely different cortical networks, although lateral OFC is commonly activated in both moral goodness and beauty. The OFC is critical for goal-directed behavior. Converging neuroimaging evidence shows differences between lateral OFC and medial OFC 43-51 , the medial OFC in general is concerned with monitoring reward values 31,52-58 , but the lateral OFC connecting the anterior insula cortex is generally involved in the processing of affectively stimuli. We found that both moral goodness and moral beauty more strongly activated the lateral OFC/anterior insula, suggesting that this brain region may be involved in encoding the emotional experiences that often occur in human interactions. As moral goodness and moral beauty judgments share a very critical characteristic, that is, the object of both judgments generally refers to the positive moral behavior. This result is not surprising given that positive moral behavior is a rewarding stimulus that evokes emotional experiences.
It is important to mention, we found that the medial OFC, with strong connections with vmPFC and ACC, not only responded to moral beauty, but also responded to moral ugliness, and there was indifferent in response intensity between them. Due to strong connections among OFC, vmPFC and ACC a network was suggested-the orbital medial prefrontal cortex (OMPFC)-related to emotion processing and social cognition. Traditionally, the mOFC has commonly been suggested as an important reward-related region involved in the processing of beauty or positive stimuli, however, recent works showed that the mOFC was activated by negative experiences, and also was involved alike by both the beautiful and the ugly stimuli. The mOFC, as a part of the medial cortical structures involved in self-referential and self-related processes, has been proved not only to be involved in self-perception in social cognition, but also engaged in coding internal motivational values, www.nature.com/scientificreports/ particularly in the absence of external prompts. Consistent with the above findings, the result that both moral beauty and moral ugliness significantly activated the OMPFC might be interpreted as the activation of neural circuits related to self vs. other-assessment. In the moral aesthetic judgment, people usually evaluate the inner beauty or ugliness of others through morally behavioral understanding, in which their mental status will be perceived and be compared with ourselves. And the mOFC or OMPFC might play a crucial role in this process.
In the present study, people usually focus on the action itselfand are emotionally unmoved 2,15,16 when they judge whether a moral behavior is good or bad. Behaviors directed by morality judgment are cognitively experienced and judged without deep emotional involvement. However, when people evaluate a person's inner beauty or ugliness according to his moral behavior, people's emotions are easily affected by his behavior: that is, if he acts good, people`s emotions are elevated; If he acts evil, people feel contempted, angry or disgusted, and the emotion is unpleasant. Both morality judgment and moral aesthetic judgment are accomplished by behavioral understanding. Understanding the behaviors of others will contribute to the harmony of people`s interpersonal relationships, and is necessary for efficient communication and collaboration, which deals primarily with what they are doing, how they are doing it (i.e., behavioral states), and why they are doing it (i.e., mental states), and it is accomplished by two processes and related brain networks: motor representation of behavior and intention representation of behavior, supported by the mirror neuron system (MNS) and the mentalizing system (MZS), respectively. Prior neuroimaging studies have shown that the MNS and the MZS are differentially activated by what/how and why questions about behaviors.
Cortical areas with motor properties, including inferior frontal gyrus (IFG), ventral premotor cortex (PMv) and inferior parietal lobule (IPL), namely, the MNS, have been observed to respond to motor behaviors when they are performed and observed, reflecting motor representation of behavior. Li et al. recently used fNIRS to examine the brain activity in the frontal, motor, parietal and occipital regions, aiming to better understand the brain correlates involved in encoding motor complexity. They found that motor complexity sensitive brain regions were present in the pars opercularis IFG/PMv, primary motor cortex (M1), IPL/supramarginal gyrus and middle occipital gyrus (MOG) during behavior execution, and in pars opercularis IFG/PMv and M1 during behavior observation, suggesting that the processing of motor complexity involves not only M1 but also pars opercularis IFG, PMv and IPL, each of which plays a critical role in behavior perception and execution. Some studies that used transcranial magnetic stimulationfound that the left IPL is involved in the representation of the manipulability of the objects, while the temporal cortex includes more abstract representation of object function. Spun et al. found that there existed a dissociation between the posterior and anterior regions of IFG in their contribution to behavior representation: The pIFG/PMv is associated with the perception of behaviors and the execution of understanding behaviors, mainly manifesting in encoding more concrete representations of behaviors-"what" and "how"; and the anterior IFG region located in the ventrolateral PFC is associated with the motive of understanding behaviors, mainly manifesting in encoding more abstract representations of behaviors-"why". In the present study, we found that compared with moral aesthetic judgment, morality judgment, both good reaction and bad reaction, recruited motor-related areas, including MFG/IFG/PMv, SMA, SPL/IPL and IOG/MOG, indicating that morality judgments might use motor representation through "body" reading, supported by the MNS 93,97-99 which is part of a larger sensorimotor brain network, to complete assessments of whether a moral behavior is good or bad and how good or bad it is. It is worth noting that the brain network for morality judgment is more consistent with that for the behavior execution than with that for the behavior observation.
Although the human MNS is reliably active during behavior understanding, several studies have shown that it is insensitive to the intentional representation of observed behavior; rather, a separate brain system known as the ToM or mentalizing system (MZS), by inferring the internal states of other persons that typically drive behaviors, such as their goals, desires, beliefs, intentions, causal attributions and traits, appears to be critical 90,100-102 . It is part of a larger default brain network, including the medial PFC, PCC/precuneus and the TPJ. We found no activation of brain regions associated with mentalizing in morality judgment, which is consistent with the result from a recent study. In this study, Yoder and Decety scanned 40 healthy adults using fMRI while these subjects were watching scenes in which people harmed or helped others. They found that subjects sensitive to righteous actions activated brain regions associated with (social) cognitive aspects of morality judgment, rather than with processing emotional aspects of a moral scenario. Mentalizing, in particular, empathy, may not be necessary for deciding whether moral behaviors are right or wrong, but it may still play an important role in the motivation of moral behavior. Morality judgment, assessing of whether a moral behavior is good or bad and how good or bad it is, seems to rely on more primitive cognitive systems.
The MZS is usually activated by why questions about behaviors is responsive to tasks of mental-state reasoning. The brain regions of the MZS are involved in the multifunctional process of distinguishing oneself from the thoughts of others and gaining an understanding of their mental status, which in turn enable people to recognize, interpret, and predict behavior. In the present study, when the subjects were required to evaluate the beautiful or ugly degree of the moral behaviors, brain regions consistent with the MZS were activated, suggesting that the processing of moral aesthetics is likely to be supported by the MZS. In the process of moral aesthetic judgment, people are likely to understand the social actor`s moral behavior and make a judgment of beauty or ugliness by reasoning the mental status that typically drive behavior, such as beliefs, desire and intentions. Moral aesthetic judgment may require psychological reflection or, more generally, a shift in attention to the social actor's intention. The neural mechanism for the question of how beautiful or ugly the moral behavior is, seems to be the same as the neural mechanism for the question of why the behavior is done. By manipulating either behavior goals or the content of a perceived behaviorwww.nature.com/scientificreports/ In addition, in moral aesthetic judgment, the beauty or ugliness of moral behavior generally refers to the perception of the beautiful or ugly of social actor's personality traits, and the traits inference always involves the activiations of TPJ and mPFC. In addition to regions implicated in mentalizing, we found that moral aesthetics also recruited the activation of anterior insula. This region is not only implicated in the meta-representation of emotional states and interoceptive awareness, but also correlates with the cognitive processing of self and others 123 , particularly is critical for the perceived intentionality of other people`s behavior. In short, moral aesthetic judgment refers to the perceiver's judgment on the inner beauty of others, which seems to be concerned with the internal processes 125 that underlie a perceiver's aesthetic experience. This process has been reported to be supported by the cortical network implicated in the mentalizing system.
In brief, previous studies mainly focused on the neural mechanism of moral process and aesthetic process respectively; the present study directly compared the two processes of morality judgment and moral aesthetic judgment under the same condition and examined the commonalities and differences in the neural correlates of these two processes through an experimental approach. However, the present study is limited in that it only used right and wrong behaviors as representatives of morality judgment; future studies using different kinds of moral judgments, such as wrongness, punishment, blame, moral dilemmas, etc., should be performed to verify our findings. In addition, to eliminate the possible confusion between the gender of the subjects and the gender of the protagonist in the experimental stimulus, we used male moral behaviors as experimental materials and recruited female subjects to evaluate behaviors. It would be interesting to determine whether the same results could be found if female moral behaviors were used as experimental materials and men were recruited as subjects. Or, whether the results obtained by recruiting subjects of different races to make moral or moral aesthetic judgments would hold up cross-culturally is also very worthy to be investigated.
# Conclusions
The perception of moral aesthetics often implies an appraisal of the moral behavior's capability to provide aesthetic pleasure, which seems to be related to the internal processes that underlie perceivers`aesthetic experience; while morality judgment usually refers to the evaluation of the moral behavior`s ability to affect the interests and well-beings of others, which seems to be related to the execution of the behavior. Both types of judgments are accomplished by behavioral understanding. In the present study, we reveal the neural underpinnings of morality judgment and moral aesthetic judgment by collecting data about subjects` brain activity as they evaluate the extent of goodness and badness, as well as the extent of beauty and ugliness of moral behaviors. By comparing two types of judgments made by the same subject on the same stimulus material, we found that morality judgment and moral aesthetic judgment recruit different cortical networks that might decode others' moral behaviors at different levels, although lateral OFC is commonly involved in the processing of goodness and beauty. These results contribute to further understanding of the essence of the relationship between morality judgment and aesthetic judgment.
## Data availability
All data analyzed in this study are available from the corresponding author upon reasonable request. |
Comparative assessment of image quality for coronary CT angiography with iobitridol and two contrast agents with higher iodine concentrations: iopromide and iomeprol. A multicentre randomized double-blind trial
Objectives To demonstrate non-inferiority of iobitridol 350 for coronary CT angiography (CTA) compared to higher iodine content contrast media regarding rate of patients evaluable for the presence of coronary artery stenoses. Methods In this multicentre trial, 452 patients were randomized to receive iobitridol 350, iopromide 370 or iomeprol 400 and underwent coronary CTA using CT systems with 64detector rows or more. Two core lab readers assessed 18 coronary segments per patient regarding image quality (score 0 = non diagnostic to 4 = excellent quality), vascular attenuation, signal and contrast to noise ratio (SNR, CNR). Patients were considered evaluable if no segment had a score of 0. Results Per-patient, the rate of fully evaluable CT scans was 92.1, 95.4 and 94.6 % for iobitridol, iopromide and iomeprol, respectively. Non-inferiority of iobitridol over the best comparator was demonstrated with a 95 % CI of the difference of [-8.8 to 2.1], with a pre-specified non-inferiority margin of -10 %. Although average attenuation increased with higher iodine concentrations, average SNR and CNR did not differ between groups. * Stephan Achenbach [email protected] on behalf of the X-ACT Study GroupConclusions With current CT technology, iobitridol 350 mg iodine/ml is not inferior to contrast media with higher iodine concentrations in terms of image quality for coronary stenosis assessment.Key Points- Iodine concentration is an important parameter for image quality in coronary CTA. - Contrast enhancement must be balanced against the amount of iodine injected. - Iobitridol 350 is non-inferior compared to CM with higher iodine concentrations. - Higher attenuation with higher iodine concentrations, but no SNR or CNR differences.
# Introduction
Coronary computed tomography angiography (coronary CTA) has become widely accepted in clinical practice [bib_ref] Cardiac CT vs. stress testing in patients with suspected coronary artery disease:..., Rahsepar [/bib_ref] [bib_ref] Utilization of cardiac computed tomography angiography and outpatient invasive coronary angiography in..., Roifman [/bib_ref] [bib_ref] Cardiac computed tomography in current cardiology guidelines, Al-Mallah [/bib_ref]. Technology progress has increased the robustness and diagnostic performance of coronary CTA, resulting in improved image quality and lower radiation exposure [bib_ref] Prospectively ECGtriggered high-pitch spiral acquisition for coronary CT angiography using dual source..., Lell [/bib_ref] [bib_ref] Image quality of ultra-low radiation exposure coronary CT angiography with an effective..., Schuhbaeck [/bib_ref] [bib_ref] Low dose CT of the heart: a quantum leap into a new..., Maffei [/bib_ref] [bib_ref] Impact of heart rate frequency and variability on radiation exposure, image quality,..., Weustink [/bib_ref] [bib_ref] Preserved diagnostic performance of dual-source CT coronary angiography with reduced radiation exposure..., Weustink [/bib_ref] [bib_ref] Optimizing radiation dose by using advanced modelled iterative reconstruction in high-pitch coronary..., Gordic [/bib_ref].
The protocols for administration of intravenous contrast media (CM) are of major importance in coronary CTA, usually performed to identify coronary artery stenoses but also calcified and non-calcified plaques [bib_ref] Non-invasive measurement of coronary plaque from coronary CT angiography and its clinical..., Dey [/bib_ref] [bib_ref] Coronary CT angiography versus intravascular ultrasound for estimation of coronary stenosis and..., Fischer [/bib_ref] [bib_ref] Coronary plaque imaging with multislice computed tomography: technique and clinical applications, Cademartiri [/bib_ref] [bib_ref] Quantification of coronary artery calcium using ultrafast computed tomography, Agatston [/bib_ref]. The optimal intravascular attenuation for coronary CT angiography is under debate [bib_ref] Multicenter comparison of high concentration contrast agent iomeprol-400 with iso-osmolar iodixanol-320: contrast..., Becker [/bib_ref] [bib_ref] Higher intracoronary attenuation improves diagnostic accuracy in MDCT coronary angiography, Cademartiri [/bib_ref]. Several publications suggested that adequate opacification of the vessel lumen for the simultaneous identification of both calcified and non-calcified plaques requires a careful contrast injection protocol that achieves a lumen opacification of at least 300 HU [bib_ref] Higher intracoronary attenuation improves diagnostic accuracy in MDCT coronary angiography, Cademartiri [/bib_ref] [bib_ref] Influence of intra-coronary enhancement on diagnostic accuracy with 64-slice CT coronary angiography, Cademartiri [/bib_ref] [bib_ref] Impact of intravascular enhancement, heart rate, and calcium score on diagnostic accuracy..., Cademartiri [/bib_ref]. In principle, the Iodine Delivery Rate (IDR) should be the reference when using different compounds for intraluminal enhancement [bib_ref] How to design contrast injection protocols for multiple detector-row CT angiography (MDCTA), Fleischmann [/bib_ref]. However, most centres do not use this approach in clinical practice.
Higher iodine concentration of the injected CM is associated with higher attenuation [bib_ref] Influence of intra-coronary enhancement on diagnostic accuracy with 64-slice CT coronary angiography, Cademartiri [/bib_ref] [bib_ref] Intravenous contrast material administration at helical 16-detector row CT coronary angiography: effect..., Cademartiri [/bib_ref]. However, increasing the total amount of iodine injected could raise safety issues for patients at risk such as contrast induced nephropathy [bib_ref] Risk of iodinated contrast material-induced nephropathy with intravenous administration, Katzberg [/bib_ref] [bib_ref] Contrast-induced nephropathy: the wheel has turned 360 degrees, Thomsen [/bib_ref] [bib_ref] Contrast-medium-Induced nephropathy correlated to the ratio between dose in gram iodine and..., Nyman [/bib_ref]. Therefore, adequate contrast enhancement must be balanced against the amount of iodine injected. Several studies have addressed the level of attenuation with different CM [bib_ref] Multicenter comparison of high concentration contrast agent iomeprol-400 with iso-osmolar iodixanol-320: contrast..., Becker [/bib_ref] [bib_ref] High iodine concentration contrast material for noninvasive multislice computed tomography coronary angiography:..., Cademartiri [/bib_ref] [bib_ref] Contrast bolus optimization for cardiac 16-slice computed tomography: comparison of contrast medium..., Rist [/bib_ref] , but few studies evaluated the impact of their concentration on image quality [bib_ref] Image quality and attenuation values of multidetector CT coronary angiography using high..., Kim [/bib_ref]. Furthermore, whether differences less than 50 mg/ml of iodine concentration could affect image quality remains to be clarified. The present study compared a CM with iodine concentration of 350 mg iodine/ml (iobitridol) to two CM with higher iodine concentrations (iopromide 370 mg/ml and iomeprol 400 mg/ml) for coronary CTA. The main objective of the study was to demonstrate the statistical non-inferiority of iobitridol 350 compared to the best of the two comparators in terms of image quality and interpretability as measured by the rate of patients with CT scans evaluable for the identification of coronary artery stenosis.
# Material and methods
## Study design and patient enrolment conditions
This study was a non-inferiority, multicentre, randomized, double-blind, clinical trial on three parallel groups. Patients were included in 23 centres from five countries between November 2010 and September 2012 and randomized on a 1:1:1 ratio to undergo clinically indicated coronary CTA after injection of iobitridol, iopromide or iomeprol. The study was approved by each local ethics committees and the National Health Authorities. Written informed consent was obtained from each participating patient.
Symptomatic adult patients with suspected coronary artery disease (CAD), and scheduled for coronary CT angiography were enrolled in this study. Patients could not be included if they had both a contraindication to β-blocker medications and a baseline heart rate above 65 beats per minute (bpm). Additional reasons for exclusion were the presence of arrhythmias or nonsinus rhythm, coronary artery bypass grafts or stents, artificial heart valves, moderate to severe aortic valve stenosis, hyperthyroidism, clinical instability, severe renal failure or previous injection of any CM within 48 hours prior to the study.
## Patient preparation
β-blockers were mandatory if heart rate was >65 bpm. The specific drug, dose and mode of administration were selected according to site routine practice. A minimum dose of 0.8 mg of sublingual nitroglycerine spray was mandatory immediately before the CT examination. Other pre-medication was permitted, given according to operator preference, and recorded.
## Injection of contrast media
For each patient, one of three CM was delivered intravenously: Iobitridol 350 mg iodine/ml (Xenetix®, Guerbet, Aulnaysous-Bois, France), iopromide 370 mg/ml (Ultravist®, Bayer Healthcare, Berlin, Germany) and iomeprol 400 mg/ml (Iomeron®, Bracco, Milan, Italy). Delivered volume and delivery rate of CM was consistent for the three CM but varied according to patient body weight (BW): 60 ml injected at 4 ml/s for a BW <60 kg, 75 ml at 5 ml/s for a BW between 60 and 80 kg, 90 ml at 6 ml/s for a BW >80 kg. Therefore, the iodine-delivery rate was lowest for iobitridol. CM was warmed and injections were followed by a 100 % saline flush of 75 ml administered at the same rate as the CM.
## Scan protocol
All coronary CTAs were performed on systems with at least 64 detector rows (single or dual source). The tube voltage was adapted to the patient's BW depending on the equipment used in the site: either 120 kVp for all patients or 120 kVp for BW ≥80 kg and 100 kVp for BW <80 kg. Each patient first received a non-contrast acquisition for the quantification of coronary calcium, followed by a high-resolution contrast-enhanced acquisition. Non-contrast scans were acquired according to local protocol, mandating the use of 120 kV tube voltage and 3 mm thickness of the reconstructed cross-sectional images.
Test bolus or bolus tracking [bib_ref] Intravenous contrast material administration at 16-detector row helical CT coronary angiography: test..., Cademartiri [/bib_ref] [bib_ref] Parameters affecting bolus geometry in CTA: a review, Cademartiri [/bib_ref] as well as dose reduction techniques (e.g. mAs modulation, scan duration, tube current and pitch) were chosen according to site specific routine. Retrospective ECG-gating was always allowed, whereas prospectively ECG triggered data acquisition was only permitted for patients with HR < 65 bpm.
Dose length product (DLP) and volume CT dose index (CTDIvol) were recorded for the non-enhanced and for the contrast-enhanced coronary CTA acquisition.
## Calcium scoring
The calcium score [bib_ref] Quantification of coronary artery calcium using ultrafast computed tomography, Agatston [/bib_ref] was calculated on-site. Patients were then classified using the following standard scale [bib_ref] Electron beam computed tomography and coronary artery disease: scanning for coronary artery..., Rumberger [/bib_ref] : 1 = normal (no calcium); 2 = mildly elevated (Agatston Score 1 to 100); 3 = moderately elevated (Agatston Score 101 to 400); 4 = highly elevated (Agatston Score 401 to 1000); 5 = severely elevated (Agatston Score > 1000).
## Image reconstruction for coronary cta
Cross-sectional image data sets were reconstructed using the thinnest possible slice thickness and standard reconstruction kernels (i.e. medium-smooth) without implementation of iterative reconstruction algorithms in order to reduce softwarerelated attenuation variability. The choice of the best temporal window was left to the local operators. Original axial DICOM data were transferred for off-site reading purposes to a dedicated image core laboratory.
## Image evaluation for detection of stenosis
All CTA images were assessed by two independent, experienced readers with more than 10 years of experience in coronary CT angiography, both fully blinded to the patients´clinical characteristics and the CM used.
The primary endpoint was the rate of patients with evaluable CT scans, i.e. CT scans that made it possible to identify or rule out coronary artery stenoses in all segments of the coronary tree with a reference diameter of 1.5 mm or more. Using a 5-point scale, off-site readers graded the quality of all 18 segments of the Society of Cardiovascular Computed Tomography (SCCT) coronary segmentation model [bib_ref] SCCT guidelines for the interpretation and reporting of coronary computed tomographic angiography, Raff [/bib_ref] (4 = excellent quality, full confidence without any doubts concerning the presence/absence of luminal stenosis; 3 = good quality, confidence concerning the presence/absence of luminal stenosis; 2 = moderate quality, relative confidence, with minor doubts concerning the presence/ absence of luminal stenosis; 1 = poor quality, some doubts concerning the presence/absence of stenosis; 0 = non diagnostic, with relevant doubts concerning the presence/absence of stenosis). A patient's CT scan was considered as evaluable for identification of coronary artery stenosis if none of the 18 coronary segments had a score of 0 (except for segments graded 0 due to an occlusion located proximally, which did not make the patient non-evaluable).
As a secondary endpoint, a per-patient image quality score was computed off-site by averaging segment quality scores within patients.
## Attenuation, signal and noise measurements
Arterial attenuation, signal-to-noise ratio (SNR) and contrastto-noise ratio (CNR) constituted secondary endpoints. Arterial vascular attenuation was measured off-site as follows: one region of interest (ROI) of 2 mm 2 minimum size and located in the lumen of the LAD, the LCX, the RCA and the LM coronary artery; one ROI of 100 mm 2 minimum size in the ascending aorta and one ROI of 50 mm 2 minimum size in the left ventricle. Attenuation of LM, ascending aorta and left ventricle was measured on pre and post contrast images.
The noise was measured in the aorta (100 mm 2 minimum size, located at the level of left main origin), air-filled cavities (50 mm 2 in bronchia or trachea) and muscle (25 mm 2 , thoracic wall) in post-contrast images and was used to derive SNR and CNR values.
## Stenosis assessment and patient management
The presence of significant stenosis (>50 % of the lumen) in the 18 SCCT segments was reported by on-site radiologists using a 5-point scale (5 = certainly yes; 4 = probably yes; 3 = doubtful; 2 = probably no ; 1 = certainly no).
On-site radiologists recorded the recommended management from the following list: no action, medication, invasive coronary angiography and other.
## Coronary track rate
In order to determine whether image quality is sufficient for automated segmentation as often used for evaluation, commercially available software ("Comprehensive Cardiac Analysis" on IntelliSpace Portal, Philips) was used to automatically track LAD, LCX and RCA up to their distal segments and the number of segments tracked per patient constituted the 'coronary track rate', which was assessed by a third independent off-site radiologist, also with more than 10 years of experience in cardiac CT.
## Clinical safety
All adverse events (AEs), including cardiac events as evidenced by ECG (performed up to 10 min post-injection), were reported from the patient's signature of the informed consent up to 30 min after the examination. The intensity of AEs was classified on a 3-point scale, based on interference with daily activities: mild (no interference), moderate (moderate interference) or severe (the subject is unable to work). The causal relationship of the AE to the CM injected was defined according to the French Method of Causality Assessment [bib_ref] Imputation of the unexpected or toxic effects of drugs. Actualization of the..., Begaud [/bib_ref].
Patient comfort and pain were assessed on-site using a selfadministered questionnaire (5-point scale ranging from 1 (worst situation) to 5 (best situation)) and a visual analogue scale (VAS), respectively.
# Statistical analysis
Assuming an expected proportion of 90 % of patients with coronary CTAs evaluable for CAD diagnosis with the three investigated contrast agents, 424 assessable patients (3x141) were needed to ensure with 80 % power and 5 % two-sided type-one error that the lower limit of the 95 % confidence interval (95 % CI) of the difference between iobitridol 350 and the best of the two comparators (iopromide 370 or iomeprol 400) is not greater than the clinical non-inferiority margin set at -10 %.
Three patient populations were defined: all-includedpatients (AIP), full analysis set (FAS), and safety set.
AIP population included all patients enrolled in the study and having signed the informed consent. FAS included all patients who underwent the examination and had available assessments of the primary endpoint for iobitridol 350, iopromide 370 or iomeprol 400 examinations. The safety set included all patients who had received at least one injection of contrast agent, regardless of the quantity.
For comparisons between the three groups, two-sided tests were performed at a 5 % level of significance. In case of multiple comparisons, significance level of each test was adjusted to ensure a 5 % overall significance level. Accuracy of estimates was computed with 95 % CIs.
Student's t-test and the F-test were used for quantitative variables, whereas the Chi-square test was used for qualitative variables.
Multiple regression models were performed to identify potential relationship between calcium scoring and stenosis assessment, image quality and territory.
# Results
## Study population
A total of 468 patients gave their consent and therefore were included (58 % male; aged 57.8 ± 12.4 years). Sixteen patients were excluded: five patients did not have a CM injection, and in 11 patients off-site image assessment was not possible due to technical failures. Therefore, 452 patients were analysed in the FAS and 463 in the safety set.
There were no significant differences between the three groups in terms of demographics, clinical symptoms, risk factors and pre-CTA heart rate. No differences were noted in terms of requirement for βblockers for the CTA procedure, calcium score and radiation dose [fig_ref] Table 1: Patients and procedure characteristics [/fig_ref].
## Image evaluation for detection of stenosis
The rate of patients with evaluable CT scans was not significantly different between the three groups (92.1 %, 95.4 % and 94.6 % of patients in the FAS, for iobitridol, iopromide and iomeprol, respectively) [fig_ref] Figure 1: Transaxial cross-sections [/fig_ref] and. The 95 % CI of the difference between iobitridol and the best of the two comparators (iopromide) was [-8.8 to 2.1], demonstrating the non-inferiority of iobitridol, when compared to other CMs, in its ability to allow CAD diagnosis through a complete assessment of coronary artery segments.
The average score for image quality per-segment (total number of segments = 6,220) was 3.5 ± 0.9, 3.5 ± 0.8 and 3.4 ± 0.9 for the iobitridol, iopromide and iomeprol groups, respectively (p > 0.05).
## Attenuation, signal and noise measurements
The average pre-contrast vascular attenuation calculated from values of the ascending aorta, LM and left ventricle was 42.2 ± 9.7 HU, without any difference between the three groups (p = 0.993;. Vascular attenuation was significantly increased in post-contrast images as compared to pre-contrast images in all three structures. Average post-contrast arterial vascular attenuation was 426.3 ± 92.9 HU, 449.8 ± 88.1 HU and 466.4 ± 104.6 HU for the iobitridol, iopromide and iomeprol groups, respectively (p = 0.001).The difference between groups was statistically significant for absolute values; however, when values accounting for noise were plotted as SNR and CNR, differences were no longer significant. Measurements of noise in the ascending aorta showed no significant difference between groups (p = 0.311).
## Other secondary endpoints
No difference was observed regarding the number of significant stenoses identified with the three CMs (p = 0.580; [fig_ref] Table 4: Secondary efficacy endpointsResults are expressed as mean ± standard deviation or n [/fig_ref].
Multivariate analyses showed a relationship between calcium scoring and stenosis assessment according to territory (p < 0.001) and between calcium scoring and image quality regardless of the territory (p = 0.007). A mean of 11 segments were automatically tracked out of a maximum of 13 segments per patient whatever the contrast media injected [fig_ref] Table 4: Secondary efficacy endpointsResults are expressed as mean ± standard deviation or n [/fig_ref].
The mean score for comfort of the examination rated by the patient was good (4.4 ± 0.6) and similar for all three groups. Patient comfort was confirmed by a low reported intensity of pain (mean score of less than 1 out of 10 cm on VAS for the three groups).
Regarding patient management, no action was required after the CTA for 73 % of the patients overall, with no significant difference between groups.
## Clinical safety
The percentage of patients experiencing post CM-injection AEs was 15.1 %; 19.5 % and 15.1 %, for the iobitridol, iopromide and iomeprol groups, respectively [fig_ref] Table 5: Incidence and characteristics of adverse events [/fig_ref]. Most AEs concerned cardiac disorders, which were reported through systematic ECG follow-up performed up to 10 min post-injection. Overall, mean heart rate was 65.0 ± 9.7 bpm 2 min after CM injection and 62.8 ± 9.3 bpm 10 min after CM injection, and was similar in all three groups.
No severe AEs were reported. Only mild events were reported with iobitridol while four and seven moderate events were reported with iopromide and iomeprol, respectively. Few post-CM AEs were considered possibly related to CM administration: two in the iobitridol group and five in the iopromide group as well as in the iomeprol group. The cardiac events considered possibly related to CM injection were bradycardia (one patient in each group) and extrasystoles (two patients in the iomeprol group). Other possibly related events were pain in the iobitridol group, injection site pain, nausea, headache and urticaria in the iopromide group, injection site pain and feeling hot in the iomeprol group.
# Discussion
The rationale for this trial was that CM iodine concentration may not play a role in the ability to visualize coronary stenoses by CTA. This study demonstrated the non-inferiority of iobitridol 350 in providing evaluable CT scans for assessment of coronary stenosis, as compared to CM with higher iodine concentrations (iopromide 370 mg iodine/ml and iomeprol 400 mg iodine/ml). These two CMs had been compared in a previous study without differences in terms of image quality [bib_ref] Image quality and attenuation values of multidetector CT coronary angiography using high..., Kim [/bib_ref]. Our study, the largest so far to address the relationship between iodine concentration and image quality, indicates the possibility of further reducing iodine content. Reflecting current clinical practice in many sites, IDR was not adjusted according to the iodine concentration of each CM.
Our results are in line with previous studies in which higher iodine concentrations were associated with higher intravascular attenuation [bib_ref] Multicenter comparison of high concentration contrast agent iomeprol-400 with iso-osmolar iodixanol-320: contrast..., Becker [/bib_ref] [bib_ref] Aortic and hepatic contrast medium enhancement at CT. Part II Effect of..., Bae [/bib_ref] [bib_ref] Aortic and hepatic contrast medium enhancement at CT. Part I. Prediction with..., Bae [/bib_ref] [bib_ref] Aortic and hepatic peak enhancement at CT: effect of contrast medium injection..., Bae [/bib_ref] [bib_ref] Multiphasic injection method for uniform prolonged vascular enhancement at CT angiography: pharmacokinetic..., Bae [/bib_ref]. However, comparison is difficult since the volume of iodine injected was not always adjusted to patient weight [bib_ref] Multicenter comparison of high concentration contrast agent iomeprol-400 with iso-osmolar iodixanol-320: contrast..., Becker [/bib_ref]. The fact that no difference was observed between CMs when considering SNR and CNR cannot be fully explained. Compared to previous studies, image noise was relatively low given the advanced technology of the CT systems that were used. This limits the influence of noise. Also, statistically, variations in image noise across patients and systems may have been larger than variations in contrast enhancement, so that lack of significance regarding the difference between the three patient groups may be a statistical effect.
Recent improvements in terms of radiation dose management also affect vascular attenuation and noise. For instance, lower kV protocols (80 kV) can now be used for patients with small body mass indices (BMIs) allowing for a reduction in radiation dose while increasing vascular attenuation [bib_ref] Low dose CT of the heart: a quantum leap into a new..., Maffei [/bib_ref]. In this study, no protocol using less than 100 kV was performed. Another recent development affecting noise is related to the FAS full analysis set, CNR contrast to noise ratio, SNR signal to noise ratio, Pre prior to contrast media intravenous administration, Post after contrast media intravenous administration, SD standard deviation [bib_ref] Low dose CT of the heart: a quantum leap into a new..., Maffei [/bib_ref] [bib_ref] Optimizing radiation dose by using advanced modelled iterative reconstruction in high-pitch coronary..., Gordic [/bib_ref]. It is therefore reasonable to expect that using lower kV settings and an iterative reconstruction algorithm may have yielded better SNR values regardless of other parameters. However, the inclusion of iterative reconstruction algorithms might have hampered comparisons to images from other scanners unless image noise would have been adapted by lowering the radiation exposure during the acquisition. Finally, newer detectors have been introduced with increased efficiency and lower noise [bib_ref] Image quality of ultra-low radiation exposure coronary CT angiography with an effective..., Schuhbaeck [/bib_ref] hence promoting the use of less iodine content (less volume or lower concentrations [bib_ref] High-pitch dualsource CT coronary angiography with low volumes of contrast medium, Lembcke [/bib_ref] [bib_ref] Optimizing contrast media injection protocols in state-of-the art computed tomographic angiography, Lell [/bib_ref] [bib_ref] Low contrast and radiation dose coronary CT angiography using a 320-row system..., Oda [/bib_ref] [bib_ref] Effect of reduced x-ray tube voltage, low iodine concentration contrast medium, and..., Yin [/bib_ref] [bib_ref] Automated tube voltage selection for radiation dose and contrast medium reduction at..., Mangold [/bib_ref] [bib_ref] Feasibility of 320-row area detector CT coronary angiography using 40 mL of..., Kim [/bib_ref]. Regarding safety, iobitridol use was associated with the lowest incidence of related AEs (i.e. 1.3 % corresponding to two patients), confirming the excellent safety profile of this CM [bib_ref] Results of the safety and efficacy of iobitridol in more than 61,000..., Petersein [/bib_ref] [bib_ref] Safety of iobitridol in the general population and at-risk patients, Vogl [/bib_ref] [bib_ref] Intravenous contrast media in uroradiology: evaluation of safety and tolerability in almost..., Wendt-Nordahl [/bib_ref].
# Study limitations
The main limitation of the study is the absence of a gold standard such as conventional angiography. Diagnostic accuracy could therefore not be compared between groups. A similar number of scans positive for stenosis in all three groups, however, indicate that there is likely no systematic difference in stenosis detection rates between the three CMs.
Another limitation is the relatively low coronary calcium burden. This is typical for the target population of the study. While severely calcified vessels may benefit from increased vessel lumen opacification for reliable evaluation [bib_ref] Influence of intra-coronary enhancement on diagnostic accuracy with 64-slice CT coronary angiography, Cademartiri [/bib_ref] [bib_ref] Impact of intravascular enhancement, heart rate, and calcium score on diagnostic accuracy..., Cademartiri [/bib_ref] , this was not specifically evaluated and the ability to identify calcified lesions was not compared across the three CMs. A systematic assessment of renal function postcontrast was not performed. Further, contrast injection rates for the three CMs used were constant and not adjusted to achieve equal IDR across the three groups. Hence IDR ranged from 1,400 to 1,600 mg iodine/s in the patients with the lowest body weight and from 2,100 to 2,400 mg iodine/ s in patients with the highest body weight. This reflects clinical practice, where protocols typically prescribe injection rates in ml/s and not IDR, and explains why higher attenuation values were observed for higher concentration CMs, but did not lead to significant differences regarding SNR, CNR or the clinical rating of image quality.
# Conclusion
With current CT technology, iobitridol 350 mg iodine/ml is not inferior to CMs with higher iodine in terms of image quality for coronary stenosis assessment by CTA. When considering image quality, SNR and CNR, iobitridol yielded similar values to iopromide and iomeprol. Iobitridol, with a lower content of iodine, holds the potential to reduce the risk of adverse reactions, as supported by its excellent safety profile. It is likely that developments in image reconstruction and detector technology may further allow improving image quality while minimizing the necessary amount of injected iodine.
[fig] Figure 1: Transaxial cross-sections (0.6-mm slice width) and curved multiplanar reconstructions of the right coronary artery, all displayed at a window level of 1,200 and width of 200 HU. (a, b) Investigation performed using iobitridol 350 mg/ml. (c, d) Investigation performed using iopromide 370 mg/ml. (e, f) Investigation performed using iomeprol 400 mg/ml. LA left atrium, LV left ventricle, RA right atrium, RV = right ventricle [/fig]
[table] Table 1: Patients and procedure characteristics [/table]
[table] Table 4: Secondary efficacy endpointsResults are expressed as mean ± standard deviation or n (%) of patients or segments VAS visual analogue scale (10 cm), F F-test; C Chi-square test *Comfort during examination was rated by the patient on a scale from 1 (very poor) to 5 (very good) [/table]
[table] Table 5: Incidence and characteristics of adverse events (safety set, N = 463) AE adverse events, CM contrast media One patient could have experienced several adverse events introduction of iterative reconstruction algorithms which are able to significantly reduce image noise thus improving SNR and CNR while keeping intravascular attenuation constant [/table]
|
Overexpression of cyclin D1 correlates with early recurrence in superficial bladder cancers.
Cyclin Dl is a cell cycle regulator essential for G1 phase progression and is frequently overexpressed in several human tumour types as a consequence of gene amplification or chromosomal rearrangements. We analysed the expression of cyclin Dl in 75 patients with transitional cell carcinoma (TCC) to investigate the possible relationship between its expression and clinical outcome as well as histopathological findings using the immunohistochemical method. We observed strong staining (++, > 50% positive cells) for cyclin Dl in 19 cases (25.3%) and weak staining (+, 5-50% positive cells) in 19 cases (25.3%). Overexpression of cyclin Dl was not associated with tumour invasion. No significant association was found between overexpression of cyclin Dl and tumour grade (P > 0.05). We assessed the differences of disease-free interval in superficial tumours and actuarial survival probability in invasive tumours according to the status of cyclin Dl expression. Tumours with (++) staining for cyclin Dl recurred much more rapidly than (-) and/or (+) staining tumours (P < 0.01 forvs ++; P < 0.05 for + vs ++). However, overexpression of cyclin Dl was not associated with a shortened overall survival of patients with invasive tumours (P< 0.1). These results suggest that genetic alteration of cyclin Dl appears to be an early event in the tumorigenesis of bladder TCC and is associated with early recurrence in superficial tumours.
Transitional cell carcinoma of the urinary bladder, like many other types of solid tumours, is expected to arise through a series of genetic changes that lead to tumour progression. The identification of the molecular events underlying urothelial cell transformation may not only expand our understanding of the natural history of the disease but may also present useful prognostic markers and potential targets for therapy.
Recent evidence suggests that amplification of the 1 1q 13 region is involved in a variety of human tumours, including bladder carcinoma [bib_ref] Amplification at chromosome 1 1q13 in transitional cell tumors of the bladder, Proctor [/bib_ref] [bib_ref] Expression of cyclin D1 and EMS 1 in bladder tumours: relationship with..., Bringuier [/bib_ref] , head and neck squamous cell carcinoma [bib_ref] Overexpression of cyclin DI correlates with recurrence in a group of forty-seven..., Michalides [/bib_ref] and carcinomas of the oesophagus and breast [bib_ref] Amplification and expression of the human cyclin D gene in esophageal cancer, Jiang [/bib_ref] [bib_ref] A clinicopathological study on overexpression of cyclin DI and of p53 in..., Michalides [/bib_ref]. In the amplified 1 1q13 region, several genes have been identified, of which cyclin Dl is most consistently amplified and overexpressed [bib_ref] Identification and cloning of two overexpressed genes, U21B3 1/PRADI and EMS 1,..., Schuuring [/bib_ref]. Cyclins are thought to be essential proteins in cell cycle regulation because of their specific and periodic expression during cell cycle progression [bib_ref] Cyclin: A protein specified by maternal mRNA in sea urchin eggs that..., Evans [/bib_ref]. Binding of the cyclins with cyclin-dependent kinase regulates their activity and contributes to cell cycle regulation . Cyclin Dl is encoded by the CCND1 gene on chromosome 1 1q13 [bib_ref] Genomic organization, chromosomal localization and independent expression of human cyclin D genes, Inaba [/bib_ref] [bib_ref] Molecular cloning and chromosomal mapping of CCND genes encoding human D-type cyclins, Xiong [/bib_ref] , which has been identified as the PRADI proto-oncogene and the most likely candidate for the BCL1 proto-oncogene [bib_ref] A novel cyclin encoded by a bcl-linked candidate oncogene, Motokura [/bib_ref] [bib_ref] TNM Classification of Malignant Tumours 3rd edn, Withers [/bib_ref].
Several studies have addressed the clinical and prognostic significance of amplification of 1lq13 loci. Amplification of cyclin Dl appears to be correlated with poor prognosis in breast carcinoma, with lymph node involvement and recurrence in head and neck squamous cell carcinoma [bib_ref] Frequent amplification of 1 1q13 DNA markers is associated with lymph node..., Muller [/bib_ref] [bib_ref] Chromosome 1 1q13 gene amplifications in head and neck squamous cell carcinomas:..., Parise 0 Jr [/bib_ref] [bib_ref] Overexpression of cyclin DI correlates with recurrence in a group of forty-seven..., Michalides [/bib_ref]. Although amplification of the 1 q13 region has been observed in bladder carcinoma, the prognostic significance of cyclin Dl overexpression has not been yet reported in bladder carcinoma. The purpose of this study was to detect the expression of cyclin DI in bladder carcinoma tissues and to investigate whether overexpression of cyclin Dl is associated with poor prognosis in patients with bladder carcinoma. For these purposes, we analysed overexpression of cyclin Dl immunohistochemically and reviewed the medical records of 75 patients retrospectively. The relationship of cyclin Dl overexpression to selected clinical variables was also analysed.
## Materials and methods patients
Primary bladder carcinomas from 75 patients were examined (age range 30-83 years; average 62 years). The tumours were staged according to the TNM pathological staging system , and the histological grade was assessed according to [bib_ref] Epithelial tumors of the bladder, Ash [/bib_ref]. The histopathological characteristics are summarized in [fig_ref] Table 1: Comparison of cyclin Dl expression, stage and grade in 75 TCC patients [/fig_ref]. Twenty of the 53 patients with superficial tumours had two or more tumours initially, and concomitant Tis (carcinoma in situ) lesions were not detected in the patients with superficial tumours. All patients were treated with curative intent and had received no prior therapy. Nine patients with solitary, low-grade and Ta tumours were treated with transurethral resection (TUR) only. Patients with multiple, high-grade Ta or TI tumours had received intravesical chemotherapy with doxorubicin (37 cases) or BCG (seven cases) after TUR. In 22 patients with invasive tumour, a total of ten patients underwent radical cystectomy, seven patients
## Immunohistochemistry
All 75 resected specimens were fixed in 10% buffered formalin for about 24 h, embedded in paraffin, and 5-jm thick sections were then deparaffinized. After the sections were heated with a microwave oven (containing 0.01 M sodium citrate buffer pH 6.0; 800 W), endogenous peroxidase was blocked with 3% hydrogen methanol. The sections were washed three times with cold 0.5 M tris-buffered saline (TBS). Inhibition of non-specific binding was accomplished by incubation with normal goat serum (Dako, Carpenteria, CA, USA) for 20 min. Monoclonal mouse antihuman cyclin Dl oncoprotein antibody (1:20 dilution, Novocastra, Newcastle, UK) was applied and incubated for 30 min. The sections were then washed three times with TBS, followed by incubation for 30 min with biotinylated antimouse IgG (Dako). After washing, peroxidase-antiperoxidase conjugate (Dako) was applied. They were then stained with diaminobenzidine tetrahydrochrolide (Dako) and counterstained with Meyer's haematoxylin. We used formalin-fixed, paraffin-embedded WI-38 cells (ATCC, Rockville, MD, USA) for positive control. Negative control sections were obtained by incubation with phosphatebuffered saline instead of monoclonal cyclin DI antibody, and they were consistently negative. Staining intensity was assessed as follows: -, no cancer cells or less than 5% of the cancer cells showed weak or ambiguous staining; +, less than 50% of the British Journal of Cancer (1997) 75(12) cancer cells were stained; ++ more than 50% of the cancer cells showed positive or strongly positive staining [bib_ref] Overexpression of cyclin DI correlates with recurrence in a group of forty-seven..., Michalides [/bib_ref]. Only nuclear staining was observed.
## Statistics
The chi-square test for trend was used to evaluate the statistical significance of the relationship between staining and prognostic variables. Survival curves were prepared using the method of [bib_ref] Nonparametric estimation from incomplete observation, Kaplan [/bib_ref]. The statistical analyses of the differences between curves were performed using the log-rank test [bib_ref] Design and analysis of randomised clinical trials requiring prolonged observation of each..., Peto [/bib_ref].
# Results
## Immunohistochemical staining for cyclin dl
Tumour cell nuclear staining of variable extent was noted in 38 tumours (50.6%), which could be readily divided into groups showing weak (+) and strong staining (++) [fig_ref] Figure 1: Immunohistochemical staining of cyclin Dl in TCC of the urinary bladder [/fig_ref]. Strong staining was observed in 19 of 75 tumours (25%) and weak staining was observed in a further 19 tumours (25%). Analysis of staining for cyclin DI in relation to the clinicopathological characteristics of the patient population is shown in [fig_ref] Table 1: Comparison of cyclin Dl expression, stage and grade in 75 TCC patients [/fig_ref]. Staining for cyclin Dl was observed in 27 of 53 superficial tumours (51%) and in 11 of 22 invasive tumours (50%). Positive staining for cyclin DI was not associated with tumour stage. Positive staining for cyclin DI was found in 21 of 40 (52.5%) well-differentiated tumours (grade I and II), 12 of 26 (46.2%) moderately well-differentiated tumours (grade III) and five of nine (55.6%) poorly differentiated tumours (grade IV). Thus, positive staining for cyclin DI was not also associated with tumour grade (P > 0.05).shows cyclin DI expression in relation to stage, grade, multiplicity and treatment in the 53 patients with superficial bladder tumours. There was little difference in the staining for
## Prognostic significance of overexpression of cyclin dl in superficial tumours
Nine of 14 superficial tumours (64%) with strong staining for cyclin DI recurred within 12 months. On the other hand, only 4 of 26 superficial tumours (15%) with negative staining and 2 of 13 superficial tumours (15%) with weak staining recurred within 12 months (CD1vs ++, P < 0.01; + vs ++, P < 0.05; [fig_ref] Figure 2: Kaplan-Meier disease-free interval curves of 53 patients with superficial bladder cancers [/fig_ref]. In nine patients treated with TUR only, three recurred during the follow-up period. Two cases with negative staining for cyclin Dl recurred in the post-operative period at 22 months and 33 months respectively, whereas one case with strong staining recurred in the postoperative period at 6 months. In 44 patients with TUR and intravesical doxorubicin or BCG therapy, 8 of 12 superficial tumours (67%) with strong staining, 2 of 10 (20%) with weak staining and 3 of 22 (14%) with negative staining recurred within 12 months (CD1 -vs ++, P < 0.005; + vs ++, P < 0.05, [fig_ref] Figure 3: Kaplan-Meier disease-free interval curves of 44 patients with superficial bladder cancer treated... [/fig_ref]. Hence, superficial tumours recurred much more rapidly when the primary tumours showed strong staining for cyclin DI compared with primary tumours showing negative or weak staining, regardless of treatment.
## Prognostic significance of overexpression of cyclin dl in invasive tumours
We assessed the actuarial survival probability in the 22 patients with invasive tumours according to the expression of cyclin Dl.
Three-year actuarial survival rate in the cases with negative staining for cyclin D1 was approximately 65%. On the other hand, in the patients with weak and strong staining for cyclin DI, 3-year actuarial survival rate was about 26%. Thus, there was a tendency for patients with stained tumours to have a poorer prognosis than British Journal of Cancer (1997) 75(12)
# Discussion
The involvement of the chromosome 11 q I human tumours has been described. Cyc candidate oncogene on the 1lq13 amplic consistently amplified and overexpresse tumours [bib_ref] Identification and cloning of two overexpressed genes, U21B3 1/PRADI and EMS 1,..., Schuuring [/bib_ref]. Bind cyclin-dependent kinases regulates their ac cell cycle regulation [bib_ref] Cell division control-related protein kinases: putative origins and functions, Kidd [/bib_ref] ; Mal Several lines of evidence lave sugges involved in the G1 to S transition of the i 1991; [bib_ref] Isolation of three novel human cyclins by rescue of GI cyclin (cln)..., Lew [/bib_ref]. Also, cyclin Dl i regulation through interactions with pRb related proteins, such as PCNA and p21 [bib_ref] Physical interaction of the retinoblastoma protein with human D cyclins, Dowdy [/bib_ref].
Several studies have reported the progn( overexpression of cyclin Dl. In breast c cyclin Dl has been related to poor prog 1992b). When studying squamous cell car neck, amplification of cyclin Dl was asso( involvement and an increased likelihood [bib_ref] Frequent amplification of 1 1q13 DNA markers is associated with lymph node..., Muller [/bib_ref] [bib_ref] Chromosome 1 1q13 gene amplifications in head and neck squamous cell carcinomas:..., Parise 0 Jr [/bib_ref] ; Mi bladder cancer, pRb and various cell cycle to be involved in tumour development ar [bib_ref] Cell division control-related protein kinases: putative origins and functions, Kidd [/bib_ref]. Furthermore, an some 1 q13 has also been found in bladdi [bib_ref] Amplification at chromosome 1 1q13 in transitional cell tumors of the bladder, Proctor [/bib_ref] ; Bringu findings suggest that alterations of cyclin role in the development of bladder cancer.
The present immunohistochemical stud' pression of cyclin Dl in approximately ha cancers. The reported frequency of amp] region was about 10-15% in bladder can [bib_ref] Expression of cyclin D1 and EMS 1 in bladder tumours: relationship with..., Bringuier [/bib_ref]. There are no available data to explain this difference in bladder cancer. However, [bib_ref] Amplification and overexpression of cyclin DI in breast cancer detected by immunohistochemical..., Gillett [/bib_ref] have shown that overexpression of cyclin Dl in breast tumours, as detected by immunohistochemistry, occurred almost twice as frequently as cyclin DI amplification. The possible explanation for this discrepancy is that immunohistochemical overexpression of cyclin Dl results from overall deregulation of cyclin Dl gene expression as well as gene amplification.
We found that the frequency of overexpression of cyclin DI in superficial tumours was almost identical to that in invasive tumours. There was no statistically significant difference between expression of cyclin DI and tumour grade, which is partly consistent with the results of Proctor et al , although they reported that amplification at l1q13 showed no correlation with tumour grade. [bib_ref] Expression of cyclin D1 and EMS 1 in bladder tumours: relationship with..., Bringuier [/bib_ref] reported that low expression of cyclin Dl mRNA correlated with more aggressive TCC. In their report, [bib_ref] Expression of cyclin D1 and EMS 1 in bladder tumours: relationship with..., Bringuier [/bib_ref]. This discrepancy may be largely influenced by the relars for cyclin Dl. But no tively small groups of tumours studied and by the sensitivity of the served between staining methods used to determine overexpression of cyclin Dl. Taken e patients with invasive together, although the frequency of overexpression of cyclin Dl in the present study was different to that reported by [bib_ref] Expression of cyclin D1 and EMS 1 in bladder tumours: relationship with..., Bringuier [/bib_ref] , the results strongly indicate that genetic alterations of cyclin Dl genes may be early events in the multistep carcinogenesis of bladder tumours and are not involved in the develop-13 region in a variety of ment of invasive tumours. :lin Dl is a prominent Bladder cancer can be clinically classified into two groups: -on, because it is most superficial tumours, which are localized in the mucosal or submud in several types of cosal layer, and invasive tumours, which infiltrate into muscular or ing of cyclin Dl with deeper layers. These two types of bladder cancer show signifitivity and contributes to cantly different clinical behaviour. At initial presentation, approxitsushime et al, 1992). mately 50-70% of bladder tumours are superficial. Although ted that cyclin Dl is metastasis is less common with superficial bladder cancers, such cell cycle tumours may progress and the majority will recur and require s involved in cell cycle additional treatment. It is generally accepted that patients with TI, ) and other cell cycle-multiple, large or high-grade tumours are at greater risk of recurfinds et al, 1992; Xiong rence [bib_ref] Superficial bladder cancer: progression and recurrence, Heney [/bib_ref] [bib_ref] Urothelial dysplasia concomitant with bladder tumors: a determinant for future new occurrences..., Wolf [/bib_ref]. In this study, we assessed the probability of disease-free survival in the 53 superostic significance of the ficial tumours according to the overexpression of cyclin Dl.
,ancer, amplification of Interestingly, superficial tumours recurred much more rapidly nosis when the tumours showed strong staining for cyclin Dl compared *cinoma of the head and with tumours showing negative or weak staining. However, over-,ciated with lymph node expression of cyclin Dl was not associated with tumour invasiveof tumour recurrence ness, grade, multiplicity and treatment in superficial tumours. Our chalides . In findings suggest that strong overexpression of cyclin Dl results in -related proteins appear the rapid recurrence of a subset of superficial bladder cancers and nd growth serves as an independent prognostic factor for the prediction of nplification of chromo-early recurrence in superficial tumours.
[er cancer With respect to prognosis, several reports on other types of tier . These carcinoma have stressed that amplification and overexpression of Dl play an important cyclin Dl correlates with poor prognosis [bib_ref] Association of int-2/hst-1 coamplification in primary breast cancer with hormone-dependent phenotype and..., Borg [/bib_ref] [bib_ref] Frequent amplification of 1 1q13 DNA markers is associated with lymph node..., Muller [/bib_ref]. The present study y demonstrated overexshows that, in patients with invasive tumours, there was a lf of 75 urinary bladder tendency for patients with overexpression of cyclin Dl to have a lification of the 1 lq13 worse prognosis than patients with negatively stained tumours for icers [bib_ref] The involvement of the chromosome I Iq13 region in human malignancies: cyclin..., Schuuring [/bib_ref] cyclin Dl, although the correlation between the overexpression of cyclin Dl and actuarial survival was not statistically significant. The weak correlation found between expression of cyclin Dl and actuarial survival in the present study of bladder cancer does not exclude a potential role as a prognostic factor. To investigate this possibility, further studies based on larger numbers of cases with complete follow-up data will be needed.
[fig] Figure 1: Immunohistochemical staining of cyclin Dl in TCC of the urinary bladder. Notice the specific and strong positive nuclear staining of the tumour cells and the absence of staining in the stroma cells (x200) received radiotherapy and five patients were treated with systemic chemotherapy. Patients were followed up for a maximum of 79 months; the median follow-up time was 35 months. [/fig]
[fig] Figure 2: Kaplan-Meier disease-free interval curves of 53 patients with superficial bladder cancers. (-) vs (++) staining, P< 0.01; (+) vs (++) staining, P < 0.05; (-) and (+) vs (++) staining, P < 0.005 [/fig]
[fig] Figure 3: Kaplan-Meier disease-free interval curves of 44 patients with superficial bladder cancer treated by intravesical chemotherapy or immunotherapy after TUR. (-) vs (++) staining, P < 0.005; (+) vs (++) staining, P < 0.05; (-) and (+) vs (++) staining, P < 0.005 cyclin D1 based on tumour invasiveness, grade, multiplicity and treatment. [/fig]
[fig] Figure 4: Kaplan-Meier actuarial survival curves bladder cancers. (-) vs (+) and (++) staining, P < the patients with negatively stained tumoui statistically significant association was obs for cyclin Dl and actuarial survival in the tumours (P < 0.1;). [/fig]
[table] Table 1: Comparison of cyclin Dl expression, stage and grade in 75 TCC patients [/table]
|
Perception of the association between erectile dysfunction and cardiovascular disease among Chinese physicians: an online survey
Objective: There is a close association between erectile dysfunction (ED) and cardiovascular disease (CVD). This study aimed to investigate Chinese physicians' understanding of this association. Methods: A total of 651 physicians, including 245 cardiologists and 406 urologists, participated in our investigation through WeChat. Results: Participants with more professional experience, a doctoral/postdoctoral degree, and an intermediate/senior title were significantly more likely to be aware of a close association between ED and CVD. Urologists had a significantly better understanding of the association of severity between both diseases, showed more positive attitudes towards phosphodiesterase type 5 inhibitor application in patients with CVD and systematic treatment, and gave greater consideration to both diseases during follow-up visits than did cardiologists. Men had a significantly better
# Introduction
Erectile dysfunction (ED), which is defined as persistent impotence to reach and maintain an erection hard enough for successful vaginal intercourse, is a common clinical entity. Despite improvements in outcomes, cardiovascular disease (CVD) remains a leading cause of morbidity worldwide.An increasing number of studies have indicated a particularly close association between ED and CVD.ED and CVD share common risk factors, including age, hypercholesterolemia, hypertension, insulin resistance and diabetes, smoking, obesity, metabolic syndrome, sedentary lifestyle, and depression.The incidence of ED in men with established chronic coronary artery disease is as high as 75%.Coronary artery disease, stroke, and peripheral artery disease usually occur after the onset of ED, following a significant period ranging from 2 to 5 years in the same patient (3 years on average). The European Guidelines on CVD and ED indicate the important clinical significance of ED as a predictor of CVD.The severity of erectile dysfunction, which is evaluated by the 5-item version of the International Index of Erectile Function, was also reported to correlate with CVD and may lead to a difference in specific CVD outcomes.Therefore, cardiologists' and urologists' perceptions of the associations among the pathogenesis, manifestation, diagnosis, and treatment of ED and CVD may help them seek cooperation to systematically manage patients with ED and CVD. ED has gradually and widely been accepted as a significant predictor of CVD.Urologists cooperating with cardiologists may evaluate the risk of CVD in patients with ED. Phosphodiesterase type V (PDE5) inhibitors, which are currently the most effective treatment for ED, have a positive effect on the cardiovascular system.Therefore, cardiologists can provide more assistance to patients with CVD and ED with the help of urologists. To the best of our knowledge (based on searches of Google Scholar, PubMed, Embase, and the Cochrane Library), few relevant studies have been conducted regarding this issue. Therefore, this study aimed to investigate Chinese physicians' understanding of the association between ED and CVD. We hope that this study will lead to more attention towards this association and create a new pattern of systematic management in patients with ED and CVD under collaboration.
# Material and methods
## Survey and participants
This study was a survey that was conducted over 3 months from November 2018 to January 2019. The English version this survey is available in the appendix located at the end of the manuscript. The website link of the questionnaire was distributed in academic WeChat Groups consisting of only physicians (in this manuscript, the term "physicians" includes medical practitioners in cardiology and urology) specializing in cardiology or urology throughout China. WeChat, which is the most popular communication tool in China with more than one billion monthly active users presently,and is a time-effective, costeffective, and convenient method, was used in our investigation. WeChat is frequently involved in doctors' daily lives for medical consultations, treatment, and follow-up,and connection and discussion of medical topics among colleagues.Participants were able to click the website link of the questionnaire shared in WeChat groups and then start the survey. Under each WeChat ID (identity), the questionnaire could be completed only once. All subjects provided written informed consent and were required to respond voluntarily and under the condition of anonymity. The investigation and all relevant protocols were evaluated and approved by the ethics committee of Xiangya Hospital, Central South University.
The survey involved two sections, including the demographic information section and the cognitive section. The demographic information section included age, sex, professional experience, educational background, professional title ("primary" referred to residents, "intermediate" referred to attending physicians, and "senior" referred to chief physicians or deputy chief physicians), and the level of the hospital. The cognitive section aimed to uncover participants' understanding of the association between ED and CVD.
# Statistical analysis
Data are presented as the mean AE standard deviation when normally distributed, median (quartiles) when non-normally distributed, and percentages for categorical data. Data were analyzed using SPSS Statistics, version 22 (IBM Corp., Armonk, NY, USA). The v 2 test was used to compare differences between participant groups. Univariate regression analysis was used to identify the effect of various factors on participants' simultaneous or systemic treatment of ED and CVD. Odds ratios (ORs) are reported with 95% confidence intervals. A P value <0.05 was considered statistically significant.
# Results
## Overview of the survey
A total of 1073 physicians visited our questionnaire through WeChat, with 664 participants from 24 provinces in China (a total of 34 provinces), including 250 cardiologists and 414 urologists who completed the interview. Thirteen questionnaires (five in cardiology and eight in urology) were excluded because of incorrect information in age or professional years. The mean age of the remaining 651 participants was 34 AE 9 years and the mean professional experience was 9.3 AE 8.5 years. A total of 63.4% of participants were men and the proportion of male sex was significantly lower in cardiology compared with urology (35.1% vs 80.5%, P <0.001). The rates of participants with primary and senior titles were 51.8% and 17.1%, respectively. A total of 17.1% and 30.6% of participants were from community and secondary hospitals, respectively, and the remaining 52.2% of participants were from tertiary hospitals. No significant differences were detected between cardiologists and urologists, except for sex.
Characteristics of the participants based on the understanding of the association between ED and CVD A total of 552 (84.8%) participants believed that ED was associated closely with the risk of CVD, while 99 (15.2%) did not. Age, sex, and hospital level did not affect participants' belief of the association between ED and the risk of CVD. Participants with more professional experience (P ¼ 0.028) or a doctoral degree (P ¼ 0.042) were significantly more likely to believe in a close relationship between ED and CVD. However a significantly higher rate of participants with a primary title considered that ED was not closely associated with CVD (P ¼ 0.010).
## Cardiologists versus urologists
Most participants from cardiology or urology were aware of the close association between ED and the risk of CVD. However, a significantly higher rate of cardiologists considered ED as a natural aging process rather than a disease compared with urologists (P ¼ 0.002). Furthermore, a significantly higher percentage of urologists believed that the severity of ED was correlated with that of CVD compared with cardiologists (P <0.001).
The most popular response when considering the most important association between ED and CVD included common risk factors (40.0%), common pathophysiology (35.1%), common concomitant diseases (12.9%), common therapeutic methods or drugs (8.9%), and others (3.1%). A significantly higher proportion of urologists considered common pathophysiology as the most important association than did cardiologists (P <0.001), while more cardiologists considered common concomitant diseases (P <0.001) . The most popular response when considering major clinical significance of the association between ED and CVD included diagnosis (56.3%), treatment (30.3%), prognosis (10.7%), and others (2.8%). A significantly higher proportion of cardiologists considered diagnosis as the major clinical significance compared with urologists (P <0.001), while significantly more urologists than cardiologists chose treatment (P <0.001) . A total of 10.6% (26/245) of cardiologists indicated that they would inquire about the problem of ED in recording the patient's history in patients with CVD and 23.3% (57/245) would provide a further test when ED was diagnosed in a patient with CVD. However, 89.7% (364/406) of urologists would take into consideration cardiovascular problems as a regular item in recording the patient's history and 75.9% (308/406) would evaluate cardiovascular problems with a further test.
When a patient was diagnosed with both ED and CVD, cardiologists appeared to manage ED and CVD together or systematically less actively than did urologists (26.9% vs 88.2%, P <0.001). A significantly higher rate of cardiologists believed that a PDE5 inhibitor was harmful to patients with CVD compared with urologists (P ¼ 0.001). Furthermore, a significantly higher rate of cardiologists expressed concern about problems in safety caused by PDE5 inhibitor use compared with urologists (P <0.001). However, a significantly higher rate of urologists expressed concern about the validity of a PDE5 inhibitor compared with cardiologists (P ¼ 0.021). During a follow-up visit, a significantly higher rate of urologists paid more attention to both disorders than did cardiologists (P <0.001).
## Male participants versus female participants
Significantly more male participants regarded ED as a disease rather than a natural aging process compared with female participants (P ¼ 0.004). Significantly more male participants believed that there was an association between the severity of ED and the risk of CVD compared with female participants (P ¼ 0.018). Additionally, a significantly higher proportion of male participants managed both diseases together actively (P <0.001) and gave consideration to both diseases during follow-up visits, compared with female participants (P <0.001).
## Factors associated with participants' systematic treatment of ed and cvd
Urologists (OR ¼ 20.228, P <0.001), !10-year professional experience (OR ¼ 2.000, P <0.001), <doctoral degree (OR ¼ 1.936, P ¼ 0.002), male sex (OR ¼ 6.091, P <0.001), and being from a tertiary hospital (OR ¼ 1.657, P ¼ 0.002) were significantly associated with conducting treatment on ED and CVD together more actively. Furthermore, those who believed that there was an association between the severity of ED and CVD (OR ¼ 1.943, P <0.001) and those who showed concern about PDE5 inhibitor administration in patients with CVD (OR ¼ 1.656, P ¼ 0.010) managed ED and CVD together more actively.
# Discussion
The existence of an association between ED and CVD is widely accepted. Not only do ED and CVD share common risk factors, but they also progress under a common pathophysiology (endothelial dysfunction).Major vascular beds are uniformly affected by endothelial dysfunction.Arteries supplying various areas differ in size (penis arteries are smaller than those in the heart, brain, and lower limbs). Therefore, penis arteries are affected first . (a) The most important association between ED and CVD according to cardiologists' and urologists' responses to question 9 in the questionnaire ("What is the most important association between ED and CVD?"). **P < 0.001. (b) Major clinical significance of the association of ED and CVD according to cardiologists' and urologists' responses to question 10 in the questionnaire ("What is the major instructive clinical significance of ED or CVD taking into account the association between these two disorders?"). **P < 0.001. by endothelial dysfunction because of their smaller size.Consequently, ED can be regarded as an early warning system or the "tip of the iceberg" of a systemic vascular disease. Novel therapies based on molecular mechanisms of ED may serve as promising therapies for CVD characterized by endothelial dysfunction.
Participants with more experience and better education are more likely to believe in a close association between ED and CVD In our study, the majority of participants (552/651, 84.8%) believed in a close association between ED and CVD. Additionally, their age and sex did not significantly affect their understanding of this association. The participants' understanding of this association may be a result of research focusing on the association between ED and CVD, and this association has been clarified in ED and CVD guidelines.Longer professional experience and a better educational background contributed to the participants' understanding of this association, but the level of their affiliated hospital did not.
## Urologists show a better understanding of the association between ed and cvd than do cardiologists
Participants in cardiology and urology mostly agreed that there was a close association of ED and CVD. In cardiologists, the most popular answer for the most important association between ED and CVD was common risk factors, whereas common pathophysiology was the most popular in the urology group. In fact, common pathophysiology (endothelial dysfunction) fundamentally leads to the close association between ED and CVD.Therefore, urologists may have a better understanding of the association between these disorders. Additionally, among cardiologists, the most popular answer for major instructive clinical significance of the association between ED and CVD was diagnosis (71.4%), whereas treatment (38.6%) was the most popular in urologists. This difference between the specialists may be a consequence of their different professional backgrounds. ED being considered a predictor of CVD has gradually been accepted by cardiologists.However, urologists may pay more attention to the benefit to patients with CVD due to treatment of ED, including modification of risk factors and drugs, such as PDE5 inhibitors. In our study, a significantly higher proportion of urologists considered that the severity of ED was associated with the extent of CVD compared with cardiologists. The risk of CVD and the number of deaths from CVD have increased steadily with the severity of ED.The number of obstructed vessels detected by coronary angiography is correlated with the severity of ED in male patients with acute myocardial infarction.Therefore, urologists may develop a better understanding than cardiologists of the association of severity of these diseases.
In the current study, only a small amount of cardiologists regularly inquired about problems of ED in cardiovascular patients and assessed ED with further tests when ED was diagnosed in a patient with CVD. In contrast, the majority of urologists inquired about CVD in patients with ED and evaluated these patients with further tests. Several factors may be responsible for this large difference between specialists. First, in our survey, most cardiologists regarded ED as a natural progression that occurs with age rather than a disorder. As a result, they may not pay attention to problems of ED, even though the problems are associated with CVD and are extremely common in men with cardiovascular problems.Second, urologists may develop a better understanding of this association, and therefore, may pay more attention to it.
We also found that cardiologists appeared to manage ED and CVD together or systematically less actively than did urologists. Additionally, a significantly lower rate of cardiologists gave consideration to both ED and CVD during the follow-up visit than did urologists. Cardiologists' negative attitude towards managing ED and CVD together or systematically in the same patient may be partly a result of their characterization of ED as a natural agerelated process. We also found that most cardiologists worried about the harmful effects of PDE5 inhibitors on patients with CVD, which may be another obstruction in the management of both disorders. In fact, a strong body of clinical data has confirmed the tolerability and established safety of PDE5 inhibitors in patients with or without CVD.Additionally, PDE5 inhibitors are not associated with an increased risk of exacerbating ischemia and worsening exercise tolerance in patients with coronary artery disease, even though PDE5 inhibitors are widely known as a contraindication to nitrates or other nitric oxide donors. Finally, lifestyle modification and pharmacotherapy, according to recommendations for reducing the risk of CVD, are effective in improving erectile function.PDE5 inhibitors, which are the first-line and most effective therapies for ED currently available,are generally reported to have a positive effect on the cardiovascular system.Urologists might manage comorbid ED and CVD more actively because they have noted improved benefits that result from systemic treatment of both diseases. Additionally, cardiologists' excessive concerns about PDE5 inhibitors may be unnecessary. Of course, appropriate clinical assessment and a medication review should be performed before administration of PDE5 inhibitors on patients with CVD.
## Male participants show a better understanding of the association between ed and cvd than do female participants
Generally, we found that male physicians showed greater understanding of the association between ED and CVD and were more active in managing ED and CVD together or simultaneously than female physicians. Several possibilities may contribute to these findings. The main reason might be that doctors find communicating about sexual functioning with an opposite sex patient unpleasant, especially when the patient is young or much older in age.ED and CVD usually coexist in older male patients; therefore, female physicians may be more uncomfortable in inquiring about details of sexual illnesses. Moreover, the conservative social atmosphere, which is rooted in traditional Chinese culture, may increase female physicians' discomfort in communicating about sexual matters with older male patients.As a result, female participants in our study showed a lower understanding of the association between ED and CVD and less active management of comorbid ED and CVD. In our survey, there was a significantly higher rate of female cardiologists than female urologists. Female participants' worse understanding of the association between ED and CVD may also be responsible for cardiologists' poorer overall understanding of this association.
Participants' department, sex, professional experience, educational background, and understanding of the association between ED and CVD affect their systematic management of patients with ED and CVD We found that department, sex, professional experience, educational background, and the level of the affiliated hospital were factors that affected participants' management of comorbid ED and CVD. However, participants with a superior education who had a better understanding of the association between ED and CVD, showed less active involvement in systematic treatment of both diseases. This finding may be because several extra years of education are required to obtain a doctoral or postdoctoral degree before clinical work, which may lead to a lack of professional experience, which is another significant factor associated with systematic treatment of ED and CVD. Additionally, whether participants managed comorbid ED and CVD appeared to be associated with their understanding of the association of severity of both diseases and concern about the administration of PDE5 inhibitors in patients with CVD. Those with a better perception of this association appeared to have better performance in managing ED and CVD together or systematically.
# Limitations
Our study had several limitations. First, our survey was conducted using an online platform, namely WeChat (non-WeChat methods were not included), which has become the most widely and frequently used social media platform in China. In fact, WeChat has nearly saturated its growth potential in China, with more than 1 billion monthly active users at present.Although WeChat users were more evenly distributed by age group in 2017 than they were in 2015, young users are still the predominant age group using the platform.Our study might have ignored a small number of participants who were not users or frequent users of WeChat, especially older participants. Second, there might have been some bias due to different professional backgrounds. Urologists and cardiologists may have had a different understanding of the same questions in our questionnaire. Third, participants may have responded to the questionnaire with what they felt was the appropriate answer rather than responding with the truth. Fourth, the questionnaire was not validated because it was not strictly a dichotomously scored questionnaire. Finally, the results may have been more accurate if the sample size had been larger.
# Conclusion
Physicians from cardiology or urology have a primary awareness of the association between ED and CVD. Urologists have a better understanding of this association than do cardiologists, and male participants had a better understanding than did female participants. Additionally, participants with more professional experience, a better educational background, and a higher professional title had better understanding of the association between ED and CVD. However, understanding of this relationship among physicians in cardiology or urology may still be insufficient and should be further improved.
## Declaration of conflicting interest
The authors declare that there is no conflict of interest.
# Funding
This study was supported by supported by the Fundamental Research Funds for the Central Universities of Central South University (2019zzts1039 to Zhang-Cheng Liao) and the National Natural Science Foundation of China (81570627 to Zheng-Yan Tang).
## Orcid id
Zheng-Yan Tang https://orcid.org/0000-0003-0693-8748 |
Spontaneous Pneumomediastinum from Running Sprints
Spontaneous pneumomediastinum (SPM) is a fairly rare condition, caused by increased intrathoracic pressure, leading to free air in the mediastinal structures. Underlying lung conditions are associated with increased incidence of SPM, including asthma, interstitial lung disease, pneumonia, bullous lung, and radiation therapy for lung cancer. It is often preceded by Valsalva maneuvers, vomiting, coughing, asthma exacerbation, sneezing, childbirth, or intense physical activity. A case of SPM is presented in a 15-yearold male, who complained of throat pain and dyspnea while running sprints at football practice. Workup revealed SPM, and he was subsequently admitted and treated conservatively. His symptoms resolved in 2 days and he was discharged and suffered no further recurrences. In contrast to secondary pneumomediastinum, SPM is usually a benign condition although life-threatening conditions can rarely arise. Differentiating between these two conditions has important prognostic indications. There is a paucity of prospectively collected data regarding SPM, and considerable variation in recommendations concerning the extent of workup.
# Introduction
Spontaneous pneumomediastinum (SPM)-pneumomediastinum without an obvious precipitating cause-is a fairly rare condition, caused by increased intrathoracic pressure, leading to free air in the mediastinal structures. Although SPM is generally considered a benign condition, secondary pneumomediastinum may be due to a serious medical condition, such as perforation of an internal organ seen in Boerhaave syndrome. This is a case of a 15-year-old male who suffered from SPM on the first day of American football practice while running sprints in a conditioning exercise. The etiology of SPM is discussed along with the common signs and symptoms associated with this condition.
## Case presentation
A fifteen-year-old male presented at 0100 to the local emergency department complaining of increasing dyspnea, chest pain and a sore throat. He denied any nasal discharge, cough, nausea or vomiting. His symptoms had started the previous afternoon on the first day of football practice. He had been running 24 100-yard sprints as a conditioning exercise, and about 2/3-way through, he started becoming symptomatic. There had been no contact drills. He had been able to complete the exercise, but his symptoms continued to worsen throughout the evening. He never had a similar episode. He did not recall any Valsalva-like maneuvers, nor did he lift any weights.
He denied being a smoker and had no past history of asthma, respiratory problems, or other significant health issues. He denied any illicit or recreational drug use.
On exam, the patient was a 6 3 , 270 lbs, large teenager who appeared slightly dyspneic, but not distressed. Blood pressure was 128/70, heart rate 80, respiratory rate 28, and temperature 97.8 F; SaO 2 was 97% on room air. HEENT exam was unremarkable, and the oral cavity showed normal appearing tonsils and a midline uvula. Speech was normal sounding, with no nasal quality. There was no erythema and no exudates. There was no subcutaneous emphysema or crepitus in the neck or the chest area. Lungs were clear to auscultation bilaterally with no adventitial lung sounds. CV showed a regular rate and rhythm with no murmurs, 2 Case Reports in Medicine clicks, gallops, or extra heart sounds, specifically no systolic crepitations. Abdomen was benign, and the rest of the exam was unremarkable.
Chest X-ray [fig_ref] Figure 1: Initial chest X-ray showed subcutaneous emphysema, with air tracking into the neck... [/fig_ref] showed subcutaneous emphysema, with air tracking into the neck area bilaterally, and a para-aortic air stripe of the left. There was also evidence of a minimal left apical pneumothorax (PNX). The trachea was midline, and the remainder of the lung fields and cardiac silhouette were normal.
The patient was diagnosed with pneumomediastinum and transferred to a tertiary care facility where he was admitted, placed on high flow oxygen, and observed. Later that morning, he was still noted to be dyspneic and complaining of chest pain. The following morning, the patient was asymptomatic, and was discharged. At a followup visit to his pediatrician 2 weeks later, the chest X-ray [fig_ref] Figure 2: Follow-up chest X-ray, expiratory view, two weeks later, shows resolution of previous... [/fig_ref] , expiration view) showed complete resolution of his previous lung findings, and he was given clearance to return to physical activity.
# Discussion
Pneumomediastinum, or mediastinal emphysema, is the presence of free air in the mediastinal structures. It is usually credited as first being described in 1819, by the pathologist R. T. Läennec, who first observed this condition in trauma patients [bib_ref] Spontaneous pneumomediastinum: clinical and natural history, Panacek [/bib_ref].
Primary or Spontaneous Pneumomediastinum (SPM) is caused by increased pressure gradient between the alveoli and pulmonary interstitium [bib_ref] Subcutaneous and mediastinal emphysema. Pathophysiology, diagnosis, and management, Maunder [/bib_ref]. Secondary causes include gas-forming microorganisms during an infection, and increasing intrathoracic pressure from trauma causing a disruption of bronchiolar, alveolar, or rarely, esophageal tissue, with leakage of air into the interstitial tissues [bib_ref] The incidence and clinical significance of pneumomediastinum found on computed tomography scan..., Molena [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref].
Spontaneous Pneumomediastinum has no apparent secondary cause, and is relatively rare in adults, (1 out of 44,511 ED visits) [bib_ref] Spontaneous pneumomediastinum: a rare complication of bronchial asthma, Faruqi [/bib_ref]. Children have increased frequency (1 in 800 to 1 in 15,150 ED visits) [bib_ref] Etiologies of spontaneous pneumomediastinum in children of different ages, Lee [/bib_ref]. The initial description and defining of this condition is credited to Hamman in 1939 [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] although reports of spontaneous pneumomediastinum date back to the 17th century, where it was observed in extreme effort during childbirth [bib_ref] Subcutaneous and mediastinal emphysema. Pathophysiology, diagnosis, and management, Maunder [/bib_ref].
The mechanism in SPM, as outlined by Macklin [bib_ref] Subcutaneous and mediastinal emphysema. Pathophysiology, diagnosis, and management, Maunder [/bib_ref] , is disruption of the alveolar tissues (the Macklin phenomenon [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref]. Increased intra-alveolar pressure causes air to leak into the interstitium and via centripetal forces (the Macklin effect [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] to collect centrally [bib_ref] Spontaneous pneumomediastinum: a rare complication of bronchial asthma, Faruqi [/bib_ref]. This process is driven by pressure differences between the mediastinum and the lung parenchyma during the respiratory cycle [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref]. Common causes of increased intra-alveolar pressure include acute asthma exacerbation, Valsalva maneuvers, childbirth, intense athletic activity, coughing or vomiting. Decrease in intrapleural pressure or in the lung interstitium, as in vasoconstriction, or loss of integrity of the alveolar-capillary membrane, as in interstitial lung disease, predisposes patients to SPM. Inhalation drugs, such as cocaine, may precipitate SPM by multiple mechanisms of vasoconstriction and interstitial lung disease [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref] [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] [bib_ref] Pneumomediastinum: is it really a benign entity? When it can be considered..., Perna [/bib_ref].
Most sports-related cases of pneumomediastinum involve contact although SPM has been reported after scuba diving and synchronized swimming [bib_ref] Pneumomediastinum and subcutaneous emphysema in a synchronized swimmer, Leiber [/bib_ref].
Pneumomediastinum differs from pneumothorax in that, in the latter, there is a disruption of the parietal pleura with collection of air in the pleural space; in pure pneumomediastinum the pleura remains intact.
The epidemiology of SPM varies depending on the study. Gender distribution has been reported as equal [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref] although many series show a predominance in males (57-87%) [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref] [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] [bib_ref] Pneumomediastinum: is it really a benign entity? When it can be considered..., Perna [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: is a chest X-ray enough? A single-center case series, Esayag [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref]. It is most common in the 2nd and 3rd decades of life [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref].
Underlying lung conditions are associated with increased incidence of SPM. Asthma (8-39%), interstitial lung disease (18%), pneumonia, bullous lung, and radiation therapy for lung cancer have all been associated with SPM [bib_ref] Spontaneous pneumomediastinum: a rare complication of bronchial asthma, Faruqi [/bib_ref] [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref].
Frequent preceding events include vomiting (24-36%), asthma exacerbation (15-24%), coughing (7-35%), sneezing (13%), breath holding, physical activity (30%), increased effort during labor (4-15%), and screaming and even snorting, a sharp exhale as a sign of psychological stress [bib_ref] Class test stress as an unusual cause of spontaneous pneumomediastinum in a..., Kribs [/bib_ref]. One series identified a 55% rate of cocaine consumption in the hours prior to presentation [bib_ref] Pneumomediastinum: is it really a benign entity? When it can be considered..., Perna [/bib_ref]. Anorexia nervosa has also been associated with SPM; it has been suggested a decrease in pulmonary tissue elasticity found in this condition may predispose patients to alveolar air leakage [bib_ref] Spontaneous pneumomediastinum: a complication of anorexia nervosa?, Fergusson [/bib_ref]. Up to 32-66% of incidences of SPM have no identified precipitating event [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref] [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: is a chest X-ray enough? A single-center case series, Esayag [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref].
## Case reports in
Symptoms include chest pain (26-70%), dyspnea (26-46%), cough (26-45%), sore throat (18%), neck pain (4-38%), rhinolalia (65%), hoarse voice (65%), neck and/or chest swelling (87%) [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] [bib_ref] Pneumomediastinum: is it really a benign entity? When it can be considered..., Perna [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref].
Signs include systolic crepitations (Hamman's sign), an abnormal crunchy or bubbling sound heard during cardiac auscultation (8-17%), subcutaneous emphysema (12-100%), neck swelling (14%), and hoarse voice and rhinolalia (65%) [bib_ref] Spontaneous pneumomediastinum: clinical and natural history, Panacek [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref] [bib_ref] Etiologies of spontaneous pneumomediastinum in children of different ages, Lee [/bib_ref] [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] [bib_ref] Pneumomediastinum: is it really a benign entity? When it can be considered..., Perna [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: is a chest X-ray enough? A single-center case series, Esayag [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref].
In an exclusive series of children (9 months to 18 years), Lee found a similar distribution of symptoms, signs, and preceding events and PNX (17%) [bib_ref] Etiologies of spontaneous pneumomediastinum in children of different ages, Lee [/bib_ref].
Diagnosis is made most commonly by chest X-ray, which reveals subcutaneous emphysema, and pneumomediastinum. Small pneumomediastinum not seen on chest Xray will be seen on chest CT although the use of CT only to confirm the diagnosis of SPM in an otherwise well-appearing patient is probably not warranted [bib_ref] Spontaneous pneumomediastinum: is a chest X-ray enough? A single-center case series, Esayag [/bib_ref]. Ruling out other causes of pneumomediastinum is important, as secondary pneumomediastinum can be associated with a poor outcome [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref].
Treatment is most often conservative; complications, although not common, include tension pneumopericardium; massive subcutaneous emphysema causing clinical effects can be drained with a silastic drain [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref]. Concurrent pneumothorax is found in 6-32% of cases in SPM [bib_ref] Pneumomediastinum: is it really a benign entity? When it can be considered..., Perna [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref] , and should be treated conventionally. Of the 20 out of 62 (32%) patients reported by Iyer to have a concurrent PNX, only 6 had tube thoracostomy. The 13 patients with PNX admitted without chest tube placement had the same length of stay as those without PNX [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref]. Rarely, tension pneumomediastinum occurs, requiring surgical drainage and placement of a drain. In the absence of gastrointestinal complaints or findings, further study of the esophagus and GI tract is not recommended [bib_ref] Spontaneous pneumomediastinum: is a chest X-ray enough? A single-center case series, Esayag [/bib_ref]. The use of antibiotics in the absence of obvious infection is also not recommended [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref].
Resolution of free air typically takes 2-4 days; recurrence is rare, but has been reported [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref] , often related to inhalant recreational drug use [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref]. Patients with underlying medical conditions, such as bullous lung disease, interstitial lung disease, or graft versus host disease have a more prolonged hospital course. In these cases, the hospital course is often determined more by an underlying condition rather than SPM.
A review of the literature reveals most articles regarding SPM are retrospective case series; as such, their information is dependent on relative completeness of documentation. Patients may have had more symptomatology then the treating physicians reported in the patient's chart. There was also significant variability in their patient populations, and in the inclusion criteria as to what constituted SPM.
Considerable disparity in recommendations exists between authors, with no clear agreement on the use of CT, hospitalization, or workup for potential esophageal disruption. Also problematic is the diagnosis of SPM cannot be made until other causes of pneumomediastinum have been ruled out. The distinction between SPM and secondary pneumomediastinum is important; causes of secondary pneumomediastinum, such as Boerhaave syndrome, can carry high mortality and morbidity rates. This may necessitate an extensive workup, including further diagnostic studies. Some authors recommend chest CT to all patients to rule out undiagnosed/unrecognized chest pathology [bib_ref] Pneumomediastinum: is it really a benign entity? When it can be considered..., Perna [/bib_ref].
This is a case of a 15-year-old male who presented with SPM following vigorous noncontact exercise. He had no identifiable underlying lung conditions. The presumed pathophysiology in this case appears to be increased intrathoracic pressure occurring during extreme physical exercise. Although the patient denied recalling any Valsalvalike maneuvers, the effect of running multiple sprints could certainly raise intrathoracic pressure. The association between extreme exertion and SPM is will documented [bib_ref] Spontaneous pneumomediastinum: diagnostic and therapeutic interventions, Al-Mufarrej [/bib_ref] [bib_ref] Diagnosis and treatment of spontaneous pneumomediastinum, Gunluoglu [/bib_ref] [bib_ref] Pneumomediastinum: is it really a benign entity? When it can be considered..., Perna [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: a comparative study and review of the literature, Caceres [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: is a chest X-ray enough? A single-center case series, Esayag [/bib_ref] [bib_ref] Spontaneous pneumomediastinum: analysis of 62 consecutive adult patients, Iyer [/bib_ref]. In the absence of GI symptoms, no further esophageal or GI workup was performed. No chest CT was performed. The patient was observed, and was asymptomatic 36 hours later and was discharged. Repeated CXR was normal after 2 weeks. There is limited prospectively acquired data regarding workup and management of SPM. In the era of cost containment, a prospective study looking at optimizing medical resources for SPM may be warranted. GI workup may be reserved for patients with significant GI symptoms [bib_ref] Spontaneous pneumomediastinum: is a chest X-ray enough? A single-center case series, Esayag [/bib_ref]. Small concomitant PNX can be managed conservatively. Most patients are asymptomatic 24-48 hours after presentations. IV antibiotics do not appear necessary in the absence of obvious infections.
[fig] Figure 1: Initial chest X-ray showed subcutaneous emphysema, with air tracking into the neck area bilaterally (chevron), and a paraaortic air stripe of the left (arrow). [/fig]
[fig] Figure 2: Follow-up chest X-ray, expiratory view, two weeks later, shows resolution of previous findings. [/fig]
|
Plastidial Phosphoglucomutase (pPGM) Overexpression Increases the Starch Content of Transgenic Sweet Potato Storage Roots
# Introduction
Starch is the main storage form of carbohydrates in plants. During the day, leaves synthesize transitory starch in the chloroplasts through photosynthesis. It is then degraded into sucrose at night and transported to non-photosynthetic organs as an energy source [bib_ref] Starch: Its metabolism, evolution, and biotechnological modification in plants, Zeeman [/bib_ref] [bib_ref] HPLC-MS/MS analyses show that the near-starchless aps1 and pgm leaves accumulate wild..., Bahaji [/bib_ref]. ADP-glucose (ADPG) is the transitory starch substrate in chloroplasts and is synthesized by the successive catalysis of plastidial glucose phosphate isomerase, phosphoglucomutase (PGM) and ADP-glucose pyrophosphorylase. The ADPG is then transformed into amylose or amylopectin by the starch synthases [bib_ref] The debate on the pathway of starch synthesis: A closer look at..., Streb [/bib_ref].
The interconversion between glucose-1-phosphate (G1P), the precursor of ADPG, and glucose-6-phosphate (G6P) is catalyzed by PGM. Two types of PGM exist in plants, Primers starting with "q" are used for the qRT-PCR. The underlined nucleotides are restriction enzyme sites.
## Expression analysis of ibppgm in the sweet potato
The total RNA was isolated from five different tissues (storage root, fibrous root, stem, leaf, and petiole) of Xu 781 plants grown in the field for approximately 100 days, and the first-strand cDNA was synthesized using PrimeScriptTM RT Reagent Kit with gDNA Eraser (Perfect Real Time) (TaKaRa, Beijing, China). The qRT-PCR was conducted on a 7500 Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) to determine the IbpPGM transcript levels using the gene-specific primers qPGM-F/R. The control, IbActin, was amplified using the primers Actin-F/R.
The Xu 781 plants that were grown in the field for approximately one month, were used to investigate the response of IbpPGM to exogenous sucrose. Briefly, the leaf-petioles (10 cm) of the Xu 781 plants were cultured in water in darkness for 1 d as a starvation treatment. Then, they were supplied with water or 175 mmol L −1 sucrose in darkness at 28 - C. The qRT-PCR was conducted to determine the transcript levels of IbpPGM in the cuttings harvested at different time points (0, 2, 4, 6, 12, 24 and 48 h) after treatment. Three cuttings were used as biological replicates per time point for each treatment.
## Prokaryotic expression of ibppgm
To determine whether IbpPGM encodes a mature protein, the gene was expressed in E. coli. The full-length ORF and the signal peptide-cleaved ORF (∆ORF) of IbpPGM were amplified with primers pET-F/R and pET-∆F/R, respectively. The sequence-verified fragments were ligated independently into the expression vector pET-28a. Then, the recombinant vectors pET-28a-IbpPGM and pET-28a-∆IbpPGM, as well as the pET-28a native vector, were introduced independently into the competent E. coli strain Transetta (DE3) cells (Transgen, Beijing, China). Fresh Luria-Bertani medium was inoculated independently with the positive clones, and they were cultured at 28 - C until the OD600 values reached 0.8. The soluble cytoplasmic proteins were prepared from the isopropyl β-D-thiogalactopyranosideinduced Transetta (DE3) cells. Then, the expressed IbpPGM protein was subjected to an SDS-PAGE analysis.
## Subcellular localization
The ORF of IbpPGM was amplified using primers 83-F/R and inserted into the pMDC83 vector between PacI and AscI cleavage sites. The recombinant vector pMDC83-IbpPGM and the native vector were introduced independently into the A. tumefaciens strain EHA105, and the positive strains were injected independently into the N. benthamiana leaf epidermal cells for the transient expression. Following the co-cultivation at 28 - C for approximately 36 h, the agroinfiltrated tobacco leaves were visualized using a laser scanning confocal microscope.
## Production of the transgenic plants
To construct the overexpression vector, the sequence-verified ORF of IbpPGM, which had been amplified using primers OPGM-F/R, was inserted into pBI121 between BamHI and SacI to replace the glucuronidase (gusA) gene. Subsequently, the expression cassette 35S-IbpPGM-NOS was excised from the pBI121-IbpPGM vector using HindIII and PstI and then ligated between the same cleavage sites in pCAMBIA3301, to generate the recombinant overexpression vector pC3301-121-IbpPGM. This plasmid was transfected into the A. tumefaciens strain EHA105. The plant transformation and regeneration were performed, as described by Wang et al. [bib_ref] A plastidic ATP/ADP transporter gene, IbAATP, increases starch and amylose contents and..., Wang [/bib_ref] using embryogenic suspension cultures of Lizixiang established, as described by Liu et al. [bib_ref] Efficient plant regeneration from embryogenic suspension cultures of sweetpotato, Liu [/bib_ref]. The putatively transgenic sweet potato plants were identified by histochemical GUS assays and PCR. Then, the positive transgenic lines were subjected to the qRT-PCR using primers qPGM-F/R, and the three lines with the highest IbpPGM expression levels were selected for further phenotypic analyses.
## Quantification of the carbohydrate contents
The starch contents in the storage roots of the transgenic and WT plants were analyzed in accordance with the method of Smith and Zeeman [bib_ref] Quantification of starch in plant tissues, Smith [/bib_ref] , and high performance liquid chromatography (HPLC) was used to determine the sucrose, glucose, and fructose contents in the storage roots with the following method. First, 30 mg of 80 - C dried storage roots was dissolved in 0.7 mL of 80% ethanol to extract the sugars. Then, the sample was thoroughly vortexed and incubated for 2 h at 70 - C. The aliquots of 0.7 mL of HPLC-grade water and 0.7 mL chloroform were added to the sample. Then, after shaking several times, the mixtures were centrifuged at 12,000 g for 10 min. Then, 0.7 mL of the aqueous supernatant was transferred into 1.5-mL Eppendorf tubes and resuspended in 0.7 mL of chloroform. Following the centrifugation at 12,000 g for 10 min, 0.5 mL of the supernatant was transferred to a glass tube for the HPLC analysis of each sugar component. The Agilent technologies HPLC column (ZORBAX Carbohydrate column; 4.6 × 150 mm, 5 µm) with a differential refraction detector was used. The mobile phase consisted of 75% acetonitrile with a flow rate of 0.8 mL min −1 and the temperature of the column was maintained at 35 - C. The sugars were identified, based on the retention time of the standards, and the sample concentrations were calculated from the external standard curve. The quantifications were carried out with three replicates for each plant line.
## Expression analysis of the starch biosynthetic genes
The transcript levels of 12 key genes in the starch biosynthetic pathways of the transgenic and WT storage roots were investigated using qRT-PCR. The 12 Ipomoea batatas genes were AGP-sTL1 and 2 (encoding the two small subunits of IbAGPase), AGP-TLI (encoding the large subunit of IbAGPase), granule-bound starch synthase I (GBSSI), soluble starch synthase I (SSI), SSII, SSIII, SSIV, starch branching enzymes I and II (SBEI and SBEII), isoamylase1 (ISA1), and pullulanase (PUL). The primers used to amplify these genes are listed in [fig_ref] Table 1: Primers used in this study [/fig_ref].
# Results
## Cloning and sequence analysis of ibppgm
The RACE method was used to clone IbpPGM from the Xu 781 sweet potato line. The cloned 2182-bp full-length IbpPGM cDNA contained a 1917-bp ORF that generated a 638-aa protein with a molecular weight of 69.3 kDa. The genomic sequence of IbpPGM was 5583 bp and contained 22 exons and 21 introns. TargertP 2.0 and ChloroP 1.1 predicted that the N-terminus of the IbpPGM protein contains a 73-aa chloroplast transit peptide. The molecular weight of IbpPGM without the transit peptide was 61.7 kDa. A multiplesequence alignment of the PGM protein showed that IbpPGM shared a conserved domain similar with those from several other plants and from Saccharomyces cerevisiae [fig_ref] Figure 1: Multiple sequence alignment of PGM proteins [/fig_ref]. Additionally, pPGM had a chloroplast transit peptide that was not found in cPGM. A GSDS analysis revealed that pPGM and cPGM evolved into two different branches, with members of each branch having identical exon-intron structures, and IbpPGM was most closely related to the pPGM from Solanum lycopersicum (80.82% homology) [fig_ref] Figure 2: Phylogenetic and gene structure analyses of IbpPGM and PGMs from other species... [/fig_ref].
## Expression of ibppgm in the sweet potato
The qRT-PCR revealed that IbpPGM was expressed in all five tested tissues of Xu 781 plants, with the highest expression occurring in the storage root, followed, successively, by the leaves, the fibrous roots, the stems, and the petioles [fig_ref] Figure 3: Expression patterns of IbpPGM [/fig_ref]. The 175-mM sucrose treatments of the leaf-petiole cuttings in darkness significantly induced the IbpPGM expression, which strongly increased at 12 h after treatment and peaked after 48 h, reaching approximately 28 times the value at 0 h [fig_ref] Figure 3: Expression patterns of IbpPGM [/fig_ref].
## Expression of ibppgm in the sweet potato
The qRT-PCR revealed that IbpPGM was expressed in all five tested tissues of Xu 781 plants, with the highest expression occurring in the storage root, followed, successively, by the leaves, the fibrous roots, the stems, and the petioles [fig_ref] Figure 3: Expression patterns of IbpPGM [/fig_ref]. The 175-mM sucrose treatments of the leaf-petiole cuttings in darkness significantly induced the IbpPGM expression, which strongly increased at 12 h after treatment and peaked after 48 h, reaching approximately 28 times the value at 0 h [fig_ref] Figure 3: Expression patterns of IbpPGM [/fig_ref].
## Expression of ibppgm in e. coli
To investigate whether IbpPGM encodes the mature protein, we constructed a recombinant vector harboring the full-length ORF of IbpPGM, as well as one carrying the ORF with the transit peptide removed (ΔIbpPGM), to eliminate the effect of the chloroplast transit peptide on the prokaryotic expression. These two vectors were then expressed independently in E. coli [fig_ref] Figure 4: Prokaryotic expression of IbpPGM in E [/fig_ref]. The full-length ORF of IbpPGM failed to express the corresponding 69.3-kDa protein, whereas ΔIbpPGM expressed a protein of approximately 61.7 kDa. This suggested that the chloroplast transit peptide inhibits the expression of IbpPGM in E. coli.
## Expression of ibppgm in e. coli
To investigate whether IbpPGM encodes the mature protein, we constructed a recombinant vector harboring the full-length ORF of IbpPGM, as well as one carrying the ORF with the transit peptide removed (∆IbpPGM), to eliminate the effect of the chloroplast transit peptide on the prokaryotic expression. These two vectors were then expressed independently in E. coli [fig_ref] Figure 4: Prokaryotic expression of IbpPGM in E [/fig_ref]. The full-length ORF of IbpPGM failed to express the corresponding 69.3-kDa protein, whereas ∆IbpPGM expressed a protein of approximately 61.7 kDa. This suggested that the chloroplast transit peptide inhibits the expression of IbpPGM in E. coli. binant vector harboring the full-length ORF of IbpPGM, as well as one carrying the ORF with the transit peptide removed (ΔIbpPGM), to eliminate the effect of the chloroplast transit peptide on the prokaryotic expression. These two vectors were then expressed independently in E. coli [fig_ref] Figure 4: Prokaryotic expression of IbpPGM in E [/fig_ref]. The full-length ORF of IbpPGM failed to express the corresponding 69.3-kDa protein, whereas ΔIbpPGM expressed a protein of approximately 61.7 kDa. This suggested that the chloroplast transit peptide inhibits the expression of IbpPGM in E. coli.
## Subcellular localization of ibppgm in n. benthamiana
Both TargertP 2.0 and ChloroP 1.1 predicted that IbpPGM contained a chloroplast transit peptide. The expression vector pMDC83-IbpPGM was, therefore, transiently expressed in the N. benthamiana epidermal cells and visualized using a laser scanning confocal microscope [fig_ref] Figure 5: Subcellular localization of the IbpPGM protein in the epidermal cells of N [/fig_ref]. In the tobacco epidermal cells, IbpPGM was observed in scattered patches and co-localized with the autofluorescence of the chloroplasts. This result concurred with the online predictions that IbpPGM localizes to the chloroplast.
## Subcellular localization of ibppgm in n. benthamiana
Both TargertP 2.0 and ChloroP 1.1 predicted that IbpPGM contained a chloroplast transit peptide. The expression vector pMDC83-IbpPGM was, therefore, transiently expressed in the N. benthamiana epidermal cells and visualized using a laser scanning confocal microscope [fig_ref] Figure 5: Subcellular localization of the IbpPGM protein in the epidermal cells of N [/fig_ref]. In the tobacco epidermal cells, IbpPGM was observed in scattered patches and co-localized with the autofluorescence of the chloroplasts. This result concurred with the online predictions that IbpPGM localizes to the chloroplast.
## Overexpression of ibppgm in the sweet potato
To functionally characterize IbpPGM, the recombinant vector pC3301-121-IbpPGM was introduced into the sweet potato cultivar Lizixiang. A total of 97 putative transgenic lines were obtained, and 10 lines were shown to be transgenic using GUS assays and PCR verification. The qRT-PCR indicated that the IbpPGM expression levels in these 10 transgenic lines were 1.3-15.3 times of the level in WT, with lines OX17, OX53, and OX85 having the three highest expression levels [fig_ref] Figure 6: Relative expression levels of IbpPGM in the IbpPGM-overexpressing [/fig_ref]. Then, the three OX lines and the wildtype plants were further propagated in vitro and we observed that the leaf area of the invitro plantlets of OX was much larger than that of WT and the growth rate of OX was also faster. However, when they were grown in the field, no differences were identified between WT and OX in leaf size and shape, vine growth vigor, or morphology of the storage roots [fig_ref] Figure 1: Multiple sequence alignment of PGM proteins [/fig_ref]. The overexpression of IbpPGM may have an influence on improving the photosynthesis of OX and under the lab tissue culture condition, this advantage of OX was magnified by the limited culture conditions, in terms of
## Overexpression of ibppgm in the sweet potato
To functionally characterize IbpPGM, the recombinant vector pC3301-121-IbpPGM was introduced into the sweet potato cultivar Lizixiang. A total of 97 putative transgenic lines were obtained, and 10 lines were shown to be transgenic using GUS assays and PCR verification. The qRT-PCR indicated that the IbpPGM expression levels in these 10 transgenic lines were 1.3-15.3 times of the level in WT, with lines OX17, OX53, and OX85 having the three highest expression levels [fig_ref] Figure 6: Relative expression levels of IbpPGM in the IbpPGM-overexpressing [/fig_ref]. Then, the three OX lines and the wild-type plants were further propagated in vitro and we observed that the leaf area of the in-vitro plantlets of OX was much larger than that of WT and the growth rate of OX was also faster. However, when they were grown in the field, no differences were identified between WT and OX in leaf size and shape, vine growth vigor, or morphology of the storage roots [fig_ref] Figure 1: Multiple sequence alignment of PGM proteins [/fig_ref]. The overexpression of IbpPGM may have an influence on improving the photosynthesis of OX and under the lab tissue culture condition, this advantage of OX was magnified by the limited culture conditions, in terms of light, temperature, humidity, and CO 2 concentration. When transplanted in the field, the photosynthetic ability and growth vigor of all plant lines were fully stimulated by the sufficient environmental factors and the phenotypic differences disappeared. lines were obtained, and 10 lines were shown to be transgenic using GUS assays and PCR verification. The qRT-PCR indicated that the IbpPGM expression levels in these 10 transgenic lines were 1.3-15.3 times of the level in WT, with lines OX17, OX53, and OX85 having the three highest expression levels [fig_ref] Figure 6: Relative expression levels of IbpPGM in the IbpPGM-overexpressing [/fig_ref]. Then, the three OX lines and the wildtype plants were further propagated in vitro and we observed that the leaf area of the invitro plantlets of OX was much larger than that of WT and the growth rate of OX was also faster. However, when they were grown in the field, no differences were identified between WT and OX in leaf size and shape, vine growth vigor, or morphology of the storage roots [fig_ref] Figure 1: Multiple sequence alignment of PGM proteins [/fig_ref]. The overexpression of IbpPGM may have an influence on improving the photosynthesis of OX and under the lab tissue culture condition, this advantage of OX was magnified by the limited culture conditions, in terms of light, temperature, humidity, and CO2 concentration. When transplanted in the field, the photosynthetic ability and growth vigor of all plant lines were fully stimulated by the sufficient environmental factors and the phenotypic differences disappeared.
## Starch and sugar contents in the transgenic sweet potato
The quantified levels of starch and sugar contents in the transgenic and WT lines are shown in [fig_ref] Table 2: Starch and soluble sugar contents in the storage roots of the plastidial... [/fig_ref]. The overexpression of IbpPGM in the sweet potato significantly increased the starch contents in the storage roots. Moreover, the sucrose contents in the transgenic storage roots decreased significantly, whereas the glucose and fructose contents increased.
## Expression profiles of the starch biosynthetic genes
The expression levels of the starch biosynthesis related genes in the transgenic lines were evaluated by the qRT-PCR [fig_ref] Figure 7: Relative expression levels of the starch biosynthesis related genes in the IbpPGMoverexpressing [/fig_ref]. All 12 detected genes were located downstream of IbpPGM in the starch biosynthesis pathway and showed increased expression levels, to different extents, in the transgenic lines, compared with WT. Among the 12 genes, IbAGP-sTL1, IbAGP-sTL2, and IbAGP-TLI were responsible for the ADPG synthesis. IbGBSSI was involved in the amylose elongation, and the other eight genes were mainly involved in the amylopectin synthesis. Thus, the overexpression of IbpPGM promoted the accumulation of the starch biosynthesis precursors and resulted in the up-regulation of related downstream genes. stream of IbpPGM in the starch biosynthesis pathway and showed increased expression levels, to different extents, in the transgenic lines, compared with WT. Among the 12 genes, IbAGP-sTL1, IbAGP-sTL2, and IbAGP-TLI were responsible for the ADPG synthesis. IbGBSSI was involved in the amylose elongation, and the other eight genes were mainly involved in the amylopectin synthesis. Thus, the overexpression of IbpPGM promoted the accumulation of the starch biosynthesis precursors and resulted in the up-regulation of related downstream genes.
# Discussion
The sweet potato is an important food crop with starchy storage roots that are raw materials for food, feed, and industrial uses. Traditional hybrid breeding results in a slow crop improvement rate and is not trait specific. Consequently, genetic engineering has become an effective way to increase the starch contents in the sweet potato storage roots, which is a primary goal of sweet potato breeding. The main functional genes in the sweet potato starch synthesis have been characterized. The overexpression of IbAATP, which is responsible for transporting ATP in the cytoplasm to plastids as an energy supply for starch synthesis, and IbSSI, which is involved in elongating the short chains of amylopectin, significantly increases the starch contents in the sweet potato storage roots [bib_ref] A plastidic ATP/ADP transporter gene, IbAATP, increases starch and amylose contents and..., Wang [/bib_ref]. Additionally, the overexpression of IbSnRK1 in the sweet potato improves the starch content, as well as the starch quality, in the storage roots [bib_ref] A sucrose non-fermenting-1-related protein kinase-1 gene, IbSnRK1, improves starch content, composition, granule..., Ren [/bib_ref].
In the plant starch synthetic pathway, PGM is responsible for the synthesis of the upstream precursor G1P. Functional studies of PGM have mainly focused on how the starch contents in the plants changed when this gene is mutated or subject to RNA interference [bib_ref] The rug3 locus of pea encodes plastidial phosphoglucomutase, Harrison [/bib_ref] [bib_ref] Alterations in growth, photosynthesis, and respiration in a starchless mutant of Arabidopsis..., Caspar [/bib_ref] [bib_ref] Molecular characterization of a new mutant allele of the plastid phosphoglucomutase in..., Kofler [/bib_ref] [bib_ref] Antisense inhibition of plastidial phosphoglucomutase provides compelling evidence that potato tuber amyloplasts..., Tauberger [/bib_ref]. However, few studies have investigated how the starch contents in plants, including the sweet potato, are altered when PGM is overexpressed. In this study, the coding and genomic sequences of IbpPGM were isolated from the Xu 781 sweet potato line using homologous cloning. A GSDS analysis showed that IbpPGM shared an identical gene structure (22 exons and 21 introns) with its counterparts from other species, such as Brassica napus and A. thaliana, whereas cPGMs shared a similar gene structure of 18 exons and 17 introns. This demonstrated that the functional differentiation of the PGM genes into plastidial and cytosolic forms was accompanied by gene structural changes. Additionally, the deletion of the chloroplast transit peptides from the cPGM proteins, compared with the pPGM proteins, also indicated that the gene function is closely related to the gene structure.
The expression level of IbpPGM was high in storage roots and leaves, which was consistent with its characterized role in starch synthesis. IbpPGM exists in amyloplasts in storage roots, whereas in leaves, it is mainly found in chloroplasts. Sucrose is an important signaling molecule in starch synthesis, and it induces the expression of starch synthesis-related genes through the abscisic acid pathway. When the sucrose contents in leaves (or in-vitro sucrose feeding) exceeds the need for respiration, starch synthesis is induced [bib_ref] Cloning and expression analysis of a potato cDNA that encodes branching enzyme:..., Koßmann [/bib_ref] [bib_ref] Impaired sucrose-induction mutants reveal the modulation of sugar-induced starch biosynthetic gene expression..., Rook [/bib_ref]. In sweet potato, the expression levels of IbGBSSI, IbAGP-sTL1, and IbSSI are all significantly induced by exogenous sucrose treatments [bib_ref] Molecular cloning and characterization of two novel isoforms of the small subunit..., Bae [/bib_ref] [bib_ref] Regulation of starch granule-bound starch synthase I gene expression by circadian clock..., Wang [/bib_ref]. In this study, the leaf-petiole cuttings were soaked in water in darkness as the starvation pretreatment to consume the endogenous carbohydrates and minimize respiration. The 0 h−6 h sucrose treatment was the recovery period for respiration, during which the IbpPGM expression was induced at a low level. When the continuous sucrose feeding exceeded the respiratory demand, the IbpPGM expression began to increase rapidly, reaching 28 times the 0 h level by the end of the treatment. However, the continuous feeding of water after the starvation pre-treatment did not induce the IbpPGM expression. In agronomic practices, spraying sucrose-based polymers on leaves may improve the yield and quality of some fruits, which may result from induced alterations in the carbohydrate metabolism in leaves and then fruits. Thus, determining whether a leaf spray of sugar-based growth regulators to increase the yield and quality of the sweet potato, would be of interest in the future.
To determine whether IbpPGM encodes a mature protein, two recombinant vectors for the prokaryotic expression were constructed. One carried the complete coding sequence of IbpPGM, and the other carried the same sequence minus the chloroplast transit peptide (∆IbpPGM). The former could not be expressed in E. coli, whereas the latter expressed a mature protein. Thus, the transit peptide of IbpPGM may form a complex secondary structure at the mRNA level that affects the initial translation in a prokaryotic system. The fluorescence of the GFP protein fused with the complete coding sequence of IbpPGM mainly targeted the chloroplasts, which indicated that the transit peptide was successfully translated in the eukaryotic system and directed the IbpPGM protein to the chloroplasts.
To characterize the function of IbpPGM, it was overexpressed in the sweet potato lowstarch cultivar Lizixiang, by infecting the embryogenic suspension cells, and 10 positive transgenic lines were obtained. The three lines having the highest IbpPGM expression levels exhibited significant 4.5%, 5.5%, and 12.0% increases in the storage-root starch content, compared with the WT line. Owing to the overexpression of IbpPGM, the precursor G1P accumulated, which up-regulated the downstream genes involved in amylose and amylopectin synthesis, leading to further increases in the starch contents [fig_ref] Figure 7: Relative expression levels of the starch biosynthesis related genes in the IbpPGMoverexpressing [/fig_ref]. In heterotrophic organs, the catalytic substrate of pPGM for G1P is G6P, which is mainly derived from the degradation of sucrose and is transported into the amyloplasts from the cytoplasm [bib_ref] Molecular characterization of a carbon transporter in plastids from heterotrophic tissues: The..., Kammerer [/bib_ref]. In this study, the overexpression of IbpPGM consumed more G6P for the G1P synthesis, which accelerated the degradation of sucrose in the cytoplasm, leading to further increases in the glucose and fructose contents.
# Conclusions
In the present study, the IbpPGM gene was cloned from the Xu 781 sweet potato line. IbpPGM was mainly expressed in the storage roots and leaves. Its expression was strongly induced by the exogenous sucrose treatments. The IbpPGM protein was subcellularly localized to the chloroplasts and was successfully expressed in E. coli when its 73-aa chloroplastic transit peptide was excised. The overexpression of IbpPGM significantly increased the starch contents of the transgenic sweet potato storage root and altered its soluble sugar levels. Meanwhile, the expression levels of starch biosynthetic genes in transgenic sweet potato storage roots showed increased expression levels to different extents. These results indicated that IbpPGM has the great potential as an important candidate gene for increasing the starch content of the sweet potato through genetic engineering.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/genes13122234/s1, [fig_ref] Figure 1: Multiple sequence alignment of PGM proteins [/fig_ref] :
[fig] Figure 1: Multiple sequence alignment of PGM proteins. AtpPGM: Arabidopsis thaliana, AF216580; StpPGM: Solanum tuberosum, NM_001288352; PspPGM: Pisum sativum, AJ250770; AtcPGM: A. thaliana, At1g23190; PscPGM: P. sativum, AJ250769; ZmcPGM: Zea mays, U89341; ScPGM: Saccharomyces cerevisiae, NP_012795. Homology in amino acids is shaded with dark blue (homology = 100%), lightblue (homology ≥ 75%) or grey (homology ≥ 50%). [/fig]
[fig] Figure 2: Phylogenetic and gene structure analyses of IbpPGM and PGMs from other species using GSDS. Brassica rapa, XP_009107139; A. thaliana, AF216580; Gossypium hirsutum, XM_016881887; Manihot esculenta, XP_021616975; Medicago truncatula, XP_003592358; P. sativum, AJ250770; Coffea arabica, XP_027103885; Solanum lycopersicum, XP_004234144; S. tuberosum, NM_001288352; A. thaliana, At1g23190; P. sativum, AJ250769; Sesamum indicum, XP_011070593; Oryza sativa, XP_015632142; Sorghum bicolor, XP_002466576; Z. mays, U89341; S. cerevisiae, NP_012795. [/fig]
[fig] Genes 2022 ,: 13, x FOR PEER REVIEW 7 of 14 [/fig]
[fig] Figure 3: Expression patterns of IbpPGM. (a) Expression of IbpPGM in different tissues of Xu 781 sweet potato. L, leaves, S, stems, P, petioles, SR, storage roots, FR, fibrous roots. (b) Induced expression of IbpPGM by a 175-mM sucrose treatment. * and ** indicate a significant difference versus 0 h at p < 0.05 and < 0.01, respectively, based on Student's t-test. [/fig]
[fig] Figure 4: Prokaryotic expression of IbpPGM in E. coli. The arrow indicates the induced protein of IbpPGM without the chloroplast transit peptide. [/fig]
[fig] Figure 5: Subcellular localization of the IbpPGM protein in the epidermal cells of N. benthamiana leaves. Bar = 50 μm. [/fig]
[fig] Figure 6: Relative expression levels of IbpPGM in the IbpPGM-overexpressing (OX) transgenic and wild-type (WT) sweet potato lines. ** indicates a significant difference versus WT at p < 0.01, based on Student's t-test. [/fig]
[fig] Figure 7: Relative expression levels of the starch biosynthesis related genes in the IbpPGMoverexpressing (OX) transgenic and wild-type (WT) sweet potato lines. IbAGP-sTL1 and -sTL2, encoding the two small subunits of ADPglucose pyrophosphorylase (GenBank accession number: Z79635 and Z79636); IbAGP-TLI, encoding the large subunit of ADPglucose pyrophosphorylase (AJ252316); IbGBSSI, granule-bound starch synthase I (AB071604); IbSSI, II, III and IV: soluble starch synthase I, II (AF068834), III and IV; IbSBEI and II, starch branching enzyme I and II (AB194725 and AB071286); IbISA1, isoamylase1 (DQ074643); IbPUL, pullulanase. * and ** indicate a significant difference versus WT at p < 0.05 and <0.01, respectively, based on Student's t-test. [/fig]
[fig] Funding: The research was funded by the National Key R&D Program of China (2019YFD1001302/ 2019YFD1001300); Natural Science Foundation of Henan Province (212300410170); Science and Technology Program of Henan Province (212102110251); Earmarked Fund for China Agriculture Research System (CARS-10-C14). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. [/fig]
[table] Table 1: Primers used in this study. [/table]
[table] Table 2: Starch and soluble sugar contents in the storage roots of the plastidial phosphoglucomutaseoverexpressing (OX) transgenic and wild-type (WT) sweet potato lines. [/table]
[bib_ref] A starchless mutant of Nicotiana sylvestris containing a modified plastid phosphoglucomutase, Hanson [/bib_ref] |
Topographic connectivity reveals task-dependent retinotopic processing throughout the human brain
The human visual system is organized as a hierarchy of maps that share the topography of the retina. Known retinotopic maps have been identified using simple visual stimuli under strict fixation, conditions different from everyday vision which is active, dynamic, and complex. This means that it remains unknown how much of the brain is truly visually organized. Here I demonstrate widespread stable visual organization beyond the traditional visual system, in default-mode network and hippocampus. Detailed topographic connectivity with primary visual cortex during movie-watching, resting-state, and retinotopic-mapping experiments revealed that visual-spatial representations throughout the brain are warped by cognitive state. Specifically, traditionally visual regions alternate with default-mode network and hippocampus in preferentially representing the center of the visual field. This visual role of default-mode network and hippocampus would allow these regions to interface between abstract memories and concrete sensory impressions. Together, these results indicate that visual-spatial organization is a fundamental coding principle that structures the communication between distant brain regions.retinotopy | hippocampus | connective field | population receptive field | naturalistic vision Significance Vision is organized retinotopically-according to the reference frame of the retina. How much of the brain is retinotopically organized remains unknown, because traditional retinotopic mapping experiments require strict fixation and sparse stimuli. Conversely, in everyday vision we use eye movements and interaction, to derive meaning from our complex surroundings. Here, I discover retinotopic activations by explaining brain-wide BOLD signals during several experiments in terms of the pattern on the surface of primary visual cortex. This revealed visually organized processing also in regions outside the visual system, in brain regions traditionally thought devoted to memory. This visual organization in default-mode network and hippocampus speaks to the joint operation of sensations and memory in everyday vision and mental life.
O ur experience of the world is ultimately based on impressions arriving through the senses. In our dominant sensory modality, vision, processing is retinotopic: organized according to the layout of the retina. That is, neighboring locations in the brain represent neighboring locations in the visual field. Retinotopic mapping experiments leverage sparse visual stimulation during fixation, allowing researchers to relate the elicited brain responses to visual space and delineate retinotopic maps in the brain. Yet, in everyday life visual inputs are not sparse, and naturalistic vision is characterized by continuous eye movements and dynamic cognitive demands. It is therefore likely that charting especially high-level visual function falls outside the scope of traditional retinotopic mapping experiments.
Retinotopic processing throughout the brain can be identified based on topographically specific connectivity with V1 (4-6), the first visual region of the cerebral cortex. There are several distinct advantages to assessing visual processing by means of retinotopic connectivity (RC), in which responses throughout the brain are explained in terms of the spatial pattern of activation on the surface of V1. First, RC is robust in the face of eye movements, because its reference frame is fixed in the brain and not the outside world. Second, because V1 harbors a map of visual space, RC patterns throughout the brain can be translated back into visual space coordinates. In effect, RC allows us to project the retinotopy of V1 into the rest of the brain. Finally, since brain responses are explained as a function of ongoing activations, RC can be estimated for any experimental paradigm. Thus, RC can be used to compare detailed visual-spatial processing across experiments and cognitive states.
I performed RC analysis on the Human Connectome Project (HCP) 7T dataset of 174 subjects in which data were col-lected during retinotopic-mapping, resting-state, and moviewatching experiments. This allowed the identification of previously unknown visual-spatial processing throughout the brain, and the quantification of how visual space is represented-even in brain regions not traditionally considered visual. Moreover, these analyses reveal how visual representations depend on cognitive state.
# Results
A parsimonious computational model for RC posits that responses arise from a localized Gaussian patch on the surface of V1, its connective field (CF). One fits the CF model by comparing model time-course predictions to ongoing blood oxygenation level dependent (BOLD) response time courses throughout the brain. To ensure that the model captures only spatially specific topographic connectivity, I correct cross-validated (CV) model prediction performance for correlation with a nontopographic null model. This null model prediction, V1's average time course, corrects for responses driven by arousal, overall contrast, or feature energy. Translating the best-fitting CF parameters into visual field locations reveals the structure of visual field maps in V2, V3, and beyond. Retinotopic maps resulting from retinotopic-mapping, movie-watching, and resting-state acquisitions are similar, with their borders in the same location. This means that in low-level visual cortex the structure and strength of this retinotopic connectivity are both stable across participants and robust against variations in experimental task, stimulation, and cognitive state. Visual processing in higher-order brain regions reveals itself through spatially specific retinotopic connectivity with V1. V1's map of visual space allows us to translate retinotopic connectivity to representations of visual space. (B) Retinotopic connectivity is quantified by modeling responses throughout the brain as emanating from Gaussian connective fields on the surface of V1. Example Gaussian CF model profiles with different size (σ) and location (v 0 ) parameters are shown on the inflated surface. (C) Predictions are generated by weighting the ongoing BOLD signals in V1 with these CF kernels. Estimating CF parameters for all locations throughout the brain captures significant variance also outside the nominal visual system (non-V1 time courses). (D) Cross-validation procedure. CV prediction performance of the CF model is assessed on a left-out, test dataset and is corrected for the performance of a nontopographic null model, the average V1 time course. (E-G) CF modeling results can be used to reconstruct the retinotopic structure of visual cortex. Polar angle preferences outside the black outline are reconstructed solely based on their topographic connectivity with V1, within the black outline. CF modeling was performed on data from retinotopic-mapping, resting-state, and movie-watching experiments separately. Visual field preferences are stable across cognitive states, as evidenced by the robust locations of polar angle reversals at the borders between V2 and V3 field maps. Additional retinotopic structure visualizations are in SI Appendix, Figs. S1 and S2.
Does this RC extend beyond the lower levels of the visual system, and, if so, how does it depend on the different cognitive states evoked in different experiments? Indeed,shows that significant portions of movie-watching BOLD fluctuations throughout the cerebral cortex are explained as resulting from RC. That is, more than half of the cerebral cortex, including large swaths of the temporal and frontal lobes, shows significant topographically specific connectivity with V1. Interestingly, the local strength of RC depends heavily on the experiment. During movie watching, mainly ventral visual and temporal brain regions exhibit RC. This retinotopic connectivity likely reflects object identity-related processing and audiovisual integration. Conversely, during resting-state scans the default-mode network (DMN) shows stronger RC, which may reflect endogenous mental imagery during mind wandering.
If retinotopic visual processing is a stable organizational property, visual field preferences derived from one experiment should predict RC from another. The detailed inspection of CF parameters can provide insights into the visual-spatial processing embodied by RC, but also allows us to understand its modulation by cognition. We can compare two regions, lateral-occipital (LO) and angular gyrus (ANG), as exemplars of high-level visual and DMN areas, respectively. In both regions, spatial sampling extent, as quantified by CF size, is stable (9) and precise (5, 10) during both retinotopy and resting state and becomes larger and more variable during movie watching. Sampling extent is strongly correlated between conditions (all linear correlations ρ > 0.36, all P < 10 −13 (LO), ρ > 0.75, all P < 10 −37 (ANG)-full statistics in SI Appendix, . This points to stable sampling of retinotopic space in both high-level visual and DMN regions.
In LO, the preferred eccentricity of cortical locationsis strongly correlated between experimental conditions (linear correlations ρ > 0.48, all P < 10 −19 ), confirming its well-known retinotopy (9, 11). The detailed differences in spatial representations between conditions, however, show the flexibility of LO's retinotopy: A very foveal bias during retinotopic mappinggives way to broader coverage of the visual field during resting state and movie watching. The across-condition robustness of ANG eccentricityis greater than that of LO (all linear correlations ρ > 0.75, all P < 10 −40 ). However, ANG shows an opposite pattern of retinotopic flexibility, with visual field preferences being more foveal specifically during the resting state. These foveopetal and foveofugal changes of eccentricity in both LO and ANG mimic the effects of topdown attentionand imagination (14) on visual field representations. This pattern is also similar to what happens to visual items of interest as they are foveated through eye movementsand follows the topographic distribution of feedback related to visual object informationin the absence of eye movements.
V1 represents contralateral visual field locations, allowing us to use the hemisphere of the best-fitting CF as a proxy for visual field representation laterality. In LO, visual representations are strongly contralateral and strongly correlated between experimental conditions (t test between hemispheres: all T (> 123) > 10, P < 10 −19 ). In the ANG region of the DMN, there is a significant contralateral bias of visual field representations during retinotopic mapping and movie watching (all T (199) > 6, P < 10 −9 ), but, presumably due to the strong foveal bias of visual field representations, not during resting state (T (199) = 0.14, P = 0.9). These findings confirm previous reports that the DMN represents visual space similarly to high-level visual brain regions in situations of visual stimulation.
Although it is not generally implicated in traditional vision science experiments, tracer-based connectivity studies place the hippocampal formation at the top of the visual-processing hierarchy. Hippocampus is thought to implement the interaction between memory-related and sensory processing or imagery, leading me to reason that both narrative understanding of naturalistic inputs and internally generated thought should evoke strong RC in hippocampus, over and above its recently discovered contralateral visual field preference during retinotopic mapping (23) (SI Appendix, .
Applying CF modeling to hippocampus BOLD time coursesreveals significant hippocampal-V1 RC in all experimental conditions. The most striking feature in the hippocampus is a gradient of movie-watching vs. resting-state RC preference along both the medial/lateral and the long, anterior/posterior axes of the hippocampus. This RC gradient corresponds in detail to previously found microstructural and functional connectivity gradients. The high power, quality, and spatial resolution of both functional and anatomical HCP images allow the quantification of cognitive state-dependent RC per hippocampal subfield. A strong prediction based on earlier findings in both miceand humansis that the Cornu Ammonis (CA) region and uncus of the hippocampus should subserve its connectivity with visual cortex during visual stimulation.
The strength and cognitive-state dependence of RC varies strongly across hippocampal subfields. Specifically, RC driven by movie watching is strongest in parasubiculum, CA1, CA3, dentate gyrus (DG), and the hippocampal-amygdalar transition area (HATA) which principally represents the hippocampal uncus. Conversely, resting-state-driven RC is strongest in the presubiculum and the hippocampal tail. Focusing on the visual space representations of hippocampus, as quantified by CF model parameters, reveals strong correlations between experimental conditions. Specifically, the detailed patterns of differences in visual field representations between experimental conditions closely resemble those that occur in the DMN.
# Discussion
CF modeling reveals retinotopic connectivity throughout the human brain, which can be quantified in terms of visual space representations. Moreover, as CF modeling can be applied to any ongoing timeseries data, it can be used to reveal changes in visual space representations across experiments.
Connective field models are driven by topographically specific BOLD responses, but do not distinguish between the many possible bottom-up and top-down sources of those responses. This is the power of the approach, but also causes interpretative challenges. Could these patterns of retinotopic connectivity be due to the complex spatiotemporal correlations that characterize the naturalistic visual inputs of, and engagement with, the moviewatching experiment? Indeed, based on just the movie-watching results, we cannot definitively assert that topographic connectivity is evidence for visual-spatial processing. But in the resting state, topographic connectivity arises from the brain's internal organization. Finding similar RC patterns in resting state and movie watching thus hints strongly at a consistent visual-spatial structure that the brain has internalized and that is used in ongoing thought. Moreover, many of these spatiotemporal correlations in visual input and behavior that characterize movie watching are explicitly avoided by the retinotopic-mapping stimulus material and paradigm (for example, the fact that usually, human heads occur on top of bodies). I construe the similar patterns of RC between movie-watching and retinotopy experiments as evidence for a joint visual frame of reference used throughout the brain. This also suggests that the degree of similarity in RC patterns between all three experiments can be seen as a gauge for how "visual" a brain region is. The gradient of RC similarity across experiments in both cortex and hippocampusaligns surprisingly well with a combination of functional connec-tivity gradients found previously in the resting state. Specifically, resting-state-movie-watching preferences behave as a combination of the primary sensory-transmodal and the multiple-demand gradients (24)-both in cortex and within the hippocampus. These large-scale gradients result from the push-pull between sensory and attentional brain systems on the one hand and the DMN on the other. Such countervailing activations and deactivations are thought to mediate a dynamic balance between outward-oriented, stimulus-driven processing on the one hand and memory-related, endogenously generated processing on the other. Importantly, the present work identifies a consistent mode of organization across both memory-related and stimulus-directed processing: In both cognitive states, connectivity is retinotopically organized. Furthermore, these results demonstrate that the push-pull between DMN and sensory regions in terms of signal amplitude (28) also involves a trade-off in foveally biased processing between regions, possibly related to recent findings of retinotopic traveling waves in visual cortex during resting state. On the basis of these findings, I propose that the detailed structure of concurrent retinotopically organized activations in visual system, DMN, and hippocampus gives rise to interactions between attentional and mnemonic processing.
The present finding, that the hippocampus shares a retinotopic mode of organization with much of the rest of the brain, is in line with canonical tracer-based network findings. As hippocampus also entertains world-centric coding of space, this solidifies the notion (19) that hippocampus is a nexus for the conjunctive coding of both world-centric and sensory reference frames.
The stability of RC structure across experiments points to the brain's use of sensory topography as a fundamental organizing principle to facilitate neural communication between distant brain regions. By casting ongoing brain responses into a common reference frame, topographic connectivity can facilitate our understanding of neural processing as resulting from canonical computational mechanisms such as divisive normalization. Future work will be able to leverage topographic connectivity to investigate the computational hierarchy that generates increasingly world-centric representations of spaceand culminates in the medial temporal lobe.
# Materials and methods
Human Connectome Project Data. HCP 7T functional MRI (fMRI) data were used, in conjunction with 3T anatomical MRI images. In total, 2.5 h from 174 subjects with full data of all 7T experiments were used, sampled at 1.6 mm isotropic resolution and a rate of 1 Hz. For all functional analyses, the Fix independent component analysis-denoised time-course data, sampled to the 59,000 vertex-per-hemisphere areal feature-based cross-subject alignment method (MSMAll) surface and 1.6-mm Montreal Neurological Institute (MNI) volume formats, were used. These data are freely available from the HCP project website.
Analysis. Hippocampal subfield segmentation was performed using FreeSurfer, after which the individual subfields were warped to the functional data's MNI space using the existing HCP warp fields with nearest-neighbor interpolation. High-resolution subfield segmentations were smoothed with a Gaussian of 0.8 mm σ to ensure representation of all subfields when resampled to the 1.6-mm resolution of the functional images. Anatomical region of interest (ROI) definitions were taken from the multimodal parcellation atlasfor surface data and FMRIB Software Library's Jülich histological atlasfor hippocampal ROIs in MNI volumetric space.
Functional data of all three experiments [movie (approximately 1 h), retinotopy (43) (30 min), and resting state (1 h)] were preprocessed identically, by means of high-pass filtering (third-order Savizky-Golay filter, 210-s period) and z scoring over time.
To create a template of retinotopic spatial selectivity, I averaged the time courses of the retinotopic-mapping experiment (bar and wedge conditions) across participants and estimated linear Gaussian population receptive field (pRF) model parameters from
[formula] g(x 0 , y 0 , σ) = exp − (x − x 0 ) 2 + (y − y 0 ) 2 2σ 2 ,[1] [/formula]
where x 0 , y 0 , and σ are the parameters that define the location (in the Cartesian {x, y} plane) and size of the pRF, respectively. The fitting procedure consisted of an initial grid fit stage, followed by an iterative fitting stage using the L-BFGS-B algorithm as implemented in scipy.optimize. The identical fitting procedure was performed on the hippocampus to create figures of visual-spatial representations shown in SI Appendix, . Gaussian connective field profiles on the surface are defined for each vertex v on the cortical manifold as
[formula] CF(v 0 , v, σ) = exp − |v − v 0 | 2 2σ 2 , [2] [/formula]
where v 0 is the center location of the connective field, and σ is the Gaussian spread in millimeters on the cortical surface. Diffusion of heat along the fiducial, or midgray, surface mesh was simulated and then used to infer geodesic distances |v − v 0 | between all V1 vertices, as implemented in pycortex(which was also used for all surface-based visualization). As the two hemispheres are two separate surface meshes, this distance matrix was calculated for each hemisphere separately. Only V1 vertices with a pRF-fit within-set R 2 of >0.2, a peak pRF position inside the stimulus aperture used in the retinotopy experiment, and a positive pRF amplitude in the above pRF analysis served as the center of a candidate CF. These conservative selection criteria were chosen to ensure that CFs are centered on visually responsive vertices within V1 and improves the interpretability of the relation between CF parameters and visual field coordinates. Separate analyses (results not reported here) served to verify that using the full V1 map as a possible CF center yielded similar RC results on the whole. Grid-fit candidate CF sizes ranged from very small (biased to the center vertex only) to evenly spanning almost the entirety of V1: [0.5, 1, 2, 3, 4, 5, 7, 10, 15, 20, 30, 40, 80] mm σ for the Gaussian CF. The predicted time course for each of the resulting 14,287 CF models was generated by taking the dot product between the CF's vertex profile and the vertex by time-courses matrix in V1. In this operation I did not apply the conservative selection criteria used for CF center vertices; all vertex time courses in V1 were used regardless of eccentricity, pRFfit R 2 , etc. Additionally, I ensured that CF models were based on only V1 time courses. That is, if the CF extended into V2, these time courses were not used to generate the CF model time course. The resulting CF model time courses were z scored and correlated with the time courses throughout the brain, without convolution with a hemodynamic response function. For each location in the brain, the connective field resulting in the highest squared correlation was selected. Subsequently, this specific connective field was used to generate model time courses for out-of-set predictions of left-out data. A fourfold scheme was used for cross-validation. As both the resting state and movie watching consist of four separate runs of approximately 15 min each, this scheme was implemented to be identical to a leave-one-run-out cross-validation pattern. CF parameters and correlation measures were then averaged across runs for further analysis.
The resulting out-of-set correlation was referenced against the correlation with a nonspatial null model, in which the time courses throughout the brain were predicted by the average V1 time course. This correction means that although corrected correlation values are no longer interpretable as correlations, they can serve to conservatively assess the presence of true topographic connectivity. Moreover, comparisons between conditions based on these corrected correlation values explicitly discount changes in nontopographic correlations between brain regions. One-sample t tests were used to compare corrected out-of-set prediction performance against 0, and reported P values are two sided. Correlations across vertices between CF parameters, as well as t tests of the differences in CF parameters across experiments (SI Appendix, , were calculated weighted by the nullmodel-corrected correlation values using the statsmodels.stats.weightstats package, which adjusts the degrees of freedom based on the applied weighting. To quantify a location's RC preference during resting state vs. movie watching, a normalized ratio of the null-model-corrected CV correlation values was used for the respective experimental conditions ρ RS ρ RS +ρ MW , where RS and MW stand for resting state and movie watching, respectively. This measure takes a value between 0 and 1, where 0.5 signifies equal RC in both experimental conditions. Because the constituent parts of this ratio are corrected for null-model performance, this measure is not sensitive to possible nontopographic differences in V1 connectivity between conditions, such as arousal.
Pilot analyses revealed that the presented results are robust to drastic changes in the employed modeling procedure: Penalized regression using V1 time courses as the design matrix (45) produces highly comparable results. That is, CF fitting is also possible when releasing the Gaussian constraint that is based on the known topographic organization in the source region. CF modeling and RC estimation in general exemplify a broader category of analysis techniques in which decomposition of signals based on their local structure represents a very efficient manner of mapping all-toall correlations between brain activations into a subspace relevant to local neural processing. Explicit CF modeling following the methods from Haak et al.was chosen here because it performs direct estimation of meaningful CF parameters.
Data Availability. All in-house analyses were implemented in python, using scientific python packages. The full list of dependencies for running the analyses, and the code itself, are available on GitHub, https:// github.com/tknapen/HCP Connective Fields. The notebooks in this repository allow users to recreate all data visualizations presented here. Subcortical atlases for hippocampusand thalamusare available through their respective publications.
fMRI/MRI data have been deposited in the Human Connectome Project (https://wiki.humanconnectome.org).
## Knapen
Topographic connectivity reveals task-dependent retinotopic processing throughout the human brain |
Graphene-Based Electrospun Fibrous Materials with Enhanced EMI Shielding: Recent Developments and Future Perspectives
As a result of advancements in electronics/telecommunications, electromagnetic interference (EMI) pollution has gotten worse. Hence, fabrication/investigation of EMI shields having outstanding EMI shielding performance is necessary. Electrospinning (ES) has recently been established in several niches where 1D nanofibers (NFs) fabricated by ES can provide the shielding of EM waves, owing to their exceptional benefits. This review presents the basic correlations of ES technology and EMI shielding. Diverse graphene (GP)-based fibrous materials directly spun via ES as EMI shields are discussed. Electrospun EMI shields as composites through diverse post-treatments are reviewed, and then different factors influencing their EMI shielding characteristics are critically summarized. Finally, deductions and forthcoming outlooks are given. This review provides up to date knowledge on the advancement of the application of graphenebased electrospun fibers/composite materials as EMI shields and the outlook for high-performance electrospun fibers/compositebased EMI shielding materials.
# Introduction
With consideration to electromagnetic waves (EMWs), it is well-known that a change in the electric field (EF) will directly result in a change in the magnetic field because both fields oscillate in a similar direction and are abrupt to one another too, hence, developing EM waves within the fluctuating field. EMWs may propagate in fluid, solid, gas, and vacuum without requiring a transfer medium, which implies they can move wave energy on a plane made of magnetic fields and/or EFs. In the meantime, the EM radiation created by EM waves incredibly affects the outer layers and cannot get back to the propagation beginning once engendered: this impact is referred to as electromagnetic interference (EMI). [bib_ref] Effects of diameter and hollow structure on the microwave absorption properties of..., Chu [/bib_ref] Due to fast-progressing information technology innovation, electronic gadgets are assuming significant parts of our everyday life. However, electronic gadgets can likewise progress to issues, such as EMI and many others. [bib_ref] Enhancing the electromagnetic interference shielding of flexible films with reduced graphene oxide-based..., Godoy [/bib_ref] The presence of EMI not only causes challenges in customary battle weapons or electronics but also makes these materials experience unusable performance; in addition, it purposes unacceptable results on accuracy of instruments, for example, the unusual activity and estimation blunders of the instruments or divulgence of private archives, subsequently influencing public safety. Additionally, EMW radiation likewise causes harm to humans. For example, the focal sensory system of the human body can be irreparably harmed, due to long haul exposure in an unnecessary electromagnetic radiation climate. The rise of EMI shields tackles a progression of previously mentioned issues brought about by EM waves: these shielding fabrics could conquer the electric/magnetic domains within the locale and viably self-control the spread of EM waves starting with any district and then onto the next. Customary EM wave shields are for the most part fabricated from metalbased components that are generally folded over the gadget to shield it from EMI. Notwithstanding, high-thickness metalbased materials are hard to handle and are easily oxidized and eroded noticeably, which genuinely hinders their utilization in EMI shielding. [bib_ref] Ultra-thin and highly flexible cellulose nanofiber/silver nanowire conductive paper for effective electromagnetic..., Chen [/bib_ref] Moreover, with the improvement of presentday electronics, such as cell phones, PCs, and other small devices under unforgiving conditions, EMI shielding materials progressed with adaptability and low thickness, and the brilliant EMI shielding performance has drawn much consideration lately. One-dimensional (1D) nanofibers (NFs) have become famous materials lately because of their customizable subatomic designs, low thickness, high porosity, and good mechanical properties. [bib_ref] Electrospun polymer nanofibers: The booming cutting edge technology, Raghavan [/bib_ref] In the meantime, 1D NFs have demonstrated incredible possibilities for planning lightweight, adaptable, and exceptional EMI shielding characteristics. [bib_ref] Morphology, thermal, mechanical properties and ageing of nylon 6, 6/graphene nanofibers as..., Maccaferri [/bib_ref] [bib_ref] Highly safe lithium-ion batteries: High strength separator from polyformaldehyde/cellulose nanofibers blend, Liu [/bib_ref] NFs are continually being created, and their design advancements are additionally improved bit by bit, for example, for stage partition, 7 stretching technique, [bib_ref] Optimizing stretching conditions in fabrication of PTFE hollow fiber membrane for performance..., Li [/bib_ref] self-gathering strategy, [bib_ref] Proteinmimetic peptide nanofibers: Motif design, self-assembly synthesis, and sequence-specific biomedical applications, Zhang [/bib_ref] electrospinning innovation, [bib_ref] Influence of pretreatment and mechanical nanofibrillation energy on properties of nanofibers from..., Balea [/bib_ref] synthetic and mechanical division, [bib_ref] Influence of pretreatment and mechanical nanofibrillation energy on properties of nanofibers from..., Balea [/bib_ref] and in situ polymerization. [bib_ref] Microwave-assisted chemical-vapor-induced in situ polymerization of polyaniline nanofibers on graphite electrode for..., Li [/bib_ref] Electrospinning is a compelling and generally utilized methodology for setting 1D NFs, inferable from its straightforward, adaptable, climate-amicable, huge scope creation, short creation cycle, and simple to control technique. 3−14 Electrospun NFs, per the literature, are generally employed within the drug delivery niche, [bib_ref] Preparation of antibacterial PCL/PVP-AgNP Janus nanofibers by uniaxial electrospinning, Li [/bib_ref] organic tissue designing, [bib_ref] Poly(lactic acid) nanofibrous scaffolds for tissue engineering, Santoro [/bib_ref] energy storage, [bib_ref] High permittivity nanocomposites fabricated from electrospun polyimide/ BaTiO3 hybrid nanofibers, Xu [/bib_ref] ecological catalysis, as well as adsorption, [bib_ref] Meso-microporous carbon nanofibers with in-situ embedded Co nanoparticles for catalytic oxidization of..., Cai [/bib_ref] wireless telecommunications, [bib_ref] Electrically conductive and fluorine free superhydrophobic strain sensors based on SiO 2..., Gao [/bib_ref] armed forces,and aviation. [bib_ref] Novel cellulose− halloysite hemostatic nanocomposite fibers with a dramatic reduction in human..., Udangawa [/bib_ref] Specifically, an enormous number of experiments on the use of electrospun filaments for EMI shielding have been done recently because of their adaptability, light weight, low thickness, and usefulness. Notwithstanding 1D NFs, aerogels and films are likewise normal EMI protecting materials: contrasted with NFs, aerogels have higher porosity and lower thickness; however, their planning interaction is substantially more confounded and energy-consuming. [bib_ref] Wood-inspired anisotropic cellulose nanofibril composite sponges for multifunctional applications, Chen [/bib_ref] Furthermore, films have basic manufacture measure; however, their little explicit surface region very much restricts their viable utilization in EM wave sheltering. [bib_ref] Wood-inspired anisotropic cellulose nanofibril composite sponges for multifunctional applications, Chen [/bib_ref] Although several reviews papers credited to electrospinning and its pertinent utilization are accounted for, no important report has been given on the graphene-reinforced polymer electrospun nanocomposite/fiber substrates and their use for EMI shielding. In this regard, we have herewith given a concise discussion on the electrospinning of graphene-based polymer nanocomposites/fibers aimed at EM wave sheltering, zeroing in on various types of electrospun filaments and associated post-treatment of spun fibers for EM wave sheltering, as well as distinctive comparison of impacting factors, as depicted in.
## Fundamentals of electrospinning (es).
The electrospinning hypothesis, projected by William Gilbert around 1600, was framed from the development of coneshaped H 2 O in the presence of applied EF. Over the course of the following 300 years, however, the connected compositions and licenses were continually distributed, and electrospinning innovation was not even additionally created because of the absence of sufficient characterization instruments. Until the early 1990s, this innovation was attributed to the creation of electron-magnifying tools to describe materials with nanoscale sizes by Darrell Reneker, Gregory Rutledge, and Wendorff. [bib_ref] A review on electrospinning nanofibers in the field of microwave absorption, Huang [/bib_ref] ES innovation can handle the fiber breadth from the nanometer phase/scale to the micron phase, which is one of the principal approaches to manufacture 1D NFs.
The electrospinning gadget principally incorporates three sections: the spinneret, high-voltage supply of power, and fiber collection gadget. [fig_ref] Figure 2: Schematic presentation of the electrospinning technique [/fig_ref] shows a schematic presentation of the electrospinning technique. The polymeric mixture for ES within the needle is moved using the needle siphon, even as a drop is framed at the nozzle/spinner, aided by high surface tension coupled with gravity (in some setups). At the point where the drop is put in the climate of a high electrostatic arena, a gradual increase of the electrostatic repulsion results in extension of the drop, forming a cone shape under the activity of steady force. When the electrostatic voltage at the spinning nozzle approaches a critical point, the repulsion due to electrostatic force balances with the droplets' surface tension, resulting in a Taylor cone. [bib_ref] Fabrication of porous fibers via electrospinning: strategies and applications, Huang [/bib_ref] With the steady rise in the static voltage, a fiber is shot out at the edge of the formed Taylor cone and deposited on the collected gadget. In the ES technique, the viscosity/concentration of polymer solution, loading voltage, feed rate, distance for fiber collection, chamber temperature, and/or humidity directly affect the fiber morphologies. [bib_ref] Electromagnetic interference shield of highly thermal-conducting, light-weight, and flexible electrospun nylon 66..., Kim [/bib_ref] Until now, many scientific investigations have been committed to investigating the electrospun strands with further developed EM wave sheltering fabrics due to high perviousness, nanoscale diameter, excellent tensile/dynamic characteristics, physical-tailed properties, and abundant skin effect/interfacial contacts aimed at dispelling EM oomph. [bib_ref] Electromagnetic interference shield of highly thermal-conducting, light-weight, and flexible electrospun nylon 66..., Kim [/bib_ref] [bib_ref] Electrospun polyimide nonwovens with enhanced mechanical and thermal properties by addition of..., Wang [/bib_ref] 1.2. Introduction to EMI Shielding. EM waves can increase within the void or a given medium. They are categorized as radio waves, microwaves, infrared light, visible light, ultraviolet light X-rays, and γ-rays per their wavelength and/or frequency. The EMW with varying wavelength and/or frequency can cause diverse influences on peripheral bodies. Whenever EMW strikes the surface of an object, some portion of the EMWs is reflected, while the residual waves pass through the incident object(s). This might as well be transmitted/reflected at the shield deuce interfaces within, resulting in uninterrupted attenuation.The SE is generally adopted to quantitatively portray the sheltering impact of EM waves in decibels (dB), as EMI shields having 20 dB (∼99% weakening of EM waves) meet business necessities. [bib_ref] Effects of diameter and hollow structure on the microwave absorption properties of..., Chu [/bib_ref] By estimating the EF strength E 0 without a shielding material and E S at a similar spot with a shielding material, SE could be determined using eq 1. SE could likewise be determined utilizing eq 2 by estimating the attractive field strength H 0 and H s without shielding material and a shielding material, individually. Essentially, SE can be determined by eq 3 by estimating the power density P 0 (without an EMI shielding material) and P s (with an EMI shielding material).
[formula] = + + SE SE SE SE A R M (4) = t f SE 8.7 A (5) = i k j j j j j y { z z z z z f SE 10 log 16 R r(6) [/formula]
Also, the SE of the eventual EM wave shields is determined by utilizing the transmission line theory, and the SE is obtained by combining the reflection loss (SE R ), absorption loss (SE A ), and multiple internal reflections (SE M ) (eq 4). SE A and SE R are shown in eqs 5 and 6, individually.From the equations as described above, f, t, r, μ r , and σ are the electromagnetic frequency, thickness, distance between the source field and the considered EMI shield(s), the shields' relative permeability, and its relative conductivity, respectively.From the above expressions, it is inferred that the EMI SE is affected by the conductivity, magnetism, thickness, and permeability of the shielding material. The addition of a magnetic/conductive phase to the substratum fabrics can make it retain magnetic as well as electric dipoles which weaken the ohmic resistance (impedance) of shields, resulting in enhanced mismatch to achieve a reflection of the EMWs. [bib_ref] Integration of MCMBs/MWCNTs with Fe 3 O 4 in a flexible and..., Chaudhary [/bib_ref] In the interim, magnetic dipoles within the shielding fabric may effectively create eddy current loss, natural loss, interfacial polarization, and magnetic loss, consequently resulting in enhanced EMI shielding. Furthermore, SE M could be insignificant when SE A ≥ 15 dB. [bib_ref] Integration of MCMBs/MWCNTs with Fe 3 O 4 in a flexible and..., Chaudhary [/bib_ref] The distance needed for the EMW constriction to 1/e or 37% inside the EMI shields is referred to as skin depth (δ), which is determined utilizing eq 7.
[formula] = f 1(7) [/formula]
The EMW disseminated by reflection on a superficial level and within the materials is essentially dispersed as heat energy. Additionally, EM waves' reflection within the shield might bring about a decrease in sent signs, which is not identified with the absorption of EM waves within the shield. In shields that are monolithic and isotropic in nature, the absorption (A) is derivatively presented from the reflectance (R) as well as estimated transmittance (T): = A I R T [bib_ref] Optimizing stretching conditions in fabrication of PTFE hollow fiber membrane for performance..., Li [/bib_ref] The EM wave shielding component incorporates its reflection loss with respect to the shield's surface, the SE A and the numerous SE R within the shielding fabric. With regard to electrospun strands, the specific 1D structural design can build a thick conductive configuration, which can prompt improved reflection of EMWs. Moreover, low impedance could remain because of the high porosity and enormous aspect ratio of electrospun NFs. Accordingly, EM waves could go into the fibrous material, and loss of EM energy will occur within its interfaces among EM waves and the electrospun fibers/strands. This extraordinary impact of damping EM waves and displaying superb shielding impact presents electrospun filaments/fibers as a promising competitors among EMI shields. A graphical representation of electromagnetic wave as well as the A segment of traveling EM wave with in-phase mutually orthogonal E and B is depicted in.
## Graphene
Graphene is a 2D single-atom-thick plate of sp 2 -hybridized carbon atoms with carbon at a chain length of 0.142 nm. [bib_ref] Band structure of graphite, Slonczewski [/bib_ref] It really is the "skinniest material", consisting of a single layer of carbon atoms assembled in an axial honeycomb crystal lattice. This same basic structural component of allotropes of carbon such as graphite, carbon nanotubes (CNTs), as well as fullerenes has always been graphene. Graphite has been around for a great many years. In a pioneering investigation, Novoselov et al. [bib_ref] Prospect of DFT utilization in polymer-graphene composites for electromagnetic interference shielding application:..., Novoselov [/bib_ref] managed to identify monolayers of graphene as well as other 2D crystals for the very first time. Graphene is a versatile material because of its excellent mechanical properties, along with a Young's modulus of 1 TPa, 33 a mean tensile strength of 13,010 GPa, outstanding thermal heat transfer (5000 W/m/K) [bib_ref] Superior thermal conductivity of single-layer graphene, Balandin [/bib_ref] and charge transport characteristics (250,000 cm 2 /V), [bib_ref] Prospect of DFT utilization in polymer-graphene composites for electromagnetic interference shielding application:..., Novoselov [/bib_ref] and good EMI shielding (>20 dB). Such extraordinary characteristics have also sparked tremendous interest in graphene for practical uses such as thermally as well as conductive reinforced nanomaterials, next-generation logic gadgets, sensing, associated electronics, transparent as well as dynamic electrodes for display systems as well as photovoltaic panels, capacitors, and lithium batteries. [bib_ref] Graphene-based liquid crystal device, Blake [/bib_ref] Hence, graphene is indeed the new stuff for nanostructured materials as well as nanodevice manufacturering in the coming years. Assorted additives have also been attempted for producing nanocomposites with improved characteristics since the revelation of polymer nanocomposites. [bib_ref] Carbon nanotube−polymer composites: chemistry, processing, mechanical and electrical properties, Okada [/bib_ref] Inorganic additives including organic montmorillonite-type layered silicate molecules or synthesized clay, nanoparticles, as well as carbon-based fillers such as carbon black, expanded graphite (EG), carbon nanotubes (CNF), and CNTs are examples. Seeing as inorganic fillers possess poor electrical and thermal conductivity, diverse carbon-based additives are becoming more appealing. CNTs, despite their high cost, have proven to be excellent conductive fillers. Graphene's tensile strength is comparable to or slightly higher than that of CNTs. Furthermore, graphene has thermoelectric conductivities significantly higher than those of CNTs. [bib_ref] Strength and breaking mechanism of multiwalled carbon nanotubes under tensile load, Yu [/bib_ref] The amazing properties of graphene@polymer nanomaterials included the amalgamation of exceptional mechanical properties, outstanding thermal conductivity, as well as electrical conductivity. As a result, graphene has emerged as a superior padding for polymer nanocomposites. Massive interest in graphene has also resulted in extensive research, and as a result, a large number of articles have been published, as well as outstanding evaluations on graphene and modified graphene-based materials. [bib_ref] Design of an efficient and selective adsorbent of cationic dye through activated..., Bhattacharyya [/bib_ref] Several processes for producing graphene sheets have been reported. Graphene is synthesized using thermal chemical vapor deposition (CVD) on metal surfaces and epitaxial growth from silicon carbide (SiC) without the use of complex mechanical or chemical treatments. 41 Plasma-enhanced CVD is another method for producing graphene sheets at a lower temperature than thermal CVD. [bib_ref] Free-standing subnanometer graphite sheets, Wang [/bib_ref] Physical exfoliation of graphite has generated graphene sheets; 32 this path is indeed recognized also as the "Scotch tape" or "peel-off" technique. From another method, graphene is created by chemically reducing graphene oxide (GO). [bib_ref] Aksay, I. A. Functionalized single graphene sheets derived from splitting graphite oxide, Schniepp [/bib_ref] Among all of these methods, reduced in the presence of graphite to GO is regarded as one of the most cost-effective as well as a viable methods of producing large quantities of graphene sheets. [bib_ref] Aksay, I. A. Functionalized single graphene sheets derived from splitting graphite oxide, Schniepp [/bib_ref] Multiple challenges must be resolved before the full potential of graphene-or graphene-oxide-based composites can be realized. The primary issue mostly in the fabrication of massive quantities of graphene has been that pristine graphene invariably tends to agglomerate to constitute stacked 3D graphitic edifices via π−π interaction but as a result cannot be processed further owing to its high hydrophobic character. Furthermore, the use of pristine graphene in polymeric composites is difficult because phase-separated nanocomposite forms tend to result in mechanical malfunction at low load conditions than in a neat polymeric matrix. [bib_ref] The chemistry of graphene, Loh [/bib_ref] That is also due to the incompatibility of graphene's surface with organic polymeric composites. Other major concerns, including (i) homogeneous distribution of nanosheets in a polymeric matrix and (ii) interfacial adhesion among graphene as well as the symbiont polymer, do have a direct impact on the nanocomposites' characteristics. Controlling the conformational changes, folding, as well as flexing of graphene nanostructures is also critical.
## Advantages and disadvantages of graphene-based electrospun polymer composite. 2.1.1. advantages of graphene-based electrospun polymer composite.
The electrochemical stability of intrinsic conducting polymers/ graphene composites has been shown to be significantly greater than that of virgin intrinsic conducting polymers in advanced materials applications such as electrodes. The improved mechanical, thermal, and electrical properties of graphene-reinforced polymer-based nanocomposites vary depending on the nanocomposites. [bib_ref] Kinetically controlled localization of carbon nanotubes in polylactide/poly (vinylidene fluoride) blend nanocomposites..., Guo [/bib_ref] Materials based on graphene or polymers have many benefits, including high specific active surface areas, excellent electron transport capabilities, and good capacitance. Because of the establishment of a "tortuous path" in the presence of graphene in the nanocomposites, nanocomposites with homogeneously dispersed graphene in the polymer matrix have also demonstrated good barrier characteristics, especially in packaging applications.
Graphene-based polymer nanocomposites have shown excellent performance in terms of EMI shielding. The EMI shielding efficiency of low weight percent graphene-filled polymeric composites can present an EMI shielding performance that is greater than that of the commercial target value of ≥20 dB. [bib_ref] Kinetically controlled localization of carbon nanotubes in polylactide/poly (vinylidene fluoride) blend nanocomposites..., Guo [/bib_ref] Graphene-based polymer nanocomposites have been proven to outperform neat polymers and conventional graphite-based composites in terms of mechanical characteristics. They also outperform their neat polymer counterparts in terms of thermal stability.
Polymer graphene-based nanocomposites increased electrical conductivity by several orders of magnitude. The development of a conducting network by graphene sheets in the polymer matrix results in a significant improvement in electrical conductivity. [bib_ref] Kinetically controlled localization of carbon nanotubes in polylactide/poly (vinylidene fluoride) blend nanocomposites..., Guo [/bib_ref] Graphene-reinforced polymer composites also outperform their metal-based counterparts in terms of light weightness.
## Disadvantages of graphene-based electrospun polymer
Composite. Nanocomposite materials containing graphene/polymer may face great challenges in future applications, especially in scientific and medical fields, due to the cytotoxic effects of graphene−polymer composites.
A major technical problem is dispersing the nanofillers uniformly within polymer matrixes. In the dearth of covalent bonding as well as existing nonbinding interactions such as π−π interactions and hydrogen bonding, poor interfacial adhesion throughout graphene/polymer nanocomposite em-phasizes the significance of platelet surface modification in the reinforcement of the need for constant improvement in the path.
## Categories of graphene-based es-fabricated fibers for emi shields
Electrospinning is an approach to create various types of filaments/fibrous strands, such as virgin filaments with one segment and composite strands with diverse fillers. In this regard, neat polymeric electrospun fibers and their composites utilized for EMI shielding will be presented.
## Graphene-based neat es-fabricated fibers for emi shields.
Graphene-based polymer electrospun nanocomposite fibers for EMI shielding applications generally can be engineered to possess dielectric tunable behavior as well as electrical conductivity (EC). Such ES of neat fibers (graphene−ceramic fibers and/or graphene−carbon fibers) have been reportedly applied for EMI shielding per the literature. 3,12,24,28,46,47 3.1.1. Graphene−Ceramic Electrospun NFs. Graphene− ceramic composite NFs such as graphene−SiC, graphene− SiO 2 , graphene−Si 3 N 4 NFs fabricated via the ES approach have benefits including high Young's modulus, low density, chemical stability, as well as tunable dielectric character, making them an excellent choice for EMI shields within the GHz range of frequency. [bib_ref] Lightweight and flexible hybrid film based on delicate design of electrospun nanofibers..., Huang [/bib_ref] Wide band gap semiconductors such as SiC materials can present the impedance match to attain high-efficiency attenuation. Also, the EM loss of the virgin silicon carbide NFs is largely derived from its state loss of available carbon and the polarization loss of nanocrystalline silicon carbide. [bib_ref] Relationship between microstructure and electromagnetic properties of SiC fibers, Mo [/bib_ref] In the first report of its kind, electrospun SiC (SiC nanocrystals)/ rGO core−shell nanowires have been utilized as highly Microwave RL for the samples of (d) S1300, (e) S1400, (f) S1500, and (g) S1600. Reproduced with permission from ref 52. Copyright 2018 Elsevier Science Ltd. efficient, lightweight EM wave absorbers/shields successfully fabricated via solution electrospinning followed by a postannealing process in a N 2 environment. The EM wave absorption performance of these nanowires was reportedly enhanced proficiently by tuning the included GO solution (1.0 mL) in the ES solutions, which displayed a SE R of −56.3 dB (optimum), as well as its actual bandwidth absorption, which could be ≤6.2 GHz covering the entire Ku band (12.2−18.0 GHz). These authors established that the explicit core@shell assembly, defects as well as rGO and C surface functionalities, along with the skin-effect/the grain boundaries between the carbon cluster and the SiC nanocrystals, all play a role in the electrospun nanowire's EM wave absorption ability. [bib_ref] SiC/rGO Core− Shell Nanowire as a Lightweight, Highly Efficient Gigahertz Electromagnetic Wave..., Wang [/bib_ref] Electrospun generated Ti 3 C 2 Tx/GO (MXene@GO) nanocomposite aerogel microparticles aimed at tunable highperformance EM wave absorption/shielding have been reported. [bib_ref] Electrospun generation of Ti3C2Tx MXene@ graphene oxide hybrid aerogel microspheres for tunable..., Li [/bib_ref] These authors reported that the fabrication of 2D nanomaterials into hybrids is an effective manner to prepare high-performance MA systems, having heterointerfaces that offer innovative loss phenomenon, thereby making up for the limitations of sole materials in EM energy attenuation, [bib_ref] Electrospun generation of Ti3C2Tx MXene@ graphene oxide hybrid aerogel microspheres for tunable..., Li [/bib_ref] with the consideration that, in practical deployments, thin and lightweight microwave-absorbing materials have excellent performance at low-frequency bands that are favorable, and they engineered hybrid aerogels fabricated with GO and Ti 3 C 2 Tx through quick freeze-assisted electrostatic spinning. [bib_ref] Electrospun generation of Ti3C2Tx MXene@ graphene oxide hybrid aerogel microspheres for tunable..., Li [/bib_ref] The differences in the conductivity GO and Ti 3 C 2 Tx abundant heterointerfaces and surface groups were generated, even as the Ti 3 C 2 Tx@GO hybrid exhibited optimized hybrid properties as well as MA. [bib_ref] Electrospun generation of Ti3C2Tx MXene@ graphene oxide hybrid aerogel microspheres for tunable..., Li [/bib_ref] They observed that the inimitable aerogel structure offered light weight along with stretching of the attenuating routes when EM waves were injected even at a relatively small reinforcing filler inclusion of 10.0 wt % at a material thickness of 1.2 mm, while optimized M@GAMS structures exhibited SE R of −49.1 dB at 14.2 GHz. [bib_ref] Electrospun generation of Ti3C2Tx MXene@ graphene oxide hybrid aerogel microspheres for tunable..., Li [/bib_ref] More notably, the M@GAMS presents effective microwave absorption at the S-band, and the RL reaches −38.3 dB at 2.1 GHz at 5.0 mm thickness. [bib_ref] Electrospun generation of Ti3C2Tx MXene@ graphene oxide hybrid aerogel microspheres for tunable..., Li [/bib_ref] With the modification of such nanocomposite systems using additional active nanoparticles and/or conducting polymers as in blends, outstanding materials can be engineered using a similar path as postulated by the authors above.
A simple approach toward the tunable fabrication of graphite/silicon carbide hybrid nanostrands through ES techniques followed by annealing at elevated temperature has been demonstrated in another study. [bib_ref] Electrospinning of graphite/SiC hybrid nanowires with tunable dielectric and microwave absorption characteristics, Wang [/bib_ref] The hybrid nanowires [fig_ref] Figure 4: Morphology and element analysis [/fig_ref] , corresponding to the reported samples, S1300, S1400, S1500, and S1600, having wider EAB of ∼4.7 GHz and very thin absorber coating thickness of 1.7 mm. The hybrid nanowires' outstanding EM wave absorption capability was attributed to their well-engineered microstructural architecture, chemical composition, and the synergistic impact between graphite and SiC. [bib_ref] Electrospinning of graphite/SiC hybrid nanowires with tunable dielectric and microwave absorption characteristics, Wang [/bib_ref] These authors also suggested that the EM wave shielding performance of the fabricated hybrid nanowires could be boosted with a much lower RL value and a broader EAB by tuning the weight ratios of PCS and PVP in situ. [bib_ref] Electrospinning of graphite/SiC hybrid nanowires with tunable dielectric and microwave absorption characteristics, Wang [/bib_ref] Furthermore, they indicated that using ultralong nanowires as reinforcing fillers in SiO 2 , SiC, Si 3 N 4 , and other ceramics to improve their EM wave absorption performance as well as mechanical qualities would be exciting; [bib_ref] Electrospinning of graphite/SiC hybrid nanowires with tunable dielectric and microwave absorption characteristics, Wang [/bib_ref] we suggest similar systems with the inclusion of graphene for highperformance EM wave shields.
There are reports of lightweight-com-flexible nanocomposites having a multiscale double-unbroken conductive system of TiO 2 @SiO 2 @PPy and sandwich structure of TiO 2 @SiO 2 @ PPy@rGO, fabricated by electrospun TiO 2 /SiO 2 delicate structured NFs. [bib_ref] Lightweight and flexible hybrid film based on delicate design of electrospun nanofibers..., Huang [/bib_ref] The hybrid system performed as an active dissipative system, resulting in excellent EMI SE accounting for ∼30 dB in the X band as well as outstanding EMI SE of ∼13,829 dB cm 2 g −1 (0.089 g cm −3 ); also, these hybrids sustained decent electrical as well as EMI SE characteristic even after repeated bending, which indicates its satisfactory flexibility being a practicable way of preparing lightweight/ flexible hybrids toward high-performance shields for flexible electronics, healthcare, armed forces, devices, etc. [bib_ref] Lightweight and flexible hybrid film based on delicate design of electrospun nanofibers..., Huang [/bib_ref] From the available literature, the structural architecture, material defects, and functional groups of graphene and/or other included reinforcing components, along with the skineffect/filler boundaries between the graphene materials cluster and the adopted matrix, all play vital roles in enhancing EM wave shielding performance of the electrospun graphene-based EMI shields. [bib_ref] SiC/rGO Core− Shell Nanowire as a Lightweight, Highly Efficient Gigahertz Electromagnetic Wave..., Wang [/bib_ref] Reports in this niche of materials are rare or limited, though there is a growing interest in the fabrication of these materials owing to their novel properties.
## Graphene-based electrospun nanocomposite fibers having fillers for em wave shields.
Even though we know that electrospun virgin graphene−polymer, graphene−ceramic, etc., nanocomposite fibrous materials present certain advantages in EM wave shields, there still lingers certain challenges to be solved. Categorization of these kinds of nanofibrous systems is constrained by flexibility, transparency, mechanical properties, etc., in some instances, depending on the material entity combinations, thereby significantly limiting their applications in real life. The inclusion of novel active and/ or nonactive fillers into the electrospun fibers has shown to be an efficient approach in solving some if not all of the aforementioned challenges along with greatly broadening the choice of raw materials in the case of composites. Conventional reinforcing or nonreinforcing fillers used in electrospun nanocomposites typically includes conductive (thermal or electrical) fillers for enhancing the EC, magnetic fillers for enhancing the magnetic performance, as well as others such as ceramics aimed at adjusting dielectric characteristics of these materials.
## Graphene-based conductive fillers electrospun composite fibers.
Thermal and electric conductive graphene-based electrospun composite fibrous materials occasionally are categorized into two categories, C-based and metal-based. Carbon nanotubes, GP, as well as transition metal oxides/carbides are among the utmost utilized C-based reinforcing filler materials for nanocomposite fiber fabrication via ES. Furthermore, metal (nano)particles are often explored as reinforcing fillers for enhancing the thermal/EC of electrospun graphene-based nanocomposite fibers. The following subsections present the diverse graphene-based electrospun nanocomposites per the available literature.
3.2.1.1. Graphene−Graphene/Graphite Electrospun Composites. Graphene−graphene electrospun composites are composites containing two or more forms of graphene in their various forms (GO, rGO, graphene (GN), GQDs, rGQDs, etc.). First, we understand that a single carbon atom has six electrons in its orbitals across the 1s 2 , 2s 2 , 2px, and 2py orbitals. [bib_ref] Sol-gel synthesized carbon nanoparticles as supercapacitor electrodes with ultralong cycling stability, Verma [/bib_ref] The amalgamation of two or more carbon atoms results in the promotion of one electron from the 2s orbital to the 2pz orbital. Graphite formation from carbon atoms results in three sp 2 orbitals created through the hybridization of carbons' 2s 2 , 2px, and 2py atomic orbitals. Carbons' sp 2hybridized orbitals form the strong in-plane bonds, whereas the 2pz orbital creates the weaker out-of-plane bond between the amalgamated graphite sheets. The splitting apart of the weak bonds between the graphite sheets results in graphene, where the 2pz electrons no longer take part in the bonding and are what give graphene its unique exceptional electrical and optical characteristics. Graphene has been well-known (labeled and marketed) as a wonder material ever since it was first unambiguously synthesized in 2004.
Graphene nanomaterials are produced either by the topdown and/or bottom-up assembly of carbon atoms, which involves the treatment of graphite ore to synthesize graphene through processes such as CVD, chemical synthesis/chemical exfoliation, and mechanical exfoliation, although certain approaches are often amalgamated in some cases. [bib_ref] Recent advances in carbon nanomaterials as electrochemical biosensors, Kour [/bib_ref] As one of the most striking EMI shielding reinforcing fillers, GP has the capacity to induce flaws and polarization losses, which could contribute to the attenuation of EM waves. Furthermore, due to its excellent thermal conductivity, electrospun fibrous materials can be used in harsh/hightemperature situations, and GP as reinforcing fillers can improve the conductivity of electrospun nanofibrous materials while also providing flexibility. [bib_ref] Ultrathin flexible graphene film: an excellent thermal conducting material with efficient EMI..., Shen [/bib_ref] [bib_ref] Ultrathin graphene: electrical properties and highly efficient electromagnetic interference shielding, Cao [/bib_ref] Reports by Guo et al. [bib_ref] Ultrathin graphene: electrical properties and highly efficient electromagnetic interference shielding, Cao [/bib_ref] on electrospun polystyrene (PS)/ graphite nanoplatelets (GNPs) nanocomposite fibers, applied as EMI shields have been published. Their investigation established that the electrospun fibers upon cold-pressing were observed to be stacked, while the hot-pressing technique was utilized to fabricate well-tuned nanocomposites. [bib_ref] Kinetically controlled localization of carbon nanotubes in polylactide/poly (vinylidene fluoride) blend nanocomposites..., Guo [/bib_ref] It was established that the EMI SE values of nanocomposites reached ∼33 dB after the inclusion of 35 wt % GNPs with a thickness of 3 mm, as a result of the creation of a conductive connective architecture as well as stratified structural architecture within the nanocomposites. [bib_ref] Kinetically controlled localization of carbon nanotubes in polylactide/poly (vinylidene fluoride) blend nanocomposites..., Guo [/bib_ref] Also, in another work, Song et al. [bib_ref] Interfacial engineering of carbon nanofiber−graphene−carbon nanofiber heterojunctions in flexible lightweight electromagnetic shielding..., Song [/bib_ref] reported an all-carbon flexible nanocomposite fibrous system for effective application as EMI shields. The authors reported that the inclusion of graphene nanosheet (GN) reinforcing fillers was associated with the polymer fibers in the ES method with the influence of van der Waals forces, as the nanocomposite fibers were carbonized to fabricate CNF−graphene nanosheet−CNF (CNF−GN−CNF) heterojunctions, as depicted in, showing the virgin/pristine surface was smooth, while the amount of CNF−GN−CNF heterojunctions was increased with increasing GN inclusion, as in Figure 5b−d. [bib_ref] Interfacial engineering of carbon nanofiber−graphene−carbon nanofiber heterojunctions in flexible lightweight electromagnetic shielding..., Song [/bib_ref] established that the inclusion of CNF−GN−CNF heterojunctions significantly influences the EC of fires due to the CNF−GN−CNF heterojunctions, which modifies the interfacial interaction of the CNF structural architectures. [bib_ref] Interfacial engineering of carbon nanofiber−graphene−carbon nanofiber heterojunctions in flexible lightweight electromagnetic shielding..., Song [/bib_ref] An EC of 800 S/m was attained by these authors for composite fiber with 17.2 wt % loading; nevertheless, it decreased with the inclusion of 31.9 wt % fiber due to excess GN loading, which resulted in the fracture of fibers and, consequently, the decrease in conductivity. [bib_ref] Interfacial engineering of carbon nanofiber−graphene−carbon nanofiber heterojunctions in flexible lightweight electromagnetic shielding..., Song [/bib_ref] As shown in, CNF−GN networks with a thickness of 0.22−0.27 mm (50−60% thicker than pristine CNF networks) apparently achieved a SE total of 25−28 dB. Flexible all-carbon structure networks, according to scientists, could play an important role in portable, flexible electronic gadgets. 57
## Graphene−nanometal electrospun composites.
The utilization of metal (nano)particles with outstanding magnetic, EC, and nanostructural properties as fillers in the fabrication of EM wave shields via the electrospun technique has recently gained momentum owing to the nanostructural architecture attained via the ES technique. Studies patterning the EMI shielding characteristics of electrospun graphenebased systems with the inclusion of diverse metallic nanoparticles such as Cd, Cu, Co, Fe, Zn, etc. should be investigated because the diverse nanometal particle loading will reveal the noticeable impact on the EC, magnetic, and EMI SE characteristics of the final EM wave shields. 40,58, [bib_ref] Temperature induced transformation of Co@C nanoparticle in 3D hierarchical core-shell nanofiber network..., Huang [/bib_ref] Graphene offers an exceptionally ultrahigh thermal conductivity in comparison to other well-known thermally conductive fillers. Thermally conductive (Ag/rGO)/polyimide (Ag/rGO)/PI) nanocomposites prepared via electrospinning of an in situ polymerized (Ag/rGO)/polyamide electrospun suspension along with a hot-press procedure has been postulated. [bib_ref] Kinetically controlled localization of carbon nanotubes in polylactide/poly (vinylidene fluoride) blend nanocomposites..., Guo [/bib_ref] Parts (a) and (b) of [fig_ref] Figure 6: Thermal conductivities of the [/fig_ref] , respectively, show the mass fraction of Ag/rGO and rGO influencing the λ values of the (Ag/rGO)/PI and rGO/PI nanocomposites and infrared thermal images of (Ag/rGO)/PI nanocomposites. The glass transition temperature (T g ), thermal conductivity (λ), as well as THRI (heat resistance index) of the (nano)composites were all enhanced by increasing the Ag/rGO nanofiller inclusion, even as the modeled data relatively agreed with the experimental findings compared to the findings obtained from reported conventional models. [bib_ref] Reduced graphene oxide heterostructured silver nanoparticles significantly enhanced thermal conductivities in hot-pressed..., Guo [/bib_ref] The establishment of the importance of the electrospinning process in helping to align rGO/Ag nanofillers to get oriented along the electrospun PI fibers as well embedding of the rGO/Ag nanofillers within or on the electrospun PI fibers was reported by these authors. They also observed that at a Ag to rGO mass ratio of 1:4, when the overall mass% of rGO/Ag nanofillers reached 15 wt %, the significant conductivity of the (rGO/Ag)/PI nanocomposites reached its peak (2.12 W/(m K)), being 8 times as extreme as virgin PI (λ of 0.27 W/(m K)). [bib_ref] Reduced graphene oxide heterostructured silver nanoparticles significantly enhanced thermal conductivities in hot-pressed..., Guo [/bib_ref] The enhanced thermal conductivity implies increased EC also, though it has not been directly corrected by these authors. These kinds of formulations, if improved upon, could serve as effective materials having outstanding properties for EMI shields and other state-of-the-art applications.
A flexible PVDF film with alternate oriented graphene/Ni nanochains for EMI shielding and thermal management with excellent EC and thermal conductivity of 76.8 S/m and 8.96 W/mK has been presented in another study. The authors also discovered that by combining high EC and synergetic electromagnetic loss, as well as multilevel electromagnetic multireflection, a film with a thickness of 0.5 mm exhibits a high EMI SE of 43.3 dB, revealing a 98.6% increase over a homogeneous film with a thickness of 21.8 dB, which could be further enhanced to 51.4 dB at a thickness of 0.7 mm. [bib_ref] Lightweight and flexible reduced graphene oxide/water-borne polyurethane composites with high electrical conductivity..., Liang [/bib_ref] Other fillers such as conductive transition metal carbides/ ceramics that typically hold high stiffness as well as efficient EC than can be used as nanofillers toward the fabrication of composite fibers via ES as EMI shields have also been reported. 62 However, there is a need for in-depth study on these composites with the inclusion of graphene materials which is limited at the moment.
Also, a group of authors has revealed from their study aimed at meeting the requirement in ultrathin portable electronic devices, flexible EMI shields "CNF−GN−CNF" have a thickness of 0.22−0.27 which exhibited a mean EMI SE around 25−28 dB. 57
## Electrospun polymer graphene nanocomposite fiber post-treatments for emi shielding
Post-treatment of electrospun polymer graphene nanocomposite fibers is aimed at amalgamating the electrospun fibers with other materials having enhanced functionality. Functionalized electrospun nanocomposite fibrous materials present better EMI shielding characteristics in comparison to virgin electrospun polymeric fibers and/or fibers with nanofillers because of the amalgamation of diverse functional entities. Conventional post-treatment approaches for processing functionalized electrospun fibrous materials could consist of diverse deposition approaches [fig_ref] Figure 7: Scheme showing diverse deposition approaches [/fig_ref] , though electroless deposition, spray deposition, dip deposition, as well as vacuum deposition have been mostly utilized. Thin film deposition is the method of applying a thin film to the surface of a substrate. It also refers to the method of depositing a fine layer of material onto another material as well as on the substratum of previous substrate material or layer upon layer. In this perspective, "thin" refers to overlay thickness within the nanometer range to a few micrometers. This is a helpful tool for developing optics, such as reflective as well as anti-reflective paints, gadgets, including laminae of insulators, semiconductors, as well as conductors for circuit boards, and packaging, such as aluminum-plated PET membranes.
As shown in [fig_ref] Figure 7: Scheme showing diverse deposition approaches [/fig_ref] , deposition approaches are divided into two main types premised on if the methodology is predominantly chemical and otherwise physical. In terms of deposition or coating of graphene layers on the surface of electrospun nanomaterials, methods depending on vapor deposition, sublimation/epitaxy, as well as mechanical exfoliation typically produce single-layer carbon products of high purity. Still, their scalability is now only recently being evidenced, albeit under specified circumstances.
## The dip deposition technique.
Dip deposition is a simplistic approach utilized toward the fabrication of fibrous electrospun nanocomposites where impregnation solution is used to wrap/coat the surface of the fibers via dip deposition. By immersion of the substrate into a solution containing hydrolyzable organic or inorganic compounds and withdrawal at constant speed into an atmosphere containing water vapor, dip-coating is a simple, inexpensive, reliable, and reproducible method for the deposition of wet liquid films. A homogeneous liquid film forms on the substrate's surface when it is removed from the solution. As the solvents are evaporated at room temperature, possible chemical reactions will take place, leaving a thin film of coating on the surface. Usually, the film is hardened/chemically transformed via heat treatment after drying in an atmosphere of water. [fig_ref] Figure 8: Schematic presentation of the dip-coating method [/fig_ref] reports a visual presentation of the dip-coating process. These kinds of fibrous polymeric materials coated with functional coatings through dip deposition are proven effective as a shield against EM waves.
Huang et al. have studied flexible and lightweight electrospun fibrous nanocomposite (TiO 2 /SiO 2 /PPy/rGo (TSPG) films (thickness of 0.26 mm and density of 0.089 g cm −3 ) having a multiscale double-incessant structural architecture and sandwich architecture by exploiting dip deposition, displaying outstanding EMI SE. [bib_ref] Lightweight and flexible hybrid film based on delicate design of electrospun nanofibers..., Huang [/bib_ref] These authors utilized the dip-coating technique, where PPy polymerized onto the surface of the fibers as influenced by FeCl 3 , as well as GO plates selfassemble on both edges toward achieving functionalized fibers. Their report established that electrospun fibers having conductive coating displayed above average EMI SE properties. [bib_ref] Lightweight and flexible hybrid film based on delicate design of electrospun nanofibers..., Huang [/bib_ref] A facile and prevailing technique for the fabrication of elastic and lightweight graphene-based nanocomposites showing high EM wave shielding properties, produced via electrospinning of waterborne polyurethane (WPU) featuring sulfonate functionalities, has been proposed. The authors revealed in their study that GO bilayers wrapped around the WPU fiber matrix displayed excellent networks steered by the electrospun WPU fibers, followed by the use of hydroiodic acid (HI) to reduce the GO to highly reduced GO (rGO), resulting in rGO/WPU nanocomposites, having significantly enhanced EC (∼16.8 S/ m) along with good EMI SE of ∼34 dB within the X band. Yuan et al. reported the electrospinning of a composite fibrous material made of Fe 3 O 4 @SiO 2 @PPy@GO (FSPG) having an innovative core−shell sandwiched amalgamated edifice fashioned via dip-disposition to be applied as EMI shields. [bib_ref] Lightweight, flexible and strong core-shell nonwoven fabrics covered by reduced graphene oxide..., Yuan [/bib_ref] They reported the SE of FSPG film as 32 dB and total SE reaching 12,608.4 dB cm 2 g −1 at ∼0.27 mm thick. The high SE of the hybrid was reported to be ascribed to the magnetic loss of Fe 3 O 4 and graphite high λ, resulting in comparatively elevated EC of FSPG composite sheaths of ∼0.71 S/cm. [bib_ref] Lightweight, flexible and strong core-shell nonwoven fabrics covered by reduced graphene oxide..., Yuan [/bib_ref] The dip deposition process could be adopted as an effective tool for turning graphene-based electrospun fibrous materials toward the attainment of ultrahigh EMI SE systems.
## Electroless deposition.
This technique is referred to as the cheap deposition approach and could be adopted to formulate an uninterrupted as well as even a glaze on the surface of the spun substrate. Dense electroless deposited layers could boost the interfacial polarization as well as enhance the spun fibrous (nano)composite polarization loss, boosting EMI sheltering characteristics of the material. In an electroless coating, an ionic metal ion is reduced to a substrate by autocatalysis or chemical means without any external power source. This excludes approaches employed to conduct coatings without a current, such as immersion plating (deposits of copper on steel dipped in copper sulfate solution or nickel on steel dipped in chloride and boric acid bath) and homogeneous chemical reduction (silvering). In contrast to immersion plating or silvering, in which the nature of the base material itself acts as a reducing agent without further assistance, autocatalytically deposited electroless metallic coatings rely on a mechanism that is distinct from immersion plating or silvering. The term "electroless plating", sometimes called "autocatalytic plating", as shown in [fig_ref] Figure 9: Schematic illustration of equipment used in electroless coating studies depicted in its... [/fig_ref] , is about as old as electroplating. Von Liebig published for the very first time in 1835 the reduction of silver salts by reducing aldehydes. Wurtz 64 described metallic nickel plating with an electroless method from an aqueous solution in the presence of hypophosphite (reducing agent) as a chemical disaster in 1988. Roux confirmed in 1911 that there was evident precipitation of metal in the form of powder. These works, however, were not utilized in any practical applications. In this field, development received little support prior to World War II. By coating a nickel−tungsten alloy in the inner regions of tubes with a citrate solution containing an insoluble anode, Brenner and Riddell produced hypophosphite, a material with unusual reducing properties. Roux's previous 1950 patent, where the procedure was unstructured and complete, was declared to be separate and different by the U.S. Patent and Trademark Office. On the other hand, Brenner and Riddell's method was based on a catalytic approach that deposited only on catalytically active surfaces dipped in the plating bath.
In a similar vein to Wurtz and Roux's attempts, Brenner later acknowledged that his invention was an unexpected result. Nevertheless, he noted that the patent was used for U.S. government protection. It is true that in 1963 a widely circulated U.S. Army-based technical study spoke largely about Wurtz and Roux's discoveries and credited them to Brenner, for whom a patent was issued in 1956. Ultimately, this phenomenon was found to be caused by a chemical reaction involving nickel ions. The rest of the electroless coatings were applied after the electroless nickel deposit was started. A 1970 patent by Bell Laboratories led to the use of thick, pure soft gold plating for semiconductors and circuit boards. Following the invention of the first electroless gold coating bath at Bell Laboratories in 1970, semiconductors and circuit boards were plated with thick, pure soft gold. [bib_ref] A review of electroless gold deposition processes, Ali [/bib_ref] According to Narcus, the electroless copper deposit was first documented, and the first profitable application was achieved by Cahill and Zeblisky et al.in tartrate baths using formaldehyde as the reducing agent. Improved electroless copper formulations 67 demonstrated a greater rate of plating with highly stable conditions under a wide range of working situations. In 1954, Brenner and Riddell 68 developed electroless cobalt together with nickel deposition. These electroless cobalt deposition technologies have been widely used in the creation of magnetic films. In particular, thin films of cobalt have also been used as data storage devices due to their thinness, high coercivity, and strong remanence. Early in 1970, the practical application of electroless silver was established. [bib_ref] Deposition of silver, Sivertz [/bib_ref] There was the development of electroless aluminum, 70 electroless platinum, 71 electroless ruthenium, and electroless rhodium. In early 1984, materials such as plastics, ceramics, polymers, and other nonconducting materials were electrolessly coated before being exposed to electroplating. This event resulted in the formation of the remaining electroless metallic deposits. Meanwhile, electroless coating chemistry has emerged as one of the most promising and important fields of surface development as well as metallic finishes.
This process is yet to be used in graphene-based electrospun (nano)composites but has been used for the fabrication of innovative cellular sheaths having ultrahigh EMI sheltering properties by the deposition of Cu or Ag on electrospun polymeric NFs with the assistance of dopamine. [bib_ref] Flexible and stretchable MXene/ Polyurethane fabrics with delicate wrinkle structure design for..., Zeng [/bib_ref] These authors reported the max EMI SE of the membranes to be ∼53 dB at a paper-thin 2.5 μm thickness and a density of 1.6 g cm −3 , while the specific SE attained was up to 232,860 dB cm 2 g −1 . [bib_ref] Flexible and stretchable MXene/ Polyurethane fabrics with delicate wrinkle structure design for..., Zeng [/bib_ref] Considering the outstanding results obtained with the adoption of this approach in other EMI shields designed per the literature, we believe it will effectively serve as a vital tool in tuning the properties of graphene-based EMI shields for enhanced property performance.
## Spray deposition.
The spray deposition technique is extensively utilized for the deposition of thin-layer metallic layer(s) onto the surface of substrates such as electrospun NFs, where metal (nano)particles are evenly attached to the exterior portion of the electrospun NFs/substrate. Boundless impact on the EM wave shielding performance of electrospun-coated fibers consisting of diverse metal nanoparticles by this technique with a difference in its volume resistivity has been explored, though largely for composite materials with nanometals without the inclusion of graphenebased materials. 73, [bib_ref] Fabrication and EMI shielding effectiveness of Ag-decorated highly porous poly (vinyl alcohol)/Fe..., Kim [/bib_ref] Spray deposition could be also performed using the electrostatic approach through electrostatic atomization where a droplet (positively charged) via an applied high voltage to the metallic nozzle is sprayed upon the substrate; this approach is also referred to as electrostatic spray deposition (ESD). The technique is separated into three steps: the generation of an aerosol from the precursor solution, the transportation of this aerosol to the substrate's surface, and the generation of the deposit as a result of the droplets' impact on the substrates. The latter two phases are critical in the creation of the microstructure of the coating because the droplet size must be well controlled in relation to the ESD control factors. This technology has various advantages over the traditional deposition procedures, including a simple setup, inexpensive and nontoxic precursors, high deposition efficiency, direct deposition under ambient conditions, and easy control of the formed layers' surface shape.
The fabrication process for graphene/polyvinylphosphonic acid/cotton nanocomposites prepared using spray layer-bylayer assembly 10 coating cycles leading to 20 mg mL −1 graphene content resulting in enhanced electrical, thermal, and EMI shielding properties have been exploited.The utilization of spray deposition of rGO-based styreneethylene/butylene-styrene (SEBS) block copolymer coating leading to improvement in the SE to ∼19.3 dB has been established. [bib_ref] Enhancing the electromagnetic interference shielding of flexible films with reduced graphene oxide-based..., Godoy [/bib_ref] They postulated that the rGO-based coating resulted in attenuation of the EMWs, thereby enhancing the EMI SE to a large extent.
Another group of researchers examined the EMI shielding properties of hydrophobic, lightweight, and flexible graphenecoated fabric MXene−graphene−PVDF nanocomposite. [bib_ref] Electromagnetic shielding by MXene-graphene-PVDF composite with hydrophobic, lightweight and flexible graphene coated..., Raagulan [/bib_ref] These authors obtained a non-3D electrospun porous architecture of MXene/graphene foam exhibiting outstanding EMI SE of ∼53.8 dB (99.999%) having a SE R of ∼13.10 dB as well as SE A of ∼43.38 dB within the X band with the solo glazed carbon cloth, which demonstrated a very good absolute SE of 35,369.82 dB·cm 2 ·g −1 , and we hereby hypothetically postulate that an electrospun nanocomposite having this same formulation composition will give better properties even at lower overall material density. This technique is highly attractive for adoption in the fabrication of electrospun graphene-based nanocomposites for EM wave sheltering applications. It has been used for nonelectrospun systems in this regard with outstanding performance. [bib_ref] Electromagnetic shielding by MXene-graphene-PVDF composite with hydrophobic, lightweight and flexible graphene coated..., Raagulan [/bib_ref] 4.4. Vacuum Deposition. Other techniques such as vacuum deposition involve the use of a vacuum filtration approach to get the deposition targeted, where the force between the support material and the focus material sediment is employed to achieve the blend. Through vacuum deposition, a conductive layer is repetitively overlaid on the substrate surface to form an efficient shielding stage, while redundant EM waves can be dispelled within the substrate. We focus on the vacuum-filtration-based deposition method and its utilization within the discussed niche with its principle, as depicted in. The depiction clearly shows that we can regulate the number as well as propagation of the operating media further into microporous material by tweaking the vacuum pressure or operation duration. In reality, of course, the only drawback is the concentration of the combination as well as the permeability of the material. The operating medium is more closely bonded with the material when it is filtered into vacuum deposition increased deposition frequency requires more understanding and research; for example, the relationships between electron density, pulse length, encasing density, as well as microstructural parameters for high-power pulsed plasma deposition microstructure as well as properties in the thickness range of 10−10,000 nm, especially for low-dimensional structures, sub-microelectronics, nanocomposites, tribological coatings, corrosion-resistant materials, reflective surfaces, and thermoelectrics 83 diverse material properties could only be created using vacuum deposition techniques, such as coatings with properties independent of thermodynamic configurational restrictions due to the extreme versatility of the range and variety of deposited materials applied ability to vary defect concentration across a broad range, tends to result in characteristics equivalent to or far excluded from traditional bulk counterparts ion energy impact as well as ion flux on coated material characteristics is not fully understood amorphous materials can be deposited at a high quench rate depositing species energy effect on interfacial interaction, nucleation, as well as development of the deposit is not fully understood controllable generation of microstructures that differ from conventionally materials, such as ultrafine (superlattice, nanostructures) to single-crystalline films enables the manufacturing of fine free-standing shapes, from brittle materials, with environmental benefits there is also a lack of understanding on the mechanical characteristics of the deposit as well as their impact by control factors on residual stress sculpted thin films as well as highly porous membranes technique costlier than atmospheric deposition technique the pores rather than accumulated upon the exterior. Ultimately, upgrading the layout in the figure is indeed not challenging. This technique has been adopted for the fabrication of graphene-based EMI shields, though not electrospun nanocomposite materials with the display of interesting performance, 78 and can be adopted for fabricating electrospun graphene-based systems as EMI shields. The advantages and disadvantages of various deposition techniques are presented in .
At present, the presence of articles solely based on electrospun graphene-based EMI shield prepared with the use of the post-treatment approaches is limited. Therefore, we have also considered articles that have also used depositionbased post-treatment processes aimed at enhanced EMI shielding performance of diverse electrospun-fabricated shields. We have observed that the order is thus: Electroless deposition (EMI SE: 53−90 db) [bib_ref] Flexible and stretchable MXene/ Polyurethane fabrics with delicate wrinkle structure design for..., Zeng [/bib_ref] [bib_ref] Scalable fabrication of highly crosslinked conductive nanofibrous films and their applications in..., Lai [/bib_ref] [bib_ref] Lightweight, flexible and strong core-shell nonwoven fabrics covered by reduced graphene oxide..., Yuan [/bib_ref] This review aims to also to ignite the interest of researchers within the niche of EMI shielding materials toward utilization of these innovative approaches.
## Factors affecting emi sheltering performance of graphene-based electrospun materials
Graphene-based electrospun nanocomposite fibers possess the benefits of flexibility, foldability, and/or low density and are deemed as one of the most favorable EM wave shielding materials. Unquestionably, there are additionally many components influencing the EM wave shielding properties of these electrospun fibrous materials, such as the thickness of the membranes consisting of electrospun fibers, its diameter, porosity, conductivity, and magnetism, along with the fiber interactions interfacially. The amalgamation of these aforementioned influences controls the EMI shielding performance of the consequent spun fibrous mat.
## Thickness.
Augmenting the width of the graphenebased spun fibrous nanocomposites as EMI shields is an effectual approach to enhancing the reflection as well as absorption operational property of the EMI shield. At a specific frequency, EM waves would be dissipated as they reaches a particular depth inside the fibrous architectural structure. An electrospun fibrous structure having a thicker layer is more capable of shielding EM waves to a higher extent. To this end, Yuan et al. comparatively studied the EMI SE of Fe 3 O 4 @ SiO 2 @polypyrrole@GO composite fibers with regard to its diverse width by assembling various pieces together and revealed that EM waves transit via thicker sandwiched structures, resulting in more loss of EM waves. [bib_ref] Lightweight, flexible and strong core-shell nonwoven fabrics covered by reduced graphene oxide..., Yuan [/bib_ref] These authors established in their study that the material thickness was directly proportional to the EMI SE and vice versa.
Based on the above discussion, we do therefore postulate that proper management of the graphene-based electrospun materials thickness up to the threshold as per the material in question will go a long way in helping researchers and industrialists to fabricate innovative EMI shields.
## Porosity coupled with fiber diameter.
It has been established that the major sheltering phenomenon of graphene-based electrospun nanocomposites in the EMI shielding mechanism is absorption, even though reflection also plays a part in EM wave shielding. EM waves are reflected on the fiber sheath surface; the residual EM waves would transit through the sheaths and are consequently consumed via interaction with the fibers. [bib_ref] Lightweight, flexible and strong core-shell nonwoven fabrics covered by reduced graphene oxide..., Yuan [/bib_ref] Narrower diameters of these fibers can result in improved interfacial interaction between the pores/voids and spun fibers, in the interim, as the porosity (voids) gets reduced. This factor "porosity" of the electrospun fibers also influences the EM waves' dissipation within its 3D structure. [bib_ref] Electrospun graphene nanosheet-filled poly (trimethylene terephthalate) composite fibers: Effects of the graphene..., Huang [/bib_ref] Studies involving the correlation between pore sizes of electrospun graphene-based nanocomposites/hybrids have been rare/limited at present. However, Nasouri et al. [bib_ref] Effects of diameter and surface area of electrospun nanocomposite fibers on electromagnetic..., Nasouri [/bib_ref] explored the relationships between the diameter of spun "MWCNTs/PVP" (we know that MWCNTs are composed of material properties similar to those of graphene) fibers regarding their ability to shield EMWs. Electrospun fibers having distinct diameters were prepared by altering the ES voltage, the aloofness between the drum collector and tip of the needle, and electrospun solution concentration. They established that the thicker fibers could be electrospun at higher solution concentration, hence boosting the porosity of the spun membranes, but with the EC revealing a declining trend. [bib_ref] Effects of diameter and surface area of electrospun nanocomposite fibers on electromagnetic..., Nasouri [/bib_ref] They attributed this effect to the differences in the surface area and aperture sizes, which resulted in variations in the passage of the flow of electricity. Moreover, higher total EMI SE was achieved with increased surface area. They concluded that smaller diameter fibers imply higher surface area, leading to higher heat losses and vice versa. Chu et al. [bib_ref] Effects of diameter and hollow structure on the microwave absorption properties of..., Chu [/bib_ref] in their study showed that the dielectric constant of their studied samples increased with decreased fiber diameters (i.e., by adjusting the fiber diameter, the porosity also changed with other properties), indicating that a thinner width will result in more dielectric loss, in order to trigger microwave loss. [bib_ref] Electrospun graphene nanosheet-filled poly (trimethylene terephthalate) composite fibers: Effects of the graphene..., Huang [/bib_ref] With regard to porosity, it is worth noting that the nanofibrous nonwoven perviousness, thought as the volume liberated from polymeric matrix over the all-out volume, can run somewhere in the range of 80 and 90% in a normal nanofibrous mat. This same phenomenon applies to graphenebased electrospun systems aimed at application in EMI shielding.
In graphene-based electrospun systems, there are few reports considering the influence of porosity on the EMI SE and/or the other parameters: higher porosity is directly proportional to the EM wave absorption as reported the literature. [bib_ref] Effects of diameter and surface area of electrospun nanocomposite fibers on electromagnetic..., Nasouri [/bib_ref] [bib_ref] Electrical conductivity of electrospun nanofiber mats of polyamide 6/polyaniline coated with nitrogendoped..., Santos [/bib_ref] 5.3. Magnetism and Conductivity. The use of conductive/magnetic fillers can improve the electrospun strands' overall conductivity and magnetic characteristics, as well as the EMI SE of the filaments. A higher conductivity indicates that the strands contain all of the freer electrons. EM waves will be reflected when they arrive at the outer layer of the material; in the interim, the higher porousness owing to the highly porous architecture of electrospun fibers containing active fillers having good conductivity and magnetic properties results in an advanced magnetization, enhancing the magnetic loss along with the absorption of the EM waves. The amalgamation of two or more kinds of fillers has multiple advantages for the performance of EMI shields.
Yuan et al. [bib_ref] Lightweight, flexible and strong core-shell nonwoven fabrics covered by reduced graphene oxide..., Yuan [/bib_ref] [bib_ref] Integration of MCMBs/MWCNTs with Fe 3 O 4 in a flexible and..., Chaudhary [/bib_ref] [bib_ref] Fabrication and EMI shielding effectiveness of Ag-decorated highly porous poly (vinyl alcohol)/Fe..., Kim [/bib_ref] With the effective and efficient amalgamation of these features with the other ones, researchers and industrialists will be able to design and engineer innovative EMI shields for even state-of-the-art applications.
## Interfacial interaction.
Interfacial interaction between the filler−matrix, filler−filler, filler−other active additives within an electrospun graphene-based EMI shielding material is of great importance, as it directly influences the overall material properties leading to the attainable EMI SE. [bib_ref] Kinetically controlled localization of carbon nanotubes in polylactide/poly (vinylidene fluoride) blend nanocomposites..., Guo [/bib_ref] [bib_ref] Electrospun graphene nanosheet-filled poly (trimethylene terephthalate) composite fibers: Effects of the graphene..., Huang [/bib_ref] Extreme interfaces can result in interactions between EMWs and shields, in this manner understanding the specific scattering of EM waves. Electrospun fibers present recompenses such as high porosity, high surface area, etc. However, its high surface construction will, in general, deliver various interfaces among strands. It has been shown that the outer layer of conductive strands brimming with moving carriers can cause partial heat losses to realize the loss of EM waves. Nonetheless, before EMWs are completely absorbed, a portion of them may reflect back and forth within the fiber due to the action of the fiber interface. 63 Because the porous graphenebased electrospun fiber has a low impedance, EM waves can strike the fibers, generating reflection and dissipating between the strands' edges, improving interfacial interactions. Furthermore, the stage and fiber in question exhibit interfacial contacts, which improve attack wave polarization loss.
## Challenges
With the interfacial interaction between EM waves and graphene-based electrospun fibrous materials, more attention should be given to the correlation between SE and the composite/hybrid materials compact structure. Moreover, the reinforcing active filler selection for the novel nanocomposite consisting of electrospun fibers also requires further explora-tional studies. Selected facets could be considered to resolve the aforementioned challenges in the forthcoming research on graphene-based electrospun fibrous materials intended to be utilized as EMI shields.
## Enhanced unique interfacial interactions (skin-effect) engineering/fabrication.
Graphene-based electrospun fibrous materials having multicomponent structures have a propensity to present numerous interfaces, thereby causing more loss of EM waves. The fabrication of multicomponent structural architecture of these composites, such as coaxial, side-by-side, and/or triaxial electrospun fibers as well as multicomponent fiber-based material architecture, will promote the shielding of EM waves by these fibrous nanocomposites. Also, the design of diverse modules within the structure is also commendable of consideration. For example, concurrent inclusion of a conductive/magnetic phase into the different structural parts could give rise to the diversity of fiber functions, hence boosting EM wave attenuation. Consequently, researching the correlation between diverse structural facets and EMI shielding performance of graphene-based EMI shields is also crucial for extra research. After these missions are investigated, ES will turn out to be a more probable technique for designing and engineering EMI shields.
## Engineering of smaller width, higher porosity electrospun fibers.
Smaller width/diameter electrospun fibers having higher specific surface area possess the ability to cause higher EM waves loss, whereas dense pore structure could result in more internal interaction between EM waves and the fibers. Conversely, the width/diameter of the fibers and their porosity are intimately correlated to/with the electrospun fibrous nanocomposite thickness. Densification of the electrospun fibrous membrane pores successfully results in a direct decrease in the infiltration of EMWs into the EMI shield. Dense structured electrospun fibrous nanocomposites in this way (based on pore structure) can be engineered by post-treatment approaches such as hot pressing technology, vacuum suction filtration, and the like.
## Novel active reinforcing (nano)filler utilization.
The utilization of novel reinforcing (nano)fillers, such as nitrides, transition metal carbides, or carbonitride(s) having excellent EC in graphene-based electrospun fibrous nanocomposites toward application in EM wave shielding, is highly encouraged. For instance, 2D layered materials such as MXenes (low density, high surface area, outstanding EC, and straightforward processing) have drawn worldwide attention from researchers and industrialists for the design/engineering of efficient EMI shields. [bib_ref] Electromagnetic property for microwave absorption and electromagnetic interference shielding, Cao [/bib_ref] Nevertheless, associated research investigations on MXenes as active reinforcing fillers in electrospun graphene-based EMI shields are rare or not reported. Hence, it will be a potential approach to integrate unique active reinforcing fillers in graphene-based spun NFs for EMI shielding materials. The challenge of reinforcing filler agglomeration in prepared solution for electrospinning that can significantly impact the ES technique along with further limiting of the large-scale fabrication of these composites requires attention. The challenge arises from poor filler dispersion/sedimentation due to a lack of proper interaction in situ between the included fillers and solution(s) and lack of sufficient electrostatic repulsion to enable the fillers to disperse more uniformly within the prepared solution. The inclusion of dispersing agent(s) or modification of the fillers through suitable addition of active surface functional groups is an effective technique toward solving the problems.
6.4. 3D Electrospun Graphene-Based Fibrous Structure. Three-dimensional electrospun graphene-based fibrous structured aerogels or sponges having highly porous architecture accelerate the entrance of EM waves into the structural network for quick dissipation, hence unveiling exceptional EM wave sheltering. [bib_ref] Superior electromagnetic interference shielding 3D graphene nanoplatelets/reduced graphene oxide foam/epoxy nanocomposites with..., Liang [/bib_ref] Though for a long time it was challenging to directly produce 3D aerogels or sponges having mechanical stable structure(s) via ES, recently, successful formation of aerogels or sponges through a mechanical cutting in solvents followed by freeze-drying of electrospun fibers has been explored. [bib_ref] Low-density open cellular sponges as functional materials, Jiang [/bib_ref] These fibrous aerogels/ sponges reportedly exhibited a lot of advantages, such as high explicit surface region, high porosity, abundant interface, low density, controllable functionality, and so on, favoring exceptional EMI shielding, along with anisotropic porous architecture vital for the EMI shielding achieved by steering freeze-drying of short electrospun fibers, signifying smart shielding which can also be realized in graphene-based spun nanocomposite shields through innovative formulations in the near future.
Up until now, ES technology has accomplished great advancement in diverse niches, though research in ES of graphene-based fibrous materials as EMI shields is rare and still in the development phase, which calls for extra efforts for additional practical applications. For the moment, there still exist some problems in the way of commercialization of these materials as EMI shields. Certain challenges and resolutions are offered.
One chief factor limiting the commercialization of electrospun graphene-based electrospun fibrous materials as EMI shields is their technical properties, notably their springiness. Graphene-based electrospun fibrous materials for EMI shielding made of rigid carbon NFs exhibit great restrain in applications in real-life. To resolve this challenge, these electrospun fibrous materials containing CNFs should be processed with softer matrices to enhance the stiffness of the EMI shields. Stretchable soft elastomers such as PDMS (polydimethylsiloxane) or others can be utilized as the matrix/comatrix in the preparation of these kinds of EMI shields.
Graphene-based electrospun EMI shields surface coating with organic/inorganic materials having exceptional resistance to the harsh environment (strong acid, high/low temperature, and/or strong base, etc.) will be an attractive viable solution to the challenge of using graphene-based electrospun EMI shields in harsh environments which is a challenge at the moment.
With a focus on commercialization, the cost of production is a worthy factor. The selection of appropriate materials and technological process optimization can greatly lower the production cost of these (nano)composite materials. For instance, modification on graphene-based electrospun composite NFs with the inclusion of bio-based materials is a good option. Further considerations should be given on searching innovative graphene-based electrospun fibrous EMI shields, with regard to the modern progress in radar technology which requires innovative EMI shields for broad-band frequency of 8−26.5 GHz in the upcoming state.
The inclusion and/or total use of functional electrospun nanofibrous, especially graphene-based materials as EMI shields in microsize electronic devices, is among the foremost target recently (as presented in [fig_ref] Table 2: Performances of Graphene-Based Electrospun Nanofibrous EMI Shields per the Literature graphene/graphite-based EMI... [/fig_ref]. High-throughput electrospinning units are also expected to open additional opportunities for graphene-based electrospun nonwoven matierals. For these purposes, the latest innovations, such as core−shelled ES, mixing and compound ES, coaxial ES, blowassisted ES, etc., have been postulated.
# Conclusions and perspectives
7.1. Conclusions. As discussed throughout the paper, nowadays the call for EM wave sheltering fabrics having the desired SE along with acceptable architectural makeup has become necessary: the design and production of innovative flexible, lightweight, and robust materials having outstanding EMI sheltering characteristics is now an imperative mission. A novel material such as 1D graphene-based electrospun fibers presents acceptable abilities in shielding EM waves, and the consolidation of nanofillers creating magnetic or dielectric dipoles could enhance microwave shielding. Posttreatment of electrospun fibrous materials to composites opens up a slew of new possibilities for creating high-performance EMI shielding materials with additional functions that are useful in a variety of applications. The compensation of materials having high specific surface area along with high aspect ratio of graphenebased spun NFs is that it makes EMI shields achieve exceptional flexibility and good porosity, along with the promotion of EM wave interactions between individual strands, accordingly, improving the ingestion of EMWs and viably acknowledging EMI protection. The presentation of the ceramic phase could further develop the impedance mismatch between the NFs. Although unadulterated carbon materials such as graphene electrospun NFs display high EMI SE, these materials' poor adaptability, as well as somewhat low dielectric character, restrains their additional utilization. The inclusion of conductive as well as magnetic nanofillers would be a viable procedure to enhance the property performance of these electrospun shields against EMI. The inclusion of conductive reinforcing (nano)fillers both enhances the dielectric performance and SE T owing to the dielectric loss, along with enhancement of the mechanical characteristics of the electrospun nanocomposite. Moreover, the amalgamation of magnetic phases could create a magnetic loss in the case of electrospun fibers, thereby enhancing the SE T while post-treatment of graphene-based electrospun fibrous materials further augment the EMI shield performance. Conventional post-treatment techniques such as spray deposition, electroless deposition, dip deposition, as well as vacuum deposition can be helpful in enhancing the effective influence on graphene-based electrospun fibrous materials as EMI shields. The EC, thickness, and magnetic properties of graphene-based EM wave shields are the vital factors influencing their EMI shielding characteristic as novel (nano)composites.
## Perspectives.
It is important to note that properties such as excellent EC, TC, in conjunction with thermal insulation are very vital and should be harnessed effectively to achieve effective EM waves shielding.
Processing techniques such as layer-by-layer have been utilized for the fabrication of graphene oxide-based nanocomposites having EC, TC, as well as thermal insulation: 97 this is then an assurance that with synergistic preparation of graphene-based hybrid materials via amalgamation of ordered electrospun layers as well as cast layers, researchers will be able to fabricate better state-of-the-art effective and efficient EMI shields for even emerging 5G signal.
With the aim of fabricating EMI shields having both excellent EMI SE as well as superior flame self-extinguishing character, carbon-based materials were endlessly dispersed in particular layered space by a group of researchers recently to construct interchanging multilayered nanocomposite at a low nanoparticle concentration, where PBS-CNT and TPU-IFR-CNT layers conversely aligned along the extrusion path. [bib_ref] Strategy for constructing electromagnetic interference shielding and flame retarding synergistic network in..., He [/bib_ref] The EMI shielding as well as flame retarding synergistic system could well be built in 8-and 16-layer nanocomposite by adjusting the number of layers. As a result of the impedance mismatch between adjacent layers, the average EMI SE result measured in the X band (8.2−12.4 GHz) exceeded 30 dB, far exceeding the commercial criterion of 20 dB. [bib_ref] Strategy for constructing electromagnetic interference shielding and flame retarding synergistic network in..., He [/bib_ref] The approach has been seconded/supported by other researchers work even with the use of different polymeric matrices. [bib_ref] Flexible and flame-retarding thermoplastic polyurethane-based electromagnetic interference shielding composites, Ji [/bib_ref] We foresee that with the amalgamation of the innovative approaches used/postulated/proven by diverse researchers with electrospinning for the fabrication of graphene-based EMI shields, robust and effective EMI shields will be fabricated by researchers which will meet the current demand of the day.
[fig] Figure 1: (a) Overview of graphene-based electrospun fibrous materials as EMI shields, with different corresponding influencing factors.(b) Schematic presentation of EM wave sheltering mechanism of graphene-based polymeric nanocomposites/fibers as well as (c) electromagnetic spectrum wavelength and frequency. [/fig]
[fig] Figure 2: Schematic presentation of the electrospinning technique. [/fig]
[fig] Figure 3: (a) Graphical presentation of electromagnetic wave and (b) A segment of traveling EM wave with in-phase mutually orthogonal E and B. Reproduced with permission from ref 30. Copyright 2022 Elsevier Science Ltd. [/fig]
[fig] Figure 4: Morphology and element analysis (a−c) of the hybrid nanowires. [/fig]
[fig] Figure 5: (a) Schematic presentation of the preparation of flexible networks and CNF−GN−CNF heterojunctions via ES. SEM images of virgin CNF networks (b) and GN/CNF composite networks with 17.2 and 31.9 wt % GN (c,d). (e) EC of the samples with different GN loadings. (f) Absolute SE of the GN/CNF nanocomposite networks with 17.2 wt % GN calculated based on its performance in virgin CNF networks. Reproduced from ref 57. Copyright 2014 American Chemical Society. [/fig]
[fig] Figure 6: Thermal conductivities of the (Ag/rGO)/PI nanocomposites. (a) Mass fraction of Ag/rGO and rGO influencing the λ values of the (Ag/ rGO)/PI and rGO/PI nanocomposites, respectively. (b) Infrared thermal images of (Ag/rGO)/PI nanocomposites. Reproduced from ref 60. Copyright 2019 American Chemical Society. [/fig]
[fig] Figure 7: Scheme showing diverse deposition approaches. [/fig]
[fig] Figure 8: Schematic presentation of the dip-coating method. [/fig]
[fig] Figure 9: Schematic illustration of equipment used in electroless coating studies depicted in its most basic form. [/fig]
[table] Table 2: Performances of Graphene-Based Electrospun Nanofibrous EMI Shields per the Literature graphene/graphite-based EMI shield cross-sectional thickness (mm) [/table]
|
Coexistent hairy cell leukaemia and hepatosplenic t-cell lymphoma: a case report
Background: Hairy cell leukaemia (HCL) is a chronic B-cell leukaemia characterized by expansion of neoplastic cells in the spleen, bone marrow and blood. Symptoms of HCL are related to pancytopenia and immune deficiency. Patients with HCL have an increased risk of second malignancy either in a form of synchronous disease or in a form of an increased incidence of a second neoplasm after the treatment of HCL. Hepatosplenic T-cell lymphoma (HSTCL) is a rare form of aggressive extranodal T-cell lymphoma. Its pathogenesis is connected to a chronic immune deficiency status and its coexistence with other neoplasms is practically non-existent.Case: We present a case of a 53-year-old female patient suffering from hepatosplenomegaly, peripheral lymphadenopathy and related B symptoms. An excisional biopsy of the enlarged axillary lymph node revealed partial infiltration with CD3+/CD56+/TIA + T cell lymphoma. Bone marrow trephine biopsy and flow cytometric immunophenotypization of bone marrow cells and peripheral blood showed presence of two types of neoplastic cells in the peripheral blood and in the bone marrow (composite lymphoma). One of them showed typical morphologic characteristics and immunohistochemical features of HCL, while another one was morphologically and immunophenotypically consistent with the diagnosis of HSTCL, respectively. The patient was treated with multivalent chemotherapy including rituximab but all treatments turned out to be only partially effective. While HCL responded to the treatment, HSTCL was refractory to the chemotherapy and the patient died 7 months after the initial diagnosis because of haematemesis induced by Mallory-Weiss syndrome. Conclusion: This is the first recorded case of coexistent HCL and HSTCL in the same patient. A multidisciplinary approach, encompassing careful morphology interpretation, immunophenotypic, cytogenetic and molecular analyses, is mandatory to obtain an accurate diagnosis of composite lymphoma.Virtual slides: The virtual slides for this article can be found here: http://www.diagnosticpathology.diagnomx. eu/vs/9354870531161685.
# Background
Hairy cell leukaemia (HCL) is a chronic B-cell leukaemia characterised by pancytopenia and clonal expansion of mature malignant B cells in the spleen, bone marrow and blood [bib_ref] Hepatosplenic T cell lymphoma, hairy cell leukemia, Jaffe [/bib_ref]. Characteristically, patients with HCL do not suffer from leukaemia itself but from symptoms that are related to pancytopenia and opportunistic infections [bib_ref] Hairy cell leukemia: biology, clinical diagnosis, unusual manifestations and associated disorders, Polliack [/bib_ref]. Hepatosplenic T-cell lymphoma (HSTCL) is a rare form of extranodal T-cell lymphoma derived from cytotoxic T cells. In up to 20% of cases, its pathogenesis is connected to a chronic immune deficiency status, such as prolonged immunosuppressive therapy after organ transplantation, long-term combined therapy with thiopurines and anti-TNF agents for chronic inflammatory bowel diseases or prolonged antigen stimulation [bib_ref] Hepatosplenic T cell lymphoma, hairy cell leukemia, Jaffe [/bib_ref] [bib_ref] A systematic review of factors that contribute to hepatosplenic T cell lymphoma..., Kotlyar [/bib_ref]. Both diseases demonstrate propensity to infiltrate the spleen, liver and bone marrow. Their coexistence has to be exceptional because we have found no reports about concurrent occurrence of HCL and HSTCL in the English literature by now. The objective of this study is to report the clinical and pathological features of a 53-year-old female patient with coexistent HCL and HSTCL who was recently treated at our institution.
## Case report
A 53-year-old female was referred to our centre because of a two-month lasting exhaustion, abdominal distension, early satiety and fever of up to 39°C. Her medical history was not significant. She had been travelling for a longer periods of time in the countries of the Middle and Far East, especially to Indochina and India. Physical examination showed splenomegaly and hepatomegaly, palpable at 15 and 6 cm below the costal margin, respectively. Peripheral lymph nodes measured up to 1 cm in the largest diameter. Petechiae and suffusions were present on the skin of the thoracic region and flanks. Laboratory tests demonstrated: (1) white blood count 3.40 × 10 9 /l (Neu 2,01 (59%) , Ly 0,99 (29%), Mono 0,17 (5%), Eos 0,03 (1%)), platelet count 64 × 10 9 / l, red blood cells 3,66 × 10 9 /l, haemoglobin level 115 g/l (2) liver function tests were raised: AST 2.21 IU/L, ALT 0.95 IU/L and γGT 1.30 IU/L with bilirubin within normal ranges-bilirubin 17 IU/L (3) LDH was raised to 15,68 IU/L; raised was also CRP to 35 (4) serology for HAV showed reactive IgG anti-HAV antibodies, whereas IgM anti-HAV, HbsAg, anti-HBc, anti-Hbs, anti-HCV and CMV were negative.
## Cytologic findings
The Giemsa-stained smears from bone marrow and peripheral blood showed presence of medium-sized atypical lymphoid cells. These cells slightly differed in size and shape, being round, oval or droplet-like with irregular cell borders. The cytoplasm was scant and basophilic, containing few small vacuoles. The nuclei were round or oval. Chromatin was immature, finely granular, with small inconspicuous nucleoli [fig_ref] Figure 1 A: -D [/fig_ref] , B, D). On flow cytometric immunophenotyping (FCI) of bone marrow and peripheral blood, these cells had immunophenotypic profile consistent with gamma-delta T-cell lymphoma (CD3 dim +, CD7 dim+, CD2+, TCRγδ dim +, CD52 dim +, CD56+, CD5-, CD4-and CD8-) [fig_ref] Figure 2: Red rectangle symbol [/fig_ref].
Besides T-cell lymphoma, a small population of monoclonal B-cells with immunophenotype characteristic of HCL (CD103+, CD25+, CD22+, surface-kappa+) [fig_ref] Figure 2: Red rectangle symbol [/fig_ref] was also present in the bone marrow and peripheral blood. After careful re-examination of peripheral blood smears, only few hairy cells were found. These cells were medium-sized and oval in shape, their cytoplasm was abundant, pale basophilic, with circumferential "hairy" projections [fig_ref] Figure 1 A: -D [/fig_ref].
## Pathologic findings
Biopsy of a slightly enlarged left inguinal lymph node and bone marrow trephine biopsy were also performed. Histology of the lymph node showed partial, predominantly intrasinusoidal infiltration by medium-sized, atypical lymphoid cells that were immunohistochemically positive for CD3, CD56, and TIA, while other Tand B-cell markers were negative. Bone marrow biopsy (Figures 3,4,5 and 6) revealed hypercellular bone marrow with reduced orthotopic haematopoiesis and intrasinusoidal infiltrates consisting of medium-sized lymphocytes with the same immunohistochemical profile as was observed on lymph node infiltrates. To continue, patchy interstitial infiltrates were also present, consisting of medium-sized lymphocytes with a round nuclear profile and fine chromatin, and they were widely separated from each other by abundant clear cytoplasm. These cells were positive for CD20, HCL, Cyclin D1 and TRAP [fig_ref] Figure 3: Bone marrow biopsy [/fig_ref]. There was also a marked increase in reticulin fibres. According to the morphology, the immunophenotype and the infiltration pattern of lymphoid infiltrates, the diagnosis of coexistent HSTCL and HCL was established.
## Follow up
The patient was treated with multivalent chemotherapy including rituximab. The effect of treatment was monitored by clinical and ultrasound examination, leukocyte and lymphocyte counts and FCI of peripheral blood. The patient received the first immunochemotherapy regimen with R-VACPE (rituximab-vincristine, doxorubicin, cyclophosphamide, prednisone, etoposide) and splenectomy [fig_ref] Figure 7: Spleen: H&E, 20×. [/fig_ref]. After that HCL regressed completely while HSCTL was refractory to the treatment. R-VACPE was therefore changed to R-DAHP (rituximab, dexamethasone, cytarabine, cisplatin) and then to R-EPOCH (rituximab-etoposide, vincristine, doxorubicin, cyclophosphamide, prednisone). However, all these treatment regiments turned out to be only partially effective. The disease progressed again in the liver, and the patient's general condition worsened. Therefore, only palliative treatment with liver irradiation was continued. The patient died soon because of haematemesis and Mallory-Weiss syndrome, 7 months after the initial diagnosis.
# Discussion
HCL is an uncommon form of chronic lymphoproliferative disease, characterised by indolent course. Hairy cells typically home to the spleen and circulate only late in the disease. Their proliferation is regulated by different cytokines and growth factors in an autocrine and paracrine fashion [bib_ref] Autocrine and paracrine regulation of neoplastic cell growth in hairy cell leukemia, Schmid [/bib_ref]. Many manifestations seen in the peripheral blood in patients with HCL are the result of interactions between the neoplastic B cells and the Tcell system. Among others, there are profound alterations in the T-cell immunity with the production of a skewed T-cell repertoire [bib_ref] Persistent clonal excess and skewed T-cell repertoire in T cells from patients..., Kluin-Nelemans [/bib_ref] , production of autoreactive T cells, and an increase in the CD3 + γδ cells as a persistent clonal excess, as well as monocytopenia, leucopenia, and so on [bib_ref] Severe decrease in peripheral blood dendritic cells in hairy cell leukemia, Bourguin-Plonquet [/bib_ref]. All of the above-mentioned factors contribute significantly to the immune deficiency found in HCL patients and to the extreme sensitivity to opportunistic infections. Since all of the above-listed factors play important roles also in the surveillance against neoplasia, it is likely that the appearance of the second neoplasms in HCL reflects an underlying immune deficiency in such patients rather than random coexistence of the two diseases. Some studies showed that patients with HCL have an increased risk of second malignancy, either in a form of synchronous disease [bib_ref] Second cancer risk in hairy cell leukemia: analysis of 350 patients, Kurzrock [/bib_ref] [bib_ref] Synchronous detection of hairy cell leukemia and HIV negative Kaposi's sarcoma of..., Aydin [/bib_ref] [bib_ref] Neoplasm associated with hairy cell leukemia, Laurent [/bib_ref] [bib_ref] Evidence for an association between hairy cell leukemia and renal cell carcinoma, Nielsen [/bib_ref] or in a form of an increased incidence of a second neoplasm after the treatment of HCL [bib_ref] Second cancer incidence and cause-specific mortality among 3104 patients with hairy cell..., Hisada [/bib_ref] [bib_ref] Increased incidence of second neoplasms in patients treated with interferon alpha 2b..., Kampmeier [/bib_ref] [bib_ref] Second malignancies in patients with hairy cell leukemia in British Columbia: a..., Au [/bib_ref]. Among malignancies with synchronous occurrence, the most frequent is the occurrence with other haematopoietic malignancies [bib_ref] Multiple myeloma and hairy cell leukemia: a rare association or coincidence?, Saif [/bib_ref] [bib_ref] Coexistent hairy cell leukemia and chronic lymphocytic leukemia, Brown [/bib_ref] [bib_ref] Simultaneous manifestation of chronic lymphocytic leukemia (CLL) and hairy cell leukemia (HCL), Sokol [/bib_ref] [bib_ref] Simultaneous diagnosis of hairy cell leukemia and chronic lymphocytic leukemia/small lymphocytic lymphoma:..., Giné [/bib_ref] [bib_ref] Coexistence of chronic myeloid leukemia and hairy cell leukemia of common clonal..., Pajor [/bib_ref] [bib_ref] The simultaneous presentation of peripheral T-cell lymphoma and hairy cell leukemia, Lawlor [/bib_ref] [bib_ref] Hairy cell leukemia associated with large granular lymphocyte leukemia: immunologic and genomic..., Marolleau [/bib_ref] [bib_ref] Simultaneous occurrence of hairy cell leukemia and Hodgkin's disease in the same..., Resegotti [/bib_ref] , B-cell lymphomas in the first place, and very rarely also with T-cell neoplasms. Contemporary literature provides only rare case reports describing the coexistence of HCL with LGL leukaemia [bib_ref] Hairy cell leukemia associated with large granular lymphocyte leukemia: immunologic and genomic..., Marolleau [/bib_ref] and simultaneous presentation of HCL and peripheral T-cell lymphoma [bib_ref] The simultaneous presentation of peripheral T-cell lymphoma and hairy cell leukemia, Lawlor [/bib_ref].
On the other hand, HSTCL is a rare and very aggressive form of extranodal lymphoma with a median survival of less than 2 years, despite therapy [bib_ref] Hepatosplenic T cell lymphoma, hairy cell leukemia, Jaffe [/bib_ref]. Its coexistence with other haematologic malignancies is practically non-existent, which could be a result of its aggressive nature and the usually quick lethal outcome. In the recent literature, there are only two case reports of HSTCL that presented together with some other type of haematologic disease. One is a case of coexistence with aggressive CD56+ CD3-TcRγδ-leukaemia [bib_ref] Coexistence of two distinct cell populations (CD56 + TcRγδ + and CD56+..., Camera [/bib_ref] , and the other is HSTCL in conjunction with systemic mastocytosis [bib_ref] Hepatosplenic gamma-delta T cell lymphoma disseminated intravascular coagulation and systemic mastocytosis: an..., Perini [/bib_ref].
However, coexistence of two or more morphologically and immunophenotypically distinct lymphoma clones in a single anatomic site is nowadays defined as composite lymphoma [bib_ref] Review article: composite lymphoma, Mokhtar [/bib_ref]. The incidence of composite lymphoma has been reported to vary between 1% and 4.7% [bib_ref] Composite lymphoma and related disorders, Kim [/bib_ref]. Many different combinations of composite lymphomas have been described [bib_ref] Review article: composite lymphoma, Mokhtar [/bib_ref]. The diagnosis of composite Tcell and B-cell lymphomas is particularly challenging because the neoplastic T-cells may outnumber the neoplastic B-cells, occasionally creating a false impression that B-cells are reactive. Similarly, the neoplastic B-cells may also outnumber the neoplastic T-cells, which may result in overlooking the T-cell lymphoma. A multidisciplinary approach, encompassing immunophenotypic, cytogenetic and molecular analyses, is therefore mandatory to obtain an accurate diagnosis. FCI can be very helpful in these settings, because it can detect small populations of neoplastic B-cells or T-cells, which could be overlooked in morphological and immunohistochemical analyses [bib_ref] Usefulness of multiparametric flow cytometry in detecting composite LymphomaStudy of 17 cases..., Demurtus [/bib_ref].
Any lymphoproliferative disease infiltrating spleen and liver can mimic HSTCL or HCL. Detailed study of morphology, together with immunophenotyping, gene rearrangement analyses, determining of EBV status and cytogenetic studies could help to establish the right diagnosis [bib_ref] Aggressive natural-killer leukemia with jaundice and spontaneous splenic rupture: a case report..., Gao [/bib_ref] [bib_ref] Identification of genuine primary pulmonary NK cell lymphoma via clinicopathologic observation and..., Gong [/bib_ref] [bib_ref] Aberrant antigenic expression in extranodal NK/T-cell lymphoma: a multiparameter study from Thailand, Pongpruttipan [/bib_ref] [bib_ref] Concurrent acute myeloid leukemia and T lymphoblastic lymphoma in a patient with..., Chang [/bib_ref] [bib_ref] Angiolymphoid hyperplasia with eosinophilia developing in a patient with history of peripheral..., Gonzalez-Cuyar [/bib_ref] [bib_ref] Peripheral T-cell lymphoma with cycline D1 overexpression: a case report, Aquino [/bib_ref]. Two diseases that show great overlapping with HSTCL regarding clinical picture as well as immunophenotype are aggressive NK-cell leukaemia and advanced extranodal NK/T-cell lymphoma, nasal type. Although immunophenotype could overlap significantly, rearranged TCR gene and no relation to EBV infection are distinguishing features that speak in the favour of the HSTCL [bib_ref] Aggressive natural-killer leukemia with jaundice and spontaneous splenic rupture: a case report..., Gao [/bib_ref] [bib_ref] Identification of genuine primary pulmonary NK cell lymphoma via clinicopathologic observation and..., Gong [/bib_ref] [bib_ref] Aberrant antigenic expression in extranodal NK/T-cell lymphoma: a multiparameter study from Thailand, Pongpruttipan [/bib_ref]. HCL can be distinguish from other B-cell neoplasms by its morphology and characteristic immunophenotype, while presence of B symptoms, organomegaly and signs of infection exclude the diagnosis of monoclonal B lymphocytosis with HCL-like phenotype in our patient [bib_ref] Diagnostic criteria for monoclonal B-cell lymphocytosis, Marti [/bib_ref].
In an attempt to explain the coexistence of HCL and HSTCL in our patient, some hypotheses can be formulated. Simultaneous appearance of HCL and HSTCL in our patient could be only fortuitous. However, since HSTCL often arises in the setting of chronic immune suppression, and HCL is a malignancy with a proven profound immune deficiency, their random coexistence is less likely. If persistent clonal excess of T-lymphocytes and skewed T-cell repertoires found in HCL are the only factors that could influence the course of T-cell oncogenesis in HCL patients, one could expect a much higher incidence of synchronous occurrence of HCL and T-cell lymphoma.
# Conclusions
In conclusion, hereby, we present the first case of coexistent HCL and HSTL that has been, to the best of our knowledge, described in the English literature so far.
## Consent
Written informed consent was obtained from the patient's parent for the publication of this report and any accompanying images.
[fig] Figure 1 A: -D. Cytological features of composite HSTCL and HCL. A: HSTCL cells in bone marrow aspirate, MGG, 60×. B: HSTCL cells in peripheral blood smear, MGG, 60×. C: HCL cell in peripheral blood, MGG, 60×. D: HSTCL cells in FNAB lymph node sample, Giemsa, 60×. [/fig]
[fig] Figure 2: Red rectangle symbol)…HSTCL, (Blue rectangle symbol)…HCL, (Dark green rectangle symbol)…reactive T-cells, (Light green rectangle symbol)…reactive B-cells. FCI findings of composite HSTCL and HCL in bone marrow aspirate. A: CD3 dim positive population of HSTCL. B: γδ T-cell receptor dim positive population of HSTCL. C: CD103 and CD22 positive population of HCL. D: Monoclonal, kappa positive population of HCL. [/fig]
[fig] Figure 3: Bone marrow biopsy: H&E, 20×. Interstitial infitrates of HCL cells and distended sinusoides containing HSTCL cells (arrows). [/fig]
[fig] Figure 4: Bone marrow biopsy: CD20 IHC, 20×. [/fig]
[fig] Figure 5: Bone marrow biopsy: CD3 IHC, 20×. [/fig]
[fig] Figure 6: Bone marrow biopsy: CD56 IHC, 20×. [/fig]
[fig] Figure 7: Spleen: H&E, 20×. [/fig]
[fig] Figure 8: Spleen: CD3 IHC, 20×. [/fig]
[fig] Figure 9: Spleen: CD56 IHC, 20×. [/fig]
|
Potential anti-EBV effects associated with elevated interleukin-21 levels: a case report
Background: Germinal center derived memory B cells and plasma cells constitute, in health and during EBV reactivation, the largest functional EBV reservoir. Hence, by reducing germinal center derived formation of memory B cells and plasma cells, EBV loads may be reduced. Animal and in-vitro models have shown that IL-21 can support memory B and plasma cell formation and thereby potentially contribute to EBV persistence. However, IL-21 also displays anti-viral effects, as mice models have shown that CD4 + T cell produced IL-21 is critical for the differentiation, function and survival of anti-viral CD8 + T cells able to contain chronic virus infections. Case presentation: We present immunological work-up (flow-cytometry, ELISA and genetics) related to a patient suffering from a condition resembling B cell chronic active EBV infection, albeit with moderately elevated EBV copy numbers. No mutations in genes associated with EBV disease, common variable immunodeficiency or pertaining to the IL-21 signaling pathway (including hypermorphic IL-21 mutations) were found. Increased (> 5-fold increase 7 days post-vaccination) CD4 + T cell produced (p < 0.01) and extracellular IL-21 levels characterized our patient and coexisted with: CD8 + lymphopenia, B lymphopenia, hypogammaglobulinemia, compromised memory B cell differentiation, absent induction of B-cell lymphoma 6 protein (Bcl-6) dependent peripheral follicular helper T cells (pT FH , p = 0.01), reduced frequencies of peripheral CD4 + Bcl-6 + T cells (p = 0.05), compromised plasmablast differentiation (reduced protein vaccine responses (p < 0.001) as well as reduced Treg frequencies. Supporting IL-21 mediated suppression of pT FH formation, pT FH and CD4 + IL-21 + frequencies were strongly inversely correlated, prior to and after vaccination, in the patient and in controls, Spearman's rho: − 0.86, p < 0.001. Conclusions: To the best of our knowledge, this is the first report of elevated CD4 + IL-21 + T cell frequencies in human EBV disease. IL-21 overproduction may, apart from driving T cell mediated anti-EBV responses, disrupt germinal center derived memory B cell and plasma cell formation, and thereby contribute to EBV disease control.
# Background
Epstein-Barr virus (EBV) infects 90-95% of adults worldwide. In health and in EBV lymphoproliferative disorders, memory B cells and plasma cells, collectively, constitute the largest functional EBV reservoir. Interestingly, the increase in EBV loads, associated with immunosuppression, can be accounted for by an increase of latently infected, germinal center (GC) derived, memory B cells. Due to the immunosuppression, a fraction of (circulating) memory B cells can now complete spontaneous lytic replication and thereby increase the EBV load. Interleukin-21 is a GC derived cytokine which may impact EBV persistence in several ways. Interleukin-21 was shown to induce, in an EBV + cell line, which resembled post-GC B cells, increased latent membrane protein 1 (LMP1) and Epstein-Barr virus nuclear antigen (EBNA1) expression, Ig secretion and plasmablast differentiation, suggesting that IL-21 may support lytic EBV replication. Mice studies have documented, that interleukin (IL)-21, contributes to the formation of T follicular helper cells and thereby supports GC derived memory B cell and plasma cell differentiation. However, IL-21 has potentially potent anti-EBV effects too. Firstly, animal models have documented, that effector CD8 + T cells, capable of killing virus infected cells, require the presence of IL-21 producing CD4 + T cells. Secondly, in-vitro studies have shown, that IL-21 is pro-apoptotic for activated and resting B cells.
We provide the first documentation that elevated CD4 + IL-21 + T cell frequencies accompany human EBV disease. Furthermore, our case indicates several anti-EBV effects of sustained and increased IL-21 levels.
## Case presentation
Due to undiagnosed progressive pulmonary infiltrates, a 67-year old lady of Danish ancestry, nonsmoker, with no infectious history, was referred to our university clinic. An open lung biopsy revealed a mixed lymphocyte cell infiltrate dominated by CD3 + T cells and by areas with centroblastic/ immunoblastic/ Hodgkin-like CD20 + B cells. The proliferation marker Ki-67 + was extensively expressed. CD30 + immunoblasts were also visible. EBV positive cells (more than 50 per high power field) were found by EBV RNA insitu hybridization using a commercial (Ventana, Tucson, USA) Epstein-Barr virus small encoded RNA (EBER) probe. PCR for immunoglobulin heavy-chain gene rearrangements identified a smaller population of B cells with clonal IgH gene-rearrangements placed on a polyclonal background. Clonal rearrangements of the TCR beta-chain were not found. Lung tissue CD3 + T cells were dominated by CD4 + T cells but perforin and granzyme B positive CD8 + T cells were also present. Our hematopathologist's concluding diagnosis was lymphomatoid granulomatosis grade III. The Easy-Sep Human B cell enrichment Kit (STEMCELL™ technologies), which functions optimally with fresh cells, was used on thawed cryopreserved patient and controls PBMC. Among negatively selected patient CD19 + B cells (suspension 1) and among the remaining cells (positively selected patient CD4 + T cells and CD16 + /CD56 + NKcells, suspension 2), we determined EBV copy numbers by DNA amplification using the Lightcycler 480 (Roche Diagnostics). DNA, isolated from 200 μL of B cell enriched suspension 1, which contained 46.000 cells (with 53% B cells), generated 13.500 EBV copies, resulting in an average EBV-copy number per cell = 0.29 copies. Similarly, DNA, isolated from 200 μL of B cell depleted suspension 2, which contained 436.000 cells (with 5% B cells), generated 65.900 EBV copies, resulting in an average EBV-copy number per cell = 0.15 copies. Assuming that the average EBV copy number, among the non-B cells in the B cell enriched suspension 1, was also 0.15 copies, suggested that the B cell EBV copy number was even higher than 0.29 copies. In comparison, both suspension 1 and 2, from a healthy IgG anti-VCA positive control, were EBV copy negative.
A (positron emission tomography) PET scan revealed mediastinal adenopathy with increased fluorodeoxyglucose (FDG) uptake. The hematologists found no indication for chemotherapy. The patient had circulating IgG, but no IgM, specific for (EBNA, concentration: 47.5 U/mL, cut-off: > 20 U/mL) and viral capsid antigen (VCA, concentration: 577 U/mL, cut-off: 20 U/ mL). EBV plasma DNA levels were repeatedly (5x over a period of three years) moderately elevated (from < 1000 copies/ml (positive) to 3905 copies/ml). By serology and PCR, she tested negative for human immunodeficiency virus, hepatitis B and C virus, cytomegalovirus, parvovirus, varicella zoster virus, and syphilis. High titer anti-nuclear antibodies (ANA) were present (homogeneous nuclear staining, titer > 1280 (+++), normal range: 0-160) . Auto-antibody specificities (anti-Cyclic Citrullinated Peptide (CCP), IgM rheumatoid factor, SS-A Ro52, SS-A Ro60, SS-B, Sm, ribo-nuclear protein (RNP), Scl-70 and Jo-1, were negative. Markers for autoimmune hepatitis (soluble liver antigen (SLA) antibodies, liver cytosol (LC1) antibodies, antimitochondrial antibodies (AMA), ACTIN, liver-kidney microsomes (LKM) antibodies) were also negative and clinical signs of autoimmunity were absent. Patient 25-Hydroxy -Vitamin D3 Immunological characteristics associated with chronic EBV reactivation. , intravenous immunoglobulin substitution was initiated. Lately, the patient has experienced a 30-40 kg weight loss, her pulmonary function is rapidly deteriorating and she constantly relies on supplemental oxygen. In agreement with informed written consent and the study protocols (S-20150176 and S-20192000-48), immunologic work-up was initiated. A peripheral blood count revealed monocytosis and flow cytometry revealed reduced CD19 + B cell concentrations, reduced frequencies of CD19 + CD27 + IgD − (GC derived) memory B cells, normal CD4 + T cell concentrations but extremely low CD8 + T cell concentrations For all three subjects, this was their first PCV vaccination. Three weeks post-vaccination, our patient's PCV titers were significantly diminished compared to age and gender matched controls and only the patient failed to generate protective antibody levels to some PCV serotypes: 4, 5, 6B and 18C (data not shown). After vaccination, the patient, in contrast to three controls, responded with a 5-fold increase in CD4 + IL-21 + frequencies but with no induction of CD4 + CD45RA − CXCR5 + CCR7 low PD-1 high (peripheral) T follicular helper cells ((p)T FH )
# Discussion and conclusions
In health and in EBV lymphoproliferative diseases, GC derived memory B cells and plasma cells constitute the primary functional EBV reservoir. Hence, by inhibiting GC derived memory B cell and subsequent plasma cell differentiation, EBV loads may be reduced. We have described the immunological effects associated with an elevated CD4 + IL21 + profile in an elderly lady suffering from chronic EBV infection where some of the clinical and immunological (severe CD8 + lymphopenia, B lymphopenia, hypogammaglobulinemia) findings resembled those of B cell chronic active EBV disease (B cell CAEBV). However, due to only moderately elevated EBV copy numbers, which is not a typical characteristic of B cell CAEBV, we decided to classify her condition, based on the hematopathologist's diagnosis, as having high grade EBV reactivation. Based on our calculations, we cannot exclude that non-B cells contributed to the patient's EBV burden. However, according to our calculations, her B cells contained on average at least twice as many EBV-copies than non-B-cells, consistent with the pathological findings. T-and NK-cell tropic EBV infections are rare entities primarily reported among East Asians (our patient was of Danish ancestry). Our histological examination provided no evidence for clonal T cell expansion. T and NK-cell EBV disease also primarily affects children or young adults who develop symptoms as fever, persistent hepatitis, skin symptoms, uveitis, hepatosplenomegaly or pancytopenia, none of which characterized our patient. Animal studies have shown, that CD4 + T cell produced IL-21 is crucial for anti-viral CD8 + T cellsand CD8 + T cells are critical for controlling EBV. Hence, the elevated CD4 + IL-21 + T cell frequencies, reported here for the first time in the context of human EBV disease, could be seen as an appropriate response to her severe EBV infection. We found no mutations in the patient's IL-21, IL-21R, IL-21 promoter or in genes related to EBV disease or CVID. We also excluded auto-immune hepatitis and hepatitis B and C as causes for the elevated patient CD4 + IL-21 + T cell frequencies. Increased IL-21 levels, likely produced by pT FH, have been observed in type I autoimmune hepatitisand chronic hepatitis B, however our patient's CD4 + IL-21 + T cells were primaly non-pT FH (CXCR5-) . Collectively, our findings suggest that our patient's elevated IL-21 profile was indeed secondary to her pulmonary EBV reactivation. Our patient's IL-21 producing CD4 + T cells were dominated by an effector profile (CCR7 − ) and showed phenotypically no signs of exhaustion (similar PD-1 expression as controls, data not shown). Collectively, this phenotypic profile was consistent with the CD4 + IL-21 + T cells being able to home to and sustain antiviral CD8 + T cell function in peripheral tissues. Our patient's CD8 + lymphopenia is a characteristic also commonly found (44%) among patients with B cell CAEBV. We excluded that her CD8 + lymphopenia was secondary to HIV, bare lymphocyte syndrome type I (lack of HLAclass I molecules), vitamin D deficiency, CD8A variants or other genetic variants associated with primary immunodeficiency. Furthermore, she had a late onset infectious history and her middle aged son had normal CD8 + T cell counts and no EBV reactivation. Collectively, this points to a secondary CD8 + T-cell penia, however one fully consistent with the effects of IL-21, since IL-21 causes accumulation of virus specific effector CD8 + T-cells in peripheral tissues. Actually, IL-21 induces expression of the gut homing receptor integrin α 4 β 7 consistent with the intestinal accumulation of CD8 + T cells. Accumulation of CD8 + T cells in lymph nodes and spleen is also observed in IL-21 transgenic mice. Our patient's modest EBV copy number elevations also indicated some preserved CD4 + and CD8 + T-cell functionality in her peripheral tissue. Despite the increased intra-and extracellular IL-21 levels (which likely explained her monocytosis), the patient generated neither pT FH nor protective protein specific antibodies, notwithstanding the central importance of IL-21 for both (p) T FH differentiationand GC derived antibody formation. T follicular helper cells (T FH ) are located in the GC and are critical for the formation of GC derived memory B cellsand plasma cells. As a proxy marker for GC located T FH , we focused on their peripheral counterparts: (CD4 + CD45RA − CXCR5 + CCR7 lo PD-1 hi ) pT FH since there are functional and developmental connections between T FH and pT FH, implying that peripheral expansion of pT FH, one week post-vaccination, correlates with GC derived antibody formation. Furthermore, both the differentiation of T FH and pT FH is dependent on the GC transcriptional repressor Bcl-6. Hence, our patient's lack of pT FH was consistent with her reduced peripheral CD4 + Bcl-6 + T cell frequencies.
Priming and differentiation of CD4 + IL-21 + T cells takes place in lymph nodes, consistent with the increased CD4 + IL-21 + T cell frequencies, observed seven days post-vaccination. IL-21 is released within the GC, or in peripheral tissues, and is bound by local IL-21R + cells (B cells, CD4 + and CD8 + T cells). This might explain why not all patient sera contained elevated IL-21 levels. While having the most pronounced CD4 + IL-21 + induction among all vaccinees, the patient had absolutely no Bcl-6 dependent pT FH induction day 7. Substantiating a potential antagonism between Bcl-6 dependent pT FH differentiation and the differentiation of CD4 + IL-21 + T cells, we observed a strong inverse correlation between pT FH and CD4 + IL-21 + formation, preand post-vaccination, in patient and controls. This could indicate that other IL-21 producing CD4 + T cells subsets, apart from pT FH /T FH, might either directly compromise the development of the latter or represent alternative GC-derived differentiation pathways. In agreement with this inverse correlation, patient CD4 + IL-21 + T cells were predominantly CXCR5 − and hence of a non-pT FH phenotype. We have not been able to address whether constantly elevated IL-21 levels per se compromised pT FH /T FH formation. Although IL-21 is a Bcl-6 stimulator, IL-21 can also suppress intra-nodal Bcl-6 through signal transducer and activator of transcription 3 (STAT3) induced Blimp-1 expression. It can therefore not be excluded, that the patient's constantly elevated CD4 + IL-21 + levels could have suppressed intra-nodal, Bcl-6 dependent, pT FH and T FH inductionand thereby compromised the patient's memory B cell formationand subsequent plasmablast generation.
Interleukin-21 also induces apoptosis in resting and activated B-cellsconsistent with the widespread B cell lymphopenia and hypogammaglobulinemia observed in this patient and in a large fraction (42%) of B cell CAEBV patients not treated with rituximab.
Collectively, our data suggests several mechanisms whereby a prolonged increased IL-21 profile might reduce EBV loads: 1) apoptosis of resting and activated B cells combined with compromised GC derived 2) memory B cell and 3) plasma cell differentiation, secondary to disruption of Bcl-6 dependent pT FH formation. Due to the very low frequencies of circulating patient memory B cells, the estimation of EBV content in this subset as well as in plasma cells (which are bone marrow resident) is technically not feasible. However, as the B cells constituted the dominant reservoir in our patient, a reduction in memory B cells and plasma cell formation could be a possible mechanism to reduce EBV loads. In addition, our patient's circulating CD4 + CD25 high FoxP3 T regs frequency was markedly reduced compared to adult controls -a finding potentially attributable to the T reg inhibitory effect of IL-21. As T regs can inhibit CD8 + effector T cell function, reduced T regs frequencies could aid CD8 + T effector cell efficacy against EBV but could also increase the risk of collateral tissue (lung) damage due to unconstrained CD8 + effector T cell activity. The latter might explain her deteriorating lung function. Consistent with the aforementioned anti-viral mechanisms, the patient had only moderately elevated EBV copy numbers. We can only speculate, as to why our patient developed high grade EBV reactivation, but her homozygosity for HLA-A* 01, a well-known risk allele for EBV+ tumorscould be implicated as well as clonal EBV escape due to extended HLA-class I homozygosity.
We present the first report of elevated CD4 + IL-21 + T cell frequencies in human EBV disease, thereby complementing animal models documenting that IL-21 is critical for containing chronic viral infections. Secondly, this case suggests that continuous IL-21 overproduction, apart from driving T cell mediated anti-EBV responses, may disrupt GC derived memory B cell and plasma cell formation and thereby diminish the two B cell compartments so important for EBV persistence. copy assessments and critically read the final paper. LDR attended the patient and critically read the paper. All authors have approved the submitted version (and any substantially modified version that involves the author's contribution to the study).
# Funding
None.
## Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate Written informed consent was obtained and the study was conducted in accordance with protocols (S-20150176 and S-20192000-48), The Regional Committees on Health Research Ethics for Southern Denmark.
## Consent for publication
Written informed consent was obtained from the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal. |
Adaptive coarse-grained Monte Carlo simulation of reaction and diffusion dynamics in heterogeneous plasma membranes
Background: An adaptive coarse-grained (kinetic) Monte Carlo (ACGMC) simulation framework is applied to reaction and diffusion dynamics in inhomogeneous domains. The presented model is relevant to the diffusion and dimerization dynamics of epidermal growth factor receptor (EGFR) in the presence of plasma membrane heterogeneity and specifically receptor clustering. We perform simulations representing EGFR cluster dissipation in heterogeneous plasma membranes consisting of higher density clusters of receptors surrounded by low population areas using the ACGMC method. We further investigate the effect of key parameters on the cluster lifetime.Results: Coarse-graining of dimerization, rather than of diffusion, may lead to computational error. It is shown that the ACGMC method is an effective technique to minimize error in diffusion-reaction processes and is superior to the microscopic kinetic Monte Carlo simulation in terms of computational cost while retaining accuracy. The low computational cost enables sensitivity analysis calculations. Sensitivity analysis indicates that it may be possible to retain clusters of receptors over the time scale of minutes under suitable conditions and the cluster lifetime may depend on both receptor density and cluster size.Conclusions:The ACGMC method is an ideal platform to resolve large length and time scales in heterogeneous biological systems well beyond the plasma membrane and the EGFR system studied here. Our results demonstrate that cluster size must be considered in conjunction with receptor density, as they synergistically affect EGFR cluster lifetime. Further, the cluster lifetime being of the order of several seconds suggests that any mechanisms responsible for EGFR aggregation must operate on shorter timescales (at most a fraction of a second), to overcome dissipation and produce stable clusters observed experimentally.
# Background
The Epidermal Growth Factor (EGF) receptor (EGFR) is a well-studied member of the ErbB family of receptor tyrosine kinases (RTKs), which are involved in cell fate decisions and are implicated in numerous human cancers [bib_ref] Untangling the ErbB signalling network, Yarden [/bib_ref]. Early studies indicated that there is a close relation between EGFR dimerization and tyrosine kinase activation [bib_ref] Signal Transmission by Epidermal Growth-Factor Receptor -Coincidence of Activation and Dimerization, Canals [/bib_ref] [bib_ref] The ErbB kinase domain: Structural perspectives into kinase activation and inhibition, Bose [/bib_ref]. EGF activates its receptor by altering the receptor's conformation and removing steric hindrances that prevent dimerization [bib_ref] EGF Activates Its Receptor by Removing Interactions that Autoinhibit Ectodomain Dimerization, Ferguson [/bib_ref]. Upon activation, EGFR forms high-density membrane clusters presumably to amplify intra-cellular signaling and stimulate endocytosis [bib_ref] Direct visualization of the binding and internalization of a ferritin conjugate of..., Haigler [/bib_ref] , which facilitates signal transduction to the nucleus [bib_ref] Modeling the signaling endosome hypothesis: Why a drive to the nucleus is..., Howe [/bib_ref] [bib_ref] Four-dimensional organization of protein kinase signaling cascades: the roles of diffusion, endocytosis..., Kholodenko [/bib_ref].
The mechanisms that contribute to the localization of EGFR [bib_ref] Modeling the signaling endosome hypothesis: Why a drive to the nucleus is..., Howe [/bib_ref] [bib_ref] Four-dimensional organization of protein kinase signaling cascades: the roles of diffusion, endocytosis..., Kholodenko [/bib_ref] are complex and remain largely elusive. Single particle tracking microscopy data have shown that the membrane skeleton creates corral structures [bib_ref] Mapping ErbB receptors on breast cancer cell membranes during signal transduction, Yang [/bib_ref] [bib_ref] Paradigm shift of the plasma membrane concept from the two-dimensional continuum fluid..., Kusumi [/bib_ref] [bib_ref] Cell surface organization by the membrane skeleton, Kusumi [/bib_ref] [bib_ref] Exploring Membrane Domains Using Native Membrane Sheets and Transmission Electron Microscopy, Wilson [/bib_ref] , which are responsible for the observed inhomogeneous diffusion of EGFR [bib_ref] Three-dimensional reconstruction of the membrane skeleton at the plasma membrane interface by..., Morone [/bib_ref]. Thus, the receptor appears to perform Brownian motion inside a corral but also hop occasionally from that corral to a neighboring one. The nature of the interactions that create such diffusional barriers between corrals is still not fully understood. Furthermore, membrane rafts (a class of structures that include caveolae [bib_ref] Paradigm shift of the plasma membrane concept from the two-dimensional continuum fluid..., Kusumi [/bib_ref] [bib_ref] Cell surface organization by the membrane skeleton, Kusumi [/bib_ref] as well as clathrin pits [bib_ref] Lipid rafts: bringing order to chaos, Pike [/bib_ref] have been shown to localize receptor in the onset of endocytocis [bib_ref] Mapping ErbB receptors on breast cancer cell membranes during signal transduction, Yang [/bib_ref].
The nature of the forces that hold the receptors together is still subject to research.
In view of these complexities, theoretical approaches can aid in understanding the mechanisms involved in these processes and hold great potential to assist with the design of cancer related pharmaceuticals [bib_ref] Clathrin-and non-clathrin-mediated endocytic regulation of cell signalling, Roy [/bib_ref]. To this end, previous modeling studies on receptor clustering and inhomogeneous diffusion have shed light on the rich dynamical behavior observed in such systems. Guo and Levine [bib_ref] Surviving cell death through epidermal growth factor (EGF) signal transduction pathways: Implications..., Henson [/bib_ref] [bib_ref] Network dynamics and cell physiology, Tyson [/bib_ref] have developed a phenomenological thermodynamic model for receptor clustering, in which the latter manifests as a first-order transition, attributed to energetic interactions between receptors. Shi [bib_ref] A thermodynamic model for receptor clustering, Guo [/bib_ref] has presented a statistical mechanical model that couples receptor clustering with signaling. In that model the interaction energy depends on the conformational state of the receptor and the presence of bound ligand. For the membrane dynamics of EGFR specifically, Mayawala et al. [bib_ref] Clustering and signalling of cell receptors, Shi [/bib_ref] have performed kinetic Monte Carlo (KMC) simulations for an EGFR model incorporating diffusion, ligand binding and dimerization. It was shown that the predominant dimerization pathway depends on receptor density as well as ligand concentration.
The models just discussed do not take into account spatial inhomogeneities in the cell membrane. Yet, the presence of the membrane skeleton partitions the membrane area into compartments, between which the receptor particles hop. Leitner et al. [bib_ref] Computational modeling reveals molecular details of epidermal growth factor binding, Mayawala [/bib_ref] have developed a stochastic corral model that captures such inhomogeneous diffusion phenomena. In that model, the segments of the cytoskeleton dissociate and re-associate acting as gates, thereby regulating protein mobility. Furthermore, Niehaus et al. [bib_ref] Regulation of protein mobility in cell membranes: A dynamic corral model, Leitner [/bib_ref] have analyzed the stochastic dynamics of receptor diffusion in corralled membranes theoretically and computationally, and have derived an effective macroscopic diffusion coefficient that lumps microscopic diffusion and inter-corral hops.
In the present work, we simulate membrane inhomogeneities that pertain to membrane rafts or clathrin pits. We present a simple biological model that includes receptor diffusion, reversible dimer formation, and the dynamics of cluster dissipation. Membrane diffusion barriers are assumed to separate high-from low-receptor-density areas. The barrier is assumed to be higher for dimers than monomers, thereby allowing monomers to diffuse faster out of the high-density-area. The justification for this choice is that the cell would need to keep the activated dimeric repressor into the pit to stimulate endocytocis, but the monomers could be left to diffuse out of the highdensity-area. Such a mechanism is not known yet, and it is difficult to experimentally isolate the effect of dimerization, since inhibitors of dimerization also affect kinase activity [bib_ref] Microscopic Simulation of Membrane Molecule Diffusion on Corralled Membrane Surfaces, Niehaus [/bib_ref]. On the other hand, several recent studies have shown that receptor recruitment is inefficient and internalization proceeds with slow rates for EGFR mutants that lack kinase activity [bib_ref] The Role of Tyrosine Kinase-Activity in Endocytosis, Compartmentation, and down-Regulation of the..., Wiley [/bib_ref] [bib_ref] Recruitment of Epidermal Growth-Factor Receptors into Coated Pits Requires Their Activated Tyrosine..., Lamaze [/bib_ref] or for wild-type receptors inactivated by small molecule inhibitors [bib_ref] Effect of tyrosine kinase inhibitors on clathrin-coated pit recruitment and internalization of..., Sorkina [/bib_ref]. It has also been explicitly argued that dimerization of EGFR is sufficient in triggering endocytocis [bib_ref] Control of epidermal growth factor receptor endocytosis by receptor dimerization, rather than..., Wang [/bib_ref]. Our model captures inefficient receptor recruitment for EGFR monomers by assuming different diffusion barriers for the monomeric and dimeric forms of the repressor. The effect of diffusion is discussed in the results and the Additional file 1.
We use the KMC method due to its ability to resolve microscopic and mesoscopic spatial (e.g., membrane rafts [bib_ref] Endocytosis and intracellular trafficking of ErbBs, Sorkin [/bib_ref] , clathrin pits, or corrals [bib_ref] Three-dimensional reconstruction of the membrane skeleton at the plasma membrane interface by..., Morone [/bib_ref] and temporal heterogeneity and correlations arising from reaction nonlinearities, track individual molecules, and account for a small number of molecules. Specifically, individual receptor locations and different domains of the membrane are discretely represented, and the spatial heterogeneities in receptor density and membrane environment can easily be captured. Furthermore, tracking of receptor locations allows straightforward comparisons to single particle tracking experiments.
However, biological systems' modeling covers usually wide time and space scales rendering the KMC method CPU intensive. Hence, a multiscale approach to bridge the separation of scales, while preserving the attributes of KMC, is needed. The coarse-grained (kinetic) Monte Carlo (CGMC) method [bib_ref] Cell surface organization by the membrane skeleton, Kusumi [/bib_ref] , which groups microscopic sites into coarse cells, is a possible multiscale framework. In the CGMC method, substantial acceleration is achieved due to the reduction of the total number of cells simulated and the longer diffusion distances executed by molecules (time acceleration). However, the method treats the receptors within a coarse cell as well mixed, an assumption that naturally leads to some loss of accuracy (its error when simulating chemical reactions is though not well understood). This necessitates the use of an adaptively refined grid [bib_ref] Coarse-grained stochastic processes and Monte Carlo simulations in lattice systems, Katsoulakis [/bib_ref] [bib_ref] Coarse-grained stochastic processes for microscopic lattice systems, Katsoulakis [/bib_ref] to improve accuracy of the CGMC simulation. In the specific system modeled herein, a microscopic grid should be used in high-density plasma membrane areas. In order to demonstrate the source of error in coarse graining and the computational advantage of the ACGMC method, we compare for the first time the accuracy and computational efficiency of the KMC, CGMC and ACGMC methods when reactions (e.g., reversible dimerization) are considered and perform sensitivity analysis for the lifetime of an EGFR cluster. We finally discuss the biological implications of our results which may give insight into the time scale of a regeneration or recruitment mechanism of clusters of receptors. The processes that are considered are shown in Tables 1 and 2 and amount to receptor hopping from an occupied site to a neighboring unoccupied one, receptor dimer formation from neighboring monomers, and dimer dissociation to form two monomers. This is a minimal set of key processes that can give insight into the EGFR membrane dynamics, as dimerization can be correlated to signaling rates.
## Microscopic model
The stochastic rates of occurrence (propensities) of the processes just noted are given as functions of the occupancy vector . Specifically, the monomer's hop propensity from site i1 of type ϕ1 to site i2 of type ϕ2 will be:
The dimer dissociation (monomerization) at site i1 of type ϕ1 will be:
The dimer formation (association) between monomers existing at sites i1 and i2 (of types type ϕ1 and ϕ2) will be:
Once all propensities have been calculated, the random time at which the next event will occur is calculated as:
where a 0 is the sum of all propensities and u is a uniformly distributed random number. The event m to be realized at time t current + τ is randomly chosen from the list of all possible events, which consists of all microscopic processes (diffusion, dimerization, monomerization) at each site. The higher the propensity of event m, the higher the probability that this event will be realized at time t current + τ. This simulation scheme is similar to the Gillespie algorithm [bib_ref] Spatially adaptive lattice coarse-grained Monte Carlo simulations for diffusion of interacting molecules, Chatterjee [/bib_ref] [bib_ref] Spatially adaptive grand canonical ensemble Monte Carlo simulations, Chatterjee [/bib_ref] but for processes occurring on a lattice. A noticeable difference in our implementation (from the Gillespie direct simulation method) includes a binary-tree search and update algorithm [bib_ref] A general method for numerically simulating the stochastic time evolution of coupled..., Gillespie [/bib_ref] [bib_ref] Exact stochastic simulation of coupled chemical reactions, Gillespie [/bib_ref] to handle the computational cost arising from the large number of microscopic sites of a lattice.
## Simulation setup
The plasma membrane often consists of small areas of high-density receptors in large areas of low-density receptors. In order to simulate such a system, in all simulations (unless otherwise noted) a single rectangular domain of side length equal to 48 nm was placed in the simulation space. This side length is within the generally accepted membrane raft size of 10-200 nm [bib_ref] Efficient Exact Stochastic Simulation of Chemical Systems with Many Species and Many..., Gibson [/bib_ref] [bib_ref] An overview of spatial microscopic and accelerated kinetic Monte Carlo methods, Chatterjee [/bib_ref]. This domain represents a potentially 'high-density' region of the membrane, which accounts for 4% of the entire simulation box. Thus, for this simulation, the 48 × 48 nm domain is enclosed in a 240 × 240 nm simulation box to which periodic boundaries were implemented. A lattice constant (the distance between lattice sites) of 6 nm was chosen following. The whole simulation space consists of 40 × 40 (1600) sites, and a high-density domain of 8 × 8 (64) sites. Larger domains (e.g., 1024 × 1024, representing a membrane section size of ~38 μm 2 ) were also simulated and showed comparable results (not shown).
To initialize the simulation, a high number of EGFR monomers are placed within one high-density region (or within multiple regions where applicable), the latter being surrounded by an area of low-density of receptors. Consequently, most reactions will happen at short time scales in the high-density region, whereas few will occur at the low-density region.
# Monte carlo methods
For the simulations of this work three frameworks were used and compared: KMC, uniform mesh CGMC (here-
[formula] s f ,i S s f ,i * = 1 s f ,i S a M Diff i i M i M i = ⋅ ⋅ → * Γ f f f f s s 1 1 2 2 1 1 2 2 , , , ,(1) [/formula]
a Diss
[formula] Diss i D = ⋅ Γ f f s 1 1 1 ,(2) [/formula]
a Assoc after abbreviated as UCGMC), and ACGMC. Exact KMC simulations fully resolve microscopic events, and can thus be extremely computationally intensive when simulating fundamental EGFR processes (diffusion, dimerization, etc.), for long times and in treating diffusion barriers from high to low-density regions. The need for long simulations arises from the necessity of simulating long-lived structures, such as clusters of receptors. To overcome these time scale issues, one has to resort to approximate methods in order to reduce the computational cost with minimal loss of accuracy. Uniform mesh CGMC (UCGMC) simulations have been shown to accelerate simulations of systems with a wide range of time and space scales [bib_ref] Detection of non-Brownian diffusion in the cell membrane in single molecule tracking, Ritchie [/bib_ref]. These methods refer to coarse-grained cells consisting of several sites. Hence, one introduces stochastic processes for the number of particles in a coarse-grained cells (the coarse variable), rather than resolving events at the microscopic (single-site) scale [bib_ref] Coarse-grained stochastic processes and Monte Carlo simulations in lattice systems, Katsoulakis [/bib_ref] [bib_ref] Coarse-grained stochastic processes for microscopic lattice systems, Katsoulakis [/bib_ref]. However, these processes exhibit error when nonlinear chemical (e.g., bimolecular) reactions occur in areas of high concentration due to averaging within each coarse cell, which results in an incorrect description of spatial correlations [bib_ref] Coarse-grained stochastic processes and Monte Carlo simulations in lattice systems, Katsoulakis [/bib_ref] [bib_ref] Coarse-grained stochastic processes for microscopic lattice systems, Katsoulakis [/bib_ref] [bib_ref] Coarse-grained stochastic processes and kinetic Monte Carlo simulators for the diffusion of..., Katsoulakis [/bib_ref]. Thus, we expect the UCGMC method to inaccurately simulate the processes in a high-density area.
[formula] Assoc i M i M = ⋅ ⋅ Γ f f f f s s 1 2 1 1 2 2 , , ,(3)t = ⋅ ⎛ ⎝ ⎜ ⎞ ⎠ ⎟ 1 0 1 a ln u (4) [/formula]
We propose that the adaptive CGMC (ACGMC) method [bib_ref] Spatially adaptive lattice coarse-grained Monte Carlo simulations for diffusion of interacting molecules, Chatterjee [/bib_ref] is one platform to overcome this problem by using a fully refined lattice for the high-density region only. In choosing a level of coarse graining of the lowdensity region one can perform UCGMC simulations at various levels of coarse-graining and compare results (e.g., dimerization rates) to those of a KMC simulation. The level of coarse-graining can then be chosen so that the error is smaller than a tolerance. The coarse grained variable that shows the number of particles of type α existing in sites of type ϕ for coarse-cell C k is:
Here qϕ ,k is the number of sites of type ϕ existing in coarse cell C k . If qϕ ,k = 1 then the microscopic events are resolved. The coarse-grained propensities are given as in the following examples. For dimer dissociation:
For dimer formation:
for the case where k = k' and ϕ1 ≠ ϕ2 (see in [bib_ref] Spatially adaptive grand canonical ensemble Monte Carlo simulations, Chatterjee [/bib_ref] [bib_ref] Coarse-grained kinetic Monte Carlo models: Complex lattices, multicomponent systems, and homogenization at..., Collins [/bib_ref] for more details on how to evaluate coarse propensities).
Simulations (not shown) indicate that the error depends on the diffusion and reaction time scales (the so called Damkohler number, Da) as well as the density of receptors. The rate of diffusion is roughly 3 orders of magnitude faster than monomer dimerization (Da < 10 -3 ), and the concentration of monomers is expected to be low (fraction of occupied sites is <0.01). Under these conditions, spatial detail in the low-density region is unimportant (the system is basically well-mixed) and any level of coarse graining is expected to return a reasonably accurate rate of reaction. In order to illustrate the limitations of each method, here we compare the results of the aforementioned methods, using three on-lattice simulation layouts shown in [fig_ref] Figure 1: KMC [/fig_ref]. Panel (a) corresponds to the traditional KMC simulation, whereby all microscopic sites are resolved. Panel (c) pertains to the UCGMC simulation, whereby the entire simulation space including the central highdensity region is represented as 25 coarse-grained cells of size 8 × 8 sites (48 nm × 48 nm). Finally, panel (b) portrays the grid of the ACGMC simulation, whereby the central (high-density) region is microscopically resolved, like a KMC simulation, while the rest of the lattice is uniformly coarse-grained into 8 × 8 (48 nm × 48 nm) CG cells. This multiscale approach attempts, as illustrated below, to combine the efficiency of the UCGMC method with the accuracy of the KMC method.
[formula] h s f a f a , , k j j C k = ∈ ∑ (5) a C Diss Diss i D i C Diss k D k k = ⋅ = ⋅ ∈ ∑ Γ Γ f f f f s h 1 1 1 1 1 1 , ,(6)a Assoc n n q q C Assoc k k k B k k k = ⋅ ⋅ − ⎛ ⎝ ′ ≠ ′ ∑ Γ f f f f f f f f f 1 2 1 2 1 2 1 1 2 , , , , ,, ⎜ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟ ⎛ ⎝ ⎜ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟ ⎟ ⋅ ⋅ ⋅ h f h f f f 1 2 1 2 , , , , k D k D q k q k [/formula]
# Results and discussion
## Monte carlo simulations for egfr cluster dissipation
In the CGMC method, the diffusion equations are accurate between coarse grained cells of different sizes when there is a single time scale of diffusion and there are no intermolecular forces between proteins (Fickian diffusion) [bib_ref] Coarse-grained kinetic Monte Carlo models: Complex lattices, multicomponent systems, and homogenization at..., Collins [/bib_ref]. However, these equations cannot be directly applied to the special case of domains separated by diffusional barriers, where the hop over a barrier is slower than the microscopic diffusion inside the domain (a two time scale diffusion process) [bib_ref] Coarse-grained stochastic processes and kinetic Monte Carlo simulators for the diffusion of..., Katsoulakis [/bib_ref]. Both the barrier and microscopic diffusion rate contribute to the effective diffusion rate between coarse cells. The effective diffusion rate between two cells separated by a single barrier, Γ effective , as a function of the barrier hopping rate and microscopic diffusion was formulated and tested in [bib_ref] Microscopic Simulation of Membrane Molecule Diffusion on Corralled Membrane Surfaces, Niehaus [/bib_ref] and is given by
Here Γ micro and Γ barrier are the microscopic and barrier hopping rates, respectively, for a single microscopic site jump. L CG is the coarse cell center-to-center distance of the two relevant coarse-grained cells perpendicular to the cell boundary over which diffusion occurs. Eq. (8) was applied to inhomogeneous membrane EGFR simulations to test the accuracy of UCGMC for the case of pure diffusion [fig_ref] Figure 2: Density of receptors vs [/fig_ref]. Microscopic diffusion and barrier hopping occur for a single species on the lattice [fig_ref] Table 1: Diffusion model in CGMC simulations [/fig_ref]. The simulation starts with concentrated receptors in the high-density region. Over time, the receptors diffuse out of this region.
The ACGMC and UCGMC methods results coincide with those of the KMC method [fig_ref] Figure 2: Density of receptors vs [/fig_ref] , confirming that Eq. (8) correctly describes the effective diffusion rate. Additionally, this shows that the CGMC method correctly handles diffusion for spatially heterogeneous systems with high-density areas separated by low-density areas (in the absence of intermolecular forces between receptors). This allows us to attribute CGMC errors, in later high-density simulations, to reactions.
## C)
With regard to simulation cost, UCGMC simulations are cheaper than KMC by over three orders of magnitude, whereas the ACGMC method is cheaper by 0.5-2.5 orders of magnitude [fig_ref] Figure 3: Instantaneous ratio of CPU time to simulated time of the KMC, ACGMC,... [/fig_ref]. The efficiency of the ACGMC method is especially pronounced at longer times when the majority of receptors have jumped out of the high-density region. The two coarse-grained simulations are faster because they only simulate large coarse hops in the low-density region, whereas the KMC simulation resolves all microscopic moves.
While the KMC and UCGMC simulations increase in cost until reaching a steady state, the ACGMC simulation reaches a maximum cost at a time that corresponds to the high-density region being half covered (initially it is fully covered). At this point the number of molecules and vacancies on the fine-grid, high-density region is equal, leading to the maximum frequency of expensive microscopic diffusion events. As receptors leave the central domain and the coverage fraction of occupied high-density sites falls below 0.5, the ACGMC cost drops by an order of magnitude. This is because there are fewer receptors in the expensive fine-grid central domain, and more receptors in the coarse-grained outer domain.
## Short time egfr diffusion-reaction simulations
The previous section demonstrated that UCGMC simulations provide the same results as the KMC simulation at a much reduced cost for a diffusion only system. We expect similar CPU savings when extended to reacting systems, but the accuracy of the simulation comes into question. In the following, we investigate the performance and accuracy of the CGMC method for diffusionreaction systems.
The reaction-diffusion model for the EGFR system is shown in [fig_ref] Table 2: EFGR diffusion-reaction model [/fig_ref]. The rates are taken from Table three of [bib_ref] Microscopic Simulation of Membrane Molecule Diffusion on Corralled Membrane Surfaces, Niehaus [/bib_ref] and represent a diffusion-controlled system. In nominal simulations, the diffusion rate of the dimer is taken to be half of that of the monomer. This choice is based on the Einstein-Stokes equation, according to which the diffusivity is inversely proportional to the radius of the diffusing particle. This may not necessarily be true in 2- [fig_ref] Table 1: Diffusion model in CGMC simulations [/fig_ref]. dimensional diffusion on membranes, in which the dependence of the diffusivity on size is generally weaker [bib_ref] Brownian motion in biological membranes, Saffman [/bib_ref]. However, as demonstrated in the Additional file 1, the results of our simulations are not sensitive to the values of these diffusivities for the parameter set chosen here.
This lack of sensitivity can be explained by identifying the rate limiting mechanism for cluster dissipation in our problem. Since the diffusivities of the receptor monomers and dimers in the inner and outer domain are much higher than the hopping rates, the rate limiting mechanism is the hopping of the monomer and the dimer across the different domains. The diffusion of monomers over the barrier separating high-and low-density regions is three orders of magnitude slower compared to the microscopic monomer diffusion inside the two regions. On the other hand, dimer diffusion over a barrier is four orders of magnitude slower than the microscopic dimer diffusion. Thus, as soon as a receptor hops out of the central domain, it quickly diffuses away, without introducing any crowding effects. Further, the values of different barrier hopping rates were chosen in accordance to intuition that once the receptors dimerize, they will be held into a clathrin pit or caveola and eventually be internalized through the corresponding endocytocis pathways. Monomeric receptors are inactive and therefore need not be internalized; instead, they are expected to diffuse easily out of the high-density domain. In the simulation, all receptors are seeded inside the high-density region in monomer form. This initial condition corresponds to the worst case scenario for the receptor cluster lifetime, namely it will give the shortest lifetimes.
Typical results for the diffusion-reaction model are shown in [fig_ref] Figure 4: Short-time [/fig_ref]. Notable is the error made by the UCGMC method due to the reactions happening inside the high-density region. At very short times [fig_ref] Figure 4: Short-time [/fig_ref] , dimerization and monomer barrier jumps dominate. As a result, a fine grid is necessary for accurate simulations. At higher densities within the high-density region, the finemeshes of KMC and ACGMC result in accurate but expensive simulations. At low-densities (in the outer region), the coarse meshes of ACGMC and UCGMC produces accurate results with low computational expense.
In these simulations, clusters of receptors are maintained for a while by stabilizing the dimer form inside. Initially, all receptors are in the monomer form and within a short time (<0.01 s), they either leave the highdensity region or dimerize. Dimers have such a reduced rate of hopping over the barrier that the dimerized receptors are essentially locked inside the high-density area. The cluster thus formed slowly dissipates via two mechanisms: i) dimers dissociate and the resulting monomers hop through the diffusional barrier before associating again; ii) dimers hop through the diffusional barrier.
Counting of the jumping events shows that the contributions of both mechanisms are of the same order of magnitude; yet, the contribution of mechanism (i) is more significant under our conditions.
## Long time egfr diffusion-reaction simulations
The short-time simulations of [fig_ref] Figure 4: Short-time [/fig_ref] show only the creation of receptor clusters. To calculate the lifetime of these clusters and the factors controlling cluster longevity, long simulations were performed.
KMC is expensive to run at this timescale; as a result, KMC comparisons with the coarse-grained simulations were done up to only 20 s. These comparisons reveal that the ACGMC and UCGMC methods produce results that are in excellent agreement with those of the KMC simulation [fig_ref] Figure 5: Long-time [/fig_ref] , and are able to easily reach the final steady state concentrations in reasonable CPU time [fig_ref] Figure 6: Long-time [/fig_ref]. The accurate ACGMC simulation reduces the cost of simulation by 2-3 orders of magnitude allowing us to obtain accurate statistics and perform a sensitivity analysis discussed below.
The time courses of receptor density simulated with the ACGMC and UCGMC methods show that the monomer coverage reaches quasi-steady-state for <1 s, due to monomer diffusion between high-density and low-density regions. On the other hand, dimer hopping over the diffusional barrier and dimer dissociation is slow, and thus, the dimers are kinetically held in the high-density area and do not relax to a uniform density until ~30 s. Due to this effect, the ACGMC method becomes more efficient as the kinetically held clusters dissipate (fewer receptors in the computationally expensive high-resolution central domain) around 1 s [fig_ref] Figure 5: Long-time [/fig_ref]. These simulations demonstrate high-density spatial receptor [fig_ref] Table 2: EFGR diffusion-reaction model [/fig_ref] heterogeneity of receptors persisting on the order of seconds.
# Sensitivity analysis
In order to investigate the dominant mechanisms controlling the properties of receptor clusters, a sensitivity analysis of the diffusion-reaction model of [fig_ref] Table 2: EFGR diffusion-reaction model [/fig_ref] was performed.
To facilitate our analysis, we used two metrics, noted in [fig_ref] Figure 7: Comparison metrics for clustering [/fig_ref] , which shows an example plot of the receptor density in the central (high-density) region (based on a weighted sum of monomers and dimers (two receptors per dimer) normalized by the overall system density (ρ HDR /ρ overall ), which remains constant, vs. time. These metrics are: i) initial cluster density and ii) cluster lifetime. The initial cluster density is defined as ρ HDR /ρ overall at 0.1 s. At this time, the initial monomers have either dimerized or left the high-density area [fig_ref] Figure 4: Short-time [/fig_ref]. This metric can also be thought of as the effectiveness of receptor trapping by the diffusional barriers when dimerization partners are readily available. The cluster lifetime is defined as the time at which the concentration of receptors within the high-density domain drops below 5 times the overall receptor density (namely ρ HDR /ρ overall = 5). The cluster lifetime illustrates how effective the diffusional barrier is at stabilizing receptors in the dimer form.
These metrics were calculated for a range of values of the following variables: size of the high-density region, overall receptor density, and dimer barrier hopping propensity. The sizes (48 nm and 24 nm) were chosen within the observed 10-200 nm range of membrane rafts on living cells [bib_ref] Microscopic Simulation of Membrane Molecule Diffusion on Corralled Membrane Surfaces, Niehaus [/bib_ref]. The overall density of receptors was also varied, since the dramatically different density of receptors in cancerous and normal cells is suspected to play important role in the dysregulation of cell communication. Finally, dimer barrier diffusion was disabled in some simulations to reflect that dimers may cross the barrier with an extremely low probability.
The results of the sensitivity analysis on these metrics appear in [fig_ref] Figure 8: Sensitivity of clustering to monomer-only barrier diffusion, different central domain sizes, and... [/fig_ref]. The initial cluster density [fig_ref] Figure 8: Sensitivity of clustering to monomer-only barrier diffusion, different central domain sizes, and... [/fig_ref] is most noticeably affected by the density of receptors in the simulation. For a fixed size of the high-density region, higher overall density simulations (833 receptors/μm 2 ) exhibit more pronounced initial clustering relative to low-density simulations. This result can be attributed to Overall density of 150 receptors/μm 2 and rates from [fig_ref] Table 2: EFGR diffusion-reaction model [/fig_ref]. Dimer (two receptors per dimer) and monomers (single receptor) counts are combined. [fig_ref] Table 2: EFGR diffusion-reaction model [/fig_ref] with a 48 nm × 48 nm high-density region. This plot begins approximately at the end time of [fig_ref] Figure 3: Instantaneous ratio of CPU time to simulated time of the KMC, ACGMC,... [/fig_ref]. The ACGMC method shows an additional gain in efficiency once the corralled cluster dissipates between 10 and 100 s. KMC simulations were only run to 20 s due to computational intensity. the higher receptor density increasing the number of dimerization partners adjacent to any given monomer (higher dimerization rate). Consequently, a higher proportion of receptors remain in the dimer form as the density of receptors in the central domain increases.
Smaller domain sizes have a higher circumference to area ratio and therefore result in a higher likelihood for a monomer to border an edge (increased monomer barrier hopping rate). The diffusion time from the center of the domain to the circumference also drops. As the above logic would suggest, smaller domains have a lower initial cluster density at the same overall densities as larger domains.
Disabling dimer barrier diffusion [fig_ref] Figure 8: Sensitivity of clustering to monomer-only barrier diffusion, different central domain sizes, and... [/fig_ref] has very little effect on the initial cluster density, reflecting that monomer barrier hopping is effectively the only path by which receptors leave the high-density region in the very early stages of the simulation. On the other hand, when dimer diffusion is taken into account, the smaller clusters have a much weaker hold on the receptors [fig_ref] Figure 8: Sensitivity of clustering to monomer-only barrier diffusion, different central domain sizes, and... [/fig_ref]. Since the former have a higher circumference to area ratio, dimers capable of jumping have more contact with the barrier and thus hop over the barriers at a faster rate, in contrast to the case of larger clusters.
Disabling dimer diffusion extends cluster lifetimes [fig_ref] Figure 8: Sensitivity of clustering to monomer-only barrier diffusion, different central domain sizes, and... [/fig_ref] by half an order of magnitude in 48 nm domains and one order of magnitude in 24 nm domains. This disproportionate increase in small domains is attributed to a higher chance of dimerization of dissociated dimers. We assume that past the initial stage (>0.01 s of [fig_ref] Figure 4: Short-time [/fig_ref] the probability of more than 2 monomers at a time existing in the high-density domain is negligible (since 2 monomers will most likely dimerize or jump a barrier long before another dimer breaks). Smaller domains (24 nm) will hold the two monomers much closer together than the larger domains (48 nm) giving the monomers a higher chance of association before one monomer hops over the diffusional barrier.
It appears that for large clusters, the receptor density plays a secondary role in the cluster lifetime [fig_ref] Figure 8: Sensitivity of clustering to monomer-only barrier diffusion, different central domain sizes, and... [/fig_ref]. On the contrary, for small clusters the density of receptors is a major factor for determining longevity. This suggests that the sensitivity of cluster lifetime to receptor density is correlated with cluster size. For large clusters, the size itself has a dominant effect on cluster lifetime, whereas for smaller clusters the lifetime is primarily a function of receptor density. Manipulating the cluster size together with receptor density has a greater overall effect on the dispersion rate of EGFR clusters than changing each variable individually.
In order to understand the effect of rate constants for monomerization, dimerization, and monomer barrier hopping (k M , k D , and D M-Barrier , respectively), we defined a new nominal case with an overall receptor density of 833 receptors/μm 2 , disabled dimer barrier hopping, and use a side length of 48 nm for the rectangular high-density region [fig_ref] Figure 8: Sensitivity of clustering to monomer-only barrier diffusion, different central domain sizes, and... [/fig_ref]. The results of this sensitivity analysis are shown in [fig_ref] Figure 9: Sensitivity of clustering to reaction and barrier diffusion rate constants [/fig_ref]. Each rate was increased by a factor of 2 or 10.
It was observed that the initial cluster density is mostly affected by increasing the rate at which monomers hop over the barrier. Initially, all receptors are in monomer form, and thus, an increase in the monomer barrier jumping rate significantly decreases the initial cluster density. Similarly, higher dimerization rates result in higher initial cluster densities, since monomers lock into the dimer state faster. The rate of monomerization (dissociation of dimers) has little influence, which is to be expected, since at short times the primary events taking place are dimerization reactions and barrier hopping.
Moreover, all kinetic rate constants affect the cluster lifetime. Changes to the rate of monomerization and dimerization influence the lifetime more than changes to the monomer barrier diffusion rate. Increasing the dimerization rate by an order of magnitude increases the kinetic lifetime of the cluster by about half an order of magnitude and well into the range of minutes. Increasing the rate of monomerization by an order of magnitude shortens the cluster lifetime by approximately an order of magnitude. The increase in the rate of monomerization seems to have a relatively greater effect on cluster lifetime than proportional changes to the rate of dimerization. These results can be explained as follows. More time spent in monomeric form directly correlates with faster cluster dissolution. This is because two monomers are only capable of associating to form a dimer if they both reside in the high-density region long enough for a dimerization event (a bimolecular reaction) to occur. If either monomer resulting from dimer disassociation (a monomolecular reaction) leaves the high-density region, the remaining monomer will most likely leave this region as well. This causes the overall probability of dimerization of two monomers to be a function of both the rate of monomer barrier hopping and the dimerization reaction rate constant, whereas the rate of monomerization is only a function of the monomerization reaction rate constant.
Given the uncertainty in kinetic and diffusion rate constants, it is quite possible that regions of high concentration of receptors could kinetically lock clusters over the period of minutes. Eventually, though, without a thermodynamic stabilization or regeneration mechanism, no long-term (beyond seconds to minutes) clustering will be observed with this model. Stabilization of these clusters would be necessary in order to effectively trigger endocytocis. The results of our simulations postulate that any stabilization or regeneration mechanism must operate on shorter timescales (possibly fraction of seconds or even shorter), to overcome dissipation and result in stable clusters seen experimentally.
# Conclusions
Adaptive Coarse-Grained kinetic Monte Carlo (ACGMC), a multiscale spatial stochastic simulation, was applied to the EGFR diffusion-reaction system. The ACGMC method properly captures the detailed spatial reactions in high receptor density regions of the membrane while efficiently and accurately simulating the lowdensity areas of the membrane with a low resolution grid. A sensitivity analysis of the density and longevity of these clusters was carried out.
Given the uncertainty in the kinetic parameters of receptor chemistry, kinetic stabilization of EGFR clusters on the order of minutes is quite plausible. EGFR clusters would only need to exist on the order of seconds or less to successfully pass on signals, or for the formation of endosomes. Our work demonstrates that diffusion barriers provide a plausible mechanism for locking clusters for short times. Any stabilization or regeneration mechanism must operate on shorter timescales (possibly fractions of seconds or even shorter), to overcome cluster dissipation. In view of our results, methods to break, strengthen, reorganize, or otherwise manipulate membrane barriers that localize trans-membrane receptors will be valuable to develop.
At relatively high receptor densities, smaller domains exhibit stronger clustering than large ones, whereas at low receptor densities clustering is weaker in smaller domains. Our results demonstrate that cluster size must be considered in conjunction with receptor density, as they synergistically affect EGFR clustering. It would thus be inappropriate to compare the behavior of cells of similar receptor densities but with different high-density region sizes or dispersions.
The ACGMC method is promising for a wide variety of multiscale and spatially heterogeneous problems. For example, it could also be conceivably expanded to three dimensions and applied to a whole new class of systems. Using the EGFR system as an example, the lifetime of an endosome, including receptor clustering, budding, and cytoplasmic transport may be simulated using 'thin' coarse cells for the membrane and coarse 3D cells representing the cytoplasm. ACGMC would allow detailed spatial resolution for high-reaction locations (local areas of the membrane), with coarse and computationally cheap cells for diffusion-heavy processes (cytoplasmic transport).
# Additional material
Authors' contributions SC developed the algorithm, set up and performed the simulations and drafted the manuscript. MS revised the manuscript, wrote the majority of the methods sections, and assessed the sensitivity of the results. DGV conceived the study, and participated in its design and coordination and helped refine the manuscript. All authors read and approved the final manuscript.
[fig] Figure 1: KMC (a), ACGMC (Adaptive) (b), and UCGMC (Uniform) (c) layouts. The central 48 nm domain starts fully covered with a local concentration of ~26000 receptors/μm 2 . A diffusional barrier separates the high-density central region from the low-density outer region. [/fig]
[fig] Figure 2: Density of receptors vs. time obtained using three simulations for the diffusion-only model. Both coarse-grained simulations perform accurately. For the UCGMC simulations equation (8) was used for the rate of hopping between the low-and high-density areas. Parameter values as in [/fig]
[fig] Figure 3: Instantaneous ratio of CPU time to simulated time of the KMC, ACGMC, and UCGMC methods in a diffusion-only system with 48 nm × 48 nm high-density region. Parameter values as in Table 1. [/fig]
[fig] Figure 4: Short-time (t = 1 -100 ms) density of receptors in monomer (M) and dimer (D) form.Overall density of 150 receptors/μm 2 , kinetic rates from [/fig]
[fig] Figure 5: Long-time (t = 0.1 -200 s) trajectory of receptor density. [/fig]
[fig] Figure 6: Long-time (t = 0.1 -200 s) CPU cost comparison of the KMC, ACGMC, and UCGMC methods in the reaction-diffusion system of [/fig]
[fig] Figure 7: Comparison metrics for clustering. ρ HDR is the receptor density in the central (high-density) region. [/fig]
[fig] Figure 8: Sensitivity of clustering to monomer-only barrier diffusion, different central domain sizes, and different overall receptor densities. (a) Initial cluster density (ρ HDR /ρ overall at t = 0.1 s). (b) Cluster lifetime (time when ρ HDR /ρ overall = 5). (c) ACGMC layout of 48 nm domain simulations. (d) ACGMC layout of 24 nm domain simulations. All receptors initially start at random locations in the central domains. [/fig]
[fig] Figure 9: Sensitivity of clustering to reaction and barrier diffusion rate constants. (a) Initial cluster density (ρ HDR /ρ overall at 0.1 s). (b) Cluster lifetime (time when ρ HDR /ρ overall = 5). [/fig]
[table] Table 1: Diffusion model in CGMC simulations. [/table]
[table] Table 2: EFGR diffusion-reaction model. [/table]
|
Use and Misuse of Alcohol Among Older Women
Older women may be especially at risk for alcohol problems because they are more likely than men to outlive their spouses and face other losses that may lead to loneliness and depression. Physiologically, women are also at greater risk for alcohol-related health problems as they age. Because of these risks, alcohol use recommendations for older women generally are lower than those set for both older men and younger women. Screening and brief intervention may be especially useful in minimizing alcohol problems in older women. Although brief intervention research with this population is limited, the findings are promising. KEY WORDS: alcohol use disorder in the elderly; female; risk factors; protective factors; prevalence; identification and screening for AODD (alcohol and other drug disorders); brief intervention; prevention goals; treatment method; treatment outcome; recommendations or guidelines FREDERIC C. BLOW, PH.D., is an associ ate professor and senior associate research scientist in the
T he growth in the number of people age 60 and older will bring a soaring increase in the amount and cost of primary and spe cialty care for this group. In 1990, those over the age of 65 comprised 13 percent of the U.S. population; by the year 2030, older adults are expected to account for 22 percent of the popula tion (U.S. Bureau of the Census 1996). Community surveys have estimated the prevalence of problem drinking among older adults to range from 1 percent to 15 percent [bib_ref] Screening for problem drinking in older primary care patients, Adams [/bib_ref] [bib_ref] A new paradigm for alcohol use in older persons, Moore [/bib_ref]. Among older women, the prevalence of alcohol misuse ranged from less than 1 percent to 8 percent in these studies. As the population age 60 and older increases, so too could the rate of alco hol problems in this age group. However, early detection efforts by health care providers can help limit the prevalence of alcohol problems and improve overall health in older adults.
Many of the acute and chronic medical and psychiatric conditions that lead to high rates of health care use by older people are influenced by the con sumption of alcohol. These conditions include harmful medication interac tions, injury, depression, memory problems, liver disease, cardiovascular disease, cognitive changes, and sleep problems [bib_ref] Brief physician advice for alcohol problems in older adults: A randomized community-based..., Fleming [/bib_ref]. For example, [bib_ref] Alcohol abuse, cognitive impairment, and mortality among older people, Thomas [/bib_ref] found that the occurrence of all types of dementia (with the exception of Alzheimer's dis ease) was higher in a sample of 2,873 people age 65 and older with definite or questionable alcohol abuse 1 com pared with those who did not abuse alcohol. At 18 months after baseline, mortality from all causes in this sample was higher among those with definite abuse (14.8 percent) or questionable abuse (20 percent) than among those with no alcohol abuse history (11.5 percent). The risk for negative alcoholrelated health effects is greater for older women than for older men at the same amounts of alcohol use.
Researchers have recently recom mended that screening and interventions focused on lifestyle factors, including the use of alcohol, may be the most appro priate way to maximize health outcomes and minimize health care costs among older adults [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref]. For example, primary health care providers can screen patients for alcohol problems and offer brief intervention-5-to 15-minute sessions of information and advice about the risks of drinking and how to reduce drinking-to help prevent atrisk drinkers from developing alcohol problems. In randomized clinical trials, women have been found to benefit most from brief interventions [bib_ref] Brief physician advice for problem alcohol drinkers: A randomized controlled trial in..., Fleming [/bib_ref] [bib_ref] Brief physician advice for alcohol problems in older adults: A randomized community-based..., Fleming [/bib_ref].
This article examines alcohol use among older women, related risk fac tors and beneficial effects, screening methods to detect alcohol problems in this population, and treatment and prevention approaches.
## Older women have increased risks for alcohol problems
Older women tend to have longer life expectancies and to live alone longer than men, and they are less likely than men in the same age group to be finan cially independent. These physical, social, and psychological factors are sometimes associated with at-risk drinking in older adulthood, so they are especially relevant for older women.
Older women have major physical risk factors that make them particularly susceptible to the negative effects of increased alcohol consumption [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref]. Women of all ages have less lean muscle mass than men, making them more susceptible to the effects of alco hol. In addition, there is an age-related decrease in lean body mass versus total volume of fat, and the resultant decrease in total body mass increases the total distribution of alcohol and other mood-altering chemicals in the body. Both men and women experience losses in lean muscle mass as they age, but women have less lean muscle mass than men throughout adulthood and, therefore, are less able to metabolize alcohol throughout their lives, includ ing into older adulthood (see [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref] for further information). Liver enzymes that metabolize alcohol and certain other drugs become less effi cient with age, and central nervous sys tem sensitivity increases with age for both genders. In sum, compared with younger adults, and with older men, older women have an increased sensi tivity to alcohol.
Older women also have a height ened response to over-the-counter and prescription medications [bib_ref] Moderate alcohol intake is associated with survival in the elderly: The Dubbo..., Simons [/bib_ref] [bib_ref] Aging and ethanol metabolism, Vestal [/bib_ref] [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref]. The use and misuse of alcohol and prescription medications are therefore especially risky for women as they age because of their specific vulnerabilities regarding sensitivity to alcohol and medications. For most patients, any alcohol consump tion coupled with the use of specific over-the-counter or prescription medi cations can be a problem. For example, combining alcohol with psychoactive medications such as benzodiazepines, barbiturates, and antidepressants can be especially problematic for this popula tion. Older women are more likely than older men to receive prescriptions for benzodiazepines in particular, and are therefore more likely to be faced with problems related to the interac tion of these medications with alcohol (see [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref] for further discussion). There is a paucity of data available on rates of the co-occurrence of alcohol and medication use in older people. This area needs more study.
Because older women generally drink less than older men or abstain from alcohol, health care providers may be less likely to recognize at-risk drink ing and alcohol problems in this popu lation. Moreover, few elderly women who abuse alcohol seek help in special ized addiction treatment settings. These problems stand in the way of effective interventions that can improve the quality of life of older women drinking at risky levels.
The following sections will first examine the prevalence of problem drinking in older women and then review the risks and benefits associated with alcohol use among older women.
The article concludes with a discussion of screening and interventions for this population.
## Prevalence of the problem
As stated above, community surveys have estimated that at-risk drinking ranges from 1 percent to 15 percent among older adults and that from 1 percent to 8 percent of older women misuse alcohol [bib_ref] Screening for problem drinking in older primary care patients, Adams [/bib_ref] [bib_ref] Brief physician advice for alcohol problems in older adults: A randomized community-based..., Fleming [/bib_ref] [bib_ref] A new paradigm for alcohol use in older persons, Moore [/bib_ref]. The wide variation of these ranges results from varying definitions of problem drinking and alcohol mis use and from the methodology used in selecting the survey respondents.
The rates of illegal drug abuse among the older population are very low. Because of the dearth of informa tion in this area, actual rates are diffi cult to measure [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref]. Future research will more completely address the use of alcohol and illegal drugs in older adulthood [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref].
Prescription drug misuse is more common and has multiple determi nants, causes, and consequences. For example, older adults may be experi encing problems related to overuse of prescription drugs because they are prescribed more medication than they can tolerate at that age, or because they are seeking prescriptions for a particular medication (e.g. benzodiazepine) from multiple providers.
## Risk and benefits associated with alcohol use by older women
Research provides evidence that one drink 2 per day (i.e., moderate drinking) is associated with certain health bene fits among older adults generally and among older women, whereas higher levels of drinking are associated with health risks. Only a few studies have either focused on women or have included sufficient numbers of older women to be conclusive about the effects for this population [bib_ref] Moderate alcohol consumption and risk of heart failure among older persons, Abramson [/bib_ref]. Therefore, this section includes studies on older adults in gen eral and on women in particular.
## Risks
A recent study of moderate and heavy drinking among older adults found that study participants reported poorer psychosocial functioning with increas ing daily alcohol consumption [bib_ref] Institute of Medicine. Broadening the Base of Treatment for Alcoholism, Graham [/bib_ref]. The frequency of drinking (drinking days per week), however, was not related to psychoso cial well-being, suggesting that the amount of alcohol consumption was a more significant factor. found that, among older women, those with a history of regular alcohol use were 2.2 times more likely to have impaired activities of daily living compared with those with no history of regular alcohol use. Alcohol use was more strongly corre lated with impairment than were smoking, age, use of antianxiety medi cation, or stroke.
Although several studies have exam ined the role of alcohol use in cardiac problems, stroke, and cancers, most of these studies have not included older women. A study using National Cholesterol Education Program data found that, among the women in the study, failure to use lipid-lowering medications was associated with alco hol consumption and smoking, among other factors [bib_ref] Adherence to National Cholesterol Education Program treatment goals in postmenopausal women with..., Schrott [/bib_ref]. In a study of postmenopausal women in the Iowa Women's Health Study, estimated the interaction of folate intake from diet and alcohol consumption at baseline for 34,393 study participants to determine the risk for specific types of breast cancer. The study compared women with low folate levels and higher alco hol consumption (i.e., more than 4 grams per day) 3 with nondrinkers who had a high folate intake. The authors found that the combination of alcohol use and low folate levels pro duced an increase in the risk of one 3 There are 12 grams of alcohol in a standard drink in the United States. type of tumor. A recent meta-analysis examined 53 epidemiological studies of the relationships between alcohol use, smoking, and breast cancer (Collabor ative Group on Hormonal Factors in Breast Cancer 2002), including 58,515 women with breast cancer and 95,067 women without the disease. This study found that, compared with women who reported drinking no alcohol, the relative risk was 1.32 for those who drank 35 to 44 grams of alcohol per day (3 to 3.6 standard drinks), and 1.46 for those who drank more than 45 grams per day (3.75 standard drinks). A relative risk of 1.32 corre sponds to a 32-percent higher risk. The relative risk of breast cancer increased by 7.1 percent for every 10 grams of alcohol consumed per day.
Epidemiological studies have clearly demonstrated that comorbidity between alcohol use and psychiatric symptoms is common in younger age groups. Less is known about comorbidity between alcohol use and psychiatric illness in later life. A few studies have indicated that a dual diagnosis with alcoholism is an important negative predictor of outcomes among the elderly [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref] [bib_ref] Heavy drinking as a risk factor for depres sion and dementia in..., Saunders [/bib_ref] [bib_ref] Alcoholism in elderly persons: A study of the psychiatric and psychosocial fea..., Finlayson [/bib_ref]. Because women are twice as likely as men to experience depression, and older women often experience several life losses that can exacerbate depression and the use of alcohol, it is important for health care providers to be aware of the potential for comorbid depression and alcoholism in this population and to keep potential comorbid factors in mind when conducting health screenings with older women, particularly when they are experiencing some of the difficult personal losses associated with aging.
## Benefits
There is growing evidence that, among otherwise healthy adults, especially middle-aged adults, moderate alcohol use may reduce risks of cardiovascular disease [bib_ref] Cognitive-behavioral treatment of older veterans with substance abuse problems, Scherr [/bib_ref] [bib_ref] Alcohol consumption and mortality among middle-aged and elderly U.S. adults, Thun [/bib_ref] , some dementing illnesses, and some cancers [bib_ref] Collaborative Group on Hormonal Factors in Breast Cancer. Alcohol, tobacco and breast..., Broe [/bib_ref] [bib_ref] Red wine, white wine, liquor, beer, and risk for coronary artery disease..., Klatsky [/bib_ref]. Simons and colleagues (2000) found that moderate alcohol intake (from 1 to 14 drinks per week) in older men and women was associated with decreased mortality. have demonstrated that older people living in the community (not in institutions) who consume moderate amounts of alcohol have fewer falls, greater mobility, and improved physical functioning when compared with nondrinkers. One of the factors affecting the disparities between the results of various studies on this topic may be the setting for the study (e.g., community, subsidized housing, assisted living situation, or institution).
In a meta-analysis of studies of alco hol's effect on coronary heart disease, [bib_ref] Alcohol's effect on the risk of coronary heart disease, Mukamal [/bib_ref] found that two drinks per day increased highdensity lipoprotein (HDL) cholesterol levels, translating to a 16.8-percent decreased risk of coronary heart disease. Additionally, a study of women with coronary heart disease found that older age, alcohol consumption, and prior estrogen use were all independently associated with higher HDL cholesterol [bib_ref] Cor relates of high HDL cholesterol among women with coronary heart disease, Bittner [/bib_ref].
The debate regarding the benefits and liabilities of alcohol use for older women continues. As new studies include larger numbers of older women, definitive recommendations regarding the relationships between alcohol use and cancers, stroke, cardiac diseases, and risk of psychiatric comor bidities will become more feasible.
Based on the risk factors associated with alcohol use by older women, drinking guidelines for this population are lower than those set for other adults, as reviewed in the next section.
## Drinking guidelines and rationale
Because of the age-related changes in how alcohol is metabolized and the potential interactions between medications and alcohol, alcohol use recommendations for older adults are generally lower than those set for adults younger than age 65. Recommendations for women are slightly lower than those for men as they age.
The National Institute on Alcohol Abuse and Alcoholism (NIAAA) and the Center for Substance Abuse Treatment (CSAT) [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref] rec ommend that people age 65 and older consume no more than one standard drink per day or seven standard drinks per week [bib_ref] Alcohol in the elderly, Dufour [/bib_ref].
These recommendations are consistent with the current evidence weighing the risks and beneficial health effects of drink ing [bib_ref] Red wine, white wine, liquor, beer, and risk for coronary artery disease..., Klatsky [/bib_ref] [bib_ref] Alcohol's effect on the risk of coronary heart disease, Mukamal [/bib_ref]. To put these recommenda tions into perspective, the guidelines for adults younger than age 65 are as follows: for women, no more than one standard drink per day; for men, no more than two standard drinks per day (U.S. Department of Health and Human Services and U.S. Department of Agriculture 1995).
## Definitions
Before discussing screening and inter vention, it is important to define the various levels of drinking. These defini tions help anchor clinical decisions regarding when and if interventions are needed. Drinking that exceeds the guidelines will not always lead to alcoholrelated problems, particularly for peo ple who are drinking a few drinks above recommended limits but not at levels that can put them at risk for alcohol dependence. It is, however, useful to consider a model indicating that the more alcohol a person consumes, the more likely that person is to have alcoholrelated problems (Institute of Medicine 1990). Categories of drinking risk pre sented here-low-risk drinking, at-risk drinking, problem drinking, and alco hol dependence-are based on that conceptualization and form a framework for understanding the spectrum of use seen in older women [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref].
Abstinence. Approximately 60 to 70 percent of older adults (70 to 80 percent of older women) abstain from drinking. Reasons for abstinence may include religious beliefs, illnesses, or previous problems with alcohol use. Alcohol-use interviews ascertain the reasons for nonuse.
Low-risk drinking is low-level alcohol use that is not problematic. Older women in this category drink within recommended drinking guidelines (less than one drink per day or seven drinks per week), are able to employ reasonable limits on alcohol consump tion, and do not drink when driving a motor vehicle or when using medica tions that may interact with alcohol.
Low-risk use of medications or other drugs would include using medications following the physician's prescription. However, screening should include a check on the number and types of medications a person is using and her concomitant use of alcohol, because interactions between medications and alcohol are not uncommon in older women.
At-risk drinking increases the chance that a person will develop drinkingrelated problems. Women age 65 and older who drink more than one drink per day are in the at-risk use category. Brief advice or brief interventions can be useful for women in this group.
Problem drinking among older women is defined as the consumption of alcohol at a level that has already resulted in adverse medical, psychological, or social consequences. Potential consequences may include injuries, medication interaction prob lems, and family problems. The pres ence of consequences, whether or not the person's drinking exceeds the recommended guideline, also suggests a need for intervention.
Alcohol abuse and dependence are disorders characterized by specific criteria. Alcohol abuse is characterized by continued drinking despite nega tive consequences and the inability to fulfill responsibilities. Alcohol depen dence, also known as alcoholism, is characterized by loss of control, preoccupation with alcohol or other drugs, and physiological symptoms such as tolerance and withdrawal (American Psychiatric Association [APA] 1994). Women age 65 and older who have alcohol abuse or dependence disorders can benefit greatly from treatment, especially elder-specific programs [bib_ref] Cognitive-behavioral treatment of older veterans with substance abuse problems, Scherr [/bib_ref].
## Screening and detection of alcohol problems in older women
CSAT [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref] has recommended that everyone age 60 and older should be screened for alcohol and prescrip tion drug use and abuse as part of regu lar health care services. People should continue to be screened yearly unless certain physical or mental health symp toms emerge during the year, or unless they are undergoing major life changes or transitions, at which time additional screenings should be conducted. The textbox lists some of the signs and symptoms of alcohol problems seen in older women. Many of these signs can be related to other problems that occur in later life, but it is important to rule alcohol use in or out of any diagnosis.
The goals of screening are to iden tify at-risk drinkers, problem drinkers, or people with alcohol abuse or depen dence disorders and to determine the need for further assessment. Screening can take place in a variety of settings including primary care, specialty care, and social service and emergency departments. Alcohol screening can be conducted because the incidence of alcohol problems is high enough to jus tify the cost, alcohol can adversely affect morbidity and mortality, and valid, cost-effective screening methods and effective treatments are available.
Systems (e.g., automatic yearly administration of alcohol screening instruments) to ensure that older women in health care settings are screened for alcohol use and conse quences are necessary for prevention and early intervention efforts. These systems must include screening for alcohol use (frequency and quantity), drinking-related consequences, medica tion use and alcohol/medication interaction problems, and depressed feelings. Screening may be conducted as part of routine mental and physical health ser vices and can be updated annually.
Screening should also take place before a patient begins taking any new medi cations or in response to problems that may be related to alcohol or medication.
Clinicians can obtain more accurate patient histories by asking questions about the recent past and by asking the alcohol use questions in the context of other health variables (e.g., exercise, weight, smoking). Alcohol (and other drug) screening for older patients should be simple and consistent with other screening procedures already in place.
Screening for alcohol use and related problems is not always standardized, and not all standardized instruments are reliable and valid with older women. The Short Michigan Alcoholism Screen ing Test-Geriatric Version (SMAST-G) [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref] , which consists of quantity and frequency questions embedded with questions about other health habits (see [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref] for a review of screening instruments for older adults), and the newer Alcohol-Related Problems Survey (Moore et al. [bib_ref] Brief Interventions and Brief Therapies for Substance Abuse, Barry [/bib_ref] are both valid and reliable instru ments with older adults. The CAGE 4 [bib_ref] Detecting alcoholism: The CAGE questionnaire, Ewing [/bib_ref] , a widely used alcohol screening test, does not have high valid ity with older adults, in particular with older women [bib_ref] Screening for problem drinking in older primary care patients, Adams [/bib_ref].
## Signs and symptoms of alcohol problems
## Prevention, brief intervention, and formal treatment with older women
For years, screening and brief interven tion have been suggested as cost-effective and practical techniques that can be used with at-risk and problem drinkers in primary care settings. CSAT has defined brief alcohol interventions as time limited (from 5 minutes to five brief sessions) and targeting a specific health behavior (at-risk drinking) [bib_ref] Brief Interventions and Brief Therapies for Substance Abuse, Barry [/bib_ref]. Over the last two decades, more research has evaluated the effectiveness of early problem detection and secondary prevention (i.e., preventing existing problems from getting worse). Such studies have evalu ated brief intervention strategies for treating problem drinkers, especially those with relatively mild-to-moderate alcohol problems who are potentially at risk for developing more severe prob lems [bib_ref] Brief physician advice for problem alcohol drinkers: A randomized controlled trial in..., Fleming [/bib_ref].
## Brief alcohol intervention goals
Brief intervention typically includes set ting flexible drinking goals that allow the patient, with guidance from the clini cian, to choose drinking moderation or abstinence. The goal of brief interven tion is to motivate at-risk and problem drinkers to change their behavior-that is, to reduce or stop alcohol consump tion. In some cases, when formal treat ment is warranted, the goal is to facilitate treatment entry. Terminology can be particularly important when working with older women. The stigma and shame associated with the term "alco holic" can be a powerful deterrent to seeking help. Avoiding pejorative terms provides a positive framework for clini cians and can help empower older women with risky alcohol or medication use to make changes, thereby reducing the negative feelings often associated with drinking problems.
Brief alcohol interventions can be conducted using guidelines and stepsadapted from work by , , and [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref]. Brief alcohol intervention protocols are designed for busy clinicians and often use a workbook that the patient can take home at the end of the session. Auxiliary issues included in the brief alcohol interven tion for older women vary based on individual patient issues and the time available for the intervention.
## Effectiveness of brief alcohol interventions with older at-risk drinkers
The spectrum of alcohol intervention for older adults ranges from prevention/ education for abstinent or low-risk drinkers and minimal advice or brief structured interventions for at-risk or problem drinkers to formal alcoholism treatment for drinkers who meet the criteria for alcohol abuse or depen dence [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref]. Although referral to formal treatment is appropriate for patients with alcohol abuse or depen dence, pretreatment strategies are also appropriate for this population. Pretreatment strategies include the use of brief interventions to help patients dis criminate between their alcohol use and the problems resulting from that use [bib_ref] Brief Interventions and Brief Therapies for Substance Abuse, Barry [/bib_ref].
Brief interventions for alcohol prob lems (for all populations) have employed various approaches to change drinking behaviors. Strategies have ranged from relatively unstructured counseling and feedback to more formal structured therapy (see [bib_ref] Brief Interventions and Brief Therapies for Substance Abuse, Barry [/bib_ref] for a review) and have relied heavily on concepts and techniques from the behavioral 4 The CAGE screening instrument [bib_ref] Detecting alcoholism: The CAGE questionnaire, Ewing [/bib_ref] consists of four questions:
- Have you ever felt you should Cut down on your drinking? - Have people Annoyed you by criticizing your drinking? - Have you ever felt bad or Guilty about your drinking? - Have you ever had a drink first thing in the morning to steady your nerves or to get rid of a hangover (Eye opener)?
self-control training literature. Several brief alcohol intervention studies conducted in primary care set tings with younger adults have shown mainly positive results. Both brief inter ventions and brief therapies (usually delivered by mental health professionals to people in substance abuse or mental health treatment) have been found to be effective in a range of clinical settings including primary care, mental health treatment, hospital, senior housing, and senior centers [bib_ref] Brief Interventions and Brief Therapies for Substance Abuse, Barry [/bib_ref]. Although fewer studies with older adults are available, two existing studies suggest that brief intervention is useful with the older population as well. and [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref] used brief interventions in ran domized clinical trials in primary care settings to reduce hazardous drinking among older adults. These studies have shown that older adults can be engaged in brief intervention, that this tech nique is acceptable in this population, and that there is a substantial reduction in drinking among at-risk drinkers receiving the interventions compared with a control group.
The first study, Project GOAL: Guiding Older Adult Lifestyles [bib_ref] Brief physician advice for alcohol problems in older adults: A randomized community-based..., Fleming [/bib_ref] , was a randomized controlled clinical trial conducted in Wisconsin with 158 older adults ages 65 to 88, 53 (34 percent) of whom were women. All patients age 65 and older in a number of primary care sites were asked to complete a screening questionnaire. Those who screened positive for at-risk drinking (i.e., those who exceeded recommended drinking guidelines) were randomized to an intervention group and a control group. One hundred forty-six subjects participated in the 12-month followup. The intervention consisted of two 10to 15-minute counseling visits during which the physician delivering the intervention followed a scripted workbook; the patients were given advice and information and asked to sign a contract designed to reinforce drinking goals. At baseline, both groups con sumed an average of 15 to 16 drinks per week. After 12 months, patients in the intervention group drank signifi cantly less than those in the control group, decreasing their consumption by about 30 percent. Because the propor tion of women in the study was small, major analyses focused on the entire sample of men and women together.
The second elder-specific study, the Health Profile Project, was conducted in primary care settings in southeast Michigan [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref]. Examining a sample that included patients age 55 and older, researchers sought to determine whether changes in drinking patterns and response to interventions occurred both in older adulthood (older than 65) and in the transitional phase from ages 55 to 65. The older-adult-specific intervention, used with both groups for consistency, included both a brief advice discussion with a psychologist or social worker and motivational interviewing techniques, and feedback. A total of 420 people participated (including those who received the intervention and the control group) in this trial, and 367 participated in 12-month followup interviews. Seventy-three women were enrolled in the study at baseline, and 69 participated in the 12-month fol lowup. The mean age of the female participants was 67.
The study found results similar to the study by for binge drinking (i.e., drinking four or more drinks per occasion) and drinking days per week, in particular, at 12-month followup. At followup, the intervention group of women averaged 7 drinks per week (within recom mended guidelines) and the control group averaged 8.2 drinks per week. Although the intervention group low ered its consumption to within NIAAA guidelines, the groups were not statisti cally different at followup. Nor did the groups differ significantly in terms of drinks per day at baseline or followup. The fact that the intervention and con trol group did not differ in drinks per drinking day at 12 months after inter vention could indicate natural minimal changes over time in behaviors for both groups. However, there were statistically significant differences between the groups in days per week (frequency) of drinking from baseline to 12 months.
On average, subjects in the intervention group decreased their drinking from 4.5 days per week at baseline to 3.1 days per week at 12 months; the control subjects drank an average of 4.3 days per week at baseline and only decreased to 3.6 days per week at 12-month followup. The intervention group showed significantly more days of abstinence per week at 12 months, indicating diminished risk. Days of abstinence are recommended for reducing risk.
These randomized controlled clinical trials extend the positive results of research on younger at-risk drinkers to the older at-risk drinking population by showing that, regardless of age, brief interventions are effective in assisting older at-risk drinkers to drink less often. The studies provide a good basis for future research focused on older women who use alcohol and on the interaction between alcohol and medications in this age group. Research is needed to determine the most effective compo nents of brief interventions with older women and the most effective venues (e.g., primary care, in-home, senior cen ter, senior housing). Research is also needed to address an under-studied area, the interaction between alcohol and medications in older women, and to determine the best methods for deal ing with this more complex problem.
Because the population of older women is increasing rapidly and rates of alcohol misuse are anticipated to increase with the aging of the Baby Boom generation, alcohol researchers need to find methods to include larger numbers of older women in studies. Randomized trials with larger sample sizes will provide a more complete pic ture of the characteristics of women who respond to brief interventions as well as the most effective education and prevention methods for this population.
## Formal, specialized treatment approaches for older women
CSAT has recommended several approaches for the effective formal treatment of older women and men with alcohol problems. These include cognitive behavioral approaches, groupbased approaches, individual counsel ing, medical/psychiatric approaches, marital and family involvement/family therapy, case management/communitylinked services and outreach, and for mal alcoholism treatment.
As with all other clinical issues, not every approach fits every older woman with alcohol abuse or dependence. Ideally, treatment should be individu alized for the specific person, taking into account his or her medical, psy chiatric, social, and cultural needs. Most of the therapeutic approaches included here have been more widely studied in younger adults [bib_ref] Brief screening for alcohol problems in elderly populations using the Short Michigan..., Blow [/bib_ref]. Only a few elder-specific studies have evaluated intervention/treatment methods other than brief intervention for at-risk drinkers and formal treat ment for people with alcohol abuse or dependence. There has been even less of a focus on older women, in part because fewer older women meet crite ria for formal treatment and because fewer women who need treatment are identified by primary providers and referred to treatment. A few examples of elder-specific studies are available, however. and Schonfeld and colleagues (2000) found that cognitive-behavioral approachessuch as teaching older adults skills necessary to rebuild social support networks and using self-management approaches for overcoming depression, grief, and loneliness-were successful in reducing or stopping alcohol use.
Research has also found that case management services are helpful for older adults receiving alcoholism treatment and may be the best way to provide outreach services. Because traditional residential alcoholism treat ment programs generally treat few older adults, small sample sizes have prevented the evaluation of formal treatment. The development of elder-specific alcoholism treatment programs in recent years has identified sufficiently large numbers of older adults with alcohol abuse or dependence disorders to begin to facili tate studies of this population (Atkin son 1995). A remaining limitation with this age group is the lack of longitudi nal studies of treatment outcomes.
In one of the few long-term studies of an elder-specific specialized alcoholism treatment program, examined multidimensional 6-month outcomes for 90 patients older than age 55. At baseline, physical health functioning was similar to that reported by seriously medically ill patients (with and without alcohol problems) in other studies, whereas psychological function ing was worse. Nearly one-third of the sample had comorbid psychiatric disor ders. Results suggested that the largest percentage of older adults who received elder-specific substance abuse treatment attained positive outcomes and that their conditions improved across a range of physical and psychosocial measures. Further research is needed in this area to determine the following:
- If elder female-specific specialized treatment is necessary, effective, or both
- If older women in elder-specific programs show better outcomes than older women in mixed-age programs
- If intervention and treatment approaches for alcohol and prescrip tion drug misuse are effective with older women.
## Summary
The growing population of older adults reflects the need for new, innovative prevention and intervention techniques and approaches targeted to older at-risk drinkers. These approaches should consider elder-specific characteristics such as alcohol-related symptoms and patterns of use, age of onset, and medi cal and mental health issues. The range of prevention and inter vention strategies available to older adults-prevention and education for people who are abstinent or low-risk drinkers, minimal advice and brief intervention for at-risk drinkers, and formal treatment for people with alco hol abuse or dependence-provides the necessary tools for health care providers to give high-quality care to older adults across the spectrum of drinking patterns.
Although some progress has been made in understanding the effective ness of alcohol screening, brief inter vention, and treatment among older women, it remains to be determined how these protocols fit into the broad spectrum of health care settings (e.g., primary care, mental health care, spe cialty physical health care, hospitals) and how to target specific interventions or treatments to appropriate subgroups of older women. The health care field must develop and test time-and costeffective methods of screening, interven tion, and treatment to provide optimal care to a vulnerable, growing, and under-recognized population of older women who are consuming alcohol and other drugs. ■ |
Quantifying the Spatial Dimension of Dengue Virus Epidemic Spread within a Tropical Urban Environment
Background: Dengue infection spread in naive populations occurs in an explosive and widespread fashion primarily due to the absence of population herd immunity, the population dynamics and dispersal of Ae. aegypti, and the movement of individuals within the urban space. Knowledge on the relative contribution of such factors to the spatial dimension of dengue virus spread has been limited. In the present study we analyzed the spatio-temporal pattern of a large dengue virus-2 (DENV-2) outbreak that affected the Australian city of Cairns (north Queensland) in 2003, quantified the relationship between dengue transmission and distance to the epidemic's index case (IC), evaluated the effects of indoor residual spraying (IRS) on the odds of dengue infection, and generated recommendations for city-wide dengue surveillance and control.Methods and Findings:We retrospectively analyzed data from 383 DENV-2 confirmed cases and 1,163 IRS applications performed during the 25-week epidemic period. Spatial (local k-function, angular wavelets) and space-time (Knox test) analyses quantified the intensity and directionality of clustering of dengue cases, whereas a semi-parametric Bayesian space-time regression assessed the impact of IRS and spatial autocorrelation in the odds of weekly dengue infection. About 63% of the cases clustered up to 800 m around the IC's house. Most cases were distributed in the NW-SE axis as a consequence of the spatial arrangement of blocks within the city and, possibly, the prevailing winds. Space-time analysis showed that DENV-2 infection spread rapidly, generating 18 clusters (comprising 65% of all cases), and that these clusters varied in extent as a function of their distance to the IC's residence. IRS applications had a significant protective effect in the further occurrence of dengue cases, but only when they reached coverage of 60% or more of the neighboring premises of a house.Conclusion: By applying sound statistical analysis to a very detailed dataset from one of the largest outbreaks that affected the city of Cairns in recent times, we not only described the spread of dengue virus with high detail but also quantified the spatio-temporal dimension of dengue virus transmission within this complex urban environment. In areas susceptible to non-periodic dengue epidemics, effective disease prevention and control would depend on the prompt response to introduced cases. We foresee that some of the results and recommendations derived from our study may also be applicable to other areas currently affected or potentially subject to dengue epidemics.
# Introduction
Dengue is a mosquito-borne infection that has re-emerged as a major international public health concern over the last four decades [bib_ref] Dengue and dengue haemorrhagic fever, Rigau-Perez [/bib_ref] [bib_ref] Dengue and dengue hemorragic fever: its history and resurgence as a global..., Gubler [/bib_ref] [bib_ref] Epidemic dengue/dengue hemorrhagic fever as a public health, social and economic problem..., Gubler [/bib_ref]. Caused by four closely related yet antigenically distinct single-stranded RNA viruses (genus Flavivirus, family Flaviridae), dengue viruses persist in a horizontal Aedes aegyptihuman transmission cycle [bib_ref] Mosquito vectors and dengue virus-vector relationships, Rodhain [/bib_ref]. Transmission and spread of dengue infection are determined by the interplay of multiple factors including the level of herd immunity in the human population; virulence characteristics of the circulating viral strain; temperature and rainfall; survival, abundance, dispersal and blood feeding behavior of female Ae. aegypti; and human density, age structure, and behavior [bib_ref] Review of the factors modulating dengue transmission, Kuno [/bib_ref]. As a consequence, contact rates between humans and mosquitoes are not random, but highly clustered in space and time [bib_ref] Identification of the people from whom engorged Aedes aegypti took blood meals..., Benedictis [/bib_ref].
The increasing trends in human population growth and urban redistribution that occurred over the past 40 years, coupled with the expansion of commercial trade and the rapid movement of humans (by air travel), have reshaped the global map of dengue transmission risk [bib_ref] Dengue: The risk to developed and developing countries, Monath [/bib_ref] [bib_ref] Geographic expansion of dengue: the impact of international travel, Wilder-Smith [/bib_ref]. Currently, about 70 to 100 million cases of classic dengue infection are reported every year, with an estimated 2.1 million cases of life-threatening disease in the form of Dengue Hemorrhagic Fever (DHF)/Dengue Shock Syndrome (DSS).
Furthermore, the number of dengue fever epidemics has increased dramatically, and an expansion of the dengue endemic and hyperendemic areas is indisputable [bib_ref] Dengue and dengue hemorragic fever: its history and resurgence as a global..., Gubler [/bib_ref] [bib_ref] Dengue: The risk to developed and developing countries, Monath [/bib_ref] [bib_ref] Geographic expansion of dengue: the impact of international travel, Wilder-Smith [/bib_ref]. Under current socioepidemiological scenarios, continued geographic expansion of epidemic dengue is expected to continue (as observed in north and central Argentina, where dengue transmission was registered for the first time in 2009, and severe dengue and DHF outbreaks are expected to follow once mixing of multiple serotypes occurs. Dengue epidemics in such areas are commonly originated by viremic travelers from endemic regions [bib_ref] Geographic expansion of dengue: the impact of international travel, Wilder-Smith [/bib_ref] , and can cover large areas leading to a large number of cases as a consequence of the limited (or null) population exposure to dengue viruses, the prevailing high vector abundances and the challenges faced by local vector control programs on dealing with massive outbreaks.
Consistent with global trends, outbreaks of dengue have become more frequent and severe in Australia, occurring exclusively in north Queensland (NQ) [bib_ref] case recognition and selective indoor residual spraying, Ritchie [/bib_ref] [bib_ref] Two contiguous outbreaks of dengue type 2 in north Queensland, Hanna [/bib_ref] [bib_ref] An epidemic of dengue 3 in far north Queensland, 1997-1999, Hanna [/bib_ref] [bib_ref] Multiple outbreaks of dengue serotype 2 in north Queensland, Hanna [/bib_ref] [bib_ref] Infection and dissemination of dengue virus type 2 in Aedes aegypti, Aedes..., Moore [/bib_ref] [bib_ref] Outbreaks of dengue in north Queensland, Hanna [/bib_ref]. Originated by viral introduction via infected travelers from endemic regions, dengue outbreaks in this region are characterized by a rapid spread both in time and space, as a result of the prevailing high Ae. aegypti populations [bib_ref] Entomological investigations in a focus of dengue transmission in Cairns, Queensland, Australia,..., Ritchie [/bib_ref] and the movement of residents and tourists within and between urban centers [bib_ref] Multiple outbreaks of dengue serotype 2 in north Queensland, Hanna [/bib_ref]. The recent occurrence of two major and widespread epidemics in 2003 and 2008-2009 has challenged local health authorities with the question of whether NQ will join the growing list of dengue endemic regions.
In Queensland, the Tropical Population Health Unit (TPHU) dependent of Queensland Health is the institution responsible for regular vector and viral surveillance, vector control, outbreak response, public education, and operational research [bib_ref] case recognition and selective indoor residual spraying, Ritchie [/bib_ref] [bib_ref] Dengue Fever Management Plan for North Queensland, Health [/bib_ref]. In late February 2003, TPHU was notified of three locally acquired dengue fever cases in a small industrial area in the city of Cairns [bib_ref] Multiple outbreaks of dengue serotype 2 in north Queensland, Hanna [/bib_ref]. Mosquito-control measures were implemented immediately and local doctors were alerted. These cases were quickly confirmed as dengue virus serotype 2 (DENV-2) [bib_ref] Multiple outbreaks of dengue serotype 2 in north Queensland, Hanna [/bib_ref]. Vectorcontrol measures began in the affected Cairns neighborhoods following the Dengue Fever Management Plan for north Queensland (Supporting Information S1), and extended until the outbreak was declared over, five-and-a-half months and 459 cases (383 of which occurred in the city of Cairns) after the primary introduction [bib_ref] Multiple outbreaks of dengue serotype 2 in north Queensland, Hanna [/bib_ref]. In following up these cases, it was learned that a PNG national introduced the virus into Cairns in late January soon after arriving from PNG into a neighborhood heavily infested by Ae. aegypti [bib_ref] Entomological investigations in a focus of dengue transmission in Cairns, Queensland, Australia,..., Ritchie [/bib_ref]. In support of such case as the epidemic's putative index case are the following: a) dengue is not endemic in Cairns, indeed, there was no confirmed dengue activity in urban Cairns before the IC arrival into Cairns from PNG; b) once transmission was confirmed in the area, the sera of the putative case (originally misdiagnosed as malaria) was retrospectively confirmed by PCR as DENV-2 positive, 49 days after its arrival from PNG [bib_ref] Multiple outbreaks of dengue serotype 2 in north Queensland, Hanna [/bib_ref] ; c) the DENV-2 from 2003 had 99.8% homology with an import from PNG into the nearby city of Townsville in April 2003 [bib_ref] Multiple outbreaks of dengue serotype 2 in north Queensland, Hanna [/bib_ref] , indicating PNG as the most likely origin of the DENV-2 introduction; d) most of the early cases occurred around the putative index case's residence.
Dengue infection spread in naive populations occurs in an explosive and widespread fashion [bib_ref] Two clustering diffusion patterns identified from the 2001-2003 dengue epidemic, Kan [/bib_ref] [bib_ref] Dengue fever in a Puerto Rican community, Neff [/bib_ref] [bib_ref] Exploratory space-time analysis of reported dengue cases during an outbreak in Florida, Morrison [/bib_ref] [bib_ref] Space-time analysis of the dengue spreading dynamics in the 2004 Tartagal outbreak,..., Rotela [/bib_ref] as a consequence of the combination of the lack of herd immunity, the population dynamics and dispersal of Ae. aegypti and the movement of individuals within the urban space [bib_ref] Multiple outbreaks of dengue serotype 2 in north Queensland, Hanna [/bib_ref] [bib_ref] Usefulness of commercially available GPS data-loggers for tracking human movement and exposure..., Vazquez-Prokopec [/bib_ref]. Knowledge on the relative contribution of such factors to the pattern of virus propagation during epidemic events has been limited. As a consequence, the spatial dimension of dengue virus spread within complex urban environments is unknown. Although a few studies had quantified the spatial pattern of dengue epidemic transmission in urban environments [bib_ref] Two clustering diffusion patterns identified from the 2001-2003 dengue epidemic, Kan [/bib_ref] [bib_ref] Exploratory space-time analysis of reported dengue cases during an outbreak in Florida, Morrison [/bib_ref] [bib_ref] Space-time analysis of the dengue spreading dynamics in the 2004 Tartagal outbreak,..., Rotela [/bib_ref] [bib_ref] Spatio-temporal tracking and phylodynamics of an urban dengue 3 outbreak in Sao..., Mondini [/bib_ref] [bib_ref] Dengue spatial and temporal patterns, Tran [/bib_ref] , difficulties in assessing where the virus has been first introduced had limited their description of the pattern of dengue infection spread. In the present study we analyzed the pattern of spread of the DENV-2 outbreak that affected the city of Cairns (NQ) in early 2003, quantified the relationship between dengue spread and the location of the epidemic's index case (IC), derived a dispersal kernel for virus spread, assessed the effects of vector control in the containment of the infection, and generated recommendations for city-wide improvement of dengue surveillance and control.
# Materials and methods
## Study area
The city of Cairns (total 2006 population: 140,347) is located in the wet tropics of northeastern Queensland, Australia (16.9u S; 145.8u E). Cairns has a tropical monsoonal climate, with respective mean daily low and high temperatures of 18uC and 26uC in winter and 24uC and 31uC in summer, and most of the rainfall (ca. 82% of the annual 1,992 mm) falling during January-April (Australian Bureau of Meteorology). Residential areas consist of a mixed housing type, with relative small 1-3 story apartment blocks interspersed with single family houses. Housing ranges from modern brick, concrete block and stucco structures to wooden houses that are 50-100 years old. These old wooden 'Queenslander' houses are elevated on wooden or concrete poles and typically feature unscreened windows to maximize air flow. Older suburbs of Cairns (i.e., Parramatta Park, Manuda, Cairns North, Edge Hill) contain mostly 'Queenslander' houses, and often are surrounded by dense tropical vegetation. Cairns also is surrounded by a series of small isolated 'beach communities' that abut the Coral Sea. The present study focused on dengue transmission dynamics in urban Cairns; beach communities and satellite towns were excluded because of their low (or absence of) dengue transmission [bib_ref] Outbreaks of dengue in north Queensland, Hanna [/bib_ref].
## Data sources and management
De-identified data with age, sex, date of onset of infection, and geographic position of the most likely place of dengue infection for each of the 383 laboratory-confirmed human dengue cases that occurred within urban Cairns in 2003 were provided by TPHU. Additional information given by TPHU included: location and
## Author summary
Global trends in population growth and human redistribution and movement have reshaped the map of dengue transmission risk, exposing a significant proportion of the world's population to the threat of dengue epidemics. Knowledge on the relative contribution of vector and human movement to the widespread and explosive nature of dengue epidemic spread within an urban environment is limited. By analyzing a very detailed dataset of a dengue epidemic that affected the Australian city of Cairns we performed a comprehensive quantification of the spatiotemporal dimensions of dengue virus epidemic transmission and propagation within a complex urban environment. Space and space-time analysis and models allowed derivation of detailed information on the pattern of introduction and epidemic spread of dengue infection within the urban space. We foresee that some of the results and recommendations derived from our study may also be applicable to many other areas currently affected or potentially subject to dengue epidemics. [fig_ref] Table 1: Adjusted dengue incidence rates per age class for the city of Cairns... [/fig_ref] , Dulles, VA) of the city of Cairns. All geographic layers were processed in a Geographic Information System (ArcGIS 9.3, ESRI, Redlands, CA) in order to hide their absolute location while preserving the relative distance between them (to protect each patient's identity).
Dengue is a notifiable diseae in Australia [bib_ref] Dengue Fever Management Plan for North Queensland, Health [/bib_ref] [bib_ref] The timeliness of notification of clinically suspected cases of dengue imported into..., Malcolm [/bib_ref]. Upon suspicion of a dengue case, medical practitioners from private clinics and the Cairns hospital are required to contact TPHU to provide details about each case (name, address, phone numbers and specific symptoms; see [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref] in Supporting Information S1 for a sample notification form). Also, all pathology laboratories in Queensland are required to promptly notify public health authorities of any laboratory result indicative of a recent dengue infection (i.e., positive dengue IGM titers or virus detection by PCR). Within 24 hours of receipt of a notification of a suspicious or a laboratory confirmed case highly trained TPHU public health nurses perform contact tracing telephonic interviews to determine a patient's travel history and to identify the origin of infection (i.e., imported or locally acquired dengue), the date of onset of infection (i.e., date of onset of symptoms minus the intrinsic incubation period of 4-7 days, the locations a patient visited while viremic and, ultimately, the most likely place where infection has occurred [bib_ref] The timeliness of notification of clinically suspected cases of dengue imported into..., Malcolm [/bib_ref]. Refer to [fig_ref] Figure 2: Spatial progression of the dengue epidemic that affected the city of Cairns,... [/fig_ref] in Supporting Information S1 for a sample dengue case report form used by TPHU nurses to obtain information on each case and to infer the most likely place of infection.
The most likely place of infection of each case was determined by epi-linking the places visited during the 4-7 days before the onset of illness. The locations were based upon a) home address, b) work address, or c) places and people who they have visited during the exposure period [fig_ref] Figure 2: Spatial progression of the dengue epidemic that affected the city of Cairns,... [/fig_ref] in Supporting Information S1). Patients were asked for addresses where they had been bitten by mosquitoes during the exposure period. Any addresses with recent (,2-4 weeks) dengue activity were considered a likely place of infection. The selection of the most likely place of infection for cases epi-linked to more than one premise was based on the amount of daytime spent by a person on each location (i.e., the premise in which a patient spent most of his/her daytime was selected as the exposure site). The information of the most likely place of infection for cases without any epidemiological link to transmission areas (based on the phone interviews) was updated if new information, such as the confirmation of previously unknown dengue cases within 2 weeks of the onset of symptoms of the initial case, was obtained. Interviews were also performed with the persons each patient identified as potential primary or secondary contacts (e.g., work mates, relatives, friends) to pre-empt the detection of further secondary infections. Based on estimates of a recent (2008-2009) epidemic, TPHU nurses generally interview approximately 3 potential cases per every confirmed case (D. Brookes, Queensland Health, personal communication). Given that TPHU only kept digital records of the most likely places of dengue infection for each case (after the change from paper to digital records, TPHU discarded all phone interview paper records) in this study we were unable to analyze the accuracy of nurses in identifying them. While we acknowledge that some of the dengue acquisition addresses may have been incorrectly identified during the interviewing process (due to patients' recall bias), we believe that most are accurate and that the general spatial nature of the outbreak is not radically perturbed by the misidentification of a small number of acquisition addresses.
Upon confirmation of an address where a patient spent time while viremic, TPHU initiated vector control activities as described in Supporting Information S1. The date and address for every IRS/source-reduction intervention were recorded in a geodatabase linked to the Cairns GIS. In the present study we only analyzed data on indoor residual spraying (IRS) because, when properly applied, it represents an effective method to supress Ae. aegypti indoor populations.
We estimated the age-adjusted incidence of infection for the 2003 epidemic by applying the direct method [bib_ref] Epidemiology, biostatistics, and preventive medicine, Jekel [/bib_ref]. This method adjusts the amount that each age group contributes to the overall rate of transmission, so that the overall rates are based on the same age structure [bib_ref] Epidemiology, biostatistics, and preventive medicine, Jekel [/bib_ref]. Such adjustment was calculated by first multiplying the age-specific disease rates by age-specific weights (i.e., the proportion of the standard 2000 Queensland population within each age group) and then summing the weighted rates across all age groups to give the overall age-adjusted dengue incidence rate for the 2003 epidemic. In this way, we eliminated possible interferences in the estimates due to heterogeneous demographic structure of the studied population.
# Data analysis
Classic, spatial, and Bayesian statistical analyses were performed to describe and quantify the spread of dengue in the 2003 Cairns epidemic. A maximum likelihood logistic regression [bib_ref] Epidemiology, biostatistics, and preventive medicine, Jekel [/bib_ref] was employed to assess the impact of IRS on the odds of secondary dengue infections at the house level. Particular attention was paid to the timing of spray versus the occurrence of cases in a house, since variation in the incubation period of dengue (,7 days when accounting for intrinsic incubation period) can affect the predicted relationship between spraying and dengue cases (by having cases infected before the spraying but showing symptoms after the spraying). Cross-correlation time series analysisbetween the weekly number of cases and the weekly number of IRS applications, weekly precipitation and average temperature were performed to assess the relationships between the temporal progression of the epidemic and selected environmental factors. Meteorological data for Cairns was obtained from the Australian Bureau of Meteorology (source station: Cairns International Airport, located ,4 km from the introduced case's residence).
Spatial analyses were performed to quantify the intensity (Local K-function) and directionality (angular wavelet analysis) of the spatial clustering of dengue cases around the introduced case (IC), and space-time analysis (Knox test) were applied to identify independent space-time clusters of dengue transmission during the 25 weeks of the outbreak. The Local K-Function (L i (d)) developed by Getis and Frankling [bib_ref] 2nd-order neighborhood analysis of mapped point patterns, Getis [/bib_ref] was applied to determine the overall distance up to which cases clustered around the IC. In our focal analysis of L i (d), i is represented by the location of the IC. Statistical inference of L i (d) is performed by comparing the observed L i (d) with the expected L i (d) generated by 999 Monte Carlo realizations under the hypothesis of Complete Spatial Randomness [bib_ref] 2nd-order neighborhood analysis of mapped point patterns, Getis [/bib_ref]. A focal spatial correlogram of L i over d was used to assess the distance up to which clustering of dengue cases around the IC was maximized (d max ). Briefly, d max represents the distance at which the value of L i is maximum. d max is interpreted as the distance beyond which the inclusion of further cases does not increase the probability that the distribution of events differ from random [bib_ref] 2nd-order neighborhood analysis of mapped point patterns, Getis [/bib_ref].
To determine if the distribution of cases around the IC occurred predominantly in a given cardinal direction (i.e., anisotropic distribution) we divided the space around the IC's house into one degree sectors, and counted the number of cases that fell within each sector. These counts were then analyzed by angular wavelet analysis [bib_ref] Wavelet analysis for detecting anisotropy in point patterns, Rosenberg [/bib_ref]. A wavelet function g(x) is a scalable windowing function. In our study we used the French Top Hat [bib_ref] Wavelet analysis for detecting anisotropy in point patterns, Rosenberg [/bib_ref] as a wavelet function. The main metric derived from fitting the wavelet function to the data is the wavelet positional variance. Peaks in this variance indicate directions where most of the cases fell relative to the IC. In order to separate true patterns from random fluctuations, the significance of the wavelet analysis was determined by comparing the observed variance with the one obtained from 999 Monte Carlo simulations [bib_ref] Wavelet analysis for detecting anisotropy in point patterns, Rosenberg [/bib_ref]. The analysis was performed for both the location of cases and the location of Cairns houses to determine if any anisotropic pattern in the distribution of cases could have been an artifact of the spatial arrangement of houses within the city. Visualization of spatial anisotropy in the occurrence of cases was performed by estimating spatial standard-deviation ellipses in ArcGIS 9.3 (ESRI).
The Knox method [bib_ref] The detection of space-time interactions, Knox [/bib_ref] was applied to quantify the space-time interaction of individual confirmed dengue cases reported during the outbreak. This method tests for possible interaction between the distance and time separating cases, based on the number of case pairs found in a particular time-space window (e.g., case pairs separated by less than M meters and T days) [bib_ref] The detection of space-time interactions, Knox [/bib_ref]. When interaction is present, distances between pairs of cases will be small, and the test statistic will be large. In our study we chose the values of M and T as 100 m and 20 days, respectively, to account for mosquito dispersal distance (i.e., Ae. aegypti seldom disperses beyond 100 m [bib_ref] Mark-release-recapture study to measure dispersal of the mosquito Aedes aegypti in Cairns, Russell [/bib_ref] [bib_ref] Dispersal of the dengue vector Aedes aegypti within and between rural communities, Harrington [/bib_ref] and virus incubation periods (i.e., maximum sum of intrinsic and extrinsic incubation periods).
The expected values of the test under the null hypothesis of random case occurrence (in space and time) were estimated by performing 999 Monte Carlo simulations.
A semiparametric Bayesian space-time structured additive regression model (STAR) [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref] [bib_ref] Structured additive regression for categorical spacetime data: A mixed model approach, Kneib [/bib_ref] was applied to assess the effects of IRS, rain and spatial autocorrelation in the odds of weekly dengue infection over the first 15 weeks of the epidemic. Since our analysis focused on weekly data we selected a discrete-time duration STAR model [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref]. Such model allows a unified treatment of time scales, linear and non-linear effects of covariates and spatially correlated random effects within a Bayesian framework [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref]. Briefly (refer to [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref] for a detailed description), a space-time geoaditive STAR model has the following predictor structure:
n it~f1 x it 1 z::::
[formula] :zf k x it k zf time t ð Þzf spat s i t À Á zu 0 it c,ð1Þ [/formula]
where u9 it c are the usual linear predictors (fixed effects) for covariate vector u, f 1..k are possibly non-linear functions of the covariates, f time is a possibly non-linear time trend and f spat is a spatially correlated (random) effect of the location s it an observation pertains to [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref]. In a Bayesian approach, all non-linear functions f j and parameters c i are considered as random variables and (1) is therefore conditional upon these random variables, having to be supplemented with appropriate prior distributions. Fixed effects were modeled by independent diffuse priors (p(c j ),const, j = 1, …, r), whereas the priors for unknown functions (f j ) were dependent on the type of covariate and on the prior knowledge about smoothness. The fixed effects (f 1..k ) and the time function (f time ) were modeled via Bayesian penalized splines (P-splines), whereas Markov random field priors were chosen as priors for the spatial effects (f spat ) [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref]. Posterior means together with confidence intervals and other parameters are obtained by drawing samples from the posterior by Markov chain Monte Carlo (MCMC) techniques [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref]. Variance parameters (t j 2 ) can be estimated by assigning additional hyperpriors to them. The most common assumption is that t j 2 are independently inverse gamma distributed, i.e., t j 2 ,IG(a j ,b j ), with hyperparameters a j and b j a priori set as a j = b j = 0.001. For updating the parameters in a MCMC sampler, a Metropolis-Hastings algorithm based on iteratively weighted least squares was applied [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref]. We performed a sensitivity analysis of the best fitted model by varying the values of a j and b j between 1 and 0.0001 as described in [bib_ref] Structured additive regression for categorical spacetime data: A mixed model approach, Kneib [/bib_ref].
Our STAR model evaluated the effect of rain, the cumulative proportion of sprays performed around a premise (cum_spr) and the spatial arrangement of premises (spat) on the odds of weekly dengue virus infection (binomial variable, 0 = no infection, 1 = infection) at the house level from weeks 0 to 15 post introduction. Analysis focused in the neighborhoods surrounding the IC's house (Parramatta Park, Manuda and Cairns North), where most of the cases and IRS applications occurred. Thyssen polygons from the centroids of all the premises within the analyzed area were generated to allow contiguity weighting schemes for the spatial random effects. Our analysis matrices had a dimension of 1,490 premises by 15 weeks (or 22,350 repeated measurements), and the structure of the evaluated models was as follows:
[formula] Logit case,0j1 ð Þf time Week ð ÞzRain fixed ð Þzf 1 cumm spr ð Þ zf spat zf 2 cumm spr à spat ð Þð2Þ [/formula]
and
[formula] Logit case,0j1 ð Þf time Week ð Þzf 1 cumm spr ð Þ zf spat zf 2 cumm spr à spat ð Þð3Þ [/formula]
We compared both models in their Deviance Information Criterion (DIC, the AIC equivalent for Bayesian models [bib_ref] Penalized structured additive regression for space-time data: A Bayesian perspective, Fahrmeir [/bib_ref] after 100,000 MCMC randomizations, and only reported the results of the model with the lowest DIC. Spatial and spatio-temporal analysis were performed with ClusterSeer 2.0 (Terraseer, Ann Arbor, MI), R 2.10.1 (http:// www.r-project.org/) and ArcGIS 9.3 (ESRI) software, whereas Bayesian STARS were performed with BayesX software (http:// www.stat.uni-muenchen.de/,bayesx) and model outputs visualized with R package BayesX (http://cran.r-project.org/web/ packages/BayesX/).
## Scientific and ethical review and approval
The protocols for storing, analyzing and reporting results on the 2003 Cairns dengue epidemic data were approved by Queensland Health's Human Research Ethics Committee (protocol HREC/ 09/QCH/52-590).
# Results
## Epidemic progression and control
Local transmission of dengue in Parramatta Park began 18 days after the onset of symptoms in the imported case (21 January). However, the TPHU was not notified of dengue activity in the area until 5 March (,43 days post-introduction) and initiated mosquito control measures the next day. It took ,49 days to retrospectively confirm that the traveler from PNG (initially misdiagnosed with malaria) was indeed the epidemic's IC. A total of 383 laboratory-confirmed symptomatic cases were registered within urban Cairns over the 25-week epidemic period that followed the initial introduction. All infections were confirmed as mild (only 5% of all cases were hospitalized) DENV-2 derived from the initial introduction; no dengue-related deaths or DHF Quantifying Dengue Virus Epidemic Spread www.plosntds.org manifestations were reported. The index case (week 0) and subsequent first wave of local transmission (weeks 2-3) are clearly distinguished by the epidemic curve [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref]. The weekly number of confirmed dengue cases then showed an exponential growth from one to 56 cases from weeks 4 to 7 post introduction (PI), followed by a stable period ranging from 44 to 47 cases during weeks 8-10 PI, and a final exponential decay thereafter [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref]. The weekly pattern of IRS applications followed the pattern of human cases, totaling 1,163 applications over a 19 week period [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref]. The overall age-adjusted incidence rate of the 2003 Cairns epidemic was 1,148 cases per 100,000 [fig_ref] Table 1: Adjusted dengue incidence rates per age class for the city of Cairns... [/fig_ref]. The highest incidence rates were observed in males 30-39 year old. Overall, young adults (30-39 years old) presented the highest incidence rates (average [SD], 258.5 and 193.6 cases per 100,000 for men and woman, respectively), but not in the younger or older ages [fig_ref] Table 1: Adjusted dengue incidence rates per age class for the city of Cairns... [/fig_ref]. The weekly number of cases was strongly and positively correlated (Cross-correlation coefficient, r.0.6) with the number of IRS applications up to a time lag of 2 weeks [fig_ref] Figure 3: Directionality analysis for the location of dengue cases within d max [/fig_ref] in Supporting Information S1). Such positive correlation was consequence of the strong and active response of TPHU to the DENV-2 epidemic once DENV-2 transmission was confirmed [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref]. The weekly variation in temperature (mean, minimum and maximum) and total precipitation during the 25-week period of the outbreak [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref] did not show any strong correlation (r,0.5) with the weekly number of cases [fig_ref] Figure 3: Directionality analysis for the location of dengue cases within d max [/fig_ref] in Supporting Information S1). However, the lack of rain during weeks 8-11 [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref] coupled with the high intensity of IRS applications during that period could have acted synergistically against local Ae. aegypti populations and help explain the sharp reduction and further interruption of dengue transmission after week 11 PI [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref]. [fig_ref] Figure 2: Spatial progression of the dengue epidemic that affected the city of Cairns,... [/fig_ref] (and Video S1) depicts the spatio-temporal pattern of dengue spread in Cairns during the 25-week transmission period. Up to week 5 PI, most transmission was contained around the IC (all but one confirmed case occurred within 100 m of the IC). Within the first 3 weeks PI, two additional cases occurred in the same house of the IC. Between weeks 6 and 10 PI, transmission continued propagating around the IC, but also expanded to other neighborhoods located further than 100 m of the IC. An increase in the distance of new cases to the IC together with a reduction in the number of confirmed cases were evident by week 15 PI. Between weeks 15 and 25 PI only a few isolated cases occurred, most of them located in the periphery of urban Cairns. The last case was contracted on week 25, and the epidemic was formally declared over three months later.
## Spatio-temporal pattern of spread
Focal clustering of DENV-2 cases was maximized at a distance of 800 m from the IC (L i (d max ) = 1,742.3; P,0.05; [fig_ref] Figure 4: Significant space-time clustering [/fig_ref] in Supporting Information S1). A total of 240 cases (63%) were located within d max , mainly in the neighborhoods Parramatta Park (PP) and Cairns North (CN) [fig_ref] Figure 4: Significant space-time clustering [/fig_ref] in Supporting Information S1). The spatial wavelet directionality analysis performed within d max showed a significant peak (Spatial Variance .1.96; P,0.05) in the orientation of cases around the IC towards the NW-SE direction [fig_ref] Figure 3: Directionality analysis for the location of dengue cases within d max [/fig_ref]. A significant peak in the distribution of houses around the IC towards the same NW-SE direction [fig_ref] Figure 3: Directionality analysis for the location of dengue cases within d max [/fig_ref] points to the orientation of the built environment (blocks were aligned NW to SE and elongated at a ratio of ca. 36 longer than wide) as the most likely cause of the anisotropic distribution of cases around the IC. The dominant wind direction during the first 2 months from the initial epidemic spread from the index case (Feb.-March) was from the SE quadrant (ie., 60-180u; [fig_ref] Figure 3: Directionality analysis for the location of dengue cases within d max [/fig_ref] , with 80% and 44% of winds at 9 AM and 3 PM from such quadrant, respectively. Such prevailing early morning and late afternoon winds could have affected dispersal of infected Ae. aegypti females, also contributing to the SE/NW oriented anisotropic distribution of dengue infections around the IC [fig_ref] Figure 3: Directionality analysis for the location of dengue cases within d max [/fig_ref].
The space-time Knox test statistic identified a total of 18 independent significant space-time clusters involving 250 (65.3%) cases [fig_ref] Figure 4: Significant space-time clustering [/fig_ref]. The remaining 133 (34.7%) cases did not show any spatio-temporal association among them or with members of the space-time clusters. [fig_ref] Table 2: Description of each space-time cluster identified for the dengue epidemic that affected... [/fig_ref] summarizes the characteristics of each identified space-time cluster, together with the average distance of each case to the original IC and the putative index case (PIdC) of each cluster, and [fig_ref] Figure 5: Using human infection as a marker of dengue virus epidemic spread [/fig_ref] in Supporting Information S1 shows the weekly distribution of the proportion of cases within each space-time cluster. The earliest and largest cluster (cluster 1) occurred around the IC, and included 129 cases (including the IC) extending over 440 m during a 13-week period [fig_ref] Figure 4: Significant space-time clustering [/fig_ref] , [fig_ref] Table 2: Description of each space-time cluster identified for the dengue epidemic that affected... [/fig_ref]. The second cluster was initiated 37 days after the onset of symptoms of the IC, and was located at an average distance of 679 m from the IC [fig_ref] Figure 4: Significant space-time clustering [/fig_ref] , [fig_ref] Table 2: Description of each space-time cluster identified for the dengue epidemic that affected... [/fig_ref]. The average distance between every cluster's case and the case that first showed symptoms (i.e., the cluster's putative index case, PIdC) was 75 m (SD: 51.2 m) whereas the average distance between cases within a cluster was 73.2 m (SD, 59.9 m) [fig_ref] Table 2: Description of each space-time cluster identified for the dengue epidemic that affected... [/fig_ref]. The spraying response (measured in days since the onset of symptoms of a cluster's first reported case) in each cluster averaged 15.3 days (range, 0-40 days) whereas the spraying coverage averaged 36.8% (range, 0-100%) for all the premises with cases and 23% (range, 0-47.1%) for all the premises found within 100 m of a premise with a confirmed case [fig_ref] Table 2: Description of each space-time cluster identified for the dengue epidemic that affected... [/fig_ref].
For each space-time cluster we identified its first confirmed case and estimated the spatial and time distances from each secondary case in the cluster to this PIdC. PIdC can be interpreted one of the most likely origins of a space-time cluster. We then considered each cluster as a replicate to assess the relationship between distance and time since the onset of symptoms of a PIdC. The distance up to which most of the secondary cases were found increased with the time since the onset of symptoms of a PIdC from 50 m to 200 m and 325 m for weeks 1-3, 4-6 and 7-8, respectively [fig_ref] Figure 5: Using human infection as a marker of dengue virus epidemic spread [/fig_ref]. Most of the cases whose onset of symptoms was within 3 weeks of the onset of symptoms of a PIdC were found within 50 m of it [fig_ref] Figure 5: Using human infection as a marker of dengue virus epidemic spread [/fig_ref]. The mean distance from a secondary case to a PIdC [fig_ref] Figure 5: Using human infection as a marker of dengue virus epidemic spread [/fig_ref]. This distance increased to 200 m by week 3. [fig_ref] Figure 5: Using human infection as a marker of dengue virus epidemic spread [/fig_ref] can be interpreted as a kernel of dengue diffusion from a PIdC after filtering out the effect of long distance human movement.
## Effect of irs on dengue infection
The odds of a secondary dengue infection at premises with confirmed dengue cases was significantly higher at unsprayed premises than at a sprayed premises (OR = 2.8; 95% CI = 1.1-6.9; P = 0.03). From the 151 unsprayed premises with confirmed dengue cases, 36 (23.8%) reported subsequent dengue cases at an average of 7 days (SD = 10.6 days; Max = 50 days) of the onset of symptoms of the first case in the premise. Whereas from the 97 sprayed premises with confirmed dengue cases, 13 (13.4%) reported subsequent dengue cases after the onset of symptoms of the first case in the premise. From such estimate are excluded 20 confirmed cases exposed to infective bites in sprayed premises within 7 days of an insecticide spray, and most likely infected before IRS applications. The main causes of lack of IRS applications were residents' refusal to grant access to the DART team due to personal matters or to aversion to pesticides, residents' absence at the time of DART visitation, and occurrence of locked entrance gates or dangerous dogs at a given premise. Bioassays using WHO cones on lambda-cyhalothrin-treated wood (SA Ritchie, unpublished data) together with an observed 90% reduction of gravid female collections in sticky ovitraps within a month of IRS applications in Parramatta Park [bib_ref] Entomological investigations in a focus of dengue transmission in Cairns, Queensland, Australia,..., Ritchie [/bib_ref] support the lack of resistance to lambda-cyhalothrin in local Ae aegypti populations.
DIC values for STAR models (2) and (3) were 2,036 and 2,025, respectively. Hence, a model without rain as a fixed effect (3) was selected as the best descriptor of the weekly spatial pattern of dengue infection. All variables in the model had an acceptance rate of 65% or more (not shown). The posterior mean (with 80% credibility intervals) of the effect of week on the odds of dengue infection is shown in [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref]. The odds of infection was negative from weeks 0-4 (the ''silent'' period before the epidemic was identified), positive from weeks 5-13 (the period of ''active'' transmission) and negative onwards (the period of effective control) [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref] , resembling the shape of the epidemic curve [fig_ref] Figure 1: The 2003 Cairns epidemic in numbers [/fig_ref]. The effect of the spatial occurrence of cases followed the observed space-time clustering of cases, with the highest posterior mean values (dark red in [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref] found around the IC's residence and the remaining areas of medium positive (orange in [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref] effect found in the periphery. The cumulative proportion of sprays around a premise had a non-linear relationship with the posterior odds of dengue infection [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref]. When spraying coverage around a premise was less than 40-60%, the odds of dengue infection was positive (IRS did not prevent further cases around a premise). Only when 60% or more of the premises around a house were sprayed with insecticides the odds of infection was significantly reduced to levels below 0, and IRS applications had a protective effect in the further occurrence of dengue cases [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref]. The spatial occurrence of IRS applications [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref] was similar to the distribution of cases [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref] , with most of the applications in the vicinity of the IC's residence and less applications in the periphery of the introduction point. The sensitivity analysis performed by changing the hyperpriors (a and b) did not show any deviation from the basic model with default values [fig_ref] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression [/fig_ref] in Supporting Information S1).
# Discussion
Dengue vector control failures are partly due to our deficiencies in understanding relationships among available interventions, virus transmission dynamics, and human behavior. In areas Quantifying Dengue Virus Epidemic Spread www.plosntds.org sporadically affected by epidemic dengue, such as northern Queensland, control interventions are generally implemented in response of the occurrence of local transmission, generally when it is too late to rapidly contain virus propagation [bib_ref] Dengue fever in a Puerto Rican community, Neff [/bib_ref] [bib_ref] Exploratory space-time analysis of reported dengue cases during an outbreak in Florida, Morrison [/bib_ref] [bib_ref] Space-time analysis of the dengue spreading dynamics in the 2004 Tartagal outbreak,..., Rotela [/bib_ref] [bib_ref] Spatio-temporal tracking and phylodynamics of an urban dengue 3 outbreak in Sao..., Mondini [/bib_ref]. By applying sound statistical analysis to a very detailed dataset from one of the largest outbreaks that affected the city of Cairns in recent times, we not only described the spread of DENV-2 with high detail but also derived important knowledge that will contribute to the understanding of how dengue virus infection propagates epidemically within complex urban environments.
Current patterns of human movement by air travel have increased the probability of introduction of dengue into Ae. aegypti infested areas [bib_ref] Geographic expansion of dengue: the impact of international travel, Wilder-Smith [/bib_ref]. However, it is recognized that only a handful of such introductions will have the potential of becoming IC's of a dengue epidemic. Therefore, the likelihood of a successful dengue introduction will depend on the juxtaposition of various events, such as the time a person remains viremic in the destination area (t v ), the number of sites he or she visited while viremic (l v ), and the number of Ae. aegypti bites he or she received across l v during t v. Lastly, the likelihood of secondary transmission in the same area will be dependent on the exposure of local residents to infective bites from the Ae. aegypti females found across l v (after accounting for extrinsic incubation period and vector dispersal). All such events are significantly modulated by the impacts of vector control interventions and the build-up of population herd immunity to DENV infection occurring both in space and time. Hence, the complex pattern of dengue introduction, establishment and further propagation relies on a complex repertoire of events that, altogether, determine the spatial dimension of virus epidemic spread.
Sequential transmission (i.e., the progressive occurrence of human cases in neighboring houses), likely attributable to mosquito-driven spread and/or short-distance mobility of viremic humans, as well as long distance propagation of infection (likely generated by human mobility) have been documented in many urban dengue epidemics [bib_ref] Dengue fever in a Puerto Rican community, Neff [/bib_ref] [bib_ref] Exploratory space-time analysis of reported dengue cases during an outbreak in Florida, Morrison [/bib_ref] [bib_ref] Possible dengue sequential infection: dengue spread in a neighbourhood during the 1996/97..., Deparis [/bib_ref]. Neff et al [bib_ref] Dengue fever in a Puerto Rican community, Neff [/bib_ref] elegantly showed how, after the introduction of dengue in a Puerto Rican village, subsequent cases occurred ''sequentially'' within the same block and also in distant parts of the village. Similarly, phylogenetic analysis of a DENV-3 epidemic affecting the city of Sao Paulo (Brazil) allowed Mondini et al. [bib_ref] Spatio-temporal tracking and phylodynamics of an urban dengue 3 outbreak in Sao..., Mondini [/bib_ref] to estimate the most likely route of viral dispersal, showing that a same lineage was ''dispersed'' within the city both at short and long distances compatible with mosquito-and human-mediated virus spread, [fig_ref] Table 2: Description of each space-time cluster identified for the dengue epidemic that affected... [/fig_ref]. doi:10.1371/journal.pntd.0000920.g004
Quantifying Dengue Virus Epidemic Spread www.plosntds.org respectively. Our detailed analysis differentiated with great detail the contribution of short and long distance propagation of dengue virus infection within Cairns. We identified the location of 18 transmission clusters within the city that, given their space-time separation, were most likely originated by the movement of viremic humans within the city. Such clusters, although different in their extent and duration of dengue infection presence, shared a similar pattern of virus propagation from their PIdC. Outdoor sticky ovitrap collections performed in PP during the initial weeks of the epidemic evidenced high dengue virus infection rates in Ae. aegypti (up to 116 per 1,000), validating the occurrence of vectormediated virus propagation within clusters. Furthermore, as most human DENV-2 infections were mild, visitation of viremic patients to their near neighbors can also help explain the rapid propagation of dengue infection within each cluster.
Build-up of immunity in the human population (a.k.a. herd immunity effect) has been postulated as an important driver in the dynamics of dengue virus infection, mainly by modulating humanmosquito infective contacts [bib_ref] Review of the factors modulating dengue transmission, Kuno [/bib_ref]. In our study we were unable to quantitatively assess the contribution of herd effect on the local dynamics of dengue transmission (no information on human population numbers per house were available). However, given the impact and extent of the epidemic, we hypothesize that the local propagation around the IC and not the overall termination of the dengue outbreak were affected by the herd immunity effect. We base this assumption on the following observations: a) given the small and sporadic introductions of dengue, previous immunity to DENV-2 can be considered negligible, and a high population susceptibility assumed; b) a significant proportion of cases occurred in close proximity to the IC, and the ratio of infected to susceptible individuals could be high enough in such area to have a detrimental effect in the local transmission of dengue; and c) city-wide transmission was not severely impacted due to the low ratio of infected individuals to the population at risk (,383/ 140,300); even assuming a 1:10 symptomatic to asymptomatic ratio such proportion would have been lower than 5%.
The 2003 Cairns epidemic had its epicenter in the neighborhood of Parramatta Park (PP) and then progressed to the neighborhoods of Cairns North (CN) and eastern Manuda (MN). Such pattern of introduction and spread was consistent with descriptions of previous Cairns epidemics [bib_ref] Two contiguous outbreaks of dengue type 2 in north Queensland, Hanna [/bib_ref] [bib_ref] An epidemic of dengue 3 in far north Queensland, 1997-1999, Hanna [/bib_ref] [bib_ref] Outbreaks of dengue in north Queensland, Hanna [/bib_ref] , and may be the product of the combination of demographic, environmental and entomologic characteristics unique to such areas. The high density of elevated old 'Queenslander' houses found in PP and CN, together with the presence of abundant tropical vegetation around them, provides optimal conditions for Ae. aegypti development and dispersal [bib_ref] Entomological investigations in a focus of dengue transmission in Cairns, Queensland, Australia,..., Ritchie [/bib_ref] [bib_ref] Mark-release-recapture study to measure dispersal of the mosquito Aedes aegypti in Cairns, Russell [/bib_ref]. As a consequence, such neighborhoods generally have the highest Ae. aegypti populations in Cairns [bib_ref] Two contiguous outbreaks of dengue type 2 in north Queensland, Hanna [/bib_ref] [bib_ref] Entomological investigations in a focus of dengue transmission in Cairns, Queensland, Australia,..., Ritchie [/bib_ref]. Given that most 'Queenslander' houses have unscreened windows, endophagy by Ae. aegypti can increase vector abundance but also vector-mediated virus propagation to neighboring premises (as observed in the 2003 Refer to [fig_ref] Figure 4: Significant space-time clustering [/fig_ref] for a geographic identification of each cluster. [bib_ref] Dengue and dengue hemorragic fever: its history and resurgence as a global..., Gubler [/bib_ref] Represents the number of days between the onset of symptoms of the PIdC and the onset of symptoms of the IC. [bib_ref] Epidemic dengue/dengue hemorrhagic fever as a public health, social and economic problem..., Gubler [/bib_ref] PIdC = putative index case of a cluster. [bib_ref] Epidemic dengue/dengue hemorrhagic fever as a public health, social and economic problem..., Gubler [/bib_ref] The mean distance between all cases belonging to a cluster. [bib_ref] Mosquito vectors and dengue virus-vector relationships, Rodhain [/bib_ref] Percentage of locations with cases that were sprayed with insecticides during the epidemic period.Number of days between the onset of symptoms of the PIdC and the date a house was sprayed within the cluster. *represents lack of sprayings. [bib_ref] Review of the factors modulating dengue transmission, Kuno [/bib_ref] Percentage of houses that were sprayed within 100 meters of all the cases belonging to a cluster. Refer to [fig_ref] Figure 4: Significant space-time clustering [/fig_ref] for a geographic representation of each cluster. doi:10.1371/journal.pntd.0000920.t002
Quantifying Dengue Virus Epidemic Spread www.plosntds.org epidemic with the propagation of infection from the IC's house -a typical 'Queenslander' house -to neighboring houses). Moreover, the high number of backpacker hostels found in such neighborhoods (particularly CN) increases the possibility of a dengue introduction by a viremic traveler in those areas (as observed in the 1997 Cairns epidemic that originated from a backpacker guesthouse located in CN [bib_ref] An epidemic of dengue 3 in far north Queensland, 1997-1999, Hanna [/bib_ref] [bib_ref] Outbreaks of dengue in north Queensland, Hanna [/bib_ref]. Hence, the patterns of introduction and further spread of dengue in Cairns appear to follow a consistent spatial pattern irradiating from a few key neighborhoods with particular conditions. Case detection, vector surveillance and disease management could be highly improved if such spatial heterogeneity in the likelihood of dengue introduction and spread is accounted in their design, particularly in the early stages of an outbreak.
A classic approach to reduce the propagation of infected vectors and suppress dengue transmission during the early stages of an epidemic (i.e., after the detection of active transmission) consists of the targeted implementation of vector control actions within a buffer distance (e.g., 50-100 m) of a confirmed case [bib_ref] Epidemiology of inapparent and symptomatic acute dengue virus infection: a prospective study..., Endy [/bib_ref]. Such approach may be insufficient if virus human or vector propagation of dengue virus occurs beyond 100 m. Mark-release-recapture (MRR) studies have shown that urban Ae. aegypti females seldom disperse beyond 100 m [bib_ref] Dispersal of the dengue vector Aedes aegypti within and between rural communities, Harrington [/bib_ref] [bib_ref] Aedes aegypti (Diptera: Culicidae) movement influenced by availability of oviposition sites, Edman [/bib_ref] [bib_ref] Aedes aegypti survival and dispersal estimated by mark-release-recapture in northern Australia, Muir [/bib_ref] [bib_ref] Analysis of survival of young and old Aedes aegypti (Diptera: Culicidac) from..., Harrington [/bib_ref]. However, given that most mosquitoes are recaptured within 3 weeks of release, MRR studies tend to collect very limited data on the dispersal ability of older females. Thus, the contribution of older, potentially viremic, females to the tail of the dispersal distribution is highly underrepresented. By using robust space-time clustering tests we were able to use human dengue infection as a marker of virus propagation and later estimate the spatial and temporal dimensions of dengue virus spread from a putative index case (PIdC). We acknowledge that human movement may also contribute to the spread of dengue virus beyond the dispersal kernel. Nonetheless, our study showed that most (95%) of the cases associated with a PIdC were within a 200 m radius of it, and that the average spread of dengue infection varied from 14 to 32 m per week. If validated with data from other epidemics and settings, such information could have significant impacts in the design and implementation of case detection and vector control activities, particularly in areas affected by epidemic dengue. A spatio-temporal unit for case detection and control actions could help determine the area around a PIdC that would need to be screened for additional cases or controlled with insecticides or source-reduction in order to halt the propagation of dengue virus infection. Rapid (and effective) response constitutes a key operational premise in such settings [bib_ref] Unforeseen costs of cutting mosquito surveillance budgets, Vazquez-Prokopec [/bib_ref].
Larval control and source reduction represent two of the most effective and widely used strategies to control Ae. aegypti populations. However, in areas sporadically affected by dengue epidemics preventive actions are generally rare, and most vector control activities occur upon the identification of an introduced case or the detection of active local transmission. In such circumstances, adulticiding (i.e., targeted control of adult Ae. aegypti by IRS or space spraying) represents an effective control action to rapidly suppress dengue transmission. Additionally, where dengue hits sporadically it is common that vector control agencies are understaffed, making the response to a dengue outbreak a daunting challenge. For instance, during the 2003 epidemic Cairns Dengue Action Response Team (DART) was composed by only 4 staff members who had to be supplemented by 'volunteers' from other health agencies once it was evident that the number of cases exceeded the capacity of these officers to respond. Under such circumstances, our analysis shows that although IRS had a significant effect in controlling Ae. aegypti and reducing the odds of dengue infection, the spraying coverage around an infected premise had to be maintained at 60% or more in order to prevent subsequent infections. This finding points to the need for a coordinated effort in the design of control interventions, since partial insecticide spraying (,40% coverage) can yield to a low control effectiveness, and a higher chance of virus spread. In Cairns, lack of personnel, the occurrence of closed or conflictive premises or of property owners refusing any control activities in their houses were the main causes of the reduced insecticide coverage found in the vicinity of many dengue-positive premises. The incorporation of database and mapping technologies to track the delivery of control actions and monitor the levels of spraying coverage would prove essential to rapidly and effectively respond to a dengue epidemic [bib_ref] Use of mapping and spatial and space-time modeling approaches in operational control..., Eisen [/bib_ref]. Preventive control in high risk areas, capacity building (particularly in GIS/mapping and spatial analysis), incorporation of trained field technicians from other public health agencies, and integration of scientifically-based and context-sensitive control actions represent key vector control components of an integrated program geared to effectively contain dengue epidemics. Underreporting and asymptomatic infections can represent an important proportion of the total number of dengue cases affecting a city during an epidemic [bib_ref] Review of the factors modulating dengue transmission, Kuno [/bib_ref]. One of the main limitations of our study has been the lack of knowledge of the relative contribution of such ''silent'' infections to the total pool of cases. Dengue infection detection and notification in northern Queensland are high (due to the quality of the health system and the history of dengue introductions) [bib_ref] Dengue Fever Management Plan for North Queensland, Health [/bib_ref] [bib_ref] Usefulness of commercially available GPS data-loggers for tracking human movement and exposure..., Vazquez-Prokopec [/bib_ref]. Hence, underreporting rates are considered low compared to the ones found in other urban centers. Nonetheless, since we analyzed data on laboratory-confirmed cases we can assume them to be the tip of the iceberg in the overall pattern of dengue spread, and that the distribution of underreported cases will follow the one of laboratory-confirmed cases.
On the other hand, the actual contribution of asymptomatic cases to the local spread of dengue is unknown, in part due to the uncertainty about their potential to infect Ae. aegypti mosquitoes while viremic [bib_ref] Mosquito vectors and dengue virus-vector relationships, Rodhain [/bib_ref]. A serologic survey or an active surveillance system could have helped assess the size of infected population (and consequently the effect of herd immunity buildup) as well as the actual attack rate and epidemic potential of dengue virus in the 2003 epidemic.
In Cairns (as in many other settings where dengue transmission is sporadic) the age-adjusted incidence was higher in young adults. Given that this age group is also one of the most mobile segments of the population their infection could have been the product of their movement (and exposure) in transmission ''hot spots''. In our study Quantifying Dengue Virus Epidemic Spread www.plosntds.org we used the most likely place of transmission (instead of the case's residence) as the positional mark of a case. Such choice allowed us to more accurately analyze the propagation of dengue virus infection, since our approach included the potential exposure locations rather than merely the residential address of a case. Although TPHU nurses are highly trained and skilled in contact tracing (some of them have been performing case detection for more than 10 years), the reliability of their assessment is not perfect. However, the lack of original records of all the locations reported by a patient (due to disposal of paper records when THPU migrated to digital records) prevented us to analyze the relationship between dengue infection and the type and number of exposure locations reported and, ultimately, the accuracy of TPHU nurses in identifying transmission locations. The integration into a GIS-based decision support system of all the locations a patient has visited during the previous week to the onset of symptoms will help TPHU nurses improve their assessment. Rapid mapping of all the locations identified by the field nurses not only will allow detect where a patient may have gotten infected, but also assess which locations he or she may have visited while viremic. Such information could improve outbreak response and control of dengue epidemics.
In summary, the 2003 Cairns outbreak exhibited a pattern common to most dengue epidemic areas: virus introduction by a viremic traveler, delayed case identification and confirmation, abundant Ae. aegypti populations, limited (or absent) preventive control measures, delayed initiation of vector control activities once transmission was confirmed, and limited personnel/infrastructure to control a spreading epidemic. All those factors acted together to generate one of the largest outbreaks Cairns has had in recent times [bib_ref] Two contiguous outbreaks of dengue type 2 in north Queensland, Hanna [/bib_ref] , and challenged health authorities to develop a plan to prevent epidemic spread following a point introduction. From our detailed analysis we generated a series of recommendations for TPHU that may help contain future dengue outbreaks in Cairns: a) upon suspicion of a dengue introduction, treat each potential case as dengue and perform containment activities covering all the premises found within a distance pre-spcified according to the time elapsed since the suspected introduction (as outlined in [fig_ref] Figure 5: Using human infection as a marker of dengue virus epidemic spread [/fig_ref] ; b) target surveillance and preventive control in the neighborhoods dominated by older, unscreened housing such as PP and CN, that are most likely to initiate epidemic transmission; c) perform insecticide resistance tests periodically; and d) incorporate GIS and space-time analysis (by training local public health specialists) as the key tools of a decision support system for dengue management in northern Queensland. We foresee that some of the analytical procedures and derived recommendations may also be applicable to other areas currently affected or potentially subject to epidemic dengue.
## Supporting information
Supporting Information S1 Supporting text and figures. Found at: doi:10.1371/journal.pntd.0000920.s001 (2.78 MB DOC)
Video S1 Animated cartographic representation of the spread of dengue virus infection within the city of Cairns, Queensland, during the 2003 epidemic. Each frame represents a week (measured since the onset of symptoms of the introduced case, IC). Red points show the week of onset of symptoms of cases in the week t whereas blue dots the cumulative cases from t0 to t-1. Found at: doi:10.1371/journal.pntd.0000920.s002 (2.89 MB WMV)
[fig] Figure 1: The 2003 Cairns epidemic in numbers. (A) Weekly number of confirmed dengue 2 cases (bars) and of indoor residual insecticide sprays (line) performed in the city of Cairns during January-August 2003, and (B) weekly variation in temperature (mean, minimum and maximum) and total precipitation over the same period. Time is measured in weeks since the onset of symptoms of the introduced case (IC). doi:10.1371/journal.pntd. [/fig]
[fig] Figure 2: Spatial progression of the dengue epidemic that affected the city of Cairns, Australia, during January-August 2003. Time is represented as weeks since the onset of symptoms of the introduced case (IC). Red circles represent cases confirmed on the week shown, whereas blue circles represent the cumulative cases until that occurred up to that week. Video S1 is an animated version of this figure. doi:10.1371/journal.pntd. [/fig]
[fig] Figure 3: Directionality analysis for the location of dengue cases within d max (the distance up to which clustering of cases around the introduced case (IC) occurred). (A) Wavelet directionality analysis (WDA) for dengue cases. (B) Standard deviation (SD) ellipses showing the spatial anisotropy of dengue cases. (C) WDA for houses. (D) Wind rose plots showing the directionality (measured in 20 degree bins from north cardinal point) of daily prevailing wind direction and speed at 9:00 AM and 3:00 PM during February-March 2003 (the first 2 months of the epidemic, when most dispersal around the index case occurred). doi:10.1371/journal.pntd.0000920.g003 [/fig]
[fig] Figure 4: Significant space-time clustering (assessed by the Knox test) of dengue cases in the city of Cairns, Australia, during January-August 2003. Red circles and numbers identify each individual space-time cluster. Detailed information about each cluster can be found in [/fig]
[fig] Figure 5: Using human infection as a marker of dengue virus epidemic spread. (A) Association between the proportion of dengue cases that occurred on a given week (colors) and the spatial and temporal distances to a putative index case (PIdC) within a space-time cluster. PIdC can be interpreted as the most likely origin of a space-time cluster. (B) Mean distance from a confirmed dengue case to a PIdC according to the time (in weeks since the onset of PIdC symptoms) such case started presenting dengue symptoms. SD represents standard deviation of the mean value. doi:10.1371/journal.pntd. [/fig]
[fig] Figure 6: Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression (STAR) model applied to the weekly dengue infection in Cairns during the 15 weeks post virus introduction. (A) effect of time (f time ) with 95% credible intervals. (B) mean posterior spatial effect (f spat ). (C) effect of IRS (measured as cumulative percentage of premises sprayed around premise i within t 0 and t) with 95% credible intervals. (D) mean posterior spatial effect of IRS. White dots in maps represent the location of the index case's (IC) residence. doi:10.1371/journal.pntd.0000920.g006 [/fig]
[table] Table 1: Adjusted dengue incidence rates per age class for the city of Cairns during January-August 2003. [/table]
[table] Table 2: Description of each space-time cluster identified for the dengue epidemic that affected the city of Cairns during January-August 2003. [/table]
|
Applicability of standard parameters in diagnostics of primary open-angle glaucoma
Background:The aim of this study was to evaluate the sensitivity, specificity, and likelihood ratio of standard diagnostic parameters C/D, NFI, and MD and their applicability in diagnostics of primary open-angle glaucoma.Material/Method:This study included 194 subjects (371 eyes), ages 30 to 65 years old, who underwent full ophthalmologic examination and HRT, GDx, and FDT examinations. The values of C/D, NFI, and MD diagnostic parameters were determined. The data were statistically evaluated to determine their sensitivity, specificity, and likelihood ratio.Results:Values of the positive and negative likelihood ratios were C/D (11.471 and 0.159), NFI (3.739 and 0.152), and MD (6.323 and 0.309), respectively.Conclusions:The C/D parameter showed the highest sensitivity and specificity, as well as high positive likelihood ratio and near-zero negative likelihood ratio. The NFI and MD parameters showed lower likelihood ratios and their applicability for the diagnosis of primary open-angle glaucoma is limited.
# Background
Contemporary diagnosis of primary open-angle glaucoma is based on evaluation of the optic disc, the neuroretinal rim, and the retinal nerve fiber layers with indirect stereoscopic examination of a dilated pupil using a slit-lamp and Volk lens. Currently, methods such as HRT laser scanning ophthalmoscopy, GDxVCC laser scanning polarimetry, and FDT perimetryare among the more commonly used.
The HRT method uses a confocal laser scanning microscope that allows obtaining a 3D morphological image and determining detailed diagnostic parameters of the optic disc, including C/D ratio. The GDx laser scanning polarimetry is based on a measurement of the delay of polarized light components and allows the evaluation of retinal nerve fiber layer thickness in the peripapillary region and determining the NFI. Evaluating the field of vision with the use of FDT enables assessment of retinal sensitivity and determination of its mean deviation [bib_ref] Visual function-specific perimetry for indirect comparison of different ganglion cell populations in..., Sample [/bib_ref].
Purpose: The aim of the study was to evaluate the applicability of C/D, NFI, and MD standard diagnostic parameters in diagnosis of primary open-angle glaucoma through determination of sensitivity, specificity, and likelihood ratios. [bib_ref] Laser Diagnostic Technologies. Operation Manual, Polaczek-Krupa [/bib_ref].
# Material and method
All the patients underwent full ophthalmologic examination and then were examined with HRT laser scanning ophthalmoscopy, GDxVCC laser scanning polarimetry, and FDT field of vision examination on Humphrey Matrix perimeter. Standard diagnostic parameters were determined. C/D ratio was determined as the square root of the surface ratio of cup to disc . NFI was determined as the probability of glaucomatous damage to the retinal nerve fiber layer . MD value was determined as the difference between retinal sensitivity in a healthy eye (taking age into consideration) and retinal sensitivity values measured in the examined patient in all examined points. Although no invasive examination methods were used, patients were informed of the aim and course of the study and gave consent to participate in it.
Criteria for qualification of patients were specified in accordance with the guidelines of the European Glaucoma Society, described in detail in previous publications [bib_ref] Laser Diagnostic Technologies. Operation Manual, Polaczek-Krupa [/bib_ref]. Exclusion criteria were: eye refraction error greater than 2 Dsph and media opacity, which lowers the quality of HRT and GDx results, and patients with very small or very large surface of the optic disc (less than 1.69 mm 2 or more than 2.82 mm 2 ) evaluated in HRT examination.
Computer-stored measurement data matrix (MS Excel spread sheet) was used for statistical analysis. The significance of differences between mean values of measured parameters in all groups was measured with the t test. Sensitivity and specificity of these parameters and likelihood ratio (LR) in relation to glaucoma patients were determined. Mean values of these parameters determined in patients with suspected glaucoma (group B) were accepted as cut-off values for calculations. Sensitivity of the method was designated as the proportion of glaucoma patients correctly detected as a result of the analysis performed, and specificity as the proportion of healthy patients correctly detected as a result of the analysis. Likelihood ratio of the method was calculated from the value of its sensitivity and specificity. If positive LR is greater than 10 and negative LR is less than 0.1, the method may be used to confirm a diagnosis, but when positive LR values are from 5 to 10 and negative LR values are from 0.1 to 0.2, the method can still be considered as a useful diagnostic device.
# Results
Results of the measurements of the nerve fiber layer thickness index (NFI), cup-to-disc ratio (C/D), and mean deviation (MD) of retinal sensitivity in patients with primary open-angle glaucoma (group A), patients with suspected primary openangle glaucoma (group B), and in healthy patients (group C) are shown in .
## Nfi c/d md
# Discussion
We found statistically significant differences between individual groups of the examined population in results of the C/D, NFI, and MD measurements . The mean value of C/D ration in Group A (glaucoma patients) was 0.65±-0.11 vs. 0.43±-0.11 in Group C (healthy control patients). These results acquired are consistent with those of Jonas et al. [bib_ref] Ranking of optic variables for detection of glaucomatous optic nerve damage, Jonas [/bib_ref] and confirm that linear C/D ratio is of great use in diagnosing primary open-angle glaucoma. The mean value of NFI in primary open-angle glaucoma patients was 37.0±22.7 vs. 15.1±-4.8 in healthy patients. This is consistent with the observations of da Pozzo et al. [bib_ref] GDx-VCC performance in discriminating normal from glaucoma eyes with early field loss...., Da Pozzo [/bib_ref] , who have shown that the NFI is lower than 18 in healthy patients and does not exceed 31 in glaucoma patients. The mean value of MD in primary open-angle glaucoma patients was -3.00±5.07 dB and -0.29±-1.94 dB in healthy controls. Bowd et al. [bib_ref] Detecting early glaucoma by assessment of retinal nerve fiber layer thickness and..., Bowd [/bib_ref] determined the MD value of -0.16 dB for an analogous group of healthy patients and -3.8 dB for primary open-angle glaucoma patients.
The analysis of sensitivity and specificity values of the standard diagnostic parameters C/D, NFI and MD and their likelihood ratio value shows that the parameters with the greatest sensitivity and specificity are the C/D ratio (85.3% and 92.6%) and NFI (88.4% and 76.4%). Much lower values were obtained for the MD parameter (72.6 and 88.5). The positive likelihood ratio for C/D parameter was 11.5 and the negative likelihood ratio was 0.16. In the case of NFI and MD parameters, the positive likelihood ratios were 3.74 and 6.32, respectively, and the negative ratios 0.15 and 0.31, respectively. The very high positive likelihood ratio obtained for the C/D parameter (greater than 10) and near-zero likelihood ratio means that confirmation of glaucomatous damage to the optic disc can be obtained through measurement of this parameter. Although the measurement of NFI and MD can reveal certain useful information, it is of lesser diagnostic value.
# Conclusions
The C/D parameter showed the highest sensitivity and specificity, as well as high positive likelihood ratio and near-zero negative likelihood ratio. The NFI and MD parameters showed lower likelihood ratios and their applicability for the diagnosis of primary open-angle glaucoma is limited. |
A female with isolated hypogonadotropic hypogonadism: A case report and review article
A B S T R A C TBackground: Isolated Hypogonadotropic Hypogonadism (IHH) is a clinical syndrome that results from gonadal failure due to abnormal pituitary gonadotropin levels, in the presence of normal baseline and reserve testing of the remaining pituitary hormones. Case presentation: An 18 years old female came with primary amenorrhea, accompanied by poor breast and pubic development, with low levels of estradiol and gonadotropins but normal levels of other anterior pituitary hormones. Imaging of the hypothalamic-pituitary region revealed hypophyseal hypoplasia due to ischemia. Sex steroids therapy was given to induce pubertal development. IHH represents a rare condition but with a good prognosis. Discussion: Early diagnosis and treatment can prevent negative physical and psychological sequelae, and restore fertility in affected patients. Constant surveillance is required due to the possibility of gonadal axis reversal and/ or relapse of gonadal axis failure and to identify any adverse effects related to therapy. Conclusion: Early identification of IHH can help in treatment efficiency.
# Introduction
Hypogonadotropic hypogonadism (HH) or secondary hypogonadism is defined as a clinical syndrome that results from gonadal failure due to abnormal pituitary gonadotropin levels. HH may result from either absent or inadequate hypothalamic Gonadotropin-Releasing Hormone (GnRH) secretion or failure of pituitary gonadotropin secretion. It is typically characterized by a lack or delay of pubertal sexual maturation and inappropriately low circulating sexual steroids associated with low serum concentrations of Luteinizing Hormone (LH) and Follicle Stimulating Hormone (FSH), which is an effect of GnRH deficiency. HH is most frequently acquired and caused by several pathological processes, but it can also occur as part of various congenital syndromes with genetic causes of central hypogonadism have been identified. Idiopathic or isolated HH (IHH) was then used to indicate cases in which secondary causes of HH had been excluded [bib_ref] Hypogonadotropic hypogonadism in women, Skałba [/bib_ref] [bib_ref] Hypogonadotropic hypogonadism revisited, Fraietta [/bib_ref] [bib_ref] Approach to the patient with hypogonadotropic hypogonadism, Silveira [/bib_ref].
Congenital HH is divided into anosmic HH (Kallmann syndrome) and congenital normosmic IHH. The incidence of congenital hypogonadotropic hypogonadism is approximately 1-10:100,000 live births, and approximately 2/3 and 1/3 of cases are caused by Kallmann syndrome (KS) and idiopathic HH, respectively [bib_ref] Hypogonadotropic hypogonadism revisited, Fraietta [/bib_ref] [bib_ref] The genetic and molecular basis of idiopathic hypogonadotropic hypogonadism, Bianco [/bib_ref]. HH is one of the rare conditions in which specific medical treatment can reverse infertility. The precise and early diagnosis of HH can prevent negative physical and psychological sequelae, preserve normal peak bone mass, and restore fertility in affected patients. Thus, clinicians should understand the characteristics of the condition and be able to recognize and treat the condition to guarantee patients' quality of life [bib_ref] Approach to the patient with hypogonadotropic hypogonadism, Silveira [/bib_ref] [bib_ref] The reproductive outcome of women with hypogonadotropic hypogonadism undergoing in vitro fertilization, Yilmaz [/bib_ref]. Based on the description above, we are interested in reporting an 18-year-old female patient with IHH. We report based on surgical case report (SCARE) 2020 guideline [bib_ref] The SCARE 2020 guideline: updating consensus surgical CAse REport (SCARE) guidelines, Agha [/bib_ref].
## Case presentation
An 18-year-old female came with primary amenorrhea, accompanied with poor breast and pubic hair development. The patient did not report eating disorders, vigorous physical activity, or psychological stress. She had no olfactory complaints. She had 1 older brother with a history of normal pubertal development. There was no history of opiates, glucocorticoid, or psychotropic drugs administration. On a physical examination, she had no apparent congenital anatomical defects with normal habitus, height: 157 cm, arm span: 156 cm, weight: 61 kg, and normal body mass index (24.7 kg/m 2 ). Pubic hair and breast development were tanner stage II-III. Her bone age was 13 years old. At the time, the patient had already received hormonal therapy from the gynecology department for 2 months (Primolut 1 × 5mg and Esthero 1 × 0.625mg), with cyclical bleeding occurring regularly following therapy. Before therapy, there was poor development of breasts and pubic hair until the age of 18, so the possibility of Constitutional Delay of Puberty can be ruled out. Both anamnesis and physical examination showed clinical characteristics suggestive of HH, with functional (drugs, eating disorder, physical activity, stress) and autoimmune causes being ruled out. Serum karyotype examination such us 46, XX (N), ruling out gonadal dysgenesis. Basal hormonal evaluation revealed low serum estradiol 7.245 pg/mL (N 11-133 pg/mL) and serum progesterone 0.08 pg/mL (N 0.21-60 pg/ mL). Laboratory result on June 2019 showed suppressed LH 0.03 mIu/ mL (N 0.7-5.6 mIu/mL) and FSH 0.23 mIu/mL (N 1. Abdominal ultrasound examination revealed uterus hypoplasia (1.4 × 2.4 × 3.4 cm). Pelvic MRI confirmed uterus hypoplasia and revealed no apparent normal ovaries [fig_ref] Figure 1: MRI of hypoplasia uterus 1 [/fig_ref]. MRI of the hypothalamic-pituitary region revealed hypophyseal hypoplasia with small vessel ischemic in right and left corona radiata and right temporal subcortical with MR Angiography showed right segment A1, ACA hypoplasia (embryonal variant). There were no radiological signs suggestive of Kallman syndrome.
The patient was assessed with isolated HH with no additional therapy from the endocrinology polyclinic and was recommended to continue her outpatient care in both endocrinology and gynaecology polyclinic regularly.
# Discussion
In both males and females, secretion of LH and FSH is episodic, with secretory bursts that occur each hour and are mediated by a concordant episodic release of GnRH. Pulsatile secretion of GnRH by hypothalamic neurons is a crucial element of the reproductive cascade, initiating the release of pituitary gonadotropins, gonadal secretion of sex steroids, pubertal development, and gametogenesis. LH and FSH bind to receptors in the ovary and testis and regulate gonadal function by promoting sex steroid production and gametogenesis. In women, LH stimulates estrogen and progesterone production from the ovary. A surge of LH in the mid-menstrual cycle is responsible for ovulation, and continued LH secretion subsequently stimulates the corpus luteum to produce progesterone by enhancing the conversion of cholesterol to pregnenolone. Development of the ovarian follicle is largely under FSH control, and the secretion of estrogen from this follicle is dependent on both FSH and LH [bib_ref] Approach to the patient with hypogonadotropic hypogonadism, Silveira [/bib_ref] [bib_ref] Hypothalamic-and pituitary-derived growth and reproductive hormones and the control of energy balance..., Blanco [/bib_ref].
The diagnosis of IHH is confirmed by the findings of low levels of sex steroids (total testosterone (T) <100 ng/dL in males and estradiol (E2) <50 pg/mL in females) in the presence of low or inappropriately normal LH and FSH serum levels with typically normal levels of other anterior pituitary hormones. Acute stimulation with GnRH (GnRH test) is of little value in differentiating between a hypothalamic and pituitary defect. Most patients with GnRH deficiency show little or no response to an initial dose of GnRH, but normal responses may be elicited after repeated injections, in the presence of a hypothalamic defect. This initial slow response is attributed to the down-regulation of GnRH receptors after prolonged GnRH deficiency. Although widely used, the practical value of the GnRH test has been questioned because of its low costeffectiveness. Indeed, the GnRH test provides no extra diagnostic information relative to baseline gonadotropin levels. The anterior pituitary function must be investigated by basal hormonal levels to rule out a more complex endocrine disorder with multiple hormone deficiencies. Thyroid function should be assessed by TSH combined with free T4. Secondary adrenal deficiency can be assessed by measuring morning cortisol and ACTH. Because sex steroids are necessary for growth, performing a bone age measurement should be considered in all hypogonadal patients, who may have delayed bone age compared with chronologic age [bib_ref] Approach to the patient with hypogonadotropic hypogonadism, Silveira [/bib_ref] [bib_ref] Isolated GnRH deficiency: a disease model serving as a unique prism into..., Balasubramanian [/bib_ref] [bib_ref] A review of hypogonadotropic hypogonadism cases followed up in our clinic in..., Yılmazel [/bib_ref].
Hormonal replacement therapy is required to induce pubertal development, to maintain normal sexual function and avoid osteoporosis. In women, replacement therapy can be started initially with lowdose estrogen (ethinyl estradiol 5 μg or conjugated estrogen 0.3 mg daily). After 6 months or sooner, if breakthrough bleeding occurs, cyclical therapy is initiated by adding a progestin for endometrial protection, and the dose of estrogen is gradually increased to 20-30 μg over a 2-to 3-year period so as to respect the normal cadence of puberty. Full replacement dose of estrogen and progesterone is attained with 0.625-1.25 mg conjugated equine estrogen daily combined with cyclic 5-10 mg medroxyprogesterone acetate or 200 mg oral micronized progesterone [bib_ref] Physiological estrogen replacement therapy for puberty induction in girls: a clinical observational..., Ankarberg-Lindgren [/bib_ref] [bib_ref] Review of hormone replacement therapy in girls and adolescents with hypogonadism, Klein [/bib_ref].
If fertility is desired, treatment with gonadotropins or GnRH is needed to induce ovulation. In women, the gonadotropin regimen is initiated with purified human Menopausal Gonadotropin (hMG) at a starting dose of 37.5 IU/d IM, the dose being carefully titrated upward to 75-150 IU/d using cycle tracking by ultrasonography. Ideally, monofollicular growth is induced, and if the leading follicle is bigger than 18 mm, human Chorionic Gonadotropin (hCG) is administered at a dose of 5,000 to 10,000 IU IM to induce ovulation. Luteal support with one to two IM injections of hCG (2000-5000 IU) is subsequently required. Pulsatile GnRH therapy can be used as an alternative to gonadotropins, being equally or more effective. GnRH pulses (75-250 ng/kg for women) are administered, SC or intravenously, by a portable infusion pump with an interpulse frequency of 120 minutes [bib_ref] GnRH and GnRH receptors in the pathophysiology of the human female reproductive..., Maggi [/bib_ref] [bib_ref] Gonadotropin regulation by pulsatile GnRH: signaling and gene expression, Stamatiades [/bib_ref].
Reversal of HH, defined as restoration of normal serum gonad concentrations after cessation of even brief treatment with sex steroids, gonadotropin, or GnRH, occurs in about 10% of all HH cases, including those with Kallman Syndrome. This post-treatment awakening of the hypothalamo-pituitary-gonadal (HPG) axis suggests the presence of hypothalamic GnRH neurons that do not function during adolescence and possibly require undefined stimuli (potentially environmental/sex steroid exposure) to initiate normal activity. The precise physiologic basis of the reversal phenomenon is yet to be fully understood but previous studies have shown a role of androgen exposure as a common factor for reversal. The androgen exposure can either be due to exogenous testosterone or via endogenous testosterone production stimulated by HCG/GnRH. It has been known that sex steroids can influence a neuronal generation. The increased androgen levels have been proposed to stimulate GnRH cells and regeneration of GnRH neurons by taking advantage of the plasticity of GnRH neurons. Another feature seen in previous studies is a relapse of hypogonadism in those who have achieved reversal due to the susceptibility of the gonadotropic axis to insult. Thus constant surveillance of patients with HH on therapy is important to look for adverse effects, or even the possibility of reversal of axis and/ or further relapse [bib_ref] Reversal of idiopathic hypogonadotropic hypogonadism, Raivio [/bib_ref] [bib_ref] Reversal and relapse of hypogonadotropic hypogonadism: resilience and fragility of the reproductive..., Sidhoum [/bib_ref] [bib_ref] Reversal of idiopathic hypogonadotropic hypogonadism: a cohort study in Chinese patients, Mao [/bib_ref] [bib_ref] Idiopathic hypogonadotropic hypogonadism reversal after testosterone replacement in a 34-year-old male, Rashid [/bib_ref].
# Conclusion
An 18 years old female patient with IHH has been reported. IHH is defined as a clinical syndrome that results from gonadal failure due to abnormal pituitary gonadotropin levels, in the presence of normal baseline and reserve testing of the remaining pituitary hormones. Clinical presentations may vary according to age, however, the patient came with complaints of primary amenorrhea and poor breasts, and pubic hair development. The diagnosis of IHH is confirmed by the findings of low levels of estradiol in the presence of low or inappropriately normal LH and FSH serum levels with typically normal levels of other anterior pituitary hormones. MRI of the hypothalamic-pituitary region revealed hypophyseal hypoplasia due to ischemia. The patient received sex steroids therapy from the gynaecology department to induce pubertal development. IHH represents a rare condition with a good prognosis in which specific medical treatments can reverse infertility. Due to the possibility of gonadal axis reversal and/or relapse of gonadal axis failure and to identify any adverse effects related to long term hormonal therapy, the patient should be closely monitored during therapy and constant surveillance is required.
# Ethical approval
Not applicable.
# Sources of funding
None.
# Author contribution
All authors contributed toward data analysis, drafting and revising the paper, gave final approval of the version to be published and agree to be accountable for all aspects of the work.
## Consent
Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request.
## Registration of research studies
Not applicable.
## Guarantor
Soebagijo Adi Soelistijo.
## Provenance and peer review
Not commissioned, externally peer reviewed.
## Declaration of competing interest
The authors declare no conflict of interest.
[fig] Figure 1: MRI of hypoplasia uterus 1. [/fig]
|
Signal Processing of MEMS Gyroscope Arrays to Improve Accuracy Using a 1st Order Markov for Rate Signal Modeling
This paper presents a signal processing technique to improve angular rate accuracy of the gyroscope by combining the outputs of an array of MEMS gyroscope. A mathematical model for the accuracy improvement was described and a Kalman filter (KF) was designed to obtain optimal rate estimates. Especially, the rate signal was modeled by a first-order Markov process instead of a random walk to improve overall performance. The accuracy of the combined rate signal and affecting factors were analyzed using a steady-state covariance. A system comprising a six-gyroscope array was developed to test the presented KF. Experimental tests proved that the presented model was effective at improving the gyroscope accuracy. The experimental results indicated that six identical gyroscopes with an ARW noise of 6.2 °/√h and a bias drift of 54.14 °/h could be combined into a rate signal with an ARW noise of 1.8 °/√h and a bias drift of 16.3 °/h, while the estimated rate signal by the random walk model has an ARW noise of 2.4 °/√h and a bias drift of 20.6 °/h. It revealed that both models could improve the angular rate accuracy and have a similar performance in static condition. In dynamic condition, the test results showed that the first-order Markov process model could reduce the dynamic errors 20% more than the random walk model.
# Introduction
Micro Electromechanical System (MEMS) gyroscopes have been used for measuring rate or angle of rotation in various inertial measurement fields thanks to their attractive advantages such as small size, low cost, possible batch fabrication and low power consumption [bib_ref] Nonresonant micromachined gyroscopes with structural mode-decoupling, Acar [/bib_ref] [bib_ref] High Performance Matched-Mode Tuning Fork Gyroscope, Zaman [/bib_ref]. However, to date the low accuracy of the MEMS gyroscope has limited its applications to tasks requiring high-precision rate signals. During the development of MEMS gyroscopes, several common methods have been explored to improve the accuracy, mostly focusing on the device itself, such as enhancement in materials, fabrication processes, structural design and interface circuits [bib_ref] Sub-Degree-Per-Hour Silicon MEMS Rate Sensor with 1 Million Q-Factor, Prikhodko [/bib_ref] [bib_ref] Dynamic characteristics of vertically coupled structures and the design of a decoupled..., Choi [/bib_ref] [bib_ref] A mode-matched silicon-yaw tuning-fork gyroscope with subdegree-per-hour allan deviation bias instability, Zaman [/bib_ref] [bib_ref] Two-scale simulation of drop-induced failure of polysilicon MEMS sensors, Mariani [/bib_ref] [bib_ref] Genetic algorithm for the design of electro-mechanical sigma delta modulator MEMS sensors, Wilcock [/bib_ref]. Other extensive research has been also carried out on the area of accuracy improvement. One of the fundamentally different methods is the virtual gyroscope technology, its principle is that several identical low-precision gyroscopes are combined together to form an array and measure the same rate. An optimal rate estimate can be obtained by the optimal filter to fuse these multiple measurements. This approach can further reduce the noise, the bias instability and improve the overall accuracy beyond the performance limitations of individual gyroscopes. Bayard and Ploen constructed a virtual gyroscope using a four-gyro array, the simulation results showed that the gyroscopes with drifts of 8.66 °/h could be combined into a virtual gyroscope with a drift of 0.062 °/h when the gyroscope array are assumed to have a correlation factor of −0.33329. In our previous work, a MEMS gyroscope array with three individual gyroscopes was studied [bib_ref] An integrated MEMS gyroscope array with higher accuracy output, Chang [/bib_ref] [bib_ref] On Improving the Accuracy of the Micromachined Gyroscopes Based on the Multi-Sensor..., Chang [/bib_ref] , and a two-level optimal filter was designed to compensate the random noises and reduce the bias drift [bib_ref] An integrated MEMS gyroscope array with higher accuracy output, Chang [/bib_ref]. Al-Majed and Alsuwaidan presented a multi-filters adaptive estimator for the MEMS gyroscope array system to improve the angular rate estimates [bib_ref] A New Testing Platform for Attitude Determination and Control Subsystems: Design and..., Al-Majed [/bib_ref]. Additionally, an extended Kalman filter (EKF) technique for combining multiple sensors was studied in [bib_ref] A Signal Processing Technique for Improving the Accuracy of MEMS Inertial Sensors, Stubberud [/bib_ref].
In recent years, the redundant MEMS inertial sensors have been utilized integrated with GPS to improve navigation performance. Numerous studies and researches have been undertaken on redundant inertial measurement unit (IMU) integration, whereby the measurements of multiple IMUs are fused [bib_ref] A review of multisensor fusion methodologies for aircraft navigation systems, Allerton [/bib_ref] [bib_ref] Noise reduction and estimation in multiple micro-electro-mechanical inertial systems, Waegli [/bib_ref] [bib_ref] Improving Accuracy with Multiple Sensors: Study of Redundant MEMS-IMU/GPS Configurations, Guerrier [/bib_ref] [bib_ref] Redundant IMUs for precise trajectory determination, Colomina [/bib_ref] [bib_ref] A low-cost redundant inertial measurement unit for unmanned air vehicles, Sukkarieh [/bib_ref] [bib_ref] Data fusion algorithms for multiple inertial measurement units, Bancroft [/bib_ref] [bib_ref] Optimum skewed redundant inertial navigators, Pejsa [/bib_ref] , into what can be called a virtual IMU (VIMU). In VIMU theory, the configuration of multiple IMUs is one of the most important considerations. Pejsa presented a theory about how multiple sensors can be placed on a cone to maximize redundancy and mathematically determined the optimal configuration for sensor axes [bib_ref] Optimum skewed redundant inertial navigators, Pejsa [/bib_ref]. In [bib_ref] Noise reduction and estimation in multiple micro-electro-mechanical inertial systems, Waegli [/bib_ref] , an approach for reducing and estimating the noises in the inertial sensors was investigated by using multiple inertial sensors, furthermore the INS/GPS integrator based on the synthetic IMU, extended IMU and geometrically constrained IMU mechanization were presented for integration of GPS with redundant MEMS-IMUs. In [bib_ref] Improving Accuracy with Multiple Sensors: Study of Redundant MEMS-IMU/GPS Configurations, Guerrier [/bib_ref] the impact of the number of sensors and sensors' orientation on the system performance were demonstrated, and the optimal configuration of the multiple IMU triads was also studied. Additionally, three mechanization approaches of the system and an observation model for the multiple IMUs integration were described in [bib_ref] Redundant IMUs for precise trajectory determination, Colomina [/bib_ref]. Furthermore, in [bib_ref] Data fusion algorithms for multiple inertial measurement units, Bancroft [/bib_ref] several fusion algorithms were developed using multiple IMUs to enhance performance in the pedestrian navigation. It can be seen that the technology of the virtual gyroscope is essentially identical with the VIMU, since both of them fuse multiple measurements to create a combined signal from a sensor array for improving the overall performance. However, in contrast with the VIMU, the virtual gyroscope usually models the rate signal directly to obtain an optimal rate estimate. Therefore, the objective of this paper, which based on the virtual gyroscope, is to study various models for modeling the rate signal to combine multiple MEMS gyroscopes.
The key of combining multiple gyroscopes for accuracy improvement lies in rate signal modeling and the optimal filter design. Therefore, how to model the rate signal is a prerequisite for constructing a virtual gyroscope system. Through analyzing the current approaches [bib_ref] An integrated MEMS gyroscope array with higher accuracy output, Chang [/bib_ref] [bib_ref] On Improving the Accuracy of the Micromachined Gyroscopes Based on the Multi-Sensor..., Chang [/bib_ref] , it can be found that these approaches usually modeled the rate signal by a random walk and then evaluated performance. However, to model the rate signal as the random walk has some limitations, which will affect the system's overall performance. The random walk process is a non-stationary process since its variance increases with time; however, in practice most of the rate signal energy is at low frequency. Furthermore, the practical dynamic characteristic of the angular rate cannot always be well represented and reflected by the random walk process in conditions of lacking more information about rate signal such as power-spectral density and bandwidth. The use of random walk models may lead to accuracy degradation, especially signal attenuation in a dynamic situation. To overcome such a problem, the rate signal should be modeled by other suitable approaches.
Markov process modeling is a powerful and commonly used technique that has been introduced for modeling inertial sensors errors for several years, but it has not been used for modeling the virtual gyroscope before. In most applications, the maneuverability characteristics of aircrafts can be regarded as contained in a certain frequency bandwidth and magnitude [bib_ref] A Quaternion-Based Orientation Estimation Algorithm Using an Inertial Measurement Unit, Kim [/bib_ref]. In addition, the angular rate signal ω k+1 for the subsequent time point t k+1 usually can be thought as related to ω k for the former time point t k . Such dynamic characteristics and property can be suitably represented by a Markov process model. Furthermore, the Markov process is a stationary process that has an exponential Autocorrelation Function (ACF). The ACF of a zero-mean first-order Markov process is defined by a decaying exponential form [bib_ref] Accuracy Improvement Using Autoregressive (AR) Modeling of INS Sensor Errors, Nassar [/bib_ref].
Although the dynamic characteristics of the rate signal can be more accurately represented by using an Autoregressive (AR) model of orders higher than one to model the rate signal, it would result in estimating more system parameters and increasing the complexity of filter. On the other hand, the first-order Markov process satisfies the requirement for establishing the system state-space model. Therefore, in this paper the first-order Markov process is used to model the rate signal, and then a complete KF is designed for obtaining the optimal rate estimate. Furthermore, the factors that affect the system performance have been analyzed. Lastly, the hardware of the virtual gyroscope system is implemented and the performance of the virtual gyroscope with two different rate signal models are tested and compared.
## Modeling of virtual gyroscope system
The structure and principle of the virtual gyroscope is shown in [fig_ref] Figure 1: Structure and principle of the virtual gyroscope [/fig_ref]. Several identical MEMS gyroscopes form a sensor array and the multiple measurements of the sensor array are collected by the data acquisition system. These measurements are used to quantify the random errors through noise modeling by the Allan variance. The optimal filter is designed to combine multiple measurements to produce a minimum variance rate estimate. Therefore, it can provide a combined signal with a higher accuracy than that of the single sensors. The key of the virtual gyroscope lies in the optimal filter, which includes the modeling of rate signal and system state-space. The optimal filter based on the first-order Markov process will be designed. In addition, the KF has the advantage of being a systematic method to ensure minimum variance rate estimation [bib_ref] A new approach to linear filtering and prediction problems, Kalman [/bib_ref]. It is suitable for dealing with dynamic data and has high real-time performance compared with other estimation approach, and widely used to improve performance of the micromechanical inertial sensors [bib_ref] Improving the responses of several accelerometers used in a car under performance..., Hernández [/bib_ref]. Therefore, the KF is used to implement the gyroscope array.
## State model for virtual gyroscope
The true rate signal can be modeled directly to improve accuracy. For most applications of gyroscope, the rate signal is propagated according to an approximate power spectral density function in which the three degree-of-freedom motion is contained. In other words, this motion is expected to be within a certain frequency bandwidth and magnitude [bib_ref] A Quaternion-Based Orientation Estimation Algorithm Using an Inertial Measurement Unit, Kim [/bib_ref]. It can be modeled in state space form as a first-order system with a process time constant set according to the system's bandwidth and the process noise related to the limit of the magnitude of the motion. Therefore, it is suitable to model the true rate signal as first-order Markov process:
[formula] 1/ n ω ω ω τ ω = − ⋅ + & (1) [/formula]
where ω is the true rate signal, ω τ is the process time constant, n ω is the zero-mean Gaussian white noise with variance parameter q ω that can be set according to the dynamic characteristics of the input rate signal. To set the true rate signal as the system state, the state model for the virtual gyroscope system can be established as:
[formula] ( ) ( ) ( ) 1/ ( ) X t F X t w t X t n ω ω τ = ⋅ + = − ⋅ + & (2) [/formula]
where system state ( )
[formula] X t ω = , coefficient scalar 1 / F ω τ = − [/formula]
, system process noise ( ) w t n ω = and
[formula] [ ( ) ( )] ( ) T E w t w t q ω τ δ τ + = . [/formula]
With such direct modeling of the rate signal, the rate signal can be estimated and obtained directly using a KF. Furthermore, the accuracy of the combined rate signal can be analyzed by the KF covariance; this can provide a reference for system improvement and parameters adjustment. In particular, it is suitable for the individual gyroscope which only includes the random noise of the white noise.
## Measurement model for virtual gyroscope
The measurement errors of MEMS gyroscope is usually composed of the known errors and random errors. The known errors can be eliminated by the testing procedure, thus only the random errors are discussed for the gyroscope. A common model for the MEMS gyroscope is widely used in many application which mainly includes the white noise denoted as angular random walk (ARW) and bias drift due to rate random walk (RRW) [bib_ref] MEMS-Based Multi-Sensor Integrated Attitude Estimation Technology for Mav Applications, Xue [/bib_ref] [bib_ref] Gyro Modeling and Estimation of Its Random Noise Sources, Lam [/bib_ref]. However, for the MEMS gyroscopes with a lower accuracy, the ARW noise is the dominant term rather than RRW; other researchers have made the same observation [bib_ref] Analysis and modeling of inertial sensors using Allan variance, El-Sheimy [/bib_ref]. In our experience, this is the case for consumer-grade and lower tactical-grade MEMS gyroscopes. Therefore, to estimate and extract the true rate signal from the noise-corrupted measurements and reduce the dominant ARW noise, it is reasonable to model the ARW alone to reduce the dimension of the KF, thus the output rate signal of such gyroscopes can be described as a simplified model:
[formula] ( ) ( ) ( ) y t t n t ω = + (3) [/formula]
where ( ) y t is the output rate signal of the gyroscope, ( ) t ω is the true rate signal, and ( ) n t is the white noise corrupting the gyroscope rate measurement but becoming the angular random walk at the gyroscope angle level. As for a gyroscope array with sensor number of N, the measurement model for the virtual gyroscope system can be expressed as:
[formula] ( ) ( ) t t ω = ⋅ + Z H v(4) [/formula]
and:
[ ]
[formula] 1 2 ( ) , , , T N t y y y = Z L , [ ] 1 1,1, ,1 T N × = H L , [ ] 1 2 ( ) , , , T N t n n n = v L (5) where ( ) t v is the measurement noise with [ ( )] 0 E t = v and [ ( ) ( )] ( ) T E t t τ δ τ + = v v R , [ ] E ⋅ is the [/formula]
mathematical expectation operator, t is an arbitrary sampling time and τ is the time lag between samples, δ(τ) is the Dirac delta function. The matrix R is the covariance matrix of the noises v(t), it will not be necessarily diagonal since the component gyroscopes may be correlated with each other. The correlation between sensor array can be interpreted as the outputs of the component gyroscopes satisfying a statistical relationship. This relationship can be characterized and indicated by a correlation factor ρ and covariance matrix R. For a gyroscope array with positive correlation factors, the noises of the component gyroscopes will show an identical statistical distribution pattern; negative correlation factors indicate the opposite statistical distribution pattern of the noises. Concretely, correlation factor ρ = 0 indicates an uncorrelated gyroscope array; the gyroscope array will have maximum correlation when ρ = ±1; when correlation factor −1< ρ <1, the gyroscope array will have a correlation in between. Suppose that a constant cross-correlation exists between the identical component sensors, then the covariance matrix R for the noises v(t) can be expressed as:
[formula] 2 1 1 1 n N N ρ ρ ρ ρ σ ρ ρ × ⎡ ⎤ ⎢ ⎥ ⎢ ⎥ = ⋅ ⎢ ⎥ ⎢ ⎥ ⎣ ⎦ R L L M M O M L(6) [/formula]
where 2 n σ is the ARW noise variance of the component gyroscopes, and ρ is the constant crosscorrelation factor.
## Optimal kf for angular rate estimate
Based on the aforementioned system state-space model of Equations (2) and (4), the continuous KF can be expressed as:
[formula] ˆˆ( ) ( ) ( )[ ( ) ( )] X t FX t t t X t = + − K Z H & (7) 1 ( ) ( ) T t P t − = K H R (8) 2 1 ( ) 2 ( ) ( ) T P t P t F P t q ω − = − + H R H &(9) [/formula]
Equations (7) to [bib_ref] An integrated MEMS gyroscope array with higher accuracy output, Chang [/bib_ref] describe the relationship between the system state X(t), filter gain K(t) and estimated covariance P(t). It can be seen that the system state X(t), i.e., the true rate signal, can be eventually estimated through solving the problem of the differential equation (9), which is called matrix Riccati differential equation (RDE). Therefore, solving the RDE is crucial to KF. From analysis, it indicates that the KF is completely observable due to the rank of observable matrix is equal to the dimension of X(t), thus the KF will be stable and solution to the RDE will approach a steady-state value, this avoids numerical integration for the RDE and updates P(t) in each step. For this case, it can get ( ) 0 P t = & and Equation (9) goes over into an algebraic equation:
[formula] 2 2 0 F P D P q ω ∞ ∞ ⋅ − ⋅ + =(10) [/formula]
where P ∞ is a scalar representing the steady-state covariance of the estimated rate signal, and
Inserting Equation (11) into Equation (8) yields:
[formula] 1 2 1 ( 1/ (1/ ) ) T D D q ω ω ω τ τ − − ∞ = − + + K H R(12) [/formula]
where ∞ K is a 1 × N row vector representing the steady-state filter gain. Using steady-state gain ∞ K , the rate signal estimate can be propagated in the continuous state-space form:
[formula] ˆ( ) ( ) ( ) ( ) X t X t t ∞ ∞ = ⋅ + K H K Z & F -(13) [/formula]
The discrete KF can be derived by directly discretizing the continuous KF of Equation (13) using a zero-order hold approximation. Assume that the measurement value Z(t) is held constant over each sampling period, we have:
[formula] 1 1ˆ( 1) AT AT k k k X e X A e − + ∞ = + − K Z(14) [/formula]
where (1/ ) A ω τ ∞ = − + K H , and T is the sampling period. Consequently, the outputs of the virtual gyroscope can be determined by the discrete KF of Equation [bib_ref] Noise reduction and estimation in multiple micro-electro-mechanical inertial systems, Waegli [/bib_ref]. The implementation of a virtual gyroscope system makes use of a discrete KF with a structure as shown in [fig_ref] Figure 2: Virtual gyroscope implementation using a discrete KF [/fig_ref]. From the above description of state-space model and KF, Equation (2) holds for constant rate input, and Equation (14) can be used to obtain a rate signal with high accuracy for the case of static and constant condition. For dynamic conditions, i.e., for time varying rate inputs, the variance q ω can be used to set an appropriate bandwidth to satisfy the dynamic requirement.
## Performance analysis of kf
The bandwidth is an important parameter for the KF, it is related to the structure and parameters of the KF such as variance q ω , number of individual gyroscopes N and noise statistical quantities of the individual gyroscopes. The correlation factor doesn't affect the bandwidth. From the continuous KF of Equation [bib_ref] A review of multisensor fusion methodologies for aircraft navigation systems, Allerton [/bib_ref] , the transfer function from the outputs of the gyroscope array to virtual gyroscope can be expressed in Laplace transform domain as:
Due to a input rate signal is detected by multiple gyroscopes, the outputs of the gyroscope array can be expressed as ( ) ( ) s t y t = ⋅ Z H without considering the sensors noises, where ( ) s y t is the output rate signal of the single gyroscope. Using Equations (12) and (15), the transfer function can be formed as:
[formula] 2 1 2 ( ) ( ( 1/ ) ) ( 1/ ( 1/ ) ) H s s Dq Dq ω ω ω ω ω τ τ τ − = + + − + +(16) [/formula]
The frequency response of the KF can be obtained as:
[formula] ( ) ( ) (0) H j A H ω ω = , (0) 1 A =(17) [/formula]
The −3 dB standard is used to define the KF bandwidth, using KF frequency response, the bandwidth can be expressed as:
[formula] 2 1 (1/ ) 2 BW Dq ω ω τ π = +(18) [/formula]
With the same derivation and analysis, the bandwidth of KF modeled by the random walk can be expressed as:
[formula] 1 2 rw BW Dq ω π =(19) [/formula]
AT A e − − Given the identical system parameters, it is obvious that BW is higher than BW rw and the bandwidth of KF modeled by a first-order Markov is wider than that of the KF modeled by the random walk. Therefore, the presented KF has an improved dynamic characteristic compared to the KF modeled by the random walk.
On the other hand, the performance of the KF and accuracy of the combined rate signal can be evaluated by the steady-state covariance. It can be seen from Equation (11) that the steady-state covariance is affected by three parameters including D , ω τ and q ω . It indicates that P ∞ reduces while decreasing the value of 1 D − . Substituting ω τ → ∞ into Equation (11) results in:
[formula] / P q D ω ∞ =(20) [/formula]
It shows that P ∞ will be equal to zero when 1 0 D − = , for this case the KF reaches the best performance and the combined rate signal with the best accuracy could be obtained. The (6) and (20), the steady-state covariance can be expressed as:
[formula] 2 1 ( 1) n N P q N ω ρ σ ∞ + − =(21) [/formula]
Therefore a minimum P ∞ can be achieved through adjusting the factors in Equation [bib_ref] Accuracy Improvement Using Autoregressive (AR) Modeling of INS Sensor Errors, Nassar [/bib_ref]. In particular, P ∞ will approach zero as 1/ ( 1) N ρ → − − . Setting parameter q ω = 1 in Equation (21), we
[formula] obtain / /[ 1 ( 1 ) ] n P N N σ ρ ∞ = + − . [/formula]
The relation between the gyroscopes noise reduction, correlation factor ρ and sensor number N is shown in [fig_ref] Figure 3: Relationship between the gyroscope noise reduction and correlation factor ρ and sensors... [/fig_ref]. It indicates that the magnitude of noise reduction increases with decreasing of factor ρ, and the reduction magnitude with a negative correlation is higher than that of a positive one. Furthermore, the correlation factor (either positive or negative) cannot degrade the individual gyroscope accuracy. In addition, Equation [bib_ref] Accuracy Improvement Using Autoregressive (AR) Modeling of INS Sensor Errors, Nassar [/bib_ref] shows that P ∞ will decrease when q ω is decreased. As for Equation (1) for modeling rate signal, the q ω can be chosen as it depends on the true rate signal characteristics. In a practical situation, the characteristics of the true rate signal cannot be accurately obtained because of dynamic behaviors of the input rate signal. However, from Equation (18), it can be seen that the KF bandwidth is related to the variance q ω . Consequently, the KF bandwidth can be used to choose an appropriate value of q ω to implement the virtual gyroscope system, that is, the requirement of various dynamic applications can be satisfied by adjusting q ω to obtain different bandwidths.
On the other hand, different values of q ω will lead to different system performances. Due to the actual statistical model of the rate signal is difficult to accurately obtain, and even q ω varies with changes of the environment, thus if the choice of q ω could accurately or closely reflect the dynamic characteristics of the input rate signal, the KF will reach the best performance and the virtual gyroscope signal with the best accuracy can be obtained; while q ω is smaller than such 'value', the performance of the KF will be degraded, it would result in a signal attenuation; while q ω is higher than such 'value', the performance of the KF will be also degraded and eventually reach the level of a simple averaging process with increasing of q ω .
# Experiments and discussion
## Correlations between the mems gyroscope array
The performance of the KF was analyzed in Section 2.4. It showed that the accuracy could be considerably improved when the gyroscope array has a negative correlation. However, in a practical implementation, the gyroscope array with an expected correlation factor has so far been hard to artificially design and fabricate. Additionally, usually the performance can be further improved through increasing the number N of the individual sensors in the array [fig_ref] Figure 3: Relationship between the gyroscope noise reduction and correlation factor ρ and sensors... [/fig_ref] ; however, one needs to consider the system complexity and uniformity between the gyroscopes. Therefore, in this work, six separate identical ADXRS300 MEMS gyroscopes are utilized to form a gyroscope array to serve as the virtual gyroscope system.
The correlation matrix can be used to indicate noise correlation between the multiple gyroscopes. Since the individual gyroscopes are described by a simple model of Equation (3), the component elements of the correlation matrix, i.e., correlation factors between every pair of sensors, can be obtained by:
[formula] ( , ) ( , ) ( , ) ( , ) i j i i j j C g g i j C g g C g g = CorrM , , 1,2, ,6 i j = L(22) [/formula]
where CorrM is a 6 × 6 correlation matrix of gyroscope array, C = cov(g i , g j ) is the covariance between the outputs noise of the ith and jth gyroscope. The correlation matrix of the six-gyroscope array was analyzed with a zero rate output recorded for 1 hour at a sampling rate of 200 Hz. Through multiple tests, the correlation matrix illustrated in [fig_ref] Table 1: Correlation matrix of noises for six-gyro array [/fig_ref] was obtained, where it is obvious that the correlations between the separate gyroscopes are relatively weak and values of correlation factor are close to zero, therefore, such a gyroscope array can be regarded as uncorrelated. It is reasonable to suppose that the main reason lies in the independent working condition for the mechanical sensing elements and controlling interface circuits of the individual gyroscopes [bib_ref] An integrated MEMS gyroscope array with higher accuracy output, Chang [/bib_ref].
## Hardware implementation
The hardware of the virtual gyroscope is mainly composed of sensor array, central processing unit, A/D data acquisition unit, serial communication unit and FLASH memory. [fig_ref] Figure 4: A prototype of the virtual gyroscope system [/fig_ref] shows the prototype of the virtual gyroscope system. Six ADXRS300 MEMS gyroscopes are utilized to form a gyroscope array. A TMS320VC5416 DSP chip is chosen as core processor for the virtual gyroscope. The A/D data acquisition unit uses a 16-bit ADS7807 to collect voltage signals from gyroscope array. The system operates in serial mode and connects with DSP directly through the 16-bit data bus. FLASH memory unit uses 4Mbit Flash chip AM29LV400 to provide storage space for external system program. The virtual gyroscope signal is exported by RS-232. The experimental results will be presented to quantify the performance of the virtual gyroscope. The geometry configuration of the sensors array should be considered to improve the estimation process. It had been shown in [bib_ref] Redundant IMUs for precise trajectory determination, Colomina [/bib_ref] [bib_ref] Data fusion algorithms for multiple inertial measurement units, Bancroft [/bib_ref] that the separation between two sensors will Gyro array DSP board Gyro array
## Dsp board
introduce an angular acceleration that relates the individual states through differential equations in fusing redundant IMUs. As the focus of this paper is not on the geometry layout of the gyroscope array, the influence of sensors geometry on the system performance will be investigated in the successive work.
## Static performance test
The ARW, bias drift and noise density of the virtual gyroscope with two KF models are tested and compared. The performances of the virtual gyroscope are evaluated using FFT analysis and root Allan variance of a zero rate output recorded for 1 hour. Due to the 40 Hz bandwidth of an individual gyroscope, and the bandwidth of the virtual gyroscope will not be higher than 40 Hz, thus the sampling rate was set to 200 Hz to satisfy the Nyquist theorem. The system parameters are chosen as τ ω = 500 s and q ω = 0.0772 deg 2 /s 3 . The comparisons of FFT plot and Allan variance measurement between the virtual gyroscope and single gyroscope are shown in [fig_ref] Figure 5: FFT plot of the virtual gyroscope compared to the single gyroscope and... [/fig_ref] , respectively. Furthermore, the accuracy of the virtual gyroscope is compared with the averaging outputs of six individual gyroscopes. The detailed results are illustrated in [fig_ref] Table 2: Static test results of the virtual gyroscope [/fig_ref].
From the FFT plot, the noise level indicates noise floor of ~0.11°/s/√Hz for the single gyroscope and 0.03°/s/√Hz for the virtual gyroscope modeled by the first-order Markov process, whereas the corresponding value estimated by the random walk model is 0.04°/s/√Hz . It also suggests a low-pass characteristic of the KF. In addition, it indicates a noise floor of ~0.05°/s/√Hz for the rate signal obtained by averaging outputs of the six individual gyroscopes. From the Allan variance plot, both the ARW noise and bias drift are reduced by fusing multiple measurements from gyroscope array. [fig_ref] Table 2: Static test results of the virtual gyroscope [/fig_ref] reveals that the ARW noise of the single gyroscope is reduced from 6.17 °/√h to 1.83 °/√h and 2.35 °/√h by two different models, indicating an ARW reduction factor of about 3.4 by the first-order Markov model. Meanwhile, the bias drift is reduced from 54.11 °/h to 16.32 °/h, which is lower than that of the 20.65 °/h estimated by the random walk model. It is clear that the performances of two models are comparable due to the static characteristics of the angular rate signal. Furthermore, the Allan variance results reveal that values of the ARW noise and bias drift for rate signal estimated by the simple averaging are 2.73 °/√h and 22.72 °/h, respectively, which are greater than that of the presented KF, thus the performance of the presented KF is higher than that of the simple averaging process. Additionally, it can be seen that the ARW reduction is greater than that of the RRW and bias drift, the main reason being that the ARW is the dominant noise in the single gyroscope and the virtual gyroscope adopts a simple error model to describe the relationship of the input angular rates and gyroscope outputs.
## Dynamic performance test
The dynamic tests are carried out on a horizontal turntable [fig_ref] Figure 4: A prototype of the virtual gyroscope system [/fig_ref]. The standard deviation (1σ) of the estimated errors is used to evaluate the accuracy of the rate signal before and after KF. The sampling rate and parameter q ω are set to 500 Hz and 1.9290 deg 2 /s 3 .
Three kinds of dynamic condition are chosen to test the presented model. Firstly, the turntable is controlled to rotate in the horizontal plane with a 40 °/s constant rate, the outputs of the individual gyroscopes and virtual gyroscope are shown in . Secondly, the turntable is controlled to rotate with a random input rate, the outputs of the individual gyroscopes and virtual gyroscope are shown in . Lastly, the turntable swings with the 40° angle amplitude and 0.25 Hz frequency, thus the input rate signal is ω = 62.8 × sin(1.57t) °/s, the results of the virtual gyroscope are shown in [fig_ref] Figure 9: Cont [/fig_ref] , here only the data of the first 30s are demonstrated to display the results more clearly. The detailed results are illustrated in [fig_ref] Table 3: Dynamic test results of the virtual gyroscope [/fig_ref]. [fig_ref] Table 3: Dynamic test results of the virtual gyroscope [/fig_ref] shows that the gyroscopes' 1σ errors are reduced from 1.45 °/s to 0.06 °/s and 0.05 °/s by two different models in the constant rate test. It can be seen that the performance improvement of two different models are nearly equivalent. In addition, from , the bound on the angular position error caused by the ARW and RRW noisy terms can be obtained by integrating the outputs of gyroscopes over time [bib_ref] Statistical modeling of rate gyros, Vaccaro [/bib_ref].
In the random rate test, the virtual gyroscope signals could well reflect the dynamic characteristic of the input angular rate , the 1σ errors are reduced from 1.61 °/s to 0.47 °/s and 0.29 °/s, respectively, by both different models. It reveals that the performance of the first-order Markov process model is higher than that of the random walk due to the input rate signal has a more dynamic characteristic than constant condition.
In the swing test, the amplitude of combined rate signal estimated by the first-order Markov model reaches to 62.29 °/s, which is basically in accordance with the experimental setting, meanwhile, the 1σ errors are reduced to 0.16 °/s that is much smaller than 1.8 °/s estimated by the random walk model. Furthermore, [fig_ref] Figure 9: Cont [/fig_ref] indicates amplitude attenuation by the random walk model, and the virtual gyroscope signal cannot accurately reproduce the dynamic characteristic of the input rate signal.
It would be useful to verify the presented KF to expand the experiments to different axes and rotations sensed by several axes in addition to the experiments. However, such experiment is difficult to conduct because of the limitation in testing condition and all the individual gyroscopes are oriented along the same axis. Therefore, the verification in the paper was only implemented through such experiments.
# Conclusions
In this paper, the first-order Markov process was used to model the rate signal for fusing multiple MEMS gyroscopes to improve the overall accuracy. It indicated that the six-gyroscope array with an ARW noise of 6.17 °/√h and a bias drift of 54.11 °/h were combined into a rate signal having an ARW noise of 1.83 °/√h and a bias drift of 16.32 °/h. The presented KF also reduced the dynamic errors by over 20% compared to the KF modeled by the random walk. It proved that the first-order Markov process is efficient for modeling rate signal to improve the system overall performance.
In the future fabrication of a number of integrated MEMS gyroscope arrays on a single chip would enhance the uniformity between the gyroscopes and the correlation between the gyroscope array still need to be further researched and explored.
[fig] Figure 1: Structure and principle of the virtual gyroscope. [/fig]
[fig] =: H R H . Since 0 P ∞ > , solving Equation (10) and the steady-state covariance can be obtained as [/fig]
[fig] Figure 2: Virtual gyroscope implementation using a discrete KF. [/fig]
[fig] 1 D: − is related to the number of individual gyroscopes, noise statistical quantities of gyroscopes and correlation factors ρ. [/fig]
[fig] Figure 3: Relationship between the gyroscope noise reduction and correlation factor ρ and sensors number N. [/fig]
[fig] Figure 4: A prototype of the virtual gyroscope system. [/fig]
[fig] Figure 5: FFT plot of the virtual gyroscope compared to the single gyroscope and averaging outputs of the gyroscope array. [/fig]
[fig] Figure 6: Allan variance results of the virtual gyroscope compared to the single gyroscope and averaging outputs of the gyroscope array. [/fig]
[fig] Figure 7, Figure 8, Figure 8, Figure 9: Constant rate test of the virtual gyroscope. (a) Outputs of the individual gyroscopes; (b) Outputs of the virtual gyroscope. Random rate test of the virtual gyroscope. (a) Outputs of the individual gyroscopes; (b) Outputs of the virtual gyroscope; (c) Estimated rate errors. ContSinusoidal rate test of the virtual gyroscope. (a) Outputs of the individual gyroscopes; (b) Outputs of the virtual gyroscope; (c) Estimated rate errors. [/fig]
[fig] Figure 9: Cont [/fig]
[table] Table 1: Correlation matrix of noises for six-gyro array. [/table]
[table] Table 2: Static test results of the virtual gyroscope. [/table]
[table] Table 3: Dynamic test results of the virtual gyroscope (unit: °/s). [/table]
|
The Role of Proprotein Convertases in Upper Airway Remodeling
Chronic rhinosinusitis (CRS) is a multifactorial, heterogeneous disease characterized by persistent inflammation of the sinonasal mucosa and tissue remodeling, which can include basal/progenitor cell hyperplasia, goblet cell hyperplasia, squamous cell metaplasia, loss or dysfunction of ciliated cells, and increased matrix deposition. Repeated injuries can stimulate airway epithelial cells to produce inflammatory mediators that activate epithelial cells, immune cells, or the epithelial-mesenchymal trophic unit. This persistent inflammation can consequently induce aberrant tissue remodeling. However, the molecular mechanisms driving disease within the different molecular CRS subtypes remain inadequately characterized. Numerous secreted and cell surface proteins relevant to airway inflammation and remodeling are initially synthesized as inactive precursor proteins, including growth/differentiation factors and their associated receptors, enzymes, adhesion molecules, neuropeptides, and peptide hormones. Therefore, these precursor proteins require post-translational cleavage by proprotein convertases (PCs) to become fully functional. In this review, we summarize the roles of PCs in CRS-associated tissue remodeling and discuss the therapeutic potential of targeting PCs for CRS treatment.
# Introduction
Chronic rhinosinusitis (CRS) is an inflammatory disease of the nose and paranasal sinuses that lasts longer than 12 weeks and is the most common upper respiratory tract disease associated with tissue remodeling. CRS has been divided into two major subtypes based on the presence (CRSwNP) or absence (CRSsNP) of nasal polyps (NPs) [bib_ref] Rhinosinusitis: establishing definitions for clinical research and patient care, Meltzer [/bib_ref]. CRSsNP comprises more than two-thirds of cases and is less likely to be managed by surgical intervention, whereas CRSwNP represents 20%-25% of cases. NPs are outgrowths of swollen inflammatory tissue that infiltrate the middle or superior meatus. They are the most severe form of pathological tissue remodeling in CRS and require surgical intervention.hypothesized that NP pathogenesis involves epithelial rupture and necrosis, leading to protrusions from the lamina propria and epithelial repair. Furthermore, [bib_ref] Excessive fibrin deposition in nasal polyps caused by fibrinolytic impairment through reduction..., Takabayashi [/bib_ref] [bib_ref] Increased expression of factor XIII-A in patients with chronic rhinosinusitis with nasal..., Takabayashi [/bib_ref] clarified that NP growth is due to the deposition of fibrin mesh within the tissue. However, why NPs only develop in some patients with CRS remains unclear.
The recent identification of appropriate CRS biomarkers has revealed new classification methods, such as the characterization of the CRS patient immune response, known as endotyping. Endotypes are classified according to distinct subsets of CD4 + T cells, namely T helper (Th)1, Th2, and Th17 cells, T cell products, infiltrating eosinophilic and noneosinophilic inflammatory cells, and remodeling markers [bib_ref] Immunopathogenesis of chronic rhinosinusitis and nasal polyposis, Schleimer [/bib_ref] [bib_ref] Use of endotypes, phenotypes, and inflammatory markers to guide treatment decisions in..., Staudacher [/bib_ref]. Age, as well as environmental and genetic factors, also influence CRS patient inflammatory endotypes [bib_ref] Increased noneosinophilic nasal polyps in chronic rhinosinusitis in US secondgeneration Asians suggest..., Mahdavinia [/bib_ref] [bib_ref] Associations between inflammatory endotypes and clinical presentations in chronic rhinosinusitis, Stevens [/bib_ref] [bib_ref] Diversity of TH cytokine profiles in patients with chronic rhinosinusitis: a multicenter..., Wang [/bib_ref] [bib_ref] Chronic rhinosinusitis in Asia, Zhang [/bib_ref]. However, some clusters cannot be classified using currently available endotype classification methods, highlighting the necessity for more specific biomarkers and indicating that CRS pathophysiology and pathogenesis remain to be fully understood. Comprehensive tissue remodeling processes, in particular, require further investigation, highlighting the necessity of elucidating regulatory mechanisms underlying tissue remodeling based on endotype classification.
Tissue remodeling is a secondary phenomenon, beginning in early-stage CRS development due to persistent inflammation [bib_ref] Mucosal remodeling and reversibility in chronic rhinosinusitis, Bassiouni [/bib_ref] [bib_ref] The development of nasal polyp disease involves early nasal mucosal inflammation and..., Meng [/bib_ref] [bib_ref] Remodeling and repair in rhinosinusitis, Watelet [/bib_ref]. Tissue remodeling in CRS is the reorganization or renovation of nasal mucosa, which can be either physiological or pathological. Nasal mucosal inflammation induces remodeling processes within the mucosa characterized by changes in extracellular matrix (ECM) protein deposition, macrophage and lymphocyte infiltration, and histological structure. Structural alterations in the nasal epithelium include goblet cell hyperplasia, squamous metaplasia, epithelial-mesenchymal transition (EMT), epithelial barrier disruption, epithelial exfoliation, and basement membrane thickening. Structural changes in the lamina propria include stromal edema, bone thickening, fibrosis, angiogenesis, and submucosal gland hyperplasia. The various CRS subgroups can be differentiated by distinct remodeling features; for example, the eosinophilic forms of both CRSwNP and CRSsNP are characterized by increased edema, resulting in more severe disease presentation than the noneosinophilic forms of CRSwNP, which are characterized by increased glandular hyperplasia and dense collagen deposition . Furthermore, fibrosis and collagen deposition, which are Th-1 biased inflammatory responses, are commonly observed in CRSsNP but not in CRSwNP. However, edema is a prominent feature of Th2-biased eosinophilic inflammation [bib_ref] Epithelial-mesenchymal transition in the pathophysiology of airway remodelling in asthma, Hackett [/bib_ref]. Eosinophilic CRS shows heightened basement membrane thickening compared with noneosinophilic CRS [bib_ref] Distinct patterns of tissue remodeling and their prognostic role in chronic rhinosinusitis, Lee [/bib_ref]. Previous studies have reported differences in tissue remodeling between polyps obtained from white and Asian patients [bib_ref] Features of airway remodeling in different types of Chinese chronic rhinosinusitis are..., Shi [/bib_ref] [bib_ref] Tissue remodeling in chronic rhinosinusitis, Van Bruaene [/bib_ref]. Eosinophilic CRS is typically more common in the EU and USA, whereas noneosinophilic CRS is more common in Asia. However, the prevalence of eosinophilic CRS has increased in Asia due to an increasingly westernized lifestyle. Interestingly, a large histopathologic study of CRS in Wuhan, China, confirmed the link between eosinophilic infiltration and edema and the association of neutrophils with fibrosis [bib_ref] Distinct immunopathologic characteristics of various types of chronic rhinosinusitis in adult Chinese, Cao [/bib_ref].
Numerous mediators are implicated in airway tissue remodeling, including growth factors, enzymes, adhesion molecules, and ECM components [bib_ref] Periostin regulates fibrocyte function to promote myofibroblast differentiation and lung fibrosis, Ashley [/bib_ref] [bib_ref] Mucosal remodeling and reversibility in chronic rhinosinusitis, Bassiouni [/bib_ref] [bib_ref] Periostin as a biomarker for nasal polyps in chronic rhinosinusitis, Maxfield [/bib_ref] [bib_ref] Upper and lower airway remodelling mechanisms in asthma, allergic rhinitis and chronic..., Samitas [/bib_ref] [bib_ref] Inflammation and remodelling patterns in early stage chronic rhinosinusitis, Van Bruaene [/bib_ref] [bib_ref] Remodeling and repair in rhinosinusitis, Watelet [/bib_ref]. For example, insulin-like growth factor-1 (IGF-1) and its receptor are involved in epithelial cell hyperplasia, mucus overproduction, and ECM deposition [bib_ref] Acute inflammation induces insulin-like growth factor-1 to mediate Bcl-2 and Muc5ac expression..., Chand [/bib_ref]. Notch signaling plays a critical role in the lineage selection of airway basal cells (BCs) during differentiation into either secretory or ciliated cells in many adults and embryonic tissues [bib_ref] Concise review: Notch signaling in stem cell systems, Chiba [/bib_ref] [bib_ref] Notch-dependent differentiation of adult airway basal stem cells, Rock [/bib_ref]. However, the sustained activation of Notch signaling promotes the transition of airway BCs to a goblet cell fate [bib_ref] Submersion and hypoxia inhibit ciliated cell differentiation in a Notch-dependent manner, Gerovac [/bib_ref] [bib_ref] Activation of NOTCH1 or NOTCH3 signaling skews human airway basal cell differentiation..., Gomi [/bib_ref] [bib_ref] Notch signaling promotes airway mucous metaplasia and inhibits alveolar development, Guseh [/bib_ref]. Schematic representation of PC processing resulting in upper airway remodeling. Depicted are proforms of the numerous PC substrates that are associated with tissue remodeling, their mature forms, and their effects on upper airway remodeling. [bib_ref] Notch-dependent differentiation of adult airway basal stem cells, Rock [/bib_ref]. Transforming growth factor β (TGF-β), platelet-derived growth factor (PDGF), fibroblast growth factor 2 (FGF-2), vascular endothelial growth factor (VEGF), and matrix metalloproteinase (MMP) all induce the pathological conversion of epithelial cells into fibroblasts, resulting in tissue fibrosis [bib_ref] Epithelial-mesenchymal transition in primary human bronchial epithelial cells is Smad-dependent and enhanced..., Câmara [/bib_ref] [bib_ref] The role of the epithelium in airway remodeling in asthma, Davies [/bib_ref] [bib_ref] Role of basic fibroblast growth factor-2 in epithelial-mesenchymal transformation, Strutz [/bib_ref]. Furthermore, VEGF promotes edema, angiogenesis, and epithelial cell growth in NPs [bib_ref] Vascular endothelial growth factor expression in nasal polyps of aspirin-intolerant patients, Fruth [/bib_ref]. Upregulation of epidermal growth factor (EGF) skews the airway BC fate toward the squamous and EMT-like phenotypes with decreased epithelial junctional barrier integrity [bib_ref] EGF shifts human airway basal cell fate toward a smoking-associated airway epithelial..., Shaykhiev [/bib_ref]. Additionally, bone morphogenetic protein 2 (BMP-2) is associated with both osteitis [bib_ref] Distinct patterns of tissue remodeling and their prognostic role in chronic rhinosinusitis, Lee [/bib_ref] and squamous metaplasia [bib_ref] Proprotein convertase 5/6A is associated with bone morphogenetic protein-2-induced squamous cell differentiation, Lee [/bib_ref] in patients with refractory CRSwNP. Importantly, numerous secreted and cell surface proteins, including the proteins mentioned above, are initially synthesized as inactive precursor proteins, requiring endoproteolytic cleavage by proprotein convertases (PCs) for activation [bib_ref] Proprotein convertases in health and disease, Artenstein [/bib_ref] [bib_ref] Proprotein and prohormone convertases: a family of subtilases generating diverse bioactive polypeptides, Seidah [/bib_ref]. Earlier work by our lab indicated that the expression of four PCs (furin, PC1/3, PC5/6, and PACE4) is significantly upregulated in CRS patient NP mucosa . These results indicate that these enzymes may play important roles in NP pathogenesis. Furthermore, PCs show promise as diagnostic markers for CRS and may ultimately be targeted by molecular therapy. We summarize the general properties and biological relevance of PCs, as well as current discoveries regarding the pathophysiological roles of furin, PC1/3, PC5/6, and PACE4 in CRS.
## Proprotein convertases
PCs are a family of calcium-dependent serine endoproteases. Examples of these enzymes include furin, PC1/3, PC2, PC4, PC5/6, PACE4, and PC7 [bib_ref] Proprotein convertases in health and disease, Artenstein [/bib_ref]. These enzymes activate precursor proteins through cleavage at doublets of the basic amino acids arginine (R) or lysine (K) or at paired basic motifs [R/K-(X)n-R/K-R↓, where the arrow indicates the cleavage site, X represents any amino acid except cysteine and n = 0, 2, 4, or 6] [bib_ref] Proprotein and prohormone convertases: a family of subtilases generating diverse bioactive polypeptides, Seidah [/bib_ref] [bib_ref] Insulin biosynthesis: evidence for a precursor, Steiner [/bib_ref]. In various organs, PCs are essential for key physiological functions, such as embryonic development and tissue homeostasis, due to their involvement in the proteolytic activation of many secretory proteins, including growth/differentiation factors and their receptors, adhesion molecules, enzymes, neuropeptides, and peptide hormones [fig_ref] Table 1: Substrates activated by PCs GDF11, PTPRM, L1CAM, α4 integrin, BMPs In vitro... [/fig_ref] [bib_ref] Proprotein convertases in health and disease, Artenstein [/bib_ref] [bib_ref] Furin at the cutting edge: from protein traffic to embryogenesis and disease, Thomas [/bib_ref] [bib_ref] Genetics of the first seven proprotein convertase enzymes in health and disease, Turpeinen [/bib_ref]. Although PC inactivation in mice and humans has revealed specific phenotypes caused by unique, tissue-specific processing events [bib_ref] The multifaceted proprotein convertases: their unique, redundant, complementary, and opposite functions, Seidah [/bib_ref] , additional investigation into the specific physiologic substrates for PCs is required. Furthermore, some PCs are associated with In vitro and ex vivo redundancy with furin and PC5/6 PC7 Transferrin receptor 1 Partial redundancy with furin, PC5/6, and PACE4 ACTH, adrenocorticotropic hormone; α-MSH, α-melanocytestimulating hormones; IGF, insulin-like growth factor; VWF, Von Willebrand factor; TNF-α, tumor necrosis factor α; TGF-β, transforming growth factor β; IGF-1R, insulin-like growth factor 1 receptor; VEGF, vascular endothelial growth factor; HIV gp 160, human immunodeficiency virus envelope glycoprotein 160; HA, hemagglutinin; GDF11, growth differentiation factor 11; PTPRM, protein tyrosine phosphatase receptor type M; L1CAM, neural cell adhesion molecule L1.
## Fig. 2. pc expression in control nasal mucosa and nasal polyps.
Western blot analysis reveals that furin, PC1/3, PC5/6, and PACE4 are expressed in both control nasal mucosa and nasal polyps, and the expression of four PCs is significantly upregulated in nasal polyps compared to the control mucosa. Right panel shows densitometric analysis of furin, normalized to β-actin and relative to control mucosa. Data represent the mean ± SEM. *P < 0.05.
various pathophysiological states, including endocrinopathies, cancer, viral/bacterial/parasitic infection, atherosclerosis, and neurodegenerative [fig_ref] Table 2: Therapeutic potential of PCs [/fig_ref] [bib_ref] Proprotein convertases in health and disease, Artenstein [/bib_ref] [bib_ref] Proprotein convertases as therapeutic targets, Chrétien [/bib_ref] [bib_ref] The biology and therapeutic targeting of the proprotein convertases, Seidah [/bib_ref] [bib_ref] Furin at the cutting edge: from protein traffic to embryogenesis and disease, Thomas [/bib_ref] [bib_ref] Genetics of the first seven proprotein convertase enzymes in health and disease, Turpeinen [/bib_ref]. These PCs therefore, represent potential therapeutic targets for the treatment of various human diseases. We summarize the reported PC inhibitors that are expected to affect human pathologies [fig_ref] Table 2: Therapeutic potential of PCs [/fig_ref]. It should be noted that none of these PC inhibitors is highly specific to only one PC.
## Inhibition of furin-mediated notch1 processing in airway basal cells promotes ciliated cell differentiation
Airway BCs are a long-lived, multipotent stem cell population responsible for normal epithelium homeostasis and regeneration after injury, which is accomplished through their capacities for self-renewal and differentiation into multiple cell types, including secretory and ciliated cells [bib_ref] Airway basal stem cells: a perspective on their roles in epithelial homeostasis..., Rock [/bib_ref]. However, chronic repetitive injuries disrupt the balance be-tween BC proliferation and differentiation and ultimately lead to pathological tissue remodeling, such as BC hyperplasia, goblet cell hyperplasia, squamous cell metaplasia, loss or dysfunction of ciliated cells, and increased matrix deposition [bib_ref] Airway basal stem cells: a perspective on their roles in epithelial homeostasis..., Rock [/bib_ref] [bib_ref] Upper and lower airway remodelling mechanisms in asthma, allergic rhinitis and chronic..., Samitas [/bib_ref]. These dramatic structural and functional changes contribute to disease susceptibility, initiation, and progression in the airway. Therefore, clarifying the fundamental mechanisms underlying BC lineage choice and differentiation during airway inflammation and remodeling is clinically relevant. The Notch signaling pathway plays an essential role in regulating the differentiation of airway BCs into secretory and ciliated cells in both the developmental and adult phases [bib_ref] Concise review: Notch signaling in stem cell systems, Chiba [/bib_ref] [bib_ref] Notch-dependent differentiation of adult airway basal stem cells, Rock [/bib_ref]. Four mammalian Notch receptors (Notch1-4) have been identified [bib_ref] Concise review: Notch signaling in stem cell systems, Chiba [/bib_ref]. Steady-state Notch signaling is present in relatively few BCs due to low epithelial turnover. However, Notch signaling is greatly increased during repair after epithelial injury [bib_ref] Notch-dependent differentiation of adult airway basal stem cells, Rock [/bib_ref]. In vivo studies have shown that the sustained activation of Notch1 signaling promotes luminal differentiation of airway BCs, primarily toward goblet cell lineages [bib_ref] Submersion and hypoxia inhibit ciliated cell differentiation in a Notch-dependent manner, Gerovac [/bib_ref] [bib_ref] Activation of NOTCH1 or NOTCH3 signaling skews human airway basal cell differentiation..., Gomi [/bib_ref] [bib_ref] Notch signaling promotes airway mucous metaplasia and inhibits alveolar development, Guseh [/bib_ref] [bib_ref] Notch-dependent differentiation of adult airway basal stem cells, Rock [/bib_ref]. A similar result was obtained from in vitro experiments using Notch signaling agonists and antagonists in air-liquid interface (ALI)-human bronchial epithelial cell cultures initiated with BCs [bib_ref] Notch signaling promotes airway mucous metaplasia and inhibits alveolar development, Guseh [/bib_ref]. Importantly, the Notch receptor is activated after cleavage by a furin-like convertase [bib_ref] The Notch1 receptor is cleaved constitutively by a furinlike convertase, Logeat [/bib_ref] [bib_ref] Calcium depletion dissociates and activates heterodimeric Notch receptors, Rand [/bib_ref]. Using an in vitro ALI-human nasal epithelial (HNE) cell culture model of airway injury [bib_ref] Airway epithelial repair, regeneration, and remodeling after injury in chronic obstructive pulmonary..., Puchelle [/bib_ref] [bib_ref] A biphasic chamber system for maintaining polarity of differentiation of culture respiratory..., Whitcutt [/bib_ref] , we found that inhibiting PC activity during BC differentiation using decanoyl-RVKR-chloromethylketone (CMK) treatment [bib_ref] Inhibition of furin-mediated cleavage activation of HIV-1 glycoprotein gpl60, Hallenberger [/bib_ref] skews differentiation toward the ciliated cell phenotype. This skewed differentiation was evidenced by increased numbers of ciliated cells and the upregulation of various genes associated with ciliated cell differentiation. Furthermore, furin knockdown resulted in suppressed Notch1 processing and increased ciliated cell numbers in ALI-HNE cell culture, indicating that furin is the enzyme responsible for Notch1 activation in HNE cells. These observations and previous studies collectively suggest that furin may play a critical role in BC lineage choice toward goblet cell lineages, as well as the pathogenesis of goblet cell hyperplasia during chronic injury. Therefore, targeting furin has potential as an attractive therapeutic approach for airway epithelial repair and regeneration after injury.
## Overexpressed pc1/3 contributes to nasal polypogenesis through emt induction
EMT is a process for an epithelial cell to undergo profound biochemical changes to acquire a mesenchymal phenotype, which includes the loss of epithelial cell-cell junctions, the generation of apicobasal polarity, interactions with the basement membrane, and the upregulation of mesenchymal markers, such as α-smooth muscle actin (α-SMA), vimentin, MMPs, collagen I, and epithelial transcriptional suppressors (Snail and Twist) [bib_ref] Epithelial-mesenchymal transition in primary human bronchial epithelial cells is Smad-dependent and enhanced..., Câmara [/bib_ref] [bib_ref] The role of the epithelium in airway remodeling in asthma, Davies [/bib_ref] [bib_ref] TGF-β-induced EMT: mechanisms and implications for fibrotic lung disease, Willis [/bib_ref]. Disrupting cell-cell adhesions during EMT allows contact between ligand/receptor pairs that do not typically interact due to segregation into either the apical or basolateral membrane domains. Additionally, this disruption initiates signal transduction cascades that affect epithelial cell activation and differentiation, resulting in tissue remodeling [bib_ref] Epithelial barrier function: at the front line of asthma immunology and allergic..., Georas [/bib_ref] [bib_ref] Apicobasal polarization: epithelial form and function, Gibson [/bib_ref] [bib_ref] Segregation of receptor and ligand regulates activation of epithelial growth factor receptor, Vermeer [/bib_ref]. Furthermore, airway epithelial injury and abnormal epithelial repair responses induce persistent epithelial cell activation by undergoing EMT, leading to a pathological process associated with fibrogenesis [bib_ref] Epithelial-mesenchymal transition in the pathophysiology of airway remodelling in asthma, Hackett [/bib_ref] [bib_ref] Induction of epithelial-mesenchymal transition in primary airway epithelial cells from patients with..., Hackett [/bib_ref] [bib_ref] Hypoxia-inducible factor 1 mediates nasal polypogenesis by inducing epithelial-to-mesenchymal transition, Shin [/bib_ref] [bib_ref] TGF-β-induced EMT: mechanisms and implications for fibrotic lung disease, Willis [/bib_ref]. Thus, understanding the precise molecular interactions underlying EMT could lead to the identification of novel therapeutic targets to treat tissue fibrosis in chronic inflammatory airway diseases. [bib_ref] Induction of epithelial-mesenchymal transition in primary airway epithelial cells from patients with..., Hackett [/bib_ref] demonstrated that TP63 + KRT5 + BCs in a multilayered, differentiated ALI-airway epithelial cell culture derived from asthmatic subjects undergo EMT after exposure to TGF-β1, which is a known major inducer of EMT. EMT was evidenced by the loss of epithelial markers, E-cadherin and zonular occludin-1, and the upregulation of mesenchymal markers, EDA-fibronectin, vimentin, α-SMA, and collagen-1 [bib_ref] Induction of epithelial-mesenchymal transition in primary airway epithelial cells from patients with..., Hackett [/bib_ref]. Another study reported that hypoxic conditions present in inflamed sinus tissue drive EMT through a Smad3-dependent mechanism [bib_ref] Hypoxia-inducible factor 1 mediates nasal polypogenesis by inducing epithelial-to-mesenchymal transition, Shin [/bib_ref]. Importantly, results from our lab revealed that both NP epithelium and ALI-HNE cell cultures undergoing TNF-α/ IL-1β-induced EMT highly express PC1/3, together with the mesenchymal marker proteins N-cadherin, collagen I, and MMP-2 [bib_ref] Overexpressed proprotein convertase 1/3 induces an epithelial-mesenchymal transition in airway epithelium, Lee [/bib_ref]. Specifically, PC1/3 expression was mostly confined to the basal and suprabasal layers in healthy control nasal epithelium but was upregulated in the entire NP epithelial layer [bib_ref] Overexpressed proprotein convertase 1/3 induces an epithelial-mesenchymal transition in airway epithelium, Lee [/bib_ref]. Because EMT is intimately linked with the acquisition of epithelial stem cell properties with greater phenotypic plasticity [bib_ref] The epithelialmesenchymal transition generates cells with properties of stem cells, Mani [/bib_ref] , differentiated epithelial cells may dedifferentiate into regressed basal/progenitor cells in disease states [bib_ref] Dedifferentiation of committed epithelial cells into stem cells in vivo, Tata [/bib_ref]. Moreover, overexpressing PC1/3 in stably transfected human lung mucoepidermoid carcinoma HCI-H292 cells resulted in decreased E-cadherin expression and increased mesenchymal marker expression (N-cadherin, vimentin, collagen I, α5 integrin, fibronectin, MMP2, Snail, and Twist), concurrent with the transition to a fibroblast-like morphology driven by actin cytoskeleton remodeling [bib_ref] Overexpressed proprotein convertase 1/3 induces an epithelial-mesenchymal transition in airway epithelium, Lee [/bib_ref]. Taken together, these observations suggest that PC1/3 contributes to tissue remodeling and CRSwNP pathogenesis and play crucial roles in EMT and fibrosis. PC1/3 likely contributes to CRSwNP pathogenesis due to altered processing of integrins, collagen I, fibronectin, neuropeptides, and MMPs [bib_ref] Proprotein convertases in health and disease, Artenstein [/bib_ref] [bib_ref] Elevated expression of proprotein convertases alters breast cancer cell growth in response..., Cheng [/bib_ref]. Further research is required to fully define the precise molecular mechanisms underlying PC1/3-mediated EMT. Greater clarification is also needed to identify the physiological PC1/3 substrates that could provide new therapeutic targets for CRSwNP treatment.
## Pc5/6a promotes the squamous differentiation of human nasal epithelial cells by activating bmp-2
Squamous metaplasia of the airway is a pathologic process by which normal, pseudostratified epithelium transdifferentiates into stratified epithelium consisting of flattened, squamous cells [bib_ref] Changes in bronchial epithelium in relation to cigarette smoking and in relation..., Auerbach [/bib_ref] [bib_ref] Airway epithelial repair, regeneration, and remodeling after injury in chronic obstructive pulmonary..., Puchelle [/bib_ref] [bib_ref] Airway basal stem cells: a perspective on their roles in epithelial homeostasis..., Rock [/bib_ref]. Thus, squamous differentiation in airway epithelial cells and epidermal differentiation share many morphological and biochemical characteristics . Chronic repetitive injuries to the airway epithelium induce tissue remodeling, such as epithelial cell hyperproliferation and squamous metaplasia, resulting in impaired mucociliary clearance [bib_ref] Airway epithelial repair, regeneration, and remodeling after injury in chronic obstructive pulmonary..., Puchelle [/bib_ref]. Interestingly, results from our lab revealed the significant upregulation of PC5/6A and BMP-2 in both the metaplastic squamous epithelium of NPs and a retinoic acid (RA) deficiency-induced squamous metaplasia model of ALI-HNE cells [bib_ref] Proprotein convertase 5/6A is associated with bone morphogenetic protein-2-induced squamous cell differentiation, Lee [/bib_ref]. RA deficiency is well-known to induce conversion from normal pseudostratified epithelium into stratified squamous airway epithelium [bib_ref] Restoration of mucociliary tracheal epithelium following deprivation of vitamin A. A quantitative..., Mcdowell [/bib_ref] [bib_ref] Tissue changes following deprivation of fat-soluble A vitamin, Wolbach [/bib_ref] [bib_ref] Secretory differentiation of serially passaged normal human nasal epithelial cells by retinoic..., Yoon [/bib_ref]. In a study by [bib_ref] Proprotein convertase expression and localization in epidermis: evidence for multiple roles and..., Pearton [/bib_ref] , four PCs, including furin, PACE4, PC5/6, and PC7, had significant roles in terminal keratinocyte differentiation in the epidermis. Additionally, BMP signaling is implicated in pathophysiological processes including wound-healing, fibrosis, and allergic inflammation in the skin and lungs [bib_ref] Bone morphogenetic proteins and their antagonists in skin and hair follicle biology, Botchkarev [/bib_ref] [bib_ref] Activation of bone morphogenetic protein/Smad signaling in bronchial epithelial cells during airway..., Rosendahl [/bib_ref] [bib_ref] Activation of the canonical bone morphogenetic protein (BMP) pathway during lung morphogenesis..., Sountoulidis [/bib_ref] [bib_ref] Epithelial to mesenchymal transition in human skin wound healing is induced by..., Yan [/bib_ref]. BMP signaling is also known to be involved in the regulation of embryonic development and adult homeostasis [bib_ref] Bone morphogenetic proteins and their antagonists in skin and hair follicle biology, Botchkarev [/bib_ref] [bib_ref] Activation of the canonical bone morphogenetic protein (BMP) pathway during lung morphogenesis..., Sountoulidis [/bib_ref]. [bib_ref] Epithelial to mesenchymal transition in human skin wound healing is induced by..., Yan [/bib_ref] demonstrated that TNF-α-induced EMT is mediated by the BMP-2 signaling pathway in wound-healing and fibrosis of human skin. A study from the Zou lab revealed the induction of a smoking-related abnormal phenotype in human airway BCs mediated by exaggerated BMP-4/BMPR1A/Smad signaling, generating squamous metaplasia [bib_ref] Exaggerated BMP4 signalling alters human airway basal progenitor cell differentiation to cigarette..., Zuo [/bib_ref]. Importantly, BMPs are known to be physiological substrates for PACE4 and PC5/6A [fig_ref] Table 1: Substrates activated by PCs GDF11, PTPRM, L1CAM, α4 integrin, BMPs In vitro... [/fig_ref] [bib_ref] SPC4, SPC6, and the novel protease SPC7 are coexpressed with bone morphogenetic..., Constam [/bib_ref] [bib_ref] Secretory proprotein convertases PACE4 and PC6A are heparin-binding proteins which are localized..., Tsuji [/bib_ref]. Our lab demonstrated that PC5/6A knockdown and pharmacological inhibition of PC activity in RA-deficient ALI cultures resulted in significant reductions in BMP-2 protein expression and processing, accompanied by the downregulation of squamous cell marker genes (cornifin/SPRR1 and involucrin) and the upregulation of secretory (MUC5AC, TFF3, and MUC5B) and ciliated cell marker genes (Tektin and DNAI1) [bib_ref] Proprotein convertase 5/6A is associated with bone morphogenetic protein-2-induced squamous cell differentiation, Lee [/bib_ref]. Conversely, PC5/6A overexpression using adenoviral-mediated transduction and exogenous BMP-2 resulted in the upregulation of squamous cell marker genes and the broad downregulation of ciliated and secretory cell differentiation genes [bib_ref] Proprotein convertase 5/6A is associated with bone morphogenetic protein-2-induced squamous cell differentiation, Lee [/bib_ref] under RA-sufficient culture conditions for the mucociliary differentiation of HNE cells [bib_ref] Secretory differentiation of serially passaged normal human nasal epithelial cells by retinoic..., Yoon [/bib_ref]. These observations suggest that PC5/6A-mediated BMP-2 maturation contributes to squamous metaplasia on the NP mucosal surface. Furthermore, [bib_ref] Distinct patterns of tissue remodeling and their prognostic role in chronic rhinosinusitis, Lee [/bib_ref] reported that in CRSwNP, BMP-2 is upregulated in NP tissues, associated with osteitis severity, advanced disease extent, and disease refractoriness after surgery. Taken together, these results indicate that PC5/6A can serve as a new CRSwNP biomarker reflecting the pathophysiology of nasal mucosa with squamous metaplasia. Targeting PC5/6A may, therefore, be a viable therapeutic strategy for treating refractory CRSwNP.
## Pace4 upregulation is associated with airway goblet cell hyperplasia
Goblet cell hyperplasia is a common feature of chronic airway diseases, which include asthma, allergic rhinitis, and CRSwNP [bib_ref] Airway goblet-cell mucus secretion, Jackson [/bib_ref] [bib_ref] Nasal mucus proteome and its involvement in allergic rhinitis, Tomazic [/bib_ref]. Recent unpublished results from our lab revealed that in the human nasal epithelium within a Th2 milieu, PACE4 upregulation is associated with goblet cell hyperplasia and mucus overproduction. In both NP epithelium and ALI-HNE cells treated with the Th2 cytokine IL-4, which induces goblet cell hyperplasia [bib_ref] SPDEF regulates goblet cell hyperplasia in the airway epithelium, Park [/bib_ref] , PACE4 expression was mostly confined to the basal and suprabasal layers. Furthermore, PACE4 knockdown in ALI-cultured BCs inhibited IL-4-induced goblet cell differentiation, which implies this enzyme is an attractive therapeutic target for CRSwNP treatment. Supporting this finding, a microarray-based study of the transcriptomes of eosinophilic CRSwNP (ECRSwNP) and noneosinophilic CRSwNP (non-ECRSwNP) showed that mRNA levels of Th2 cytokines and PCSK6, which is the gene PACE4, are significantly increased in ECRSwNP [bib_ref] Transcriptome analysis reveals distinct gene expression profiles in eosinophilic and noneosinophilic chronic..., Wang [/bib_ref]. These results suggest PACE4 involvement in Th2 inflammation in ECRSwNP. ECRSwNP exhibits a poorer outcome compared to non-ECRSwNP. Indeed, ECRSwNP exhibits greater objective disease severity and a high recurrence rate after surgery [bib_ref] Mucosal eosinophilia and recurrence of nasal polyps-new classification of chronic rhinosinusitis, Nakayama [/bib_ref]. Therefore, PACE4 may increase the risk of ECRSwNP, making it a potential diagnostic and prognostic biomarker and treatment target. Further investigation into the mechanism of how PACE4 is involved in ECRSwNP and the potential therapeutic benefits of targeting PACE4 in ECRSwNP is required.
# Conclusion
There is still much more to learn about pathological endotyping or subphenotyping of tissue remodeling features in CRS patients, which has the potential to identify patients at a higher risk of recurrent or persistent disease. Here, we assert that PCs have crucial impacts on various types of CRS pathological tissue remodeling, including goblet cell hyperplasia, fibrosis, and squamous metaplasia. Therefore, PCs could be considered promising diagnostic and prognostic biomarkers in CRS patients. Targeting PCs has great potential to treat CRS. However, PC substrate specificity remains unknown in both the physiological and pathophysiological context. This lack of knowledge is largely due to substantial redundancies in the substrates and functions among PCs and the co-expression of some PCs in cells. Therefore, further studies to elucidate the precise mechanisms of PC activity and PC substrate specificity in tissue remodeling and CRS pathogenesis will enable the development of specific biomarkers for disease progression and more individualized treatment strategies. The challenge of identifying potent and safe PC inhibitors has great potential to yield an alternative CRS therapeutic option that could ultimately improve human health.
[table] Table 1: Substrates activated by PCs GDF11, PTPRM, L1CAM, α4 integrin, BMPs In vitro and ex vivo redundancy with furin and PACE4 PACE4 Nodal, Lefty, L1CAM, MMPs, BMPs [/table]
[table] Table 2: Therapeutic potential of PCs [/table]
|
Study of the Impact of Industrial Restructuring on the Spatial and Temporal Evolution of Carbon Emission Intensity in Chinese Provinces—Analysis of Mediating Effects Based on Technological Innovation
# Introduction
Since the industrial revolution, the world has witnessed rapid socioeconomic development, and the concentration of greenhouse gases such as carbon dioxide and the global average temperature have also risen sharply [bib_ref] Global warming and economic externalities, Rezai [/bib_ref] [bib_ref] Increase of the transient climate response to cumulative carbon emissions with decreasing..., Tachiiri [/bib_ref]. According to the "State of the Global Climate 2020" released by the World Meteorological Organization, the global average temperature in 2020 increased by about 1.2 - C compared to before the industrial revolution. Climate warming has an irreversible multi-dimensional impact on the earth's ecosystem and human habitat. The frequency and intensity of extreme weather events such as sea level rise, heat waves, droughts, and floods continue to increase. Together, these 2 of 18 factors lead to potential conflict, friction, and even war between major global powers [bib_ref] Optimal GHG Emission Abatement and Aggregate Economic Damages of Global Warming, Kanellos [/bib_ref] [bib_ref] Shortened Duration of Global Warming Slowdowns with Elevated Greenhouse Gas Emissions, Gao [/bib_ref]. China's carbon emissions have remained a challenging issue [bib_ref] Driving factors of carbon emissions in China's municipalities: A LMDI approach, Liu [/bib_ref]. According to the relevant statistics provided in the BP Statistical Review of World Energy 2020, China's total carbon emissions in 2019 ranked first in the world, accounting for 28.76% of the world's total, more than the sum of the EU and the United States' carbon emissions (24.28%). Facing the huge pressure of carbon reduction, China took the initiative to undertake international responsibilities in line with its national conditions and made a series of commitments to the international community. At the Paris Climate Conference in 2015, China proposed to achieve a carbon peak around 2030 and reduce the carbon emissions per unit GDP by 60%-65% compared with 2005 [bib_ref] Climate risk and corporate environmental performance: Empirical evidence from China, Ren [/bib_ref]. In 2020, President Xi Jinping promised in the general debate of the 75th United Nations General Assembly that China would strive to reach the peak of carbon emissions by 2030 and achieve carbon neutrality by 2060 [bib_ref] Towards carbon neutrality: A study on China's long-term low-carbon transition pathways and..., He [/bib_ref].
Since the reform and opening up, China's economic development has relied on an energy structure dominated by coal, and the industrial structure has become more industrialized, with high carbon emissions [bib_ref] The Impact of Rationalization and Upgrading of Industrial Structure on Carbon Emissions..., Gu [/bib_ref]. According to some scholars, the contribution rate of China's industrial restructuring to the achievement of carbon intensity goals is more than 70% [bib_ref] Adjustment of industrial structure and the potential assessment of energy saving and..., Wang [/bib_ref]. Industrial restructuring, together with more environmentally friendly and efficient energy-saving and emission-reduction technologies, can bring about advanced and new production technologies. The improvement of production technology and the increased technical content of products can effectively improve production efficiency, elevate energy efficiency, reduce factor inputs and pollution emissions per unit of output, and weaken the negative environmental externalities caused by production processes [bib_ref] Research on the spatial effect and threshold effect of industrial structure upgrading..., Li [/bib_ref] [bib_ref] Have carbon emissions been reduced due to the upgrading of industrial structure?..., Zhu [/bib_ref]. Therefore, industrial restructuring is one of the important means to achieve economic transformation and reduce carbon emissions [bib_ref] The nonlinear impact of industrial restructuring on economic growth and carbon dioxide..., Zhou [/bib_ref]. China's scientific and technological innovation capabilities have improved in recent years, making it the first driver for high-quality economic development. Technological innovation can reconfigure production factors, change the existing design, production processes, and technological processes of enterprises, develop and utilize low-carbon clean technologies, promote clean energy and new energy, improve the recycling efficiency of existing resources and energy, and reduce the emission intensity of unit carbon dioxide [bib_ref] The effects of environmental regulation and industrial structure on carbon dioxide emission:..., Chen [/bib_ref]. In view of this, this study takes 30 provinces in China as the research object (except Hong Kong, Macau, Taiwan and Tibet) and analyzes the relevant literature and theories in order to achieve the following objectives: this study takes carbon emission intensity as the explanatory variable, industrial structural transformation in each province as the core explanatory variable, and science and technology innovation as the mediating variable, and selects economic development, population size, urbanization rate and environmental regulation as control variables to clarify the direct and indirect effects of industrial structural transformation on carbon emission intensity, explore whether industrial structural transformation can achieve energy saving and emission reduction through science and technology innovation, and analyze the mediating role played by science and technology innovation in the model.
The remainder of this article is organized as follows: Section 2 reviews the relevant literature; Section 3 presents the theoretical analysis and research hypotheses; Section 4 describes the data sources and evaluation models used for the present study; Section 5 exhibits the spatiotemporal evolution characteristics of China's carbon emission intensity; Section 6 analyzes the research results; and Section 7 presents the conclusions and recommendations.
## Literature review
With the increasingly severe greenhouse gas effect, carbon emissions have drawn widespread attention from governments and scholars worldwide. Promoting industrial restructuring and improving technological innovation are considered the main measures to achieve energy conservation, emission reduction, and transformation of economic development models. At present, scholarly research on this aspect mainly focuses on the following two aspects:
The impact of industrial restructuring on carbon emissions has been analyzed mainly from the following three aspects. [bib_ref] Global warming and economic externalities, Rezai [/bib_ref] The relationship between industrial restructuring and carbon emissions: most researchers believe that industrial restructuring can significantly reduce carbon emission intensity and optimize the regional environment [bib_ref] Industrialization, servicification, and environmental Kuznets curve: Nonlinear panel regression analysis, Tasdemir [/bib_ref] , while some scholars hold that there is a significant nonlinear relationship between industrial structure and carbon emission intensity [bib_ref] Uncovering the impacts of industrial transformation on low-carbon development in the Yangtze..., Zhang [/bib_ref]. [bib_ref] Increase of the transient climate response to cumulative carbon emissions with decreasing..., Tachiiri [/bib_ref] The methods used to explore the impact of industrial restructuring on carbon emissions: the widely used method is the LMDI decomposition model. For example, Simbi et al. used this method to examine the main sources and potential driving forces of carbon dioxide in 20 African countries from 1984 to 2014. They found that the industrial structure and emission efficiency partially offset carbon dioxide growth [bib_ref] Decomposition and decoupling analysis of carbon dioxide emissions in African countries during..., Habimana Simbi [/bib_ref]. The second approach uses a panel data model, spatial econometric model, and other measurement methods to empirically analyze the impact of industrial restructuring on carbon emissions [bib_ref] Industrial output, services and carbon emissions: The role of information and communication..., Nwani [/bib_ref] [bib_ref] Can Industrial Structural Adjustment Improve the Total-Factor Carbon Emission Performance in China?, Cheng [/bib_ref]. For example, Cheng et al. utilized a dynamic spatial panel model to show spatial differences in the impact of industrial structure on China's carbon intensity [bib_ref] Industrial structure, technical progress and carbon intensity in China's provinces, Cheng [/bib_ref]. Other researchers have looked at the carbon reduction effect of industrial restructuring using the input-output table, STIRPAT, and other models [bib_ref] Realizing low-carbon development in a developing and industrializing region: Impacts of industrial..., Tian [/bib_ref] [bib_ref] Examining industrial structure changes and corresponding carbon emission reduction effect by combining..., Li [/bib_ref] [bib_ref] Structural decomposition of variations of carbon dioxide emissions for the United States,..., Sesso [/bib_ref].
(3) The impact of industrial restructuring on carbon emissions at different scales: Scholars have addressed this issue from the global, national, regional, and urban scales [bib_ref] Global characteristics and trends of research on industrial structure and carbon emissions:..., Sun [/bib_ref] [bib_ref] Addressing Key Drivers of Regional CO 2 Emissions of the Manufacturing Industry..., Matsumoto [/bib_ref] [bib_ref] Investigating whether the environmental Kuznets curve hypothesis holds for sectoral CO 2..., Htike [/bib_ref].
The research on the impact of technological innovation on carbon emissions can be reviewed from the following three aspects. (1) The carbon reduction effects of different technological innovations [bib_ref] Do innovation in environmental-related technologies asymmetrically affect carbon dioxide emissions in the..., Xin [/bib_ref] [bib_ref] The effects of environmental innovations on CO 2 emissions: Empirical evidence from..., Mongo [/bib_ref] : for example, You et al. investigated the direct and indirect effects of technological progress types on industrial carbon efficiency based on industrial data from 30 provinces in China from 2001 to 2016. From the perspective of direct effects, energy technology progress is more conducive to improving carbon efficiency than carbon emission technology progress, and neutral technological progress is more beneficial to carbon efficiency than capital-implied technological progress. In terms of indirect effects, the progress of capital technology affects the improvement of carbon efficiency. The progress of energy technology has a positive and significant impact on carbon efficiency through the green upgrade of the industrial structure. [bib_ref] Increase of the transient climate response to cumulative carbon emissions with decreasing..., Tachiiri [/bib_ref] The impact of technological innovation on carbon emissions in different industries [bib_ref] A comparative analysis of the relationship between innovation and transport sector carbon..., Demircan Cakar [/bib_ref] [bib_ref] Dynamic Nexus between Technological Innovation and Building Sector Carbon Emissions in the..., Erdogan [/bib_ref] : for example, Erdogan et al. believe that long-term innovation has no significant impact on carbon emissions in energy, transportation, and other industries; the improvement of industrial innovation can reduce carbon emissions, while construction innovation will increase carbon emissions [bib_ref] The effects of innovation on sectoral carbon emissions: Evidence from G20 countries, Erdogan [/bib_ref]. (3) The impact mechanism of technological innovation on carbon emissions at different scales [bib_ref] Do globalization and renewable energy contribute to carbon emissions mitigation in Sub-Saharan..., Acheampong [/bib_ref] [bib_ref] Consumption-based and territory-based carbon emissions intensity: Determinants and forecasting using club convergence..., Bhattacharya [/bib_ref] [bib_ref] Renewable energy, technological innovation and the environment: A novel dynamic auto-regressive distributive..., Ulucak [/bib_ref] : for instance, Obobisa et al. used panel time series data from 2000 to 2018 to examine the impact of green technology innovation on CO 2 emissions in 25 African countries. The estimated results showed that green technology innovation significantly reduced carbon emissions in these countries [bib_ref] The impact of green technological innovation and institutional quality on CO 2..., Obobisa [/bib_ref].
To sum up, scholars have conducted fruitful work both theoretically and practically on the impact and measurement of industrial restructuring and technological innovation on carbon emissions. However, most studies still focus on the relationship between industrial structure, technological innovation, and carbon emissions, and few have incorporated the three into one analytical framework. Due to the significant impact of industrial restructuring on technological innovation, the interaction path between the two has an important impact on reducing carbon emission intensity. Given this, the present study measured China's carbon emission intensity, analyzed its spatiotemporal evolution characteristics, and accounted for the direct and indirect impact of industrial restructuring and technological innovation on carbon emission intensity using provincial panel data from 2001 to 2019. This study makes contributions from the following two aspects. Firstly, it incorporates industrial restructuring, technological innovation, and carbon emission intensity into one analytical framework and illuminates their relationships from the perspectives of industrial structure rationalization and optimization, which is theoretically helpful to clarify the mechanisms of action among the three. Secondly, this study employs the fixed effect model and the mediating effect model to empirically investigate the direct impact of industrial restructuring on carbon emissions and the mediating role of technological innovation in industrial restructuring affecting carbon emission intensity.
## Theoretical analysis and research hypotheses
## Direct mechanism of industrial restructuring on carbon emission intensity
Industrial restructuring refers to the process or trend in which the industrial structure develops from unreasonable to rational and from low-level to high-level. It is the organic unity of industrial structure rationalization and industrial structure optimization. Therefore, this study analyzed the impact of industrial structure on carbon emission intensity from the perspectives of rationalization and optimization [bib_ref] The Impact of Rationalization and Upgrading of Industrial Structure on Carbon Emissions..., Gu [/bib_ref]. In terms of the impact of industrial structure rationalization on carbon emission intensity, as resources flow from sectors with lower productivity to sectors with higher productivity, low-end enterprises with extensive production methods and serious ecological damage in the industrial structure are eliminated, and they gradually evolve to eco-friendly, high-tech industrial sectors [bib_ref] The effects of environmental regulation and industrial structure on carbon dioxide emission:..., Chen [/bib_ref] [bib_ref] Examining industrial structure changes and corresponding carbon emission reduction effect by combining..., Li [/bib_ref]. By adjusting the ratio between industries, the proportion of resource-intensive industries, labor-intensive industries, and other high-energy-consuming industries tends to decrease; the shares of technology-and knowledge-intensive industries are increased; resources within the industries are rationally allocated and effectively utilized, thereby promoting the production efficiency and resource utilization efficiency of enterprises and reducing the carbon emission intensity per unit product [bib_ref] Industrial structure, technical progress and carbon intensity in China's provinces, Cheng [/bib_ref] [bib_ref] The impact of economic structure to the environmental Kuznets curve (EKC) hypothesis:..., Dogan [/bib_ref]. Regarding the impact of industrial structure optimization on carbon emission intensity, since most energy-intensive sectors are concentrated within the secondary industry, its carbon emissions are also the highest among the three industries [bib_ref] Can Industrial Structural Adjustment Improve the Total-Factor Carbon Emission Performance in China?, Cheng [/bib_ref]. As the industrial structure shifts towards low carbonization and service orientation, the output value of the tertiary industry continues to increase; the output value of the secondary industry gradually declines; the energy utilization efficiency improves; the consumption and demand for energy resources due to economic growth decrease; and the carbon emission intensity declines. On the other hand, within the industrial structure, traditional industries and low-tech industries have enormous energy demands and significant carbon emissions due to the low level of technological innovation. Industrial structure optimization promotes the internal upgrading of the industrial structure whereby emerging and high-tech industries gradually replace traditional and low-tech industries. Production efficiency and energy utilization efficiency are correspondingly improved [bib_ref] Have carbon emissions been reduced due to the upgrading of industrial structure?..., Zhu [/bib_ref]. Therefore, without considering other goals, carbon emissions can be effectively reduced by restricting high-emitting industries and expanding low-emitting industries [bib_ref] Resource abundance, industrial structure, and regional carbon emissions efficiency in China, Wang [/bib_ref]. Therefore, we propose Hypothesis 1: Industrial structure rationalization and optimization have a significantly inhibitory effect on carbon emission intensity.
## Indirect mechanism of industrial restructuring on carbon emission intensity
Industrial restructuring not only has a direct impact on carbon emission intensity but also indirectly affects it through technological innovation. From the perspective of factor allocation, the transformation and upgrading of industrial structure promote the reorganization and reallocation of production factors among different industries and sectors. Under the guidance of current policies and market mechanisms, production factors gradually shift from low-productivity sectors to high-productivity sectors [bib_ref] The cyclical impact of innovation in green and sustainable technologies on carbon..., Khattak [/bib_ref]. In this process, the spatial reset and circulation of innovation resources will inevitably affect the efficiency of technological innovation. Sectors with high productivity often enjoy higher technical levels. When production factors gradually flow to these sectors, along with the circulation and concentration of innovation resources, the technical level between industries is further improved, thus driving the rapid development of high-tech industries. Industrial input has changed from high-carbon input to low-carbon input, and improving industrial output efficiency leads to an increase in unit energy output benefits, thereby reducing carbon emissions [bib_ref] The effects of environmental innovations on CO 2 emissions: Empirical evidence from..., Mongo [/bib_ref]. From the viewpoint of industrial linkages, the transformation and upgrading of industrial structure enhance technological innovation within and between industries through the accumulation of capital, labor, and technology flows [bib_ref] Do innovation in environmental-related technologies asymmetrically affect carbon dioxide emissions in the..., Xin [/bib_ref]. Meanwhile, industrial restructuring has a remarkable radiation effect on the upstream and downstream industries through the correlation effect, improving the technological innovation capabilities of the upstream and downstream industries, resulting in the optimization of the industrial input-output structure of enterprises, and substantially reducing the factor inputs and carbon emissions per unit output [bib_ref] Do environmental technologies help to reduce transport sector CO 2 emissions? Evidence..., Alataş [/bib_ref]. From the standpoint of supply and demand, industrial restructuring promotes the upgrading of residents' consumption demand, which increases the market demand for products with good environmental quality and high technology content. Driven by the market and profits, enterprises will increase the intensity of R&D investment, enhance technological innovation, upgrade existing production equipment and production processes, introduce new, more advanced and environmentally friendly technologies and methods, and launch new processes and new products to meet the diversified and personalized product demand of the market. This will promote the transformation of enterprises from the original factor-driven model to the innovation-driven development model and indirectly promote carbon reduction [bib_ref] The cyclical impact of innovation in green and sustainable technologies on carbon..., Khattak [/bib_ref] [bib_ref] How do environmental innovations and energy productivity affect the environment? Analyzing the..., Ahmad [/bib_ref].
Therefore, we propose Hypothesis 2: Industrial restructuring promotes technological innovation and acts on carbon emission intensity.
In summary, the influence path of industrial restructuring and technological innovation on carbon emissions can be summarized as the positive transmission mechanism of "industrial restructuring → technological innovation → carbon emission intensity reduction", and the research logic model is shown in [fig_ref] Figure 1: The research logic model [/fig_ref].
further improved, thus driving the rapid development of high-tech industries. Industrial input has changed from high-carbon input to low-carbon input, and improving industrial output efficiency leads to an increase in unit energy output benefits, thereby reducing carbon emissions [bib_ref] The effects of environmental innovations on CO 2 emissions: Empirical evidence from..., Mongo [/bib_ref]. From the viewpoint of industrial linkages, the transformation and upgrading of industrial structure enhance technological innovation within and between industries through the accumulation of capital, labor, and technology flows [bib_ref] Do innovation in environmental-related technologies asymmetrically affect carbon dioxide emissions in the..., Xin [/bib_ref]. Meanwhile, industrial restructuring has a remarkable radiation effect on the upstream and downstream industries through the correlation effect, improving the technological innovation capabilities of the upstream and downstream industries, resulting in the optimization of the industrial input-output structure of enterprises, and substantially reducing the factor inputs and carbon emissions per unit output [bib_ref] Do environmental technologies help to reduce transport sector CO 2 emissions? Evidence..., Alataş [/bib_ref]. From the standpoint of supply and demand, industrial restructuring promotes the upgrading of residents' consumption demand, which increases the market demand for products with good environmental quality and high technology content. Driven by the market and profits, enterprises will increase the intensity of R&D investment, enhance technological innovation, upgrade existing production equipment and production processes, introduce new, more advanced and environmentally friendly technologies and methods, and launch new processes and new products to meet the diversified and personalized product demand of the market. This will promote the transformation of enterprises from the original factor-driven model to the innovation-driven development model and indirectly promote carbon reduction [bib_ref] The cyclical impact of innovation in green and sustainable technologies on carbon..., Khattak [/bib_ref] [bib_ref] How do environmental innovations and energy productivity affect the environment? Analyzing the..., Ahmad [/bib_ref].
Therefore, we propose Hypothesis 2: Industrial restructuring promotes technological innovation and acts on carbon emission intensity.
In summary, the influence path of industrial restructuring and technological innovation on carbon emissions can be summarized as the positive transmission mechanism of "industrial restructuring → technological innovation → carbon emission intensity reduction", and the research logic model is shown in [fig_ref] Figure 1: The research logic model [/fig_ref].
## Industrial restructuring technological innovation carbon emission intensity
Path 2：Promote Path 2：Inhibit
Path1：Direct Inhibitory [fig_ref] Figure 1: The research logic model [/fig_ref]. The research logic model.
## Research design
## Modeling
In this study, the fixed-effect panel model was used to test the direct impact mechanism of industrial restructuring on carbon emission intensity. The benchmark regression model for the test is as follows:
[formula] 0 1 2 it it it i it CI INS X α α α μ ε = + + + + (1)(1) [/formula]
In Equation (1), CIit represents the carbon emission intensity of province i in year t; INSit stands for industrial restructuring, which is decomposed into industrial structure rationalization (RIS) and industrial structure optimization (OIS); α0 refers to the intercept; α1 and α2 are the regression coefficients of INS and X, respectively; μi expresses the individual effect; εit is the random disturbance term; X means the relevant control variable.
As mentioned above, industrial restructuring not only directly impacts carbon emission intensity but may also indirectly impact the reduction of carbon emission intensity through technological innovation. To confirm whether technological innovation plays a mediating role in this process, the following mediating effect model was constructed about the study by Wen et al. [bib_ref] Monotonicity of effect sizes: Questioning kappa-squared as mediation effect size measure, Wen [/bib_ref].
## Research design
## Modeling
In this study, the fixed-effect panel model was used to test the direct impact mechanism of industrial restructuring on carbon emission intensity. The benchmark regression model for the test is as follows:
[formula] CI it = α 0 + α 1 I NS it + α 2 X it + µ i + ε it(1) [/formula]
In Equation (1), CI it represents the carbon emission intensity of province i in year t; INS it stands for industrial restructuring, which is decomposed into industrial structure rationalization (RIS) and industrial structure optimization (OIS); α 0 refers to the intercept; α 1 and α 2 are the regression coefficients of INS and X, respectively; µ i expresses the individual effect; ε it is the random disturbance term; X means the relevant control variable.
As mentioned above, industrial restructuring not only directly impacts carbon emission intensity but may also indirectly impact the reduction of carbon emission intensity through technological innovation. To confirm whether technological innovation plays a mediating role in this process, the following mediating effect model was constructed about the study by Wen et al. [bib_ref] Monotonicity of effect sizes: Questioning kappa-squared as mediation effect size measure, Wen [/bib_ref].
[formula] TEC it = β 0 + β 1 I NS it + β 2 X it + µ i + ε it(2)CI it = λ 0 + λ 1 I NS it + λ 2 TEC it + λ 3 X it + µ i + ε it(3) [/formula]
In Equations (2) and (3), TEC it represents the technological innovation in year t of province i. Equation (2) is dedicated to the impact of industrial restructuring on technological innovation, and Equation (3) focuses on the combined impact of industrial restructuring and technological innovation on carbon emission intensity. If the coefficients β 1 and λ 2 in Equations (2) and (3) are significant at the same time, technological innovation is a mediating variable; if λ 1 is not significant, technological innovation is a complete mediating variable; if λ 1 is significant and is smaller than λ 3 , technological innovation is a partial mediating variable. The difference between the partial mediating effect and the complete mediating effect is that the core explanatory variable of the former can directly affect the explained variable. At the same, the latter must rely on the action of the mediating variable.
## Variables
## Explained variables
Carbon emission intensity (CI): Carbon intensity refers to carbon emissions per unit GDP. To date, there has been sufficient research on carbon emissions, but carbon intensity can reflect the actual economic development and CO 2 emissions of a country (region) more scientifically than carbon emissions [bib_ref] Spatial effects of carbon emission intensity and regional development in China, Wang [/bib_ref]. There is no unified measurement method for carbon emissions. Therefore, the annual carbon emissions of each province [fig_ref] Table 1: The carbon emission coefficients of various fossil fuels [/fig_ref] were estimated based on the carbon emission coefficients of eight types of fossil fuels provided in the "2006 IPCC Guidelines for National Greenhouse Gas Inventories" and "China Energy Statistical Yearbook" as well as the energy consumption of fossil fuels in each province using Equation [bib_ref] Optimal GHG Emission Abatement and Aggregate Economic Damages of Global Warming, Kanellos [/bib_ref]. On this basis, the carbon emissions of each province (city and autonomous region) were obtained. In Equation (4), C co2 is the carbon dioxide emissions; k (k = 44/12) refers to the carbon dioxide to carbon molecular weight ratio; E i represents the consumption of the i th fossil fuel; δ i stands for the emission coefficient of the i th fossil fuel.
[formula] C co 2 = k · n ∑ i=1 E i · δ i (4) [/formula]
## Explanatory variables
Industrial restructuring: Industrial restructuring was decomposed into two parts: industrial structure optimization (OIS) and industrial structure rationalization (RIS) [bib_ref] How to achieve a win-win situation between economic growth and carbon emission..., Dong [/bib_ref]. RIS is expressed by the Theil index using the following Equation:
[formula] RIS = n ∑ i=1 Y i Y ln Y i L i Y L = n ∑ i=1 Y i Y ln Y i Y i L i L (5) [/formula]
In Equation (5), Y is the industrial output value; L represents the number of employees in the industry; i stands for the industry; n expresses the number of industrial sectors; Y/L stands for the productivity; Y i /Y means the output structure; L i /L refers to the employment structure. According to the basic assumptions of classical economics, when the economy is in equilibrium, the efficiency of industrial sectors is at an equalized level, Y i /L i = Y/L, namely, RIS = 0. Conversely, if RIS is not equal to 0, the more the industrial structure deviates from the equilibrium, the more unreasonable the industrial structure is. Since this indicator is negative, it was reciprocated in the subsequent regression. The ratio of the added value of the tertiary industry and that of the secondary industry in a certain region was used as a proxy indicator for OIS. If the value is rising, the regional economy is moving towards servitization, and the industrial structure is more advanced.
## Mediating variables
Technological innovation (TEC): Technological innovation refers to the invention and creation ability of a region, enterprise, or individual in a certain technical field [bib_ref] Do innovation in environmental-related technologies asymmetrically affect carbon dioxide emissions in the..., Xin [/bib_ref]. The number of patents can indicate the level of invention and innovation and the degree of innovation activity of a region or enterprise and reflect the real technological innovation achievements of a specific region in that year. The data are accurate, authoritative, and easy to obtain. In this study, the number of patent applications accepted (take the logarithm) was used to measure the level of technological innovation.
## Control variables
Previous studies have established that a country or region's carbon emission intensity is affected not only by industrial restructuring and technological innovation but also by population size, the urbanization process, economic development level, the intensity of environmental regulation, and other factors [bib_ref] Exploring the linkage among energy intensity, carbon emission and urbanization in Pakistan:..., Shah [/bib_ref]. In this study, population size (POP), urbanization rate (URB), economic development (RGDP), and environmental regulation (ER) were used as control variables that affect the intensity of regional carbon dioxide emissions.
The population affects urban infrastructure construction and industrial development through the agglomeration effect and scale effect, which in turn affect the carbon emission intensity effect. The number of permanent residents represents the population at the end of the year.
Urbanization leads to the improvement of urban infrastructure and environmental management. On the other hand, it also leads to numerous urban problems. Urbanization is expressed by the proportion of the permanent urban population to the total population.
According to the environmental Kuznets hypothesis, the level of regional economic development is closely related to environmental pollution. The environmental quality will show an inverted U-shaped trend of deterioration followed by improvement as the level of economic development increases. RGDP is represented by the real GDP per capita.
The implementation of environmental regulations can force enterprises to adopt cleaner production methods and more advanced energy-saving and emission-reducing technologies for production to reduce the cost of environmental management and the demand for high-carbon energy. Environmental regulation is expressed by the proportion of industrial pollution control investment in GDP.
## Data sources and descriptive statistics
In view of the differences in statistical calibers and methods of data from Hong Kong, Macao, and Taiwan, and the severe lack of data from Tibet, and considering the completeness and availability of data, this study finally selected 30 provinces (municipalities and autonomous regions) in China as the study area. The data were derived mainly from "China Statistical Yearbook", "China Industrial Statistical Yearbook", "China Environmental Statistical Yearbook", "China Science and Technology Statistical Yearbook", "China Energy Statistical Yearbook", the website of the National Bureau of Statistics, and the statistical yearbooks of these provinces (municipalities and autonomous regions) between 2002 and 2020. For the missing data of individual years, the interpolation method was used to fill in the data, and the data on price variables were deflated with 2001 as the base period. According to research needs, the 30 provinces (cities and autonomous regions) were divided into the eastern region (Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, Hainan), central region (Jilin, Heilongjiang, Shanxi, Anhui, Jiangxi, Henan, Hubei, Hunan), and western region (Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang). [fig_ref] Table 2: The descriptive statistics of variables [/fig_ref] shows the descriptive statistics of each variable.
## Spatiotemporal evolution characteristics of carbon emission intensity
From the perspective of temporal evolution, the carbon emission intensity of the whole country and the three major regions showed a continuous downward trend with certain stages from 2001 to 2019 [fig_ref] Figure 2: Variation trend of carbon emission intensity in China and the three major... [/fig_ref]. The national carbon emission intensity dropped from 4.50 t/CNY 10,000 to 2.68 t/CNY 10,000, with an average annual decrease of 2.84%. To be specific, it dropped from 3.32 t/CNY 10,000 to 1.76 t/CNY 10,000, with an average annual decrease of 3.48% in the eastern region; it dropped from 5.20 t/CNY 10,000 to 2.35 t/CNY 10,000, with an average annual decrease of 4.32% in the central region; it dropped from 5.18 t/CNY 10,000 to 3.86 t/CNY 10,000, with an average annual decrease of 1.63% in the western region. In terms of stages, 2001-2006 was a relatively stable stage, and the decline in carbon emission intensity in the whole country and the three major regions was the same with little change; 2006-2011 was a slow decline stage, with an average annual decline of 3.78% in the national carbon emission intensity, specifically, 3.54% in the eastern region, 5.78% in the central region and 2.83% in the western region; 2011-2019 was a stage of rapid decline, with an average annual decline of 4.66% in the national carbon emission intensity, specifically, 5.34% in the eastern region, 5.37% in the central region and 3.99% in the western region. The overall carbon emission intensity in the eastern region was relatively low and always lower than the national average level. The carbon emission per unit in the central region was relatively high before 2008. After 2008, the average annual decrease in carbon emission per unit was faster and lower than the national average level. The carbon emission intensity in the western region was relatively high and always higher than the national average level. The overall carbon emission intensity of the three major regions presented a pattern of "western region > central region > eastern region", with regional nonequilibrium characteristics, but the regional differences in carbon emission intensity were gradually narrowing, and the convergence trend was becoming more and more obvious. The overall carbon emission intensity in the eastern region was relatively low and always lower than the national average level. The carbon emission per unit in the central region was relatively high before 2008. After 2008, the average annual decrease in carbon emission per unit was faster and lower than the national average level. The carbon emission intensity in the western region was relatively high and always higher than the national average level. The overall carbon emission intensity of the three major regions presented a pattern of "western region > central region > eastern region", with regional non-equilibrium characteristics, but the regional differences in carbon emission intensity were gradually narrowing, and the convergence trend was becoming more and more obvious. On the whole, China's carbon emission intensity continued to decrease as a whole from 2001 to 2019, indicating that China's green economic transformation, energy conservation, and emission reduction policies were productive, laying a solid foundation for high-quality economic development in the new era. In particular, since the new development concepts of "innovation, coordination, green, openness, and sharing" were put forward at the 18th National Congress of the Communist Party of China, governments at all levels have vigorously promoted industrial restructuring, eliminated outdated process equipment, promoted energy-saving technology products, and built clean, low-carbon, safe and efficient energy systems, thereby improving the efficiency of energy utilization and reducing the carbon emission intensity per unit of GDP continuously.
To explore the spatial evolution characteristics of China's carbon emission intensity, a K-means cluster analysis was carried out using SPSS26.0 based on the level of carbon emission intensity in each province from 2001 to 2019. Then, ArcGIS10.8 was employed to visualize the carbon emission intensity of the 30 provinces. The carbon emission intensity was divided into four types according to the clustering results: low-level area, relative low-level area, relative high-level area, and high-level area [fig_ref] Figure 3: The spatial distribution pattern of carbon emission intensity in China [/fig_ref]. From 2001 to 2019, China's carbon emission intensity showed a spatial pattern of high in the west and low in the east, high in the north and low in the south. The provinces with high carbon emission intensity were mainly concentrated in the northern and northwestern regions, while the carbon emission intensity in the eastern region was relatively low. From the perspective of different years, the regional differences in carbon emission intensity in 2001 were obvious, and the spatial agglomeration characteristics were significant, mainly expressed in lowlevel and relative high-level areas. There were 13 provinces in the low-level area, mainly located in the coastal areas and the Sichuan-Chongqing area. There were ten provinces with relatively high carbon emission intensity, mainly in the central and northeastern regions. Shanxi Province was a low-level area, while Guizhou, Ningxia, Liaoning, Gansu, Inner Mongolia, and Xinjiang were relatively high-level. The change in the overall spatial pattern in 2007 was not obvious compared with 2001, and low-level areas and relative high-level areas still dominated it. Shandong, Yunnan, and Hainan changed from low-level areas of carbon emission intensity to high-level areas, while Shanghai changed from a relatively high-level area to a high-level one. The spatial differentiation pattern of carbon emission intensity changed significantly in 2013 compared with 2007. The low-level areas increased significantly from 11 to 20, whereas the high-level areas decreased remarkably from 12 to 6 and gradually shifted and shrank to the western inland areas. In 2019, the low-level carbon areas further increased to 23 from 20 in 2013. The number of low-level areas did not change. The high-level areas mainly included Liaoning, Inner Mongolia, Shaanxi, Gansu, and other provinces. Shanxi and Xinjiang had relatively high carbon emission intensity and were considered high-level areas.
level areas increased significantly from 11 to 20, whereas the high-level areas decreased remarkably from 12 to 6 and gradually shifted and shrank to the western inland areas. In 2019, the low-level carbon areas further increased to 23 from 20 in 2013. The number of low-level areas did not change. The high-level areas mainly included Liaoning, Inner Mongolia, Shaanxi, Gansu, and other provinces. Shanxi and Xinjiang had relatively high carbon emission intensity and were considered high-level areas.
# Empirical results
# Benchmark regression results
To represent the correlation between industrial restructuring and carbon emission intensity more vividly, a scatter-fitting diagram of industrial structure rationalization, industrial structure optimization, and carbon emission intensity was drawn [fig_ref] Figure 3: The spatial distribution pattern of carbon emission intensity in China [/fig_ref]. According to the trend of the fitting line, there was a negative correlation between industrial structure rationalization, industrial structure optimization, and carbon emission intensity,
# Empirical results
# Benchmark regression results
To represent the correlation between industrial restructuring and carbon emission intensity more vividly, a scatter-fitting diagram of industrial structure rationalization, industrial structure optimization, and carbon emission intensity was drawn [fig_ref] Figure 4: Linear fitting of industrial structure rationalization, industrial structure optimization, and carbon emission... [/fig_ref]. According to the trend of the fitting line, there was a negative correlation between industrial structure rationalization, industrial structure optimization, and carbon emission intensity, which could support Hypothesis 1 to a certain extent. However, the real impact of the two on carbon emission intensity still needs to be further tested using a model. which could support Hypothesis 1 to a certain extent. However, the real impact of the two on carbon emission intensity still needs to be further tested using a model. The present study used a panel model to verify the impact mechanism of industrial restructuring on carbon emission intensity. LM and Hausman tests revealed that the fixedeffect model was better than the mixed effect and random effect models, and the fixedeffect model was the optimal explanatory model. To ensure the stability of the panel regression results, the estimated results of random effects and OLS were also listed [fig_ref] Table 3: The overall regression results [/fig_ref].
Judging from the estimation results at the national level, industrial structure rationalization had a significant negative impact on carbon emission intensity. With other conditions remaining unchanged, every 1% increase in the level of industrial structure rationalization reduced carbon emission intensity by 0.547%. This showed that industrial structure rationalization promoted the benign flow of production factors among various industrial sectors, improved the effective allocation and rational utilization of resources, enhanced the coupling quality of the input structure and output structure of inter-industry factors, and significantly elevated the efficiency of resource utilization and the level of comprehensive utilization, thereby effectively reducing carbon emission intensity. Industrial structure optimization had a significant negative impact on carbon emission intensity. When related variables were controlled, every 1% increase in the level of industrial structure optimization reduced the carbon emission intensity by 1.117%. This indicated that the optimization of the industrial structure led to the gradual elimination or transfer of industries with high pollution and high energy consumption and the accelerated development of high value-added and low energy consumption industries such as producer services and modern manufacturing, thus improving the energy structure, reducing fossil energy consumption, and lowering carbon emission intensity. Therefore, industrial structure rationalization and industrial structure optimization could dramatically reduce the carbon emission intensity, but from the panel estimation results, industrial structure optimization had better carbon reduction effects than industrial structure rationalization. For this reason, in the process of regional industrial restructuring, it is necessary not only to improve the utilization efficiency and coupling degree of inter-industry factor resources and promote the coordinated and balanced development of various industries and within industries but also to accelerate the development of modern service industries, high-tech industries, and strategic emerging industries and improve the advanced level of the industrial structure.
From the perspective of control variables, the driving effect of population size on carbon emission intensity was very significant, passing the 1% significance test, indicating that with the expansion of population size, the energy demand for production and living increased, thus increasing the unit carbon emissions. The impact of urbanization rate on The present study used a panel model to verify the impact mechanism of industrial restructuring on carbon emission intensity. LM and Hausman tests revealed that the fixedeffect model was better than the mixed effect and random effect models, and the fixed-effect model was the optimal explanatory model. To ensure the stability of the panel regression results, the estimated results of random effects and OLS were also listed [fig_ref] Table 3: The overall regression results [/fig_ref]. Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, here and below.
Judging from the estimation results at the national level, industrial structure rationalization had a significant negative impact on carbon emission intensity. With other conditions remaining unchanged, every 1% increase in the level of industrial structure rationalization reduced carbon emission intensity by 0.547%. This showed that industrial structure rationalization promoted the benign flow of production factors among various industrial sectors, improved the effective allocation and rational utilization of resources, enhanced the coupling quality of the input structure and output structure of inter-industry factors, and significantly elevated the efficiency of resource utilization and the level of comprehensive utilization, thereby effectively reducing carbon emission intensity. Industrial structure optimization had a significant negative impact on carbon emission intensity. When related variables were controlled, every 1% increase in the level of industrial structure optimization reduced the carbon emission intensity by 1.117%. This indicated that the optimization of the industrial structure led to the gradual elimination or transfer of industries with high pollution and high energy consumption and the accelerated development of high value-added and low energy consumption industries such as producer services and modern manufacturing, thus improving the energy structure, reducing fossil energy consumption, and lowering carbon emission intensity. Therefore, industrial structure rationalization and industrial structure optimization could dramatically reduce the carbon emission intensity, but from the panel estimation results, industrial structure optimization had better carbon reduction effects than industrial structure rationalization. For this reason, in the process of regional industrial restructuring, it is necessary not only to improve the utilization efficiency and coupling degree of inter-industry factor resources and promote the coordinated and balanced development of various industries and within industries but also to accelerate the development of modern service industries, high-tech industries, and strategic emerging industries and improve the advanced level of the industrial structure.
From the perspective of control variables, the driving effect of population size on carbon emission intensity was very significant, passing the 1% significance test, indicating that with the expansion of population size, the energy demand for production and living increased, thus increasing the unit carbon emissions. The impact of urbanization rate on carbon emission intensity was significantly negative, showing that with the increase in urbanization rate, the agglomeration effect and scale effect of urbanization gradually appeared, which showed a significant inhibitory effect on carbon emission intensity. Economic development had an obvious negative inhibitory effect on carbon emission intensity, indicating that the improvement in economic development level drew more funds to be invested in the research and development of energy conservation, emission reduction, and cleaner production, effectively reducing carbon emission intensity. However, the effect of environmental regulation on carbon emission intensity did not pass the significance test, exhibiting that current environmental regulation was weak and the effect on carbon emission had not yet appeared.
## The mediating effect of technological innovation
The mediation effect test was further used to analyze the transmission mechanism of the impact of industrial restructuring on carbon emission intensity. The panel fixed effect model was employed to explore the mediating effect of technological innovation and verify whether it acts as a mediating variable in industrial restructuring affecting carbon emission intensity. The first step was the regression result of Equation (1), see [fig_ref] Table 3: The overall regression results [/fig_ref] , and the second and third steps were the regression of Equations (2) and (3), see [fig_ref] Table 4: Regression results from the mediating effect of technological innovation [/fig_ref]. Columns 1 and 2 are the regression results of industrial structure rationalization and industrial structure optimization on technological innovation, and columns 3 and 4 are the regression results of industrial structure rationalization and technological innovation, industrial structure optimization, and technological innovation on carbon emission intensity. Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, here and below.
Judging from the regression results of columns 1 and 2 in [fig_ref] Table 4: Regression results from the mediating effect of technological innovation [/fig_ref] , the impact of industrial structure rationalization and industrial structure optimization on technological innovation passed the 1% significance test. The influence coefficients were 0.244 and 0.387, respectively. In other words, every 1% increase in industrial structure rationalization and industrial structure optimization increased technological innovation by 0.244% and 0.387%, respectively, indicating that the core explanatory variables had a significant impact on the explained variables; that is, both industrial structure rationalization and industrial structure optimization could significantly promote regional technological innovation. Judging from the regression results in columns 3 and 4, the regression coefficients of technological innovation passed the 1% and 5% significance tests, and the influence coefficients were -0.527 and -0.273, respectively, showing that technological innovation had a significant inhibitory effect on carbon emission intensity. The regression coefficients of industrial structure rationalization and industrial structure optimization were -0.429 and -1.011, and both passed the 1% significance test, indicating that technological innovation was mediating in the process of industrial restructuring affecting carbon emission intensity. Industrial restructuring could improve the energy structure and reduce energy consumption by promoting technological innovation, thereby significantly inhibiting the unit carbon emission intensity. As such, Hypothesis 2 proves valid.
## Regional differences
Since there are significant differences in industrial resource endowments, industrial bases, and economic development levels in different regions, the present study further explored the impact of industrial restructuring on carbon emission intensity in the three major regions. Judging from the estimation results at the regional level [fig_ref] Table 5: Regression results by region [/fig_ref] , at the 1% significance level, the estimation results of both industrial structure rationalization and industrial structure optimization passed the significance test. The influence coefficients of industrial structure rationalization on the carbon emission intensity of the three regions were negative, at −0.169, −3.566, and −2.400, respectively, indicating that for every 1% increase in industrial structure rationalization, the carbon emission intensity of the three regions decreased by 0.169%, 3.566%, and 2.400%, respectively. The influence coefficients of industrial structure optimization on the carbon emission intensity of the three major regions were negative, which were −1.939, −2.197, and −1.815, respectively, showing that for every 1% increase in industrial structure optimization, the carbon emission intensity in the three regions dropped by 1.939%, 2.197%, and 1.815%, respectively. Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, here and below.
It could also be seen from the table that industrial structure rationalization and industrial structure optimization had a greater impact on carbon emission intensity in the central and western regions than in the eastern region. On the one hand, industrial restructuring had a marginal decreasing effect on carbon emission intensity, which weakened with the continuous optimization of the industrial structure. Compared with the eastern region, the industrial level in the central and western regions was relatively low; the industrial structure was relatively simple; the secondary industry had dominated for a long time; the secondary industry had more high-carbon sectors and was thus more likely to be positively affected by the energy-saving and emission-reduction effects brought about by changes in the industrial structure. Therefore, the impact of industrial restructuring on the carbon emission intensity of the central and western regions was more significant. On the other hand, most provinces in the eastern region had low carbon emission intensity, with little room for further decline, and the carbon reduction effect of industrial restructuring had been released in advance. In contrast, the economic foundation of the central and western regions was relatively weak; industry was still in the early stage of transformation; economic development was still highly dependent on traditional energy and high-energy-consuming industries; and there was still much room for the reduction of carbon emission intensity.
## Robustness test
The econometric model constructed herein may have endogeneity problems caused by variable omission, measurement error, and bidirectional causality, which might affect the robustness of the regression results. Therefore, to ensure the reliability of the regression results, the one-period lag of each explanatory variable was used as an instrumental variable, and the two-stage least squares method was utilized to test the robustness of the overall regression results and the mediating effect regression results [fig_ref] Table 6: The robustness test results [/fig_ref]. From the regression results of robustness, it could be seen that the direction and significance of the regression coefficients of the benchmark regression test (model 1-model 2) and the mediating effect test (model 3-model 4) were not significantly different from the previous ones, indicating that the empirical conclusions drawn above were robust and reliable. Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, here and below.
# Discussions
From the direct effect, the relationship between industrial restructuring and carbon emission intensity is negative, which implies that industrial restructuring plays a positive role in the improvement of carbon emission reduction. This conclusion is supported by the results of Dong al. and Zhang [bib_ref] How to achieve a win-win situation between economic growth and carbon emission..., Dong [/bib_ref] [bib_ref] Changing industrial structure to reduce carbon dioxide emissions: A Chinese application, Chang [/bib_ref]. From the general rule of industrial structure evolution, the shift of industrial structure center of gravity from primary to tertiary industries is often accompanied by the gradual replacement of traditional industries by low-pollution, low-energy-consuming and high-value-added new industries, which not only improves factor production efficiency but also reduces energy consumption intensity, thus positively influencing the reduction of carbon emissions [bib_ref] Research on the spatial effect and threshold effect of industrial structure upgrading..., Li [/bib_ref] [bib_ref] Have carbon emissions been reduced due to the upgrading of industrial structure?..., Zhu [/bib_ref]. Therefore, it is recommended that Chinese provinces continue to accelerate the pace of industrial transformation and upgrading, promote the industrial structure in the direction of advanced development, encourage the transformation and upgrading of traditional industries, improve industrial efficiency, and strive to achieve low energy consumption and high output while accelerating the formation of knowledge-and technology-intensive industries such as strategic emerging industries and high-tech industries.
In terms of the mediating effect, the test results of this study confirm that industrial restructuring has an indirect effect on carbon emission intensity through technological innovation. Industrial restructuring can re-match innovation resources, change the factor supply and demand structure, and force enterprises to improve science and technology innovation and the transformation ability of science and technology achievements by increasing R&D investment [bib_ref] Carbon Dioxide Emissions Reduction through Technological Innovation: Empirical Evidence from Chinese Provinces, Liu [/bib_ref]. Technological innovation will promote substituting other factors such as new energy and capital for traditional energy through the transformation of results, which will reduce energy consumption [bib_ref] Industrial structure, technical progress and carbon intensity in China's provinces, Cheng [/bib_ref]. Therefore, we suggest paying attention to the interaction between technological innovation and industrial restructuring to reduce carbon emission intensity jointly. At the same time, since there are obvious differences in the regional development of the eastern, central and western regions, it is necessary to grasp their development positioning accurately, combine the actual economic and social development of the region, and formulate a targeted carbon emission intensity control strategy in line with the region. The eastern region should fully play the core role of technological innovation in industrial structure upgrading while pursuing advanced industrial structure. The central and western regions should accelerate the enhancement of enterprise innovation capacity, focus on introducing production capital and technologyintensive industries, and use technology to drive the transformation and upgrading of industrial structure.
Of course, there are certain limitations to this paper. First, low-carbon pilot industrial restructuring may also affect carbon emissions through other indirect channels, and this paper only chose technological innovation in its examination. Industrial restructuring may also affect carbon emissions through channels such as government preferences and fiscal decentralization. Second, although this paper applied econometric models as much as possible and considered a series of factors that may influence the estimation results, it still could not fully solve the endogeneity problem of the model. Furthermore, due to the difficulty of data collection, Tibet, Hong Kong, Macau, and Taiwan were not included in this paper. Future research on the effect of industrial restructuring and technological innovation on carbon emissions needs to be followed up continuously. Finally, further empirical investigation on regional and administrative hierarchy heterogeneity is still needed due to the limitation of space.
# Conclusions
This study analyzed the spatiotemporal evolution characteristics of carbon emission intensity in 30 Chinese provinces using ArcGIS and the energy consumption data from 2001 to 2019, concerning the IPCC carbon emission accounting method. Industrial restructuring, technological innovation, and carbon emission intensity were incorporated into a unified research framework. The panel fixed effect and mediating effect models were used to empirically test the direct and indirect driving mechanisms of industrial restructuring and technological innovation on carbon emission intensity. The main conclusions are as follows. First, from 2001 to 2019, China's global carbon emission intensity showed a continuous decline. Second, the results of the benchmark regression showed that the rationalization of industrial structure and the advanced industrial structure have significant adverse inhibitory effects in influencing carbon emission intensity, and there is regional heterogeneity. Third, The mediating effect analysis revealed that technological innovation was an important channel for industrial restructuring to affect carbon emission intensity. Industrial restructuring reduced carbon emission intensity through direct effects and the indirect effect of technological innovation.
[fig] Figure 1: The research logic model. [/fig]
[fig] Figure 2: Variation trend of carbon emission intensity in China and the three major regions (2001-2019). [/fig]
[fig] Figure 3: The spatial distribution pattern of carbon emission intensity in China (2001-2019). [/fig]
[fig] Figure 4: Linear fitting of industrial structure rationalization, industrial structure optimization, and carbon emission intensity. [/fig]
[table] Table 1: The carbon emission coefficients of various fossil fuels.Note: The data are from the "2006 IPCC Guidelines for National Greenhouse Gas Inventories" and "China Energy Statistical Yearbook". [/table]
[table] Table 2: The descriptive statistics of variables. [/table]
[table] Table 3: The overall regression results. [/table]
[table] Table 4: Regression results from the mediating effect of technological innovation. [/table]
[table] Table 5: Regression results by region. [/table]
[table] Table 6: The robustness test results. [/table]
|
Cardiovascular safety and efficacy of metformin-SGLT2i versus metformin-sulfonylureas in type 2 diabetes: systematic review and meta-analysis of randomized controlled trials
Diabetes is a serious threat to global health and among the top 10 causes of death, with nearly half a billion people living with it worldwide. Treating patients with diabetes tend to become more challenging due to the progressive nature of the disease. The role and benefits of combination therapies for the management of type 2 diabetes are well-documented, while the comparative safety and efficacy among the different combination options have not been elucidated. We aimed to systematically synthesize the evidence on the comparative cardiovascular safety and efficacy of combination therapy with metformin-sodium-glucose cotransporter-2 inhibitors versus metforminsulfonylureas in patients with type 2 diabetes. We searched MEDLINE-PubMed, Embase, Cochrane Library, and ClinicalTrials.gov up to 15 August 2019 without restriction in the year of publication. We included randomized controlled trials of patients with type 2 diabetes who were on metforminsodium-glucose cotransporter-2 inhibitors or metformin-sulphonylureas combination therapy at least for a year. The primary endpoints were all-cause mortality and serious adverse events, and the secondary endpoints were cardiovascular mortality, non-fatal myocardial infarction, non-fatal stroke, hypoglycemia, and changes in glycated hemoglobin A1c (HbA1c), body weight, fasting plasma glucose, blood pressure, high-density lipoprotein cholesterol, and low-density lipoprotein cholesterol. We used a random-effects meta-analysis model to estimate mean differences for continuous outcomes and risk ratio for dichotomous outcomes. We followed PICOS description model for defining eligibility and the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) 2015 guidelines for reporting results. Of 3,190 citations, we included nine trials involving 10,974 participants. The pooled analysis showed no significant difference in all-cause mortality (risk ration [RR] = 0.93, 95% CI [0.52, 1.67]), serious adverse events (RR = 0.96, 95% CI [0.79, 1.17]) and adverse events (RR = 1.00, 95% CI [0.99, 1.02]) between the two, but in hypoglycemia (RR = 0.13, 95% CI [0.10, 0.17], P < 0.001). Participants taking metformin-sodium glucose cotransporter-2 inhibitors showed a significantly greater reduction in HbA1c (mean difference [MD] = − 0.10%, 95% CI [− 0.17, − 0.03], body weight (MD = − 4.57 kg, 95% CI [− 4.74, − 4.39], systolic blood pressure (MD = − 4.77 mmHg, 95% CI [− 5.39, − 4.16]), diastolic blood pressure (MD = − 2.07 mmHg, 95% CI [− 2.74, − 1.40], and fasting plasma glucose (MD = − 0.55 mmol/L, 95% CI [− 0.69, − 0.41]), p < 0.001. Combination therapy of metformin and sodium-glucose cotransporter-2 inhibitors is a safe and efficacious alternative to combination therapy of metformin and sulphonylureas for patients with type 2 diabetes who are at risk of cardiovascular comorbidity. However, there remains a need for additional long-term randomized controlled trials as available studies are very limited and heterogeneous. OPEN
## Scientific reports
| (2021) 11:137 | https://doi.org/10.1038/s41598-020-80603-8 www.nature.com/scientificreports/ formin-SGLT2Is versus metformin-sulfonylureas combination therapies for patients with T2DM. We included RCTs without restriction on year of publication, but published in English language, up to 15 August 2019. The RCTs were needed to have at least a 1-year follow-up of patients. The keywords we used in different combinations using Boolean Operators were: metformin, biguanide, sodium-glucose co-transporter-2 inhibitors, SGLT-2 inhibitors, dapagliflozin, canagliflozin, empagliflozin, ertugliflozin, sulfonylurea, gliclazide, glimepiride, glyburide, glibenclamide, glipizide, tolbutamide, type 2 diabetes mellitus, T2DM, cardiovascular outcomes, and randomized controlled trials (Supplementary file 1). All potentially eligible studies were considered for this review, irrespective of the primary outcomes. Manual searching was performed to find out additional eligible trials from the reference lists of key articles.
Eligibility. We formulated the study's eligibility criteria using the PICOS (participants, interventions, comparison, outcomes, and study designs) description model [bib_ref] GRADE guidelines: 2. Framing the questions and deciding on important outcomes, Guyatt [/bib_ref] :
- Participants
- Patients with T2DM
- Man or woman of any age - Who was taking a combination therapy of metformin-sulfonylurea or metformin-SGLT2Is
- Intervention - A combination of metformin with any of the SGLT2Is, which could be dapagliflozin, canagliflozin, empagliflozin, or ertugliflozin.
## - comparator
- A combination of metformin with any of sulfonylureas compounds, which could be gliclazide, glipizide, glyburide, glibenclamide, or glimepiride. - Outcomes
- Primary endpoints -All-cause mortality -Serious adverse events (SAEs). Study selection. Two independent authors examined the title and abstract of all searched studies. From the title and abstract of all studies identified by the database search, those studies duplicated and not meet the inclusion criteria were excluded. The full texts of the remaining studies were further reviewed. Disagreements were resolved by consensus and if persisted, we arbitrated through discussion with a third author. Data extraction. Two independent authors extracted the needed data from each RCTs. These include name of the first author, year of publication, trial registration, mean age of the participant, baseline average body weight, HbA1c, interventions, comparators, number of participants randomized, duration of follow-up, and patient-important outcomes. Data on the mean change of HbA1c (%), body weight (Kg), FPG (mmol/L), SBP (mmHg), DBP (mmHg), HDL-C (mmol/L), and LDL-C (mmol/L) were collected from baseline for continuous outcomes. The status and number of events were captured for the two groups, which include, death, hypoglycemia, adverse events (AEs), SAEs, SAEs related to study drugs, genital mycotic infection (GMI), and adverse cardiovascular events. www.nature.com/scientificreports/ Assessment of risk of bias. We used the Cochrane risk of bias toolto assess the risk of bias in each included study and the risks were judged by two independent authors as "Low", "Unclear", or "High" based on the critical domains, including random sequence generation, allocation concealment, blinding, incomplete outcome data, and selective reporting. Disagreements were resolved by consensus and if persisted, we arbitrated through discussion with a third author.
# Statistical analysis.
We carried out meta-analysis using the computer software packages RevMan 5.3 46 to compare the cardiovascular safety and efficacy between the two combination therapies. Pooled results of continuous patient-important outcomes i.e., HbA1c, FPG, blood pressure, body weight, HDL-C, and LDL-C were reported using a mean difference (MD) with 95% confidence interval (CI). Pooled results of binary outcomes i.e. all-cause mortality, AEs, AEs related to study drug, SAEs, SAEs related to study drug, hypoglycemic event, worsening of coronary artery disease, acute myocardial infarction (AMI), aortic aneurism, coronary artery occlusion (CAO) and GMI were summarized using risk ratio (RR) with 95% CI. We used Mantel-Haenszel method [bib_ref] Mantel-Haenszel estimators of a common odds ratio for multiple response data, Suesse [/bib_ref] to pool effect estimates of dichotomous outcomes and inverse variance for continuous outcomes. The analysis was conducted using a random-effects meta-analyses model as it assumes the observed estimates of the treatment can vary across studies because of the real differences in the treatment effect in each study as well as sampling variability (chance) [bib_ref] Interpretation of random effects meta-analyses, Riley [/bib_ref]. We used Cochrane Q test [bib_ref] Measuring inconsistency in meta-analyses, Higgins [/bib_ref] to assess heterogeneity between studies, and I 2 testing 50 to quantify heterogeneity between studies, with values > 50% representing moderate-to-high heterogeneity. We carried out sensitivity and subgroup analysis by duration of the RCTs. We couldn't conduct funnel plot and Egger test to check any possible reporting bias because the number of studies included in the meta-analyses are insufficient (less than 10 trials). We considered statistical analysis with a p-value < 0.05 statistically significant.
# Results
Search results. We searched a total of 3,190 citations through the databases, of which we assessed 30 fulltext studies for eligibility and found nine of them 51-59 fulfilled the inclusion criteria. We excluded the rest 21 full-text articles 23,24,60-78 mainly for they did not include SGLT2Is or sulfonylureas as a combination therapy; included a combination of more than two glucose-lowering drugs; included single glucose-lowering drug; data were driven from review or post hoc analysis of previous RCTs; had no active comparator; or had a duration of interventions less than a year [fig_ref] Figure 1: PRISMA flow diagram of the study [/fig_ref]. summarizes the characteristics of the nine RCTs included. Four RCTs [bib_ref] Efficacy and safety of canagliflozin versus glimepiride in patients with type 2..., Cefalu [/bib_ref] [bib_ref] Results of VERTIS SU extension study: safety and efficacy of ertugliflozin treatment..., Hollander [/bib_ref] [bib_ref] Canagliflozin provides durable glycemic improvements and body weight reduction over 104 weeks..., Leiter [/bib_ref] [bib_ref] Ertugliflozin compared with glimepiride in patients with type 2 diabetes mellitus inadequately..., Hollander [/bib_ref] used two different doses of SGLT2Is. The meta-analysis included all results of both doses for dichotomous outcomes but only a high dose of SGLT2Is for continuous outcomes. who were in either of the two combination therapies at least for a year .
## Study characteristics.
## Methodological quality and risk of bias.
The studies were found to be "low risk of bias" when these studies were subjected to the Cochrane Collaboration's Tool for Quality Assessment of Randomized Controlled Trials [fig_ref] Figure 2: 'Risk of bias' summary [/fig_ref].
Efficacy and safety assessments. All-cause mortality. All the nine included RCTs involving 10,974
participants assessed all-cause mortality/death events between intervention and control groups. Our meta-analysis of pooled results revealed no significant difference in all-cause mortality/death events between patients with T2DM who were on metformin-SGIT2Is and metformin-sulphonylureas combination therapies (RR = 0.93, 95% CI [0.52, 1.67], p = 0.81), with statistically non-significant heterogeneity between studies (I 2 = 9%) [fig_ref] Figure 3: Comparison of all-cause mortality/death events between patients with T2DM who were on... [/fig_ref]. Subgroup analysis by duration of follow-up showed no significant difference in risk of death between the two groups, year 1 (RR = 0. [bib_ref] Effects on the incidence of cardiovascular events of the addition of pioglitazone..., Vaccaro [/bib_ref] Two trials 51,56 assessed the change in HDL-C and LDL-C levels from baseline between the two arms. Both HDL-C and LDL-C levels were reduced in patients who received metformin-sulphonylureas combination therapies. However, the pooled effect was not statistically significant between the two arms in HDL-C (MD = 4.32, 95% CI [− 4.00, 12.64] mmol/L, p = 0.32), I 2 = 99%) and LDL-C (MD = 3.63, 95% CI [− 3.96, 11.22] mmol/L, P = 0.35, I 2 = 88%) (S14 & S15) respectively.
Adverse events. With pooled data from the nine trials, we found no statistically significant difference in the risk of developing adverse events between the two arms (RR = 1.00, 95% CI [0.99, 1.02], p = 0.66, I 2 = 0% . We performed a sensitivity analysis by removing the highest weighted study [bib_ref] Empagliflozin compared with glimepiride in metformin-treated patients with type 2 diabetes: 208-Week..., Ridderstråle [/bib_ref] Five trials assessed the risk of serious adverse events related to study drug. The pooled result of these trials showed that serious adverse events related to study drug were less frequent on patients taking metforminsulphonylureas combination, but the pooled result was not significant at 95% CI (RR = 1.34 [0.75, 2.36] (S7).
Hypoglycemic events. We analyzed hypoglycemic events on pooled results of the nine trials involving 10,794 T2DM patients, considering the occurrence of at least one hypoglycemic event during the follow-up period. Patients under metformin-SGIT2I combination therapy were found to experience significantly fewer hypoglycemic events as compared to patients under metformin-sulphonylureas combination therapy (RR = 0.13, 95% CI [0.10, 0.17], P < 0.001) I 2 = 67%, p = 0.002) [fig_ref] Figure 8: Risk of hypoglycemia between patients who were on metformin-SGIT2I and metforminsulphonylureas combination... [/fig_ref]. We performed sensitivity analysis by removing two low weighted trials [bib_ref] Long-term glycaemic response and tolerability of dapagliflozin versus a sulphonylurea as add-on..., Del Prato [/bib_ref] [bib_ref] Ertugliflozin compared with glimepiride in patients with type 2 diabetes mellitus inadequately..., Hollander [/bib_ref]. However, the result of the remaining trials was similar to the nine trials in risk of hypoglycemia (RR = 0.13, 95% CI [0.10, 0.18, P < 0.001) with increased between study heterogeneity (I 2 = 73%, P = 0.009). To see the robustness of the result, we did subgroup analysis at different duration of follow-up. However, the risk of experiencing hypoglycemic events was consistently more frequent under patients on metformin-sulfonylureas combination therapy at 95% CI RR = 0. [bib_ref] Health system performance for people with diabetes in 28 low-and middle-income countries:..., Manne-Goehler [/bib_ref] Bodyweight. All trials included in the meta-analysis assessed the change in body weight from baseline between the two groups. The pooled result showed that body weight of patients in the metformin-SGLT2I was significantly reduced from baseline compared to patients in the metformin-sulfonylureas (MD = − 4.57, 95% CI [fig_ref] Figure 1: PRISMA flow diagram of the study [/fig_ref]. We conducted a sensitivity analysis by removing the low weighted [bib_ref] Canagliflozin provides durable glycemic improvements and body weight reduction over 104 weeks..., Leiter [/bib_ref]
# Discussion
The result of this systematic review and meta-analysis showed that the risk of all-cause mortality/death events was not statistically significant between patients in the metformin-SGLT2I and metformin-sulfonylureas combination therapies. In agreement with our findings, a previous study reported a non-significant difference in all-cause mortality and cardiovascular death between the two arms. Similarly, study 77 also revealed a lower and statistically nonsignificant cardiovascular death and all-cause mortality in T2DM patients who were on metformin-sulfonylurea combination therapy. However, another study [bib_ref] Sulfonylureas as second line drugs in type 2 diabetes and the risk..., Douros [/bib_ref] reported that the use of sulfonylureas as a second-line drug has significantly associated with an increased risk of myocardial infarction and all-cause mortality. www.nature.com/scientificreports/ The goal of durable glycemic control is to reduce the long-term risk of diabetes-related cardiovascular morbidity and mortality [bib_ref] Durability of glycaemic efficacy over 2 years with dapagliflozin versus glipizide as..., Nauck [/bib_ref]. Recent studies showed that SGLT2Is have been shown to decrease cardiovascular events in treating high-risk patients with T2DM 34 . Likewise, the addition of empagliflozin to metformin therapy improves in patients with established CVD or heart failure [bib_ref] The safety of empagliflozin plus metformin for the treatment of type 2..., Scheen [/bib_ref]. Even though the current meta-analysis of metformin-SGLT2I combination showed a favorable effect on cardiovascular outcomes (acute myocardial infarction, aortic aneurism, coronary artery occlusion, and atherosclerosis), the pooled analysis was not statistically significant.
Lowering blood pressure is significantly important to reduce the risk of CVD and death in many patients with T2DM. A 10-mmHg reduction in systolic blood pressure decreased the risk of major CVD events by 20% [bib_ref] Blood pressure lowering for prevention of cardiovascular disease and death: A systematic..., Ettehad [/bib_ref]. SGLT2Is induced a greater reduction in systolic and diastolic blood pressure [bib_ref] Sodium-glucose co-transporter 2 inhibitors compared with sulfonylureas in patients with type 2..., Chen [/bib_ref]. Similar with the other findings, the pooled result of our study showed a reduction in both systolic and diastolic blood pressure with patients in the metformin-SGLT2Is combination therapy as compared with metformin-sulfonylureas, which might be due to the effect of SGLT2Is reducing renal glucose reabsorption and increasing urinary excretion [bib_ref] Efficacy and safety of canagliflozin versus glimepiride in patients with type 2..., Cefalu [/bib_ref]. Consistent with www.nature.com/scientificreports/ our findings, previous studies reported a greater increase from baseline in HDL-C and LDL-C in patients who were on SGLT2Is groups as compared with sulfonylureas [bib_ref] Efficacy and safety of canagliflozin versus glimepiride in patients with type 2..., Cefalu [/bib_ref] [bib_ref] Results of VERTIS SU extension study: safety and efficacy of ertugliflozin treatment..., Hollander [/bib_ref] [bib_ref] Canagliflozin provides durable glycemic improvements and body weight reduction over 104 weeks..., Leiter [/bib_ref].
The result of this meta-analysis showed that the overall incidence of AEs was similar across the two arms, which is consistent with a trial reported by elsewhere [bib_ref] Ertugliflozin compared with glimepiride in patients with type 2 diabetes mellitus inadequately..., Hollander [/bib_ref]. However, metformin-SGLT2Is combination was associated with a higher episode of GMI which is similar to other meta-analysis reported an increased risk of GMI with SGLT2Is 73 . In addition, for both genders, the proportion of patients reporting symptoms of GMI was higher in the SGLT2Is group than in the sulfonylureas [bib_ref] Long-term glycaemic response and tolerability of dapagliflozin versus a sulphonylurea as add-on..., Del Prato [/bib_ref]. However, these infections respond to standard antimicrobial treatment and their incidence declines with time [bib_ref] Durability of glycaemic efficacy over 2 years with dapagliflozin versus glipizide as..., Nauck [/bib_ref]. SAEs such as ketoacidosis, bone fractures, and pyelonephritis [fig_ref] Figure 1: PRISMA flow diagram of the study [/fig_ref]. Comparison of change in body weight (Kg) from baseline between patients who were on metformin-SGIT2I and metformin-sulphonylureas combination therapies. SD standard deviation; CI confidence interval; MD = mean difference; Met metformin; SGLT2I sodium-glucose co-transporter 2 inhibitor; SU sulfonylurea. [fig_ref] Figure 1: PRISMA flow diagram of the study [/fig_ref]. Comparison of change in FPG (mmol/L) from baseline between patients who were on metformin-SGIT2I and metformin-sulphonylureas combination therapies. SD standard deviation; CI confidence interval; MD mean difference; Met metformin; SGLT2I sodium-glucose co-transporter 2 inhibitor; SU sulfonylurea. www.nature.com/scientificreports/ were rarely reported with metformin-SGLT2Is 53 , while the pooled result of the current study did not find a statistically significant difference between metformin-SGLT2I and metformin-sulfonylureas combinations. Even though the pooled analysis did not find statistically significant results, AEs related to study drug was observed more in patients with metformin-SGLT2Is, but more SAEs related to study drug in metformin-sulfonylureas. This might be due to the hypoglycemic effect of sulfonylureas and GMI as a result of SGLT2Is respectively. Sulfonylureas as second-line drugs are associated with an increased risk of severe hypoglycemia [bib_ref] Blood pressure lowering for prevention of cardiovascular disease and death: A systematic..., Ettehad [/bib_ref] , and our study is in support of this. The hypoglycemic effect of sulfonylureas might be due to its insulin-dependent mechanism of action, while the less hypoglycemic effect of SGLT2Is is due to the non-insulin dependent mechanism of action.
Obesity is one of the main risk factors for T2DM and representing a major worldwide health problem. Lowering body weight is an important part of T2DM management. SGLT2Is has been associated with an added benefit of weight loss in patients with T2DM, whereas sulfonylureas are reported to increase body weight [bib_ref] Body weight Considerations in the management of type 2 diabetes, Apovian [/bib_ref]. In support of this evidence, our study showed a 4.57 kg weight loss in metformin-SGLT2Is group than the metformin-sulfonylureas. The weight loss caused by SGLT2Is is probably due to the loss of calories via urine and glucose-induced osmotic diuresis [bib_ref] Efficacy and safety of canagliflozin versus glimepiride in patients with type 2..., Cefalu [/bib_ref].
Long term glycemic control is the major goal of diabetes management to prevent both microvascular and macrovascular complications of DM. Both metformin-SGLT2Is and metformin-sulfonylureas combinations are effective to control HbA1c for a short duration of follow-up. However, for a long duration of follow-up, metformin-SGLT2Is are more effective than metformin-sulfonylureas 73 . Our finding is consistent with this evidence where both groups were equally effective for a one-year duration of follow-up; however, as the duration of follow-up increases to 4 years, metformin-SGLT2Is combination showed a significant reduction from baseline in HbA1c. Eight previously conducted trials reported a higher reduction of FPG in patients randomized to metformin-SGLT2Is 51-54,56-59 , and our pooled result is in support of their finding as it showed reduction of FPG under patients on metformin-SGLT2Is.
This study reported important information about cardiovascular safety, efficacy, and cardiovascular risk factors control between the two combination therapies. However, we acknowledge that available studies are very limited and heterogeneous. There remains a need for additional long-term trials comparing the overall safety, efficacy, and cost-effectiveness of metformin-SGLT2Is and metformin-sulfonylureas combination therapies.
# Conclusion
Combination therapy of metformin and sodium-glucose cotransporter-2 inhibitors are a safe and efficacious alternative to combination therapy of metformin and sulfonylureas for patients with T2DM who are at risk of cardiovascular comorbidity. However, there remains a need for additional long-term randomized controlled trials as available studies are very limited and heterogeneous. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.
[fig] Figure 1: PRISMA flow diagram of the study. [/fig]
[fig] Glycated hemoglobin A1c: All the nine RCTs involving 9,180 participants assessed the changes in HbA1c (%) from baseline between the two arms. The pooled data of these trials showed significant difference in the mean difference of HbA1c between patients on metformin-SGLT2I and metformin-sulfonylureas combination therapies; MD = − 0.10, 95% CI [− 0.17, − 0.03] %, p = 0.005), I 2 = 63%, p = 0.006)(Fig. 9).Interestingly, subgroup analysis by duration of follow-up showed a reduction of HbA1c from baseline was not statistically significant between the two groups at year-one, with MD = − 0.01[− 0.13, 0.11] %, I 2 = 67%. However, metformin-SGLT2I induced a greater reduction in HbA1c from baseline after 2 years (MD = − 0.12[− 0.20, − 0.05] %, p = 0.001) and after 4 years (MD = − 0.24[− 0.37, − 0.10] %, p = 0.0007) (S9). [/fig]
[fig] Figure 2: 'Risk of bias' summary: review authors' judgments about each 'risk of bias' item for included trials. Green cells = 'low risk'; blank cells = 'unclear risk'; red cells = 'high risk' . [/fig]
[fig] Figure 3: Comparison of all-cause mortality/death events between patients with T2DM who were on metformin-SGIT2I and metformin-sulphonylureas combination therapies. RR risk ratio; CI confidence interval; df degrees of freedom; Met metformin; SGLT2I sodium-glucose co-transporter 2 inhibitor; SU sulfonylurea. [/fig]
[fig] Figure 4: Comparison of changes in systolic blood pressure between patients who were on metformin-SGIT2I and metformin-sulphonylureas combination therapies. CI confidence interval; df degree of freedom; MD mean difference; Met metformin; SGLT2I sodium-glucose co-transporter 2 inhibitor; SU sulfonylurea; SD standard deviation. [/fig]
[fig] Figure 5, Figure 6: Comparison of changes in diastolic blood pressure between patients who were on metformin-SGIT2I and metformin-sulphonylureas combination therapies. CI confidence interval; df degree of freedom; MD mean difference; Met metformin; SGLT2I sodium-glucose co-transporter 2 inhibitor; SU sulfonylurea; SD standard deviation. Risk of adverse events between patients who were on metformin-SGIT2I and metforminsulphonylureas combination therapies. RR risk ratio; CI confidence interval; df degrees of freedom; Met metformin; SGLT2I sodium-glucose co-transporter 2 inhibitor; SU sulfonylurea. [/fig]
[fig] Figure 7: Risk of serious adverse events between patients who were on metformin-SGIT2I and metforminsulphonylureas combination therapies. RR risk ratio; CI confidence interval; df degrees of freedom; Met metformin; SGLT2I sodium-glucose co-transporter 2 inhibitor; SU sulfonylurea.Scientific Reports | (2021) 11:137 | https://doi.org/10.1038/s41598-020-80603-8 [/fig]
[fig] Figure 8: Risk of hypoglycemia between patients who were on metformin-SGIT2I and metforminsulphonylureas combination therapies. RR risk ratio; CI confidence interval; df degrees of freedom; Met metformin; SGLT2I sodium-glucose co-transporter 2 inhibitor; SU sulfonylurea. [/fig]
[fig] Figure 9: Comparison of change in HbA1c (%) from baseline between patients who were on metformin-SGIT2I and metformin-sulphonylureas combination therapies. SD standard deviation; CI confidence interval; MD mean difference; Met metformin; SGLT2I sodium glucose co-transporter 2 inhibitor; SU sulfonylurea.Scientific Reports | (2021) 11:137 | https://doi.org/10.1038/s41598-020-80603-8 [/fig]
|
Human GTPBP5 is involved in the late stage of mitoribosome large subunit assembly
Human mitoribosomes are macromolecular complexes essential for translation of 11 mitochondrial mRNAs. The large and the small mitoribosomal subunits undergo a multistep maturation process that requires the involvement of several factors. Among these factors, GTP-binding proteins (GTPBPs) play an important role as GTP hydrolysis can provide energy throughout the assembly stages. In bacteria, many GTPBPs are needed for the maturation of ribosome subunits and, of particular interest for this study, ObgE has been shown to assist in the 50S subunit assembly. Here, we characterize the role of a related human Obg-family member, GTPBP5. We show that GTPBP5 interacts specifically with the large mitoribosomal subunit (mt-LSU) proteins and several late-stage mitoribosome assembly factors, including MTERF4:NSUN4 complex, MRM2 methyltransferase, MALSU1 and MTG1. Interestingly, we find that interaction of GTPBP5 with the mt-LSU is compromised in the presence of a non-hydrolysable analogue of GTP, implying a different mechanism of action of this protein in contrast to that of other Obg-family GTPBPs. GTPBP5 ablation leads to severe impairment in the oxidative phosphorylation system, concurrent with a decrease in mitochondrial translation and reduced monosome formation. Overall, our data indicate an important role of GTPBP5 in mitochondrial function and suggest its involvement in the late-stage of mt-LSU maturation.
# Introduction
Ribosomes are conserved macromolecular structures that catalyse protein synthesis in all organisms. Mitochondria retain their own translation machinery, consisting of the mitoribosomes, which produces 13 essential polypeptide components of the oxidative phosphorylation (OxPhos) complexes. Human mitoribosomes sediment as 55S particles and comprise the 39S large subunit (mt-LSU), formed by 16S mt-rRNA, mt-tRNA Val and 52 mitoribosomal proteins (MRPs), and the 28S small subunit (mt-SSU), formed by 12S mt-rRNA and 30 MRPs (1-3). Mitoribosomes exhibit unique features due to the acquisition of both mitochondrial-specific proteins and mitochondrial N-and C-terminal extensions of the universal ribosomal proteins, and drastic reduction in mt-rRNAs size. Together, these differences result in a higher protein:rRNA ratio when compared to bacterial ribosomes, reflecting a functional and structural divergence that remains to be fully understood. Furthermore, the dual genetic origin of the mitoribosome (rRNAs and tRNA are encoded by the mitochondrial DNA (mtDNA) whereas all 82 MRPs are encoded in the nucleus) suggests complex mechanisms are needed for coordinated production and assembly of mitoribosomal components [bib_ref] Kinetics and mechanism of mammalian mitochondrial ribosome assembly, Bogenhagen [/bib_ref].
Mutations in both MRP genes and factors which contribute to mitoribosome biogenesis have been reported to cause severe mitochondrial disorders and therefore, understanding the molecular mechanisms underlying these processes is of crucial importance [bib_ref] Mitochondrial transcript maturation and its disorders, Van Haute [/bib_ref]. Recent advances in cryo-electron microscopy (cryo-EM) have allowed for a characterization of the complexity of the mitoribosome structure [bib_ref] The structure of the human mitochondrial ribosome, Amunts [/bib_ref] , including association with a few accessory factors involved in mitoribosome maturation, such as MALSU1, L0R8F8 and mt-ACP (mitochondrial acylcarrier protein) [bib_ref] Structures of the human mitochondrial ribosome in native states of assembly, Brown [/bib_ref]. However, a complete understanding of which factors are involved, and how such accessory proteins co-ordinate assembly, remains to be determined.
GTP-binding proteins (GTPBPs) are important regulatory proteins present in all domains of life, with functions implicated in a variety of cellular processes, including receptor signaling, intracellular signal transduction, cytoskeleton organization and protein synthesis. These proteins are a family of NTPases defined as molecular switches, named as such as they are generally active when bound with GTP and inactive when bound with GDP. GTP binding and hydrolysis take place via the G-domain, consisting of five conserved motifs (G1-G5), which is present in all GTPases and conserved across species [bib_ref] Structure of small G proteins and their regulators, Paduch [/bib_ref] [bib_ref] The universally conserved prokaryotic, Verstraeten [/bib_ref].
Over the past decades, GTPBPs have been extensively studied in bacteria and represent one of the major groups of ribosomal assembly factors [bib_ref] Role of GTPases in ribosome assembly, Karbstein [/bib_ref] [bib_ref] Role of GTPases in bacterial ribosome assembly, Britton [/bib_ref]. Additionally, the repertoire of GTPBPs known to function in humans in mitoribosome assembly has been increasing in recent years. NOA1 (C4orf14) was shown to localize to mitochondria, where it was found to bind to the mt-SSU in a GTP-dependent manner, with the depletion of NOA1 leading to a mitochondrial translation defect [bib_ref] NOA1 is an essential GTPase required for mitochondrial protein synthesis, Kolanczyk [/bib_ref] [bib_ref] Human C4orf14 interacts with the mitochondrial nucleoid and is involved in the..., He [/bib_ref]. Era-like 1 (ERAL1) also associates with the mt-SSU, binding directly to 12S mt-rRNA [bib_ref] Human ERAL1 is a mitochondrial RNA chaperone involved in the assembly of..., Dennerlein [/bib_ref] [bib_ref] ERAL1 is associated with mitochondrial ribosome and elimination of ERAL1 leads to..., Uchiumi [/bib_ref]. Another GTPase, MTG1 (GTPBP7), couples mt-LSU maturation with the formation of an intersubunit bridge, acting as a quality-control checkpoint in mitoribosome assembly [bib_ref] Human G-proteins, ObgH1 and Mtg1, associate with the large mitochondrial ribosome subunit..., Kotani [/bib_ref] [bib_ref] MTG1 couples mitoribosome large subunit assembly with intersubunit bridge formation, Kim [/bib_ref]. In addition, GTPBP10 has been shown to act in the mt-LSU assembly process [bib_ref] MitoRibo-Tag mice provide a tool for in vivo studies of mitoribosome composition, Busch [/bib_ref] [bib_ref] The human Obg protein GTPBP10 is involved in mitoribosomal biogenesis, Lavdovskaia [/bib_ref] [bib_ref] Human GTPBP10 is required for mitoribosome maturation, Maiti [/bib_ref].
Bacterial GTPBPs belonging to the Obg-family have been reported to have a role in assembly of the 50S subunit in many prokaryotic species. The Obg-family belongs to the TRAFAC (translation-factors-related) class and can be divided in subfamilies of high molecular mass GTPBPs (Obg, EngD, DRG, NOG). Obg proteins possess a highly conserved glycine-rich N-terminal domain in addition to the highly conserved GTP-binding domain and a C-terminal domain [bib_ref] Classification and evolution of P-loop GTPases and related ATPases, Leipe [/bib_ref] [bib_ref] Structure of small G proteins and their regulators, Paduch [/bib_ref]. In Escherichia coli, the recently resolved structure of the 50S-ObgE-GMPPNP complex has revealed that ObgE is an important factor in the final stages of the LSU assembly, acting as a quality-control checkpoint [bib_ref] Structural and functional insights into the mode of action of a universally..., Feng [/bib_ref]. Furthermore, in S. cerevisiae, Obg-family protein Mtg2p localizes to mitochondria, and has been reported to participate in the assembly of the 54S subunit [bib_ref] The yeast GTPase Mtg2p is required for mitochondrial translation and partially suppresses..., Datta [/bib_ref].
In humans, two Obg-family proteins are present, GTPBP10 and GTPBP5 (Mtg2, ObgH1), both of which have previously been shown to localize to mitochondria. Ablation of GTPBP10 has been shown to lead to a severe mitochondrial translation defect and the factor is likely to be involved in the late stages of mt-LSU maturation [bib_ref] MitoRibo-Tag mice provide a tool for in vivo studies of mitoribosome composition, Busch [/bib_ref] [bib_ref] The human Obg protein GTPBP10 is involved in mitoribosomal biogenesis, Lavdovskaia [/bib_ref] [bib_ref] Human GTPBP10 is required for mitoribosome maturation, Maiti [/bib_ref]. In a similar manner to the interaction of ObgE with the bacterial large subunit [bib_ref] Structural and functional insights into the mode of action of a universally..., Feng [/bib_ref] , interaction of GTPBP10 with the mt-LSU was found to be stabilized by the incubation with non-hydrolysable analogue GMPPNP [bib_ref] MitoRibo-Tag mice provide a tool for in vivo studies of mitoribosome composition, Busch [/bib_ref].
In contrast to GTPBP10, characterization of GTPBP5 function has been limited. In addition to revealing mito-chondrial localization of GTPBP5, work by Kotani et al. [bib_ref] Human G-proteins, ObgH1 and Mtg1, associate with the large mitochondrial ribosome subunit..., Kotani [/bib_ref] showed that GTPBP5 has the capacity to co-sediment with the mt-LSU in vitro, through sucrose gradient centrifugation analysis. However, knock-down of GTPBP5 using siRNA had no detectable detrimental effect on the OxPhos system or mitochondrial translation.
Our study herein, builds upon the initial characterization performed by show that GTPBP5 interacts with the mt-LSU in cultured HEK293T cells. Our interactome analysis reports a specific interaction of GTPBP5 with several late-stage mt-LSU assembly factors. Surprisingly, we find that GTP hydrolysis is required for binding of GTPBP5 to the mt-LSU, evidenced by the loss of GTPBP5mt-LSU interaction in the presence of a non-hydrolysable analog of GTP, suggesting a different mechanism of action from GTPBP10. The loss of GTPBP5 causes a severe defect in OxPhos function, associated with a reduction in the synthesis rate of mtDNA-encoded proteins and decreased monosome formation.
While we prepared the revision of this work, similar findings were reported by Maiti et al. [bib_ref] Human GTPBP5 (MTG2) fuels mitoribosome large subunit maturation by facilitating 16S rRNA..., Maiti [/bib_ref]. Our study, together with Maiti et al., extend on the previously limited knowledge of GTPBP5 function, to implicate it in the assembly of the human mitoribosome.
# Materials and methods
## Generation and maintenance of flp-in t-rex stable mammalian cell lines
The Flp-In T-Rex human embryonic kidney 293 (HEK293T) cell line (Invitrogen) was used to generate stable mammalian cell lines that allow for doxycyclineinducible expression of C-terminal FLAG-tagged GTPBP5 (GTPBP5::FLAG) in a dose-dependent manner. GTPBP5 cDNA was obtained from the hORFeome Database (Internal ID:12579) and subsequently cloned into pcDNA5/FRT/TO. Primers used for cloning of GTPBP5 and GTPBP5 mutant are listed in . HEK293T cells were cultured in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% (v/v) tetracycline-free fetal bovine serum (FBS), 2 mM Glutamax (Gibco), 1× Penicillin/Streptomycin (Gibco), 50 g/ml uridine, 10 g/ml Zeocin (Invitrogen) and 100 g/ml blasticidin (Gibco) at 37 - C under 5% CO 2 atmosphere. Prior to transfection, cells were seeded in a six-well plate and grown in culture medium lacking selective antibiotics. Cells were then transfected with pcDNA5/FRT/TO-GTPBP5::FLAG, and pOG44 constructs using Lipofectamine 3000 according to manufacturer's instructions. At 48 h post-transfection cells were selected by addition of hygromycin (100 g/ml, Invitrogen) and blasticidin (100 g/ml) to the culture medium. After two to three weeks, selected colonies were picked and cultured as single clonal populations. To confirm inducible expression of GTPBP5, cells were incubated with 100 ng/ml doxycycline, and cell pellets were harvested for western blot analysis 24 or 48 h later.
For cell growth measurements, cells were seeded at a density of 2 × 10 4 in six-well plates and grown in the presence or absence of 50 ng/ml of doxycycline in glucose-free DMEM supplemented with 0.9 g/l galactose, 10% (v/v) FBS, 1 mM sodium pyruvate, 2 mM Glutamax and 1× Penicillin/Streptomycin. Cell number was measured at 48 hourly intervals with EVE™ Automated Cell Counter (Na-noEnTek).
For SILAC-based quantitative proteomic analysis, cells were grown in Iscove's modified Dulbecco's medium (IMDM) supplemented with 'heavy' ( 15 N-and 13 C-labelled) or 'light' Arg, Lys and 10% dialysed FCS (Thermo Scientific HyClone).
## Generation of gtpbp5 knock-out cell line
GTPBP5 is encoded by the MTG2 gene. Generation of the HEK293T knock-out cell line for MTG2 was performed using CRISPR/Cas9 system as described in . Two pairs of gRNAs targeting exon 1 of MTG2 were designed and cloned into the pSpCas9(BB)-2A-Puro (pX459) V2.0 vector to generate out-of-frame deletions. Cells were transfected with the pX459 variants using Lipofectamine 3000 according to manufacturer's recommendations. Transfected cell populations were selected by puromycin treatment at a final concentration of 1.5 g/ml for 48 h. Subsequent to this, cells were diluted to achieve single-cell derived clones on 96-well plates. Resultant clones were screened by Sanger sequencing to assess knockout and loss of GTPBP5 in selected clones was confirmed by western blotting.
# Immunocytochemistry analysis
Human 143B osteosarcoma (HOS) cell line, HEK293T, HEK293T overexpressing GTPBP5::FLAG and GTPBP5 KO cells with re-expression of GTPBP5 (GTPBP5 RESCUE ) were used for immunocytochemistry (ICC) analysis. Cells were cultured in DMEM supplemented with 10% (v/v) FBS, 2 mM Glutamax and 1× Penicillin/Streptomycin at 37 - C under 5% CO 2 atmosphere. 143B cells were transiently transfected with pcDNA5/FRT/TO-GTPBP5::FLAG using Lipofectamine 3000 according to manufacturer's instructions, while HEK293T cells, HEK293T cells overexpressing GTPBP5::FLAG and GTPBP5 RESCUE cells were induced with 50 ng/ml of doxycycline 48 h prior the analysis.
The ICC experiment was performed on fixed cells as previously described [bib_ref] MRM2 and MRM3 are involved in biogenesis of the large subunit of..., Rorbach [/bib_ref]. For 143B cells, TOM20 was used to detect the mitochondrial network while for the remaining cell lines MitoTracker Red CMXRos was supplemented to the medium for 15 min prior to fixation. Images were acquired using Zeiss LSM800 confocal microscope.
## Immunodetection of proteins
For immunoblot analysis, either 12 or 20 g aliquots of total cell lysate protein were used to detect steady state levels of OxPhos complexes and mitoribosomal proteins, respectively. To detect GTPBP5, 20 g of lysed isolated mitochondria was used. The lysates were subjected to SDS-PAGE and wet transferred onto PVDF membranes (Millipore) at 4 - C for 1 h and 10 min at 300 mA. Blocking solution of 5% nonfat milk (Semper) in PBS-T (1× phosphate buffered saline supplemented with 1% Triton X-100) was applied to the membranes for 1 h at RT. The membranes were later incubated with specific primary antibodies in 5% milk in PBS-T overnight. Subsequently, membranes were washed three times with PBS-T and incubated with the HRP-conjugated secondary antibodies (GE Healthcare) in 5% milk in PBS-T for 1 h at RT and, after three washes with PBS-T, visualized using ECL (Bio-Rad). lists the antibodies used in this study.
## Immunoprecipitation of proteins
Isolation of mitochondria was performed from HEK293T cells expressing GTPBP5::FLAG as described in Rorbach et al. [bib_ref] The human mitochondrial ribosome recycling factor is essential for cell viability, Rorbach [/bib_ref]. Pelleted mitochondria were suspended in lysis buffer (Sigma-Aldrich, St. Louis, MO, USA) freshly supplemented with 1× PIC (cOmplete, Mini, EDTA-Free, Protease Inhibitor Cocktail, Roche) and 5 mM MgCl 2 . For specific analysis, 5 -guanylyl-imidodiphosphate trisodium salt hydrate (GMPPNP) at a final concentration of 20 mM and RNase A at a final concentration of 20 g/ml were also added. Immunoprecipitation was performed using the FLAG-immunoprecipitation (IP) kit (Sigma-Aldrich) following manufacturer's instructions. Wash and elution buffers were freshly supplemented with 5 mM MgCl 2 . Elution was performed using 3X FLAG peptide according to manufacturer's instructions.
## [ 35 s]-metabolic labelling of mitochondrial proteins
To label newly-synthesized mtDNA-encoded proteins, cells were seeded into a six-well dish at 80-90% confluency. Cell lines overexpressing GTPBP5 or GTPBP5 S238A were induced with 50 or 0.5 ng/ml of doxycycline prior to the experiment. Firstly, two washing steps of 5 min each in methionine/cysteine-free DMEM were performed. Subsequently, cells were incubated with fresh methionine/cysteine-free DMEM supplemented with Glutamax 100× (Gibco), Sodium Pyruvate 100× (Gibco), 10% dialysed FBS and 100 g/ml emetine (Sigma-Aldrich) for 20 min at 37 - C. Labeling was performed with addition of 166,7 Ci/ml of EasyTag EXPRESS [ 35 S] protein labelling mix (methionine and cysteine) (Perkin Elmer) for 30 min at 37 - C. Following labelling, cells were washed with 1 ml of PBS three times and the final pellets were collected by centrifugation and stored at -20 - C. Cells were lysed in 1× PBS-PIC with addition of 50 units of benzonase (Life Technologies) with incubation on ice for 20 min, followed by addition of SDS to 1% f.c. and further incubation on ice for 30 min. The suspended pellets were stored at -20 - C for 1 h to ease lysis. After cell lysis, 30 g total protein, assayed by DC assay (Bio-Rad), were separated on Bolt 12% Bis-Tris Plus (Invitrogen) SDS-PAGE gels. Gels were then incubated in Imperial Protein Stain (Thermo Fisher) for 1 h at RT and with fixing solution (20% methanol, 7% acetic acid, 3% glycerol) for 1 h at RT. Finally, gels were vacuum-dried at 65 - C for 2 h. The resultant gel was exposed to storage Phosphor screens and visualized with Typhoon FLA 7000 Phosphorimager.
## Rna isolation and northern blotting
Total RNA was isolated from GTPBP5 KO cell line, HEK293T cell line and sucrose gradient fractions using TRIzol (Invitrogen) according to manufacturer's instructions. RNA concentration was determined using Nan-oDrop ND-1000 UV-Vis Spectrophotometer (Thermo Scientific). For northern blot analysis, RNA was solubilized in NorthernMax™-Gly sample loading dye and 4 g of RNA were resolved into 1.2 % (w/v) agarose gels containing 0.7% formaldehyde and 1× NorthernMax MOPS buffer (Life Technologies). Subsequently, gels were transferred onto Hybond-N+ membrane overnight. Membranes were then UV crosslinked and hybridized with DNA probes. [ 32 P]-labelled DNA probes were prepared using the Prime-It II Random Primer Labelling Kit (Agilent Technologies) using 50 Ci of dCTP (PerkinElmer) according to manufacturer's instructions. Primers used to prepare DNA templates for [ 32 P]-labelled DNA probes are listed in Supplementary . Following probing, membranes were washed with 1× SSC buffer (150 mM NaCl, 15 mM trisodium citrate (pH 7.0)). Probed membranes were exposed to a storage Phosphor screen and visualized using Typhoon FLA 7000 Phosphorimager.
## Sucrose gradient centrifugation analysis
Isolation of mitochondria was performed as described in Richter et al. [bib_ref] Translation termination in human mitochondrial ribosomes, Richter [/bib_ref]. 1 mg of mitochondria were lysed in lysis buffer (10 mM Tris-HCl pH 7.5, 100 mM KCl or 50 mM KCl, 20 mM MgCl 2 , 1× PIC, 260 mM sucrose, 1% Triton X-100) freshly supplemented with 0.4 U/l final concentration of RNase Block Ribonuclease Inhibitor (Agilent), then loaded onto a linear sucrose gradient (10-30% (w/v), 11 ml total volume) in 1× gradient buffer (20 mM Tris-HCl pH 7.5, 100 mM KCl or 50 mM KCl, 20 mM MgCl 2 , 1× PIC) and then centrifuged for 15 h at 79 000 × g at 4 - C (Beckman Coulter SW41-Ti rotor). A total of 25 fractions with a volume of 450 l each were collected via pipetting from the top of the gradient, and 15 l of each fractions 3-17 inclusive were used for western blot analysis. For fractions 1 and 2, and for fractions 18 and 19, 7.5 l of each fraction were combined and resolved together.
For SILAC-based proteomics, HEK293T or GTPBP5 KO cells and GTPBP5 RESCUE or GTPBP5 KO cells were grown in 'heavy' or 'light' labelled media for more than seven doublings as described in Van Haute et al. [bib_ref] METTL15 introduces N4-methylcytidine into human mitochondrial 12S rRNA and is required for..., Van Haute [/bib_ref]. Cell lines were pooled and mitochondrial isolation and sucrose gradient analysis were performed as described above. All buffers used for gradient analysis contained 50 mM KCl, instead of 100 mM KCl, to better preserve interaction with auxiliary factors.
## Analysis of mitochondrial respiratory complexes activity using bn-page
BN-PAGE and in-gel activity measurements were performed on isolated mitochondria for Complex I, Complex IV and Complex V. Mitochondria were pelleted, resuspended in ACNA (1.5 M aminocaproic acid, 50 mM Bis-Tris pH 7.00) and quantified using Qubit. Following quantification, mitochondria were lysed using 4% (w/v) digi-tonin for 20 min on ice and subsequently centrifuged at maximum speed for 30 min at 4 - C. 50 g of proteins were then separated by Blue-Native Polyacrylamide Gel Electrophoresis (BN-PAGE) for about 5 h. The activity of mitochondrial respiratory complexes was measured in-gel by addition of buffers containing OxPhos substrates specific for Complex I (2 mM Tris-HCl, pH 7.4, 0.1 mg/ml NADH, 2.5 mg/ml iodonitrotetrazolium chloride), Complex IV (0.5 mg/ml DAB, 50 mM phosphate buffer pH 7.4, 1 mg/ml cytochrome c, 0.2 M sucrose, 20 g/ml (1 nM) catalase) and Complex V (50 mM glycine, 5 mM MgCl 2 , 1% Triton X-100, 1.5 mM lead nitrate, 2 mM ATP).
## High-resolution mitochondrial respiration analysis
Mitochondrial respiration was measured by sequential addition of complex I and complex II substrates, uncoupler and inhibitors using an Oxygraph-2K oxygen electrode (Oroboros, Innsbruck, Austria) in MiR05 respiration buffer (110 mM D-sucrose, 20 mM taurine, lactobionic acid 60 mM, 10 mM KH 2 PO 4 , 3 mM MgCl 2 , 0.5 mM EGTA, 1 mg/ml fatty acid free bovine serum albumin (BSA), 20 mM HEPES pH 7.1).
Basal rate was first measured in intact cells (Basal) followed by permeabilization of cells using digitonin (10 g/ml) and addition of complex I substrates glutamate (10 mM), malate (2 mM), pyruvate (2.5 mM) and ADP (2.5 mM) to measure complex I-linked respiration (CI-ADP; Complex I respiration--respiratory rate measured in the presence of complex I substrates (glutamate, malate, pyruvate) and ADP). Succinate (10 mM; complex II substrate) was then added to measure convergent complex I + II respiratory rate (CI+II-ADP; Complex I+II respiration--respiratory rate measured in the presence of complex I (glutamate, malate, pyruvate), complex II substrate (succinate) and ADP). Uncoupled, maximal respiration was subsequently measured in the presence of the uncoupler carbonyl cyanide m-chlorophenyl hydrazone (CCCP, 1.0 M; Max U/C; Maximal respiratory rate--respiratory rate measured in the presence of complex I (glutamate, malate, pyruvate), complex II substrate (succinate), ADP and the uncoupler CCCP) and nonmitochondrial residual oxygen consumption rate (ROX) was measured after addition of rotenone (0.5 M) and antimycin A (2.5 M). Oxygen consumption rates were calculated using DatLab software and expressed as O 2 flow (pmol/s/10 6 cells). Differences between wild-type, KO and rescue cell lines were determined by two-tailed Student's t tests.
# Mass spectrometry analysis
Preparation of peptides from GTPBP5-IP for mass spectrometry. Proteins obtained from coimmunoprecipitations were prepared according to a modified protocol from Busch et al. [bib_ref] MitoRibo-Tag mice provide a tool for in vivo studies of mitoribosome composition, Busch [/bib_ref] described below. Briefly, proteins obtained from GTPBP5::FLAG and HEK293T co-immunoprecipitations (n = 3) were resuspended in 200-300 l of 0.6 M Guanidium chloride buffer (5 mM Tris-HCl pH 7.5, 0.1 mM Tris(2carboxyethyl)phosphine (TCEP; ThermoFisher), 0.5 mM 2-chloroacetamid (Merck, cat. no. 8024120100)). Peptides were generated by overnight digest with 300 ng trypsin gold (Promega, cat. no. V5280, resuspended in 50 mM acetic acid) added directly to the resuspended protein samples. On the next day, peptides were desalted using home-made StageTips (Empore Octadecyl C18, 3M; [bib_ref] Stop And Go Extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS..., Rappsilber [/bib_ref] and eluted with 80-100 l of 60% acetonitrile/0.1% formic acid buffer. The eluted peptides were subsequently dried using a vacuum concentrator plus (Eppendorf) and resuspended with 0.1% formic acid for mass spectrometry.
Preparation of peptides from SILAC sucrose gradient experiment for mass spectrometry. Peptides from SILAC sucrose gradient centrifugation experiments were prepared from fractions 1 and 2 joined, 3 and 4 joined, and 5-17 individually. Collected fractions were precipitated in 20 × 100% ice-cold ethanol overnight at -20 - C. Pelleted proteins were resuspended in 6 M GuHCl/Tris pH 8.0 solution and sonicated for 5 min at maximum output (10 s on/off cycles). After a 5 min incubation at RT, samples underwent a second round of sonication and were later centrifuged at maximum speed for 10 min. DTT at a final concentration of 5 mM was added to the obtained supernatants and incubated for 30 min at 55 - C followed by incubation with 15 mM chloroacetamide for 15 min at RT in the dark. Prior to digestion, protein quantification was performed and trypsin (Pierce, trypsin protease MS-grade, Thermo Fisher Scientific) was added accordingly. Protein digestion was performed at 37 - C overnight with mild shaking. After 12-14 h, trypsin was inactivated using 1.2% formic acid and samples were spun down at 3000 × g for 10 min at RT. Samples were desalted, using pre-packed C18 desalting columns (Thermo Fisher Scientific) previously equilibrated and washed respectively with 100% acetonitrile (ACN) and 0.5% formic acid, and eluted (0.5% formic acid, 50% ACN). Peptides were consequently dried using a SpeedVac Vacuum Concentrator and resuspended in 0.5% formic acid for mass spectrometry.
# Lc-ms/ms analysis.
Peptides were separated on a 25 cm, 75 m internal diameter PicoFrit analytical column (New Objective) packed with 1.9 m ReproSil-Pur 120 C18-AQ media (Dr Maisch,) using an EASY-nLC 1200 (Thermo Fisher Scientific). The column was maintained at 50 - C. Buffer A and B were 0.1% formic acid in water and 0.1% formic acid in 80% acetonitrile. Peptides were separated on a segmented gradient from 6% to 31% buffer B for 45 min and from 31% to 50% buffer B for 5 min at 200 nl/min. Eluting peptides were analysed on QExactive HF mass spectrometer (Thermo Fisher Scientific). Peptide precursor m/z measurements were carried out at 60 000 resolution in the 300-1800 m/z range. The ten most intense precursors with charge state from 2 to 7 only were selected for HCD fragmentation using 25% normalized collision energy. The m/z values of the peptide fragments were measured at a resolution of 30 000 using a minimum AGC target of 2e5 and 80 ms maximum injection time. Upon fragmentation, precursors were put on a dynamic exclusion list for 45 s.
For the analysis of the sucrose gradient fractions, peptides were separated using a segmented gradient from 6% to 31% buffer B and from 31% to 44% for 5 min. Eluting peptides were analysed on an Orbitrap Fusion Tribrid mass spectrometer (Thermo Fisher Scientific). Peptide precursor m/z measurements were carried out at 60 000 resolution in the 350-1500 m/z range. The 10 most intense precursors with charge state from 2 to 7 only were selected for HCD fragmentation using 27% normalized collision energy. The m/z values of the peptide fragments were measured at a resolution of 50 000 using a minimum AGC target of 2e5 and 86 ms maximum injection time. Upon fragmentation, precursors were put on a dynamic exclusion list for 45 s.
Protein identification and quantification. The raw data were analysed with MaxQuant version 1.5.3.8 (32) using the integrated Andromeda search engine [bib_ref] Andromeda: a peptide search engine integrated into the MaxQuant environment, Cox [/bib_ref]. Peptide fragmentation spectra were searched against the canonical sequences of the human reference proteome (proteome ID UP000005640, downloaded September 2018 from UniProt). Methionine oxidation and protein N-terminal acetylation were set as variable modifications; cysteine carbamidomethylation was set as fixed modification. The digestion parameters were set to 'specific' and 'Trypsin/P,' The minimum number of peptides and razor peptides for protein identification was 1; the minimum number of unique peptides was 0. Protein identification was performed at a peptide spectrum matches and protein false discovery rate of 0.01. The 'second peptide' option was on. Successful identifications were transferred between the different raw files using the 'Match between runs' option. Labelfree quantification (LFQ) (34) was performed using an LFQ minimum ratio count of two. LFQ intensities were filtered for at least two valid values in at least one group and imputed from a normal distribution with a width of 0.3 and down shift of 1.8. Protein quantification was performed using the t-test function in Perseus version 1.5.2.4 [bib_ref] The MaxQuant computational platform for mass spectrometry-based shotgun proteomics, Tyanova [/bib_ref].
For SILAC labelling analyses, the raw data from the sucrose gradient fractions was analysed with MaxQuant version 1.6.1.0. Heavy labels were set to Arg10 and Lys8 and Label min. ratio count to 1. Match between runs was enabled between the same fraction number of the two sucrose gradients. For analyses, R version 3.5.2 and ggplot2 3.2.1 were used. Intensity values or normalized H/L ratios were extracted from the MaxQuant proteinGroups.txt file. Within each replicate, intensities of control (HEK293T or GTPBP5 RESCUE ) and GTPBP5 KO were normalized to each other against the median of the control, and then log2 transformed for plotting. Assembly factors associated with either the mt-SSU or the mt-LSU were manually curated.
qPCR For qPCR analysis, total RNA was extracted from HEK293T WT and GTPBP5 KO cells using TRIzol reagent according to manufacturer's instructions and RNA concentration was determined using Qubit RNA Broad Range assay kit (Thermo Fisher). Aliquots of 2.5 g total RNA were treated with TURBO DNase (Thermo Fisher) to digest contaminating DNA, and 50% of treated RNA was transferred for cDNA synthesis using the High-Capacity cDNA Reverse Transcription kit (Thermo Fisher). Resultant cDNA was diluted 1:20 with nuclease free water and 1% of cDNA was used in each 10 l qPCR reaction to assess steady-state levels of mt-mRNAs and mt-rRNAs. . To determine steady-state levels of mt-mRNAs and mt-rRNAs, the 2 − Ct method was applied to determine fold change of mt-RNAs relative to the mRNA level endogenous house-keeping control ACTB. Data analysis was performed using Microsoft Excel software. Data represented combination of four biological replicates and error bars represents standard error of the mean (SEM).
## Statistics
The data represented from IVT, oxygraphy, growth curve and quantification analysis are derived from the mean values. Error bars represent the standard deviation. Statistical analysis was done by a two-tailed, unpaired-unequal variance Student's t-test. The significance threshold was set at P < 0.05; indicated as * for P < 0.05, ** for P < 0.01 and *** for P < 0.001.
# Results
## Gtpbp5 interacts with the mt-lsu and late-stage assembly factors
GTPBP5 has previously been described as a mitochondrial protein in human cells [bib_ref] Human G-proteins, ObgH1 and Mtg1, associate with the large mitochondrial ribosome subunit..., Kotani [/bib_ref] , a finding which we were able to reproduce via confocal microscopy, by confirming mitochondrial localization of an exogenously expressed GTPBP5 protein with a C-terminal FLAG tag (GTPBP5::FLAG) in 143B human osterosarcoma (HOS) cells (Supplementary [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref].
As part of our study, we aimed to assess the mitochondrial protein interactome of GTPBP5 to discern whether the previously observed in vitro interaction of GTPBP5 with the mt-LSU occurs in living cells. We therefore generated a HEK293T Flp-In T-REx cell line, which allows for expression of the GTPBP5::FLAG in a doxycycline inducible manner. The GTPBP5::FLAG overexpressing cells, together with the control HEK293T, without FLAG protein expression, were tested via confocal microscopy to confirm mitochondrial localization of GTPBP5 (Supplementary [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref] , upper and middle panel). Following 48 h of GTPBP5::FLAG expression, we combined mitochondrial isolation and FLAG-IP to assess GTPBP5 protein interactors through the application of label-free quantitative mass spectrometry, relative to a negative control (HEK293T without FLAG protein expression; n = 3) [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref]. Through this approach, we observed enrichment of 51 out of 52 constituent proteins of the mt-LSU in the GTPBP5::FLAG pulldown [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref] and Supplementary [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref]. Interestingly, the only protein of the mt-LSU not enriched was bL36m, known to be incorporated to the mt-LSU at a very late stage of assembly [bib_ref] Structures of the human mitochondrial ribosome in native states of assembly, Brown [/bib_ref]. Of particular note, we did not observe enrichment of mt-SSU core proteins over that of global proteins [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref] , suggesting that GTPBP5 binds exclusively to the mt-LSU or mt-LSU late stage assembly intermediate and not to the 55S monosome particle.
Western blotting was performed for the input and eluate from the GTPBP5::FLAG IP experiments, which supported the specific interaction of GTPBP5::FLAG with constituent proteins of mt-LSU (mL37, bL28m), and not of the mt-SSU (uS15m, uS17m) [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref].
As a negative control, to exclude the possibility that mitoribosomal proteins may be non-specifically enriched during our FLAG-IP procedure, we performed FLAG-IP of mitochondrially-targeted luciferase (MtLuc::FLAG) under identical experimental conditions. We observed no enrichment of mt-LSU or mt-SSU proteins for our MtLuc::FLAG pulldown, assayed via western blotting (Supplementary [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref] , supporting the specific binding of GTPBP5 to the mt-LSU under these experimental conditions.
In addition to the enrichment of core mt-LSU proteins, GTPBP5::FLAG pulldown highly enriched known mt-LSU accessory factors, which have been implicated in mitoribosome biogenesis, including mitochondrial rRNA methyltransferase 2 (MRM2) [bib_ref] The human mitochondrial ribosome recycling factor is essential for cell viability, Rorbach [/bib_ref] [bib_ref] C7orf30 specifically associates with the large subunit of the mitochondrial ribosome and..., Wanschers [/bib_ref] , MTG1 (GTPBP7) [bib_ref] MTG1 couples mitoribosome large subunit assembly with intersubunit bridge formation, Kim [/bib_ref] , mitochondrial assembly of ribosomal large subunit 1 (MALSU1) [bib_ref] MTERF4 regulates translation by targeting the methyltransferase NSUN4 to the mammalian mitochondrial..., Cámara [/bib_ref] [bib_ref] NSUN4 is a dual function mitochondrial protein required for both methylation of..., Metodiev [/bib_ref] and L0R8F8 (7) [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref]. MRM2 has been shown to perform 2 -O-ribose methylation at position U(1369) in 16S mt-rRNA [bib_ref] The human mitochondrial ribosome recycling factor is essential for cell viability, Rorbach [/bib_ref] [bib_ref] C7orf30 specifically associates with the large subunit of the mitochondrial ribosome and..., Wanschers [/bib_ref] , and its high enrichment here may suggest that GTPBP5 binds to the maturing mt-LSU at a similar stage of assembly to MRM2. In addition, MTG1 was recently found to bind to a late stage intermediate of mt-LSU assembly [bib_ref] MTG1 couples mitoribosome large subunit assembly with intersubunit bridge formation, Kim [/bib_ref]. Recently, have reported that MALSU1, together with L0R0F8 and mt-ACP, bind to the mt-LSU at the subunit interface to prevent premature formation of the 55S monosome particle [bib_ref] Structures of the human mitochondrial ribosome in native states of assembly, Brown [/bib_ref]. The high enrichment of these assembly factors [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref] further support the suggestion that GTPBP5 may bind to a late stage assembly intermediate of the mt-LSU.
Of particular note, we observed the greatest logFC in abundance of peptides related to NSUN4 (NOL1/NOP2/Sun domain family member 4) and MTERF4 (mitochondrial transcription termination factor 4). MTERF4 and NSUN4 have been found together to form a stoichiometric complex that binds to the mt-LSU and has been shown previously to be required for formation of the 55S monosome particle [bib_ref] C6orf203 is an RNA-binding protein involved in mitochondrial protein synthesis, Gopalakrishna [/bib_ref] [bib_ref] Human GTPBP5 (MTG2) fuels mitoribosome large subunit maturation by facilitating 16S rRNA..., Maiti [/bib_ref]. Additional to its interaction with MTERF4, NSUN4 functions alone as a m 5 C rRNA methyltransferase, methylating the C841 residue in 12S mt-rRNA [bib_ref] NSUN4 is a dual function mitochondrial protein required for both methylation of..., Metodiev [/bib_ref]. In this role, NSUN4 would be expected to be bound to the mt-SSU, and therefore the enrichment observed for NSUN4 here is instead likely linked to its role in partnership with MTERF4, as we did not enrich for any mt-SSU proteins in our GTPBP5::FLAG-IP [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref]. The high enrichment of both MTERF4 and NSUN4 upon GTPBP5::FLAG pulldown may indicate that GTPBP5 is binding to a mt-LSU particle that is nearing translational competency.
Western blotting was further used to confirm the specific enrichment of those mt-LSU assembly factors identified in LFQ proteomic analysis (MTERF4, NSUN4 MTG1, MALSU1, [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref] , whilst additional known mt-LSU interacting factors (e.g. MRM3, FASTKD4), were not identified via either LFQ or western blotting [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref]. Through our immunoprecipitation analysis, therefore, we show that GTPBP5 interacts with the specific components of the late stage mt-LSU assembly intermediate within a biological system.
## Gtpbp5 ablation results in a mitochondrial gene expression defect
To further investigate GTPBP5 function, we produced a knockout cell line using plasmid-based CRISPR/Cas9 delivery. Two sgRNAs were designed to produce an out-offrame deletion targeted to the first protein-coding exon of the MTG2 gene, which encodes for the GTPBP5 protein (Supplementary [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref]. PCR analysis across the expected deletion site in our isolated GTPBP5 clone, revealed that the deletion was present on both alleles (Supplementary [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref]. To assess GTPBP5 expression in the produced knockout clone, we performed western blotting, which confirmed loss of expression in GTPBP5 KO [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref]. We assessed proliferation of GTPBP5 KO cells in high glucose medium, which resulted in a reduced rate of proliferation relative to that of control HEK293T cells [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref]. We further cultured both lines in galactose medium, which forces cells to rely on OxPhos as their major source of ATP production. A striking inability of GTPBP5 KO cells to proliferate in galactose media was observed [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref] , suggesting that a profound OxPhos defect is present in the GTPBP5 KO line.
We sought to further investigate the suspected OxPhos defect in the GTPBP5 KO line. Western blotting to assess levels of OxPhos complex proteins revealed a strong reduction in steady-state levels of Complex I protein NDUFB8 (nuclear-encoded), Complex IV proteins MTCO1, MTCO2 and MTCO3 (mtDNA-encoded) and Complex V protein ATP6 (mtDNA-encoded) whilst levels of nuclear-encoded SDHB (Complex II) and ATP5A (Complex V) were unaffected [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref].
To further support the specificity of the observed OxPhos phenotype to the loss of GTPBP5 expression, we derived a GTPBP5 KO line complemented with the GTPBP5::FLAG cDNA (GTPBP5 RESCUE ), which was stably introduced in a doxycycline-inducible manner using the Flp-In T-REx system. Expression of the GTPBP5::FLAG protein in the GTPBP5 RESCUE line was confirmed via western blotting [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref] and confocal microscopy [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref].
To understand the impact of GTPBP5 loss on mitochondrial bioenergetic function, we measured oxygen consumption. Mitochondrial respiration in the presence of Complex I or Complex II-linked substrates was significantly decreased in GTPBP5 KO cells [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref] , consistent with the decreased steady-state protein levels of Ox-Phos subunit NDUFB8 [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref]. Furthermore, mitochondrial oxygen consumption was partially recovered in GTPBP5 RESCUE cells [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref].
We next analysed in-gel activities of OxPhos complexes for control, GTPBP5 KO and GTPBP5 RESCUE lines [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref]. BN-PAGE analysis revealed a reduction in activities of Complexes I, IV and V in their independent forms, and within supercomplexes in the case of Complexes I and IV [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref]. In addition, for Complex V, we observed an accumulation of subassemblies with a reduced molecular mass relative to the F 1 F 0 holoenzyme in the GTPBP5 KO line that retained ATP hydrolysis activity and represent F 1 -containing subcomplexes. Similar subassemblies are reported frequently where mtDNA expression is compromised [bib_ref] Loss of LRPPRC causes ATP synthase deficiency, Mourier [/bib_ref] [bib_ref] Subcomplexes of mitochondrial complex V reveal mutations in mitochondrial DNA, Smet [/bib_ref] [bib_ref] Guanine nucleotides stabilize the binding of Bacillus subtilis Obg to ribosomes, Zhang [/bib_ref].
Crucially, the reduction in activity of Complexes I, IV and V observed in the GTPBP5 KO line is rescued upon reexpression of GTPBP5::FLAG cDNA [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref] , further supporting that the observed reduction in complex activities is specific to loss of GTPBP5, and likely is due to a general defect to mitochondrial gene expression.
## Ablation of gtpbp5 results in a severe mitochondrial translation defect
We decided to further study mitochondrial translation within the GTPBP5 KO line and observed a substantial decrease in [ 35 S]-Met/Cys labelled proteins derived from mitochondrial translation, to ∼50% of that observed in the HEK293T control line [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref]. This defect in mitochondrial translation was rescued upon expression of GTPBP5::FLAG in the GTPBP5 RESCUE line, confirming that the translation defect is linked to loss of GTPBP5 expression. Of note, induction of GTPBP5 re-expression was performed with 50 ng/ml doxycycline (as described in 37,39), leading to very high GTPBP5 expression levels [fig_ref] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression [/fig_ref] , however, this did not alter significantly mitochondrial translation. The same observation was made when testing mitochondrial translation in GTPBP5::FLAG overexpressing cell line [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref]. Nevertheless, we cannot exclude that high overexpression of GTPBP5 might, to some extent, inhibit protein synthesis, as previously observed for ObgE [bib_ref] Structural and functional insights into the mode of action of a universally..., Feng [/bib_ref].
Next, we investigated the steady-state levels of mt-mRNAs by quantitative real-time PCR [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref] and northern blotting (Supplementary . No major changes in stability of the transcripts were detected, apart from the increased steady-state levels of ND2, suggesting that the general translation defect was not caused by a lowered availability of mitochondrial RNAs to translate.
We therefore sought to explore whether the mitochondrial translation defect observed in the GTPBP5 KO line was linked to a defect in the mitoribosome. Steady-state analysis of mt-LSU and mt-SSU proteins revealed no significant reduction in the levels of mt-LSU and mt-SSU proteins overall, although some (bL28m, mL37) were mildly affected [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref].
Steady-state level analysis of mt-LSU accessory factors, including those enriched in our GTPBP5::FLAG pulldown experiments, revealed an upregulation in steady-state levels of a number of factors [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref] , with the most striking increase observed for MTERF4, which was highly enriched in our GTPBP5::FLAG-IP experiment [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref]. The increase in steady-state levels of several factors may reflect a compensatory mechanism to promote mt-LSU assembly under conditions of absence of GTPBP5. The lack of significant increase in NSUN4 levels, however, is notable, as it is thought to function as a heterodimer with MTERF4 on the mitoribosome to promote monosome formation. Total lysates from HEK293T and GTPBP5 KO cell lines were resolved via SDS-PAGE, immunoblotting was performed, and membranes were probed with OxPhos antibody cocktail against nuclear encoded proteins of complexes I (NDUFB8), complex II (SDHB) and complex V (ATP5A) and mtDNA-encoded complex IV protein MTCO2. In addition, membranes were probed with ATP6, MTCO3 and MTCO1 antibodies. GAPDH and SDHA were used as loading control. (C) Western blotting was performed on the control HEK293T (WT), GTPBP5 KO cells and KO cells expressing GTPBP5::FLAG (GTPBP5 RESCUE ) that were used for Oxygraph in (E). Antibody staining against FLAG was performed to confirm expression of the GTPBP5::FLAG protein. Complex II protein SDHA was used as loading control. (D) Mitochondrial respiration was measured in intact (basal) and in digitonin-permeabilized cells by the sequential addition of substrates and inhibitors to measure ADP-stimulated complex I respiration (CI-ADP; ADP, glutamate, malate, pyruvate) or ADP-stimulated complex I + II respiration (CI+II-ADP; + succinate), the complex I + II uncoupled maximal respiration rate (Max U/C; + CCCP) and the residual oxygen consumption (ROX; + rotenone, + antimycin A). Oxygen consumption is expressed in pmol of oxygen/s/10 6 cells and measured by Oroboros oxygraph. Data is presented as mean ± 1SD (n ≥ 3). Significant differences between wild-type and KO or between wild-type and rescue cell lines were determined by two-tailed Student's t tests. *P < 0.05, **P < 0.01. Pairwise P-values are presented in the following format for each conditions (HEK293T WT versus GTPBP5 [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref] , mt-SSU (uS15m, uS16m, uS17m, mS35 and bS18m) and to accessory proteins to the mitoribosome (MTERF4, NSUN4, MTG1, MRM2, MRM3, MALSU1). Antibodies to SDHA and GAPDH were used to assess loading. (D) Quantification of steady-state levels of mitoribosomal proteins was performed from three biological replicates using ImageJ software, with normalization of signal of protein to either SDHA (mL37, uS16m, mS35, uL15m, bS18m, uL4m, uL12m, mL45, mL49) or GAPDH (uL3m, uS15m, bL28, uS17m). Student's two-tailed t-test, P-values are as follows: (HEK293T vs GTPBP5
## Loss of gtpbp5 results in a defect in monosome formation
To further investigate the role of GTPBP5 in mitoribosome function, we performed sucrose gradient density fractionation of mitochondrial lysates derived from HEK293T control, GTPBP5 KO and GTPBP5 RESCUE lines followed by western blot analysis. Through this, we observed a reduction in the quantity of 55S monosome particles in GTPBP5 KO , whilst expression of the GTPBP5::FLAG protein recovered the abundance of 55S particles , B and Supplementary . In addition, we performed sucrose gradient centrifugation analysis followed by northern blotting to investigate 16S and 12S mt-rRNAs migration and observed a decrease of both rRNA levels in the monosome fractions in GTPBP5 KO cells, consistent with the Western Blot results (Supplementary . This suggests that GTPBP5 is important for mt-LSU assembly, with loss of GTPBP5 stalling the developing mt-LSU in a state which reduces its capacity to bind to the mt-SSU.
To better determine the stage of mt-LSU assembly at which GTPBP5 acts and therefore the potential mt-LSU intermediate that accumulates in its absence, we combined sucrose gradient centrifugation experiment with stable isotope labelling with amino acids in cell culture (SILAC). This allowed us to measure quantitative changes in the steady-state levels of assembly factors associated with the mt-SSU and mt-LSU. GTPBP5 KO cells in combination with HEK293T cells or GTPBP5 RESCUE cells were grown in media with 'light' amino acids or stable 'heavy' isotope labelled amino acids. Cells were pooled and isolated mitochondria were subjected to sucrose gradient centrifugation with collected fractions analysed by mass spectrometry. The mitoribosome assembly factors identified were quantified using the observed intensities of the relative peptides. Comparison of HEK293T sample with GTPBP5 KO revealed an increase in MTERF4 levels associated with the mt-LSU in GTPBP5 KO , as well as a slight increase of other mt-LSUassociated assembly factors such as MTG1 and MALSU1 , D). MRM2 and GTPBP10 levels associated with the mt-LSU did not change between control and GTPBP5 KO . Interestingly, DDX28 and MRM3 levels were decreased in GTPBP5 KO samples compared to control, suggesting they might be involved in different steps of assembly, not trapped in GTPBP5 KO . The levels of FASTKD5 and TRMT61B as well as some of the mt-SSU assembly factors (ERAL1, NOA1, TFB1M, METTL15) remained largely unchanged . Next, we extended our analysis by comparing GTPBP5 RESCUE sample with GTPBP5 KO . We did not observe major changes in mt-LSU-associated levels of MTERF4 between the two samples, nor several other factors (e.g. NSUN4 (previously not detected), MTG1, GTPBP10) (Supplementary [fig_ref] Figure 6: mt-LSU intermediate, and may possibly act downstream in the assembly pathway [/fig_ref] , suggesting that the factors trapped in GTPBP5 KO may also interact with the mt-LSU assembly intermediate containing GTPBP5.
## Interaction of gtpbp5 with the mt-lsu requires gtp hydrolysis
Previously, we observed an enrichment of core proteins of the mt-LSU upon GTPBP10::FLAG-IP when lysates were pre-incubated with non-hydrolysable GTP analog (GMPPNP) (Supplementary , and previously reported in Busch et al. [bib_ref] MitoRibo-Tag mice provide a tool for in vivo studies of mitoribosome composition, Busch [/bib_ref]. This indicated that GTP hydrolysis is required for release of GTPBP10 from the maturing mt-LSU, in line with what has been observed for the homologous ObgE protein in B. subtilis [bib_ref] Guanine nucleotides stabilize the binding of Bacillus subtilis Obg to ribosomes, Zhang [/bib_ref].
To determine if a similar mechanism of action is present for GTPBP5, we performed GTPBP5::FLAG pulldown experiment in the presence and absence of GMPPNP. In stark contrast to GTPBP10::FLAG, pre-incubation with GMPPNP prior to GTPBP5::FLAG-IP resulted in a complete loss of interaction of the mt-LSU subunits with GTPBP5::FLAG, as assessed via western blotting , B and Supplementary , suggesting that GTP hydrolysis is important for GTPBP5 binding to the mt-LSU, not for its release. Similar results were observed when we preincubated the lysates with GTP instead of GMPPNP, with a loss of interaction with the mt-LSU/MTERF4:NSUN4 for GTPBP5 and an enhanced interaction with the mt-LSU for GTPBP10 (Supplementary . Notably, GTPBP10 was not found to bind to MTERF4 in any of the conditions tested (Supplementary , reinforcing the hypothesis of a specific link between GTPBP5 and MTERF4:NSUN4 complex.
To further confirm the importance of the G-domain for GTPBP5 function, we designed a S238A mutant form of GTPBP5::FLAG protein at a highly conserved position in the Walker A motif in the G-domain [bib_ref] The C-terminal ␣-helix of YsxC is essential for its binding to 50S..., Wicker-Planquart [/bib_ref]. GTPBP5 S238A ::FLAG pull-down experiment showed no interaction of GTPBP5 S238A with the mt-LSU, nor with MTERF4 (Supplementary .This is in line with the GMPPNP experiments suggesting that the interaction is dependent on GTP hydrolysis.
To further assess the importance of GTP hydrolysis for GTPBP5 function, we generated a GTPBP5 KO line complemented with the GTPBP5 S238A ::FLAG cDNA (GTPBP5 RESCUE(S238A) ), which was stably introduced in a doxycycline-inducible manner using the Flp-In T-REx system, as previously performed for the generation of GTPBP5 RESCUE cell line. GTPBP5 RESCUE(S238A) cells were grown in galactose media and displayed a growth rate comparable to GTPBP5 KO cells, which was significantly lower when compared to HEK293T and GTPBP5 RESCUE cells (Supplementary . Additionally, a decrease in mitochondrial protein synthesis was observed in GTPBP5 RESCUE(S238A) and GTPBP5 KO cells compared to HEK293T and GTPBP5 RESCUE (Supplementary , therefore confirming that this mutation is important for GTP hydrolysis and consequently for GTPBP5 function.
To investigate the nature of GTPBP5 interaction with MTERF4 and NSUN4, we performed FLAG-IP from mitochondrial lysate, which had been pre-treated with RNase A prior to IP. Upon RNase A treatment, we anticipated that the rRNA within the mitoribosome would be degraded and that mt-LSU proteins would dissociate [bib_ref] Loss of LRPPRC causes ATP synthase deficiency, Mourier [/bib_ref]. Under these conditions, we observed a lack of enrichment of mt-LSU proteins upon GTPBP5::FLAG pulldown (mL37, bL28m) , which are clearly enriched in the absence of RNase A treatment [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref]. In addition, the enrichment of accessory factors MTERF4 and NSUN4 was also lost , suggesting that the interaction of GTPBP5 with the MTERF4:NSUN4 complex relies on the integrity of the mt-LSU or that of a late-stage mt-LSU intermediate.
# Discussion
Despite recent breakthroughs in our structural understanding of the mitoribosome and the identification of a number of accessory factors which have been implicated in mitoribosome biogenesis, the complete repertoire of required factors is likely far from complete and the mechanistic details of the function of already known factors need to be revealed. Here, we confirm that Obg-superfold family protein GTPBP5 can be added to the growing list of factors whose function relates to the maturation of the mt-LSU. During the review of our manuscript, a characterization of GTPBP5 has been reported by . Consistent with our data, Maiti et al. also found that GTPBP5 is involved in the late stages of the mtLSU assembly and GTPBP5 knock out affects mitoribosome maturation and consequently mitochondrial translation.
## The role of obg family gtpases in ribosome assembly
The presence of two Obg family GTPases in human mitochondria raises questions related to whether each protein has a divergent role in mt-LSU assembly and whether they bind to exclusive positions on the mt-LSU. Alignment of their primary sequences to that of E. coli ObgE, indicates an overall high level of conservation , as also shown in previous studies [bib_ref] The human Obg protein GTPBP10 is involved in mitoribosomal biogenesis, Lavdovskaia [/bib_ref] [bib_ref] Human GTPBP10 is required for mitoribosome maturation, Maiti [/bib_ref]. However, whilst loops 1, 2 and 3 of the Obg domain in GTPBP5 are predicted to structurally resemble those of ObgE (Supplementary , closer inspection of the primary sequence of GTPBP5 in loops 1 and 3 reveals differences in the key basic residues which have previously been shown in ObgE to be required for LSU binding, likely through direct rRNA coordination to the peptidyl transferase center (PTC) in 50S LSU (23) (shown in bold and underlined in Supplementary . This may reflect a distinct binding mechanism of GTPBP5 to the LSU compared to ObgE. Moreover, comparisons between the predicted GTPBP10, GTPBP5 and the resolved ObgE structures reveal distinct deletions and expansions, in the Obgand G-domains, respectively, of GTPBP10 (Supplementary . Also, the C-terminal domains (CTD) of each protein's primary sequence are highly divergent (Supplementary , not shown in the structural predictions due to their intrinsically disordered nature) [bib_ref] Structural and functional insights into the mode of action of a universally..., Feng [/bib_ref] [bib_ref] Assignment of 2 -O-methyltransferases to modification sites on the mammalian mitochondrial large..., Lee [/bib_ref]. Together, these differences may reflect on functional divergence between GTPBP5 and GTPBP10.
The role of Obg family member ObgE (E. coli) and other bacterial homologues in ribosome assembly has been extensively characterized and may provide clues to the ac-tion of GTPBP5 on the maturing mt-LSU [bib_ref] Structural and functional insights into the mode of action of a universally..., Feng [/bib_ref] [bib_ref] Assignment of 2 -O-methyltransferases to modification sites on the mammalian mitochondrial large..., Lee [/bib_ref] [bib_ref] C7orf30 is necessary for biogenesis of the large subunit of the mitochondrial..., Rorbach [/bib_ref] [bib_ref] The GTP-binding protein YlqF participates in the late step of 50 S..., Matsuo [/bib_ref] [bib_ref] The Caulobacter crescentus CgtAC protein cosediments with the free 50S ribosomal subunit, Lin [/bib_ref]. ObgE has been implicated in binding to a late stage assembly of the bacterial LSU (49), causing conformational changes at the interface of the 50S that are likely to prevent association of the LSU with the SSU, precluding 70S formation [bib_ref] Structural and functional insights into the mode of action of a universally..., Feng [/bib_ref].
Similar to ObgE, GTPBP5 appears to be excluded from the 55S monosome, evidenced by a lack of pulldown of components of the mt-SSU in our GTPBP5-IP [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref]. Moreover, bL36m, a protein incorporated at the very last stages of mt-LSU assembly (7) is also missing from the GTPBP5 pull-down experiments. Therefore, GTPBP5 is likely to be involved in the very late stages of assembly, and must be released before 55S monosome particles are formed.
Both GTPBP5 or GTPBP10 were found to functionally complement a loss of ObgE protein in E. coli [bib_ref] Human small G proteins, ObgH1, and ObgH2, participate in the maintenance of..., Hirano [/bib_ref]. However, through our investigation of GTPBP5 function, and through previous studies by ourselves and others [bib_ref] MitoRibo-Tag mice provide a tool for in vivo studies of mitoribosome composition, Busch [/bib_ref] [bib_ref] The human Obg protein GTPBP10 is involved in mitoribosomal biogenesis, Lavdovskaia [/bib_ref] [bib_ref] Human GTPBP10 is required for mitoribosome maturation, Maiti [/bib_ref] , we observed that the loss of function of either GTPBP5 or GTPBP10 results in a severe defect in mitochondrial translation, suggesting that whilst they are both GTPases of the Obg-superfamily, GTPBP5 and 10 have functions which are not redundant, although both are likely related to late stage maturation of the mt-LSU. Notably, with our experimental set up, we observed a 50% reduction of translation in GTPBP5 KO cells, while the steady state levels and activities of the OxPhos components were more drastically reduced. A similar phenotype has been observed in the case of GTPBP10 deficiency [bib_ref] The human Obg protein GTPBP10 is involved in mitoribosomal biogenesis, Lavdovskaia [/bib_ref] and in the recent study of GTPBP5 by where it was shown that while COX2 and COX3 steady-state levels were not detectable by western blot analyses, mitochondrial translation rates were decreased only by 50-60% [bib_ref] Human GTPBP5 (MTG2) fuels mitoribosome large subunit maturation by facilitating 16S rRNA..., Maiti [/bib_ref]. Taking that into consideration, we cannot rule out additional role of GTPases, downstream of mitoribosome maturation.
Mitoribosome biogenesis is not necessarily a linear process, and although both GTPBP5 and GTPBP10 are thought to bind to late state intermediates of the mt-LSU, it is unclear whether they function in sequential steps in maturation, or in concert at different regions of the complex. It is intriguing to note that no published datasets of pulldown of GTPBP10, or our presented data of GTPBP5, revealed reciprocal pulldown of the two Obg GTPases, despite both proteins being implicated in binding to late stage assemblies of the mt-LSU. It cannot be excluded, therefore, that the Obg domain of either protein may facilitate binding to the same site on the maturing mt-LSU, in a mutually exclusive manner, at different stages of assembly.
We are currently limited to structural predictions of GTPBP5 and GTPBP10. However, recent advances in cryo-EM structures have allowed the structural interrogation of assembly factors in concert with the mitoribosome (7). Obtaining similar structures of either GTPBP5 or GTPBP10 bound to the mt-LSU particle would allow direct determination of the binding sites of either protein in the maturing mt-LSU, including insight into how the interaction of the Obg-domain is coordinated with the mt-LSU and whether binding induces conformational changes to the mt-LSU, in a similar manner to ObgE.
## Mechanism of gtp hydrolysis of gtpbp5
There are conformational differences between the GTP and GDP-bound forms of Obg-family GTPases, due to alteration in the switch elements of the G-domain [bib_ref] Molecular modeling study for interaction between Bacillus subtilis obg and nucleotides, Lee [/bib_ref] , which allows for alterations in the angle between the Obg fold and the G-domain itself. This conformational change between different guanine nucleotide-bound states may be of particular significance for the role of both GTPBP5 and GTPBP10 in aiding conformational changes to mt-LSU. Here in, we found that pre-incubation of mitochondrial lysates with non-hydrolysable GMPPNP led to a loss of interaction of GTPBP5 with the mt-LSU under our experimental conditions , B and Supplementary . This was a surprising finding, since it has previously been shown that for the homologous ObgE, GTPase activity is stimulated when it is already bound to the LSU (23), with GTP hydrolysis linked to the release of ObgE from the LSU particle. Also, GTPBP10 shows an increased interaction with the mt-LSU following pre-incubation with GMPPNP prior to FLAG-IP (19) (Supplementary , indicating that GTP hydrolysis is linked to GTPBP10 release from the mt-LSU [fig_ref] Figure 6: mt-LSU intermediate, and may possibly act downstream in the assembly pathway [/fig_ref]. Interaction of another GTPase, NOA1 (C4orf14), with the mt-SSU has also been described to be dependent on GTP in a similar manner to ObgE and GTPBP10 [bib_ref] Human C4orf14 interacts with the mitochondrial nucleoid and is involved in the..., He [/bib_ref]. The behavior of GTPBP5 seems instead to reflect a necessity for GTP hydrolysis in promoting GTPBP5 binding to the ribosome itself [fig_ref] Figure 6: mt-LSU intermediate, and may possibly act downstream in the assembly pathway [/fig_ref] , suggesting that GTPBP5 association with the mt-LSU might induce potential conformational changes important for mitoribosome maturation. GTPBP5 release from the mt-LSU may instead be mediated through the action of additional protein factors, or through additional conformational changes in the mt-LSU at a later stage in assembly.
## Interplay of gtpbp5 with other mt-lsu assembly factors
The enrichment of both MALSU1:L0R8F8 and MTERF4:NSUN4 complexes with the mt-LSU particle obtained via GTPBP5::FLAG pulldown strongly supports the notion that GTPBP5 is interacting with a late stage mt-LSU assembly that is reaching maturity.
The increased steady-state levels of MTERF4 and its accumulation in the mt-LSU-containing sucrose gradient fractions observed in the GTPBP5 KO line is a potential compensatory effect to overcome the loss of function of GTPBP5 [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref]. It is intriguing that whilst MTERF4 and NSUN4 are required to function together to promote monosome formation [bib_ref] NSUN4 is a dual function mitochondrial protein required for both methylation of..., Metodiev [/bib_ref] , steady-state levels of NSUN4 is not upregulated in GTPBP5 KO , perhaps indicating that NSUN4 is not a limiting factor in the formation of the MTERF4:NSUN4 complex, and therefore upregulation of NSUN4 steady-state levels is not likely to provide a compensatory effect. This compensatory upregulation of MTERF4 and other accessory factors may explain why there is residual monosome formation and translation in the absence of GTPBP5. On the contrary, DDX28 levels decreased in GTPBP5 KO compared to control, suggesting that this factor is not associated with GTPBP5 The mechanisms of overexpression of accessory factors alleviating defects in ribosome assembly have previously been observed, indicating that ribosome assembly is not a linear process. For example, ObgE overexpression in E. coli was able to rescue ribosome assembly in a null mutation of RrmE/RrmJ/FtsJ (ribosomal RNA large subunit methyltransferase E), likely through stabilization of the conformation of maturing 50S subunits, without the requirement to rescue the methylation status of 23S rRNA [bib_ref] Overexpression of two different GTPases rescues a null mutation in a heat-induced..., Tan [/bib_ref]. Similarly, through experiments performed in yeast, overexpression of Mtg2p, the yeast homologue of GTPBP5, was found to recover the deleterious effect on mitochondrial translation of a model lacking the rRNA methyltransferase Mrm2p (24), the yeast homologue of RrmE.
In addition to MTERF4:NSUN4, in our GTPBP5::FLAG pulldown experiments we detected high enrichment of the mitochondrial rRNA methyltransferase MRM2 [fig_ref] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4 [/fig_ref]. Furthermore, we observed increased levels of MRM2 in GTPBP5 KO cells and a clear association with the GTPBP5 KO mt-LSU intermediate [fig_ref] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels... [/fig_ref]. Whilst we do not focus on the possible interplay between the functions of the GTPBP5 protein and MRM2 methyltransferase on the 16S mt-rRNA in this study, it is notable, as mentioned above, that in yeast the homologue of GTPBP5 was able to partially rescue MRM2 knockout [bib_ref] The yeast GTPase Mtg2p is required for mitochondrial translation and partially suppresses..., Datta [/bib_ref]. It is possible, therefore, that GTPBP5 functions in a similar way in the human mt-LSU, to promote adoption of a conformation that allows MRM2 binding or methylation of 16S. Cryo-EM analysis previously revealed that ObgE binds to a site adjacent to the methylation site of FtsJ (23S rRNA position U2552, PDB: 4CSU) (23), the bacterial homolog of MRM2. Importantly, in recently published study by , BioID experiments confirmed a strong interaction of MRM2 and GTPBP5 [bib_ref] Human GTPBP5 (MTG2) fuels mitoribosome large subunit maturation by facilitating 16S rRNA..., Maiti [/bib_ref]. Moreover, the levels of methylation of the 16S rRNA at positions U1369 and G1145 that are normally catalysed by MRM2 and MRM3, respectively, were shown to be partially reduced in the absence of GTPBP5 [bib_ref] Human GTPBP5 (MTG2) fuels mitoribosome large subunit maturation by facilitating 16S rRNA..., Maiti [/bib_ref] , suggesting an important interplay between MRM2 and GTPBP5.
Together, interaction of GTPBP5 with several assembly factor may play a role in quality control mechanisms, promoting ordered maturation of the mt-LSU during last stages of mitoribosome biogenesis.
## Data availability
Proteomic data from SILAC-based sucrose gradient analysis are deposited in ProteomeXchange, PXD017384.
## Supplementary data
[fig] Figure 1: GTPBP5 interacts with the mt-LSU proteins and the MTERF4:NSUN4 in the mitoribosome assembly complex. (A) Quantitative mass spectrometry analysis of proteins interacting with GTPBP5::FLAG. Following FLAG-IP, eluates from HEK293T expressing GTPBP5::FLAG and control HEK293T without FLAG protein expression (WT) were subjected to label-free quantitative mass spectrometry (LFQ) (n = 3). Proteins with a positive log 2 (fold change) (logFC > 3) and -log 10 (P-value) of >1.3 (adjusted P-value < 0.05) are identified. Identified proteins are as indicated in the key. A fully annotated version of this dataset is presented as SupplementaryFigure S2AandTable S5. (B) Boxplot comparison of the logFC of core proteins of the 39S (mt-LSU) or 28S (mt-SSU) relative to global proteins from dataset in A. Stated P-values indicate pair-wise significance of difference in logFC between global proteins, mt-LSU proteins or mt-SSU proteins as determined via Welch's unequal variances t-test. (C) Immunoblotting of GTPBP5::FLAG-IP. Input mitochondrial lysates and eluate of FLAG-IP from HEK293T expressing GTPBP5::FLAG and control HEK293T without FLAG protein expression (WT) were resolved via SDS-PAGE, western blotting was performed and subsequent membranes were probed with antibodies against FLAG and for proteins of either the mt-LSU (bL28m, mL37) or mt-SSU (uS15m, uS17m) (left panel), or for known mt-LSU assembly factors (right panel). The extra band in mL37 panel (*) represents NSUN4 signal from previous probing. [/fig]
[fig] Figure 2: GTPBP5 knockout leads to alteration in mitochondrial gene expression. (A) Western blotting to confirm knockout of GTPBP5. Total lysate from HEK293T WT and GTPBP5 KO clones was resolved via SDS-PAGE, immunoblotting was performed and membranes were probed with anti-GTPBP5 antibody and anti-SDHA as a loading control. The asterisk represents a shorter version of GTPBP5 protein. (B) Proliferation of GTPBP5 KO cells is compromised. Growth curves of WT HEK293T, and GTPBP5 KO cells in DMEM containing either 4.5g/l glucose or 0.9 g/l galactose (three biological replicates were performed and mean average cell numbers at each time point is indicated, error bars = ±1 SD). Student's two-tailed t-test. P-values are as follows: (HEK293T versus GTPBP5 KO ) Glucose: 2d, 0.0227; 4d, 0.0241; 6d, 0.0241. Galactose: 2d, 0.405; 4d, 0.00944; 6d, 0.000997. (C) Western blotting to assess steady state levels of OxPhos subunits in GTPBP5 KO . [/fig]
[fig] Figure 3: GTPBP5 knockout causes reduction in mitochondrial translation without affecting the steady-state levels of mitochondrial rRNAs/mRNAs and MRPs. (A) [ 35 S]-labeling of mitochondrial translation in wild type HEK293T cells, GTPBP5 KO cells, GTPBP5 RESCUE , and parental WT HEK293T cells overexpressing GTPBP5::FLAG. Following inhibition of cytosolic translation, cells were cultured for 30 minutes in the presence of a [ 35 S]-methionine and cysteine mix to specifically label products of mitochondrial translation. Total cell extracts (30 g) were resolved via 12% SDS-PAGE and visualized via autoradiography. Presented experiment is representative of three independent biological replicates. Relative quantification of [ 35 S]-labelling signal (right panel) for the lane relative to Coomassie blue stain as loading control, using Image J. Mitochondrial translation in GTPBP5 KO cells is compared to HEK293T WT, GTPBP5 RESCUE and HEK293T expressing GTPBP5::FLAG, as in the left panel, for three independent biological replicates. Student's two-tailed t-test, P-value: 0.0219. (B) Quantitative real-time PCR to determine steady-state levels of mt-RNAs. Relative steady-state level comparison of mitochondrial transcripts in HEK293T WT cells and GTPBP5 KO cells by qPCR (n = 4 biological replicates, error bars represent SEM). Student's two-tailed t-test, P-values are as follows: (HEK293T vs GTPBP5 KO ), 12S: 0.807, 16S: 0.126, ND1: 0.335, ND2: 0.111, ND3: 0.189, ND4: 0.397, ND5: 0.0721, ND6: 0.0230, COX1: 0.124, COX2: 0.526, COX3: 0.379, ATP6: 0.681, ATP8: 0.626, CYTB: 0.730. (C) western blotting to assess steady-state levels of proteins of the mitoribosome and accessory factors. Total lysate from HEK293T WT and GTPBP5 KO clones was resolved via SDS-PAGE, immunoblotting was performed and membranes were probed with antibodies to the mt-LSU ( [/fig]
[fig] Figure 4, Figure 5: GTPBP5 is important for monosome formation and its absence leads to accumulation of some late-stage mt-LSU assembly factors. (A) Sedimentation of mitochondrial ribosomes on 10-30% isokinetic sucrose gradients for HEK293T control, GTPBP5 KO and GTPBP5 RESCUE . Mitochondria were isolated from cells, and lysates were loaded onto gradients. Following centrifugation, obtained fractions were analysed by western blotting with antibodies against proteins of the mt-LSU (mL37), the mt-SSU (uS15m). (B) Quantification of the gradient distribution of mitoribosomal components (mL37 and uS15m) in HEK293T WT, GTPBP5 KO and GTPBP5 RESCUE cell lines. Measurements were performed using ImageJ software, indicating averages for three independent biological experiments. The y-axis represents the fraction of the total signal in each set of fractions relative to the total signal for HEK293T WT gradient. Error bars represent SEM. (C) SILAC-based proteomic analysis of mitoribosome assembly factors in sucrose gradient fractions. Heatmap of the identified assembly factors associated with the mt-LSU and mt-SSU were plotted as log2 transformed and normalized intensity values as indicated colours. (D) Heatmap plotted as described in (C) for the label swap experiment. GTP hydrolysis activity of GTPBP5 is required for binding to mitochondrial ribosomes and MTERF4:NSUN4 module. (A) FLAG-IP of GTPBP5 in the presence and absence of non-hydrolysable GTP analog, GMPPNP. FLAG-IP was performed from HEK293T expressing GTPBP5::FLAG and control HEK293T, where mitochondrial lysates were treated with or without GMPPNP (20 mM) during the lysis stage prior to IP protocol. Resulting input, mitochondrial lysates, and eluates were resolved via SDS-PAGE, western blotting was performed and subsequent membranes were probed with antibodies against FLAG and for proteins of either the mt-LSU (uL3m, bL28m) or mt-SSU (uS15m, uS17m). In addition, immunoblotting was performed for known accessory factors of the mt-LSU (MTERF4, NSUN4, MALSU1, MRM3, FASTKD4 and NGRN). (B) FLAG-IP as in (A), where mitochondrial lysates were treated with or without RNase A (20 g/ml), or with GMPPNP (20 mM), during lysis prior to IP protocol. Immunoblotting was performed with antibodies against FLAG and for proteins of either the mt-LSU (mL37, bL28m) or mt-SSU (uS15m, uS17m), and mt-LSU accessory factors MTERF4 and NSUN4. The extra band in GTPBP5 panel (indicated with *) might represent a shorter version of GTPBP5 protein, with its MTS removed, while the one in uS15m panel (*) a possible degradation product of GTPBP5. The extra band in mL37 panel (*) represents the FLAG signal from previous probing. [/fig]
[fig] Figure 6: mt-LSU intermediate, and may possibly act downstream in the assembly pathway (Figure 4C, D). Of note, when comparing GTPBP5 KO with GTPBP5 RESCUE , no major changes in the 368 Nucleic Acids Research, 2021, Vol. Model for GTPases action during mt-LSU biogenesis. (A) Mechanism of action of GTPBP10 during mitoribosome biogenesis. GTP-bound GTPBP10 binds to the premature mt-LSU and GTP hydrolysis is required for release of GTPBP10 from the mt-LSU. During this step, the energy released during GTP hydrolysis can potentially be used for driving conformational changes of the mitoribosome and/or binding/release of other assembly factors. (B) Mechanism of action of GTPBP5. GTP-bound GTPBP5 binds to a late-stage intermediate of the mt-LSU in a manner that requires GTP hydrolysis. The energy released during GTP hydrolysis can potentially be used for driving conformational changes of the mitoribosome. During later stages, GTPBP5 dissociates from the mt-LSU. mt-LSU-associated assembly factors levels were observed, consistent with our co-IP experiments showing association of GTPBP5 with MTERF4:NSUN4 complex, MRM2, MALSU1 and MTG1. [/fig]
|
A brief positive psychological intervention prior to a potentially stressful task facilitates more challenge‐like cardiovascular reactivity in high trait anxious individuals
When confronted with stress, anxious individuals tend to evaluate the demands of an upcoming encounter as higher than the available resources, thus, indicating threat evaluations. Conversely, evaluating available resources as higher than the demands signals challenge. Both types of evaluations have been related to specific cardiovascular response patterns with higher cardiac output relative to peripheral resistance indicating challenge and higher peripheral resistance relative to cardiac output signaling threat. The aim of this research was to evaluate whether a brief positive psychological exercise (best possible selves intervention) prior to a potentially stressevoking task shifted the cardiovascular profile in trait anxious individuals from a threat to a challenge type. We randomly assigned 74 participants to either a best possible selves or a control exercise prior to performing a sing a song stress task and assessed their level of trait anxiety. Cardiac output (CO) and total peripheral resistance (TPR) were continuously recorded through baseline, preparation, stress task, and recovery, respectively, as well as self-reported affect. Trait anxiety was related to higher CO in the best possible selves group and lower CO in the control group.While high trait anxious individuals in the control group showed increasing TPR reactivity, they exhibited a nonsignificant change in the best possible selves group.Moreover, in the latter group a stress-related decrease in positive affect in high trait anxious participants was prevented. Findings suggest that concentrating on strengths and positive assets prior to a potentially stressful encounter could trigger a more adaptive coping in trait anxious individuals. K E Y W O R D S best possible selves, biopsychosocial model, CO, TPR, trait anxiety 2 of 16 | SCHWERDTFEGER ET al. | 3 of 16 SCHWERDTFEGER ET al.
# | introduction
High anxious individuals are characterized by showing elevated anxiety symptoms, including cognitive (e.g., worry, confusion) and emotional (e.g., feeling nervous, agitated) signs of anxiety in diverse negative situations. Of note, the stronger affective stress experience in high anxious individuals might be rooted in cognitive evaluations of the stressor. Specifically, previous research has shown that anxious individuals appraise an upcoming stressor as more threatening and rate their coping resources as limited as compared to low anxious individuals, thus, suggesting a neurobiological sensitivity for threat. It might thus be assumed that a positive psychological exercise aiming at strengthening resources could ameliorate the intensified threat processing in anxious individuals, a notion that was tested in this study.
The amount of available resources and the demands imposed by the context constitute major foundations of evaluating an upcoming encounter as challenge or threat. Specifically, according to the biopsychosocial model (e.g.,, challenge results when available resources outweigh demands and threat results when demands outweigh resources. It should be noted that both challenge and threat have been considered cognitive appraisals playing a major role in the coping process as outlined by. They differentiate challenge from threat appraisals as part of the primary appraisal process contributing to the self-relevance of a given situation. Deviating from this view, the biopsychosocial model of challenge and threat emphasizes that challenge or threat represents the outcome of the coping process when the individual is engaged in a task of high self-relevance. Therefore, it has been suggested to use the term evaluations rather than appraisals when referring to challenge and threat as related to the biopsychosocial model. Of note, challenge and threat evaluations have been considered to reflect rather dynamic and mostly unconscious processes sensitive to various intrapersonal and contextual aspects (e.g., novelty or familiarity with the task, safety vs. danger cues, skill level, ability, and expertise;. Thus, challenge and threat evaluations can be modified rather than representing stable trait-like differences, although a habitual tendency toward either challenge or threat evaluations may exist .
Importantly, the biopsychosocial model suggests that challenge and threat evaluations could modulate cardiac and vascular reactivity in response to stress, particularly in situations of high self-or goal-relevance requiring active instrumental responding (so-called motivated performance tasks, like public speaking or mental stress). In general, engagement in these stress tasks, independent of cognitive evaluations is accompanied by increases in sympathetic and decreases in parasympathetic nervous system activity, as indicated by increases in heart rate (HR) and shortening of the time elapsed between the depolarization of the left ventricle and ventricular ejection (i.e., pre-ejection period; e.g.,. While this response allows the individual to engage in the respective task, the subjective evaluations of the demands posed by a motivated performance situation and the available intraindividual resources may considerably modify this response.
Precisely, according to, activation of the sympathetic adrenal medullary axis (SAM) enhances cardiac performance with elevated left ventricular contractility, increased cardiac output (CO; liters of blood pumped by the heart in 1 min, also referred to as cardiac efficiency) and reduced vascular resistance (total peripheral resistance, TPR; resistance of peripheral arteries, which determines the amount of oxygenated blood in the periphery). This response has been referred to as "toughness" and is characterized by high-quality task performance, emotional stability, and low anxiety. According to, this response could also be labeled active coping and corresponds to evaluating aversive situations as challenge rather than threat. Hence, when perceived resources outweigh demands a challenge response could be expected with comparably high CO and lower TPR. On the contrary, coactivation of the SAM and the hypothalamic-pituitary adrenal axis inhibits decreases in systematic peripheral resistance due to sensitizing effects of cortisol on the artery walls. This response pattern has been associated with threat evaluations. More specifically, when perceived demands outweigh available resources the resulting stress response is dominated by a vascular profile, that is a higher vasoconstriction (i.e., TPR) relative to CO. It should be emphasized that challenge and threat evaluations and their cardiovascular concomitants have been suggested to represent two anchors of a bipolar dimension, implying that relative rather than absolute differences are important.
Previous research found robust evidence for this model. For example, a cardiovascular challenge profile has been associated with better task performance; for a review, see, thus, substantiating the behavioral concomitants of challenge appraisals. Other studies examined associations between psychological concepts and cardiovascular indicators of challenge and threat. For example, individuals with stable high self-esteem showed a challenge-type cardiovascular response to failure feedback while those with unstable high self-esteem exhibited a threat-type cardiovascular response. Importantly, correlational evidence for the validity of the biopsychosocial model is further advanced by experimental studies manipulating resource and threat appraisals and analyzing their relative impact on the cardiovascular indices of challenge and threat, respectively (e.g.,. For example, manipulating resource and threat appraisals via brief instructions led to cardiovascular challenge and threat patterns, respectively. Furthermore, a cardiovascular challenge profile could be triggered by a positive performance feedback, which also resulted in enhanced visual attention toward gains. In a study by, reflecting on an upcoming self-relevant speech task using non-first-person language (self-distancing) resulted in a challenge-type cardiovascular response pattern as compared to using first-person language.
Conversely, a cardiovascular threat profile could be provoked by threat appraisals, stigma, stereotype threat, and benevolent sexism, among others. Furthermore, when social status got threatened and the status was unlikely to change, a cardiovascular threat pattern emerged, which changed to challenge, when instability was expected. Taken together, while challenge evaluations seem to benefit adaptive behavior and adaptive cardiovascular reactivity, thus, suggesting salutogenic effects, threat evaluations seem to foster vulnerability.
Of note, high anxiety proneness and related traits like neuroticism may be accompanied by subjective evaluations of threat rather than challenge (e.g.,and by elevated cardiac stress reactivity and exaggerated vascular stress responses (e.g.,. Correspondingly, in a study on social anxiety,could show that socially anxious women (scoring high on trait social anxiety) exhibited a cardiovascular threat response (i.e., higher TPR as compared to CO) during a social interaction task relative to their low socially anxious counterparts. It could thus be assumed that high anxious women confronted with a motivated performance task tend to evaluate their resources as lower and the demands as higher, thus, resulting in a cardiovascular pattern indicating threat.
Due to the modifiable nature of challenge and threat evaluations, previous research examined if brief instructions and manipulations prior to a stress task could modify cardiovascular reactivity from a threat-type response to a challenge-type response and performance, respectively. Specifically, reminding participants of their personal resources and evaluating a task as less demanding could shift the individual's evaluation from threat to challenge. In line with this reasoning, reappraising anxious arousal prior to "Karaoke" singing, public speaking, or a math performance task resulted in more challenge-oriented behavior and better performance. A similar challenge-inducing effect of a brief arousal reappraisal manipulation was reported by. Consequently, brief instructions and low-threshold interventions could prove useful to foster a positive, challenge-oriented mindset.
Strategies focusing on individual's strengths and resources are the main focus of positive psychological interventions, which have begun to attract attention among researchers for several years. A prominent example of a positive intervention is the best possible selves exercise. This kind of intervention asks participants to elaborate on their best possible future outcomes, thus, facilitating personalized representations of life goals. Meta-analyses and narrative reviews have shown that best possible selves interventions increase positive affect, optimism, and life satisfaction (e.g., .
Importantly, although such interventions have been designed as repeated exercises to foster an individual's well-being chronically, research also suggests that the immediate effects of a brief positive psychological intervention (so-called micro-intervention) are quite robust. In particular, following a best possible selves session, individuals have been found to exhibit increases in happiness, positive affect, and subjective well-being. However, the mechanisms involved in the effectiveness of a best possible selves intervention remain largely speculative. It has been hypothesized that writing about life goals may enable individuals to actively engage and pursue these goals, which requires visualization, self-awareness, and self-regulation. Other research identified reduced goal ambivalenceand-depending on the context-more approach-oriented behavioras key mechanisms of a best possible selves intervention.
Grounding on the biopsychosocial model of challenge and threat,identified several antecedents of challenge and threat evaluations. Specifically, increased self-efficacy, perceived control, and achievement-related/approach-oriented goals may all contribute to evaluating resources as higher than demands, potentially resulting in a cardiovascular challenge response. Hence, it could be assumed that a best possible selves intervention is particularly well-suited to trigger more challenge-oriented than threat-oriented evaluations, ultimately leading to the respective cardiovascular response patterns. More specifically, a best possible selves intervention may facilitate self-regulation, optimism, and active coping, and may thus shift an individual's evaluation in a motivated performance task toward challenge rather than threat. It should be noted in this respect that both optimism (e.g.,and life satisfaction (e.g.,have been associated with a challenge-type instead of a threat-type cardiovascular stress profile to a motivated performance task.
It could be assumed that a brief best possible selves intervention could be particularly beneficial for individuals with elevated anxiety, who have been found to show compromised possible selves and future outlook. In accordance with this,found-among others-that individuals scoring high in neuroticism seemed to benefit more from a best possible selves intervention relative to controls. Specifically, writing about personalized life goals buffered individuals high in neuroticism from decreases in happiness. Thus, the positive effects of a best possible selves exercise are expected to be particularly pronounced in individuals showing a tendency toward threat evaluations, such as emotionally unstable, high anxious individuals. Correspondingly,could observe that anxious individuals, who interpreted anxiety as facilitative appraised an upcoming exam as challenging rather than threatening. Even more,found that socially anxious individuals who were instructed to reframe stress arousal as functional evidenced higher CO and lower TPR to a social-evaluative stress task, thus, suggesting that instructing high anxious individuals in positive coping could lead to a more challenge-oriented cardiovascular stress response.
Hence, the aim of this research was to evaluate the effects of a brief best possible selves intervention on cardiovascular markers of challenge and threat in individuals with different levels of trait anxiety. In accordance with previous research, we expected that while high trait anxious relative to low trait anxious individuals should show a cardiovascular response profile associated with threat rather than challenge following an active control task (i.e., writing about their furniture), they were supposed to exhibit a challenge-type rather than threat-type response after performing a brief best possible selves writing task (i.e., higher CO relative to TPR).
# | method
## | participants
Sample size calculation was grounded on the assumption of medium effect sizes of micro-interventions on psychological indicators of healthas well as mediumto-large sized effects of a brief arousal manipulation on cardiovascular indicators of challenge and threat in socially anxious individuals. As power calculation in mixed effects models is difficult to ascertain when assumptions about random variances, slope variances, and intraclass correlations, among others, are lacking solid foundations, power analysis was conducted using a 2 × 4 mixed analysis of variance with medium effect size (f = 0.25) using GPower. The resulting sample size was N = 62. Overall, 82 individuals took part in the study, of whom eight were excluded due to excessive artifacts or multiple ectopic beats. Participants were recruited via oral communication, flyers at the university campus and the surroundings, and email announcements. The final sample for analysis comprised of 74 individuals (49 women, 25 men). Their mean age was 23.27 years (SD = 4.42) and the mean waist to hip-ratio (WHR) was 0.81 (SD = 0.10). Participants were eligible for study entry if they were free of physical and psychological medication and free of chronic and acute illnesses. Exclusion criteria were verified prior to the start of the experiment. Ethical approval by the institutional review board (GZ. 39/25/63 ex 2018/19) was obtained.
## | experimental manipulation
The study examined the effects of the best possible selves intervention in an experimental between subjects' design. Participants were randomly assigned to either a best possible selves intervention group (n = 36) or a control intervention group (active control, writing about their furniture; n = 38). Randomization was accomplished by assigning odd and even numbers to participants. Precisely, odd numbers were assigned to the best possible selves condition and even numbers to the active control condition.
In accordance with, the best possible selves condition was introduced as follows: "Now, please think about what your life will look like in the future (5-10 years from now). Imagine that everything went as great as possible. You have worked hard and successfully achieved all your life goals, both professionally as well as privately. Imagine that all your dreams have come true and that you live the best possible life. Write down this scenario as detailed as possible." After 10 min, they were further asked to elaborate in detail on the actions necessary to reach the goals for 5 min ("Now, please list concrete ideas that help to realize your goals. For example, activities that you could tackle, contacts you could make, ways in which you could use your time as efficiently as possible. There are no limits to your creativity here").
The active control group was instructed to describe the furniture of their apartment in detail (from entering the door to each room). After 10 min participants were further asked | 5 of 16 SCHWERDTFEGER to elaborate on the changes and adaptations in furniture since their moving in (for 5 min; the instructions are available upon request). Following the brief best possible selves or active control intervention, state affect was assessed.
## | stress task
The sing-a-song stress testwas introduced as potentially stressful task, which requested participants to (a) prepare and (b) sing a song in front of a camera for 3 min each. It should be noted that a similar "Karaoke" task proved sensitive in response to an arousal reappraisal intervention. A second informed consent was obtained asking about permission of video recording. During the preparation phase of 3 min the song ("Hollywood Hills" by "Sunrise Avenue") was delivered via a loudspeaker while the lyrics were presented on the computer screen. Subsequently, participants were asked to rate their affective well-being and how threatening and challenging the upcoming task was perceived (i.e., threat and challenge appraisal on a 4-point Likert scale ranging from 1 = "not at all" to 4 = "very much"). Then, participants were instructed to sing along to the song (stress phase of 3 min.). They were informed about their behavior being evaluated with respect to their performance, self-confidence, and talent. Importantly, the time-course of the BSP and the active control group was identical and the experimental procedure was administered fully automatized by means of the software PsychoPy.
## | variables and instruments
## | manipulation check
In order to analyze if the best possible selves intervention was conducted as per instruction, transcripts of the written elaborations in the best possible selves and active control group were made and analyzed with the linguistic word inquiry software LIWC (vers. 2015;. The total number of words, positive emotion words, positive feeling words, future-related words, and achievement-related words, among others, were analyzed.
## | trait anxiety
Trait anxiety was assessed using the trait version of the State-Trait Anxiety and Depression Inventory (STADI;. The STADI aims to assess both state and trait aspects of anxiety and depression. It is based on the State-Trait Anxiety Inventory (STAI;, but allows a reasonable separation of anxiety and depression symptoms. The trait version asks for anxiety and depression symptoms on a 4-point frequency scale ranging from 1 (= almost never) to 4 (= nearly always). Anxiety is assessed via 10 items, which can be further subdivided into worry (five items; e.g., "I think of the worst") and emotionality (five items; e.g., "I am bustling"). The STADI has proven reliable and valid in previous research as could be shown in healthy and clinical populations. The mean across the 10 items was calculated. Reliability of the scale was good (Cronbach's alpha = 0.87).
## | state affect
Positive affect (PA) and negative affect (NA) were assessed via the German version of the Positive and Negative Affect Schedule (PANAS;. PA and NA is assessed via 10 adjectives each (PA: e.g., active, strong, excited, inspired, enthusiastic; NA: e.g., guilty, upset, scared, nervous, ashamed) to be rated on a 5-point Likert scale ranging from 1 (not at all) to 5 (extremely). The mean score across the 10 items was calculated for both dimensions. PA and NA were assessed at baseline, post-intervention, immediately before the stress task and at recovery, respectively. Reliability coefficients (Cronbach's alpha) varied between .84 and .92 for PA and .62-.85 for NA.
## | self-reported challenge and threat
We assessed ratings of challenge ("How challenging do you consider the upcoming task?") and threat ("How threatening do you consider the upcoming task?") via unipolar items ranging from 1 = not at all to 4 = very much so. It should be noted that previous research examining the subjective indicators of challenge and threat used multiple items of task demands (e.g., uncertainty, low skills, and perceived danger) and resources (e.g., confidence, ability, familiarity;. The present research did not follow this convention, making the calculation of a challenge/threat-ratio not feasible. Although the single items-approach applied in this study does not have existing evidence that it is related to differentiating cardiovascular patterns of challenge and threat, it was applied on an exploratory basis as an alternative approach.
## | cardiovascular indicators of challenge and threat
CO and TPR were continuously recorded on a beatto-beat basis using the hemodynamic cardiovascular monitor Finometer® PRO (FMS, Finapres Medical Systems, Amsterdam, The Netherlands). The Finometer records blood pressure continuously via finger and height sensors (vascular unloading technique;. Initially, absolute accuracy of the brachial artery blood pressure is calibrated by a well-validatedprocedure (i.e., return-to-flow calibration; e.g.,using an upper arm cuff. The continuous brachial artery blood pressure is reconstructed by an initial calibration procedure using an upper arm cuff (i.e., return-to-flow calibration; e.g.,. Importantly, the Finometer® has an additional on board "Physiocal" calibration feature, which accounts for potential drifts in the data by an automatized and repeated application (max. every 70 beats). The Finometer® finger cuff was attached at participant's middle finger of the nondominant hand and the upper arm cuff was attached on the same side. CO (via modelflow estimate;and TPR were analyzed offline using BeatScope software (vers. 2.10), by considering individual's gender, age, body mass, and weight. CO is quantified as liters per minute and TPR as the ratio of the mean arterial pressure to cardiac output and is expressed as centimeter-gram-second units (dyn.s/ cm 5 ).
HR was assessed via an electrocardiogram applying a chest lead (modified Eindhoven II-lead; AccuSync® 72), which was recorded with a Biopac MP150 amplifier system (1,000 Hz). HR was analyzed offline via Kubios premium software (vers. 3.2;applying artifact correction (automatic artifact correction algorithm) if necessary.
## | procedure
Upon arrival at the laboratory, participants signed informed consent. Then, their waist and hip circumferences were measured. Subsequently, the electrodes were attached and the devices calibrated. Participants then worked on questionnaires to assess demographic variables and trait anxiety. Then, baseline recording started (3 min). In order to obtain basal resting values, participants viewed different landscape photographs during baseline (for a similar approach, see, which was followed by an affective state questionnaire. Subsequently, the instruction for the best possible selves group or the active control group was presented and participants engaged in writing for 15 min. Handwriting was required in both groups. Thereafter, state affect was assessed. Following the preparation period (3 min), the sing-a-song task was initialized (3 min), after which state affect was assessed again. Finally, the recovery period was followed (5 min), during which participants were instructed to fixate a green dot in the center of the computer screen. We aimed for a 5 min recovery period in order to account for the comparably slow recovery in cardiovascular measures (e.g.,. Presentation of a fixation cross was implemented during this period in order to standardize participants' attention, thus, mirroring a vanilla baseline approach (e.g.,. Then, state affect was assessed and subsequently the electrodes were detached and participants debriefed. The procedure is visualized in .
Participants were tested individually in a separated quiet room. Each intervention was conducted for 15 min prior to the stress task. The experiment lasted up to 1 hr and 15 min, and participants received course credit if applicable. Participants were instructed to refrain from intense physical activity and alcohol for 4 hr and from nicotine (if smokers) for 2 hr prior to the scheduled time.
## | data parametrization and analyses
Beat-to-beat values of CO and TPR were aggregated for each period. Values were visually checked for artifacts and outliers, which were removed if necessary (baseline: 0.28%, preparation: 0.38%, stress: 0.36%, recovery: 0.28%). Due to a skewed distribution TPR was logarithmized prior to analysis (lnTPR).
Mixed effects models were calculated accounting for random factors. We specified a random intercept model, thus, accounting for individual differences in intercepts. Several predictor variables were entered as follows.
F I G U R E 1 Study procedure with the main study phases and the main variables. CO, cardiac output; HR, heart rate; NA, negative affect; PA, positive affect; TPR, total peripheral resistance | 7 of 16 SCHWERDTFEGER ET al.
## | effects of the best possible selves versus active control exercise on cardiovascular indicators of challenge and threat
For predicting CO and TPR, trait anxiety (grand mean centered) was entered as a continuous predictor variable and group (active control vs. best possible selves) as a dichotomous variable. Task period (0 = baseline, 1 = preparation, 2 = stress, 3 = recovery) served as a factorize variables and the three-way interaction between group, task period and trait anxiety was entered. Additional models including covariates (sex age, WHR, physical activity, and smoking) were also calculated, but did not reveal different findings. Hence, we report the more parsimony models in order to avoid model over-specification.
## | effects of best possible selves versus active control exercise on psychological indicators of state affect
For predicting PA and NA, trait anxiety (grand mean centered) was entered as a continuous predictor variable and group (active control vs. best possible selves) as a dichotomous variable. Task period (0 = baseline, 1 = postintervention, 2 = stress, 3 = recovery) served as factorized predictor variable and the three-way interaction between group, task period and trait anxiety was entered in a final step.
In case of significant interactions involving trait anxiety or group, references were changed accordingly in order to allow for simple slope analyses. Specifically, trait anxiety was centered at the standard deviation (±1 SD) to allow separating trajectories for both low and high anxious participants. The reference for group was changed to 0 and 1 in order to explore both groups separately. Significance (p-)values were derived using Satterthwaite's degrees of freedom method. Models were analyzed using the statistical computing software R (vers. 3.5.3; R Core Team, 2019) with the packages lme4 (vers. 1.1-21;and lmer.Test (vers. 3.1-0;. The level of significance was fixed at p < .05 (two-tailed). Of note, in order to evaluate the robustness of the effects, we also applied post hoc power simulations using the R package simr, applying 10,000 Monte Carlo simulations.
# | results
## | sample characteristics
Participants of the active control and the best possible selves group were compared on a number of measures . Of note, both groups did not differ in trait anxiety, age, sex, WHR, smoking status, alcohol intake, regular physical exercise, baseline HR, CO, and lnTPR, thus, ensuring comparability between groups.
## | manipulation checks
Importantly, the total number of words in the writing tasks was comparable between groups (active control: M = 270.11, SD = 60.17; best possible selves: M = 252.78, SD = 50.88; t(72) = 1.33, p = .186). However, positive emotion words and positive feeling words were significantly more prevalent in the best possible selves as compared to the control Relative to the control group, the best possible selves group also evidenced a higher number of I-related words, future-related words, social words, achievement-related words, occupational words, insight words, communication-related words, and cognitive mechanism words, among others (Bonferroniadjusted p's < .0006; data available upon request). The active control group, on the contrary, scored higher on home-related words, physical words, as well as up and down-related words (Bonferroni-adjusted p's < .0006; data available upon request). Taken together, individuals in both groups performed the tasks in accordance with instructions. Subjective threat ratings did not differ significantly between groups as analyzed via t tests for independent samples (t(72) = 0.71, p = .944). However, subjective challenge ratings were significantly higher in the active control (M = 3.29, SD = 0.61) as compared to the best possible selves group (M = 2.86, SD = 0.76; t(72) = 2.68, p = .009). Subjective threat ratings were significantly positively associated with trait anxiety in the active control (r = .41, p = .012), but unrelated to anxiety in the best possible selves group (r = .26, p = .126). Subjective challenge ratings were unrelated to trait anxiety in both groups (active control: r = .26, p = .109; best possible selves: r = .23, p = .178). Finally, across groups subjective challenge ratings were significantly positively associated with CO reactivity to stress (r = .27, p = .021) and unrelated to TPR reactivity (r = −0.10, p = .402), while no significant relationships emerged between subjective threat ratings and both TPR reactivity (r = .01, p = .943) and CO reactivity (r = .08, p = .524).
In order to verify the motivated performance nature of the sing-a-song stress task, the change from baseline to stress was evaluated for HR using an analysis of variance with repeated measures (baseline vs. stress) and group as a between-subjects factor (active control vs. best possible selves). There was a significant and large-sized main effect for task period (F(1, 72) = 124.69, Wilk's Λ = .366, p < .001, η p 2 = .63), documenting a significant increase in HR to the sing-a-song task. Descriptive data are reported in . Importantly, there was no significant two-way interaction between task period and group (F(1, 72) = 1.70, Wilk's Λ = .977, p = .197, η p 2 = .02). Together these findings suggest that the sing-a-song task resulted in significant cardiac stress reactivity, thus, justifying its use as a motivated performance task for both groups.
## | predicting cardiovascular indicators of challenge (co) and threat (tpr)
The mixed effects models for predicting CO and lnTPR are depicted in . In general, CO increased from baseline to both preparation (b = 0.782, p < .001) and stress (b = 1.022, p < .001) and returned to baseline levels during recovery (b = −0.068, p = .505), thus, documenting an effective stress response. Importantly, a significant three-way interaction between group, trait anxiety and stress (b = 0.070, p = .001) emerged and suggested group-related differences in the relationship between CO reactivity and trait anxiety. Consequently, post hoc simple slope analyses were conducted by changing reference levels for both groups (active control as reference vs. best possible selves as reference) and trait anxiety (±1 SD) and rerunning the models. For the control group, the interaction between stress (vs. baseline) and trait anxiety was significantly negative (b = −0.079, p = .022), thus, suggesting lower CO increases with increasing trait anxiety. Of note, this relationship was reversed in the best possible selves group (b = 0.062, p = .014). Hence, anxiety was associated with lower CO reactivity in the active control and with higher CO reactivity in the best possible selves group. Further analyses documented that high trait anxious participants showed a lower CO in the active control as compared to the best possible selves group (interaction of group and stress: b = 0.376, p = .012), whereas for low anxious individuals this effect was reversed (b = −0.343, p = .022), thus, indicating a lower CO increase in the best possible selves relative to the control group. Taken together, these results suggest that trait anxiety significantly moderated the effects of the best possible selves intervention such that with increasing levels of anxiety CO reactivity became stronger in the best possible selves relative to the control group. A power simulation applying p < .05 resulted in a power of 91%, thus, suggesting a relatively robust effect.
For lnTPR there were significant increases from baseline to preparation (b = 0.071, p < .001), stress (b = 0.113, p < .001) and recovery (b = 0.154, p < .001), respectively. These findings document that the stress task resulted in elevated peripheral resistance, which endured into the recovery period. Moreover, a significant three-way interaction between group, trait anxiety and stress was found (b = −0.008, p = .037), which showed a similar descriptive trend when analyzing the raw untransformed variable (b = −12.19, t = −1.67, p = .097). Again, post hoc simple slope analyses were conducted to further elucidate this effect. Importantly, while for the control group the interaction between stress reactivity and trait anxiety was not significant (b = 0.007, p = .231), it was marginally significant for the best possible selves group (b = −0.009, p = .057), indicating a nonsignificant but potentially meaningful descriptive trend that with increasing anxiety levels peripheral resistance reactivity tended to be lower relative to the control group. Further analyses showed that for low trait anxious participants (−1 SD) lnTPR significantly increased from baseline to stress in both groups (active control: b = 0.097, p = .018; best possible selves: b = 0.135, p < .001). However, for high trait anxious participants, lnTPR significantly increased from baseline to stress in the control Unstandardized regression estimates; active control group = writing about one's furniture, best possible selves group = writing about one's best possible selves.
*** p < .001; **p < .01; *p < .05.
T A B L E 3 Linear mixed effects models for predicting cardiac output (CO; left side) and total peripheral resistance (TPR, logarithmized units; right side) in the course of the experiment group only (b = 0.171, p < .001), while it was virtually unchanged in the best possible selves group (b = 0.049, p = .163). Together these findings suggest that the best possible selves relative to the active control group resulted in lower peripheral resistance reactivity in high anxious individuals, thus, indicating a negative relationship in the best possible selves group in contrast to the expected positive relationship in the active control group with a noneffective treatment. Of note, a power analysis resulted in a power of 56%, thus, suggesting a rather fragile effect. The slopes for both models (CO and lnTPR) throughout the experimental periods in both conditions are depicted in (CO) and b (ln TPR). Of note, because previous studies often restricted analyses to the last minute of the baseline period and the first minute of the motivated performance task (e.g.,, we reanalyzed the outcome measures accordingly. In brief, analyses confirmed the findings presented above. In particular, the three-way interactions of group, trait anxiety and stress were significant for both CO (b = 0.086, t = 3.55, p < .001) and lnTPR (b = −0.01, t = −2.61, p = .01), thus, supporting the reliability of the effects.
## | predicting state affect (pa and na)
The mixed effects model for the prediction of PA and NA is depicted in . For PA there was a significant interaction of group and post-intervention (b = 0.221, p = .004). Follow-up analyses indicated that the control intervention resulted in a nonsignificant change in PA relative to baseline (b = 0.0005, p = .996). However, for the best possible selves group a significant increase in PA relative to baseline could be observed (b = 0.442, p < .001), thus, suggesting efficacy of the experimental manipulation. Moreover, there were significant three-way interactions between group, trait anxiety and stress (b = 0.036, p = .023) and group, trait anxiety and recovery (b = 0.048, p = .002), respectively. Post hoc simple slope-analyses documented that the interaction between trait anxiety and stress (vs. baseline) was not significant in the best possible selves (b = 0.025, p = .175), but tended to be negative in the active control group (b = −0.046, p = .067). Hence, trait anxiety tended to be negatively associated with PA only in the control group. When comparing high (+1 SD) versus low trait anxious individuals (−1 SD) it turned out that high trait anxious individuals showed a significant decrease in PA from baseline to stress in the control (b = −0.374, p = .024) and no significant change in the best possible selves group (b = 0.164, p = .253). Moreover, in the latter group high trait anxious participants showed a significant increase in PA during recovery (b = 0.295, p = .040), but they evidenced no change in recovery relative to baseline in the control group (b = 0.004, p = .981). Low trait anxious individuals, on the contrary, showed a significant increase in PA during recovery (relative to baseline) in the active control (b = 0.573, p < .001), but no such change in the best possible selves group (b = −0.115, p = .421). Taken together, the best possible selves intervention prevented a decline in PA relative to the active control group and led to a significant increase in PA during recovery among high trait anxious individuals, while low trait anxious individual appeared to be relatively unaffected by the best possible selves intervention.
For NA, there were significant main effects for stress (b = 0.484, p < .001) and recovery (b = 0.216, p < .001), respectively, documenting elevated NA relative to baseline. F I G U R E 2 Slopes (and confidence intervals) for cardiac output (a, left side) and total peripheral resistance (b, right side) for each period of the experiment. Slopes are derived from the mixed effects models | 11 of 16 SCHWERDTFEGER ET al.
Moreover, trait anxiety was associated with elevated levels of NA (b = 0.023, p = .015), which however, was dependent on stress (two-way interaction between trait anxiety and stress: b = 0.030, p = .002). Hence, the relationship between trait anxiety and NA was stronger during stress as compared to baseline. A significant interaction between group and recovery (b = 0.140, p = .003) further suggested that recovery values in NA were not significantly different from baseline in the active control (b = 0.076, p = .249), but significantly higher in the best possible selves group (b = 0.356, p < .001). There were no interactions involving group, trait anxiety and stress or recovery, respectively.
# | discussion
The aim of this research was to examine if trait anxiety moderates the efficacy of a well-evaluated positive psychological micro-intervention on cardiovascular indicators of challenge and threat. Specifically, in accordance with the biopsychosocial modeland previous research on brief instructions to facilitate challenge-type cardiovascular responses (e.g.,, we expected a shift to a more challenge-like cardiovascular response consisting of relatively higher CO and/or lower TPR in trait anxious individuals following a brief best possible selves exercise as compared to a group following a neutral writing exercise, during which anxiety was supposed to be associated with an increased vascular responding. Importantly, the findings confirmed expectations. The relationship between trait anxiety and CO increase from baseline to stress was positive in the best possible selves and negative in the active control group. In line with this, the relationship of trait anxiety with TPR reactivity was negative in the best possible selves and positive in the active control group. Correspondingly, trait anxiety was positively associated with subjective threat ratings in the control group only and affective responding was more positive following the best possible selves writing task in high anxious individuals as compared to the control writing task. Specifically, the decrease in PA reactivity to the stress task in high trait anxious participants could be prevented by the best possible selves exercise and they evidenced elevated PA during recovery in the best possible selves group only. Hence, this study complements previous research, which suggested T A B L E 4 Linear mixed effects models for predicting positive affect (PA; left side) and negative affect (NA; right side) in the course of the experiment that neuroticism (a more generalized concept of negative affectivity, including trait anxiety) moderated the effects of a best possible selves intervention on psychological outcomes . Importantly, the finding that a brief positive psychological exercise could change cardiovascular indicators of stress-related evaluations from threat and behavioral inhibition to a more approach-oriented, challenge type in high trait anxious individuals complements previous research suggesting that brief instructions regarding reappraisal of emotional arousal (e.g.,, positive performance feedback (e.g.,, mindfulness practice, focusing on gains rather than losses (e.g.,, or even nasal oxytocin applicationcould facilitate a challenge-type response profile.
Deviating from several previous studies, however, we aimed to identify trait anxiety as a possible moderator of the effectiveness of a brief positive intervention on cardiovascular indicators of challenge and threat. Of note, the findings confirm results of, who found that socially anxious individuals who were instructed to reframe stress arousal as a positive coping tool evidenced a challenge-type cardiovascular stress response as compared to those receiving no instruction. Thus, in line withit seems that focusing on resource appraisals like self-efficacy, perceived control, and approach-oriented goals facilitates a challenge-type stress response at least in a subset of individuals (namely those scoring high on (trait) anxiety).
The findings of the present research were restricted to cardiovascular indices of challenge and threat and were not entirely supported by subjective ratings of challenge and threat. Although we failed to show that anxiety is positively associated with subjective challenge ratings in the best possible selves condition, there was a positive association between trait anxiety and subjective threat ratings in the control group, thus, supporting previous findings of a close connection between anxiety and threat sensitivity (e.g.,. Moreover, while both groups did not differ in subjective threat ratings, the control group exhibited higher challenge ratings than the best possible selves group. This finding seems to contradict the assumption that focusing on positive assets and life goals should facilitate challenge rather than threat evaluations. However, it should be noted that challenge may have different meaning for different individuals. Challenge might be more closely associated with stress than threat, especially among young adults. Nonetheless, at least for subjective challenge ratings there was evidence for validity as became evident by the positive correlation with CO reactivity. It should be noted though that subjective challenge and threat evaluations as defined by the biopsychosocial model were not directly assessed as representing resources and demands, respectively, but rather via single items. Consequently, our approach to assess subjective indicators of challenge and threat did not allow a more finegrained and broad-banded assessment of task demands and resources as suggested by.
Importantly, the best possible selves exercise was particularly effective in individuals with elevated trait anxiety, but failed to benefit low anxious individuals on both the subjective and cardiovascular level. Of note, low anxious participants appeared to show the opposite response pattern than high anxious individuals, namely the expected more challenge-oriented response following the control writing task (i.e., higher CO reactivity and PA during recovery), but a relative threat-type response in the best possible selves writing task (i.e., lower CO). These findings are somehow challenging to interpret.have extensively elaborated on how and why positive psychological activities could backfire in some individuals. They note that activity overdose, overvaluing happiness, person-activity misfit, failed mediators, and social costs of positive activities could all contribute to adverse effects of positive psychological interventions. The unexpected findings in low anxious individuals could possibly be attributed to an overdose or ceiling effect in the best possible selves group. Low anxious individuals tend to evaluate upcoming stressors as less threatening and more controllableas became evident in this study by stronger CO reactivity and elevated PA during recovery in the control group. A further increase of approach motivation via the best possible selves micro-intervention could have led to an overly high allocation of resources to this exercise, thus, undermining motivational approach to the forthcoming motivated performance task. Certainly, further studies are needed to analyze when and for whom positive psychological interventions might have detrimental effects.
Importantly, manipulation checks suggested validity of the interventions. Specifically, positive, future-and achievement-related words were more prevalent in the best possible selves than in the control group. Moreover, PA increased more strongly following the best possible selves as compared to the control intervention. Some other findings warrant further discussion. First, both groups evidenced significant increases in HR from baseline to stress with a comparably large effect size, thus, suggesting that the sing-a-song task constituted a motivated performance situation as a necessary prerequisite for analyzing cardiovascular indicators of challenge and threat. Second, TPR increased consistently from baseline to preparation and stress, and further increased during recovery, thus, documenting a failure to return to baseline levels. Further research seems warranted to analyze the consequences of this failure to recover in such a young, healthy sample. Third, NA increased during recovery relative to baseline in the best possible selves group only, thus, documenting an increase in negative feeling states toward the end of the study. This finding may | 13 of 16 SCHWERDTFEGER ET al.
indicate a contrast effect following a positive psychological exercise or a kind of disillusion. As mentioned previously, there is increasing awareness about detrimental effects of positive psychological interventions, and given the novelty of the field and evidence for publication bias, this issue definitely warrants further research.
Although findings provide preliminary support for beneficial effects of a brief best possible selves intervention in trait anxious individuals, some limitations should be mentioned. First, replication of the results utilizing larger sample sizes is certainly warranted before further implications can be drawn. Although power simulations showed a comparably high power for the analysis of CO, associations with TPR appeared rather fragile and results must be regarded preliminary. Second, a major limitation constitutes the compromised assessment of subjective indicators of challenge and threat. Although we found some evidence for the validity of the single-time challenge rating, researchers are advised to assess the subjective perceptions of task demands and intraindividual coping resources in more detail and in accordance with the suggestions derived from the biopsychosocial model; see alsofor examples). Third, this study was not designed to evaluate stress-related long-term effects of a best possible selves exercise. Previous research could show that the effects of positive psychological micro-interventions diminish in the long range (e.g.,. Thus, further research is warranted to analyze the timeline of the effects as well as the generalizability to other motivated performance tasks. Fourth, the sample comprised of rather young and healthy adults. Hence, findings cannot be generalized to older populations or clinical (anxious) individuals and potential clinical implications remain speculative. Finally, findings are restricted to subjective and particularly physiological (i.e., cardiovascular) indicators of stress and coping and behavioral consequences of the intervention were not evaluated, although important. For example, there is evidence that vocal indicators of confidence and a more confident and dominant behavior are associated with a challenge-type stress response (e.g.,. Hence, it would be worthwhile to verify if anxious participants manifest more behavioral signs of motivational approach behavior (and possibly better performance) during the stress task following the best possible selves exercise.
# | conclusion
Notwithstanding the abovementioned limitations, the results of this study may inform about the efficacy of positive psychological interventions in trait anxious individuals exposed to diverse motivated performance situations. In particular, a brief positive psychological exercise prior to a stressful encounter could shift a cardiovascular response profile in high trait anxious individuals from a threat-related pattern with elevated TPR and lower CO to a challenge-type pattern with elevated CO and attenuated TPR. Thus, findings suggest that high trait anxious individuals could engage in a more adaptive coping when elaborating on personal life goals and positive assets prior to a stressful encounter. Using the well-established biopsychosocial model of challenge and threat, this study contributes to previous research indicating that brief instructions or interventions could reliably modify subjective and cardiovascular concomitants of coping. Future studies should analyze the health-related long-term effects of such interventions and why some individuals (e.g., low trait anxious) seem to not benefit from positive interventions. |
The Landscape of COVID-19 Vaccination in Zimbabwe: A Narrative Review and Analysis of the Strengths, Weaknesses, Opportunities and Threats of the Programme
# Introduction
The World Health Organisation (WHO) instigated emergency use authorisations (EUA) for the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) vaccines in December 2020 . Messenger ribonucleic acid (mRNA) based vaccines were the first to be authorised . This was the first time these vaccines were authorised for public health use globally. These were beyond the reach of many Sub-Saharan Africa (SSA) countries, especially as a number of them were not part of the COVID-19 Vaccines Global Access (COVAX) initiative. COVAX refers to an alliance of several established global health institutions that aimed to improve worldwide access to COVID-19 vaccines. The WHO called for equitable vaccine distribution, realising the considerable possibility of vaccine nationalisation and hoarding. As an unprecedented historical achievement, over one billion COVID-19 vaccines had been administered by the end of March 2021, just about four months from the onset of global vaccination programmes [bib_ref] Unprecedented achievement': Who received the first billion COVID vaccinations?, Kreier [/bib_ref]. Not surprisingly, most of these were administered in developed countries, while many countries from SSA were still to initiate vaccination. Despite substantial progress in vaccination by more powerful economies [bib_ref] Unprecedented achievement': Who received the first billion COVID vaccinations?, Kreier [/bib_ref] , SSA countries substantially remain behind, and the possibility that they will continue lagging remains high if vaccine nationalism continues [bib_ref] COVID-19 Vaccines: How to Ensure Africa has Access, Nkengasong [/bib_ref] [bib_ref] COVID-19 vaccine access in Africa: Global distribution, vaccine platforms, and challenges ahead, Loembé [/bib_ref].
In realising the challenges in accessing the vaccines that had EUAs, and not being a part of the COVAX initiative, Zimbabwe sought alternative sources of vaccines. By way of donations, the country received its first batch of 200,000 doses of Sinopharm vaccines from China, and subsequently more donations from China, Russia and India, kick-starting an ambitious vaccine rollout programme ahead of many of its African counterparts. The country then procured more vaccines for its population from China, which has continued as the major supplier of COVID-19 vaccines to date. Though the government later approved Johnson and Johnson vaccines for use, these have not become available. Remarkably, as of 5 January 2022, 7.34 million vaccine doses had been administered in Zimbabwe, with 3.18 million people (21.4% of the eligible population) fully vaccinated. The trajectory of the country's national vaccine programme since its inception in February 2021 to December 2021 is shown in [fig_ref] Figure 1: The trajectory of Zimbabwe's national vaccine programme since its in inception in... [/fig_ref]. This is commendable progress, placing Zimbabwe ahead of its many SSA counterparts. When Zimbabwe started the vaccination programme in February 2021, none of the vaccines used had obtained EUA from the WHO. However, in March 2021, the WHO extended its EUA to include Sinopharm vaccines. Subsequently, the Sinovac vaccines were also authorised by the WHO. accines 2022, 10, x FOR PEER REVIEW 2 of 12 of these were administered in developed countries, while many countries from SSA were still to initiate vaccination. Despite substantial progress in vaccination by more powerful economies [bib_ref] Unprecedented achievement': Who received the first billion COVID vaccinations?, Kreier [/bib_ref] , SSA countries substantially remain behind, and the possibility that they will continue lagging remains high if vaccine nationalism continues [bib_ref] COVID-19 Vaccines: How to Ensure Africa has Access, Nkengasong [/bib_ref] [bib_ref] COVID-19 vaccine access in Africa: Global distribution, vaccine platforms, and challenges ahead, Loembé [/bib_ref].
In realising the challenges in accessing the vaccines that had EUAs, and not being a part of the COVAX initiative, Zimbabwe sought alternative sources of vaccines. By way of donations, the country received its first batch of 200 000 doses of Sinopharm vaccines from China, and subsequently more donations from China, Russia and India, kickstarting an ambitious vaccine rollout programme ahead of many of its African counterparts. The country then procured more vaccines for its population from China, which has continued as the major supplier of COVID-19 vaccines to date. Though the government later approved Johnson and Johnson vaccines for use, these have not become available. Remarkably, as of 5 January 2022, 7.34 million vaccine doses had been administered in Zimbabwe, with 3.18 million people (21.4% of the eligible population) fully vaccinated. The trajectory of the country's national vaccine programme since its inception in February 2021 to December 2021 is shown in [fig_ref] Figure 1: The trajectory of Zimbabwe's national vaccine programme since its in inception in... [/fig_ref]. This is commendable progress, placing Zimbabwe ahead of its many SSA counterparts. When Zimbabwe started the vaccination programme in February 2021, none of the vaccines used had obtained EUA from the WHO. However, in March 2021, the WHO extended its EUA to include Sinopharm vaccines. Subsequently, the Sinovac vaccines were also authorised by the WHO. SARS-CoV-2 variants have continued emerging, and other waves of the COVID-19 pandemic are imminent. To date, variants with increased transmissibility from the ancestral wild type have emerged, namely the Beta, Gamma, Delta and more recently, in November 2021, the Omicron variant [bib_ref] Highly mutated SARS-CoV-2 Omicron variant sparks significant concern among global experts-What is..., Poudel [/bib_ref]. These have had properties sufficient to warrant being termed variants of concern (VOCs) by the WHO. As VOCs continue to emerge, with further associated epidemic waves, the COVID-19 pandemic may continue causing marked disruption to populations' socioeconomic well-being and health across the globe. Economic repressions, disturbances to academic learning activities and sporting activities, as well as suppression of global tourism activities are some of the negative socioeconomic consequences of the COVID-19 pandemic. SARS-CoV-2 variants have continued emerging, and other waves of the COVID-19 pandemic are imminent. To date, variants with increased transmissibility from the ancestral wild type have emerged, namely the Beta, Gamma, Delta and more recently, in November 2021, the Omicron variant [bib_ref] Highly mutated SARS-CoV-2 Omicron variant sparks significant concern among global experts-What is..., Poudel [/bib_ref]. These have had properties sufficient to warrant being termed variants of concern (VOCs) by the WHO. As VOCs continue to emerge, with further associated epidemic waves, the COVID-19 pandemic may continue causing marked disruption to populations' socioeconomic well-being and health across the globe. Economic repressions, disturbances to academic learning activities and sporting activities, as well as suppression of global tourism activities are some of the negative socioeconomic consequences of the COVID-19 pandemic.
Vaccination inarguably is the best public health strategy to limit the spread of rapidly progressive infectious diseases, reduce the associated morbidity, mortality and burden to healthcare systems, and allow a return to normal life activities. Population benefits of vaccination will be realised when an adequate proportion of the population has been vaccinated for SARS-CoV-2, which could be as high as 80-95% [bib_ref] Addressing challenges to rolling out COVID-19 vaccines in African countries, Nachega [/bib_ref]. Zimbabwe is still to reach this target, with less than 50% of the population currently fully vaccinated. To move the vaccination programme towards achieving this big but surmountable target, it is crucial to understand the strengths, weaknesses, opportunities and threats to the programme, to inform public health policy and strategy meaningfully. This review aimed to present some of the strengths, weaknesses, opportunities and threats to the Zimbabwean national COVID-19 vaccination programme since its rollout in February 2021.
# Methodology
We conducted a review of the current literature on the COVID-19 vaccination program in Zimbabwe. We searched for articles published in English on the WHO website; Google Scholar and PubMed; official public health websites operated by the government of Zimbabwe; and newspaper articles written and published within Zimbabwe. We used the following keywords: COVID-19; response; COVID-19 vaccine; SARS-CoV-2 vaccine; Zimbabwe; and other subject specific terms such as hesitancy; public health communication; pharmacovigilance; surveillance; vaccine access; vaccine distribution; policy. We used the Boolean operators AND and OR to separate the keywords. For instance, the search strategy Consistent with standard literature review methodology, some steps, such appraising evidence quality (a standard step in systematic reviews) were omitted.
To allow for a well-rounded review, the information gathered was structured and is presented according to the pre-established themes, strengths, weaknesses, opportunities and threats.
## Findings
The findings of this review are illustrated schematically in [fig_ref] Figure 2: Strengths, weaknesses, opportunities and threats to the national COVID-19 vaccination programme [/fig_ref].
Vaccines 2022, 10, x FOR PEER REVIEW 4 of 12
The findings of this review are illustrated schematically in [fig_ref] Figure 2: Strengths, weaknesses, opportunities and threats to the national COVID-19 vaccination programme [/fig_ref].
## Covid-19 vaccine swot analysis
# Strengths
The government of Zimbabwe demonstrated its political will and commitment to the vaccination programme by looking for alternative sources of vaccines earlier on when the WHO-approved vaccines were expensive and inaccessible. Pre-existing good relations between Zimbabwe and countries such as China, India and Russia enabled the government to access vaccines early, including at least 500,000 donated doses [bib_ref] Dynamics and Trends in Vaccine Procurement and Distribution in Zimbabwe, Maketo [/bib_ref]. Among the first to be vaccinated in the country was the Vice President, followed by other high-ranking officials. Healthcare workers were among the first targets of vaccination in Zimbabwe, and when some of them received the vaccines they shared their positive experiences on social and mainstream media. Such activities might have helped to build up vaccine confidence and uptake by the general population. People had been largely sceptical about receiving vaccines whose safety and effectiveness profiles were largely unknown.
Zimbabwe rolled out its phased COVID-19 vaccination programme in February 2021. Phase 1 of the programme targeted healthcare workers and frontline workers such as those working at ports of entry into the country [bib_ref] Implementing national COVID-19 vaccination programmes in sub-Saharan Africa-early lessons from Zimbabwe: A..., Murewanhema [/bib_ref]. The uptake of the vaccine by healthcare workers is believed to have positively influenced the rest of the population to get vaccinated [bib_ref] Commentary on the Zimbabwean People's Response towards the Anticipated COVID-19 Vaccine, Chigevenga [/bib_ref]. There are five vaccine choices approved for use in Zimbabwe‚ including China's Sinovac and Sinopharm approved by WHO for emergency use‚ Russia's Sputnik V and India's Covaxin. This gives the population options to choose from. There are high levels of trust in the World Health Organization and the country's Ministry of Health and Child Care (MoHCC) as sources of information. As such, the endorsement of the Sinopharm and Sinovac vaccines for emergency use by the WHO and their acceptance by the MoHCC positively impacted vaccine acceptance and uptake in Zimbabwe. With the availability of more doses, the vaccination programme entered into the subsequent phase 2 and phase 3, aiming at those with chronic conditions, other essential workers such as in the education sector, and, subsequently, the rest of the adult population above 18 years of
# Strengths
The government of Zimbabwe demonstrated its political will and commitment to the vaccination programme by looking for alternative sources of vaccines earlier on when the WHO-approved vaccines were expensive and inaccessible. Pre-existing good relations between Zimbabwe and countries such as China, India and Russia enabled the government to access vaccines early, including at least 500,000 donated doses [bib_ref] Dynamics and Trends in Vaccine Procurement and Distribution in Zimbabwe, Maketo [/bib_ref]. Among the first to be vaccinated in the country was the Vice President, followed by other high-ranking officials. Healthcare workers were among the first targets of vaccination in Zimbabwe, and when some of them received the vaccines they shared their positive experiences on social and mainstream media. Such activities might have helped to build up vaccine confidence and uptake by the general population. People had been largely sceptical about receiving vaccines whose safety and effectiveness profiles were largely unknown.
Zimbabwe rolled out its phased COVID-19 vaccination programme in February 2021. Phase 1 of the programme targeted healthcare workers and frontline workers such as those working at ports of entry into the country [bib_ref] Implementing national COVID-19 vaccination programmes in sub-Saharan Africa-early lessons from Zimbabwe: A..., Murewanhema [/bib_ref]. The uptake of the vaccine by healthcare workers is believed to have positively influenced the rest of the population to get vaccinated [bib_ref] Commentary on the Zimbabwean People's Response towards the Anticipated COVID-19 Vaccine, Chigevenga [/bib_ref]. There are five vaccine choices approved for use in Zimbabwe‚ including China's Sinovac and Sinopharm approved by WHO for emergency use‚ Russia's Sputnik V and India's Covaxin. This gives the population options to choose from. There are high levels of trust in the World Health Organization and the country's Ministry of Health and Child Care (MoHCC) as sources of information. As such, the endorsement of the Sinopharm and Sinovac vaccines for emergency use by the WHO and their acceptance by the MoHCC positively impacted vaccine acceptance and uptake in Zimbabwe. With the availability of more doses, the vaccination programme entered into the subsequent phase 2 and phase 3, aiming at those with chronic conditions, other essential workers such as in the education sector, and, subsequently, the rest of the adult population above 18 years of age, provided free of charge and on a voluntary basis. Due to good relations with vaccine producing countries, the country received 943,200 COVID-19 vaccine doses from the global COVAX facility in October 2021 to supplement current national vaccine distribution. The government has publicized its intentions to start vaccinating Zimbabwe's teenagers over the age of 16 against COVID-19 in November 2021 [bib_ref] Teenagers Now Eligible for Vaccination, Mohcc [/bib_ref]. Through their political campaigns in various parts of Zimbabwe, politicians have been trying to encourage hesitant people to get inoculated against COVID-19 amid pervasive misinformation. Several organisations in Zimbabwe, including parastatals, commercial companies and public health affiliates, have embarked on an operation to encourage their members to get vaccinated amid diffidence triggered by anti-vaccination posts on social media and other platforms [bib_ref] Commentary on the Zimbabwean People's Response towards the Anticipated COVID-19 Vaccine, Chigevenga [/bib_ref].
Zimbabwe has a well-run programme of expanded immunisation (ZEPI). Essential lessons to kick-start COVID-19 vaccination may have been obtained from ZEPI experiences, and the pre-existing cold-chain system provided a starting point for the national COVID-19 vaccination programme [bib_ref] Strengthening the expanded programme on immunization in Africa: Looking beyond 2015, Machingaidze [/bib_ref]. Strong collaborations with development partners such as the WHO and UNICEF are already in place, including platforms for training and monitoring and evaluation. A strong surveillance system set up by the MoHCC with support from the WHO and other partners enables close monitoring of the vaccination programme and provides vaccination statistics daily as part of the COVID-19 situation reports. The provincial coverage of COVID-19 vaccination in the country is illustrated in [fig_ref] Figure 3: Coverage of first and second COVID-19 vaccine doses by province in Zimbabwe [/fig_ref] and shows that the country has made great strides in ensuring equitable distribution across the ten provinces.
## Weaknesses
Circulating theories, mythologies, fictions and misunderstandings about the roots of COVID-19 and the intentions of vaccination continue to be big drivers of vaccine indecision and substantial obstacles to vaccine uptake in the general population. Zimbabwe has vaccines from China, Russia and India but not from Western nations. Acceptance of these vaccines has been sluggish and marred with wariness of Chinese-produced vaccines, disinformation, and averseness to the vaccines by some healthcare workers and prominent government officials not disclosing their vaccination status. Reported repudiations to accept the Chinese vaccines by some politicians and healthcare workers received farreaching attention in the media and might have negatively influenced public trust in these vaccines [bib_ref] Commentary on the Zimbabwean People's Response towards the Anticipated COVID-19 Vaccine, Chigevenga [/bib_ref]. The government has since penned a new law demanding all its employees to get vaccinated or face punitive action.
Vaccination centres occasionally run out of supplies, and poor urban settlements and rural areas have frequently been famished of doses [bib_ref] Zimbabwe's COVID-19 vaccination roll-out: Urgent need to rethink strategies to improve the..., Musuka [/bib_ref]. Citizens have oftentimes waited for hours, only to be told the vaccination centre was closing early due to shortages of doses or administering staff. Those entitled to second jabs have also complained of being turned back from centres giving preference to those pursuing a first jab. Long queues disheartened several unvaccinated residents due to the presence of few centres and the slow pace at which the healthcare workers are working to give the doses [bib_ref] Addressing challenges to rolling out COVID-19 vaccines in African countries, Nachega [/bib_ref]. There have been pervasive social media stories of COVID-19 vaccine hesitancy amongst pregnant and breastfeeding females, proliferated by circulating mythologies, delusions and rumours concerning the safety of vaccinations in this populace [bib_ref] Strengthening the expanded programme on immunization in Africa: Looking beyond 2015, Machingaidze [/bib_ref]. Presently, there is no policy on administering COVID-19 vaccines to pregnant and breastfeeding mothers, with subsequent turning away from this population from vaccination centres. Additionally, these lack of clear admissibility procedures led to lack of uniform practices, with some centres turning away clients, whereas others were vaccinating them. Vaccines 2022, 10, x FOR PEER REVIEW 6 of 12
## Weaknesses
Circulating theories, mythologies, fictions and misunderstandings about the roots of COVID-19 and the intentions of vaccination continue to be big drivers of vaccine indecision and substantial obstacles to vaccine uptake in the general population. Zimbabwe has vaccines from China, Russia and India but not from Western nations. Acceptance of these vaccines has been sluggish and marred with wariness of Chinese-produced vaccines, disinformation, and averseness to the vaccines by some healthcare workers and prominent government officials not disclosing their vaccination status. Reported repudiations to accept the Chinese vaccines by some politicians and healthcare workers received far-reaching attention in the media and might have negatively influenced public The population seems to have been more confident in the Western vaccines such as the Pfizer BionTech, Moderna and Astra Zeneca, none of which were available to the Zimbabwean population. The supplies of the vaccines have not been consistent, with episodes of stock ruptures, and those willing to be vaccinated not able to access the vaccines. Vaccination centres were reportedly limited and faced human resource challenges. Zimbabwe has been affected by a high level of healthcare worker migration to other countries that offer better remuneration and working conditions. Consequently, people have had to queue for long hours or days or move from one place to another searching for the vaccines. Sometimes, those who received the first dose have struggled to access the second dose, resulting in prolonging the period between the first and the second dose. A good proportion, close to 50% of the people who received the Covaxin vaccines from India, could not access their second dose after India ran into a harsh Delta variant driven epidemic wave and could not meet its obligations of supplying vaccines to other countries. The MoHCC did not decide on an alternative vaccine for the second dose for about six months for these clients who had failed to access their second COVAXIN dose. Logistical challenges in ensuring vaccines reach their destinations were initially widely reported [bib_ref] Zimbabwe's COVID-19 vaccination roll-out: Urgent need to rethink strategies to improve the..., Musuka [/bib_ref] , but this is expected with the inception of a programme, and the situation has since improved.
Lack of clear local guidelines regarding vaccination for specific groups such as the elderly, those with chronic conditions, and pregnant and breastfeeding mothers has resulted in confusion and frustration, as some were turned away from vaccination centres. Similar to authorities such as the WHO and the Royal College of Obstetricians and Gynaecologists, the MoHCC must come with its guidance adapted to the local context. Lack of clear information and guidelines can propagate vaccine hesitancy. An additional group for consideration of the vaccination programme are school-going children, as currently, the vaccination programme excludes all those below 16 years of age. As calls to make the school environment safer widen, and the population is frustrated with protracted school calendar disruptions, it is critical to urgently consider this important aspect of public health.
## Opportunities
The private sector in Zimbabwe is willing to participate in the national vaccination programme for very marginal profits. The government must leverage this to ensure wider vaccine access to the population by ensuring those who access the private sector are vaccinated there. However, it is critical to protect the population from exploitation and profiteering by private players. Several development partners have collaborated with the government of Zimbabwe in several of its public health programmes, including HIV/AIDS, malaria, ZEPI and maternal and child health. Such partnerships can facilitate training, monitoring and evaluation, vaccine advocacy and risk communication and community engagement, and help build vaccine confidence and uptake. There is a need for a urgent conversation centred on addressing the challenge posed by vaccine uncertainties in expecting and breastfeeding women and of apparent unanimity and guiding principles to back healthcare personnel involved in administering vaccines [bib_ref] Vaccination hesitancy among women of reproductive age in resource-challenged settings: A cause..., Murewanhema [/bib_ref]. This should be complemented by an all-encompassing education for healthcare personnel and the public to decrease misperception, upsurge confidence and increase vaccine uptake [bib_ref] Dealing with vaccine hesitancy in Africa: The prospective COVID-19 vaccine context, Afolabi [/bib_ref].
COVID-19 vaccination programs could be included in standard antenatal care procedures and baby clinics as potential solutions to attendance concerns at vaccination centres. Furthermore, antenatal clinics could give an excellent chance to provide targeted communications and counselling to expectant mothers [bib_ref] COVID-19 vaccine rumors and conspiracy theories: The need for cognitive inoculation against..., Islam [/bib_ref]. More than half of health facilities have maternal waiting shelters, which can become possible platforms for spreading COVID-19 vaccination information through verbal communications during consultations, or through printed material in vernacular langauges that the local residents understand. Health education targeted on the safety and effectiveness of the vaccines might positively influence vaccine intentions and subsequent uptake, mainly if communication emanates from trustworthy sources.
Zimbabwe has several medical associations, such as the Zimbabwe College of Public Physicians and Zimbabwe Society of Obstetricians and Gynaecologists. Close collaboration between the different societies provides an opportunity to develop clear evidence-based COVID-19 vaccination consensus guidelines for special populations such as the elderly, those with chronic conditions and pregnant and breastfeeding women. Consensus guidelines can help improve vaccine uptake and the efficiency of programmes. The recent approval of the Janssen vaccines by the Medicines Control Authority of Zimbabwe is a step in the right direction. The Janssen vaccines will provide an opportunity for those members of the population who were sceptical in taking the existing vaccines, widen the basket to choose from, and provides alternatives for those who may have allergies to the current vaccines.
## Threats
Vaccine hesitancy remains the biggest threat to the COVID-19 vaccination programmes globallyand is not a new phenomenon [bib_ref] COVID-19 vaccine hesitancy in South Africa: How can we maximize uptake of..., Cooper [/bib_ref]. Drivers of vaccine hesitancy are contextual [bib_ref] Dealing with vaccine hesitancy in Africa: The prospective COVID-19 vaccine context, Afolabi [/bib_ref] , and therefore qualitative explorations are essential to understand the drivers unique to specific populations. Zimbabwe is essentially a religious country, and religious influencers who have spoken widely against the vaccines have fuelled hesitancy. Several circulating myths, misconceptions and rumours regarding the origins of SARS-CoV-2 and the dangers of the vaccines in the population have circulated widely on diverse social media platforms, with an extensive reach [bib_ref] COVID-19 vaccine rumors and conspiracy theories: The need for cognitive inoculation against..., Islam [/bib_ref]. Around the start of the programme in Zimbabwe, there was news of a healthcare worker who had died in Masvingo after receiving a dose of the SARS-CoV-2 vaccine, and this scared the population.
The emergence of SARS-CoV-2 VOCs is a global threat to vaccination efforts, including Zimbabwe. VOCs have contributed significantly to subsequent epidemic waves in Zimbabwe. The second epidemic wave that occurred from December 2020 to January 2021 was mainly attributable to the Beta variant, with over 90% of the cases during the wave attributable to Beta [bib_ref] Drivers of the third wave of COVID-19 in Zimbabwe and challenges for..., Murewanhema [/bib_ref]. The subsequent third epidemic wave between June and August 2021 was 98% attributable to the Delta variant [bib_ref] Drivers of the third wave of COVID-19 in Zimbabwe and challenges for..., Murewanhema [/bib_ref] , and though genomic sequencing results are not yet published it is postulated that the Omicron variant was the major driver of the short-lived fourth epidemic wave of COVID-19 in Zimbabwe in December 2021. In Zimbabwe and beyond, the emergence of VOCs has been a major driver of these epidemic waves as they have increased transmissibility and the potential to evade public health interventions [bib_ref] The origins and potential future of SARS-CoV-2 variants of concern in the..., Otto [/bib_ref]. The lack of adequately vaccinated populations as well as populations with compromised immunity such as people living with HIV and AIDS (PLWHA) who are prevalent in Zimbabwe and sub-Saharan Africa has been perceived as a fertile breeding ground for new VOCs and breakthrough infections among the fully vaccinated [bib_ref] Insights from Zimbabwe's SARS-CoV-2 genomic surveillance, Dzinamarira [/bib_ref].
VOCs have the potential to propagate widespread community transmission, which can easily overwhelm fragile public health sectors such as in Zimbabwe. However, the greatest concern with the newer variants is their impact on vaccination programmes. VOCs may result in reduced vaccine effectiveness and shifts in herd immunity thresholds and accelerate the need for booster shots on a global scale. Evidence from other settings showed reduced effectiveness of mRNA vaccines against the delta variant, especially in terms of reducing incident infections, though the vaccines remained very effective against reducing hospitalizations and adverse outcomes from severe COVID-19 including mortality. Reduced effectiveness of the chemically inactivated vaccines against the Omicron variant has been reported as well, and these are the mainstays of Zimbabwe's COVID-19 vaccination programme. Messages of reduced vaccine effectiveness can reduce vaccine confidence and uptake and worsen the pre-existing vaccine hesitancy. Others have argued that natural immunity is a feasible goal with a highly transmissible variant such as Omicron and therefore do not see the need for getting vaccinated.
In the context of SARS-CoV-2, others have begun to argue that herd immunity may be an impossible goal with the continued emergence of VOCs.This raises the need for continued reinvigorating of the standard infection prevention and control practices such as the wearing of facemasks, hand hygiene and physical distancing. Unfortunately, this comes at a time when pandemic fatigue and complacency among populations due to a protracted battle with the COVID-19 pandemic are on the rise, which makes effective vaccination the most critical control strategy for the pandemic. The richer countries' demand for booster shots might lead to more vaccine hoarding and nationalisation, resulting in global shortages and supply chain disruptions. Low-income countries, including Zimbabwe, which are still to provide first doses for significant proportions of their populations, may lag behind significantly.
Zimbabweans are very mobile, primarily because of work and trade. European and other developed countries have started discussing the possibility of vaccine passports. Unfortunately, the vaccines used in Zimbabwe so far primarily have no recognition among the countries concerning the case. This may lead to those Zimbabweans intending to travel to the European Union countries and America shunning the current vaccines hoping for the approved vaccines. The approval of the Janssen vaccines is therefore welcome, but they are not yet available. The government must accelerate its efforts to procure these for the population urgently. This is more so as a heterologous approach to booster vaccination is being considered to be more superior an approach than homologous vaccination, especially against the emerging VOCs.
## Recommendations
We recommend local studies to objectively identify the strengths, weaknesses, opportunities, and threats to the COVID-19 vaccination in Zimbabwe to adequately inform public health and policy. Research in the country regarding COVID-19 policy and strategies is largely insufficient. The public health stakeholders need to understand and address adequately the drivers of vaccine hesitancy in the three main categories of complacency, confidence and convenience factors as categorised by the WHO. Critical to success is improving vaccine supplies, availability and accessibility, making sure that the most vulnerable groups in society are reached. The different professional medical organisations within the country must urgently come up with consensus guidelines to address especially special groups' concerns. To these ends, the government must leverage on the existing collaborations with the different development partners that operate within and outside the country and continue strengthening the bilateral relations with countries such as China, India and Russia.
# Conclusions
Though the Zimbabwe COVID-19 vaccination programme has several strengths and there are opportunities to strengthen it, vaccine hesitancy and the emergence of newer VOCs continue to pose significant threats to the current tremendous vaccination efforts in the country. There is a need to reinvigorate the vaccination programme in Zimbabwe, before further waves of COVID-19, to reduce the potential impacts of widespread community transmission on the public health sector. The public healthcare sector is highly fragile, and vaccination will serve to reduce strain by reducing hospitalisations, morbidity and mortality from COVID-19. Therefore, the government must urgently leverage on the strengths and opportunities to build vaccine confidence and uptake and address the weaknesses and threats adequately. There is an urgent need to continue procuring more vaccines to avoid stock outages, acquire other vaccines, especially the mRNA-based vaccines to widen the basket to choose for the consumers, and for strengthened messages through the various media available to build vaccine confidence and uptake. Addressing vaccine hesitancy is extremely important moving forward. Therefore, to achieve the targets, an urgent multisectoral intersection between the different relevant stakeholders in public health is required, as well as strengthening ties with the private sector. To be able to acquire other vaccine types, given the limited financial resources in the country, there is need for better cooperation with international partners, including joining pre-existing arrangements such as the COVAX initiative.
[fig] Figure 1: The trajectory of Zimbabwe's national vaccine programme since its in inception in February 2021. [/fig]
[fig] vv 2 variantsFigure 2: Effective health promotion strategies v Pre-existing cold-chain system v Effective vaccine procurement systems v Vaccine confidence v Political will to roll-Vaccine hesitancy v Conspiracy theories v Vaccine supply chain disruptions v Lack of transparency on the side effects of vaccine v Religious objections to vaccination v Easy accessibility of social media with wider reach of rumors, falsehoods, myths, misconceptions v Unavailability of second COVAXIN doses v The emergence of new SARS-CoV-Strengths, weaknesses, opportunities and threats to the national COVID-19 vaccination programme. [/fig]
[fig] Figure 2: Strengths, weaknesses, opportunities and threats to the national COVID-19 vaccination programme. [/fig]
[fig] Figure 3: Coverage of first and second COVID-19 vaccine doses by province in Zimbabwe. [/fig]
[fig] Author: Contributions: Conceptualization, G.M. (Grant Murewanhema) and T.D.; methodology, G.M. (Grant Murewanhema) and T.D.; writing-original draft preparation, (Grant Murewanhema).; writing-review and editing, G.M. (Grant Murewanhema), G.M. (Godfrey Musuka), K.D., I.C. and M.P.M. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: Not applicable. [/fig]
|
Targeting the SLIT/ROBO pathway in tumor progression: molecular mechanisms and therapeutic perspectives
## Slit/robo pathway as an oncogene in tumor progression
In the last decade, studies have demonstrated the dual roles of SLITs and ROBOs as axon guidance cues in the developing nervous system, where they both attract and repel neuronal migration. This bifunctionality is also observed in cancers as both oncogenes and tumor suppressor genes. The expression of SLIT and ROBO is altered in a wide variety of cancer types, identifying them as potential therapeutic targets.
Previous studies have shown that high levels of SLITs and ROBOs are expressed in many types of tumors and SLIT/ROBO signaling has a positive effect on tumor growth. For instance, by means of microarray analysis and real-time polymerase chain reaction (PCR), Groene et al.showed that the expression of ROBO1 and ROBO4 was significantly upregulated in colorectal carcinoma compared with normal tissue, while SLIT2 showed no differential expression between colorectal carcinoma and normal tissue. ROBO1 expression was mainly in tumor cells, whereas ROBO4 was located primarily in the endothelial cells of tumor vessels.Zhou et al.demonstrated that the expression of SLIT2 and ROBO1 was significantly associated with an increased metastatic risk and poorer overall survival in colorectal carcinoma patients. ROBO1 and SLIT2 mRNAs were detected in breast cancer cell lines and breast cancer tissues, and ROBO1 expression was elevated in breast carcinoma compared with normal breast mammary epithelial cells.There was also a striking increase in SLIT1 expression in prostate tumors.Owing to the overexpression of ROBO1 in hepatocellular carcinoma and the shedding of ROBO1 into serum in humans, this receptor of SLITs was thought to be a potential new serological marker for hepatocellular carcinoma.In addition, SLIT2 and ROBO1 were overexpressed in human mucoepidermoid carcinoma Mc3 cells.SLIT2/ROBO1 pathway promoted the Mc3 cells proliferation and the treatment of Mc3 cells with the monoclonal antibody R5 which can interrupt the SLIT2/ROBO1 pathway caused significantly suppressed cell growth and proliferation and markedly lowered the expression of PCNA.Furthermore, SLIT2 expression was correlated with the loss of basement membrane in the samples of human skin squamous cell carcinoma at different stages of disease progression. The SLIT2-Tg mice were found to develop significantly more skin tumors than wild-type mice, the skin tumors that occurred in SLIT2-Tg mice were significantly larger than those in the wildtype mice after 7,12-dimethylbenz[a]anthracene initiation until the end of the experiment. SLIT2 also could promote the invasive ability of the squamous cell carcinoma cell line A431 and this effect could be significantly repressed by the antibody R5.Tumor metastasis, the process of tumor cells migrating to other distant organs, invading blood and lymphatic vessels and leading to secondary tumor formation, is important for cancer development. Recent studies indicated that the SLIT/ ROBO signaling pathway plays a role in promoting tumor cell migration and, thus, promotes tumor metastasis. SLIT2 functions as a potent chemoattractant for breast cancer cells, inducing migration of cells expressing ROBO1. Furthermore, the SLIT2/ROBO1 signal was shown to upregulate MMP-9 to enhance breast cancer cell invasion.Low mRNA expression of the ROBO2 was associated with poor patient survival, whereas high mRNA expression of ROBO3, a known inhibitor of ROBO2 signaling, demonstrated an appropriate reciprocal inverse association with poor survival in pancreatic cancer.ROBO3 expression is upregulated in pancreatic cancer tissue samples and pancreatic cancer cell lines. Overexpression of ROBO3, which was associated with activated Wnt/β-catenin and GSK-3β, and other markers indicating epithelial-mesenchymal transition (EMT), promotes pancreatic cancer cell growth, invasion, and metastasis in vitro and in mouse xenograft tumor models. MiR-383 was also identified as a suppressor of ROBO3, and its expression was inversely correlated with ROBO3.To further explore the SLIT/ROBO signaling in tumor progression, Yang et al.intercrossed SLIT2 transgenic mice with a nonmetastatic RIP1-Tag2 mouse tumor model. They found that transgenic overexpression of SLIT2 significantly enhanced tumor lymph angiogenesis and subsequently promoted mesenteric lymph node metastasis of pancreatic islet tumors. SLIT2 was expressed minimally in normal and hyperplastic mucosa, moderately in dysplastic mucosa, and highly in neoplastic mucosa obtained from hamster buccal pouch in a multistage model of 7,12-dimethyl-1,2-benzanthraceneinduced squamous cell carcinoma, and increased SLIT2 expression was associated with higher tumor angiogenesis. 47 Importantly, interruption of the SLIT2-ROBO1 interaction using R5 inhibited tumor angiogenesis and growth in an in vivo model. Therefore, targeting SLIT/ROBO signaling may offer a novel approach for oral cancer therapy.
As is well known, the EMT is one of the initiating steps that play a key role during tumor invasion and metastasis. In colorectal epithelial carcinoma cells, recombinant SLIT2 inducing ROBO1 expression recruited a ubiquitin ligase Hakai for E-cadherin ubiquitination and lysosomal degradation, thus promote the EMT, tumor growth, and liver metastasis. Moreover, this effect can be attenuated by knockdown of Hakai.Therefore, the SLIT2, ROBO1, and ROBO3 may function as oncogenes that promote cancer proliferation and metastasis which may provide potential target structures for the antitumorigenic and antiangiogenic therapy of these special carcinomas.
## Slit/robo pathway as tumor suppressor genes
Compared with several studies on promoting tumor progression, SLITs and ROBOs are tumor suppressor genes in some special tumors. Here we further elucidate the anticancer function of the SLIT/ROBO pathway.
It has been shown that SLIT/ROBO pathway genes are frequently inactivated by promoter region's hypermethylation, resulting in downregulated gene expression in many human cancers. 58,60-62 SLIT2 was methylated in 71% (5/7) of glioma cell lines and in 59% (37/63) of other tumors and the SLIT2 expression was downregulated in methylated gliomas tumor samples, which indicated that SLIT2 was frequently inactivated by promoter region CpG island hypermethylation in gliomas and might be a good candidate for a glioma tumor suppressor gene. 59 SLIT2 expression was reduced in CRC tissues because of hypermethylation of the SLIT2 gene in CRC cells, and SLIT2 could inhibit CRC cell migration that required USP33 by deubiquitinating and stabilizing ROBO1.Another study also showed that USP33 was downregulated in lung cancer patients and that low expression of USP33 was associated with poor prognosis, which may be associated with reduced protein stability of ROBO1 in lung cancer cells.
## Slit2/robo1 pathway interrupts the hgf/c-met mediating cancer progression
The hepatocyte growth factor (HGF) and its receptor, the transmembrane tyrosine kinase c-MET, promote cell proliferation, survival, motility, and play a crucial role both in tumor progression.The shRNA-mediated depletion of SLIT2 or ectopic expression of a soluble decoy ROBO enhance HGF-induced migration, matrix invasion, accompany with the upregulation of Cdc-42 and the downregulation of Rac-1 activities. Accordingly, autocrine overexpression or exogenous administration of SLIT2 prevent HGF-induced motile responses, reduce Cdc-42 activation, and stimulation of Rac-1.In addition, medulloblastoma invasion was inhibited after recombinant SLIT2 protein treatment, which was accompanied with downregulation of activated Cdc42. 67
## Slit2/robo1 pathway prevents the cxcl12/ cxcr4 induced cancer progression
Upregulation of CXCR4 is associated with poor prognosis in breast cancer and pancreatic cancer. showed that loss of SLITs (SLIT2, SLIT3) or their ROBO1 receptor in murine mammary gland or human breast carcinoma cells resulted in coordinate upregulation of the CXCL12 and CXCR4 signaling axis, which was accompanied by hyperplastic changes in cells and desmoplastic alterations in the surrounding stroma. Furthermore, SLIT overexpression downregulated CXCR4 and dominantly suppressed tumor growth in a xenograft model.SLIT2 treatment inhibited CXCL12/CXCR4-induced breast cancer cell metastasis. SLIT2 inhibited CXCL12-induced tyrosine phosphorylation of focal adhesion components such as RAFTK/Pyk2 at residues 580 and 881, focal adhesion kinase at residue 576, and Paxillin. It was also found that SLIT2 inhibited CXCL12-induced phosphatidylinositol 3-kinase (PI3K), p44/42 MAPK, and MMP-2 and MMP-9 activities, but it did not have an effect on JNK and p38 MAPK activities.These studies classified SLITs as negative regulators of the CXCL12/CXCR4 pathway and identified a molecular signature in hyperplastic breast lesions that signified inappropriate upregulation of key prometastatic genes.In addition, there were some studies that revealed that SLIT/ROBO signaling might not play an important role in regulating human cancer cell proliferation and migration. Dai et al. found that three major members (SLIT2, SLIT3, and ROBO1) of the SLIT/ROBO family were widely expressed in the human normal and malignant ovarian tissues and in OVCAR-3 and SKOV-3 cells.However, recombinant human SLIT2 did not significantly affect SKOV-3 cell migration, and OVCAR-3 and SKOV-3 cell proliferation. SLIT2 also did not induce ERK1/2 and AKT1 phosphorylation in OVCAR-3 and SKOV-3 cells. Sanz-Pamplona et al. also demonstrated that no significant association between SLIT2 or SLIT3 level of expression and prognosis was found in colon cancer.In short, these inconsistent results showed the complexity of the role of the SLIT/ROBO signaling pathway in tumor proliferation. More thorough and careful studies are needed to clarify the precise role of this pathway. . They further showed that srGAPs inactivated Cdc42, resulting in reconstruction of cytoskeletal proteins via regulation of actin polymerization to inhibit cell migration in the mammalian nervous system. Ubiquitin-specific protease 33 (USP33)/VDU1 was also shown to be involved in ROBO1-USP33 interaction and participate in SLIT/ROBO signaling in cancer cell migration.Another study also showed that srGAP1 is an important downstream molecule of Slit2 signaling in CRC, and mediates the antimigration function of Slit2 by inhibiting Cdc42.SrGAP2 protein expression is reduced or absent in a subset of primary osteosarcoma samples, srGAP2 and other axon guidance proteins likely play a role in osteosarcoma metastasis. 74 The intracellular signal transduction pathway that includes Abelson kinase, Enabled protein, GAPs, and the Rho family of small GTPases may also play a role in endothelial cells. In addition, PI3K is important for endothelial cell responses to SLIT2, but the mechanism remains unknown.Apart from this, it was shown that cellular protrusive activity was inhibited via a SLIT2-ROBO4paxillin-GIT1 network.The study led by Zhang and Zhou demonstrated that downregulation of ROBO1 using small interfering RNA inhibited mesenchymal stem cell (MSC) proliferation.In addition, four miRNAs (miR), including miR-218, miR-29a, miR-146, and miR-148, inhibited the protein expression of ROBO1 in the MSCs, with miR-29 having the most marked effect. ROBO1 was identified as a novel target of miR-29a with a luciferase reporter assay. Overexpression of miR-29a suppressed the protein expression levels of ROBO1 and SLIT2 and inhibited the viability and proliferation of the MSCs. By contrast, overexpression of ROBO1 partly rescued these inhibitory effects of miR-29a on the MSCs. These results indicated that the miR29a/ROBO1 axis was crucial for the regulation of MSC viability and proliferation, suggesting that miR29a may serve as a potential clinical target for MSC expansion and stem cell transplantation. In addition, miR-218 suppressed nasopharyngeal cancer progression through downregulating the SLIT2-ROBO1 pathway in a negative feedback loop manner. 56
## Downstream and crosstalk of the slit/robo
## The role of slit/robo in the tumor microenvironment
The tumor microenvironment is composed of tumor cells, stromal cells, and other cellular components as well as extracellular matrix and various molecules that mediate intercellular interactions. These components form a complex tumor microenvironment that prominently affects the occurrence, growth, invasion, metastasis, and drug resistance.Here we further investigate the role of SLIT/ROBO pathway in the tumor microenvironment.
## Slit/robo pathway in tumor angiogenesis
The SLIT/ROBO signal is critical for axon guidance and neuronal precursor cell migration in the nervous system. Evidence suggested that classical neuronal guidance cues also regulated vascular development.Expression of ROBO1 and ROBO4 has been observed in vascular endothelial cells. They may serve different functions in SLIT/ ROBO signaling due to differential expression in various phenotypes of endothelial cells, ROBO1 induced long and thin actin fibers, whereas ROBO4 induced short and thick actin bundles along with membrane ruffles. 79 SLIT3, which was also observed in endothelial cells and vascular smooth muscle cells, promoted angiogenesis but decreased neurogenesis.Here, we focus on SLIT/ROBO signaling in angiogenesis via SLIT/ ROBO1 and SLIT/ROBO4. The SLIT2 was expressed in a broad spectrum of tumor cell lines and interacted with human umbilical vein endothelial cells (HUVECs) and tumor-associated endothelial cells in the presence of ROBO1.
Tumor cells appeared to secrete the SLIT2 protein, which formed a gradient field for the attraction of endothelial cells through interaction with ROBO1 on endothelial cell surface. Thus, endothelial cells migrated toward tumor and formed new blood vessels.The SLIT/ROBO signaling inhibited tumor angiogenesis and growth in a model of chemically induced squamous cell carcinoma when blocked by R5, a monoclonal antibody against the first immunoglobulin domain of ROBO1.The observation that ROBO1 was highly expressed in the early postnatal days of C57BL/6J mice correlated with superficial vascular and deep vascular plexus formation, indicating that ROBO1 might participate in retinal neovascularization.Thus, SLIT/ROBO1 signaling may promote angiogenesis in tumors.
ROBO4 is a vascular-specific receptor and mainly expressed in active angiogenesis, especially tumor vessels.Previous studies showed that ROBO4 inhibited endothelial cell migration through interaction with SLIT2, suggesting that ROBO4 might negatively regulate new vessel formation. This point was supported by several studies focusing on HUVECs, microvascular endothelial cells,and an animal model of ocular angiogenesis.In addition, Acevedo et al. suggested that SLIT2 could both positively and negatively regulate angiogenesis by binding to ROBO1 and ROBO4, respectively, and activation of ROBO4 blocked vascular endothelial growth factor (VEGF)induced angiogenesis and vascular permeability.However, a study by suggested that ROBO4 expressed much in fibrovascular membranes (FVMs), and loss of ROBO4 disturbed tube formation.Knockdown or overexpression of ROBO4 impaired intersomitic vessels formation in zebrafish showed contradictory functions journals.sagepub.com/home/tam 9 of ROBO4 in angiogenesis.Furthermore, using siRNA knockdown of both ROBO1 and ROBO4 decreased endothelial motility and disrupted tube formation.Taking these studies together, it is suggested that SLIT/ROBO1 and SLIT/ROBO4 signaling might interact with each other and cooperate in tumor angiogenesis .
## Slit/robo signaling in inflammation
SLIT/ROBO signaling influences neuronal migration, axon guidance, and also functions in a similar way in directing inflammatory cells discrepantly. SLIT2 is involved in inflammation modulating in several ways. SLIT2/ROBO1 selectively impaired directional migration of neutrophils and T cells toward chemoattractant especially CXCL12/ CXCR4-induced chemotaxisand neutrophil recruitment,and also mediated inhibition of Langerhans cell migration resulting in suppression of contact hypersensitivity responses.However, during lung inflammation, SLIT2/ROBO1 led to enhancement of eotaxin-induced eosinophil chemotaxis, and exaggeration of allergic airway inflammation.Furthermore, through SLIT2/ROBO4 interaction, SLIT2 could regulate endothelial related inflammation by inhibiting vascular leak and stabilizing the vasculature by downregulating lipopolysaccharide or VEGF.In patients with acute kidney injury (AKI), higher levels of plasma ROBO4 suggested a link between endothelial dysregulation and onset of AKI.SLIT3 took a role in increasing the spontaneous and chemoattractant-induced migration of primary monocytes by inducing a chemokinetic effect.SLIT/ROBO particles have shown their potentials as markers or targets in treating inflammatory diseases such as periodontitis, AKI, and crescentic glomerulonephritis .However, another group showed that SLIT2 was expressed by cancerassociated fibroblasts (CAFs), increasing neurite outgrowth from dorsal root ganglia neurons as well as from Schwann cell migration or proliferation. Inhibition of SLIT2/ROBO signaling disrupted this stromal/neural connection.The controversial effects of SLIT/ROBO on tumor neural invasion reflect its complicated role in tumor progression and need to further be explored.
# Conclusion
There is a close relationship between abnormal signaling and tumor development. To develop effective cancer therapeutics, it is useful to understand how SLIT/ROBO signaling affects tumor formation and angiogenesis and the mechanisms by which SLIT/ROBO signaling exerts these effects. To develop an effective therapeutic approach for cancer treatment, additional studies are required, also taking into account the important regulation mediated by SLIT/ROBO pathway.
# Author contributions
Zhengdong Jiang and Gang Liang conceived the concept for this paper. All authors (Ying Xiao, Tao Qin, and Xin Chen) collected the available literature. Erxi Wu, Qingyong Ma, and Zheng Wang revised the manuscript critically. Zhengdong Jiang and Gang Liang contributed equally to this work.
# Funding
This work was supported by the National Natural Science Foundation of China (Grant Numbers 81672434 and 81472248).
## Conflict of interest statement
The authors declare that there is no conflict of interest.
## Orcid id
Qingyong Ma https://orcid.org/0000-0001 -8977-320X . Mechanisms of the SLIT/ROBO pathway in tumor inflammation process. SLIT2 is involved in inflammation modulated by ROBO1, ROBO4, and AKT pathway. The downstream pathways of ROBO1 and ROBO4 including PI3K and nuclear factor (NF)-kB pathways involve in chemotaxis of inflammatory cells (T cells, Langerhans cells, neutrophils, and eosinophils) and angio-associated components (endothelial cells, platelets). SLIT3 regulates CXCL12/CXCR4 and RhoA in modulating inflammation, chiefly monitoring inflammatory migration and chemotaxis.
## Wang j and
|
Association of Lipoprotein(a) With Atherosclerotic Plaque Progression
BACKGROUND Lipoprotein(a) [Lp(a)] is associated with increased risk of myocardial infarction, although the mechanism for this observation remains uncertain.OBJECTIVES This study aims to investigate whether Lp(a) is associated with adverse plaque progression.METHODS Lp(a) was measured in patients with advanced stable coronary artery disease undergoing coronary computed tomography angiography at baseline and 12 months to assess progression of total, calcific, noncalcific, and lowattenuation plaque (necrotic core) in particular. High Lp(a) was defined as Lp(a) $ 70 mg/dL. The relationship of Lp(a) with plaque progression was assessed using linear regression analysis, adjusting for body mass index, segment involvement score, and ASSIGN score (a Scottish cardiovascular risk score comprised of age, sex, smoking, blood pressure, total and high-density lipoprotein [HDL]-cholesterol, diabetes, rheumatoid arthritis, and deprivation index).RESULTSA total of 191 patients (65.9 AE 8.3 years of age; 152 [80%] male) were included in the analysis, with median Lp(a) values of 100 (range: 82 to 115) mg/dL and 10 (range: 5 to 24) mg/dL in the high and low Lp(a) groups, respectively. At baseline, there was no difference in coronary artery disease severity or plaque burden. Patients with high Lp(a)showed accelerated progression of low-attenuation plaque compared with low Lp(a) patients (26.2 AE 88.4 mm 3 vs À0.7 AE 50.1 mm 3 ; P ¼ 0.020). Multivariable linear regression analysis confirmed the relation between Lp(a) and lowattenuation plaque volume progression (b ¼ 10.5% increase for each 50 mg/dL Lp(a), 95% CI: 0.7%-20.3%). There was no difference in total, calcific, and noncalcific plaque volume progression.CONCLUSIONS Among patients with advanced stable coronary artery disease, Lp(a) is associated with accelerated progression of coronary low-attenuation plaque (necrotic core). This may explain the association between Lp(a) and the high residual risk of myocardial infarction, providing support for Lp(a) as a treatment target in atherosclerosis. (J Am Coll Cardiol 2022;79:223-233)
Structurally, Lp(a) is a low-density lipoprotein (LDL) to which apolipoprotein(a) is covalently bound; the latter carrying proinflammatory oxidized phospholipids (OxPLs) [bib_ref] Oxidized phospholipids, Lp(a) lipoprotein, and coronary artery disease, Tsimikas [/bib_ref]. The atherogenic mechanisms of this multifaceted particle are thought to include accumulation of the LDL component in atherosclerotic plaque, prothrombotic effects due to interference of the apolipoprotein(a) tail with plasminogen activation, as well as induction of a multilevel proinflammatory response mediated by OxPL-cargo. The prothrombotic and proinflammatory effects of Lp(a) have been suggested to promote plaque destabilization leading to plaque rupture and atherothrombotic events [bib_ref] Plasma lipoprotein(a) concentration predicts future coronary and cardiovascular events in patients with..., Nestel [/bib_ref]. Previous studies have shown an association between high serum levels of Lp(a) and high baseline atherosclerotic plaque volumes and the presence of adverse plaque features in patients with coronary artery disease [bib_ref] Lipoprotein (a) level is associated with plaque vulnerability in patients with coronary..., Muramatsu [/bib_ref]. However, data on the interaction between Lp(a) and progression of coronary plaque volumes and composition in contemporary patients are lacking.
Advanced quantitative plaque assessments on coronary computed tomography angiography (CCTA) now allow the assessment of both calcific and noncalcific atherosclerotic plaque types facilitating the tracking of disease progression as well as changes in coronary plaque morphology with good reproducibility [bib_ref] Reproducibility of quantitative plaque measurement in advanced coronary artery disease, Meah [/bib_ref]. Moreover, recent studies have shown a strong association between plaque composition determined by quantitative computed tomography (CT) analysis and clinical events [bib_ref] A boosted ensemble algorithm for determination of plaque stability in high-risk patients..., Al'aref [/bib_ref] [bib_ref] Coronary atherosclerotic precursors of acute coronary syndromes, Chang [/bib_ref]. In particular, the burden of low-attenuation plaque (attenuation density <30 HU), which serves as a marker of necrotic core, provides powerful prediction of future myocardial infarction outperforming clinical risk scores, severity of luminal stenosis, and CT calcium scoring [bib_ref] Coronary artery plaque characteristics associated with adverse outcomes in the SCOT-HEART study, Williams [/bib_ref].
In this study, we assessed whether high concentrations of serum Lp(a) are associated with progression of adverse plaque phenotype in a cohort of patients with advanced stable coronary artery disease who were already using guideline-directed preventative therapies. C T p l a q u e a n a l y s i s . Plaque measurements were performed using previously validated semiautomated software version 2.5 (AutoPlaque) [fig_ref] FIGURE 1: LowCoronary computed tomography angiography [/fig_ref] by a trained observer blinded to the patient's clinical status [bib_ref] Automated threedimensional quantification of noncalcified coronary plaque from coronary CT angiography: comparison..., Dey [/bib_ref]. We have previously shown excellent reproducibility for these measurements [bib_ref] Reproducibility of quantitative plaque measurement in advanced coronary artery disease, Meah [/bib_ref].
# Methods
Coronary artery centerlines were extracted in a semiautomated fashion for each major artery and any tributary of >2-mm diameter with visually observed disease. A region of interest was placed in the aorta to define blood pool attenuation. Coronary artery segments were defined manually according to Society of Cardiovascular Computed Tomography guidance.
All suitable vessel segments were manually identified and vessel wall and plaque constituents were automatically determined using scan-specific thresholds with manual adjustments made as required. S t a t i s t i c a l a n a l y s i s . [fig_ref] TABLE 1 Patient: Values are mean AE SD, median [/fig_ref]. Respective plaque burdens were also similar between high-and low-Lp(a) groups [fig_ref] TABLE 1 Patient: Values are mean AE SD, median [/fig_ref]..
## Accelerated progression of adverse plaque
Results were also similar regarding fibro-fatty Abbreviations as in [fig_ref] TABLE 2: Change in Coronary Plaque Volumes on Repeat CCTA in Patients With High... [/fig_ref].
## Funding support and author disclosures
[fig] FIGURE 1: LowCoronary computed tomography angiography (CCTA) of a patient with low serum lipoprotein(a) [Lp(a)] concentration (9.2 mg/dL) showing atherosclerotic plaque in the left circumflex artery (A and C) with evidence of mixed plaque on automated plaque assessment (red overlay). Low-attenuation plaque is visualized in bright orange on the 3-dimensional reconstruction (B and D) and does not appear to progress on serial scanning (baseline volume 18.6 mm 3 , 1-year volume 17.8 mm 3 ). severity by visual estimation. A segment involvement score was calculated as the total number of coronary artery segments exhibiting plaque, irrespective of the degree of luminal stenosis within each segment (minimum ¼ 0; maximum ¼ 16) (16). [/fig]
[fig] FIGURE 2: Low2 patients with high serum Lp(a) concentrations (82.2 and 152 mg/dL, respectively). In patient 1, atherosclerotic plaque in the mid-right coronary artery at baseline (A, noncalcific highlighted with red overlay) and after 1 year (C). Low-attenuation plaque is visualized in bright orange on the 3-dimensional reconstructions (B and D) showing progression from a volume of 81 mm 3 to 133 mm 3 1 year later. Similar representative images are seen in patient 2 with mixed atherosclerotic plaque in the mid-right coronary artery at baseline (E) and 1 year (G). Low-attenuation plaque progressed on serial scanning from a volume of 14.4 mm 3 (F) to 30.8 mm 3 after Coronary artery calcium scores on noncontrast CT were comparable between the high-and low-Lp(a) groups: 378 (IQR: 137 to 650) AU versus 371 (IQR: 101 to 894) AU (P ¼ 0.902). On CCTA, there were no baseline differences between high-and low-Lp(a) groups in terms of total, calcific, noncalcific, and low-attenuation plaque volumes [/fig]
[fig] DISCUSSION: For the first time, we have used repeat CCTA to investigate the association between serum Lp(a) concentrations and progression of coronary plaque volume and phenotype in patients with advanced multivessel coronary atherosclerosis. Patients with high serum Lp(a) concentrations ($70 mg/dL) showed [/fig]
[fig] FIGURE 3: Effect depicted as betas with 95% CIs for the percentage change in plaque volume from baseline to follow-up CCTA, standardized for each 50 mg/dL increase in Lp(a). Lp(a) was associated with low-attenuation plaque progression in univariable (b ¼ 10.2%, P ¼ 0.031) and multivariable (b ¼ 9.6%, P ¼ 0.048) analyses, and with fibro-fatty plaque progression in univariable analysis (b ¼ 6.7%, P ¼ 0.034), showing a trend in multivariable analysis (b ¼ 6.0%, P ¼ 0.062). Abbreviations as in [/fig]
[table] TABLE 1 Patient: Values are mean AE SD, median (IQR), or n (%). Bold P values <0.05 show statistically significant differences. with coronary stents were excluded from the analysis.Coronary atherosclerotic plaque volumes were measured for total plaque, calcific and noncalcific plaque, fibro-fatty plaque, and low-attenuation plaque as a marker of necrotic core. After delineating the lumen, the software automatically adjusted the cutoff points for different plaque types based on attenuation measured in the thoracic aorta. Plaque volumes (measured in mm 3 ) for each plaque type were measured across all coronary segments and summed to generate the total plaque volume on a per-patient level. Baseline plaque burdens were calculated by dividing plaque volumes by the coronary vessel volume and multiplying by 100. Plaque progression was defined as the difference in plaque volumes between the baseline and the follow-up CCTA scan. We used plaque volume for this progression analysis as this is a less-derived measurement than plaque burden; therefore, it is more suitable for assessing changes in plaque status in individual patients over time. [/table]
[table] TABLE 2: Change in Coronary Plaque Volumes on Repeat CCTA in Patients With High and [/table]
|
Femur Fracture in a Premature Infant: An Unusual Association of Sickle Cell Disease with Osteogenesis Imperfecta
# Background
Osteogenesis imperfecta (OI) is a genetic disorder with either recessive or dominant inheritance. It is caused by mutations in genes that encode type 1 collagen. These genes include COL1A1, COL1A2, CRTAP, and P3H, with COLA1A and COLA12 being the most common. Type 1 collagen is known to synthesize bone, skin, and other connective tissues, and when collagen synthesis is defective, it can lead to brittle bones and various abnormalities [bib_ref] Osteogenesis imperfecta: A pediatric orthopedic perspective, Franzone [/bib_ref] [bib_ref] Current approach to diagnosis and treatment of children with osteogenesis imperfect, Burtsev [/bib_ref]. The incidence of OI is 6 to 7 per 100 000, and at least 8 different subtypes of the disease are known. Types 1 and 4 have a mild-moderate phenotype. Types 2 and 3 are the most severe forms [bib_ref] A newborn with multiple fractures in osteogenesis imperfecta: A case report, Bayram [/bib_ref] [bib_ref] Osteogenesis imperfecta: Diagnosis and treatment, Biggin [/bib_ref].
The most common presentation of OI is spontaneous fracture in a patient with no history of trauma [bib_ref] Current approach to diagnosis and treatment of children with osteogenesis imperfect, Burtsev [/bib_ref]. Other clinical features include blue sclera, sensorineural hearing loss, dentinogenesis imperfecta, and short stature. Severe OI also can cause neurological, respiratory, and cardiac complications [bib_ref] Cardiopulmonary status in adults with osteogenesis imperfecta: Intrinsic lung disease may contribute..., Khan [/bib_ref] [bib_ref] Association between ribs shape and pulmonary function in patients with Osteogenesis Imperfecta, Sanchis-Gimeno [/bib_ref] [bib_ref] A multicenter study to evaluate pulmonary function in osteogenesis imperfecta, Tam [/bib_ref].
Sickle cell disease (SCD) is an autosomal-recessive genetic disorder. The substitution of valine for glutamic acid in the sixth position of the beta globin chain leads to the formation of sickle hemoglobin and results in a major hemolytic anemia. There are several different phenotypes of SCD. The most common and more severe are hemoglobin SS disease (SS), hemoglobin SC disease (SC), and hemoglobin SB 0 (Beta Zero) thalassemia (S beta-thalassemia). The less common and less severe phenotypes are hemoglobin SD, hemoglobin SE, and hemoglobin SO.
Increased blood viscosity and sickling of red blood cells lead to complications such as painful vaso-occlusive crisis, dactylitis, acute chest syndrome, avascular necrosis, priapism, and strokes. Splenic dysfunction contributes to impaired immunity, placing patients at risk of developing overwhelming sepsis. Patients with SCD are placed on prophylactic penicillin for the first 5 years of life to prevent infections. Research is ongoing on ways to reduce morbidity and mortality. The discovery of new medications to address pain and other complications has improved the quality of life for many patients with SCD. Stem cell transplants and gene therapy offer hope for cure. Many patients with SCD, however, face a lifetime of pain and the psychological toll of chronic illness [bib_ref] Standard management of sickle cell disease complications, Abboud [/bib_ref] [bib_ref] Vitamin D supplementation for sickle cell disease, Soe [/bib_ref].
We present the case of an infant with dual diagnoses of OI and SCD. A review of the literature shows 1 case report about a patient diagnosed with both conditions [bib_ref] Sickle cell disease with osteogenesis imperfecta, Patil [/bib_ref]. The case presented here was reviewed and acknowledged by the NewYork Presbyterian-Brooklyn Methodist Hospital Institutional Review Board.
## Case report
The patient was born at 26 6/7 weeks gestational age via normal spontaneous vaginal delivery to a 40-year-old G3P1011 mother who had a history of sickle trait. The infant was admitted to the Neonatal Intensive Care Unit (NICU) for prematurity. Her NICU course was complicated by respiratory distress syndrome, chronic lung disease, suspected sepsis with metabolic acidosis, anemia, suspected necrotizing enterocolitis, and cow's milk protein allergy. Newborn screening done on Day 10 of life revealed a diagnosis of SCD. A repeat hemoglobin electrophoresis at 5 months of age confirmed SCD-SS type.
At 83 days of life, the infant was noted to be irritable. Examination revealed swelling of the left thigh with tenderness and minimal active movement. The affected limb was held abducted and flexed. An X-ray of the extremities revealed a non-displaced fracture of the left femoral shaft with periosteal reaction [fig_ref] Figure 1: Radiograph of the fracture [/fig_ref]. An orthopedist was consulted and a splint was placed. Imaging of the limb 10 days later showed a healing fracture [fig_ref] Figure 2: Follow-up radiograph of the fracture 10 days later [/fig_ref]. Upon further work-up, calcium, phosphorus, and vitamin D levels were all normal. Four days before, an alkaline phosphatase level was 599 units/L and its peak was 699 units/L at 2 weeks of life. Calcium and phosphorus levels were normal throughout the infant's NICU stay. An endocrinologist was consulted and recommended adding 400 units of vitamin D supplementation to the regimen. Further examination of the infant revealed bluish sclera, which was confirmed by an ophthalmologist. The patient then was evaluated for OI. Genetic testing confirmed the diagnosis of OI type 1.
At 18 months of age, the patient was in the 13th percentile for length and below the third percentile for weight. She has not had any further fractures. She was hospitalized for 2 sickle cell pain crises, at 1 year and 16 months of age. Both times, she presented with irritability and inability to bear weight on her right lower extremity. Imaging performed during the admissions did not reveal new fractures.
# Discussion
This is a rare example of 2 unrelated major genetic disorders coexisting in the same patient. Currently there is only 1 such case reported in the literature [bib_ref] Sickle cell disease with osteogenesis imperfecta, Patil [/bib_ref]. A number of factors potentially influenced our patient's bone health, based on the diagnoses and treatment during her NICU stay. One previous study looked at how caffeine given to patients in the NICU is associated with osteopenia of prematurity. During pregnancy, about 80% of bone mineralization occurs during the third trimester. Therefore, premature infants are born with decreased bone mineral content compared with full-term infants as well as decreased bone mineral density (BMD). Prematurity was already a risk factor predisposing our patient to fractures. In another study, 335 infants with a gestational age of <31 weeks and birth weight of <1500 g were assessed who were given caffeine during NICU stays to treat apnea of prematurity. Chest X-rays showed that 51% of the infants had osteopenia of prematurity and 8% had fractures [bib_ref] Caffeine is a risk factor for osteopenia of prematurity in preterm infants:..., Ali [/bib_ref]. Our patient was treated with caffeine from birth until Day 62 of life, which may have increased her risk of femoral fracture on Day 83 of life.
Our patient was diagnosed with OI type 1, which is a milder form of the disease. Treatment consists of prevention of fractures by maintaining bone strength and BMD. Low 25(OH) vitamin D and BMD have been described in children with either OI or SCD. Multiple hospitalizations and impaired mobility due to pain also can contribute to decreased BMD, thus possibly increasing the rate of fractures [bib_ref] Vitamin D supplementation for sickle cell disease, Soe [/bib_ref].
Patients with SCD are prone to chronic bone complications, such as vaso-occlusive crises, avascular necrosis, and osteomyelitis. Vitamin D is important for maintaining bone mineralization and patients with SCD often become vitamin D-deficient. Two previous studies looked at the effect of vitamin D supplementation on patients with SCD. The first study compared a group of patients with SCD who received vitamin D3 supplementation with a group that received placebo for 6 weeks. At the end of the study, the group that received vitamin D supplementation had significantly higher serum vitamin D levels and fewer days during which they experienced pain compared with the placebo group [bib_ref] Vitamin D supplementation for sickle cell disease, Soe [/bib_ref] [bib_ref] Serum 25-hydroxyvitamin D and diet mediates vaso-occlusive related hospitalizations in sickle-cell disease..., Mccaskill [/bib_ref]. A similar study was conducted in patients with SCD to investigate the effects of vitamin D supplementation in the treatment and placebo groups. The participants in both groups were treated for 6 weeks and then were followed for another 6 months. The result was fewer pain days and better physical function in the treatment group as vitamin D levels rose compared with the placebo group [bib_ref] High dose vitamin D therapy for chronic pain in children and adolescents..., Osunkwo [/bib_ref]. Given the history of SCD, it appears that it would be beneficial to prescribe vitamin D supplementation for patients with it to help them maintain bone health and possibly decrease their risk of fractures.
A shared feature in both SCD and OI is short stature. In patients with SCD, vasculopathy, transfusional iron overload, and end-organ failure can lead to endocrinopathies, such as thyroid dysfunction, poor growth, and pubertal delay. In a study of 52 patients with SCD, height, weight, vitamin D level, maturity, and growth hormone (GH) levels were monitored over 6 months. Of the patients, 48 had at least 1 endocrine variation. Vitamin D deficiency was seen in 84% and 3% had GH deficiency. Specifically, patients with the HbSS genotype had lower levels of insulin-like growth factor (IGF)-1 than those with the HbSC genotype of SCD [bib_ref] Endocrine and metabolic complications in children and adolescents with sickle cell disease:..., Mandese [/bib_ref]. Short stature can be caused by nutritional deficiencies, the hypermetabolic state in SCD, and decreased synthesis of IGF-1 and GH. GH replacement therapy would be beneficial in patients diagnosed concurrently with concurrent SCD and OI who have short stature.
Fortunately, our patient is clinically well. She has not sustained any new fractures since discharge. Her physical examination is normal except for blue sclera. Her height is in the 13th percentile and her weight is below the third percentile. She is on folic acid, prophylactic penicillin, and vitamin D supplementation. She had her first admission for an SCD-related pain crisis at age 1 year and a pain crisis admission at 16 months. During both hospitalizations, no fractures were found and the patient improved with fluids and nonsteroidal anti-inflammatory drugs. She has been referred to a comprehensive center for OI at a tertiary care hospital. Graff et al. studied growth patterns in 117 patients aged 2 to 18 with OI type 1. Data from healthy children, including height, weight, and body mass index (BMI), were used as a reference. This was a retrospective study that analyzed height measurements, except in those with comorbidities, on bisphosphonates, or who had vertebral compression fractures or surgical intervention for prior bone trauma. The authors found that children aged 2 to 3 years with OI were slightly shorter than children without OI (SD -1.2). Between ages 4 to 7 years, children with OI catch up to their peers (SD -0.5). In adolescents and young adults, growth slows down (SD -2.7). Overall body weight and BMI were similar in children with and without OI [bib_ref] Developmental charts for children with osteogenesis imperfecta, type I (body height, body..., Graff [/bib_ref].
Our patient's height is at the 13 th percentile and her weight is below the third percentile. More research is needed on predicting height in patients with OI.
# Conclusions
We report a rare case of concomitant SCD and OI in a premature infant who sustained a fracture in the NICU. Premature infants are at a higher risk for decreased BMD. Other comorbidities, such as the use of caffeine in apnea of prematurity, add to this risk. Multiple therapeutic interventions are put in place in the NICU, such as monitoring calcium and phosphorus and supplementation with vitamin D, to optimize bone health and mitigate complications from decreased BMD. Despite that, fractures still occur.
This case report emphasizes the importance of a thorough physical examination and a high index of suspicion to determine the factors that contribute to fracture in a premature infant. The overlap of symptoms may present challenges to diagnosis and care. Aggressive management of bone health, vitamin D levels, and attention to nutrition and exercise will help. The possibility of cure for both disorders with stem cell transplant or gene therapy offers hope and impetus for future research and clinical trials.
[fig] Figure 1: Radiograph of the fracture. [/fig]
[fig] Figure 2: Follow-up radiograph of the fracture 10 days later. [/fig]
|
Atypical Tibial Fracture in a 63-Year-Old Woman With Intermittent Use of Bisphosphonate Unmasking Hypophosphatasia
ORCiD numbers: 0000-0003-4899-7228 (U. H. Malabu); 0000-0002-6175-985X (J. Lockett); 0000-0002-1737-9470 (E. Lyster).We report an unusual case of atypical proximal tibial stress fracture (APTF) associated with intermittent use of bisphosphonates (BPs) and persistently low serum alkaline phosphatase (ALP) levels. We describe the case of a 63-year-old white woman who had experienced an APTF after 4 years of intermittent exposure to alendronate given for recurrent metatarsal stress fractures. BP administration was stopped after the diagnosis of the APTF. A review of her previous serum ALP levels revealed they had been consistently low. Adult hypophosphatasia (HPP) was diagnosed by the low serum ALP activity and elevated urine phosphoethanolamine levels. She was treated conservatively with analgesics. Adult HPP is an underrecognized condition associated with atypical insufficiency fractures, and BP use compounds this risk. To the best of our knowledge, we report the first case of intermittent BP exposure preceding an APTF in an adult patient with HPP, highlighting the uncommon site of the proximal tibia for BP-associated atypical insufficiency fractures, the need to screen for HPP in those with persistently low ALP levels before they begin BP therapy, and the importance of avoiding BP use in those with HPP.
history revealed repeated left and right metatarsal stress fractures that had occurred from 2009 to 2013. Her relevant medication history showed usage of alendronate (70 mg weekly) that had begun in June 2014. However, she admitted to taking alendronate intermittently, often missing the medication for #11 months at a time [fig_ref] Table 1: Prescription Record Indicating Occasional Use of Alendronate From 2014 to Fracture in... [/fig_ref]. She had no history of oral glucocorticoid use, and she had no dental issues. Her other medical history was pertinent for laparoscopic resection of a right adrenal aldosterone-producing adenoma in 2011. Postoperatively, her antihypertensive medications were stopped. On examination, she was in good general condition with a body mass index of 36.1 kg/m 2 . Tenderness was present on palpation along the upper shaft of the right tibia. A conventional radiograph showed periosteal reaction with no evidence of cortical beaking or transverse (stress) fractures [fig_ref] Figure 1: Radiographs showing evolution of APTF in HPP [/fig_ref]. A technetium-99m bone scan showed increased uptake in the region of the right proximal tibial shaft [fig_ref] Figure 2: Technetium-99m bone scan showing [/fig_ref]. The findings from a bone biopsy were consistent with a healing fracture with no evidence of malignancy or infection. Her T-scores for bone mineral density on dual-energy x-ray absorptiometry were 21.19 at the lumbar spine and 21.87 at the neck of femur. The serum carboxy-terminal collagen crosslinks and N-terminal propeptide of type I procollagen levels were normal at 430 ng/L (normal range, ,800 ng/L) and 49 mg/L (normal range, 15 to 90 mg/L), respectively. Serum and urine protein electrophoresis did not detect a monoclonal immunoglobulin.
At the time of the fracture, her serum biochemical parameters of bone metabolism were normal: calcium, 2.38 mmol/L; phosphate, 1.32 mmol/L; ALP, 39 U/L (normal range, 30 to 115 U/ L); parathyroid hormone, 5.6 pmol/L (normal range, 1.6 to 6.9 pmol/L); 25-hydroxy vitamin D, 59 nmol/L (normal range, 50 to 150 nmol/L); and estimated glomerular filtration rate, 83 mL/ min. However, on reviewing her bone metabolic profile from 2009, all the parameters were unremarkable, except for her serum ALP levels, which were frequently low, ranging from 23 to 28 U/L (normal range, 30 to 115 U/L). The higher ALP level at her presentation was likely in response to her proximal tibial insufficiency fracture. The urine phosphoethanolamine level was elevated at 13 mmol/mol creatinine (normal range, ,5), consistent with a diagnosis of HPP. Her serum pyridoxal 5 0 -phosphate (vitamin B 6 ) was, however, within the normal range at 110 nmol/ L (normal range, 20 to 190 nmol/L). She was advised to stop taking the alendronate and was treated conservatively with simple analgesics, calcium carbonate, vitamin D, and nonweightbearing with progressive, successful healing of the fracture. With resolution, her ALP and carboxy-terminal collagen crosslinks levels had gradual decreased (31 U/L and 370 ng/L, respectively). A repeat serum pyridoxal 5 0 -phosphate test when she was not taking vitamin B 6 supplements showed that the level had returned to the upper limit of normal (190 nmol/L).
# Discussion
To the best of our knowledge, we report the first case of APTF in a patient with adult HPP who had been using only intermittent BP. Our report is unique regarding the location of the fracture and that, unlike previously reported cases of AFFs in patients with HPP and longterm regular exposure to BP [bib_ref] Atypical femoral fractures" during bisphosphonate exposure in adult hypophosphatasia, Sutton [/bib_ref] , our patient was taking the BP only occasionally. In our patient, the serum ALP levels had been previously low, although not recognized by the treating physicians. The primary clinical utility of ALP is that elevated levels are used to diagnose bone diseases associated with high bone turnover, such as osteomalacia and Paget disease. Little attention has been given to low ALP values [bib_ref] Clinical spectrum of hypophosphatasia diagnosed in adults, Berkseth [/bib_ref] [bib_ref] Absence of recognition of low alkaline phosphatase level in a tertiary care..., Maman [/bib_ref]. The clinical features of adult HPP include a propensity to fracture, the early loss of deciduous teeth, and the presence of musculoskeletal pain [bib_ref] Clinical spectrum of hypophosphatasia diagnosed in adults, Berkseth [/bib_ref]. A spectrum of disease severity in HPP exists that is inversely related to the age of symptom onset (categorized into five forms according to symptom onset), ranging from in utero fetal death with perinatal HPP to adult HPP. Odontohypophosphatasia is the mildest variant (isolated early shedding of deciduous teeth) [bib_ref] Hypophosphatasia-aetiology, nosology, pathogenesis, diagnosis and treatment, Whyte [/bib_ref]. The disease is caused by loss-of-function mutations in the tissue nonspecific ALP gene. More than 300 mutations have been identified, with both autosomal dominant and autosomal recessive inheritance patterns [bib_ref] Hypophosphatasia-aetiology, nosology, pathogenesis, diagnosis and treatment, Whyte [/bib_ref]. The low ALP activity results in extracellular accumulation of inorganic pyrophosphate, a potent inhibitor of hydroxyapatite formation and, thus, bone mineralization. BPs are structurally similar to inorganic pyrophosphate, with earlier nonamino agents (e.g., etidronate) shown to inhibit bone mineralization through a similar mechanism with toxicity presenting similar to that of HPP [bib_ref] Atypical femoral fractures" during bisphosphonate exposure in adult hypophosphatasia, Sutton [/bib_ref] [bib_ref] Development of bisphosphonates, Fleisch [/bib_ref] [bib_ref] Hypophosphatasia: enzyme replacement therapy brings new opportunities and new challenges, Whyte [/bib_ref]. Newer agents are thought to inhibit bone remodeling through alternative mechanisms [bib_ref] Development of bisphosphonates, Fleisch [/bib_ref]. AFFs were first recognized as a rare complication of BP treatment in 2005, and the American Society of Bone and Mineral Research (ASBMR) released a position statement on AFF in 2010 that was revised in 2014 [bib_ref] Atypical subtrochanteric and diaphyseal femoral fractures: second report of a task force..., Shane [/bib_ref]. The pathogenesis has not been fully elucidated. However, these fractures share features with stress and insufficiency fractures in which accumulated, unrepaired microtrauma leads to fracture, suggesting AFFs might be related to an inability to remodel and heal the damage after prolonged usage of BPs [bib_ref] Atypical femur fractures: review of epidemiology, relationship to bisphosphonates, prevention, and clinical..., Black [/bib_ref]. The ASBMR taskforce on AFF reported a case definition that clearly delineates AFF from other minimaltrauma subtrochanteric fractures; it does not make reference to nonfemoral fractures [bib_ref] Atypical subtrochanteric and diaphyseal femoral fractures: second report of a task force..., Shane [/bib_ref]. Nevertheless, fractures with similar morphology at other sites, including the proximal tibia, have infrequently been reported with BP use and have been postulated to have the same pathophysiology [bib_ref] Atypical femur fractures: review of epidemiology, relationship to bisphosphonates, prevention, and clinical..., Black [/bib_ref] [bib_ref] Atypical insufficiency fracture of the tibia associated with long-term bisphosphonate therapy, Breglia [/bib_ref] [bib_ref] Antiresorptive-associated spontaneous fractures of both tibiae, followed by an atypical femur fracture..., Tan [/bib_ref]. Patients with severe HPP are known to sustain AFFs [bib_ref] Hypophosphatasia and the risk of atypical femur fractures: a case-control study, Bhattacharyya [/bib_ref]. Furthermore, case reports have described AFFs in patients with HPP taking BPs [fig_ref] Table 2: Previous Reported Cases of Atypical Fractures in Patients With Hypophosphatasia Taking BisphosphonatesAn... [/fig_ref] [bib_ref] Atypical femoral fractures" during bisphosphonate exposure in adult hypophosphatasia, Sutton [/bib_ref] [bib_ref] Teriparatide treatment in an adult patient with hypophosphatasia exposed to bisphosphonate and..., Righetti [/bib_ref] [bib_ref] Teriparatide treatment in adult hypophosphatasia in a patient exposed to bisphosphonate: a..., Doshi [/bib_ref] [bib_ref] Incidence of mutations in the ALPL, GGPS1, and CYP1A1 genes in patients..., Peris [/bib_ref]. A further mechanism of insufficiency fractures in patients with HPP taking BPs has only recently been proposed. BPs, which are analogs of calcium pyrophosphate, could inhibit the already reduced activity of ALP by binding to Zn 21 and Mg 21 ; both are needed for ALP activity [bib_ref] Atypical femoral fractures" during bisphosphonate exposure in adult hypophosphatasia, Sutton [/bib_ref]. The combination of defective bone mineralization and pharmacologically impaired remodeling in these patients likely precipitates these fractures. Therefore, BPs would be especially contraindicated for patients with HPP, demonstrating the importance of screening for HPP before prescribing BPs to minimize the risk of insufficiency fractures.
Adult HPP can be easily overlooked, because patients typically present with recurrent, poorly healing metatarsal stress fractures [bib_ref] Clinical spectrum of hypophosphatasia diagnosed in adults, Berkseth [/bib_ref]. Adult HPP should be suspected especially in patients with a diagnosis of osteoporosis and low serum ALP levels before they started taking BPs, after the exclusion of other causes of hypophosphatasemia [bib_ref] Hypophosphatasia-aetiology, nosology, pathogenesis, diagnosis and treatment, Whyte [/bib_ref]. Additional biochemical evidence to support a diagnosis of HPP includes elevated levels of ALP substrates (i.e., plasma pyridoxal 5 0 -phosphate, plasma or urine phosphoethanolamine, and plasma or urinary inorganic pyrophosphate) [bib_ref] Atypical femoral fractures" during bisphosphonate exposure in adult hypophosphatasia, Sutton [/bib_ref]. The single finding of a low serum ALP level alone is insufficient to diagnose HPP and, thus, requires confirmation by measuring the substrates of this enzyme and, where available, mutational analysis of the tissue nonspecific ALP gene [bib_ref] Hypophosphatasia-aetiology, nosology, pathogenesis, diagnosis and treatment, Whyte [/bib_ref]. In our patient, the serial serum ALP levels were consistently low, along with an elevated urinary level of phosphoethanolamine and high to normal pyridoxal 5 0 -phosphate. Our finding is in keeping with the report by McKiernan et al. [bib_ref] Mutational and biochemical findings in adults with persistent hypophosphatasemia, Mckiernan [/bib_ref] , which showed 61% of those with HPP with low serum ALP levels having elevation of only one ALP substrate, either phosphoethanolamine or pyridoxal 5 0 -phosphate, at diagnosis. A urinary inorganic pyrophosphate level would be helpful additional information. Although essential to confirm the diagnosis, genetic studies were not conducted for our patient. Neither of these tests are readily available in Australia. The findings from the biochemical studies and the history of recurrent metatarsal stress fractures in our patient were suggestive of a diagnosis of adult HPP [bib_ref] Clinical spectrum of hypophosphatasia diagnosed in adults, Berkseth [/bib_ref].
The incidence of mild vs moderate forms of HPP in the adult population is thought to be high, with an estimated prevalence of carriers of mild to moderate mutations of 1 in 6370 in the European population [bib_ref] Hypophosphatasia: enzyme replacement therapy brings new opportunities and new challenges, Whyte [/bib_ref]. The prevalence of asymptomatic HPP carriers could even be as great as 1 in 250 to 300 persons [bib_ref] Hypophosphatasia: enzyme replacement therapy brings new opportunities and new challenges, Whyte [/bib_ref]. These data indicate the need to consider an assessment for HPP for patients presenting with minimal trauma or stress fractures, especially when evaluating antiosteoporotic treatment with BP. Similar to our study, Sutton et al. [bib_ref] Atypical femoral fractures" during bisphosphonate exposure in adult hypophosphatasia, Sutton [/bib_ref] reported the case of a patient with HPP initially misdiagnosed with osteoporosis, who had developed a bilateral AFF after 4 years of treatment with BPs. Osteoporosis had also been initially misdiagnosed in our patient after recurrent bilateral metatarsal fractures and treatment with oral BPs. She had developed the fracture after 4 years of intermittent BP treatment. The ASMBR definition for atypical insufficiency fractures is limited to the subtrochanteric femur and diaphysis [bib_ref] Atypical subtrochanteric and diaphyseal femoral fractures: second report of a task force..., Shane [/bib_ref]. Except for the location in the proximal tibia, some of the ASBMR criteria were evident in our patient, including a noncommunuted tibial fracture after no trauma, prodromal pain for several weeks leading up to the fracture, and delayed fracture healing. Involvement of a nonfemoral site highlights the risk in other weight-bearing bones and calls for broadening the definition of "atypical fractures" [bib_ref] Atypical subtrochanteric and diaphyseal femoral fractures: second report of a task force..., Shane [/bib_ref] [bib_ref] Atypical femur fractures: review of epidemiology, relationship to bisphosphonates, prevention, and clinical..., Black [/bib_ref].
In conclusion, HPP is a relatively common, infrequently recognized, entity that is often misdiagnosed as osteoporosis. The diagnosis of HPP should be suspected in those with a low serum ALP level before BP treatment and should be excluded as a cause of minimal trauma fractures. BP treatment should be avoided, owing to the increased risk of atypical fractures (not only AFFs). The findings from the present case have highlighted these salient practice points. Further studies are needed to elucidate the pathophysiology of atypical fractures.
## Additional information
Correspondence: Usman H. Malabu, MD, FRACP, James Cook University College of Medicine and Dentistry, 1 James Cook Drive, Townsville, Queensland 4811, Australia. E-mail: usman.malabu@ jcu.edu.au.
Disclosure Summary: The authors have nothing to disclose. Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.
[fig] Figure 1: Radiographs showing evolution of APTF in HPP. (A) Radiograph at her initial presentation with pain showing only periosteal reaction along the proximal tibial shaft. (B) Radiograph 1 mo later showing a nondisplaced pathological fracture through the proximal tibial shaft and extending inferiorly through the metaphysis (arrows) with periosteal reaction in the lateral tibial cortex. (C) Radiograph 2 mo after presentation showing a more conspicuous fracture line, with evidence of early callus formation. (D) Radiograph 6 mo after fracture showing bony union with mature callus. [/fig]
[fig] Figure 2: Technetium-99m bone scan showing (A) prompt perfusion and (B) hyperemia in the region of the right proximal tibial shaft. [/fig]
[table] Table 1: Prescription Record Indicating Occasional Use of Alendronate From 2014 to Fracture in December 2018The patient had been without treatment for #30 mo. [/table]
[table] Table 2: Previous Reported Cases of Atypical Fractures in Patients With Hypophosphatasia Taking BisphosphonatesAn incomplete fracture was present before the patient had begun BP therapy, with progression occurring with medication use. [/table]
|
Ventilator-dependent pulmonary nocardiosis in a patient with chronic obstructive pulmonary disease
# Introduction
Nocardia is aerobic Gram-positive bacteria that can cause opportunistic infections through direct inhalation in the lungs, brain, and skin of immunocompromised individuals. [bib_ref] Pulmonary nocardiosis in the acquired immunodeficiency syndrome, computed tomographic findings: A case..., Sarcinelli-Luz [/bib_ref] [bib_ref] Pulmonary nocardiosis in a patient with chronic obstructive pulmonary disease -Case report..., Anderson [/bib_ref] Due to its inherent inoculation pathway, 90% of nocardiosis involves the pulmonary system. [bib_ref] Pulmonary nocardiosis in the acquired immunodeficiency syndrome, computed tomographic findings: A case..., Sarcinelli-Luz [/bib_ref] Pulmonary nocardiosis (PN) has been reported to occur mostly in the following populations: HIV patients, lymphoreticular cancer patients, transplant recipients, alcoholics, and diabetics. Disseminated nocardiosis has been reported to even mimic a metastatic malignancy in the immunosuppressed. [bib_ref] Clinical spectrum and outcome of pulmonary nocardiosis: 5-year experience, Singh [/bib_ref] However, 4 cases of PN in patients with chronic obstructive pulmonary disease (COPD) as the only risk factors have been reported, suggesting a close association for PN and COPD. [bib_ref] Disseminated nocardiosis masquerading as metastatic malignancy, Arjun [/bib_ref] Underlying pulmonary disease was the second most common predisposing condition for the development of PN in one cohort of patients. [bib_ref] Pulmonary nocardiosis in Chronic Obstructive Pulmonary Disease: A new clinical challenge, Castellana [/bib_ref] The potential mechanism behind this association is the recurrent or persistent bacterial infection in the respiratory system and long-term exposure symptoms were accompanied by intermittent high fever, excessive thirst, polyuria, diarrhea, and anorexia. Significant medical history included moderate degree of COPD managed with inhaled albuterol 2.5 mg PRN and tiotropium bromide 18 mcg. Her blood pressure, heart rate, respiration rate, temperature, and oxygen saturation at room air were 190/114 mmHg, 149/min, 40/min, 37.4°C, and 64%. Her white blood cell and lactic acid levels were elevated at 13.6 × 10 9 /L (70% granulocytes) and 21.3 mg/dL, respectively. Her sodium level was found to be low at 128 mmol/L. Arterial blood gas showed hypoxemic respiratory failure (pH 7.21, CO 2 43.0 mmHg, O 2 57.6 mmHg, HCO 3 18.8 mmol/L, and O 2 saturation 85.8%). She was expeditiously placed on 2 L of oxygen, but she continued to complain of worsening dyspnea and desaturated to 84%. In light of rapidly progressing hypoxemic respiratory failure, she was intubated and placed on mechanical ventilation. For acute management of her tachycardia and hypertension, metoprolol was given.
Chest X-ray revealed multilobar pneumonia [ [fig_ref] Figure 1: Roentgenogram of chest demonstrating multilobar pneumonia [/fig_ref] ]. Computed tomography (CT) of her chest revealed extensive consolidating pneumonia throughout the lung field bilaterally. This finding was most extensive in the lower lung fields, but there was considerable ground glass infiltrate in the right upper lobe and to a lesser extent, the left upper lobe [ [fig_ref] Figure 2: Computerized tomography of chest visualizing extensive consolidating pneumonia throughout the lung field... [/fig_ref] ]. She was started on vancomycin, cefepime, and methylprednisolone for empirical treatment. Her sputum culture grew Nocardia asteroides. The diagnosis of nocardiosis prompted the medical team to obtain a CT of her brain, which showed no signs of abscess or infection. Based on the result from the culture, antibiotic regimen was switched to piperacillin/tazobactam, sulfamethoxazole/trimethoprim, and azithromycin; however, the patient did not show any improvement on the combinative antibiotic therapy. A third antibiotic regimen with levofloxacin, ampicillin-sulbactam, meropenem, and sulfamethoxazole/trimethoprim was tried for a week without success. The fourth combinative antibiotic therapy consisted of vancomycin, imipenem, metronidazole, and sulfamethoxazole/trimethoprim, also with no success.
Due to lack of improvement, further investigation with bronchoscopy and video-assisted thoracoscopy (VAT) was performed. Bronchoscopy showed an inflamed airway with thin frothy secretions, mainly from the right hemithorax. Bronchoalveolar lavage showed nonspecific acute inflammation and blood. VAT of the right lung showed patchy infiltrates of all 3 lobes, the lower lobe most prominently. A modest amount of serous pleural effusion was also noted. Lung biopsy revealed acute fibrinous and organizing pneumonia [ [fig_ref] Figure 3: Lung biopsy [/fig_ref] ] consistent with the diagnosis of multifocal necrotizing pneumonia, and a large thickened vessel with an organizing fibrin thrombus [ [fig_ref] Figure 4: Lung biopsy [/fig_ref] ]. Overall, the lung was described as bobby and friable.
Following the investigation, the patient was treated with imipenem, metronidazole, and sulfamethoxazole/ trimethoprim. Three days later, amikacin was added. Over a 1 month period, the patient clinically improved. Repeat blood and sputum cultures were negative for Nocardia. Despite this, three attempted extubations were unsuccessful, and the patient remained ventilator-dependent. A tracheal tube was placed on her 31 st day on the ventilator to prevent vocal cord damage and tracheal stricture. The patient failed a swallowing test after tracheal tube placement and subsequently had a percutaneous endoscopic gastrostomy tube inserted for feeding purposes. During her admission, she developed a non-ST elevation myocardial infarction, possibly from demand ischemia. Coronary angiography was deferred due to the severe infection. The patient was started on 6 weeks of clopidogrel. Other cardiovascular conditions included DVT of her right arm resulting in a pulmonary embolism [ [fig_ref] Figure 5: Computerized tomography of chest showing pulmonary embolism [/fig_ref] ]. She was started on a heparin drip, and later bridged to warfarin. Another complication was persistent hyponatremia secondary to syndrome of inappropriate antidiuretic hormone secretion (SIADH). Throughout admission, her sodium level fluctuated from 120 to 131 mmol/L. She was treated with tolvaptan, furosemide, fluid restriction, and sodium chloride tablets. At the time of discharge, her sodium level was at 130 mmol/L. Two days before the patient's discharge, her WBC was still elevated at 17.2 × 10 9 /L, thus her antibiotic regimen was changed to linezolid, tobramycin, and sulfamethoxazole/ trimethoprim. On the 60 th day of hospitalization, she was discharged to a long-term acute care facility with a peripherally inserted central catheter for continuous antibiotic therapy. The patient expired 9 months following discharge due to failure to thrive.
# Discussion
Clinical presentation of PN is nonspecific and can resemble acute exacerbation of COPD. Common symptoms include chest pain, coughing, sputum production, dyspnea, fatigue, and loss of appetite. [bib_ref] Pulmonary nocardiosis in Chronic Obstructive Pulmonary Disease: A new clinical challenge, Castellana [/bib_ref] Common, nonspecific radiographic findings include infiltrates, cavitation, pleural effusion, or bony erosion. [bib_ref] Pulmonary nocardiosis in a patient with chronic obstructive pulmonary disease -Case report..., Anderson [/bib_ref] The diagnosis of PN can be challenging due to the rarity of the disease entity, radiographic pleomorphism, and lack of pathognomonic clinical signs or serologic markers. Only isolation of the organism through smears or culture is diagnostic. [bib_ref] Specific clinical manifestations of Nocardia: A case report and literature review, Yu [/bib_ref] Nocardia species grow on aerobic media over 5-21 days, and can exhibit acid-fast characteristics. Sensitivity of cultures from sputum and bronchoalveolar lavage are approximately 90% and 100%, respectively. [bib_ref] Pulmonary nocardiosis in a patient with chronic obstructive pulmonary disease -Case report..., Anderson [/bib_ref] This slow method greatly delays the diagnostic process. Improvements on current diagnostic methods are urgently needed.
Following a diagnosis of PN, it is critical to start treatment immediately because about 50% of untreated PN can spread to other parts of the body through the blood or lymphatic systems. [bib_ref] Specific clinical manifestations of Nocardia: A case report and literature review, Yu [/bib_ref] Treatment of PN consists of systematic antibiotic therapy. Sulfonamide-based antibiotics, such as trimethoprim-sulfamethoxazole, are the first line of treatment. For patients with sulfa allergies or nocardiosis refractory to sulfonamide-based regimen, the following alternative treatment options may be adopted: minocycline, doxycycline, amoxicilline-clavulanate, carbapenem, amikacin, cefuroxime, ceftriaxone, clarithromycin, ofloxacin, linezolid, and inhaled aminoglycoside. The recommended treatment duration is 3-12 months. [bib_ref] Pulmonary nocardiosis in a patient with chronic obstructive pulmonary disease -Case report..., Anderson [/bib_ref] [bib_ref] Pulmonary nocardiosis in a kidney transplant recipient: A case report and review..., Yaich [/bib_ref] SIADH is a known complication from PN, but the treatment of nocardiosis and fluid restriction typically resolves hyponatremia. [bib_ref] Pulmonary nocardiosis with a rare association of syndrome of inappropriate antidiuretic hormone..., Paritala [/bib_ref] Proper electrolyte balance is critical to improving patient outcomes.
The lesson to learn here is that clinicians should be aware of the association between PN and structural lung disease such as COPD, cystic fibrosis, and bronchiectasis. Broader awareness of this association would help reduce misdiagnosis and decrease the mortality rate from PN. Immunocompromised patients on long-term steroid therapy are also very vulnerable to PN.
## Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
[fig] Figure 1: Roentgenogram of chest demonstrating multilobar pneumonia [/fig]
[fig] Figure 2: Computerized tomography of chest visualizing extensive consolidating pneumonia throughout the lung field bilaterally [/fig]
[fig] Figure 3: Lung biopsy (H and E, ×10 power) showing a large blue sheet of predominantly neutrophilic active pneumonia, and organizing pneumonia consisting of amorphous pink fibrin bands in alveolar spaces [/fig]
[fig] Figure 4: Lung biopsy (H and E, ×20 power) showing a large thickened blood vessel with organizing fibrin thrombus [/fig]
[fig] Figure 5: Computerized tomography of chest showing pulmonary embolism [/fig]
|
Hazards and risks in oncology: radiation oncology
Adverse effects and hazards which have their origin from radiation using conventional techniques like 3-D conformal radiotherapy and total radi-Thomas G. Wendt 1 ation doses are well known. However little is known about the sprectum of especially late toxicity after radiation using new technologies like intensity modulated radiotherapy (IMRT) combined with novel target volume and dose concepts. Since IMRT allows for selective protection of the large salivary glands this technique improves the intermediate term quality of life and is the standard of care despite many details need further prospective evaluation. Combining cytotoxic drugs and radiotherapy yield improved survival in well-defined high risk patients. However morbidity and mortality of these protocols are high and deserve special expertise and supportive therapy. EGF-receptor antibodies have gained well defined indications, albeit specific toxicities in combination with irradiation deserve prospective studies and special attention.
# Introduction
Radiation oncology is a mainstay in treating extracranial head and neck cancers. Its impact both as a single modality but also in combination with surgery and/or systemic therapy has been elucidated in numerous studies of different levels of evidence. A longstanding tradition over ten decades has led to an accumulated knowledge upon therapeutic and adverse effects of radiotherapy performed in conventional 2 dimensional and recently 3 dimensional radiation technique. Initially conventional xrays and telecobalt gamma rays were used and with the advent of electron accelerators in the second half of last century with photons and electrons. These developments led a to a large body of knowledge upon the volumes to be treated and their total and single radiation doses to be given to obtain high local control rates and keep acute and chronic adverse effects at a reasonable level. During last three decades unconventional fractionation schedules have been studied in prospective randomized studies in order to optimize tumor control but also to clinical course of acute and chronic adverse effects at normal tissue inevitably exposed to ionizing radiation dose causing treatment associated acute and late morbidity. Shortly after computer tomography and magnetic resonance tomography revolutionized diagnostics of tumors both became the basis of individualized anatomy based radiation dose distribution planning. Another step forward was computerization including digital processing of linac accelerators used in medicine. New algorithms were established for individual dose calculation including Monte Carlo calculation which allowed much more sophisticated dose distribution in tumor and surrounding normal tissues and organs. This led to the clinical implementation of in-tensity modulated radiotherapy allowing different radiation doses within one target volume, e.g. higher doses to hypoxic subvolumes of a tumor (dose painting) [bib_ref] Towards multidimensional radiotherapy (MD-CRT): biological imaging and biological conformality, Ling [/bib_ref]. Positron emission tomography has recently been integrated not only in the staging process [bib_ref] Modification of staging and treatment of head and neck cancer by FDG-PET/CT..., Abramyuk [/bib_ref] but also in the radiation therapy planning process. Different radiopharmaceuticals e.g. deoxyglucose, fmisonidazol were used to detect tumor subvolumes with different radiobiologically relevant chracteristics e.g. proliferation or tumor hypoxia. These characteristics may be exploited in order to optimize an individual dose distribution plan and thus increase local tumor control probability (biologic radiation planning) [bib_ref] Positron emission tomography-guided, focal-dose escalation using intensitymodulated radiotherapy for head and neck..., Madani [/bib_ref]. However correlation of tumor size in excised specimen and metabolic seems loose . New dose distribution concepts and fractionation schedules including integrated boost harbour which has an at least in part unknown toxicity profile which may impact negatively on chronic morbidity particularly in long term surviving patients. Another new therapeutic avenue became clinically important with increasing knowledge upon interaction between cytotoxic drugs and more recently target drugs and ionizing radiation on a cellular basis (e.g. radiosensitization of cells). Much experience with regard to tumor control and adverse effects has been accumulated with simultaneous and sequential chemoradiation after extensive clinical use in numerous studies. However this is not true at the same degree for drugs used in target therapy.
## Intensity modulated radiotherapy (imrt)
Radiation induced xerostomia is a major source for chronic morbidity after primary or postoperative radiother-apy. Conventional 3D conformal radiotherapy does not allow sparing of the major salivary glands exept for few cases. Acute and chronic xerostomia causes a couple of adverse effects which impairs the quality of life in particular in long term survivors. Among others the most important are malnutrition and subsequent loss of weight, dental decay and chronic infection of the oral cavity. The technologically challenging intensity modulated radiotherapy allows for dedicated sparing of one or both parotid glands, which consequently results in lower probability of stimulated xerostomia provided radiation dose constraints, e.g. median organ dose of 26 Gy or less have been fulfilled [5], . IMRT is associated with higher global quality of life compared to conventional 3D conformal radiotherapy [7], [8] which is particularly observed with increasing length of survival. The retrospective pattern of care data of the SEER showed equivalent tumor control and suvival figures at 5-years obtained in 1,600 patients treated with both methods . For the individual not being treated in a radiation oncology department not experienced with IMRT eventually means an adverse factor with respect of his further quality of life [10]. However with increasing experience with IMRT it became clear that the departure from conventional radiation techniques may carry risks hitherto not known. With 3D conformal radiotherapy in primaries other than larynx and hypopharynx the lower neck nodes have been irradiated by a direct ventral portal and by shielding the spine the lower part of the paryngeal muscles have been shielded too. At the beginning of the IMRT era sparing of these muscles have not been recognized as important for preventing from dysphagia and aspiration and were initially not considered organs at risk. Not surprising that in a series of studies an increase of dysphagia after primary high dose chemoradiation using IMRT technique has been discribed . Recent concepts focus on sparing of pharyngeal musculature, esophageal sphinctger and supraglottic larynx. Total doses of less than 60 Gy should reduce the risk of dysphagia and aspiration [12], [13].
## Chemoradiation and radioimmunotherapy advantages and risks of chemoradiation
Conventionally used high total radiation doses (applied with 3D conformal technique), e.g. 70 Gy to advanced primaries and/or neck nodes result in local control rates of around 30%. Since these results cannot be improved by increased total radiation doses, cytotoxic drugs have been integrated in protocols giving chemotherapy simultaneously or sequentially to irradiation or more recently both. When 5-fluorouracil, mitomycin C and cis-platinum are given simultaneously to radiotherapy, not only the cytotoxic effect contributes to tumor cell kill but an radiosensitizing effect of drugs is exploited.
Simultaneous chemoradiation using 5-fluorouracil and cis-platinum or mitomycin C or carboplatinum augmented loco-regional control rates in locally advanced cancer often considered inoperable due to local extension compared to radical radiotherapy alone. Since not only 3-5 years tumor control rates but also overall survival has been increased simultaneous chemoradiation has become the standard of care [bib_ref] Hyperfractionated accelerated chemoradiation with concurrent fluorouracil-mitomycin is more effective than dose-escalated hyperfractionated..., Budach [/bib_ref] , [bib_ref] A meta-analysis of hyperfractionated and accelerated radiotherapy and combined chemotherapy and radiotherapy..., Budach [/bib_ref] , [bib_ref] Randomized trial of radiation therapy versus concomitant chemotherapy and radiation therapy for..., Calais [/bib_ref] , [bib_ref] Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): an update on..., Pignon [/bib_ref] , [bib_ref] Efficacy of intensified hyperfractionated and accelerated radiotherapy and concurrent chemotherapy with carboplatin..., Semrau [/bib_ref] , [bib_ref] Simultaneous radiochemotherapy versus radiotherapy alone in advanced head and neck cancer: a..., Wendt [/bib_ref]. Induction chemotherapy preceding definitive radiotherapy results in an increased overall treatment duration of typically 12 to 15 weeks (in which tumor surviving stem cells may proliferate) and higher cumulated drug doses. The two metaanalyses however showed only a smaller increase of tumor control rates compared to simultaneous chemoradiation . Noteworthy recent protocols including docetaxel have not been considered. When triplets consisting of taxane containing induction chemotherapy followed by radical radiotherapy and simultaneous taxane containing chemotherapy are given, increased median progression and overall survival span are reported compared to 5-fluorouracil based induction chemotherapy at the expense of more frequent hemato-logic°3/°4 toxicity [bib_ref] Gruppo di Studio Tumori della Testa e del Collo XRP 6976 F/2501..., Paccagnella [/bib_ref] , [bib_ref] TAX 324 Study Group. Cisplatin and fluorouracil alone or with docetaxel in..., Posner [/bib_ref] , [bib_ref] EORTC 24971/TAX 323 Study Group. Cisplatin, fluorouracil, and docetaxel in unresectable head..., Vermorken [/bib_ref]. In these studies increase progression free survival was essentially gained by better loco-regional tumor control. Although total drug amount delivered increased the risk of distant metastases has not been reduced [bib_ref] Gruppo di Studio Tumori della Testa e del Collo XRP 6976 F/2501..., Paccagnella [/bib_ref] , [bib_ref] TAX 324 Study Group. Cisplatin and fluorouracil alone or with docetaxel in..., Posner [/bib_ref] pointing to distant metastases being the achilles' heel of new concepts. In a randomized comparison of immediate chemoradiation compared to induction chemotherapy followed by chemoradiation the latter resulted in a marginal increase in progression free survival but no increase in overall survival at 3 years [bib_ref] Gruppo di Studio Tumori della Testa e del Collo XRP 6976 F/2501..., Paccagnella [/bib_ref]. Acute toxicity of induction chemotherapy can jeopardize chemoradiation [bib_ref] Neoadjuvant docetaxel, cisplatin, 5-fluorouracil before concurrent chemoradiotherapy in locally advanced squamous cell..., Ghi [/bib_ref]. Postoperative radiotherapy is a standard of care in patients with loco-regionally advanced tumors with distinct pathohistological risk factors. In two randomized studies of the RTOG and EORTC patients with risk factors R1status, extracapsular spread and more than 3 involved lymph nodes benefitted from the additional simultaneously applied chemotherapy, [bib_ref] Radiation Therapy Oncology Group 9501/Intergroup. Postoperative concurrent radiotherapy and chemotherapy for high-risk..., Cooper [/bib_ref]. However a subgroup analysis of the RTOG study showed increased survival only in patients with R1-status and extracapsular spread of involved lymph nodes [bib_ref] Long-term follow-up of the RTOG 9501/intergroup phase III trial: postoperative concurrent radiation..., Cooper [/bib_ref]. Since chemoradiation exposes patients with a treatment associated lethality of 1-4%, it is crucial to restrict chemotherapy to those who are likely to benefit. An organ sparing strategy is persued particularly in patients who would need a total laryngectomy for cure of their laryngeal or hypopharyngeal cancer. Radiotherapy can contribute to perserve organ and function. Historical data showed a concordant responsiveness of cancers to chemo-and radiotherapy. Therefore in actual protocols a short induction chemotherapy select those cancers suitable for further chemoradiation and sort out those which need early salvage surgery. Also for smaller tumors simultaneous, sequential chemoradiation and radiotherapy alone was studied in a RTOG protocol. Like in other primary tumor sites simultaneous chemoradiation was superior [bib_ref] Long-term results of RTOG 91-11: a comparison of three nonsurgical treatment strategies..., Forastiere [/bib_ref]. This protocol however carries a higher risk of late laryngeal toxicity compared to sequential therapy (15 vs. 10%).
## Therapy associated morbidity of simultaneous chemoradiation
In loco-regionally advanced cancers simultaneous chemoradiation results in a 6-8% gain in 5-year survival compared to irradiation alone at the expense of increased acute morbidity. In these protocols the clinically leading toxicity is (often reversible) hematologic. When the treatment is intensified beside hematologic toxicity other not dominating adverse effects like nausea and emesis, alopecia, dysphagia, stomatitis, neuropathy and others also increase in frequency and serverity. However they are less likely assessed. This "underreporting" leads to an unjustified favourable estimation of intensified protocols [bib_ref] TAME: development of a new method for summarising adverse events of cancer..., Trotti [/bib_ref]. Not only acute morbidity is augmented but also subacute adverse effects are reported to be more frequent 3 months after end of chemoradiation compared to radiation alone [bib_ref] Early morbidity after radiotherapy with or without chemotherapy in advanced head and..., Bieri [/bib_ref]. Beside acute morbidity late adverse effects, e.g. 5 years after end of chemoradiation gain increasing interest. However data are more scarce compared to those of acute morbidity, [bib_ref] Factors associated with severe late toxicity after concurrent chemoradiation for locally advanced..., Machtay [/bib_ref]. Among chronic symptoms aspiration and dysphagia, both vital risks, are increasingly investigated. Among all cytotoxic drugs cis-platinum, carboplatin, 5-fluorouracil and taxanes are the drugs most used in modern protocols. When radiotherapy and cis-platinum are combined, acute adverse effects increase. Given in postoperative protocols°3+ (WHO) adverse effects increase from 34% to 77% compared to irradiation alone. Hence only 60% of all patients receive all 3 scheduled courses [bib_ref] Radiation Therapy Oncology Group 9501/Intergroup. Postoperative concurrent radiotherapy and chemotherapy for high-risk..., Cooper [/bib_ref]. The toxicity profile change when cytotoxic drugs are administered concomitantly. Neuropathy and nephropathy are emerging risks which forces on intensified clinical monitoring but also intensified supportive care including optimized enteral feeding [bib_ref] A diseasespecific enteral nutrition formula improves nutritional status and functional performance in..., Fietkau [/bib_ref]. Cis-platinum is administered in anglo-saxonian protocols often on day 1 and 29 at a dosage of 100mg/kg body weight. In central europe the application of the same dose over 5 days (5 times 20 mg/kg body weight day 1-5, 29-33) has become standard in protocols. With the latter dosage the same tumor control rates are achieved with lower hematologic side effects [bib_ref] Comparison of four cisplatin-based radiochemotherapy regimens for nonmetastatic stage III/IV squamous cell..., Rades [/bib_ref]. In patients with preexisting diabetes mellitus more variation of their blood glucose profile have been observed when treated with cis-platinum for head and neck cancer. This may be due to an altered glucose metabolism. An intensive blood glucose monitoring is mandatory and existing medical antidiabetic therapy has to be adapted or initiated if no medication was applied before [bib_ref] Altered glucose metabolism during chemoradiation for head and neck cancer, Nguyen [/bib_ref]. A cisplatinum associated hyperglycemic koma has been described [bib_ref] Cisplatininduced hyperglycemic hyperosmolar coma, Komdeur [/bib_ref]. The large salivary glands are highly radiosensitive. Due to their proximity to the target volume (tumor or tumorbed and safety margins) it is often impossible to spare them from high radiation dose applied to the target volume. If one parotid gland is exposed to a median cumulated dose of 26 Gy or more irreversible subacute and chronic xerostomia must be anticipated . Even lower median doses have been advocated . If cis-platinum is given simultaneously to irradiation the tolerance of salivary glands and residual salivation further decreases [bib_ref] Effect of Cisplatin on parotid gland function in concomitant radiochemotherapy, Hey [/bib_ref]. If cisplatinum is used in a simultaneous chemoradiation protocol, a radiation technique should be applied which allows an optimal sparing of at least one parotid gland. In these cases the target median dose should probably be less than 26 Gy in order to reduce the risk of permanent xerostomia. This is best accomplished by sophisticated intensity modulated radiotherapy which should be regarded as a standard of care in chemoradiation protocols. Further improvement of parotid gland sparing is achieved by radiotherapy with protons [bib_ref] Comparison of intensitymodulated radiotherapy, adaptive radiotherapy, proton radiotherapy, and adaptive proton radiotherapy..., Simone [/bib_ref] , which however is not available for a larger population. Moreover chemoradiation protocols are associated with a treatment related mortality of 1-4% [bib_ref] Hyperfractionated accelerated chemoradiation with concurrent fluorouracil-mitomycin is more effective than dose-escalated hyperfractionated..., Budach [/bib_ref] , [bib_ref] Simultaneous radiochemotherapy versus radiotherapy alone in advanced head and neck cancer: a..., Wendt [/bib_ref] , [bib_ref] Long-term follow-up of the RTOG 9501/intergroup phase III trial: postoperative concurrent radiation..., Cooper [/bib_ref] , [bib_ref] Long-term results of RTOG 91-11: a comparison of three nonsurgical treatment strategies..., Forastiere [/bib_ref] , [bib_ref] Intensified hyperfractionated accelerated radiotherapy limits the additional benefit of simultaneous chemotherapy--results of..., Staar [/bib_ref]. Particularly in polymorbid patients these hazards need to be counterbalanced critically against the possible gain in survival achieved by addtional chemotherapy.
## New concepts by antibodies and tyrosinkinase inhibitors
Considerable improvements have been achieved by radical and postoperative chemoradiation protocols. Despite intensification of treatment no further gain can be expected due to concomitant morbidity of most patients, which limits their tolerance and compliance. A particular issue are distant metastases, which are experienced more often as loco-regional control increases. As a consequence new drugs with more specific tumoricidal acitivity and less hematological and mucosal toxicity are investigated. Among the vast body of molecular strategies up to now only a few gained clinical significance or are studied in phase III trials: 1. monoclonal antibodies against the extracellular domain of the epidermal growth factor receptor (EGFR inhibitors), 2. the inhibition of intra-and extracellular domains with tyrosinkinase inhibitors with low molecular weight ("small molecules") (EGFR tyrosinkinase inhibitors) and 3. antiangiogenetic drugs against vascular endothelial growth factor (VGFR inhibitors) blocking the neoangiogenesis in tumors [bib_ref] Targeted therapy in head and neck cancer, Kundu [/bib_ref]. Currently in Europe only Cetuximab for EGFR inhibition has been approved in combination with radiotherapy of squamous cell head and neck cancers.
## Cetuximab
More than 90% of all squamous cell carcinoma of the head and neck show an overexpression of the epidermal growth factor receptor, which is associated with lower survival [bib_ref] Impact of epidermal growth factor receptor expression on survival and pattern of..., Ang [/bib_ref]. Recent data show versatile acitivies of cetuximab at different cellular levels and clinical response does not closely correlate with the degree of cellular 3/7 GMS Current Topics in Otorhinolaryngology -Head and Neck Surgery 2013, overexpression of EGFR. The exact mode of action of cetuximab remains unclear. Despite that cetuximab serves as an alternative to cis-platinum in clinical chemoradiation protocols for head and neck cancers. In the TREMPLIN protocol [bib_ref] Sequential chemoradiotherapy (SCRT) for larynx preservation (LP): Results of the randomized phase..., Lefebvre [/bib_ref] , a larynx preservation protocol, cis-platinum was randomized against cetuximab. Equivalent 2-year survival data without laryngectomy were reported (cis-platinum: 79%, cetuximab: 71%, n.s.). However cetuximab was associated with less protocol violation and less late morbidity. Low acute toxicity is a charactistic of treatment with cetuximab. However the toxicity profile is different for that of cis-platinum. One typical side effect is the higher risk of radiation dermatitis°3-4 including skin necrosis, which is confined to the irradiated skin an differs from akneiform rash. Various studies observed radiation dermatitis up to 23% of patients [bib_ref] Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck, Bonner [/bib_ref] , [bib_ref] Severe cutaneous reaction during radiation therapy with concurrent cetuximab, Budach [/bib_ref] , [bib_ref] Severe radiation dermatitis in patients with locally advanced head and neck cancer..., Koutcher [/bib_ref]. For management of cetuximab associated radiation dermatitis recently multidisciplinary guidelines have been published [bib_ref] Interdisciplinary management of EGFRinhibitor-induced skin reactions: a German expert opinion, Potthoff [/bib_ref] , [bib_ref] Consensus guidelines for the management of radiation dermatitis and coexisting acne-like rash..., Bernier [/bib_ref]. When antibodies against EGFR cetuximab or panitumumab are infused acute bronchospasm is occasionally observed and therefore monitoring is mandatory. Case reports of bronchiolitis and lung fibrosis are published [bib_ref] Pulmonary toxicities from targeted therapies: a review, Barber [/bib_ref]. Cetuximab may rarely cause hypocalemia and hypomagnesemia. When misdiagnosed the latter may proceed and cause convulsions and arrhythmia [bib_ref] Hypomagnesaemia and targeted anti-epidermal growth factor receptor (EGFR) agents, Costa [/bib_ref].
## Cardiac toxicity of new drugs
A large hospital based survey admission diagnosis and prevalent comorbidity was analysed. Patients admitted with head and neck cancer had a higher incidence of cardivascular (41% vs. 27%) and obstructive lung (12% vs. 5%) disease compared to patients admitted due to a non-cancer diagnosis. During a 12 months follow-up patients treated with chemotherapy for their cancer experienced 1.7 fold chronic anemia and 2.6 fold a second cancer compared to the non-cancer patients. For cetuximab no cardiac or hematopoetic adverse effects are documented. Therefore cetuximab may be the preferred drug for in comorbid head and neck cancer patients.
## Hypothyroidism
Due to proximity of the thyroid gland and the target volume of radiotherapy of most tumors a large part of the gland inevitably will be exposed to high radiation doses. 6 months after exposure of the gland thyroid stimulation hormone will rise [bib_ref] A prospective analysis of subacute thyroid dysfunction after neck irradiation, Nishiyama [/bib_ref]. The incidence of a subclinical or clinical hypothyroidism will increase over years and has been observed even 20 years after radiotherapy. However with increasing time span, other etiologic factors may contribute [bib_ref] Radiotherapy-induced thyroid disorders, Jereczek-Fossa [/bib_ref]. Hemithyroidectomy and radiotherapy were predictive for hypothyroidism an a multivariate analysis [bib_ref] Hypothyroidism in patients treated with total laryngectomy. A multivariate study, Léon [/bib_ref]. Typically 10 years after radiotherapy a subclinical or clinical hypothyroidism is diagnosed in 20-25% in all irradiated head and neck cancer patients.
Advanced radiation techniques like the IMRT are able to spare at least parts of the thyroid gland. However long term experience upon the incidence of hypothyroidism after IMRT is lacking. Irrespective of radiation technique and total dose used all patients need a yearly follow-up for their functional status. Radiogenic thyroid diseases are treated like spontaneously ones.
## Notes
## Competing interests
The author received honorarium for oral presentations from Serono-Merck.
[table] 4 9: Burri RJ, Rangaswamy B, Kostakoglu L, Hoch B, Genden EM, Som PM, Kao J. Correlation of positron emission tomography standard uptake value and pathologic specimen size in cancer of the head and neck. Int J Radiat Oncol Biol Phys. 2008 Jul;71(3):682-8. DOI: 10.1016/j.ijrobp.2007.10.055 5. Amosson CM, Teh BS, Van TJ, Uy N, Huang E, Mai WY, Frolov A, Woo SY, Chiu JK, Carpenter LS, Lu HH, Grant WH 3rd, Butler EB. Dosimetric predictors of xerostomia for head-and-neck cancer patients treated with the smart (simultaneous modulated accelerated radiation therapy) boost technique. Int J Radiat Oncol Ten Haken RK, Kim HM, Marsh LH, Ship JA. Dose, volume, and function relationships in parotid salivary glands following conformal and intensity-modulated irradiation of head and neck cancer. Int J Radiat Oncol Biol Phys. 1999 Chen AM, Farwell DG, Luu Q, Vazquez EG, Lau DH, Purdy JA. Intensity-modulated radiotherapy is associated with improved global quality of life among long-term survivors of head-and-neck cancer. Int J Radiat Oncol Biol Phys. 2012 Sep;84(1):170-5. DOI: 10.1016/j.ijrobp.2011.11.026 8. Vergeer MR, Doornaert PA, Rietveld DH, Leemans CR, Slotman BJ, Langendijk JA. Intensity-modulated radiotherapy reduces radiation-induced morbidity and improves health-related quality of life: results of a nonrandomized prospective study using a standardized follow-up program. Int J Radiat Oncol Biol Phys. 2009 May;74(1):1-8. DOI: 10.1016/j.ijrobp.2008.07.059 Yu JB, Soulos PR, Sharma R, Makarov DV, Decker RH, Smith BD, Desai RA, Cramer LD, Gross CP. Patterns of care and outcomes associated with intensity-modulated radiation therapy versus conventional radiation therapy for older patients with head-andneck cancer. Int J Radiat Oncol Biol Phys. 2012 May;83(1):e101-7. DOI: 10.1016/j.ijrobp.2011.11.067 10. Jabbari S, Kim HM, Feng M, Lin A, Tsien C, Elshaikh M, Terrel JE, Murdoch-Kinch C, Eisbruch A. Matched case-control study of quality of life and xerostomia after intensity-modulated radiotherapy or standard radiotherapy for head-and-neck cancer: initial report. Int J Radiat Oncol Biol Phys. 2005 Nov;63(3):725-31. DOI: 10.1016/j.ijrobp.2005.02.045 11. van der Molen L, van Rossum MA, Burkhead LM, Smeele LE, Hilgers FJ. Functional outcomes and rehabilitation strategies in patients treated with chemoradiotherapy for advanced head and neck cancer: a systematic review. Eur Arch Otorhinolaryngol. 12. Jensen K, Lambertsen K, Grau C. Late swallowing dysfunction and dysphagia after radiotherapy for pharynx cancer: frequency, intensity and correlation with dose and volume parameters. Radiother Oncol. 2007 Oct;85(1):74-82. DOI: 10.1016/j.radonc.2007.06.004 13. Li B, Li D, Lau DH, Farwell DG, Luu Q, Rocke DM, Newman K, Courquin J, Purdy JA, Chen AM. Clinical-dosimetric analysis of measures of dysphagia including gastrostomy-tube dependence among head and neck cancer patients treated definitively by intensity-modulated radiotherapy with concurrent chemotherapy. Radiat Oncol. 2009;4:52. DOI: 10.1186/1748-717X-4-52 [/table]
|
Dissociable sources of erogeneity in social touch: Imagining and perceiving C-Tactile optimal touch in erogenous zones
Previous research points to two major hypotheses regarding the mechanisms by which touch can be experienced as erotogenic. The first concerns the body part to which touch is applied (erogenous zones) and the second the modality of touch (sensual touch optimal in activating C Tactile afferents). In this study, we explored for the first time the relation between those two mechanisms in actual and imagined social touch. In a first experiment, we randomly assigned "Giver" and "Receiver" roles within 19 romantic couples (20 females, 18 males, age 32.34 ± 8.71SD years) and asked the "Giver" to apply CT-optimal (3 cm/s) vs. CT-suboptimal (18 cm/s) touch on an erogenous (neck) vs. non-erogenous zone (forehead) of their partner. We then obtained ratings of pleasantness and sexual arousal from both "Receivers" and "Givers". In a second experiment, 32 healthy females (age 25.16 ± 5.91SD years) were asked to imagine CT-optimal vs. CT-suboptimal stimulation (stroking vs. patting) and velocity (3 cm/s vs. 18 cm/s) on different erogenous vs. non-erogenous zones and rate pleasantness. While both erogenous body part and CT-optimal, sensual touch were found to increase pleasant and erotic sensations, the results showed a lack of an interaction. Furthermore, pleasantness was induced by mere imagination of touch without any tactile stimulation, and touch that was sexually arousing for the receiver was rated as more sexually arousing for the giver as well, pointing to top-down, learned expectations of sensory pleasure and erogeneity. Taken together, these studies provide the first direct evidence that while both the body location to which touch is applied and the mode of touch contribute to pleasant and erotic sensations, these two factors appear to mediate subjective pleasantness and erogeneity by, at least partly, independent mechanisms.OPEN ACCESSCitation: Panagiotopoulou E, Filippetti ML, Gentsch A, Fotopoulou A (2018) Dissociable sources of erogeneity in social touch: Imagining and perceiving C-Tactile optimal touch in erogenous zones. PLoS ONE 13(8): e0203039. https://doi.org/ 10.
# Introduction
Social touch is a fundamental part of intimate human relationships, signalling and communicating a wide range of distinct emotions [bib_ref] Touch communicates distinct emotions, Hertenstein [/bib_ref]. Skin-to-skin touch in adults can also be PLOS ONE | https://doi.org/10.1371/journal.pone.0203039 August 24, 2018 1 / 13 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 erotogenic, that is, it can generate feelings of sexual arousal. However, there is disagreement regarding the neurophysiological and psychological mechanisms by which the erogeneity of touch is experienced and communicated. Specifically, there are currently two major hypotheses regarding such mechanisms, the first emphasising the (erogenous) body part to which touch is applied and the second stressing the (affective) modality of touch that is used. To our knowledge however, no study has addressed the relation between these two hypotheses. We first outline each of these hypotheses below, and then we present how this study aimed to address their relation. The concept of bodily 'erogenous zones' refers to certain body parts, such as the genitalia, that are relevant to sexual behaviours and hence tactile stimulation of these body parts leads to heightened, erotic responses. Yet, various body parts that are not near the genitalia can elicit strong erotic sensations when being touched. To explain this paradox, it was initially proposed that erotic sensations from these body parts arise as a consequence of their adjacency to the genitals in the somatotopic map of the primary somatosensory cortex (S1). Despite this intriguing possibility, a recent systematic investigation of erogenous zones, involving 41 different body areas, provided evidence against this hypothesis [bib_ref] Reports of intimate touch: Erogenous zones and somatosensory cortical organization, Turnbull [/bib_ref]. Specifically, there were no significant inter-correlations between ratings of sexual arousal for nearby S1 sites, suggesting that sexual arousal is unrelated to proximity of cortical representation of body parts in S1. Moreover, there were surprisingly low sexual arousal ratings for the feet, which are adjacent to the cortical mapping for genitals in S1. Therefore, it was suggested that the somatotopy of erogenous zones may be coded elsewhere in the brain, possibly in the insular cortex [bib_ref] Reports of intimate touch: Erogenous zones and somatosensory cortical organization, Turnbull [/bib_ref] , as this area is involved in the re-representation of interoception offering the basis for subjective feelings [bib_ref] Interoception: the inside story-a model for psychosomatic processes, Cameron [/bib_ref] [bib_ref] How do you feel? Interoception: the sense of the physiological condition of..., Craig [/bib_ref]. Nevertheless, there has been no direct exploration of this hypothesis as yet.
Interestingly, the insula is also implicated in a second, alternative hypothesis about the mechanisms of tactile erogeneity. This hypothesis links erogeneity to the functional role of a specific tactile modality. In fact, there is evidence for a specialised system coding for dynamic touch, involving slow-conducting, unmyelinated peripheral nerve fibers, the so-called CT (C Tactile) afferents that have been differentiated from the fast-conducting, myelinated peripheral nerve fibers (Ab fibers) typically studied in discriminatory touch [bib_ref] Unmyelinated tactile afferents signal touch and project to insular cortex, Olausson [/bib_ref] [bib_ref] Soma-totopic organization of gentle touch processing in the posterior insular cortex, Bjornsdotter [/bib_ref] [bib_ref] On the role of C fiber mediated touch in interoception, Bjornsdotter [/bib_ref]. This type of touch is referred to as "affective" or "sensual" touch. Studies using single unit microneurography, a neurophysiological method employed to record impulse conduction directly from peripheral nerve fibers of the skin, have shown that CT-afferents respond optimally to dynamic touch applied at very low indentation forces ranging from 0.3 to 2.5 mN [bib_ref] Unmyelinated afferents. constitute a second system coding tactile stimuli of the human..., Vallbo [/bib_ref]. Further, there is a distinct spatial pattern due to the fact that CT fibers are found only in hairy skin, while they are lacking in glabrous skin, suggesting that this type of touch is perceived differently across body regions. Moreover, they are temperature specific-showing a preference for a temperature of approximately 32˚C, which corresponds to interpersonal skin-to-skin contact [bib_ref] Human C-tactile afferents are tuned to the temperature of a skin-stroking caress, Ackerley [/bib_ref]. Crucially, the CT-afferents respond optimally to a specific range of stimulation velocities (1-10 cm/s) and less strongly at slower or faster velocities leading to an inverted U shape response pattern [bib_ref] Coding of pleasant touch by unmyelinated afferents in humans, Löken [/bib_ref]. Stimulating the skin at CT-optimal velocities is also accompanied by stronger feelings of subjective pleasure as compared to stimulating the skin in suboptimal velocities [bib_ref] Coding of pleasant touch by unmyelinated afferents in humans, Löken [/bib_ref]. This correlation between CT sensitivity and subjective ratings of pleasantness suggests that CT-afferents represent a distinct peripheral ascending pathway for pleasant tactile stimulation [bib_ref] Coding of pleasant touch by unmyelinated afferents in humans, Löken [/bib_ref]. However, CT-optimal touch is not merely associated with feelings of pleasantness, but it has also been found to be associated with feelings of erotic sensation [bib_ref] Unmyelinated tactile cutaneous nerves signal erotic sensations, Jönsson [/bib_ref] [bib_ref] Mediated Erotic Touch Perception Relates to Sexual Desire and Performance in a..., Bendas [/bib_ref]. Specifically, the same inverted U-shaped relationship found between stroking velocities and pleasantness ratings has also been found between stroking velocities and eroticism ratings, with CT-optimal velocities (1-10cm/s) leading to higher levels of perceived eroticism than slower or faster velocities [bib_ref] Unmyelinated tactile cutaneous nerves signal erotic sensations, Jönsson [/bib_ref] [bib_ref] Mediated Erotic Touch Perception Relates to Sexual Desire and Performance in a..., Bendas [/bib_ref].
Furthermore, research suggests that the discriminative and affective properties of touch are also processed differently in the brain. Functional magnetic resonance imaging (fMRI) analysis during CT stimulation has shown activation not only in the classic somatosensory areas S1 and S2 (as with discriminative touch) but also in the posterior insular cortex [bib_ref] Unmyelinated tactile afferents signal touch and project to insular cortex, Olausson [/bib_ref] [bib_ref] Soma-totopic organization of gentle touch processing in the posterior insular cortex, Bjornsdotter [/bib_ref] [bib_ref] On the role of C fiber mediated touch in interoception, Bjornsdotter [/bib_ref]. Furthermore, when applying soft stroking to neuropathy subjects lacking Ab afferents, only the posterior insular cortex is activated [bib_ref] Unmyelinated tactile afferents signal touch and project to insular cortex, Olausson [/bib_ref] [bib_ref] Functional role of unmyelinated tactile afferents in human hairy skin: sympathetic response..., Olausson [/bib_ref]. Neural responses in other brain regions have also been correlated with soft stroking, such as the superior temporal sulcus, the orbitofrontal cortex, the medial prefrontal cortex and the anterior cingulate cortex [bib_ref] Brain mechanisms for processing affective touch, Gordon [/bib_ref] [bib_ref] Touching and feeling: Differences in pleasant touch processing between glabrous and hairy..., Mcglone [/bib_ref] [bib_ref] The representation of pleasant touch in the brain and its relationship with..., Francis [/bib_ref]. Moreover, repetitive transcranial magnetic stimulation (rTMS) studies have shown that inhibition of right hemisphere S1 and S2 does not affect pleasantness perception associated with affective touch [bib_ref] Encoding of Touch Intensity But Not Pleasantness in Human Primary Somatosensory Cortex, Case [/bib_ref] [bib_ref] Inhibitory rTMS of secondary somatosensory cortex reduces intensity but not pleasantness of..., Case [/bib_ref]. Thus, these studies confirm the idea that affective touch is associated with activation of a multimodal neural network of brain areas mapping body state representation and emotion, including the insula, that bypasses the somatosensory areas S1 and S2 (but see [bib_ref] Primary somatosensory cortex discriminates affective significance in social touch, Gazzola [/bib_ref].
Despite the fact that both proposals regarding the origins of erogeneity, namely somatotopy and CT-afferents, stress the role of the insular cortex in the perception of erotic touch, they differ as regards to whether the body part or the modality of touch are the critical factors in the erogeneity of touch. Yet to our knowledge, no previous study has assessed the relationship between these two hypotheses when either receiving or, giving touch. The primary aim of this study was to assess this question by manipulating in the same design both the erogeneity of body parts and the modality of touch used to elicit pleasant and erotic sensations.
Moreover, a secondary aim of this study was to assess the relationship between body part and modality of touch in eliciting pleasant sensations when tactile stimulation was merely imagined, rather than actually perceived. In an intimate context, tactile interactions may be influenced not only by current sensations, but also by expectations, imagination and memories. Indeed, it has been shown that erotic mental imagination is strong enough to cause sexual arousal [bib_ref] Sexual fantasy, Leitenberg [/bib_ref]. Interestingly, there is evidence that the anterior insula responds both when CToptimal touch is experienced and imagined, without any sort of tactile or even visual stimulation [bib_ref] Dissociating the Neural Correlates of Experiencing and Imagining Affective Touch, Lucas [/bib_ref]. However, no study has tested whether imagining CT-optimal (i.e. sensual) touch in erogenous versus non-erogenous zones can also give rise to subjective feelings of pleasantness. Accordingly, a secondary aim of the present study was to assess the relationship between body part and modality of touch in eliciting pleasant sensations when tactile stimulation was merely imagined.
In sum, previous research points to two major hypotheses regarding the mechanisms by which touch can be experienced as erotogenic. The first concerns the body part to which touch is applied (erogenous zones) and the second the modality of touch (CT-optimal, sensual touch). In this study, we explored for the first time the relation between these two hypotheses both when touch was actually given and received but also when it was merely imagined. In a first experiment, we investigated actual (active and passive) stimulation of an erogenous vs. non-erogenous zone with CT-optimal versus non-optimal touch and measured both pleasantness and erotic perception of touch in romantic couples. We hypothesised that CT-optimal, sensual touch would be rated as more pleasant and more arousing than neutral touch, and that sensual touch on an erogenous zone would be more pleasant and arousing that sensual touch on a non-erogenous zone. In a second experiment, we investigated pleasantness perception of imagined sensual vs. neutral touch in erogenous vs. non-erogenous zones. We hypothesised that CT-optimal, sensual touch would be perceived as more pleasant than neutral touch, and that pleasantness would be higher when it is imagined being applied on an erogenous zone as compared to a non-erogenous zone.
## Experiment 1: pleasant and erotic perception of active and passive sensual touch on erogenous zones
Method Ethical declarations. The study was approved by the Ethics Committee of the Research Department of Clinical, Educational and Health Psychology, University College London.
Participants. Twenty-three romantic couples gave their informed consent to participate in the study. Participants were visitors of a public event on haptics at the Royal Institution in London. Three couples were excluded as they failed to follow the experimental instructions during administration and one couple was excluded due to an experimental error during administration. The final sample comprised of 19 couples, 20 females and 18 males, for a total of N = 38 participants (age 32.34 ± 8.71SD years). All subjects gave written, informed consent.
Design. The experiment employed a 2x2x2 design with 2 within-subject factors: 1) Body Part (neck vs. forehead), 2) Velocity (3 cm/s, i.e. CT-optimal vs. 18 cm/s, i.e. CT-suboptimal) and one between-subject factor, Role (Giver vs. Receiver).
Apparatus and material. The experimental paradigm (the velocity factor, in particular) was explicitly developed based on the large body of evidence revealed by microneurography (see Introduction) on CT-fiber properties underlying interpersonal touch [bib_ref] Coding of pleasant touch by unmyelinated afferents in humans, Löken [/bib_ref]. Before the experimental task, the experimenter assigned touch "Giver" and "Receiver" roles within the couple, following a standard randomisation procedure. These roles where then kept constant throughout the experiment. The experimenter trained the Giver to apply the different types of touch (CT-optimal and CT-suboptimal) using the index and middle fingers of the participant's dominant hand, with a velocity of either 3cm/s (slow/CT-optimal) or 18 cm/s (fast/CT-suboptimal). To help the Giver better understand the difference between slow and fast stroke velocities, the experimenter applied both touches on the Giver's forearm during the training phase. The body parts to which touch was applied were selected from a list of 41 body parts generated by a recent systematic review [bib_ref] Reports of intimate touch: Erogenous zones and somatosensory cortical organization, Turnbull [/bib_ref]. This list was divided into 3 almost-equal parts and we subsequently selected one body part from the top 14 (i.e. Nape of Neck) and one body part from the bottom 14 (i.e. Forehead) to represent erogenous and non-erogenous zones, respectively. The highest-rated body parts were the ones that are strongly associated with sexual content (i.e. genitals, breasts) and were not selected for ethical and practical considerations. Both participants were given a booklet, which contained the pleasantness and sexual arousal ratings, to be completed after each trial.
Procedure. Prior to the main experimental phase, participants were familiarized with procedures and all rating scales. The Receiver was asked to sit on a chair facing an empty wall to avoid any external distraction. The experimenter then identified the stroking areas on the nape of the neck and forehead of the Receiver, each measuring 9 cm long × 4 cm wide, and marked them with a washable marker on their skin. In each trial, the Giver was asked to deliver a "stroking-type" touch of one of the two trained velocities, either on the neck or forehead of their partner, as instructed by a checklist that the experimenter held. After each trial, both the Giver and the Receiver were asked to silently rate both the pleasantness and sexual arousal value of the touch in a continuous VAS scale (-10 to 10; not at all to extremely) using the booklet provided to each so that they could not see, or hear each other's ratings. That is, Givers were asked to rate how pleasant and sexually arousing it was to apply the touch, whereas Receivers were asked to rate how pleasant and sexually arousing was to feel the touch. Before starting the 12 trials task, the couple underwent 2 practice trials.
Data analysis. As data were not normal, we computed z-scores for both the pleasantness and sexual arousal ratings. We performed two separate analyses of variance (ANOVA) for each dependent variable (Pleasantness and Arousal ratings). In both analyses, the Body Part ("Neck" vs. "Forehead") and Velocity ("Slow" vs. "Fast") were entered as within-subject variables, whereas Role ("Giver" vs. "Receiver") and Gender were the between-subject factors. Greenhouse-Geisser correction was applied when sphericity could not be assumed (Mauchly's test for sphericity, p = 0.05). Comparisons were assessed for significance using planned twotailed t-tests. Level of significance was set to 0.05. Information on the relationship between pleasantness and sexual arousal is included in Supplementary Material (S1 File).
# Results
Pleasantness ratings. The ANOVA revealed significant main effects of Velocity, F(1,34) = 109.07, p < 0.001 and Body Part, F(1,34) = 15.10, p < 0.001, with CT optimal touch being rated as more pleasant than CT-suboptimal touch, t(37) = 10.44, p < 0.001 and touch on the neck being judged as more pleasant compared to touch on the forehead, t(37) = 3.05, p = 0.004. All other interactions were not significant (S1 File). The interaction Body Part x Role was significant, F[1,34] = 5.20, p = 0.029, with Receivers rating the touch on the neck as more pleasant than the touch on the forehead, t(18) = 3.05, p = 0.007, but no difference between these body parts was found in Givers, t(18) = 1.09, p = 0.288. We also found a significant interaction Body Part x Role x Gender, F[1,34] = 11.17, p = 0.002. To better understand this three-way interaction, we averaged ratings across Velocity and investigated the judged pleasantness among Givers and Receivers separately. We found that, among Givers, there was no significant difference between males and females in pleasantness ratings on the neck or the forehead, Body Part x Gender: F [1,17] = 0.087, p = 0.771. However, among Receivers we found a significant interaction between Body Part and Gender, F[1,17] = 7.61, p = 0.013, suggesting that females preferred significantly more to be touched on the neck compared to the forehead, t(8) = -3.77, p = 0.006 [fig_ref] Fig 1: Pleasantness ratings of givers and receivers [/fig_ref].
Sexual arousal ratings. Gender did not reveal any significant effects, and therefore data were averaged across this factor. The ANOVA revealed significant main effects of Velocity, F [1,36] = 107.60, p < 0.001 and Body Part, F[1,36] = 4.14, p < 0.001. Couples rated as more arousing to feel and give touch on the neck, compared to the forehead, t(37) = 5.81, p < 0.001, and to feel and give the CT optimal touch, compared to the CT non-optimal touch, t(37) = 10.42, p < 0.001. We also found an interaction trend between Body Part and Role, F[1,36)] = 4.14, p = 0.049. Planned independent samples t-tests however showed that no significant difference in arousal ratings between neck and forehead was present among the two groups (Givers, t(36) = 0.581, p = 0.565; Receivers, t(36) = -1.040, p = 0.306) [fig_ref] Fig 2: Arousal ratings of givers and receivers for the neck and forehead [/fig_ref]. All the other interactions were not significant (S1 File).
## Experiment 2: pleasant perception of imagined sensual touch on erogenous zones
Method Ethical declarations. The study was approved by the Ethics Committee of the Research Department of Clinical, Educational and Health Psychology, University College London.
Participants. Thirty-two right-handed healthy females (age 25.16 ± 5.91SD years) were recruited from UCL Psychology Subject Pool and took part in the study for course credit or £8/hour. Even though this experiment did not involve actual touch, a female experimenter was instructing participants to imagine different types of touch. Research has shown that the hedonic value of touch varies according to the gender of both receiver and giver [bib_ref] Primary somatosensory cortex discriminates affective significance in social touch, Gazzola [/bib_ref] , therefore, to control for gender effects related to touch perception and for practical considerations, only heterosexual females took part in the study. Participants had no known physical or mental illness and normal sense of touch. All subjects gave written, informed consent.
Design. We employed a 2x2x2 within-subjects design with 3 factors: 1) Type of Stimulation (stroking, i.e. CT-optimal vs. patting, i.e. CT-suboptimal, 2) Velocity (3 cm/s, i.e. CT-optimal vs. 18 cm/s, i.e. CT-suboptimal) and 3) Body Part (erogenous vs. non-erogenous zone). The dependent measure was reported pleasantness of imagined touch on a scale from -100 (not at all) to +100 (extremely). The experiment also included two manipulation checks. First, at the beginning and end of each trial we asked participants to report their subjective pleasantness and sexual arousal in response to video clips showing different types of touch. Here, we explored the perceptual effects of seen touch irrespective of body part, and manipulated "Type of Stimulation" and "Velocity". Second, we explored whether the sample perceived the chosen body parts as highly or minimally erogenous. Hence, we manipulated "Body part" (erogenous vs. non-erogenous) and measured the reported sexual arousal.
Apparatus and material. The experimental paradigm was explicitly developed based on the large body of evidence on CT-fiber properties underlying interpersonal touch [bib_ref] Coding of pleasant touch by unmyelinated afferents in humans, Löken [/bib_ref]. As with experiment 1, the body parts were selected from the same list of 41 body parts [bib_ref] Reports of intimate touch: Erogenous zones and somatosensory cortical organization, Turnbull [/bib_ref]. Three body parts were selected from the top 14 and three body parts from the bottom 14 to represent erogenous and non-erogenous zones, respectively. Therefore, we selected the erogenous zones "Nape of neck", "Lower back" and "Stomach", whereas the non-erogenous zones were "Forearm", "Forehead" and "Elbow".
To familiarise participants with the different types of touch, we presented four videos for the four types of touch (slow stroking, fast stroking, slow patting, fast patting) (S1 In all familiarisation videos, the stimulated area was a human forearm (6 cm length and 4 cm width), with light pressure (targeting at indentation forces of < 0.3 N). Each video clip lasted 4 sec. The number of pats was matched to the number of strokes for both the slow and fast condition, with 2 touches in the "slow touch" videos and 12 touches in the "fast touch" videos. The four conditions (slow stroking, fast stroking, slow patting, fast patting) were repeated three times each, for a total of 12 trials, the order of which was randomised.
Procedure. Computer-generated stimulation was controlled by a customized software program (Presentation software, Neurobehavioral Systems Inc.) and presented on the screen, which was placed at a viewing distance of approximately 80 cm. Each of the 12 trials began with one of the familiarisation videos in order to demonstrate the type of touch to imagine. After seeing the video, participants were asked to rate the pleasantness of the touch they watched ("How pleasant is this type of touch?") in a continuous VAS scale (-100 to 100; not at all to extremely). Right after rating each video, the main experimental task took place, whereby participants were asked to imagine the same type of touch they saw in the video being applied to various body parts ("Imagine how pleasant the touch would be on each of the following six body parts") and rate pleasantness in a continuous VAS scale (-100 to 100; not at all to extremely).
After the main experimental task, participants were presented again with each of the four familiarisation videos showing the different types of touch and were asked to rate the sexual arousal of the touch irrespective of body part ("How sexually arousing is this type of touch?"), in a continuous VAS scale (-100 to 100; not at all to extremely). Then, they were asked to rate the sexual arousal value of each of the six body parts irrespective of type of touch ("How sexually arousing is this body part?"), once again in a continuous VAS scale (-100 to 100; not at all to extremely). Data analysis. Statistical analyses were performed using SPSS v. 23 (IBM, Chicago, IL, USA). For the main experimental task (imagined touch), pleasantness ratings were averaged across the 3 trials of each condition. A repeated-measures ANOVA was conducted with Type of Stimulation, Velocity and Erogeneity of Body Part as within-subject factors. Pleasantness was measured and post hoc analyses were performed using Bonferroni correction. Level of significance was set to 0.05. Information on how data of our two manipulation checks were analysed is included in Supplementary Material (S1 File).
# Results
Based on the findings of the second manipulation check (details in S1 File), "Stomach" and "Forearm" were excluded from this analysis. Therefore, pleasantness ratings for imagined touch were averaged for the 2 high-rated and the 2 low-rated erogenous zones. The ANOVA revealed significant main effects of Type of Stimulation, F [bib_ref] Touch communicates distinct emotions, Hertenstein [/bib_ref] Post hoc analyses revealed that for both high and low erogenous zones, pleasantness ratings for slow stroking were significantly higher than fast stroking, slow patting and fast patting (all p-values < 0.001). For slow stroking and fast patting, pleasantness ratings for erogenous zones were significantly higher than for non-erogenous zones, [slow stroking: t(31) = 2.87, p = 0.007]; [fast patting: t(31) = 3.17, p = 0.003] [fig_ref] Fig 3: Pleasantness ratings for different types of imagined touch in erogenous vs [/fig_ref].
# Discussion
Interpersonal sensual touch between individuals can convey different meanings, depending on how it is performed and in which body part it is applied. At present, the neurophysiological and psychological mechanisms by which the erogeneity of touch is experienced and communicated are still a matter of debate. In fact, it is unclear whether it is the (erogenous) body part to which touch is applied [bib_ref] Reports of intimate touch: Erogenous zones and somatosensory cortical organization, Turnbull [/bib_ref] or the (affective) modality of touch that is used that drives our pleasant and erotic perception [bib_ref] Unmyelinated tactile cutaneous nerves signal erotic sensations, Jönsson [/bib_ref] [bib_ref] Mediated Erotic Touch Perception Relates to Sexual Desire and Performance in a..., Bendas [/bib_ref]. This study explored the relationship between body part and tactile modality in eliciting pleasant and erotic sensations, by using the socially-relevant CTafferent system tuned to affective touch, in order to examine the perceptual effects of imagined and actual sensual vs. neutral touch in erogenous vs. non-erogenous zones. The graph shows that slow stroking was perceived as more pleasant than all other types of touch and that for slow stroking and fast patting pleasantness was higher for erogenous as compared to non-erogenous zones. Error bars indicate standard error.
In Experiment 1 we explored actual stimulation of an erogenous (neck) vs. non-erogenous zone (forehead), with the aim of assessing subjective pleasantness and erogeneity. We tested romantic couples, which allowed us to address the reciprocal property of touch in an ecologically valid setup. In line with previous research, the results showed that sensual touch was perceived as more pleasant and more sexually arousing than neutral touch. Similarly, we also found that touch on the neck (erogenous) was rated as more pleasant and sexually arousing compared to touch on the forehead (non-erogenous). However, the lack of interaction suggests that CT-optimal stroking (i.e. modality of touch) and touch on erogenous zones (i.e. body part) may mediate subjective pleasantness and erogeneity by at least partly independent mechanisms.
In Experiment 2, we explored pleasantness perception of imagined sensual (CT-optimal) vs. neutral (CT-suboptimal) stimulation of erogenous vs. non-erogenous zones. Our findings provide the first direct evidence that imagining sensual touch was perceived as more pleasant than all other types of touch, suggesting that pleasantness induced by merely imagining touch may be driven, at least in part, by top-down processes, guided by learned expectations of sensory pleasure [bib_ref] Active interpersonal touch gives rise to the social softness illusion, Gentsch [/bib_ref] [bib_ref] Vicarious responses to social touch in posterior insular cortex are tuned to..., Morrison [/bib_ref]. In fact, interpersonal touch is generally present in romantic interactions, therefore most people have memories of the pleasant state that sensual, CT-optimal stimulation elicits. In contrast to previous findings [bib_ref] Unmyelinated tactile cutaneous nerves signal erotic sensations, Jönsson [/bib_ref] [bib_ref] Quantitative assessment of pleasant touch, Essick [/bib_ref] , we also found that pleasantness was higher when participants were asked to imagine sensual stimulation being applied on erogenous as compared to non-erogenous zones. The lack of an interaction in this second experiment also suggests that CT-optimal stroking (i.e. modality of touch) and touch on erogenous zones (i.e. body part) may mediate subjective pleasantness of imagined touch by at least partly independent mechanisms.
In addition to the findings regarding our main hypotheses, in Experiment 1 we found no differences between touch-receivers and givers in how they reported their subjective sexual arousal ratings for CT-optimal vs. CT-suboptimal touch and for neck vs. forehead. In our sample, arousal ratings had higher variance compared to pleasantness ratings, which could explain the absence of significant differences in our analyses. Variability in arousal across subjects is not surprising, given the known impact of contextual and individual factors on erotic perception [bib_ref] Effects of positive and negative mood on sexual arousal in sexually functional..., Mitchell [/bib_ref] [bib_ref] Ambivalent affect and sexual response: The impact of co-occurring positive and negative..., Peterson [/bib_ref]. Interestingly however, and indeed despite this individual variability, our additional correlation analyses (S1 File) show that both across body parts and stroking velocity, participants that rated the touch as highly arousing also found the same touch as pleasant. Additionally, across participants we found positive associations in arousal ratings of givers and receivers in both erogenous and non-erogenous zones, as well as for CT-optimal stroking velocity. Previous research has reported hedonic benefits of CT-optimal touch for the giver [bib_ref] Active interpersonal touch gives rise to the social softness illusion, Gentsch [/bib_ref]. More specifically, there is evidence that stroking others' skin feels softer and smoother than stroking one's own skin. This so-called "social softness illusion" is selectively present when touch is applied according to the optimal properties of the receiver's CT-based affective touch system, i.e. optimal velocities and hairy skin only [bib_ref] Active interpersonal touch gives rise to the social softness illusion, Gentsch [/bib_ref]. Here we suggest that similar mechanisms might be in place in the context of CT-mediated erotic touch. Activating the neurophysiological system for sensual touch in the receiver, appeared to induce feelings of sexual arousal to the giver as well, reflecting a mechanism of sensual sharing between individuals.
In addition, gender differences were found in the way males and females reported the experience of receiving touch in different body parts. Women rated touch on the neck as more pleasant than similar touch on the forehead. This finding is in line with the recent evidence that while participants of both genders equally rated the forehead as not sexually arousing, the mean score of level of arousal for the neck was higher for females [bib_ref] Reports of intimate touch: Erogenous zones and somatosensory cortical organization, Turnbull [/bib_ref] , pointing to potentially higher levels of sensitivity to the neck in women. However, considering our relatively small sample size, future research should replicate and provide support to these findings.
Our data support the presence of an interplay between affective touch and erogenous zones, possibly mediated by other factors, such as gender differences and context. Neuropsychological and behavioural studies have now widely demonstrated the effects of affective tactile interactions for affiliative behaviour, social bonding, and emotion regulation [bib_ref] Touch communicates distinct emotions, Hertenstein [/bib_ref] [bib_ref] The Skin as a Social Organ, Morrison [/bib_ref] ]. Yet touch can have different impacts on others, depending on contextual factors and situational variables. Research has shown that CT touch might scaffold rather than determine erotic perception, by creating a sort of emotional backdrop to sexual feelings [bib_ref] Unmyelinated tactile cutaneous nerves signal erotic sensations, Jönsson [/bib_ref]. Therefore, other factors should be taken into account when considering erotic touch perception and this might be especially true in the context of a romantic relationship. For example, tactile interactions in a newly-established romantic relationship may be perceived as more erotic compared to the same touch in longer-term couples. In fact, previous studies have found that sexual arousal ratings are negatively associated with relationship length and that people in a long-term relationship are less sexually active than newly established couples [bib_ref] Unmyelinated tactile cutaneous nerves signal erotic sensations, Jönsson [/bib_ref].
Neuroimaging studies have shown that CT-optimal touch specifically activates the insula [bib_ref] Unmyelinated tactile afferents signal touch and project to insular cortex, Olausson [/bib_ref] , and that a map in this brain region could potentially offer an explanation for the arrangement of the erogenous zones [bib_ref] Reports of intimate touch: Erogenous zones and somatosensory cortical organization, Turnbull [/bib_ref]. The insula has been recently identified as the cortical hub for the primary interoceptive information about the physiological condition of the body . Evidence also points to the insula in integrating interoceptive states with other exteroceptive, cognitive and social information, providing the basis for all subjective feelings from the body and even emotional awareness [bib_ref] Interoception: the inside story-a model for psychosomatic processes, Cameron [/bib_ref]. While pleasant and erotic perception of sensual touch in erogenous zones could be associated with activity in anterior insula, as mentioned above, the absence of an interaction between touch modality and body part in both experiments (i.e. with actual and imagined touch) suggests that different cognitive mechanisms might be responsible for the conscious experience of pleasantness and erogeneity. More specifically, it seems that CT-optimal touch on non-erogenous zones, and vice versa, is still perceived to be pleasant and erotic, suggesting that pleasantness and erogeneity of touch can be elicited by both mechanisms, together or in isolation. Future studies could establish whether these two sources of erogeneity are influenced differentially by contextual factors, such as gender or relationship nature/quality. Moreover, the involvement of neural circuits was not directly tested, hence a conclusion about the neural pathways for conscious perception of touch sensuality cannot be reached. Further neuroimaging research is, therefore, required to address this.
Altogether, this study provides the first direct evidence that while how we touch and where we touch are both crucial in eliciting pleasant and erotic sensations, they represent dissociable sources of pleasantness and erogeneity in social touch. Furthermore, sensory pleasure and erogeneity seem to be driven at least partly by top-down, learned expectations. In fact, pleasantness is induced by mere imagination of touch without any tactile stimulation. Nevertheless, for this experiment only women were tested to control for gender effects related to touch [bib_ref] Primary somatosensory cortex discriminates affective significance in social touch, Gazzola [/bib_ref] , thus, future studies could investigate whether the present results also extend to women instructed by male experimenter and men instructed by female and male experimenters. Further evidence for the role of top-down factors in erotic perception is provided by the fact that there seems to be a strong sensual reciprocity of giving and receiving touch among romantic couples. This intersubjective communication and the subsequent erotogenic benefits for both receiver and giver may therefore act as a reinforcer for interpersonal touch and affiliative behaviours. Future studies could use autonomic measures, such as skin conductance and heart rate, to explore the mechanisms of interpersonal synchronisation between touch-givers and receivers, which has been found to be crucial for social behaviours [36] and tease apart other top-down factors (e.g. expectations) that could influence pleasant and erotic perception of touch.
## Supporting information
[fig] Fig 1: Pleasantness ratings of givers and receivers. The graph shows that, among Givers, there was no significant difference between males and females in pleasantness ratings on the neck or the forehead. Among Receivers, females preferred more to be touched on the neck compared to the forehead. Error bars indicate standard error.https://doi.org/10.1371/journal.pone.0203039.g001 [/fig]
[fig] Fig 2: Arousal ratings of givers and receivers for the neck and forehead. The graph shows that there was no preference in arousal ratings between Receivers and Givers. Error bars indicate standard error. https://doi.org/10.1371/journal.pone.0203039.g002 Dissociable sources of erogeneity in social touch PLOS ONE | https://doi.org/10.1371/journal.pone.0203039 August 24, 2018 [/fig]
[fig] Fig 3: Pleasantness ratings for different types of imagined touch in erogenous vs. non-erogenous zones. [/fig]
|
Review of Recent Developments on Using an Off-Lattice Monte Carlo Approach to Predict the Effective Thermal Conductivity of Composite Systems with Complex Structures
Here, we present a review of recent developments for an off-lattice Monte Carlo approach used to investigate the thermal transport properties of multiphase composites with complex structure. The thermal energy was quantified by a large number of randomly moving thermal walkers. Different modes of heat conduction were modeled in appropriate ways. The diffusive heat conduction in the polymer matrix was modeled with random Brownian motion of thermal walkers within the polymer, and the ballistic heat transfer within the carbon nanotubes (CNTs) was modeled by assigning infinite speed of thermal walkers in the CNTs. Three case studies were conducted to validate the developed approach, including three-phase single-walled CNTs/tungsten disulfide (WS 2 )/(poly(ether ether ketone) (PEEK) composites, single-walled CNT/WS 2 /PEEK composites with the CNTs clustered in bundles, and complex graphene/poly(methyl methacrylate) (PMMA) composites. In all cases, resistance to heat transfer due to nanoscale phenomena was also modeled. By quantitatively studying the influencing factors on the thermal transport properties of the multiphase composites, it was found that the orientation, aggregation and morphology of fillers, as well as the interfacial thermal resistance at filler-matrix interfaces would limit the transfer of heat in the composites. These quantitative findings may be applied in the design and synthesis of multiphase composites with specific thermal transport properties.
# Introduction
Polymer composites can combine the merits of fillers and polymer matrix in a hybrid system. They can even replace the traditional materials (e.g., glass-based materials, metals and alloys) in a number of applications, due to their unique optical, mechanical, thermal and electrical properties [bib_ref] Progress in green polymer composites from lignin for multifunctional applications: A review, Thakur [/bib_ref]. During recent years, carbon-based two-phase polymer composites, such as carbon nanotube (CNT)-polymer composites and graphene-polymer composites, have attracted much attention from both academy and industry, owing to their superior strength, high thermal conductivity and good electrical conductivity [bib_ref] Carbon nanotubes: Present and future commercial applications, Volder [/bib_ref] [bib_ref] Graphene-multilayer graphene nanocomposites as highly efficient thermal interface materials, Shahil [/bib_ref] [bib_ref] Advanced multifunctional graphene aerogel-Poly(methyl methacrylate) composites: Experiments and modeling, Fan [/bib_ref]. In general, the thermal conductivity and electrical conductivity of carbon additives (e.g., CNTs, graphene, carbon black, carbon fiber) are both much higher than those of polymer matrices.
With the additions of carbon fillers into polymers, both the thermal and electrical conductivity of composites can be significantly enhanced, which may limit the applications of composites in some specific fields. For instance, in electronic packaging, the packaging materials require high thermal conductivity but low electrical conductivity. Two-phase composites with only one type of carbon filler are difficult to achieve this requirement. Multi-types of fillers are required to enhance thermal conductivity but retain electrical insulation for composites. For example, Dai et al. developed polyimide (PI) composites containing one-dimensional (1D) SiC nanowires and two-dimensional (2D) graphene nanosheets as hybrid fillers [bib_ref] Enhanced thermal conductivity and retained electrical insulation for polyimide composites with sic..., Dai [/bib_ref]. As the 1D SiC nanowires prevent the graphene nanosheets forming networks, the three-phase composite achieves a low electrical conductivity (~1.32ˆ10´1 0 S/cm) but a high thermal conductivity (~0.577 W/m¨K) [bib_ref] Enhanced thermal conductivity and retained electrical insulation for polyimide composites with sic..., Dai [/bib_ref].
On the other hand, as electronic devices are developing to the integrated and micro-scale level, thermal interface materials (TIMs) with high thermal conductivity are necessary to effectively dissipate the heat and then to protect electronic devices [bib_ref] Thermal properties of graphene and few-layer graphene: Applications in electronics, Yan [/bib_ref] [bib_ref] Alumina-coated graphene sheet hybrids for electrically insulating polymer composites with high thermal..., Qian [/bib_ref]. Traditional TIMs with one type of filler require high loading of fillers to achieve the expected thermal conductivity of 0.5-5 W/m¨K, which dramatically increases the cost and the viscosity of the composite. Recently, it has been found that combining multiple fillers in polymer may significantly enhance the thermal conductivity of the composites owing to the synergistic effect of fillers. Zhao et al. obtained a three-phase composite by adding 0.2 vol % graphene foam (GF) and 2.7 vol % multilayer graphene flakes (MGFs) into polydimethylsiloxane (PDMS). Due to the synergistic effect of GF and MGFs, the MGF/GF/PDMS composites achieved a thermal conductivity of up to 1.08 W/m¨K, which is 440%, 184% and 80% higher than those of pure PDMS, 0.2 vol % GF/PDMS and 2.7 vol % MGF/PDMS composites, respectively [bib_ref] High thermal conductivity of flexible polymer composites due to synergistic effect of..., Zhao [/bib_ref].
In addition to the above multiphase composites, other diverse types of multiphase composites have been developed for their specific applications, such as CNT/inorganic nanoparticle/polymer [bib_ref] New hybrid nanocomposites containing carbon nanotubes, inorganic fullerene-like WS 2 nanoparticles and..., Naffakh [/bib_ref] [bib_ref] Morphology and thermal properties of novel poly(phenylene sulfide) hybrid nanocomposites based on..., Naffakh [/bib_ref] , CNT/graphene/polymer [bib_ref] Thermal conductivity of a graphene oxide-carbon nanotube hybrid/epoxy composite, Im [/bib_ref] [bib_ref] Multi-walled carbon nanotube-graphene-polyaniline multiphase nanocomposite with superior electromagnetic shielding effectiveness, Gupta [/bib_ref] [bib_ref] Carbon hybrid fillers composed of carbon nanotubes directly grown on graphene nanoplatelets..., Yu [/bib_ref] , graphene/inorganic nanoparticle/polymer [bib_ref] Effect of nano-fillers on the thermal conductivity of epoxy composites with micro-Al..., Gao [/bib_ref] [bib_ref] Novel conductive epoxy composites composed of 2-D chemically reduced graphene and 1-D..., Luan [/bib_ref] , and polymer blends [bib_ref] Current advances in the carbon nanotube/thermotropic main-chain liquid crystalline polymer nanocomposites and..., Cheng [/bib_ref] [bib_ref] Enhancement of fracture toughness, mechanical and thermal properties of rubber/epoxy composites by..., Wang [/bib_ref]. Such multiphase composites can combine the merits of fillers and matrix to achieve advanced properties for diverse applications. Experimental measurements have shown that the effective thermal conduction properties of multiphase composites can be significantly enhanced due to the synergistic effects of fillers. However, computational studies in this field are still limited due to the complex structure and composition of multiphase composites. The effective medium theory has been used to estimate the effective thermal conductivity (K eff ) of multiphase composites [bib_ref] Thermal transport phenomena and limitations in heterogeneous polymer composites containing carbon nanotubes..., Gong [/bib_ref]. The K eff is calculated in separated steps, which neglects the synergistic effect of fillers. An effective modeling of heat transfer in multiphase composites can reveal the thermal transport phenomena and limitations in them, which may help to design multiphase composites with specific thermal properties.
In this article, we review an effective approach to model the heat transfer in multiphase composites with complex structures based on our work about the computational studies on the effective thermal properties of multiphase composites. The paper is organized as following: In Simulation Methods, an off-lattice Monte Carlo approach for modeling heat transfer in a three-phase composite is presented in detail. Three case studies are discussed in Results and Discussion to demonstrate the novelty, accuracy and superiority of the developed approach. The case-study systems include three-phase single-walled CNT (SWNT)/tungsten disulfide (WS 2 )/poly(ether ether ketone) (PEEK) composite, four-phase SWNT/SWNT bundle/WS 2 /PEEK composite, and complex graphene/poly(methyl methacrylate) (PMMA) composite. We quantitatively studied the factors influencing the thermal transport in multiphase composites, such as the morphology, dispersion, alignment and concentration of fillers, as well as the effects of the interfacial thermal resistance between any pair of components. Such quantitative findings may guide researchers to design and synthesize multiphase composites with specific thermal transport properties.
# Simulation methods
In this section, we chose a three-phase SWNT/WS 2 /PEEK composite to illustrate the off-lattice Monte Carlo approach. The off-lattice Monte Carlo approach regards the heat flow as the result of the movement of discrete thermal walkers with random motion [bib_ref] Random walks in nanotube composites: Improved algorithms and the role of thermal..., Duong [/bib_ref]. The Monte Carlo approach has been successfully applied to model heat or mass transport in flow through porous materials [bib_ref] Efficient lagrangian scalar tracking method for reactive local mass transport simulation through..., Voronov [/bib_ref] [bib_ref] Transport properties of random arrays of freely overlapping cylinders with various orientation..., Tomadakis [/bib_ref] and in convective flows [bib_ref] Turbulent transport from continuous sources at the wall of a channel, Papavassiliou [/bib_ref] [bib_ref] On the prandtl or schmidt number dependence of the turbulent heat or..., Mitrovic [/bib_ref]. A three-dimensional (3D) model was built based on the SWNT/WS 2 /PEEK composite fabricated by Naffakh et al. [bib_ref] New hybrid nanocomposites containing carbon nanotubes, inorganic fullerene-like WS 2 nanoparticles and..., Naffakh [/bib_ref]. As shown in [fig_ref] Figure 1: A schematic plot of the SWNT/WS2/PEEK model [/fig_ref] , SWNTs (500 nm length and 2 nm diameter) and WS 2 spherical nanoparticles (110 nm diameter) were randomly and uniformly dispersed in PEEK matrix (925 nm side length). The mass fractions of SWNTs, WS 2 nanoparticles and PEEK were 0.5 wt. %, 0.5 wt. % and 99.0 wt. %, respectively. The 3D model is a representative volume element (RVE) of the three-phase composite, which can be repeated to accurately replicate the realistic composite samples. In this section, we chose a three-phase SWNT/WS2/PEEK composite to illustrate the off-lattice Monte Carlo approach. The off-lattice Monte Carlo approach regards the heat flow as the result of the movement of discrete thermal walkers with random motion [bib_ref] Random walks in nanotube composites: Improved algorithms and the role of thermal..., Duong [/bib_ref]. The Monte Carlo approach has been successfully applied to model heat or mass transport in flow through porous materials [bib_ref] Efficient lagrangian scalar tracking method for reactive local mass transport simulation through..., Voronov [/bib_ref] [bib_ref] Transport properties of random arrays of freely overlapping cylinders with various orientation..., Tomadakis [/bib_ref] and in convective flows [bib_ref] Turbulent transport from continuous sources at the wall of a channel, Papavassiliou [/bib_ref] [bib_ref] On the prandtl or schmidt number dependence of the turbulent heat or..., Mitrovic [/bib_ref]. A three-dimensional (3D) model was built based on the SWNT/WS2/PEEK composite fabricated by Naffakh et al. [bib_ref] New hybrid nanocomposites containing carbon nanotubes, inorganic fullerene-like WS 2 nanoparticles and..., Naffakh [/bib_ref]. As shown in [fig_ref] Figure 1: A schematic plot of the SWNT/WS2/PEEK model [/fig_ref] , SWNTs (500 nm length and 2 nm diameter) and WS2 spherical nanoparticles (110 nm diameter) were randomly and uniformly dispersed in PEEK matrix (925 nm side length). The mass fractions of SWNTs, WS2 nanoparticles and PEEK were 0.5 wt. %, 0.5wt. % and 99.0 wt. %, respectively. The 3D model is a representative volume element (RVE) of the three-phase composite, which can be repeated to accurately replicate the realistic composite samples. is placed in the center of a PEEK cube with a side length of 925 nm, while 317 SWNTs (2 nm diameter and 500 nm length) are randomly distributed in the PEEK cube. The WS2 particle is painted red and the nanotubes are black in the figure. Constant heat flux is applied along the x direction by creating a hot surface and a cooled surface (Reproduced with permission from [bib_ref] Thermal transport phenomena and limitations in heterogeneous polymer composites containing carbon nanotubes..., Gong [/bib_ref]. Copyright Elsevier, 2015); (b) a contour plot of thermal walker distribution in the center xy plane of the SWNT/WS2/PEEK model at the thermal steady state (Reproduced with permission from [bib_ref] Inter-carbon nanotube contact and thermal resistances in heat transport of three-phase composites, Gong [/bib_ref]. Copyright American Chemical Society, 2015).
To model the heat transfer in the composite, a constant heat flux was applied along x-direction by continuously releasing a large quantity (e.g., 40,000) of thermal walkers from both sides in each time step (i.e., hot walkers from x = 0, and cold walkers from the other side). The two sides can be treated as a hot and a cold surface, as presented in [fig_ref] Figure 1: A schematic plot of the SWNT/WS2/PEEK model [/fig_ref]. The hot walkers and cold walkers (with negative energy) have same absolute value of energy, so the energy of the whole system is conserved. A thermal walker jumps randomly following a Brownian motion after released from the surfaces. The Brownian motion is described by position changes of thermal walkers in each direction. The position changes take values from a normal distribution with a zero mean and a standard deviation, σ, expressed as [bib_ref] The electrodynamics of moving bodies, Einstein [/bib_ref] :
[formula] m Dt (1) [/formula]
where m D is the thermal diffusivity of the matrix and t is the time step duration of the simulation.
Thermal walkers travel randomly from the initial release surface. Once a thermal walker jumps to the interface between PEEK and a SWNT, it is allowed to either jump into the SWNT with a probability of fm-SWNT, or still remains in the PEEK matrix with a probability of (1 − fm-SWNT). The fm-SWNT is related to the interfacial thermal resistance (known as Kapitza resistance) between the PEEK matrix and SWNTs, which can be estimated from the acoustic mismatch theory, as follows [bib_ref] Thermal-boundary resistance, Swartz [/bib_ref] : is placed in the center of a PEEK cube with a side length of 925 nm, while 317 SWNTs (2 nm diameter and 500 nm length) are randomly distributed in the PEEK cube. The WS 2 particle is painted red and the nanotubes are black in the figure. Constant heat flux is applied along the x direction by creating a hot surface and a cooled surface (Reproduced with permission from [bib_ref] Thermal transport phenomena and limitations in heterogeneous polymer composites containing carbon nanotubes..., Gong [/bib_ref]. Copyright Elsevier, 2015); (b) a contour plot of thermal walker distribution in the center xy plane of the SWNT/WS 2 /PEEK model at the thermal steady state (Reproduced with permission from [bib_ref] Inter-carbon nanotube contact and thermal resistances in heat transport of three-phase composites, Gong [/bib_ref]. Copyright American Chemical Society, 2015).
To model the heat transfer in the composite, a constant heat flux was applied along x-direction by continuously releasing a large quantity (e.g., 40,000) of thermal walkers from both sides in each time step (i.e., hot walkers from x = 0, and cold walkers from the other side). The two sides can be treated as a hot and a cold surface, as presented in [fig_ref] Figure 1: A schematic plot of the SWNT/WS2/PEEK model [/fig_ref]. The hot walkers and cold walkers (with negative energy) have same absolute value of energy, so the energy of the whole system is conserved. A thermal walker jumps randomly following a Brownian motion after released from the surfaces. The Brownian motion is described by position changes of thermal walkers in each direction. The position changes take values from a normal distribution with a zero mean and a standard deviation, σ, expressed as [bib_ref] The electrodynamics of moving bodies, Einstein [/bib_ref] :
[formula] σ " a D m ∆t (1) [/formula]
where D m is the thermal diffusivity of the matrix and ∆t is the time step duration of the simulation. Thermal walkers travel randomly from the initial release surface. Once a thermal walker jumps to the interface between PEEK and a SWNT, it is allowed to either jump into the SWNT with a probability of f m-SWNT , or still remains in the PEEK matrix with a probability of (1´f m-SWNT ). The f m-SWNT is related to the interfacial thermal resistance (known as Kapitza resistance) between the PEEK matrix and SWNTs, which can be estimated from the acoustic mismatch theory, as follows [bib_ref] Thermal-boundary resistance, Swartz [/bib_ref] :
[formula] f m-SWNT " 4 ρ m C P m ν m R bd(2) [/formula]
where ρ m , C P m , ν m and R bd are the density of the PEEK matrix, the specific heat capacity of PEEK, the speed of sound in PEEK, and the interfacial thermal resistance between PEEK and SWNT, respectively. When a thermal walker jumps into a SWNT, it is assumed to travel with an infinite speed, due to the ballistic phonon transport and ultrahigh thermal conductivity of SWNTs [bib_ref] Off-lattice monte carlo simulation of heat transfer through carbon nanotube multiphase systems..., Gong [/bib_ref]. The implementation of this assumption occurs by randomly placing the walker anywhere within the SWNT. The random placement is based on a uniform distribution function and it occurs in a single time step. The diffusivity inside the CNTs would need to be considered explicitly (using a second Gaussian random motion) if the CNTs were on the same order of magnitude or longer than the wavelength of a phonon. However, this case would present itself for much longer CNTs than the ones consider presently. When inside a SWNT, thermal walkers may exit the SWNT based on another probability, designated as f SWNT-m . This probability is determined from f m-SWNT , as [bib_ref] Random walks in nanotube composites: Improved algorithms and the role of thermal..., Duong [/bib_ref] :
[formula] V SWNT f SWNT-m " C f´SWNT σA SWNT f m-SWNT (3) [/formula]
where V SWNT and A SWNT are the volume and surface area of SWNTs, and C f´SWNT is a thermal equilibrium factor at the PEEK-SWNT interface which depends on the geometry of SWNTs and the interfacial area between PEEK and a SWNT. The thermal equilibrium factor is introduced in order to preserve the second law of thermodynamics at thermal equilibrium. The above relation between f SWNT-m and f m-SWNT can be explained as follow: When reaching the thermal steady state, the heat flux exiting a SWNT should be equal to that entering the SWNT. In a time step, all thermal walkers inside a SWNT may travel to the PEEK matrix owing to their infinite speed. However, among the thermal walkers in PEEK, only those around the SWNT surface may jump into the SWNT due to the Brownian motion. In order to maintain a balanced heat flux exiting and entering a SWNT, the relation as described in Equation (3) should be satisfied. When inside a WS 2 nanoparticle, a thermal walker jumps randomly, similar to the way it travels in the PEEK matrix, but with a different thermal diffusivity (that of WS 2 ) used in Equation (1). At the PEEK-WS 2 nanoparticle interface, thermal walkers from either the PEEK or the WS 2 behave similarly to the walkers crossing the PEEK-SWNT interface from the PEEK side. However, the two probabilities ( f m-WS 2 and f WS 2´m ) have different relation from that described in Equation (3), due to the different motion of walkers in WS 2 nanoparticles and SWNTs. As thermal walkers have Brownian motion in WS 2 nanoparticles, f m-WS 2 and f WS 2´m are related as: where f WS 2´m and f m-WS 2 are the walker travelling probabilities from the WS 2 nanoparticles to the PEEK matrix and the reverse. Variables r, σ m and σ WS 2 are the radius of the WS 2 nanoparticles, the standard deviation of Brownian motion in the PEEK matrix, and the standard deviation of Brownian motion in the WS 2 nanoparticles, respectively. The parameter C f´WS 2 is the thermal equilibrium factor at the PEEK-WS 2 nanoparticle interface, which can be numerically determined in the same manner as C f´SWNT [bib_ref] Off-lattice monte carlo simulation of heat transfer through carbon nanotube multiphase systems..., Gong [/bib_ref] [bib_ref] Inter-carbon nanotube contact in thermal transport of controlled-morphology polymer nanocomposites, Duong [/bib_ref] [bib_ref] Mesoscopic modeling of cancer photothermal therapy using single-walled carbon nanotubes and near..., Gong [/bib_ref]. Thermal walkers will be bounced back when they jump outside of the model to maintain a constant heat flux. Periodic boundary conditions are applied in the y and z directions. The computational domain was divided into 300ˆ300ˆ300 grids to calculate the temperature profile, which can be obtained by counting the number of hot walkers and then subtracting the number of cold walkers in a grid cell. [fig_ref] Figure 1: A schematic plot of the SWNT/WS2/PEEK model [/fig_ref] shows a typical contour plot of thermal walker distribution in the thermal steady state. With constant heat flux applied along the x direction, the temperature along the x direction should be a straight line which has a slope inversely proportional to the thermal conductivity of the SWNT/WS 2 /PEEK composite. A reference model of pure PEEK matrix was also built to estimate the effective thermal conductivity (K eff ) of the composite. With the same heat flux and boundary conditions, the temperature profiles along x direction in the composite model and the pure matrix model are related as:
[formula] f WS 2´m " C f´WS 2 pr`σ m q 3´r3 r 3´`´σ WS 2˘3 f m-WS 2 (4)q 2 "´K eff dT c dx "´K m dT m dx (5) [/formula]
where q", T c and T m are the applied constant heat flux, the temperature in the composite and the temperature in the pure PEEK matrix, respectively. K m is the thermal conductivity of the pure PEEK matrix, which is known to be 0.23 W/m¨K from the literature [bib_ref] High-performance nanocomposites based on polyetherketones, Diez-Pascual [/bib_ref]. Thus, as expressed in Equation (5), the effective thermal conductivity of the composite (K eff ) can be calculated based on the temperature profiles in the composite and the pure PEEK.
# Results and discussion
In this section, three case studies are presented and discussed to demonstrate the capabilities of the developed off-lattice Monte Carlo approach. The case study systems include three-phase SWNT/WS 2 / PEEK composites, SWNT/WS 2 /PEEK composites with SWNT bundles and graphene/PMMA composites with different sized graphene sheets. The typical parameters influencing or limiting the thermal transport properties of the composites were quantitatively investigated. The quantitative findings may provide an overview of the thermal transport phenomena and limitation mechanisms in multiphase composites.
## Model of three-phase swnt/ws 2 /peek composites
## Validation of the developed off-lattice monte carlo approach
The developed approach was validated by comparing the simulation results with the measured K eff of SWNT/WS 2 /PEEK composites with different compositions [bib_ref] New hybrid nanocomposites containing carbon nanotubes, inorganic fullerene-like WS 2 nanoparticles and..., Naffakh [/bib_ref] , as presented in . The different compositions (e.g., 0.5/0.5/99.0) were marked with the mass fractions of the SWNTs (0.5%), the WS 2 nanoparticles (0.5%) and the PEEK matrix (99.0%). In the models for different compositions, the dimensions of the SWNTs and the diameter of the WS 2 nanoparticle were maintained, while the side length of the PEEK cube was varied. As shown in , the simulation results from our approach are in good agreement with the experimental data, which validates the developed approach.
Keff of SWNT/WS2/PEEK composites with different compositions [bib_ref] New hybrid nanocomposites containing carbon nanotubes, inorganic fullerene-like WS 2 nanoparticles and..., Naffakh [/bib_ref] , as presented in . The different compositions (e.g., 0.5/0.5/99.0) were marked with the mass fractions of the SWNTs (0.5%), the WS2 nanoparticles (0.5%) and the PEEK matrix (99.0%). In the models for different compositions, the dimensions of the SWNTs and the diameter of the WS2 nanoparticle were maintained, while the side length of the PEEK cube was varied. As shown in , the simulation results from our approach are in good agreement with the experimental data, which validates the developed approach. Effective medium theory approaches (EMT) are commonly applied to predict the K eff of composites [bib_ref] Single wall carbon nanotube/polyethylene nanocomposites: Thermal and electrical conductivity, Haggenmueller [/bib_ref] [bib_ref] Measurement and model on thermal conductivities of carbon nanotube nanorefrigerants, Jiang [/bib_ref] [bib_ref] A simple model for thermal conductivity of carbon nanotube-based composites, Nan [/bib_ref] [bib_ref] Model for the enective thermal conductivity of carbon nanotube composites, Xue [/bib_ref]. For comparison with our developed approach, two widely-used EMTs (i.e., the Maxwell-Garnett model (MG) [bib_ref] A simple model for thermal conductivity of carbon nanotube-based composites, Nan [/bib_ref] [bib_ref] Anomalous thermal conductivity enhancement in nanotube suspensions, Choi [/bib_ref] and Nan et al.'s model [bib_ref] Interface effect on thermal conductivity of carbon nanotube composites, Nan [/bib_ref] were utilized to predict the K eff of the SWNT/WS 2 /PEEK composites. The comparison among different models is presented in . The MG-EMT overestimated the K eff of the SWNT/WS 2 /PEEK composites, which is likely due to the neglect of the interfacial thermal resistance at the SWNT-PEEK interface. On the contrary, Nan et al.'s EMT model underestimated the K eff of the SWNT/WS 2 /PEEK composites. This is because that model does not account for the synergistic effects of SWNTs and WS 2 nanoparticles in the composites.
## Effects of interfacial thermal resistances on the k eff of swnt/ws 2 /peek composites
The interfacial thermal resistance (R bd ) at the nanofiller-matrix interface (i.e., SWNT-PEEK, WS 2 -PEEK) is caused by the difference in the vibrational phonon spectra of each side, and by interfacial defects [bib_ref] Thermal conductivity of carbon nanotubes and their polymer nanocomposites: A review, Han [/bib_ref] [bib_ref] Role of thermal boundary resistance on the heat flow in carbon-nanotube composites, Shenogin [/bib_ref]. The R bd is found to greatly limit the K eff of SWNT/WS 2 /PEEK composites [bib_ref] Morphology and thermal properties of novel poly(phenylene sulfide) hybrid nanocomposites based on..., Naffakh [/bib_ref]. The R bd at the SWNT-PEEK (R SWNT-PEEK ) and WS 2 -PEEK (R WS 2´P EEK ) interfaces was varied to investigate the effects on the K eff of SWNT/WS 2 /PEEK composites, and the results are presented in [fig_ref] Figure 3 .: Effects of the interfacial thermal resistances at [/fig_ref]. The R bd is interpreted by an average phonon transmission probability at the interfaces in our approach, as described already above. It has been reported in the literature that the R bd in the nanoscale falls into the range of 1.0ˆ10´9-1.0ˆ10´6 m 2¨K /W [bib_ref] Thermal boundary resistance at the graphene-oil interface, Konatham [/bib_ref] [bib_ref] Anisotropic heat transfer of single-walled carbon nanotubes, Maruyama [/bib_ref] [bib_ref] Interfacial heat flow in carbon nanotube suspensions, Huxtable [/bib_ref] [bib_ref] Thermal conductivity and interface thermal conductance of amorphous and crystalline Zr 47..., Shukla [/bib_ref] , so the R SWNT-PEEK was varied between 1.158ˆ10´9 and 1.158ˆ10´6 m 2¨K /W, which corresponds to an average phonon transmission probability of 0.001-1.0. Similarly, the R WS 2´P EEK ranged from 2.32ˆ10´1 0 to 2.32ˆ10´8 m 2¨K /W, corresponding to an average phonon transmission probability from 0.005 to 0.5.
is interpreted by an average phonon transmission probability at the interfaces in our approach, as described already above. It has been reported in the literature that the Rbd in the nanoscale falls into the range of 1.0 × 10 −9 -1.0 × 10 −6 m 2 ·K/W [bib_ref] Thermal boundary resistance at the graphene-oil interface, Konatham [/bib_ref] [bib_ref] Anisotropic heat transfer of single-walled carbon nanotubes, Maruyama [/bib_ref] [bib_ref] Interfacial heat flow in carbon nanotube suspensions, Huxtable [/bib_ref] [bib_ref] Thermal conductivity and interface thermal conductance of amorphous and crystalline Zr 47..., Shukla [/bib_ref] , so the PEEK -SWNT R was varied between 1.158 × 10 −9 and 1.158 × 10 −6 m 2 ·K/W, which corresponds to an average phonon transmission probability of 0.001-1.0. Similarly, the PEEK -WS 2 R ranged from 2.32 × 10 −10 to 2.32 × 10 −8 m 2 ·K/W, corresponding to an average phonon transmission probability from 0.005 to 0.5. As shown in [fig_ref] Figure 3 .: Effects of the interfacial thermal resistances at [/fig_ref] , the K eff of SWNT/WS 2 /PEEK composites decreases with both R SWNT-PEEK and R WS 2´P EEK . In [fig_ref] Figure 3 .: Effects of the interfacial thermal resistances at [/fig_ref] , compared with that in the parallel SWNT case, the effect of R WS 2´P EEK on the K eff of SWNT/WS 2 /PEEK composites is much weaker in the random and perpendicular SWNT cases. R SWNT-PEEK significantly impedes the transfer of heat between SWNTs and the PEEK matrix, weakening the enhancement of SWNTs on the K eff . At a high R SWNT-PEEK , SWNTs with three different orientations lead to a similar K eff , which is close to the K PEEK , indicating that SWNTs do not enhance the heat conduction in the composite. A low R SWNT-PEEK is desired to obtain the composite with high K eff , which may be achieved by proper functionalization of SWNTs to couple the phonon spectra of SWNTs and PEEK [bib_ref] Effect of chemical functionalization on thermal transport of carbon nanotube composites, Shenogin [/bib_ref] [bib_ref] Modeling of interfacial modification effects on thermal conductivity of carbon nanotube composites, Clancy [/bib_ref] [bib_ref] The effect of non-covalent functionalization on the thermal conductance of graphene/organic interfaces, Lin [/bib_ref]. The influence of R WS 2´P EEK on the K eff of composites is much weaker than that of R SWNT-PEEK . This may be because: (i) the ultrahigh thermal conductivity of SWNTs allows them to dominate the heat transfer through composites; (ii) the much larger interfacial area of SWNTs make the influence of the SWNT-PEEK interface more significant than that of the WS 2 -PEEK interface for heat transfer; (iii) the long cylindrical SWNTs are more effective than the spherical WS 2 nanoparticles to enhance the K eff of the composites [bib_ref] Particle aspect-ratio effects on the thermal conductivity of micro-and nanoparticle suspensions, Cherkasova [/bib_ref] , offering a larger characteristic length scale over which heat can be transferred.
## Effects of the morphology of swnts on the k eff of swnt/ws 2 /peek composites
Since SWNTs dominate the heat transfer in the three-phase composites, the morphology of SWNTs (i.e., length and diameter) was varied to study its effect on the K eff of SWNT/WS 2 /PEEK composites. The length was varied from 100 to 900 nm while keeping a constant diameter of 2 nm. The diameter was varied from 2 to 8 nm while the length was kept at 500 nm. As presented in [fig_ref] Figure 4: Effects of [/fig_ref] , when SWNTs are parallel or random to the heat flux, longer SWNTs induce higher K eff , which is consistent with previous studies [bib_ref] Thermal conductivity of carbon nanotubes and their polymer nanocomposites: A review, Han [/bib_ref] [bib_ref] The influence of carbon nanotube aspect ratio on thermal conductivity enhancement in..., Kapadia [/bib_ref]. Due to the ballistic phonon transport in SWNTs, longer SWNTs are more effective than short ones to transport heat through the composites, leading to a higher K eff [bib_ref] Particle aspect-ratio and agglomeration-state effects on the effective thermal conductivity of aqueous..., Cherkasova [/bib_ref]. For the study of SWNT diameter, a composite with randomly oriented SWNTs was chosen for investigation as it represents most realistic composites. As shown in [fig_ref] Figure 4: Effects of [/fig_ref] , the K eff of SWNT/WS 2 /PEEK composites increases with the decrease of SWNT diameter, which may be ascribed to the interfacial area of SWNTs. With same volume fraction of SWNTs, a larger interfacial area between SWNTs and PEEK is obtained for SWNTs having smaller diameter. The larger interfacial area can afford more effective heat transfer channels between SWNTs and PEEK, thus inducing a larger enhancement of the K eff . The diameter used here covers the range of SWNTs (ď2 nm), double-walled CNTs (2-4 nm), and multi-walled CNTs (ě4 nm). Therefore, it can be speculated that SWNTs are more efficient than other CNTs to enhance the K eff of SWNT/WS 2 /PEEK composites.
SWNTs are more effective than short ones to transport heat through the composites, leading to a higher Keff [bib_ref] Particle aspect-ratio and agglomeration-state effects on the effective thermal conductivity of aqueous..., Cherkasova [/bib_ref]. For the study of SWNT diameter, a composite with randomly oriented SWNTs was chosen for investigation as it represents most realistic composites. As shown in [fig_ref] Figure 4: Effects of [/fig_ref] , the Keff of SWNT/WS2/PEEK composites increases with the decrease of SWNT diameter, which may be ascribed to the interfacial area of SWNTs. With same volume fraction of SWNTs, a larger interfacial area between SWNTs and PEEK is obtained for SWNTs having smaller diameter. The larger interfacial area can afford more effective heat transfer channels between SWNTs and PEEK, thus inducing a larger enhancement of the Keff. The diameter used here covers the range of SWNTs (≤2 nm), doublewalled CNTs (2-4 nm), and multi-walled CNTs (≥4 nm). Therefore, it can be speculated that SWNTs are more efficient than other CNTs to enhance the Keff of SWNT/WS2/PEEK composites. The length was varied from 100 to 900 nm, corresponding to an aspect ratio from 50 to 450. The diameter was varied from 2 to 8 nm while the length was kept as 500 nm. Different compositions with randomly orientated SWNTs were chosen to study the effect of SWNT diameter. The error bars represent the standard deviation of the results obtained from 3 separate simulations with different distribution of SWNTs. Reproduced with permission from [bib_ref] Thermal transport phenomena and limitations in heterogeneous polymer composites containing carbon nanotubes..., Gong [/bib_ref]. Copyright Elsevier, 2014. [fig_ref] Figure 4: Effects of [/fig_ref]. Effects of (a) length and (b) diameter of SWNTs on the K eff of SWNT/WS 2 /PEEK composites. The length was varied from 100 to 900 nm, corresponding to an aspect ratio from 50 to 450. The diameter was varied from 2 to 8 nm while the length was kept as 500 nm. Different compositions with randomly orientated SWNTs were chosen to study the effect of SWNT diameter. The error bars represent the standard deviation of the results obtained from 3 separate simulations with different distribution of SWNTs. Reproduced with permission from [bib_ref] Thermal transport phenomena and limitations in heterogeneous polymer composites containing carbon nanotubes..., Gong [/bib_ref]. Copyright Elsevier, 2014.
## Model of swnt/ws2/peek composites with swnt bundles
## Model of swnt/ws 2 /peek composites with swnt bundles
SWNTs tend to aggregate into bundles during the composite synthesis process due to strong van der Waals forces. The thermal conductivity of SWNT bundles is generally lower than that of an individual SWNT in the bundle, due to the interfacial thermal resistance among adjacent SWNTs [bib_ref] Thermal transport measurements of individual multiwalled nanotubes, Kim [/bib_ref] [bib_ref] Thermal boundary resistance and thermal conductivity of multiwalled carbon nanotubes, Prasher [/bib_ref] [bib_ref] Scaling laws and mesoscopic modeling of thermal conductivity in carbon nanotube materials, Volkov [/bib_ref]. Thus, the SWNT bundles may limit heat transfer in the composites [bib_ref] Thermal conductivity of carbon nanotubes and their polymer nanocomposites: A review, Han [/bib_ref]. A SWNT/WS 2 /PEEK composite model with SWNT bundles was built to shed some light on the influence of SWNT bundles on the thermal transport properties of the composites. As shown in [fig_ref] Figure 5: Schematic plot of the SWNT/WS2/PEEK model with SWNT bundles [/fig_ref] , there were individual SWNTs, SWNT bundles, WS 2 nanoparticles, and PEEK matrix in the system. In the present work, only the straight bundles with line contacts were taken into account. The CNT bundles with complicated interconnected networks were out of the scope of this article. SWNTs tend to aggregate into bundles during the composite synthesis process due to strong van der Waals forces. The thermal conductivity of SWNT bundles is generally lower than that of an individual SWNT in the bundle, due to the interfacial thermal resistance among adjacent SWNTs [bib_ref] Thermal transport measurements of individual multiwalled nanotubes, Kim [/bib_ref] [bib_ref] Thermal boundary resistance and thermal conductivity of multiwalled carbon nanotubes, Prasher [/bib_ref] [bib_ref] Scaling laws and mesoscopic modeling of thermal conductivity in carbon nanotube materials, Volkov [/bib_ref]. Thus, the SWNT bundles may limit heat transfer in the composites [bib_ref] Thermal conductivity of carbon nanotubes and their polymer nanocomposites: A review, Han [/bib_ref]. A SWNT/WS2/PEEK composite model with SWNT bundles was built to shed some light on the influence of SWNT bundles on the thermal transport properties of the composites. As shown in [fig_ref] Figure 5: Schematic plot of the SWNT/WS2/PEEK model with SWNT bundles [/fig_ref] , there were individual SWNTs, SWNT bundles, WS2 nanoparticles, and PEEK matrix in the system. In the present work, only the straight bundles with line contacts were taken into account. The CNT bundles with complicated interconnected networks were out of the scope of this article. [fig_ref] Figure 6: perpendicular to the heat flux, the number of SWNTs in each bundle... [/fig_ref]. The models with different SWNT orientations (parallel, random and perpendicular to the heat flux) were all built, as illustrated in [fig_ref] Figure 5: Schematic plot of the SWNT/WS2/PEEK model with SWNT bundles [/fig_ref]. As shown in [fig_ref] Figure 6: perpendicular to the heat flux, the number of SWNTs in each bundle... [/fig_ref] , when SWNT bundles are parallel or randomly oriented relative to the direction of the heat flux, the Keff of
## Effects of the morphology of swnt bundles on the k eff of swnt/ws 2 /peek composites
The number of SWNT bundles was changed from 0 to 48 (no individual SWNTs) to investigate the effect of SWNT dispersion state on the K eff of the SWNT/WS 2 /PEEK composites, and the results are presented in [fig_ref] Figure 6: perpendicular to the heat flux, the number of SWNTs in each bundle... [/fig_ref]. The models with different SWNT orientations (parallel, random and perpendicular to the heat flux) were all built, as illustrated in [fig_ref] Figure 5: Schematic plot of the SWNT/WS2/PEEK model with SWNT bundles [/fig_ref]. As shown in [fig_ref] Figure 6: perpendicular to the heat flux, the number of SWNTs in each bundle... [/fig_ref] , when SWNT bundles are parallel or randomly oriented relative to the direction of the heat flux, the K eff of the composites slightly decreases with an increase of SWNT bundles. This may be caused by the non-uniform distribution of SWNTs with the presence of SWNT bundles, as well as the SWNT-SWNT thermal resistance within the bundles, which prevents heat from conducting along the heat flux direction [bib_ref] Influence of dispersion states of carbon nanotubes on physical properties of epoxy..., Song [/bib_ref]. Different from the above two SWNT orientations, more SWNT bundles perpendicular to the heat flux can induce a higher K eff of the SWNT/WS 2 /PEEK composites. The K eff slightly increases by 40% when the bundle number increases to 48. This is likely due to the bigger diameter of SWNT bundles compared with individual SWNTs, which accelerates the heat transfer in the radial direction (also the heat flux direction).
The number of individual SWNTs in each bundle was varied from 10 to 25, at an increment of 5, to study its influence on the K eff of the composites. The total number of SWNTs (960) and the number of SWNT bundles (36) were kept constant. As shown in [fig_ref] Figure 5: Schematic plot of the SWNT/WS2/PEEK model with SWNT bundles [/fig_ref] , when SWNT bundles are parallel or randomly oriented relative to the heat flux, more SWNTs in each bundle induce a lower K eff . With the same number of bundles, more SWNTs per bundle lead to a worse distribution of SWNTs, which may reduce the heat transfer along the heat flux. On the other hand, more SWNTs in a bundle increase the stiffness of the bundle, which weakens the phonon coupling between SWNTs and PEEK via low-frequency vibrations [bib_ref] Importance of interfaces in governing thermal transport in composite materials: Modeling and..., Roy [/bib_ref] , thus leading to a larger SWNT-PEEK thermal resistance [bib_ref] Interfacial thermal resistance between carbon nanotubes: Molecular dynamics simulations and analytical thermal..., Zhong [/bib_ref]. When SWNT bundles are perpendicular to the heat flux, the number of SWNTs in each bundle has no apparent effect on the K eff of the composites. will exist among SWNTs within the bundles. Previous models cannot take into account this resistance for Keff prediction, whereas the developed approach can investigate this resistance by controlling the motion of thermal walkers at the SWNT-SWNT interface [bib_ref] Inter-carbon nanotube contact and thermal resistances in heat transport of three-phase composites, Gong [/bib_ref]. The SWNT SWNT R was varied from 6.153 × 10 −10 to 6.153 × 10 −7 m 2 ·K/W, corresponding to an average phonon transmission probability of 0.001-1.0 [bib_ref] Morphology effects on nonisotropic thermal conduction of aligned single-walled and multi-walled carbon..., Duong [/bib_ref]. As presented in , the Keff of the SWNT/WS2/PEEK composites decreases with the increase of the 3.2.2. Effects of the SWNT-SWNT Thermal Resistance on the K eff of SWNT/WS 2 /PEEK Composites When SWNTs aggregate into bundle structures, SWNT-SWNT thermal resistance (R SWNT-SWNT ) will exist among SWNTs within the bundles. Previous models cannot take into account this resistance for K eff prediction, whereas the developed approach can investigate this resistance by controlling the motion of thermal walkers at the SWNT-SWNT interface [bib_ref] Inter-carbon nanotube contact and thermal resistances in heat transport of three-phase composites, Gong [/bib_ref]. The R SWNT-SWNT was varied from 6.153ˆ10´1 0 to 6.153ˆ10´7 m 2¨K /W, corresponding to an average phonon transmission probability of 0.001-1.0 [bib_ref] Morphology effects on nonisotropic thermal conduction of aligned single-walled and multi-walled carbon..., Duong [/bib_ref]. As presented in , the K eff of the SWNT/WS 2 /PEEK composites decreases with the increase of the R SWNT-SWNT . A higher R SWNT-SWNT more greatly prevents heat from transferring among the bundled SWNTs, leading to a lower K eff of the composites.
A critical SWNT-SWNT thermal resistance, R c , was found to dominate the heat transfer in the composites, which was estimated to be 0.155ˆ10´8 m 2¨K /W. When R SWNT-SWNT < R c , more SWNT bundles may induce a higher K eff of the composite. This is because at low R SWNT-SWNT , heat prefers to transfer through SWNT-SWNT contacts, thus the SWNT bundles would be more effective heat transfer channels than the mono-dispersed SWNTs. It can be inferred that the detrimental effect of the SWNT bundles could be reduced by decreasing the R SWNT-SWNT to be less than R c , which may be achieved by covalent functionalization to enhance the phonon coupling and weaken the phonon scattering at a SWNT-SWNT interface [bib_ref] Modeling of thermal conductance at transverse CNT-CNT interfaces, Varshney [/bib_ref] [bib_ref] Transversal thermal transport in single-walled carbon nanotube bundles: Influence of axial stretching..., Gharib-Zahedi [/bib_ref].
will exist among SWNTs within the bundles. Previous models cannot take into account this resistance for Keff prediction, whereas the developed approach can investigate this resistance by controlling the motion of thermal walkers at the SWNT-SWNT interface [bib_ref] Inter-carbon nanotube contact and thermal resistances in heat transport of three-phase composites, Gong [/bib_ref]. The SWNT SWNT R was varied from 6.153 × 10 −10 to 6.153 × 10 −7 m 2 ·K/W, corresponding to an average phonon transmission probability of 0.001-1.0 [bib_ref] Morphology effects on nonisotropic thermal conduction of aligned single-walled and multi-walled carbon..., Duong [/bib_ref]. As presented in , the Keff of the SWNT/WS2/PEEK composites decreases with the increase of the
## R
, heat prefers to transfer through SWNT-SWNT contacts, thus the SWNT bundles would be more effective heat transfer channels than the mono-dispersed SWNTs. It can be inferred that the detrimental effect of the SWNT bundles could be reduced by decreasing the SWNT SWNT R to be less than c R , which may be achieved by covalent functionalization to enhance the phonon coupling and weaken the phonon scattering at a SWNT-SWNT interface [bib_ref] Modeling of thermal conductance at transverse CNT-CNT interfaces, Varshney [/bib_ref] [bib_ref] Transversal thermal transport in single-walled carbon nanotube bundles: Influence of axial stretching..., Gharib-Zahedi [/bib_ref].
## Model of graphene/pmma with complex structure
In the graphene-based polymer composites, graphene sheets have a large size distribution in the x-y plane (50-500 nm) due to the oxidation and reduction of graphene sheets during the composite fabrication [bib_ref] Graphene-multilayer graphene nanocomposites as highly efficient thermal interface materials, Shahil [/bib_ref]. The existing models commonly ignore the size distribution of graphene sheets, resulting in an inaccurate prediction of the thermal conductivity. In the developed approach, one can take into account the size distribution (various length, width and thickness), volume fraction and orientation of graphene sheets, as well as the interfacial thermal resistance at the graphene-polymer interface.
In this subsection, a graphene/PMMA composite model was built to validate that the develop approach can be applied to study graphene/polymer composites. Graphene sheets with various length (50-500 nm), width (50-500 nm), thickness (2.4-9.0 nm) and different orientations (parallel, random and perpendicular to the heat flux) were generated in the model.
The developed graphene/PMMA model was validated by comparing the predicted K eff with the measured results, as shown in. As the interfacial thermal resistance (R bd ) at the graphene-PMMA interface was an input value, the model was validated as follows: Estimate the R bd by matching the simulated K eff with the measured value for one composite, and then use this estimated R bd as input to calculate the K eff of other composites. The predicted K eff showed a good agreement with the experimental data, validating the developed model. The R bd was estimated to be 1.906ˆ10´8 m 2¨K /W. For comparison, a modified effective medium theory (EMT) was used to calculate the K eff of the graphene/PMMA composites, and the results are presented in. The estimated K eff from the modified EMT is much higher than the experimental results, and even higher than the predicted K eff of the model with parallel graphene. This is likely because the modified EMT cannot take into account the size distribution of graphene sheets. In the modified EMT, the length and width were treated as infinite when compared with the thickness of graphene sheets, which failed to take into account the graphene sheet with relatively short length and width (e.g.,~50 nm) [bib_ref] Electronic transport properties of individual chemically reduced graphene oxide sheets, Gomez-Navarro [/bib_ref] [bib_ref] Evaluation of solution-processed reduced graphene oxide films as transparent conductors, Becerril [/bib_ref] [bib_ref] Strong, conductive, lightweight, neat graphene aerogel fibers with aligned pores, Xu [/bib_ref]. It should be noted that if given the K eff of a specific composite, the interfacial thermal resistance between the nanofillers and the matrix can be estimated by using the developed approach and an inverse calculation procedure.
Rbd as input to calculate the Keff of other composites. The predicted Keff showed a good agreement with the experimental data, validating the developed model. The Rbd was estimated to be 1.906 × 10 −8 m 2 ·K/W. For comparison, a modified effective medium theory (EMT) was used to calculate the Keff of the graphene/PMMA composites, and the results are presented in. The estimated Keff from the modified EMT is much higher than the experimental results, and even higher than the predicted Keff of the model with parallel graphene. This is likely because the modified EMT cannot take into account the size distribution of graphene sheets. In the modified EMT, the length and width were treated as infinite when compared with the thickness of graphene sheets, which failed to take into account the graphene sheet with relatively short length and width (e.g., ~50 nm) [bib_ref] Electronic transport properties of individual chemically reduced graphene oxide sheets, Gomez-Navarro [/bib_ref] [bib_ref] Evaluation of solution-processed reduced graphene oxide films as transparent conductors, Becerril [/bib_ref] [bib_ref] Strong, conductive, lightweight, neat graphene aerogel fibers with aligned pores, Xu [/bib_ref]. It should be noted that if given the Keff of a specific composite, the interfacial thermal resistance between the nanofillers and the matrix can be estimated by using the developed approach and an inverse calculation procedure. More details of the experimental set-up can be found in Reference [bib_ref] Advanced multifunctional graphene aerogel-Poly(methyl methacrylate) composites: Experiments and modeling, Fan [/bib_ref] ; (b) the thermal conductivity of GA-PMMA composites as a function of graphene volume fraction. The interfacial thermal resistance between graphene sheets and PMMA was estimated to be R bd = 1.906ˆ10´8 m 2¨K /W in the developed model. The same value was utilized in the composites with parallel-and perpendicular-oriented graphene. In the modified EMT, the utilized thermal conductivity of graphene and the R bd of graphene-PMMA were 100 W/m¨K and 1.0ˆ10´8 m 2¨K /W, respectively. The error bars represent the standard deviation of the results from separate measurements of thermal conductivity of graphene/PMMA composites. Reproduced with permission from [bib_ref] Advanced multifunctional graphene aerogel-Poly(methyl methacrylate) composites: Experiments and modeling, Fan [/bib_ref]. Copyright Elsevier, 2015.
# Conclusions
In summary, the developed off-lattice Monte Carlo approach has proved to be accurate as a computational model for heat transfer phenomena and heat transfer mechanisms for multiphase composites with complex structures. The developed approach not only provides a more accurate method to predict the K eff of the multiphase composites than existing EMT models, but also offers an effective computational approach to estimate the interfacial thermal resistance between the nanofillers and the matrix. The quantitative findings presented herein showed that multiphase composites with higher K eff can be obtained by (a) reducing the interfacial thermal resistances at filler-matrix interfaces; (b) aligning the fillers along the heat flux direction; (c) using fillers with larger interfacial area; and (d) improving the dispersion of fillers to be more uniform rather than forming bundles. Through proper modifications of the geometry and thermal properties of the components in the model, the developed approach may be applied to study the thermal transport properties of other multiphase systems, such as CNT/graphene/polymer composites, graphene stabilized polymer blends, CNT stabilized emulsions and other multiphase organic or inorganic composites.
[fig] Figure 1: A schematic plot of the SWNT/WS2/PEEK model: (a) a WS2 nanoparticle (110 nm diameter) [/fig]
[fig] Figure 2, Figure 2: Validation of the developed approach by comparing the simulation results with the experimental data from Reference[9]. The side lengths of the PEEK cubes in 0.were 760, 925 and 1580 nm, respectively. The interfacial thermal resistance at SWNT-PEEK interface was used as 1.0 × 10 −8 m 2 ·K/W for Nan et al.'s effective medium theory (EMT)[32]. The error bars represent the standard deviation of the results obtained from 3 Validation of the developed approach by comparing the simulation results with the experimental data from Reference[9]. The side lengths of the PEEK cubes in 0.were 760, 925 and 1580 nm, respectively. The interfacial thermal resistance at SWNT-PEEK interface was used as 1.0ˆ10´8 m 2¨K /W for Nan et al.'s effective medium theory (EMT)[32]. The error bars represent the standard deviation of the results obtained from 3 separate simulations with different distribution of SWNTs. Reproduced with permission from[18]. Copyright Elsevier, 2014. [/fig]
[fig] Figure 3 .: Effects of the interfacial thermal resistances at (a) SWNT-PEEK and (b) WS2-PEEK interfaces on the Keff of SWNT/WS2/PEEK composites. The 0.5/0.5/99.0 composition was used for this quantitative study. The models with different SWNT orientation (e.g., SWNTs parallel to the heat flux, SWNTs randomly orientated to the heat flux, and SWNTs perpendicular to the heat flux) were built to study the effect of SWNT orientations. The error bars represent the standard deviation of the results from 3 separate simulations with different distribution of SWNTs. Reproduced with permission from[18]. Copyright Elsevier, 2014.As shown inFigure 3, the Keff of SWNT/WS2/PEEK composites decreases with both InFigure 3b, compared with that in the parallel SWNT case, the effect ofPEEK -WS 2 R on the Keff of SWNT/WS2/PEEK composites is much weaker in the random and perpendicular SWNT cases. PEEK -SWNT R significantly impedes the transfer of heat between SWNTs and the PEEK matrix, weakening the enhancement of SWNTs on the Keff. [/fig]
[fig] Figure 3: Effects of the interfacial thermal resistances at (a) SWNT-PEEK and (b) WS 2 -PEEK interfaces on the K eff of SWNT/WS 2 /PEEK composites. The 0.5/0.5/99.0 composition was used for this quantitative study. The models with different SWNT orientation (e.g., SWNTs parallel to the heat flux, SWNTs randomly orientated to the heat flux, and SWNTs perpendicular to the heat flux) were built to study the effect of SWNT orientations. The error bars represent the standard deviation of the results from 3 separate simulations with different distribution of SWNTs. Reproduced with permission from[18]. Copyright Elsevier, 2014. [/fig]
[fig] Figure 4: Effects of (a) length and (b) diameter of SWNTs on the Keff of SWNT/WS2/PEEK composites. [/fig]
[fig] Figure 5: Schematic plot of the SWNT/WS2/PEEK model with SWNT bundles: (a) A WS2 nanoparticle (painted red) with 110 nm diameter is located in the center of a PEEK cube (925 × 925 × 925 nm 3 ). A total of 960 SWNTs (2 nm diameter and 500 nm length, painted black) are randomly dispersed in the model, forming 45 bundles with 20 SWNTs in each bundle and 60 unbundled SWNTs. Constant heat flux is applied along × direction; (b) composite with individual SWNTs and SWNT bundles oriented parallel to the heat-flux direction. Reproduced with permission from [24]. Copyright American Chemical Society, 2015. 3.2.1. Effects of the Morphology of SWNT Bundles on the Keff of SWNT/WS2/PEEK Composites The number of SWNT bundles was changed from 0 to 48 (no individual SWNTs) to investigate the effect of SWNT dispersion state on the Keff of the SWNT/WS2/PEEK composites, and the results are presented in [/fig]
[fig] Figure 6: perpendicular to the heat flux, the number of SWNTs in each bundle has no apparent effect on the Keff of the composites. Effects of the morphology of SWNT bundles on the Keff of the SWNT/WS2/PEEK composites: (a) bundle number; and (b) the number of individual SWNTs in per bundle. The results for SWNTs with different orientations (parallel, random and perpendicular to the heat flux direction) are all presented. The error bars represent the standard deviation of the results obtained from 3 separate simulations with different distribution of SWNTs and SWNT bundles. Reproduced with permission from [24]. Copyright American Chemical Society, 2015. 3.2.2. Effects of the SWNT-SWNT Thermal Resistance on the Keff of SWNT/WS2/PEEK Composites When SWNTs aggregate into bundle structures, SWNT-SWNT thermal resistance ( SWNT SWNT R ) [/fig]
[fig] Figure 7, Figure 7: Effect of the SWNT-SWNT thermal resistance on the Keff of the SWNT/WS2/PEEK composites with 12-48 SWNT bundles. The SWNTs were randomly distributed in the composites. The individual Effect of the SWNT-SWNT thermal resistance on the K eff of the SWNT/WS 2 /PEEK composites with 12-48 SWNT bundles. The SWNTs were randomly distributed in the composites. The individual SWNT number in each bundle was kept at 20. The critical TBR (dashed line) was estimated to be R c = 0.155ˆ10´8 m 2¨K /W by intersecting the K eff curves of different SWNT bundles, as shown in the insert figure. Reproduced with permission from[24]. Copyright American Chemical Society, 2015. [/fig]
[fig] Figure 8: (a) Validation of the developed graphene/PMMA model by comparing the simulation results with the experimental data for different volume fractions of graphene sheets (0.67%, 1.34%, 2.01% and 2.50%). The insert figure is the set-up scheme of the comparative infrared microscopy technique for measuring the thermal conductivity. More details of the experimental set-up can be found in Reference[4]; (b) the thermal conductivity of GA-PMMA composites as a function of graphene volume fraction. The interfacial thermal resistance between graphene sheets and PMMA was estimated to be Rbd = 1.906 × 10 −8 m 2 ·K/W in the developed model. The same value was utilized in the composites with parallel-and perpendicular-oriented graphene. In the modified EMT, the [/fig]
[bib_ref] Thermal properties of graphene and few-layer graphene: Applications in electronics, Yan [/bib_ref] |
Cocultures of human colorectal tumor spheroids with immune cells reveal the therapeutic potential of MICA/B and NKG2A targeting for cancer treatment
Background: Immunotherapies still fail to benefit colorectal cancer (CRC) patients. Relevant functional assays aimed at studying these failures and the efficacy of cancer immunotherapy in human are scarce. 3D tumor cultures, called tumor organoids or spheroids, represent interesting models to study cancer treatments and could help to challenge these issues.Methods: We analyzed heterotypic cocultures of human colon tumor-derived spheroids with immune cells to assess the infiltration, activation and function of T and NK cells toward human colorectal tumors in vitro.Results: We showed that allogeneic T and NK cells rapidly infiltrated cell line-derived spheroids, inducing immunemediated tumor cell apoptosis and spheroid destruction. NKG2D, a key activator of cytotoxic responses, was engaged on infiltrating cells. We thus assessed the therapeutic potential of an antibody targeting the specific ligands of NKG2D, MICA and MICB, in this system. Anti-MICA/B enhanced immune-dependent destruction of tumor spheroid by driving an increased NK cells infiltration and activation. Interestingly, tumor cells reacted to immune infiltration by upregulating HLA-E, ligand of the inhibitory receptor NKG2A expressed by CD8 and NK cells. NKG2A was increased after anti-MICA/B treatment and, accordingly, combination of anti-MICA/B and anti-NKG2A was synergistic. These observations were ultimately confirmed in a clinical relevant model of coculture between CRC patients-derived spheroids and autologous tumor-infiltrating lymphocytes.Conclusions: Altogether, we show that tumor spheroids represent a relevant tool to study tumor-lymphocyte interactions on human tissues and revealed the antitumor potential of immunomodulatory antibodies targeting MICA/B and NKG2A.
# Background
T cells and Natural Killer (NK) cells are major effectors of antitumor immune responses [bib_ref] Natural killer cell-mediated immunosurveillance of human cancer, Malmberg [/bib_ref] [bib_ref] An early history of T cell-mediated cytotoxicity, Golstein [/bib_ref]. Immunotherapies aimed at enhancing their activity have now reached bedside, with some outstanding effects reported in melanoma, breast or kidney cancers [bib_ref] T-cell and NK-cell infiltration into solid tumors: a key limiting factor for..., Melero [/bib_ref] [bib_ref] The future of immune checkpoint therapy, Sharma [/bib_ref]. Nevertheless, large proportions of patients fail to respond to these treatments [bib_ref] Resistance mechanisms to immune-checkpoint blockade in cancer: tumorintrinsic and -extrinsic factors, Pitt [/bib_ref] while other tumor types such as colorectal (CRC) and pancreatic cancers remain poorly responsive to immunomodulation [bib_ref] The immune revolution in gastrointestinal Tumours: leading the way or just following?, Puzzoni [/bib_ref]. There is thus a need to study the mode of action of existing immunotherapeutic agents and to dissect the cellular and molecular events underlying failures. This will allow the development of new immunomodulators for non-responsive cancers.
One of the challenges resides in the development of relevant models to human pathologies. Indeed, mice models hardly mimic treatment resistance observed in patients and human models mainly consist of monolayer cultures that poorly recapitulate the complex features of tumor environment [bib_ref] Intrinsic resistance of solid tumors to immune checkpoint blockade therapy, Zhao [/bib_ref] [bib_ref] In vitro tumor models: advantages, disadvantages, variables, and selecting the right platform, Katt [/bib_ref]. Recently, three-dimensional tumor cultures, called tumor organoids or spheroids, have been developed from tumor cell lines or primary tumor samples cultured in non-adherent conditions [bib_ref] In vitro tumor models: advantages, disadvantages, variables, and selecting the right platform, Katt [/bib_ref] [bib_ref] Long-term expansion of epithelial organoids from human colon, adenoma, adenocarcinoma, and Barrett's..., Sato [/bib_ref] [bib_ref] Dangles-Marie V. Spherical cancer models in tumor biology, Weiswald [/bib_ref] [bib_ref] Modeling development and disease with organoids, Clevers [/bib_ref]. Many studies have highlighted the similarities of these spheroids to human tumors regarding their necrotic and proliferative zones, physicochemical parameters or maintenance of mutated pathways [bib_ref] Preserved genetic diversity in organoids cultured from biopsies of human colorectal cancer..., Weeber [/bib_ref] [bib_ref] Characterization of genetic intratumor heterogeneity in colorectal cancer and matching patient-derived spheroid..., Árnadóttir [/bib_ref] [bib_ref] Unsuccessful mitosis in multicellular tumour spheroids, Molla [/bib_ref] [bib_ref] Cancer cells growing on perfused 3D collagen model produced higher reactive oxygen..., Liu [/bib_ref]. Consequently, spheroids are now widely used to study the efficacy of cytotoxic treatments in various cancer types [bib_ref] Opportunities and challenges for use of tumor spheroids as models to test..., Mehta [/bib_ref] [bib_ref] Shortterm spheroid culture of primary colorectal cancer cells as an in vitro..., Jeppesen [/bib_ref] [bib_ref] A novel three-dimensional heterotypic spheroid model for the assessment of the activity..., Herter [/bib_ref] [bib_ref] Personalized Cancer medicine: an organoid approach, Es [/bib_ref].
Interactions of tumor spheroids with T and NK cells started to be explored through heterotypic cocultures [bib_ref] In vitro" 3D models of tumor-immune system interaction, Hirt [/bib_ref]. These studies brought information about NK [bib_ref] Cetuximab reconstitutes pro-inflammatory cytokine secretions and tumor-infiltrating capabilities of sMICA-inhibited NK cells..., Klöss [/bib_ref] [bib_ref] Cytotoxicity and infiltration of human NK cells in in vivo-like tumor spheroids, Giannattasio [/bib_ref] [bib_ref] Umbilical cord blood CD34+ progenitor-derived NK cells efficiently kill ovarian cancer spheroids..., Hoogstad-Van Evert [/bib_ref] [bib_ref] Activated human primary NK cells efficiently kill colorectal cancer cells in 3D..., Lanuza [/bib_ref] [bib_ref] A novel three-dimensional immune oncology model for high-throughput testing of tumoricidal activity, Sherman [/bib_ref] and T cells [bib_ref] A three-dimensional tumor cell defect in activating autologous CTLs is associated with..., Dangles-Marie [/bib_ref] [bib_ref] A novel mechanism for anti-EGFR antibody action involves chemokinemediated leukocyte infiltration, Hoffmann [/bib_ref] [bib_ref] Combination therapy with bispecific antibodies and PD-1 blockade enhances the antitumor potency..., Chang [/bib_ref] [bib_ref] The influence of stromal cells and tumor-microenvironment-derived cytokines and chemokines on CD3+CD8+..., Koeck [/bib_ref] [bib_ref] Increasing tumorinfiltrating T cells through inhibition of CXCL12 with NOX-A12 synergizes with..., Zboralski [/bib_ref] [bib_ref] CDK4/6 inhibition augments antitumor immunity by enhancing T-cell activation, Deng [/bib_ref] [bib_ref] Ex vivo profiling of PD-1 blockade using organotypic tumor spheroids, Jenkins [/bib_ref] capacity to infiltrate and kill tumor spheroids in various context. Yet, they did not deepen the characterization of the infiltrating cells, the mechanisms of tumor-lymphocytes interactions through activating or inhibitory pathways, nor the impact of immune modulators on tumor spheroids fate.
NK cell receptors provide activating or inhibitory signals to induce direct cytotoxicity, antibody-dependent cellular cytotoxicity (ADCC) or tolerance directed against target cells. Among them, NKG2D and NKG2A are expressed by NK cells, significant proportions of CD8 T cells and subsets of CD4 T cells [bib_ref] CD4 +NKG2D+ T cells in Crohn's disease mediate inflammatory and cytotoxic responses..., Allez [/bib_ref] [bib_ref] Phenotype of NK-like CD8(+) T cells with innate features in humans and..., Barbarin [/bib_ref]. Interactions between NKG2D and its ligands MICA/B activate NK cell cytotoxicity and have been reported to be an important costimulatory signal for T cells [bib_ref] Roles of the NKG2D immunoreceptor and its ligands, Raulet [/bib_ref] [bib_ref] NKG2D in NK and T cell-mediated immunity, Ogasawara [/bib_ref]. NKG2A binding to the MHC-like molecule HLA-E provides powerful inhibitory signaling in both T and NK cells [bib_ref] The role of CD94/NKG2 in innate and adaptive immunity, Gunturi [/bib_ref] [bib_ref] Clinical and immunological significance of HLA-E in stem cell transplantation and cancer, Wieten [/bib_ref]. NKG2D-MICA/ B and NKG2A-HLA-E pathways have been reported by our group and others as engaged in CRC and could be a potential immunotherapy for patients with CRC [bib_ref] NKG2D ligand expression in human colorectal cancer reveals associations with prognosis and..., Mcgilvray [/bib_ref] [bib_ref] Possible association of decreased NKG2D expression levels and suppression of the activity..., Shen [/bib_ref] [bib_ref] Phenotypic analysis of T cells infiltrating colon cancers: correlations with oncogenetic status, Chirica [/bib_ref] [bib_ref] Human leukocyte antigen-E protein is overexpressed in primary human colorectal cancer, Levy [/bib_ref] [bib_ref] Loss of Fas expression and high expression of HLA-E promoting the immune..., Huang [/bib_ref].
We therefore developed cocultures of CRC tumors and immune cells to study the infiltration, activation and function of immune cells toward human tumors with a particular focus on NKG2D and NKG2A pathways. We set up this model in allogeneic conditions, using healthy donor blood cells (HD PBMCs) and HT29 tumor cell line, and more importantly in autologous conditions using primary tumor-derived spheroids and tumor infiltrating lymphocytes (TILs) from the same patients.
We showed that activated/memory T and NK cells were able to infiltrate spheroids, kill tumor cells and disrupt the three-dimensional structure. Both NKG2D-MICA/B and NKG2A-HLA-E pathways were involved in these processes. Accordingly, we showed that anti-MICA/B antibodies efficiently stimulated antitumor responses. While NKG2A blockade did not show significative impact alone, combination of both MICA/B and NKG2A targeting was synergistic, inducing a strong immune-dependent destruction of patients-derived spheroids during cocultures with autologous TILs.
# Methods
## Cell line
Mycoplasma-free HT-29 and DLD1 tumor cell lines were obtained from ATCC (cat. HTB-38 and CCL-221) and cultured in HEPES-containing RPMI 1640 (Thermo-Fisher) complemented with 10% FCS (Eurobio), 50 U/ mL penicillin, 50 μg/mL streptomycin, 2 mM GlutaMAX and 1 mM Sodium Pyruvate (all from ThermoFisher), thereafter named complete RPMI.
## Healthy donors (hd) blood cells
Heparinized venous blood was collected from healthy donors, diluted 1:2 with PBS and then layered on a density gradient (Lymphocytes separation medium, Eurobio). Peripheral blood mononuclear cells (PBMCs) were collected from the interface after centrifugation, washed with PBS, and resuspended in complete RPMI.
## Cell sorting
T and NK cells were enriched from HD PBMCs by magnetic cell depletion of B cells and monocytes using anti-CD19 and anti-CD14 microbeads with the MACS technology (Miltenyi), according to the manufacturer procedures. Particular CD4, CD8, and NK cells enrichments were respectively achieved by positively sorting CD4+, CD8+ and CD8-CD56+ cells from HD PBMCs with the MACS technology.
## Coculture protocol
HT29 (or DLD1) spheroids were generated by seeding 10 4 cells per well on Nunclon Sphera (ThermoFisher) or Costar ultra-low attachment (Corning) round bottom 96 wells plates in complete RPMI. 5 days later, spheroids contained 3. 10 4 cells and cocultures were started by adding 3.10 5 total or CD19 − CD14 − sorted HD PBMCs per well, together with stimulatory or inhibitory molecules. In , 3.10 4 cells of each subset were added with the spheroids to perform cocultures. For flow cytometry analyses, 6 wells per condition were seeded. OUT and IN compartments were isolated by first pooling the 6 cocultures wells in eppendorf tubes. Spheroids were gently resuspended and left to sediment to the bottom of the eppendorf. Supernatant cell suspension constituted the non-infiltrating immune cells (=OUT). These steps were repeated 2 times with PBS in order to wash the spheroids from the non-infiltrating immune cells. Spheroids were then trypsinized to obtain a single cell suspension (=IN) further analyzed by flow cytometry.
## Cytokines and functional antibodies
The following commercial stimulatory or inhibitory molecules were used in our cocultures: Interleukin 15 (IL-15, used at 10 ng/mL, Miltenyi), anti-IFNγ blocking antibodies (B27 clone, used at 2 μg/mL, BD Biosciences) and anti-NKG2D blocking antibodies (1D11 clone, used at 5 μg/mL, BD Biosciences). Anti-NKG2A (monalizumab, hIgG4) and ADCC-enhanced anti-MICA/B (IPH4301, hIgG1) were provided by Innate Pharma along with corresponding isotype controls.
## Antibodies and flow cytometry analyses
Cells were stained with saturating amounts of various fluorescent-labeled antibody combinations including anti-EpCAM (EBA-1 clone), anti-CD45 (HI30 clone), anti-CD3e (UCHT1 clone), anti-CD56 (B159 clone), anti-CD4 (SK3 clone), anti-CD8 (SK1 clone), anti-CD25 (M-A215 clone), anti-CD107a (H4A3 clone), anti-CD45RO (REA611 clone), anti-NKG2A (REA110 clone), anti-NKG2D (BAT221 clone), anti-CD137 (4B4-1 clone), anti-CD16 (3G8 clone), anti-HLA-E (3D12 clone), Annexin V and co-stained with DAPI (all from BD Biosciences or Miltenyi). Cells were analyzed with the Attune NxT flow cytometer (Life Technologies) and further analyses were performed with FlowJo software (Tree Star).
## Spheroid volume calculation
Before trypsinization, pooled spheroids were placed in 96 wells plate and pictured using the EVOS FLc microscope at a 2x to 4x magnification. Images were then analyzed using the Icy software by measuring the length (L) and width (W) of each spheroid. Spheroid volumes were then calculated as follows: V = (L x W x W) / 2.
## Live imaging of spheroid apoptosis
This was achieved by adding CellEvent caspase-3/− 7 green dye (ThermoFisher) in the coculture wells and imaging green fluorescence across time using the Incucyte S3 system (Essen BioScience). Single pictures of each well were acquired every hour during 1 week and analyzed using Incucyte S3 software.
## Immunofluorescence (if)
Before starting the cocultures, CD19 − CD14 − sorted HD PBMCs were stained with CFSE (ThermoFisher) according to manufacturer procedure. Infiltrated spheroids were isolated at 24 h and embedded in Tissue-Tek OCT compound (Sakura). Spheroid blocs were then 20 μm-thick sliced using CM1520 cryostat (Leica), the slides were mounted with DAPI-containing Fluoromount-G (Thermofisher) and imaged at a 5x magnification using an epifluorescence microscope (Axio Imager 2, Zeiss). Quantification of CFSE+ cells infiltrating the spheroids was done using the H-K means plugin of the Icy software.
## Immunohistochemistry (ihc)
The formalin-fixed spheroids were first embedded in Histogel (Thermo Scientific) and then in paraffin. Blocks were sliced in 5 μm-thick sections and immunostainings performed on a Discovery Ultra automaton (Ventana). After pre-treatment with cell conditioning 1 (Ventana), sections were incubated 1 h at 37°C with anti-MICA/B (clone MIA4, Innate Pharma) or anti-HLA-E (clone MEM-E/02, Exbio) primary antibodies at 2 μg/mL and 1 μg/mL, respectively. Anti-mouse IgG detection system (discovery OmniMap anti-mouse HRP, Ventana) was used for HLA-E staining and an additional amplification step using tyramide was used for MICA/B staining (discovery Amp HQ kit, Ventana). After revelation with 3,3-diaminobenzidine and counterstaining with hematoxylin, sections were washed, dehydrated, cleared and mounted using a coverslipper (ClearVue, Thermo Scientific). Stained sections were finally scanned on a slide scanner (S60 Nanozoomer, Hamamatsu).
## Colorectal cancer patients (crc) samples
Between February 2017 and February 2018, 41 patients who underwent resection of colon cancer at the Saint Louis Hospital in Paris were prospectively included. This study was approved by the French ethical committee (approval n°2016/45), and all subjects gave written informed consent.
Patients' blood was manipulated the same way than HD blood (see above). Tumor samples were washed with PBS (ThermoFisher) and cut into pieces of 2 to 3 mm. Fragments were incubated for 20 min at room temperature with an antibiotic cocktail containing fungizone, normocin and gentamicin (all from ThermoFisher) to avoid contamination and then enzymatically digested for a total of 21 min at 37°C in complete RPMI containing collagenase IV (Sigma-Aldrich) and DNase I (Roche). The supernatant was then filtered and the immune cells extracted with lymphocyte separation medium (Eurobio). Resulting tumor-infiltrating lymphocytes (TILs) were either directly analyzed by flow cytometry or cultured in complete RPMI complemented with IL-15 and IL-7 (10 ng/mL, both from Mitenyi) in order to keep them alive before coculture starting with autologous tumor spheroids.
Generation of patient-derived tumor spheroids was achieved by first culturing digested and filtered tumor in classic adherent culture flasks (BD Falcon) in complete RPMI. Cells were then washed every 2 days and cultured for up to 10 days. This allowed the isolation of adherent cells that were then trypsinized and seeded in ultra-low attachment 96 wells plates to create a unique uniform spheroid in each well. Cell numbers and subsequent setup and analyses of autologous cocultures were done using the same protocol than cocultures between HT29 cell line and HD PBMCs.
## Statistical analyses
Data are expressed as Mean ± SEM. Statistically significance of differences was analyzed using GraphPad Prism 7 (GraphPad Software, La Jolla, USA) by paired Student's t test, two-way ANOVA or Wilcoxon matched-pairs signed rank test when appropriate. A p value < 0.05 was considered as statistically significant.
# Results
## Activated/memory t cells and nk cells infiltrate colon cancer cell line-derived spheroids
We generated colon cancer spheroids from HT29 cell line that we cocultured with peripheral blood immune cells from healthy donors (HD PBMCs), depleted of B cells and monocytes in order to enrich for T and NK cells. After coculture, infiltrating cells (IN) and cells remaining in the medium (OUT) were mechanically separated and analyzed [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref].
To assess whether immune cells entered in the spheroids or remained in the culture medium, we analyzed cocultured spheroids by immunofluorescence (IF) and observed a deep and homogeneous infiltration of pre-stained immune cells [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref] , upper panels). By flow cytometry, we could detect infiltrating cells 24 h after coculture initiation [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref] , reaching about 1500 total immune cells per spheroid. Comparing cellular proportions IN and OUT of the spheroids, we observed lower T cell and higher NK cell proportions in the tumor structure [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref] , upper panels). We also found a reversal of CD4 to CD8 subpopulations ratio in the spheroids [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref] , bottom panels), pointing that NK and CD8 T cells could have a particular advantage for spheroid infiltration.
Infiltrating CD8 T cells showed increased proportions of CD45RO+ memory cells. Additionally, NK and CD8 T cells displayed increased expression of the activation marker CD25 and of the degranulation marker CD107a in the spheroids compared to OUT cells [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref]. Conversely, CD4 T cells phenotype did not seem to be impacted by spheroid infiltration, with no change in memory or activation markers observed inside the spheroids. CD4 Tregs proportions were found unchanged as well in this context (Additional file 1: [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref]. These results suggest that memory CD8 T cells and NK cells are prone to infiltrate and get activated in spheroids.
## Immune stimulation increases ht29 spheroid infiltration
We tested the possibility to modulate the immune response in this system by adding IL-15 to the cocultures. IL-15 induced a strong increase of spheroid infiltration by both T and NK cells [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref] , reaching up to 4500 total immune cells infiltrated per spheroid. It also increased the proportions of infiltrating CD8 T cells [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref] and enhanced the activation of all infiltrating cells, as witnessed by increased proportions of CD45RO + memory CD4 and CD8 T cells and increased expression of CD25 and CD107a by T and NK cells in the spheroids [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref]. This shows that known immune mediators, such as IL-15, can modulate the activity of T and NK cells within HT29 spheroids.
## Spheroid infiltration leads to tumor cell apoptosis and spheroid destruction
We explored the functional consequences of this immune cell infiltration and activation in spheroids using several methods. We first examined spheroid integrity by monitoring their volume through picture-based measurements . HT29 spheroids cocultured with T and NK cells-enriched HD PBMCs were smaller than control spheroid at 48 h, suggesting that immune cells disrupted the tumor structure . To examine the cause of such disruption, we assessed the global activation of caspase-3 and -7 in the spheroids by live-imaging and Additional file 2: Movie S1, Additional file 3: Movie S2, Additional file 4: Movie S3) and quantified Annexin V and DAPI staining of tumor cells by flow cytometry . These complementary analyses showed that spheroid destruction was correlated to an active apoptosis process in tumor cells, involving caspase-3 and -7 activation, which resulted in the accumulation of AnnexinV+DAPI+ apoptotic tumor cells.
IL-15 enhanced these apoptotic processes and induced an almost complete spheroid destruction at 48 h . Conversely, IFNγ blockade in our system resulted in decreased spheroid destruction, tumor cell apoptosis and spheroid infiltration by immune cells compared to control conditions (Additional file 1: . Similar results were obtained using another colon cell line, DLD-1 (data not shown). Together, these results show that infiltration of tumor spheroid by activated/memory immune cells triggers tumor cell killing and spheroid destruction, which can be enhanced or dampened by immune modulation.
## Cd4 t cells, cd8 t cells and nk cells are self-sufficient to exert antitumor effects during the cocultures
To test the intrinsic capacity of CD4 T cells, CD8 T cells and NK cells to react toward tumor spheroids we purified and cocultured them, alone or combined, with HT29 tumor spheroids. We observed that the three cell types were self-sufficient to disrupt the three-dimensional structure, induce tumor cell apoptosis and infiltrate tumor spheroids . In this context, infiltrating cells showed increased levels of CD25 and CD107a compared to cells outside the spheroid, as well as higher proportion of CD45RO expression in T cells . This demonstrates the capacity of the three cell types to react to the tumor cells upon infiltration, and confirms our previous observations of an increased infiltration capacity of memory T cells as compared to naïve T cells.
Overall, we also observed that NK cells were prompter than T cell subsets to infiltrate spheroids and kill tumor cells, but also that combination of the three cell types had an additive impact on the particular activation of each of them. This implies a possible cellular cooperation process between T cells and NK cells to destroy tumor spheroids in our model.
Adding of IL-15 during these cocultures then allowed us to refine our previous conclusions by showing that IL-15-mediated effects were mainly driven by NK cells
## Nkg2d-mica/b pathway is engaged during the cocultures
We observed that NKG2D expression was increased on infiltrating CD4 T cells, unchanged on CD8 T cells and decreased on NK cells [fig_ref] Figure 4: NKG2D-MICA/B pathway is engaged during the cocultures [/fig_ref] , advocating for an engagement of NKG2D in the spheroids [bib_ref] Phenotypic analysis of T cells infiltrating colon cancers: correlations with oncogenetic status, Chirica [/bib_ref]. In parallel, immunohistochemistry (IHC) analyses showed strong expression of NKG2D ligand MICA/B on tumor cells in both control and cocultured spheroids [fig_ref] Figure 4: NKG2D-MICA/B pathway is engaged during the cocultures [/fig_ref].
We tested the functional relevance of NKG2D-MICA/ B pathway by using anti-NKG2D blocking antibodies in the cocultures. Spheroid infiltration by immune cells as well as spheroid destruction and tumor cell apoptosis were significantly decreased in this condition [fig_ref] Figure 4: NKG2D-MICA/B pathway is engaged during the cocultures [/fig_ref] , d and e, respectively), suggesting that MICA/B are major targets of cytotoxic cells infiltrating spheroids.
Anti-MICA/B antibodies exert antitumor effects MICA/B expression by tumor cells and NKG2D pathway engagement prompted us to test the antitumor potential of an ADCC-enhanced anti-MICA/B antibody [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref]. Compared to its control isotype, anti-MICA/B induced an increased immune-dependent spheroid destruction and tumor cell death, as observed by microscopic data, live-imaging and flow cytometry [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref]. No changes were observed in the absence of immune cells (not shown).
This tumor destruction was associated with an increase of overall spheroid infiltration by T and NK cells [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref] , with unchanged proportions of infiltrating T cells [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref] and Additional file 1: and B) but increased proportions of infiltrating NK cells [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref]. Interestingly, NKG2D expression on infiltrating NK cells was partially restored [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref] , confirming that anti-MICA/B also blocked NKG2D interaction with its ligands in this system (Innate Pharma unpublished results). Concurringly, CD8 T cells upregulated NKG2D inside the spheroids during anti-MICA/B treatment (Additional file 1: Anti-MICA/B induced a strong upregulation of the early activation marker CD137 on NK cells [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref] , advocating for their increased stimulation. The Fc receptor CD16 was strongly downregulated at the NK cell surface [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref] , exhibiting the engagement of this receptor during ADCC [bib_ref] Obinutuzumab-mediated high-affinity ligation of FcγRIIIA/CD16 primes NK cells for IFNγ production, Capuano [/bib_ref]. Taken together, these results show that anti-MICA/B antibodies induce a strong NK-mediated immune antitumor response in this spheroid model.
## Nkg2a blockade enhances the antitumor effects of anti-mica/b antibodies
Interestingly, we observed an increased expression of the inhibitory receptor NKG2A on infiltrating NK cells after anti-MICA/B treatment [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref]. This led us to study the expression patterns of NKG2A and its ligand HLA-E during the cocultures. We first observed that NKG2A expression was increased on infiltrating T cells and downmodulated in infiltrating NK cells compared to OUT cells . Tumor cells did not express HLA-E in control spheroids but its expression was strongly induced by the presence of immune cells . Moreover, both NKG2A and HLA-E were significantly enhanced when immune cells were stimulated with IL-15 . This suggests that tumor cells in the spheroids could evade the immune response through NKG2A-HLA-E interactions.
We thus assessed the antitumor potential of an anti-NKG2A blocking antibody. Surprisingly, this antibody used alone did not exert antitumor effect in these conditions when looking at microscopic, live-imaging and flow cytometry readouts .
However, NKG2D and NKG2A pathways showed interesting interactions in our system, prompting us to test the efficacy of anti-MICA/B and anti-NKG2A combination. We observed increased spheroid alteration , tumor cells apoptosis and spheroid infiltration when anti-NKG2A was added to anti-MICA/B, compared to anti-MICA/B alone or isotype controls. No changes were observed in the absence of immune cells (not shown). These results suggest that anti-MICA/B and anti-NKG2A antibodies have a synergistic effect on immune mediated anti-tumor response in this spheroid model. This effect was not linked to changes in T and NK cells proportions (Additional file 1: [fig_ref] Figure 4: NKG2D-MICA/B pathway is engaged during the cocultures [/fig_ref] and B) but associated with an increased expression of CD137 by infiltrating CD8 T cells (Additional file 1: [fig_ref] Figure 4: NKG2D-MICA/B pathway is engaged during the cocultures [/fig_ref]. These results confirm that NKG2A-HLA-E crosstalk is an interesting resistance mechanism to target in CRC and suggest that NKG2A additional blockade could reinforce CD8 T cells activation.
## Nkg2a-hla-e and nkg2d-mica/b pathways are relevant targets in crc patients
Our in vitro model suggests that the NKG2D-MICA/B pathway is a potential target in CRC treatment. Additionally, NKG2A-HLA-E interaction is a functional inhibitory mechanism arising during antitumor immune responses that could be targeted to improve treatment efficacy. To assess the relevance of these pathways in CRC, we investigated the expression of NKG2A and NKG2D on T and NK cells in the blood and the tumor of CRC-bearing patients . We observed that T and NK cell populations presented respectively higher and lower levels of NKG2A in the tumor compared to blood . These results were comparable to what we observed between the OUT and IN compartments in the coculture model [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref] and suggested that NKG2A pathway could be engaged in tumor-infiltrating cells during CRC.
NKG2D expression was also found lower on tumor-infiltrating CD8 T cells in CRC patients compared to blood, as already described [bib_ref] Phenotypic analysis of T cells infiltrating colon cancers: correlations with oncogenetic status, Chirica [/bib_ref]. This was also observed on NK cells , while NKG2D levels were found higher on tumor-infiltrating than on blood CD4 T cells. Intriguingly, these variations of NKG2D expression in patients were similar to those observed in our coculture model . Together, these results pointed to a potential functional relevance of NKG2A and NKG2D receptors in CRC patients.
## Tumor-infiltrating lymphocytes (tils) do not exert antitumor effects against autologous crc patient-derived spheroids without stimulation
To test this hypothesis, we derived primary CRC tumor-spheroids that we cocultured with autologous TILs and Additional file 1: [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref]. In these settings we observed the impact of adding autologous TILs and/or IL-15 on spheroid volume. Interestingly, we showed that adding autologous TILs did not significantly impact the spheroid structure . However, IL-15 together with autologous TILs induced an immune-mediated destruction of tumor In d to f panels, spheroid volumes are normalized to the culture condition without stimulation. g panel represents fold changes between anti-MICA/B and its isotype, and between combination and combination of isotypes. In d to g panels, each patient is represented by a specific symbol. n = 5 independent experiments. Statistical significance of a and b panels was analyzed Wilcoxon matched-pairs signed rank test, the other panels using paired t test (* p < 0.05; ** p < 0.005, *** p < 0.001, **** p < 0.0001) spheroids compared to control condition or IL-15 alone, showing that TILs stimulation can exert antitumor effect in these autologous conditions. We observed no correlation between the magnitude of these effects and respective clinical data nor tumor cell content in the spheroids, as well as TILs composition in T cells and NK cells (Additional file 1: Tables S1 and S2).
Anti-MICA/B exert antitumor effect in autologous cocultures and synergizes with anti-NKG2A to disrupt tumor structure
We then tested the potential of anti-NKG2A and anti-MICA/B antibodies, alone or in combination, in these autologous conditions. Addition of these antibodies did not affect spheroid size in the absence of TILs , except for one patient for which spheroid growth was highly variable when cultured alone. During the cocultures anti-NKG2A alone did not show a reproducible effect while anti-MICA/B, alone or combined with anti-NKG2A, induced a significant immune-mediated decrease of spheroid size compared to isotype control. By calculating the changes in spheroid volume induced by anti-MICA/B alone or combined with anti-NKG2A compared to their respective isotype control conditions, we also showed that anti-NKG2A increased the spheroid destruction induced by autologous TILs in the presence of anti-MICA/B antibodies. These results show that anti-MICA/B antibodies are able to elicit an immune-mediated destruction of CRC patients-derived spheroids in an autologous context, and that additional targeting of NKG2A inhibitory receptor enhances this effect.
# Discussion
In this work, we describe that coculture of tumor spheroid with immune cells is a relevant and powerful tool to study human antitumor immune responses in vitro and ex vivo. We aimed at generating individual uniform spheroids in matrix-and growth factors-free medium to allow a rapid, precise and reproducible manipulation of coculture settings, including effectors to target cells ratio and treatments conditions. These models allowed us to develop innovative in vitro readouts such as measurements of tumor volume, shape and infiltration by immune cells as well as easy characterization of infiltrated cells, which cannot be achieved through classical monolayer culture models. By doing so, we showed that tumor spheroids can be infiltrated by allogeneic activated/ memory T and NK cells which induce tumor cell death and spheroid destruction, notably through IFNγ and NKG2D-MICA/B pathways. The latter has been identified for a long time as an important pathway in antitumor immunity but remained controversial in regard to its pro-or anti-tumorigenic role [bib_ref] NKG2D ligands in tumor immunity: two sides of a coin, Zhang [/bib_ref]. This controversy has been recently clarified by studies that identified NKG2D-MICA/B engagement as a major mechanism of tumor immune surveillance, subverted by tumors through proteolytic shedding of surface MICA/B which at the same time lower membrane-bound MICA/B recognition by NKG2D and induce the saturation of NKG2D with soluble MICA/B [bib_ref] Roles of the NKG2D immunoreceptor and its ligands, Raulet [/bib_ref] [bib_ref] NKG2D ligands in tumor immunity: two sides of a coin, Zhang [/bib_ref]. Recent efforts to disrupt this tolerance mechanisms showed that inhibiting MICA/B shedding was a successful strategy to boost antitumor immune responses [bib_ref] Antibody-mediated inhibition of MICA and MICB shedding promotes NK cell-driven tumor immunity, De Andrade [/bib_ref]. Our results support these concepts, by showing that NKG2D blockade prevents immune infiltration and activation in tumor spheroids and that exploiting MICA/B as a tumor antigen to induce ADCC is a feasible and efficient approach.
Interestingly, our results point at another resistance mechanisms used by tumor cells that try to evade immune recognition. This is illustrated by HLA-E upregulation on tumor cells upon spheroid infiltration, associated with NKG2A increase on infiltrating immune cells. NKG2A-HLA-E pathway has already been described as a potential inhibitor of antitumor immune responses [bib_ref] Loss of Fas expression and high expression of HLA-E promoting the immune..., Huang [/bib_ref] [bib_ref] Effects of anti-NKG2A antibody administration on leukemia and normal hematopoietic cells, Ruggeri [/bib_ref]. In our settings, we observe that NKG2A blockade alone has no impact up to 48 h after coculturing immune cells and tumor spheroids, but that it could serve as a useful combinatorial treatment to avoid NKG2A-mediated tumor resistance to immunotherapy.
Our results are in line with other studies that used tumor spheroids to examine tumor-lymphocytes interactions in various contexts [bib_ref] A three-dimensional tumor cell defect in activating autologous CTLs is associated with..., Dangles-Marie [/bib_ref] [bib_ref] A novel mechanism for anti-EGFR antibody action involves chemokinemediated leukocyte infiltration, Hoffmann [/bib_ref] [bib_ref] Combination therapy with bispecific antibodies and PD-1 blockade enhances the antitumor potency..., Chang [/bib_ref] [bib_ref] The influence of stromal cells and tumor-microenvironment-derived cytokines and chemokines on CD3+CD8+..., Koeck [/bib_ref] [bib_ref] Increasing tumorinfiltrating T cells through inhibition of CXCL12 with NOX-A12 synergizes with..., Zboralski [/bib_ref] [bib_ref] CDK4/6 inhibition augments antitumor immunity by enhancing T-cell activation, Deng [/bib_ref] [bib_ref] Ex vivo profiling of PD-1 blockade using organotypic tumor spheroids, Jenkins [/bib_ref]. The specificity of our model resides in its setup simplicity and high versatility, that allows for multiple culture conditions and readouts to analyze dynamic responses in a controlled environment. To our knowledge, we are also one of the first group reporting the precise flow cytometry phenotyping of T and NK cells infiltrating human tumor spheroids, and the study of new immune modulators on human autologous antitumor responses.
We chose to focus on the response of T cells and NK cells in this context because these cells are the main effectors of antitumor immune response. Nevertheless, we observed that blood monocytes and dendritic cells are also able to efficiently infiltrate HT29 spheroids (data not shown). Heterotypic cocultures of tumor spheroids with other immune cells types could thus permit to expand our knowledge on human antitumor immune responses.
We adapted our allogeneic model to autologous cocultures derived from primary CRC tumor samples to generate a clinically relevant functional assay. Despite interindividual size variability, we were able to grow spheroids on a per-patient basis within 2 weeks. These spheroids were composed of mixed tumor-associated fibroblast and tumor cells (Additional file 1: [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref] , thus mimicking tumor structure. Autologous cocultures with patients-derived spheroids showed that TILs were unable to reproducibly destroy autologous CRC tumors. This was expected, as tumors are edited to develop in front of an activated immune system that becomes unable to eradicate tumor cells [bib_ref] Resistance mechanisms to immune-checkpoint blockade in cancer: tumorintrinsic and -extrinsic factors, Pitt [/bib_ref]. However, we also showed that stimulating these TILs induce an immune response capable of destroying tumor structures. This has been observed using non-specific stimulation with IL-15 and more importantly using specific immune modulatory anti-MICA/B and anti-NKG2A monoclonal antibodies. Hence, spheroid cocultures are functionally relevant to the study of immunotherapies, echoing a recent study from Voest group, and pave the way for the study of anti-MICA/B and anti-NKG2A for cancer treatment. Our results also highlight the powerful antitumor potential of IL-15-based treatments for which cocultured spheroids could help deeply characterize the efficacy and mode of action.
# Conclusions
Overall, we show that human tumor spheroids cocultured with immune cells are relevant in vitro and ex vivo tools for human functional assays. As such, they will be a major addition to our scientific arsenal to study tumor immunology and will also help the scientific community to refine, reduce and complement animal experimentation. This field of research is rather new and thus widely opened to technical and scientific advances. Active research regarding the automation, miniaturization and adaptation of spheroid coculture models to all human tumor types will allow to dynamically study the antitumor immune response arising in patients and ultimately anticipate treatments efficacy in a personalized manner.
## Additional files
Additional file 1: [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref]. Tregs infiltration in HT29 spheroids. Percentages of Foxp3+CD25+ Tregs among CD4+ T cells analyzed in [fig_ref] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor... [/fig_ref] and d. . IFNg blockade decreases spheroid infiltration and destruction by immune cells. (A) Pictures and analyses of (B) spheroid volume, (C) tumor cell apoptosis, and (D) spheroid infiltration 48h after coculturing HT29 spheroids with CD19-CD14-PBMCs in the presence or not of anti-IFNg blocking antibodies. . T cell subsets and NKG2D expression by CD8 T cells after MICA/B treatment. (A) CD4 and (B) CD8 T cells proportions as well as (C) NKG2D expression by CD8 T cells relative to experiments in [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref] to k. [fig_ref] Figure 4: NKG2D-MICA/B pathway is engaged during the cocultures [/fig_ref]. T and NK cells proportions and CD137 expression by CD8 T cells after combination therapy. Proportions of (A) T and NK cells, of (B) CD4 and CD8 T cells subsets and (C) CD137 expression by CD8 T cells relative to to h. [fig_ref] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation... [/fig_ref]. Pictures of primary CRC tumors cultures. Pictures of (A) primary CRC tumor cultured in adherent culture flasks and (B) tumorderived spheroids used in autologous cocultures. . Clinical characteristics of the patients used for autologous cocultures. .
[fig] Figure 1: Allogeneic activated/memory T and NK cells are able to infiltrate HT29 tumor spheroids. a Scheme of the coculture (CC) protocol between HT29 spheroids and CD19-CD14-sorted PBMCs from healthy donors. b Immunofluorescence (n = 2 independent experiments) and flow cytometry (n = 19 independent experiments) analyses of spheroid immune infiltration in the presence or not of IL-15 at 24 h. c Flow cytometry analyses of T and NK cells (respectively gated CD3+ and CD3-CD56+ among live single cells lymphocytes) as well as CD4+ and CD8+ T cells subsets (respectively gated CD4 + CD8-and CD4-CD8+ among CD3+) percentages in the IN and OUT compartments, in the presence or not of IL-15 at 24 h. n = 19 independent experiments. d Flow cytometry analyses of CD25, CD107a and CD45RO expression by CD4+ T cells, CD8+ T cells and NK cells in the IN and OUT compartments in the presence or not of IL-15 at 24 h. n = 8 to 18 independent experiments. Statistical significance of immunofluorescence experiments was analyzed using the Mann-Whitney test, the others using Wilcoxon matched-pairs signed rank test (* p < 0.05; ** p < 0.005, *** p < 0.001, **** p < 0.0001) in this context. However, IL-15 increased tumor cell apoptosis induced by the combination of CD4 and CD8 T cells, implying that IL-15 is not exclusively sensed and active through NK cells. [/fig]
[fig] Figure 2 0, Figure 3: Immune infiltration and activation in HT29 spheroids lead to tumor cell apoptosis and spheroid destruction. HT29 spheroids were cocultured or not with CD19-CD14-PBMCs and IL-15 and we monitored a spheroid volume using microscopic pictures, b the dynamic cleavage of caspase-3/− 7 in the spheroids using Incucyte live imaging system, and c the staining by Annexin V and DAPI dyes of EpCAM+ cells by flow cytometry at 48 h. n = 4 to 12 independent experiments. Statistical significance of A panel was analyzed using 2-way ANOVA, the other panels using the Wilcoxon matched-pairs signed rank test (* p < 0.05; ** p < 0.005, *** p < 0.001, **** p < Specific effects of CD4, CD8 and NK cells cocultured with HT29 spheroids alone or combined. HT29 spheroids were cocultured with sorted CD4 T cells, CD8 T cells and NK cells at a 1:1 ratio, alone or in combination and with or without IL-15. We analyzed a spheroid destruction through analysis of macroscopic pictures and b tumor cell apoptosis using flow cytometry after 48 h as well as c immune infiltration and d expression of CD25, CD107a and CD45RO by CD4 T cells, CD8 T cells and NK cells in IN and OUT compartments after 24 h. n = 4 to 5 independent experiments. Statistical significance was analyzed using the Wilcoxon matched-pairs signed rank test (# p = 0.06, as compared to spheroid alone in A to C panels and to OUT cells in D panels) indicating a possible engagement of NKG2D on CD8 T cells as well. [/fig]
[fig] Figure 4: NKG2D-MICA/B pathway is engaged during the cocultures. a NKG2D expression by CD4 T cells, CD8 T cells and NK cells in the IN and OUT compartments, in the presence or not of IL-15, as measured by flow cytometry at 24 h. n = 18 independent experiments b MICA/B expression by tumor cells in the spheroids cocultured or not with CD19-CD14-PBMCs, as measured by immunohistochemistry at 24 h. Representative pictures of 1 experiment. c to e Analyses of c spheroid volume, d tumor cell apoptosis and e spheroid infiltration 48 h after coculturing HT29 spheroids with CD19-CD14-PBMCs in the presence or not of anti-NKG2D blocking antibodies. n = 3 to 4 independent experiments.Statistical significance of A panel was analyzed using the Wilcoxon matched-pairs signed rank test, C to E panels using paired t test (* p < 0.05; ** p < 0.005, *** p < 0.001, **** p < 0.0001) [/fig]
[fig] Figure 5: Anti-MICA/B induce immune-mediated anti-tumor effects through increased NK cell infiltration and activation in HT29 spheroids. HT29 spheroids were cocultured or not with CD19-CD14-PBMCs in the presence or not of anti-MICA/B antibodies or corresponding control isotype. We analyzed tumor cell death and spheroid destruction using (a, n = 10) microscopic pictures at 48 h, (b, n = 4) Incucyte live imaging or (c, n = 10) flow cytometry at 48 h, as well as (d and e) overall or (f and g) particular T and NK cells infiltration using flow cytometry (d, f and g, n = 10) and immunofluorescence (e, n = 1) at 24 h. We also analyzed by flow cytometry the expression of (h, n = 8) NKG2D, (i, n = 6) CD137, (j, n = 7) CD16 and (k, n = 8) NKG2A by NK cells in the IN and OUT compartments at 24 h. Statistical significance of immunofluorescence experiment was analyzed using the Mann-Whitney test, A panel using 2-way ANOVA and the other panels using the Wilcoxon matched-pairs signed rank test (* p < 0.05; ** p < 0.005, *** p < 0.001, **** p < 0.0001) [/fig]
[fig] Figure 6 0, Figure 7: NKG2A-HLA-E pathway is engaged during the cocultures and NKG2A blockade synergizes with anti-MICA/B antibodies. a NKG2A expression by CD4 T cells, CD8 T cells and NK cells in the IN and OUT compartments and in the presence or not of IL-15, analyzed by flow cytometry at 24 h. n = 17 independent experiments. b HLA-E expression by tumor cells in the spheroids cocultured or not with CD19-CD14-PBMCs, analyzed by flow cytometry (n = 16 independent experiments) and immunohistochemistry (representative pictures of 1 experiment) at 24h. c to e Analyses of (c, n = 4) caspase-3/− 7 activity, (d, n = 10) spheroid volume, and (e, n = 10) spheroid infiltration 48 h after coculturing HT29 spheroids with CD19-CD14-PBMCs in the presence or not of anti-NKG2A blocking antibodies or corresponding control isotype. f to h Analyses of (f, n = 7) spheroid volume, (g, n = 7) tumor cell apoptosis, and (h, n = 7) spheroid infiltration 48 h after coculturing HT29 spheroids with CD19-CD14-PBMCs in the presence or not of anti-MICA/B antibodies alone or combined with anti-NKG2A blocking antibodies, or with corresponding control isotype antibodies alone or combined. Statistical significance of c panel was analyzed using 2-way ANOVA, the other panels using the Wilcoxon matched-pairs signed rank test (* p < 0.05; ** p < 0.005, *** p < 0.001, **** p < Cocultures between patients-derived spheroids and autologous TILs confirm therapeutic potential of MICA/B and NKG2A targeting. a and b Flow cytometry analyses of a NKG2A and b NKG2D expression by CD4 T cells, CD8 T cells and NK cells in the blood and tumor of colorectal cancer patients. n = 41 independent experiments. c Scheme of the protocol used to create primary colon tumor-derived spheroids and maintain autologous TILs in culture before cocultures. d Representation of the change in spheroid volume 48 h after coculturing or not primary CRCderived spheroids and autologous TILs in the presence or not of IL-15, measured using microscopic pictures. e to g Change in spheroid volume 48 h after culturing primary CRC-derived spheroids with (f and g) or without (e) autologous TILs, in the presence or not of anti-MICA/B, anti-NKG2A or corresponding control isotypes. [/fig]
|
The influence of food, beverages and NSAIDs on gastric acid secretion and mucosal integrity.
Gastric acid secretion is stimulated by all foods, especially proteins, and many beverages, the most potent beverages are milk and fermented substances such as beer and wine. The effects of food on mucosal integrity have been little studied, whereas non-steroidal anti-inflammatory drugs are well known to induce tissue injury.
# Introduction
The effects of orally ingested substances on gastric function has been the subject of human investigation almost since the "discovery" of hydrochloric acid in gastric juice. From chilis and peppers to vodka and cognac, a variety of substances have been tested. Furthermore, the recognition of their ulcerogenic potential has sparked an interest in the effects of NSAIDsb on acid secretion and the gastric mucosa. This review focuses on the effects of foods, beverages, and NSAIDs on gastric acid secretion and mucosal integrity. Cigarette smoking, ulcer therapeutic agents and other drugs will not be discussed.
GASTRIC ACID SECRETION Foods and Beverages All foods are capable of stimulating gastric acid secretion through distention of the stomach, but proteins are the the major stimulants. Digested protein in the form of peptides, peptones, and amino acids act primarily through the stimulation of gastrin from antral G cells [bib_ref] Role of gastrin heptadecapeptide in the acid secretory response to amino acids..., Feldman [/bib_ref]. The aromatic amino acids are the most potent of the amino acids. There is also evidence that amino acids absorbed into the circulation may directly stimulate parietal cells [bib_ref] Intravenous infusion of amino acids stimulates gastric acid secretion in man, Isenberg [/bib_ref]. Protein is the most important physiologic stimulus to acid secretion in man.
The effects of various beverages on acid secretion have been assessed by McArthur and colleagues [bib_ref] Relative stimulatory effects of commonly ingested beverages on gastric acid secretion in..., Mcarthur [/bib_ref]. When compared to water control, soft drinks, coffee and tea, beer and milk were all substantial and significant stimulants of acid secretion. The highest outputs occurred with beer and milk, each producing acid output at or above the maximal acid output in response to pentagastrin. The authors found no correlation between pH, osmolality, caloric content, buffering capacity, ionized calcium levels or caffeine content and the degree of secretory response. For example, although 7-Up has no caffeine or protein and only minimal amounts of calcium, it is a significant stimulant of acid secretion.
Various investigators have studied individual classes of beverages in greater detail. Ippoliti found that low-fat or non-fat milk stimulated as much acid secretion as whole milk in both duodenal ulcer patients and normal subjects [bib_ref] The effect of various forms of milk on gastric acid secretion, Ippoliti [/bib_ref]. Low-calcium milk stimulated acid output in duodenal ulcer patients but not normal subjects. bAbbreviations: NSAID, non-steroidal anti-inflammatory drug. milk responsible for acid secretion is protein. The observation that milk was a stimulant of acid secretion, as well the atherogenic properties of fat and the potential for development of the milk-alkali syndrome, certainly hastened its demise as ulcer therapy.
Feldman and his collaborators at CURE found that decaffeinated coffee was more potent than a peptone meal in stimulating both gastrin release and acid secretion, with the coffee producing acid output equal to 70 percent of peak acid output to pentagastrin [bib_ref] Gastric acid and gastrin response to decaffeinated coffee and a peptone meal, Feldman [/bib_ref].
It remains unknown what component of decaffeinated coffee is the secretagogue. The amount of protein in coffee is only 12 percent of that in the peptone meal, and the amount of calcium is minimal.
The acid secretory response to intragastric ethanol is very weak, if at all [bib_ref] Effect of intragastric infusions of ethanol and wine on serum gastrin concentration..., Peterson [/bib_ref] [bib_ref] Alcohol and gastric acid secretion in humans: a short review, Singer [/bib_ref]. In one well controlled study, ethanol in concentrations of five percent, 12 percent, and 36 percent (concentrations found in popular beverages) produced no more acid than isovolumetric saline controls [bib_ref] Effect of intragastric infusions of ethanol and wine on serum gastrin concentration..., Peterson [/bib_ref]. Similar results have been found for distilled spirits (e.g., whisky, cognac) [bib_ref] Effect of intragastric infusions of ethanol and wine on serum gastrin concentration..., Peterson [/bib_ref]. On the other hand, beer and wine (red or white) are quite potent stimulants of acid secretion [bib_ref] Effect of intragastric infusions of ethanol and wine on serum gastrin concentration..., Peterson [/bib_ref] [bib_ref] Alcohol and gastric acid secretion in humans: a short review, Singer [/bib_ref]. In one study, the mean acid output in response to 300 ml of 12 percent red or white wine was higher than the peak acid output to pentagastrin [bib_ref] Effect of intragastric infusions of ethanol and wine on serum gastrin concentration..., Peterson [/bib_ref]. In the latter study, it was noted that there was a highly significant correlation (r = .91) between the rise in serum gastrin concentrations and acid output in response to red or white wine. Beer and wine differ from whisky and cognac in that they are fermented rather than distilled. It seems reasonable that the stimulatory substances in fermented beverages are amino acids and amines, both of which are present in beer and wine. Histamine is one amine that does not play a role, however, since white wine has very low or absent levels of the amine and is still a potent stimulant of acid secretion. NSAIDs Only a few studies have been performed during the past 20 years measuring the effect of NSAIDS on gastric acid secretion or gastric acidity. In two studies involving healthy volunteers, indomethacin was shown to increase mean basal acid secretion [bib_ref] Effect of indomethacin on gastric acid and bicarbonate secretion in humans, Feldman [/bib_ref] [bib_ref] Effect of indomethacin on basal and histamine stimulated human gastric acid secretion, Levine [/bib_ref] , increase histamine stimulated-acid secretion [bib_ref] Effect of indomethacin on basal and histamine stimulated human gastric acid secretion, Levine [/bib_ref] , but have no effect on meal-stimulated acid secretion [bib_ref] Effect of indomethacin on gastric acid and bicarbonate secretion in humans, Feldman [/bib_ref]. Where acid secretion was increased, however, there was marked variation in the response from subject to subject. A more recent study found no significant differences in median 24 h pH in rheumatoid arthritis patients infected with H. pylori [bib_ref] Effects of indomethacin on intragastric pH and meal-stimulated serum gastrin secretion in..., Janssen [/bib_ref].
## Mucosal integrity foods and beverages
Literature on the effects of specific foods on gastric mucosal integrity in humans is sparse. The Singapore group has reported that chilis do not produce macroscopic gastric mucosal damage [bib_ref] Role of gastrin heptadecapeptide in the acid secretory response to amino acids..., Feldman [/bib_ref] and may even protect against aspirin-induced mucosal injury [bib_ref] Chili protects against aspirin-induced gastroduodenal mucosal injury in humans, Yeoh [/bib_ref]. Graham and colleagues have shown that red or black peppers produce gastric mucosal injury comparable to that seen with aspirin [bib_ref] Effect of red pepper and black pepper on the stomach, Myers [/bib_ref] , but that highly spiced meals (i.e., mexican food with jalapeno peppers or a pepperoni pizza) do not lead to endoscopically demonstrable gastroduodenal mucosal damage [bib_ref] Spicy food and the stomach: evaluation by videoendoscopy, Graham [/bib_ref].
The effect of alcohol on human gastric mucosa was first described in 1833 by William Beaumont who observed erosions and hemorrhages through the gastric fistula of Alexis St. Martin. Others have confirmed that the acute ingestion of alcohol produces erythema, sub-epithelial hemorrhages and erosions [bib_ref] Gastritis: a reevaluation, Palmer [/bib_ref] [bib_ref] Alcohol-induced gastric and duodenal lesions in man, Gottfried [/bib_ref] [bib_ref] Alcohol injury to the normal human gastric mucosa: endoscopic, histologic, and functional..., Tamawski [/bib_ref]. Laine and Weinstein describe subepithelial hemorrhages ("blood under plastic wrap") in 20 of 125 actively drinking alcoholic patients undergoing endoscopy for upper gastrointestinal bleeding [bib_ref] Histology of alcoholic hemorrhagic "gastritis": A prospective evaluation, Laine [/bib_ref]. Histologically, they found extensive foveolar hemorrhage in the target lesions, but only edema in the surrounding mucosa. Inflammation was not a major component of alcohol-induced injury.
## Nsaids
NSAIDs are well-recognized to produce two types of lesions: acute gastric mucosal erosions and subepithelial hemorrhages or frank gastric ulceration. The first are most likely related to a topical effect of the NSAID on the gastric mucosa, whereas ulceration is related primarily to prostaglandin depletion. However, both mechanisms may well play roles in both situations. The incidence of new gastric ulcers in patients taking NSAIDs ranges from 10 percent to 20 percent and the incidence of duodenal ulcers ranges from four percent to 10 percent [bib_ref] Prevention of NSAID-induced gastric ulcer with misoprostol: multicentre, double-blind, placebo-controlled trial, Graham [/bib_ref] [bib_ref] Prevention of gastroduodenal damage induced by non-steroidal anti-inflammatory drugs: controlled trial of..., Ehsanullah [/bib_ref] [bib_ref] Effect of ranitidine on gastroduodenal mucosal damage induced by nonsteroidal antiinflammatory drugs, Robinson [/bib_ref] [bib_ref] and The Misoprostal Study Group. Duodenal and gastric ulcer prevention with misoprostol..., Graham [/bib_ref]. It should be noted, however, that only a small proportion of these are clinically important, producing either pain or complications. It appears that even very low doses of NSAIDs can lower prostaglandin levels and produce ulceration. Current investigation is ongoing to find compounds with relatively low ulcerogenicity or novel means of administering NSAIDs so to minimize gastric injury. |
Diffuse Alveolar Hemorrhage in a 39-year-old Woman: Unusual Initial Presentation of Microscopic Polyangiitis
Microscopic polyangiitis (MPA) is a necrotizing vasculitis involving the small vessels without granulomatous inflammation. Most MPA initially presents with renal involvement without pulmonary involvement. Isolated and initially presenting alveolar hemorrhage is very rare. The patient was a 39-year-old female with a progressive cough, dyspnea, and blood-tinged sputum for the previous 5 days. We determined that her condition was MPA though VATS lung biopsy and renal biopsy. After 2 months of steroid therapy, the chest lesions had improved. We report here a rare case of MPA with isolated and initial involvement of the lung with a review of the literature.
## Case report
A 39-year-old woman visited the emergency room with a progressive cough, dyspnea, and blood-tinged sputum for the previous 5 days. The patient did not use tobacco or alcohol, nor did she have any specific medical history. At initial presentation, her blood pressure was 107/54 mmHg, pulse rate 74 beats per minute, respiration rate 22 breaths per minute, and body temperature 36.8 o C. The patient was wide awake and well oriented. There were no other suggestive symptoms.
Here CBC count was as follows: white blood cell count 8,420/mm 3 (75.1% granulocytes, 17.7% lymphocytes, 5.6% monocytes and 0.4% eosinophils), hematocrit 25.2 mg/dL and platelets 374 K/mm 3 . Her serum chemistry and liver function were normal. There was no serologic evidence of hepatitis. A urine analysis was non-specific. A chest PA revealed an in-creased opacity throughout both lower lung fields and a chest CT revealed extensive ground-glass opacities in both whole lung fields without any lymphadenopathy [fig_ref] Figure 1: dence of hepatitis [/fig_ref]. A sputum study was negative for acid-fast bacilli and blood culture findings were negative. Bronchoscopy findings were compatible with pulmonary alveolar hemorrhage at RML and RLL.
At first we considered the case to be a simple alveolar hemorrhage associated with pneumonia, so we treated her conservatively and her symptoms, including hemoptysis, improved. On the 5th day from admission, her urine color changed abruptly to bloody and we performed further evaluation for simultaneous hematuria and hemoptysis. Serum protein electrophoresis revealed a decrease in total protein albumin but a normal immunotyping pattern. There was no serologic evi- pneumonia, and mycoplasma tuberculosis titers were all negative, Complement levels (C3, C4, and CH50) were normal but anti MPO antibody (＞100) and p-ANCA (1 : 160) were positive. For accurate diagnosis, we performed a VATS lung biopsy on the 9th hospital day and renal biopsy on the 14th hospital day. The pathology report showed the lung lesion was intraralveolar hemorrhage suggesting capillaritis and the kidney lesion was focal segmental necrotizing glomerulonephritis without antibody deposition at the capillary basement membrane [fig_ref] Figure 2: Lung specimen showed intraalveolar hemorrhage with aggregation of pigment-laden macrophages and a... [/fig_ref]. Therefore, we diagnosed this condition as microscopic polyangiitis presenting with initial pulmonary involvement and we started her on steroid therapy. After 2 months of steroid therapy, her chest lesions had improved but hematuria remained [fig_ref] Figure 4: Chest X-ray after steroid therapy for 2 months showed improved infiltration in... [/fig_ref]. More treatment will be needed.
# Discussion
There are many causes for hemoptysis or diffuse alveolar hemorrhage. One of them is vasculitis, which is rarely encountered in the thoracic surgery field [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] [bib_ref] Intrapulmonary hemorrhage in collagen-vascular diseases includes a spectrum of underlying conditions, Kobayashi [/bib_ref]. Vasculitis is an inflammatory destructive process affecting blood vessels and surrounding tissue through diverse mechanisms. Microscopic polyangiitis (below MPA) is a necrotizing vasculitis involving the small vessels without granulomatous inflammation [bib_ref] Clinical features and outcomes of microscopic polyangiitis in Korea, Oh [/bib_ref].
MPA is the most common antineutrophilic cytoplasmic antibody (ANCA)-associated small vessel vasculitis and is characterized by a lack of immune deposits in the involved vessels [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. The incidence of MPA is about 1 : 100,000 with a slight predominance in men. The mean age of onset is about 40 or 50 years [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. The kidney is the most frequently involved organ in 90% of patients. The lung is involved in 22%, and alveolar hemorrhage occurs in 11% [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref] [bib_ref] A case of pulmonary fibrosis with microscopic polyangiitis, Jeong [/bib_ref]. Other organs and systems including the skin, GI tract, musculoskeletal system, heart, and nervous system can also be involved [bib_ref] Clinical features and outcomes of microscopic polyangiitis in Korea, Oh [/bib_ref] [bib_ref] Microscopic polyangiitis presenting as a "pulmonary-muscle" syndrome: is subclinical alveolar hemorrhage the..., Birnbaum [/bib_ref].
Most MPA initially presents as renal involvement without pulmonary involvement, while isolated and initially presenting alveolar hemorrhage is very rare [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. The cause of MPA remains unknown, but some researchers have proposed genetic, environmental, and immunologic factors such as ANCA in the pathogenesis of MPA [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. Chest radiographic findings consist of ground-glass opacity, consolidation, thickened bronchovascular bundles and a honeycomb appearance caused by alveolar hemorrhage [bib_ref] Intrapulmonary hemorrhage in collagen-vascular diseases includes a spectrum of underlying conditions, Kobayashi [/bib_ref] [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. An ANCA test is used to aid in diagnosis and to monitor inflammatory activity in vaculitis including MPA, Wegener's granulomatosis, and Churg-Strauss syndrome [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] [bib_ref] Clinical features and outcomes of microscopic polyangiitis in Korea, Oh [/bib_ref]. Previous studies have reported that 50% to 80% of MPA patients test positive for ANCA [bib_ref] Clinical features and outcomes of microscopic polyangiitis in Korea, Oh [/bib_ref] [bib_ref] A case of pulmonary fibrosis with microscopic polyangiitis, Jeong [/bib_ref].
Because lung biopsy can aggravate conditions of critically ill patients and pulmonary involvement usually follows after renal involvement, lung biopsy is not usually recommended in the case of vasculitis-associated alveolar hemorrhage if the diagnosis can be obtained by biopsy of other organs and by a serum immunology study [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. In our case, because of the unusual presentations, specifically, initial pulmonary involvement without renal involvement and hemoptysis improvement after conservative treatment, we needed a lung biopsy for accurate diagnosis. Lung specimens of MPA usually show only necrosis, nonspecific inflammation, and hemorrhage pulmonary capillaritis [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. Renal specimens usually show focal segmental necrotizing glomerulonephritis without any immune deposition [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. Our lung specimen showed intraalveolar hemorrhage with aggregation of pigment-laden macrophages and a few neutrophils, suggesting capillaritis and organizing pneumonia, and our renal specimen showed focal segmental necrotizing glomerulonephritis without immune deposits in the basement membrane. These lung and kidney specimen pathologies were consistent with pulmonary and renal involvement of MPA.
The differential diagnosis should include Goodpasture syndrome. Goodpasture syndrome is a rare but severe immunologic disease, which is characterized by rapidly progressing glomerulonephritis and diffuse intraalveolar hemorrhage [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] [bib_ref] Intrapulmonary hemorrhage in collagen-vascular diseases includes a spectrum of underlying conditions, Kobayashi [/bib_ref]. However, in our case there was immune deposition at the capillary basement membrane [bib_ref] Pulmonary vasculitis, Brown [/bib_ref] , so we could rule out Goodpasture syndrome.
## Diffuse alveolar hemorrhage due to mpa
[formula] − 451 − [/formula]
The mainstay of treatment of MPA is systemic corticosteroids and cyclophosphamide administration [bib_ref] Clinical features and outcomes of microscopic polyangiitis in Korea, Oh [/bib_ref] [bib_ref] Microscopic polyangiitis presenting as a "pulmonary-muscle" syndrome: is subclinical alveolar hemorrhage the..., Birnbaum [/bib_ref]. The treatment duration of at least 12 months of combined therapy has been recommended based on empirical research. Few papers have studied the course and prognosis of pulmonary involvement in MPA [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref]. Alveolar hemorrhage usually improves within several days after the start of steroid therapy but can continue longer [bib_ref] Intrapulmonary hemorrhage in collagen-vascular diseases includes a spectrum of underlying conditions, Kobayashi [/bib_ref] [bib_ref] Pulmonary involvement in microscopic polyangiitis, Collins [/bib_ref].
We describe a 39-year-old woman presenting with a progressive cough, dyspnea, and hemoptysis. We determined that her condition was MPA though VATS lung biopsy and renal biopsy. After 2 months of steroid therapy, her chest lesions had improved. We report here a rare case of MPA with isolated and initial involvement of the lung, along with a review of the literature.
[fig] Figure 1: dence of hepatitis. Serum antinuclear antibody, anti-glomerular basement membrane antibody (anti-GBM), mycoplasma Diffuse Alveolar Hemorrhage Due to MPA − 449 − Chest X-ray (A) and Chest CT (B) showed extensive groundglass opacities in both whole lung fields without any lymphadenopathy. [/fig]
[fig] Figure 2: Lung specimen showed intraalveolar hemorrhage with aggregation of pigment-laden macrophages and a few neutrophils suggesting capillaritis (hematoxylin-eosin stain, x100). [/fig]
[fig] Figure 3: Renal specimen shown was focal segmental necrotizing glomerulonephritis without antibody deposition at the capillary basement membrane (A: hematoxylin-eosin stain, ×400, B: immunofluorescence stain, ×400). [/fig]
[fig] Figure 4: Chest X-ray after steroid therapy for 2 months showed improved infiltration in both lung fields. [/fig]
|
The frequency of CD25+CD4+ and FOXP3+ regulatory T cells in ectopic endometrium and ectopic decidua
Background: The presence of regulatory T (Treg) cells in human endometrium is crucial for maintaining immunological homeostasis within the uterus. For this study we decided to evaluate the subpopulations of Treg cells in conditions where a disturbance in the immunological equilibrium in ectopic endometrium and decidua has been observed, such as in cases of ovarian endometriosis (involving local immune cell suppression) and ectopic pregnancy (involving an increase in local immune system activity). We then compared these findings to what we observed in the normal eutopic endometrium of women during the secretory phase of the menstrual cycle (with immune cells under individual control).Methods:The endometrium tissue samples evaluated in our study were obtained from 47 women during one of two kinds of laparoscopic procedures. 16 of the women underwent laparoscopies due to Fallopian tube pregnancies (EP), and 16 due to ovarian endometrioma, while 15 women made up a control group. The presence of regulatory T cells in these tissue samples was evaluated by FACS. Results: In our study, the percentages of FOXP3+ cells within the subpopulation of CD4+ T lymphocytes found in the decidua of the patients treated for Fallopian tube pregnancies were statistically significantly lower than both those observed in the ovarian endometriosis tissue samples and those found in the secretory eutopic endometrium samples of the control group.Conclusion: The disturbance in the immunological equilibrium observed in ectopic endometrium and decidua would seem to be related to the alteration in the Treg cell population that occurs in these ectopic tissues.
# Background
During the normal menstrual cycle the endometrium within the uterus is widely infiltrated by immune cells [bib_ref] Variation during the menstrual cycle of immune cell populations in human endometrium, Starkey [/bib_ref] [bib_ref] Human uterine natural killer cells, King [/bib_ref] [bib_ref] Dendritic cells in the human decidua, Gardner [/bib_ref]. The specific activities of these immune cells are crucial for the proper course of such reproductive processes as menstruation and implantation. The suppressive profile of the endometrium is what enables the respective activities of these infiltrating immune cells. This suppressive phenomenon of the endometrium is a result both of the ability of endometrial cells to suppress immune cells and of the infiltration of the endometrium by such immune regulatory cells as regulatory T (Treg) cells or suppressive macrophages (B7H4) [bib_ref] The role of the T-regulatory cells in pregnancy and cancer, Wilczynski [/bib_ref] [bib_ref] B7-H4 expression identifies a novel suppressive macrophage population in human ovarian carcinoma, Kryczek [/bib_ref] [bib_ref] The analysis of Metallothionein, RCAS1 immunoreactivity regarding immune cells concentration in endometrium..., Wicherek [/bib_ref]. In some cases, however, the endometrium, which constitutes an active form of immunoregulatory tissue, is located outside the uterus. In other cases, as in ectopic pregnancy (e.g., within the Fallopian tube), the trophoblast cells may be located outside the uterus, even though the infiltrating trophoblast cells within the tubal wall are linked to the decidualization of epithelial and stromal cells. In cases of ovarian endometriosis and of ectopic pregnancy, mononuclear immune cells are recruited to the microenvironment of ectopic lesions; after this Treg cells infiltrate the tissue [bib_ref] Metallothionein and RCAS1 expression in comparison to immunological cell activity in endometriosis,..., Wicherek [/bib_ref] [bib_ref] A conditional mouse model for human MUC1-positive endometriosis shows the presence of..., Budiu [/bib_ref].
Treg cells are crucial for the tolerance and prevention of autoimmunity and have been described as pivotal both for the survival of allogeneic organ grafts and for the evasion of immune surveillance by fetal cells within the uterus [bib_ref] The role of the T-regulatory cells in pregnancy and cancer, Wilczynski [/bib_ref] [bib_ref] B7-H4 expression identifies a novel suppressive macrophage population in human ovarian carcinoma, Kryczek [/bib_ref] [bib_ref] The analysis of Metallothionein, RCAS1 immunoreactivity regarding immune cells concentration in endometrium..., Wicherek [/bib_ref] [bib_ref] Metallothionein and RCAS1 expression in comparison to immunological cell activity in endometriosis,..., Wicherek [/bib_ref] [bib_ref] A conditional mouse model for human MUC1-positive endometriosis shows the presence of..., Budiu [/bib_ref] [bib_ref] Decidual and peripheral blood CD4+CD25+ regulatory T cells in early pregnancy subjects..., Sasaki [/bib_ref] [bib_ref] CD25+CD4+ regulatory T cells prevent graft rejection: CTLA4-and IL-10-dependent immunoregulation of alloresponses, Kingsley [/bib_ref] [bib_ref] Foxp3: a critical regulator of the development and function of regulatory T..., Hori [/bib_ref] [bib_ref] CD4(+)CD25high regulatory T cells in human pregnancy, Saito [/bib_ref] [bib_ref] Regulatory T cells and regulatory NK cells play an important roles in..., Saito [/bib_ref]. Moreover, it has been observed that the number of Treg cells in the peripheral blood not only increases in women during pregnancy but also that their level of infiltration within the endometrium and decidua becomes elevated [bib_ref] Foxp3: a critical regulator of the development and function of regulatory T..., Hori [/bib_ref]. While Tilburgs et al. have shown that fetus-specific Treg cells are recruited to the decidua from the periphery [bib_ref] Evidence for a selective migration of fetus-specific CD4+CD25bright regulatory T cells from..., Tilburgs [/bib_ref] , Arruvito et al. have demonstrated that the changes that take place in the Treg cell population in the peripheral blood over the different menstrual cycle phases are crucial for the proper course of reproductive processes. More specifically, the authors of the latter study found that the deregulation of the frequency of Treg cells in women with recurrent spontaneous abortion (RSA) may contribute to reproductive failure [bib_ref] Expansion of CD4+CD25+and FOXP3+ regulatory T cells during the follicular phase of..., Arruvito [/bib_ref]. Recently, Schumacher et al. have shown that Treg cell attraction to decidua is associated with the high HCG concentration that occurs during intrauterine pregnancy development [bib_ref] Human chorionic gonadotropin attracts regulatory T cells into the fetal-maternal interface during..., Schumacher [/bib_ref]. By contrast, the Treg cell concentration within the decidua derived from women with ectopic pregnancy or who suffer from spontaneous abortion is lower because the concentration level of beta HCG is lower in such conditions than it is in normal pregnancy [bib_ref] Human chorionic gonadotropin attracts regulatory T cells into the fetal-maternal interface during..., Schumacher [/bib_ref]. Moreover, the infiltration level of Treg cells within decidua is decreased in patients suffering from spontaneous abortion compared with the level of infiltration found in the decidua of those undergoing induced abortions [bib_ref] Decidual and peripheral blood CD4+CD25+ regulatory T cells in early pregnancy subjects..., Sasaki [/bib_ref] , and spontaneous abortion is strongly associated with an increase in the activity of the maternal immune system [bib_ref] Th1/Th2/Th17 and regulatory T-cell paradigm in pregnancy, Saito [/bib_ref] [bib_ref] Progesterone-induced blocking factor (PIBF) modulates cytokine production by lymphocytes from women with..., Raghupathy [/bib_ref] [bib_ref] Inflammation and implantation, Dekel [/bib_ref]. Furthermore, it has been shown that the rupture of the tubal wall is a result not only of trophoblast cell infiltration, but more importantly of the accumulation of immune cytotoxic cells within the ectopic ovum microenvironment [bib_ref] The analysis of Metallothionein, RCAS1 immunoreactivity regarding immune cells concentration in endometrium..., Wicherek [/bib_ref]. In contrast to the situation with ectopic decidua in the Fallopian tube, local suppression of immune cytotoxic cells has been observed within ectopic endometrium (such as ovarian endometrioma) during the development of endometriosis [bib_ref] Apoptosis in endometrial glandular cells in women with and without endometriosis, Dmowski [/bib_ref]. Recently, Berbic et al. have demonstrated the presence of Treg cells within both the ectopic and eutopic endometrium of women suffering from peritoneal endometriosis. They showed that the FOXP3 expression within the eutopic endometrium derived from women with peritoneal endometriosis differs from that observed in healthy women [bib_ref] The role of Foxp3+ regulatory T-cells in endometriosis: a potential controlling mechanism..., Berbic [/bib_ref]. An alteration in Treg lymphocyte infiltration generally disrupts the immunological equilibrium. Since the potential role of Treg cells in maintaining homeostasis in ectopic endometrium and decidua has not been precisely established in the literature, from a clinical standpoint, both endometriosis and ectopic pregnancy constitute unresolved problems. For this reason, we have chosen the specific type of ectopic endometrium and decidua found at the beginning of decidualization as a research model in order to demonstrate the potential role of Treg cells in the development of the suppressive profile of the endometrium. Thus to reiterate, the aim of our present study has been to evaluate the subpopulations of Treg cells in ectopic endometrium and decidua.
# Methods
## Patients
The endometrium tissue samples evaluated in our study were obtained from 47 women during one of two kinds of surgical procedures. 16 of the women underwent laparoscopic procedures because of unruptured Fallopian tube pregnancies (EP group, mean age 30.7 (± 6.1) years) and 16 because of ovarian endometrioma, (OE group, mean age 32.3 ± 5.2) while 15 (constituting the control group, mean age 35.5 ± 8.1) underwent curettage as an additional procedure during uterine cervix biopsy on account of a diagnosis of cervical intraepithelial neoplasia or CIN (it should be noted that the biopsy of the endometrium was an additional procedure for which the patient's consent was obtained).
The ectopic endometrium tissue samples were derived from the women who underwent laparoscopy due to ovarian endometrioma while the ectopic decidua samples were derived from those patients who underwent the procedure because of ectopic pregnancy. Meanwhile, the eutopic endometrium tissue samples were derived from the patients on whom curettage was performed. The patients included in the study were treated between January 2007 and June 2010 in the Department of Gynecology and Oncology at the Jagiellonian University, Krakow, Poland. The patients referred to the control group were randomly selected from the group who had curettage following uterine cervix biopsy because they were diagnosed with cervical intraepithelial neoplasia. The patients from the control group were not found to have any other pathological lesions within the reproductive tract and had normal, regular menstrual periods.
Patients with intrauterine infections (as confirmed by histopathological examination) were excluded from the study. Patients with concomitant systemic diseases, diseases of the thyroid gland, diabetes or hypertension, or the presence of any other lesions as detected by gynecological, ultrasound, cytology, or colposcopy examinations were also excluded. Furthermore, none of the patients included in our study received any hormonal treatment. The tissue samples were then independently evaluated using routinely stained (hematoxylin and eosin) slides prepared from paraffin-embedded tissue material, enabling accurate diagnosis: the presence of ectopic pregnancy or ovarian endometrioma was confirmed in each of the patients. All of the patients were then classified into subgroups according to the particular menstrual cycle phase as determined by the histopathological analysis of the tissue samples obtained during surgery and the results of clinical examination together with an assessment of the levels of progesterone, estradiol, and follicle-stimulating hormone (FSH) in the blood sera. (Responsible OE group-8 patients during proliferative cycle phase FSH mlU/ml 5.62 (± 2.4) mlU/ml; estradiol 140.1 (± 34.2) pg/ml; progesterone 0.87 (± 1.9) ng/ml; OE group-8 patients during secretory cycle phases FSH 6.5 (± 4.1) mlU/ml; estradiol 159.3 (± 66.3) pg/ml; progesterone 6.4 (± 3.2) ng/ml; Control group FSH 8.2 (± 5.13) mlU/ml; estradiol 131.4 (± 53.2) pg/ml, progesterone 7.56 (± 2.1) ng/ml.).
The patient's consent was obtained in each case. Prior to the present study we also obtained the approval of the Jagiellonian University Ethical Committee for our research program (DK/KB/CM/0031/447/2010, KBET/ 135/B/2007).
## Decidual mononuclear cell isolation
The decidua and endometrium samples were cut into small fragments and disintegrated by being smashed through a 40 μm cell-strainer. The realized cells were then spun down and the resulting pellet was subjected to ammonium chloride lysis in order to get rid of any contaminating red blood cells. Following lysis, the cells were washed in PBS. They were then re-suspended in PBS, counted, and used either immediately for staining or frozen for future analysis (cells were frozen in a medium consisting of10% DMSO, 5% albumin, and 85% of a cell culture medium).
## Flow cytometry
The cell phenotype was analyzed with the panel of mAb-CD4 FITC, CD25 APC, and FOXP3 PE (Pharmingen). Briefly stated, to the 1 × 10 6 cells suspended in 60 μl of staining buffer (PBS, 2% FBS) 20 μl of each mAb (CD4, CD25) was added. Next the cells were incubated in the dark for 30 min at 4°C. After incubation the cells were washed twice in PBS and were permeabilized with FoxP3 permeabilization buffer (Becton Dickinson; USA) for 10 min at room temperature in the dark. Following this, they were stained with anti-FOXP3 antibodies for 30 min at 4°C in the dark. The stained cells were then washed and collected using the FACSC anto-cytometer (Becton Dickinson; USA), and finally were analyzed with FACS Diva software (Becton Dickinson; USA). Each time 3 × 10 4 events were saved for analysis. Logical gates were used to analyze particular populations of cells. The cells were gated first according to SSC and CD4FITC parameters (gate P1). After this, the cells from gate P1 were analyzed according to CD25PE and CD4FITC parameters, and gate P2 was established on double positive CD4FITC CD25PE cells. The double positive cells from gate P2 were analyzed for the presence of FoxP3 antigen. The gate FoxP3+ was set on cells which had stained positively with FoxP3 APC antibodies [fig_ref] Figure 1: Treg cell gating strategy [/fig_ref].
# Statistical analysis
The statistical significance of the recorded differences among the groups was determined by the Kruskal-Wallis analysis of variance (ANOVA). All statistical analyses were carried out with the Statistica 8.0 software program (StatSoft Inc., Tulsa, OK, USA). A p value < 0.05 was considered indicative of statistical significance.
# Results
CD25+CD4+FOXP3+ T cells were found in all the examined endometriosis tissue samples. These same cells were observed in 72% of the eutopic endometrium tissue samples but were present in only 29% of the tissue samples taken from patients who had ectopic pregnancies [fig_ref] Figure 1: Treg cell gating strategy [/fig_ref].
We did not observe any differences over the course of the menstrual cycle in the percentage of FOXP3+ positive cells within the subpopulation of CD4+T lymphocytes in ectopic endometrium in the tissue samples from the group of patients with ovarian endometriosis. The median for OE during the proliferative and secretory cycle phases were 8.5 (16.7) and 6.3 (5.9) respectively.
The percentage of FOXP3+ cells within the subpopulation of CD4+ lymphocytes was statistically significantly higher in the tissue samples from the patients with endometriosis as compared with the percentage found in the tissue samples from the patients who had developed ectopic pregnancies. Likewise, the percentage of FOXP3+ cells within the subpopulation of CD4+ lymphocytes was higher than in the endometrium from the control group, but these differences were not statistically significant. The percentages of these cells observed in the endometrium from the control group were, however, statistically significantly higher than those found in the ectopic deciduas . Similar results were obtained from analysing FOXP3 positive cells within the subpopulation of CD4+CD25+ lymphocytes .
# Discussion
In our study, the percentages of FOXP3+ cells in the subpopulation of CD4+ T lymphocytes found in the decidua of the patients treated for Fallopian tube pregnancy were statistically significantly lower than both those observed in the ovarian endometriosis samples and those found in the secretory eutopic endometrium of the control group.
Single statements derived from the literature on the subject indicate that Treg lymphocytes are present within ovarian endometrial tissue [bib_ref] A conditional mouse model for human MUC1-positive endometriosis shows the presence of..., Budiu [/bib_ref] [bib_ref] The role of Foxp3+ regulatory T-cells in endometriosis: a potential controlling mechanism..., Berbic [/bib_ref]. Furthermore, it has been demonstrated that the percentage of FOXP3 positive lymphocytes found within the subpopulation of T cells in the peripheral blood of women suffering from ovarian endometriosis does not differ from that observed in healthy women [bib_ref] A conditional mouse model for human MUC1-positive endometriosis shows the presence of..., Budiu [/bib_ref]. Berbic et al. previously demonstrated the presence of Treg cells within peritoneal ectopic endometrium [bib_ref] The role of Foxp3+ regulatory T-cells in endometriosis: a potential controlling mechanism..., Berbic [/bib_ref]. In that study the authors demonstrated that there was an alteration in the FOXP3 expression within the eutopic endometrium that was derived from women suffering from peritoneal endometriosis during the secretory cycle phases when compared to what was observed in the tissue taken from healthy women [bib_ref] The role of Foxp3+ regulatory T-cells in endometriosis: a potential controlling mechanism..., Berbic [/bib_ref]. For fertile women the significant decrease in the Treg cell population in the peripheral blood typically occurs just after ovulation [bib_ref] Expansion of CD4+CD25+and FOXP3+ regulatory T cells during the follicular phase of..., Arruvito [/bib_ref]. By contrast, the women suffering from recurrent spontaneous abortion (RSA) were characterized by an absence of fluctuation in the number of Treg cells [bib_ref] Expansion of CD4+CD25+and FOXP3+ regulatory T cells during the follicular phase of..., Arruvito [/bib_ref]. The increase in the number of Treg cells depends upon estrogen levels, and estrogens are also responsible for the increase in the immune suppressive potential of these cells. Estrogen levels generally seem to be linked with a decrease in TH1 response. Furthermore, a local increase in estrogen levels is characteristic of patients with ovarian endometrioma lesions. In our study the percentages of FOXP3+ cells in the subpopulation of CD4+ T lymphocytes found in eutopic endometrium were slightly lower than those found in ovarian endometrioma tissue. Additionally, no differences in the percentage of Treg cells within the T lymphocyte subpopulation were observed over the course of the menstrual cycle in the ovarian endometriosis tissue samples. Most likely, the absence of Treg cell fluctuation can be linked to an immune defect arising with the development of endometriosis. Although in 95% of cases the endometrial cells migrate along with retrograde menstruation from the endometrial cavity to the peritoneal cavity, only 15% of them actually survive in ectopic localization despite the host immune system response [bib_ref] Immunobiology of endometriosis, Lebovic [/bib_ref]. Since they are able to regulate the activity of immune cells, Under normal physiological conditions the intensity of the immune suppressive activity of the endometrium depends on the menstrual cycle phase and differs significantly between the proliferative and secretory phases (as has been observed by Arruvito et al. in a study on the percentage of Treg cells in the late proliferative and early secretory cycle phases [bib_ref] Expansion of CD4+CD25+and FOXP3+ regulatory T cells during the follicular phase of..., Arruvito [/bib_ref] or in different studies concerning the expression of the factors responsible for immune cell suppression, such as RCAS1 expression within endometrial cells) [bib_ref] Alterations in RCAS1 serum concentration levels during menstrual cycle in patients with..., Wicherek [/bib_ref]. In patients with endometriosis, however, the suppressive activity of the endometrium does not fluctuate. In our previous studies we have demonstrated that both the expression of RCAS1 and HLA-G in ovarian endometriosis and its concentration in the peripheral blood does not differ between the proliferative and secretory cycle phases [bib_ref] Alterations in RCAS1 Serum Concentration Levels during the Normal Menstrual Cycle and..., Wicherek [/bib_ref] [bib_ref] Differences in the blood serum levels of soluble HLA-G concentrations between the..., Basta [/bib_ref]. Both protein RCAS1 and HLA-G are important immunosuppressive factors [bib_ref] Rouas-Freiss N: Human leukocyte antigen-G is expressed in advanced-stage ovarian carcinoma of..., Menier [/bib_ref] [bib_ref] Inhibition of cell growth and induction of apoptotic cell death by the..., Nakashima [/bib_ref] [bib_ref] Association between RCAS1 expression and microenvironmental immune cell death in uterine cervical..., Sonoda [/bib_ref] [bib_ref] The biological role of the unique molecule RCAS1: a bioactive marker that..., Sonoda [/bib_ref]. The suppressive profile of the endometrium therefore depends not only on the absence of physiological changes-as has been observed from the Treg cell population-but also on the over-expression of suppressive factors within ectopic endometrium. For example, the ovarian endometriosis tissue samples were characterized by an over-expression of mRNA for FOXP3 in comparison to normal endometrial tissue [bib_ref] A conditional mouse model for human MUC1-positive endometriosis shows the presence of..., Budiu [/bib_ref]. Additionally, women experiencing primary unexplained infertility have significantly lower FOXP 3 expression within the endometrium than do fertile women [bib_ref] Primary unexplained infertility is associated with reduced expression of the T-regulatory cell..., Jasper [/bib_ref]. These studies thus show how important the proper regulation of immune cell activities is for maintaining homeostasis during pregnancy. reported that about 7% of CD4+ cells in the decidua derived from patients suffering spontaneous abortion are CD4+CD25 bright cells and that this percentage is lower than in the tissue of women experiencing normal pregnancies [bib_ref] Decidual and peripheral blood CD4+CD25+ regulatory T cells in early pregnancy subjects..., Sasaki [/bib_ref]. Furthermore, Shansham et al. have shown that patients suffering unexplained spontaneous abortion were typified by a smaller proportion of CD4 +CD25+FOXP3+ lymphocytes within decidua than women experiencing normal pregnancies. Recently, Arruvito et al have demonstrated that the Treg lymphoctes of women suffering RSA typically exhibit lower than normal suppressive activity [bib_ref] A physiological role for inducible FOXP3(+) TREG cells Lessons from women with..., Arruvito [/bib_ref]. Thus the decrease in the level of Treg cell infiltration within decidua may result in an increase in both cytotoxic T lymphocyte and NK cell activity. During pregnancy Treg cells are recruited to the decidua [bib_ref] CD25+CD4+ regulatory T cells prevent graft rejection: CTLA4-and IL-10-dependent immunoregulation of alloresponses, Kingsley [/bib_ref] , but the percentage of Treg cells in both the decidua and the peripheral blood increases only until spontaneous labor begins [bib_ref] Evidence for a selective migration of fetus-specific CD4+CD25bright regulatory T cells from..., Tilburgs [/bib_ref]. When recruited to the endometrium, Treg cells are directly able to suppress the immune response to both fetus-specific and fetus non-specific antigens [bib_ref] Evidence for a selective migration of fetus-specific CD4+CD25bright regulatory T cells from..., Tilburgs [/bib_ref] [bib_ref] Trophoblast Research Award Lecture: Unique properties of decidual T cells and their..., Tilburgs [/bib_ref]. The beginning of spontaneous labor, however, is associated with a decrease in the percentage of Treg cells within decidua along with a subsequent increase in the immune response [bib_ref] Comparison of decidual leukocytes following spontaneous vaginal delivery and elective cesarean section..., Sindram-Trujillo [/bib_ref] [bib_ref] Differential distribution of NK cells in decidua basalis compared with decidua parietalis..., Sindram-Trujillo [/bib_ref] [bib_ref] Increased T-cell activation in decidua parietalis compared to decidua basalis in uncomplicated..., Sindram-Trujillo [/bib_ref] [bib_ref] Changes in the subpopulation of CD25+ CD4+ and FOXP3+ regulatory T cells..., Galazka [/bib_ref] , as has also been observed during spontaneous abortion [bib_ref] Th1/Th2/Th17 and regulatory T-cell paradigm in pregnancy, Saito [/bib_ref]. The activity of immune cells within the uterus is precisely controlled The changes in the percentage of FOXP3+ cells within the subpopulation of CD4+ lymphocytes. The comparison of the percentages of FOXP3+ cells within the subpopulation of CD4+ lymphocytes in the ectopic endometrium (OE) tissue samples derived from patients with ovarian endometriosis and in ectopic decidua (EP) derived from patients with Fallopian tube pregnancy with the percentages found in the eutopic endometrium samples derived from patients during the secretory cycle phases (E). The data is presented as a median ± IQR (Intraquartile Range).
## Figure 3
The changes in the percentage of FOXP3+ cells within the subpopulation of CD4+CD25+ lymphocytes. The comparison of the percentages of FOXP3+ cells within the subpopulation of CD4+CD25+ lymphocytes in the ectopic endometrium (OE) tissue samples derived from patients with ovarian endometriosis and in ectopic decidua (EP) derived from patients with Fallopian tube pregnancy with the percentages found in the eutopic endometrium samples derived from patients during the secretory cycle phases (E). The data is presented as a median ± IQR (Intraquartile Range). by many factors, and the appearance of such pathology as preeclampsia, abortion, placental abruption, and retained placental tissue results from a disturbance in the regulation of maternal immune cells [bib_ref] RCAS1 decidual immunoreactivity in severe pre-eclampsia: immune cell presence and activity, Wicherek [/bib_ref] [bib_ref] RCAS1 Decidual Immunoreactivity during Placental Abruption: Immune Cell Presence and Activity, Wicherek [/bib_ref] [bib_ref] The role of the endometrium in the regulation of immune cell activity, Wicherek [/bib_ref]. Just such a situation has also been observed during the development of ectopic Fallopian tube pregnancy where the tubal rupture is seen as a consequence of the increase in maternal immune cell activity [bib_ref] The analysis of Metallothionein, RCAS1 immunoreactivity regarding immune cells concentration in endometrium..., Wicherek [/bib_ref] [bib_ref] Metallothionein expression and infiltration of cytotoxic lymphocytes in uterine and tubal implantation..., Wicherek [/bib_ref]. In our previous study we have demonstrated that the development of Fallopian tube pregnancy is related both to the accumulation of cytotoxic immune cells and NK cells within ectopic decidua and to a continued increase in the activity of these cells within the tubal wall [bib_ref] The analysis of Metallothionein, RCAS1 immunoreactivity regarding immune cells concentration in endometrium..., Wicherek [/bib_ref] [bib_ref] Metallothionein expression and infiltration of cytotoxic lymphocytes in uterine and tubal implantation..., Wicherek [/bib_ref]. In our current study, however, we observed that the Treg cell accumulation within the tubal wall was not as dense as it was within the eutopic endometrium during the secretory cycle phase. This is not unlike what occurs in eutopic endometrium with Arias Stella reaction where the Treg cell population decreases in comparison to that of eutopic endometrium during the secretory cycle phases [bib_ref] The characterization of the subpopulation of suppressive B7H4(+) macrophages and the subpopulation..., Wicherek [/bib_ref]. The decrease in the accumulation of Treg cells in ectopic decidua within the Fallopian tube wall observed both in the Schumacher et al. [bib_ref] Human chorionic gonadotropin attracts regulatory T cells into the fetal-maternal interface during..., Schumacher [/bib_ref] study and in our current one would therefore seem to be associated with an increase in the activity of the immune cells infiltrating these tissues that leads finally to tubal rupture. This contrasts with the conditions linked to ectopic endometrium; ovarian endometrioma development, for example, is associated with an immune defect. This observation would seem to accord with our own results showing that the percentage of the Treg cells within the overall cell population was the highest in the ovarian endometrioma tissue samples.
The proper balance between immune cell activity and the intensity of the suppressive profile of the endometrium or decidua is what enables the proper course of physiological reproductive processes. The correct balance is a result of the accumulation of Treg lymphocytes within the endometrium and decidua respectively. Such processes as the development of ectopic pregnancy and ovarian endometrioma are associated with a disturbance in the suppressive profile of the endometrium resulting from the alteration in the Treg cell population.
# Conclusion
The disturbance in the immunological equilibrium observed in ectopic endometrium and decidua is associated with the alteration in the Treg cell population that occurs in these ectopic tissues.
[fig] Figure 1: Treg cell gating strategy. The characterization of the subpopulation of Treg cells (CD4+CD25+FOXP3+) within the subpopulation of T lymphocytes in eutopic endometrium during the secretory cycle phases (E), ovarian endometriosis (OE), and Fallopian tube pregnancy (EP). Left panel shows the gating strategy for CD4-positive cells (P1). Middle panel is used to set the gate on double positive CD4 + CD25 + cells (P2). On the right panel the histogram representing FoxP3 staining is shown (FoxP3 + ). themselves evade immune surveillance. The immunomodulatory activity of endometrial cells is crucial for the proper course of reproductive processes. [/fig]
|
Musashi2 promotes the progression of pancreatic cancer through a novel ISYNA1‐p21/ZEB‐1 pathway
| INTRODUC TI ONPancreatic cancer (PC) is known as one of the most fatal tumours due to its aggressive malignant aggression in local invasion and distant metastasis. Nowadays, the 5-year survival rate of PC remains 7%-8% and the incidence rate remains increasing although surgery technique and radio-chemotherapy are greatly improved. 1,2 It is imperative to find new molecular biomarkers which can predict the malignant biology and provide new therapeutic strategies for PC patients. MSI2, a translational repressor, is of great importance for self-renewal and pluripotency in embryonic stem cells via regulation of cloning efficiency and differentiation. 3 Abnormal MSI2 expression has been found in chronic myeloid leukaemia, gastric cancer, hepatocellular cancer and colorectal cancers. 4-7 The findings in our former study indicated overexpression of MSI2 facilitated the progression Abstract Our previous studies found overexpression of Musashi2 (MSI2) conduced to the progression and chemoresistance of pancreatic cancer (PC) by negative regulation of Numb and wild type p53 (wtp53). Now, we further investigated the novel signalling involved with MSI2 in PC. We identified inositol-3-phosphate synthase 1 (ISYNA1) as a novel tumour suppressor regulated by MSI2. High MSI2 and low ISYNA1 expression were prevalently observed in 91 PC tissues. ISYNA1 expression was negatively correlated with MSI2 expression, T stage, vascular permeation and poor prognosis in PC patients. What's more, patients expressed high MSI2 and low ISYNA1 level had a significant worse prognosis. And in wtp53 Capan-2 and SW1990 cells, ISYNA1 was downregulated by p53 silencing. ISYNA1 silencing promoted cell proliferation and cell cycle by inhibiting p21 and enhanced cell migration and invasion by upregulating ZEB-1. However, MSI2 silencing upregulated ISYNA1 and p21 but downregulated ZEB-1, which can be rescued by ISYNA1 silencing. Moreover, reduction of cell migration and invasion resulting from MSI2 silencing was significantly reversed by ISYNA1 silencing. In summary, MSI2 facilitates the development of PC through a novel ISYNA1-p21/ZEB-1 pathway, which provides new gene target therapy for PC. K E Y W O R D S cell biology, ISYNA1, MSI2, pancreatic cancer
and chemoresistance of PC through negative regulation of Numb and wild type p53 (wtp53). [bib_ref] Musashi2 promotes the development and progression of pancreatic cancer by down-regulating Numb..., Sheng [/bib_ref] [bib_ref] Cooperation of Musashi-2, Numb, MDM2, and P53 in drug resistance and malignant..., Sheng [/bib_ref] Myoinositol, a kind of water-soluble vitamin synthesized in cells, exhibits various functions, such as seawater acclimation, glucose-lipid metabolism and tumour suppression. [bib_ref] Myo-inositol phosphate synthase expression in the European eel (Anguilla anguilla) and Nile..., Kalujnaia [/bib_ref] [bib_ref] Potential role and therapeutic interests of myo-inositol in metabolic diseases, Croze [/bib_ref] [bib_ref] Myo-inositol treatment increases serum plasmalogens and decreases small dense LDL, particularly in..., Maeba [/bib_ref] [bib_ref] Reduced IL-6 levels and tumor-associated phospho-STAT3 are associated with reduced tumor development..., Unver [/bib_ref] Inositol-3-phosphate synthase 1 (ISYNA1) encodes a necessary limit-enzyme in myoinositol biosynthesis and is a direct target of wtp53 in HCT116 and HepG2 cells. [bib_ref] Regulation of myo-inositol biosynthesis by p53-ISYNA1 pathway, Koguchi [/bib_ref] Meanwhile, ISYNA1 is closely related with the chemotherapeutic resistance in malignant melanoma. [bib_ref] Proteomics analysis of melanoma metastases: association between S100A13 expression and chemotherapy resistance, Azimi [/bib_ref] However, the coordinate role of ISYNA1 and MSI2 in the initiation of PC, to our knowledge, has not been reported, which would be investigated in current study.
## | material s and me thods
## | tissue samples
## | cell lines and culture
We purchased Miapaca-2, PANC-1, AsPC-1, BxPC-2 and SW1990 cells from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China), and Capan-2 cell was received from the American Type Culture Collection. All PC cells were cultured with recommended 1640 or DMEM media including 10% foetal calf serum (HyClone), respectively. As described previously, 9 SW1990 and Capan-2 cells expressed wtp53 whereas Miapaca-2, PANC-1, AsPC-1 and BxPC-2 cells expressed mutant p53 (mtp53).
## | immunohistochemistry
According to the methods described previously, [bib_ref] Expression of SGLT1, Bcl-2 and p53 in primary pancreatic cancer related to..., Casneuf [/bib_ref] [bib_ref] Low expression of Claudin-4 is associated with poor prognosis in esophageal squamous..., Sung [/bib_ref] paraffin samples were used to perform immunohistochemistry (IHC). The intensity of staining was divided into 4 levels: 0 (negative), 1 (weak), 2 (medium), and 3 (strong), while extent of staining was divided into 5 levels: 0 (0%), 1 (1%-25%), 2 (26%-50%), 3 (51%-75%) and 4 (76%-100%) based on the percentage of the positively stained area, and three professional pathologists scored the final IHC staining. In addition, we only scored pancreatic ductal cells; pancreatic acinar and islet cells were excluded when scoring. The products of extent and intensity score were considered as final scores (0-12). The final score > 4 was defined as high ISYNA1 and MSI2 expression.
## | western blot
The whole protein lysates extracted by BCA method were boiling with loading buffer, added into 10% SDS-polyacrylamide gels for electrophoresis, next proteins were transferred into polyvinylidene difluoride membranes(PVDM), then membranes were and β-actin (Proteintech) antibodies at 4°C at least 6 hours or overnight. Membranes were immersed with secondary antibodies (Santa Cruz) at room temperature for 2 hours and finally detected by the ECL kit. The grey values of each band were analysed by Quantity One software. The experiments were repeated three times.
## | qrt-pcr
1ml TRIzol reagent was added into each cell and tissue sample for extraction of total RNA under the manufacturer. Using nucleotide determination, RNA level of each sample was firstly maintained at the same level. Then total RNA was used to synthesize cDNA by the Expand Reverse Transcriptase Kit. Finally, ISYNA1 and GAPDH expression were detected by a Light Cycler 2.0 with the Light Cycler kit. The reaction conditions were denaturation at 95°C for 30 seconds, 45 cycles of amplification at 95°C for 5 seconds and dissolution at 60°C for 45 seconds. The primer sequences were seen in . The relative ISYNA1 expression was counted by ΔCT: ΔCT(ISYNA1)-ΔCT(GAPDH). The experiments were repeated three times.
## | immunofluorescence staining assay
## | construction of msi2 silencing cell lines and isyna1 rna interference
The sequence of Crispr-cas9 and Crispr-sgRNA were designed and synthesized by Genechem (Genechem Co, Ltd), while ISYNA1siRNA, p53siRNA and siRNA control were purchased from GenePharma (GenePharma Co, Ltd). Firstly, Capan-2 and SW1990 cells were infected with Crispr-cas9, next these cells were filtrated by puromycin (Sigma). Then the stable sublines were infected with MSI2-sgRNA (sgMSI2-1/sgMSI2-2) and sgRNA control (scramble) to specifically silence target genes. Capan-2, SW1990 and Miapaca-2 cells were transiently transfected with ISYNA1siRNA, p53siRNA and siRNA control for 48-72 hours with lipofectamine 3000 (Invitrogen) according to the manufacturer. The silencing efficiency of MSI2, ISYNA1 and p53 was verified by WB. All of target sequences were seen in Table S1.
## | flow cytometric analysis of cell cycle
## | cell proliferation assays
After transfection with siRNA control and si-ISYNA1-1 for 48-72 hours, the Capan-2, SW1990 and Miapaca-2 cells were inoculated to 96-well plates at a suitable density (5000-6000 cells per well) and cultured for 1-4 days. Next, 20 μL MTT solution (5 mg/ mL) was added into each well for 4 hours. Then removing cell supernatants and cells were treated with 100 μL dimethyl sulphoxide. The optical density (OD) value at 560 nm was finally calculated by the microplate reader. The experiments were repeated three times. calculated under five random fields at 200× magnification. 5 × 10 6 cells/mL and 2.5 × 10 6 cells/mL of SW1990 cells and 2 × 10 6 cells/ mL and 1 × 10 6 cells/mL of Capan-2 cells were used to perform cell migration and invasion experiments, respectively. The experiments were repeated three times.
## | cell migration and invasion assays
# | statistical analysis
Statistical analyses were conducted by SPSS software 20.0 (SPSS).
The differences between expression of MSI2 and ISYNA1 in pancreatic cancer and adjacent normal pancreatic tissues were analysed by t test. The clinicopathological significance of ISYNA1 and MSI2,
and their relationship were estimated by chi-squared and correlation analysis, respectively. Survival of patients was analysed by the Kaplan-Meier curve, while Log-rank test was used to evaluate the differences. The differences of cell proliferation and cycle, cell migration and invasion were analysed by t test. P < .05 was considered to be statistically significant.
## | re sults
## | exceptional isyna1 expression in pancreatic cancer tissues and cells
In 91 PC tissues and 57 adjacent normal pancreas, ISYNA1 was positively expressed in 36 PC tissues (36/91, 39.5%), which was much lower than that in 57 normal pancreas (33/57, 57.9%) (P < .01).
Similarly, in 57 paired specimens, ISYNA1 expression in PC tissues cells compared with other cell lines . Immunofluorescence staining showed ISYNA1 was expressed in both nucleus and cytoplasm in pancreatic cancer cell lines , which was consistent with the results in IHC assays.
## | association between msi2 and isyna1 expression with patients' parameters and survival time
In our previous study, we proved that MSI2 overexpression was positively associated with tumour size, UICC stage and poor prognosis of pancreatic cancer patients. [bib_ref] Musashi2 promotes the development and progression of pancreatic cancer by down-regulating Numb..., Sheng [/bib_ref] In current study, ISYNA1 expression was negatively related to T stage (P = .035) and vascular permeation (P = .030) in 91 PC patients . Moreover, high expression of MSI2 with low ISYNA1 expression was found in serial pancreatic cancer sections and vice versa . Correlation analysis identified a negative correlation between their expression in pancreatic cancer tissues (P = .032) . Kaplan-Meier indicated that patients expressed high MSI2 or low ISYNA1 had a worse survival time (P = .005 and P = .015, respectively) .
Moreover, patients expressed high MSI2 and low ISYNA1 underwent a much worse prognosis (P = .001) . In addition, univariate analysis showed lymph nodes metastasis and UICC stage were also related to the prognosis of pancreatic cancer patients. And in multivariate analysis, the expression of MSI2 and ISYNA1 were independent unfavourable prognostic indicators for pancreatic cancer .
## | isyna1 was downregulated by wtp53 in capan-2 and sw1990 cells
Capan-2 and SW1990 cells expressed wtp53 while Miapaca-2 cell expressed mutant p53. The relationship between p53 and ISYNA1 in PC cells remained unclear. WB showed p53 silence decreased ISYNA1 expression in wtp53 Capan-2 and SW1990 cells [fig_ref] 3 91: ± 2 [/fig_ref] , but not in mtp53 Miapaca-2 cell [fig_ref] 3 91: ± 2 [/fig_ref]. And similar outcomes were also obtained in PCR assays [fig_ref] 3 91: ± 2 [/fig_ref]. It indicated that wtp53 but not mtp53 had a positive regulation targets ISYNA1.
## | isyna1 silencing promoted cell cycle by repressing p21 in capan-2 and sw1990 cells
## Ta b l e 2 association analysis between msi2 and isyna1 in pc
## | isyna1 silencing promoted cell migration and invasion by upregulating zeb-1 in capan-2 and sw1990 cells
Transwell assay demonstrated that cell migration and invasion was prominently increased in si-ISYNA1 group in comparison to siRNA control group in Capan-2 and SW1990 cells [fig_ref] Total 5 × 10: 6 cells/mL in all pancreatic cancer cells were used for flow... [/fig_ref].
Myoinositol, as a kind of water-soluble vitamin synthesized, acts as a tumour suppressive role in various cancers. However, its function remains unclear in PC. We found that myoinositol significantly suppressed cell migration and invasion in vitro [fig_ref] Total 5 × 10: 6 cells/mL in all pancreatic cancer cells were used for flow... [/fig_ref]. WB
showed an evident upregulation of ZEB-1 in si-ISYNA1 groups compared with control, but other proteins (MSI2, p53, E-cadherin, β-Catenin, N-cadherin and snail-1) were unchanged [fig_ref] Total 5 × 10: 6 cells/mL in all pancreatic cancer cells were used for flow... [/fig_ref].
Altogether, ISYNA1 silence promoted cell migration and invasion by upregulating ZEB-1 in Capan-2 and SW1990 cells.
## | isyna1 silencing reversed reduction of cell migration and invasion caused by msi2 silence in capan-2 and sw1990 cells
Crispr-cas9 mediated MSI2 silencing Capan-2 and SW1990 stable cells were successfully constructed , which were cotransfected with ISYNA1siRNA. Transwell assays indicated cell migration was prominently decreased in MSI2 silencing Capan-2
and SW1990 cells compared with control group. However, ISYNA1
silence obviously reversed reduction of cell migration caused by MSI2 silence . Similarly, ISYNA1 silence reversed the reduction of cell invasion resulted from MSI2 silence in Capan-2 and SW1990 cells . Taken together, MSI2 promoted cell migration and invasion by downregulating ISYNA1 protein in Capan-2 and SW1990 cells.
## | msi2 and isyna1 coordinately regulated p21 and zeb-1 in vitro
We next researched the potential relationship of MSI2, p53,
[formula] ISYNA1 [/formula]
## | d iscuss i on
Our previous studies indicated that MSI2 overexpression made for the progression of pancreatic cancer by negative regulation of Numb and wtp53 with the stimulation of gemcitabine or cisplatin. [bib_ref] Musashi2 promotes the development and progression of pancreatic cancer by down-regulating Numb..., Sheng [/bib_ref] [bib_ref] Cooperation of Musashi-2, Numb, MDM2, and P53 in drug resistance and malignant..., Sheng [/bib_ref] In current study, we further investigated the novel signalling in terms of MSI2 in PC tissue and cell level. We identified ISYNA1 as a novel candidate tumour suppressor regulated by MSI2 and wtp53, which has to been reported, to our knowledge.
ISYNA1 encodes an inositol-3-phosphate synthase enzyme and plays a vital role in cellular myoinositol biosynthesis, which is closely related to the tumorigenesis. 14 In current study, ISYNA1
was downregulated in pancreatic cancer tissues and was negatively related to MSI2 expression, T stage, vascular permeation and poor prognosis of PC patients. Meanwhile, patients expressed high MSI2 and low ISYNA1 underwent a worse prognosis all which is not reported previously to our knowledge. Previous studies focus on ISYNA1 function in gene differential methylation and isoforms identification. [bib_ref] Differential methylation of the gene encodingmyo-inositol 3-phosphate synthase (Isyna1) in rat tissues, Seelan [/bib_ref] [bib_ref] Identification of myo-inositol-3-phosphate synthase isoforms, Seelan [/bib_ref] Recently, a high concentration of myoinositol in tumours is found related to high levels of ISYNA1. [bib_ref] Early prediction of response to Vorinostat in an orthotopic glioma rat model, Wei [/bib_ref] However, studies involved the role of ISYNA1 in malignant tumours remain scarce and controvertial. The expression of ISYNA1 and myoinositol in gliomas was significantly higher than that in primary central nervous system lymphomas. [bib_ref] Myo-inositol concentration in MR spectroscopy for differentiating high grade glioma from primary..., Nagashima [/bib_ref] Overexpression of ISYNA1 suppressed cell proliferation in HCT116 and HepG2 cells. [bib_ref] Regulation of myo-inositol biosynthesis by p53-ISYNA1 pathway, Koguchi [/bib_ref] The inconsistent outcomes might be due to the difference in cancer types and environment.
Our previous study reported that MSI2 silencing upregulated wtp53 only under chemotherapy stimulus. [bib_ref] Cooperation of Musashi-2, Numb, MDM2, and P53 in drug resistance and malignant..., Sheng [/bib_ref] Here, we first found that MSI2 silencing upregulated ISYNA1 and identified ISYNA1 as a direct target of wtp53 in Capan-2 and SW1990 cells. However, in mtp53 Miapaca-2 cell, p53 silence had no effects on ISYNA1. Also, MSI2 silence had no effect in both wtp53 and mtp53 in normal PC cells in our current and previous studies. [bib_ref] Cooperation of Musashi-2, Numb, MDM2, and P53 in drug resistance and malignant..., Sheng [/bib_ref] Thus, ISYNA is a down- ZEB-1, a member of ZHF family, was closely associated with epithelial-mesenchymal transition in many tumours, such as colorectal cancer and PC. [bib_ref] Claudin-1 up-regulates the repressor ZEB-1 to inhibit E-cadherin expression in colon cancer..., Singh [/bib_ref] [bib_ref] The transcription factor ZEB1 (δEF1) promotes tumour cell dedifferentiation by repressing master..., Aigner [/bib_ref] [bib_ref] Epithelial-mesenchymal transition and mesenchymal-epithelial transition via regulation of ZEB-1 and ZEB-2 expression..., Kurahara [/bib_ref] Its abnormal expression contributed to the invasion and metastasis of carcinoma by inducing EMT. [bib_ref] ZEB1: at the crossroads of epithelial-mesenchymal transition, metastasis and therapy resistance, Zhang [/bib_ref] Here, we firstly reported that ISYNA1 silence promoted cell migration and invasion via upregulating ZEB-1 in PC. Furthermore, ISYNA1 silence reversed the reduction of cell migration, cell invasion and ZEB1 protein resulted from MSI2 silence in vitro. Thus, the overexpression of MSI2
facilitates cell migration and cell invasion in pancreatic cancer by regulating ISYNA1-ZEB-1 pathway.
In summary, MSI2 accelerates the development and progression of pancreatic cancer through a novel ISYNA1-p21/ZEB-1 pathway, which supplies a novel gene targets in the PC treatment.
## Ack n owled g em ents
The study was sustained by the Chinese National Science
## Data ava i l a b i l i t y s tat e m e n t
All data will be available upon reasonable request.
## O rci d
Ming Dong https://orcid.org/0000-0003-2685-1420
[fig] Total 5 × 10: 6 cells/mL in all pancreatic cancer cells were used for flow cytometry. 1 mL precooled 75% ethanol was added to fix the cells at 4°C overnight after transfection. Then cells were added with 500 μL premixed staining solution (500 μL buffer, 25 μL propidium iodide and 10 μL RNase A) and underwent water bath at 37°C for 30 minutes in darkness. CytoFLEX flow cytometry was applied to detect cell cycle with red fluorescence at wavelength of 488 nm and Flow Jo 7.6 software was used to analyse the results. Experiments were repeated three times. [/fig]
[fig] 3 91: ± 2.38) was lower than that in paired normal tissues (5.72 ± 3.16) (P < .01) (Figure 1A,B). Meanwhile, both qRT-PCR and WB showed ISYNA1 mRNA and protein level in 29 paired fresh pancreatic cancer tissues were lower than that in corresponding normal pancreatic F I G U R E 2 The relationship of MSI2 and ISYNA1 with the survival of 91 postoperative PC patients in IHC and Kaplan-Meier analysis. A, IHC showed negative expression of MSI2 and ISYNA1 in PC tissues (200×). B and C, Kaplan-Meier survival curves in PC patients with high and low ISYNA1, MSI2 expression levels, respectively. D, The negative relationship of MSI2 and ISYNA1 was plotted against overall survival time tissues, respectively (P < .01; P = . [/fig]
[fig] 01 F: of ISYNA1 silencing on G1 phase and S + G2 phase in Capan-2 and SW1990 cells, respectively. E, Effects of ISYNA1 silencing on cell proliferation in Capna-2,SW1990 and Miapaca-2 cells. F and G, Knocking down of ISYNA1 downregulating p21 in Capan-2 and SW1990 cells, respectively. Bars indicate Means ± SE,*P < .05, **P < .100×). E and F, WB analysis showed ZEB-1 was upregulated in ISYNA1 knocked down PC cells. Bars indicate Means ± SE,*P < .05, **P < .01 [/fig]
[fig] Foundation: No. 81672835). Lei Zhou performed the research; Lei Zhou and WeiWei Sheng wrote the paper and discussed the data; Xiaoyang Shi, Chao Jia and Rongxian Cao added additional experiments and performed the statistical analysis; Guosen Wang, Yiheng Lin, Fang Zhu and Qi Dong were performed for administrative, technical and material support; Ming Dong and Weiwei Sheng work for the study design and concept. Ming Dong supervised the total project.CO N FLI C T O F I NTE R E S TNo conflicts of interest are confirmed.AUTH O R CO NTR I B UTIO N Lei Zhou: Conceptualization (equal); Data curation (lead); Formal analysis (lead); Investigation (equal); Methodology (equal); Software (equal); Writing-original draft (lead); Writing-review & editing (lead). WeiWei Sheng: Conceptualization (lead); Data curation (lead); Formal analysis (lead); Methodology (equal); Validation (equal); Visualization (equal); Writing-original draft (equal); Writing-review & editing (equal). Chao Jia: Data curation (equal); Formal analysis (equal); Methodology (equal); Visualization (equal). Xiaoyang Shi: Investigation (equal); Methodology (equal); Software (equal). Rongxian Cao: Data curation (equal); Methodology (equal). Guosen Wang: Formal analysis (supporting); Methodology (supporting). Yiheng Lin: Formal analysis (supporting); Methodology (supporting). Fang Zhu: Supervision (supporting); Validation (supporting). Qi Dong: Supervision (supporting); Validation (supporting). Ming Dong: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Project administration (equal); Resources (equal); Supervision (equal); Validation (equal); Writing-original draft (equal); Writing-review & editing (equal). [/fig]
|
Current Status of Research on Osteoporosis after Solid Organ Transplantation: Pathogenesis and Management
Improved survival following organ transplantation has brought to the forefront some long-term complications, among which osteoporosis and associated fractures are the major ones that adversely affect the quality of life in recipients. The pathogenesis of osteoporosis in transplant recipients is complex and multifactorial which may be related to increased bone resorption, decreased bone formation, or both. Studies have shown that the preexisting underlying metabolic bone disorders and the use of immunosuppressive agents are the major risk factors for osteoporosis and fractures after organ transplantation. And rapid bone loss usually occurs in the first 6-12 months with a significant increase in fracture risk. This paper will provide an updated review on the possible pathogenesis of posttransplant osteoporosis and fractures, the natural history, and the current prevention and treatment strategies concerning different types of organ transplantation.
# Introduction
Solid organ transplantation (SOT) has become an effective and established clinical therapy for end-stage renal, liver, cardiac, and pulmonary disease over the past two decades. Meanwhile, patient and graft survivals have been greatly improved due to tremendous advances in immunosuppressive agents and surgical techniques. However, improved survival has also made the adverse effects of some major complications apparent, such as osteoporosis and osteoporotic fractures.
The most rapid decrease in bone mineral density (BMD) frequently occurs in the first 6-12 months after transplantation and is accompanied with a marked increase in fracture risk. A recent study found that, compared with the non-SOT patients, the overall hazard ratio (HR) of osteoporosis after SOT was 5.14 (95% CI, 3. , and the HR of related fractures was 5.76 (95% CI, 3.80-8.74) [bib_ref] Osteoporosis and fractures after solid organ transplantation: a nationwide population-based cohort study, Yu [/bib_ref]. The highest risk of osteoporosis and fractures was observed in lung transplant recipients, followed by other types of SOT [bib_ref] Osteoporosis and fractures after solid organ transplantation: a nationwide population-based cohort study, Yu [/bib_ref].
Studies have shown that pretransplant bone disease and posttransplant immunosuppressant therapy are the two major risk factors for osteoporosis. In addition, other factors including aging, tobacco, alcohol, nutritional deficiencies, immobility, and hypogonadism may also contribute to the risk of osteoporosis.
In this review, we will summarize the prevalence of osteoporosis and fractures and the current understanding on its etiology and natural history and will focus on the latest prevention and treatment strategies for osteoporosis and fractures in SOT recipients.
## The effects on skeleton of
Immunosuppressive Agents 2.1. Glucocorticoids. The therapeutic use of glucocorticoids (GCs) in patients has been shown related to an increase in bone loss and fracture risk [bib_ref] Prevalence of vertebral deformities and symptomatic vertebral fractures in corticosteroid treated patients..., De Nijs [/bib_ref]. Trabecular bone is usually more affected than cortical sites in glucocorticoid-induced osteoporosis (GIO) [bib_ref] Advanced imaging assessment of bone fragility in glucocorticoid-induced osteoporosis, Kalpakcioglu [/bib_ref]. Patients who underwent organ transplantation are usually treated with high-dose glucocorticoid in the first few months and then taped gradually. However studies found that substantial bone loss occurring in early posttransplantation period has been largely attributed to treatment with high-dose glucocorticoid immediately after transplantation [bib_ref] Approach to the patient with transplantationrelated bone loss, Ebeling [/bib_ref]. Daily doses of 7.5 mg prednisone equivalent are generally considered a major cause for bone loss, even 2 BioMed Research International though lower doses have been reported as harmful to bone health [bib_ref] Risk of fractures with inhaled corticosteroids in COPD: systematic review and meta-analysis..., Loke [/bib_ref] [bib_ref] Use of oral corticosteroids and risk of fractures, Van Staa [/bib_ref]. Glucocorticoids in pharmacological dosages have an adverse effect on bone metabolism via both direct and indirect mechanisms [bib_ref] Glucocorticoid-induced osteoporosis and osteonecrosis, Weinstein [/bib_ref] [bib_ref] Glucocorticoids, osteocytes, and skeletal fragility: the role of bone vascularity, Weinstein [/bib_ref]. The direct effects of glucocorticoids on bone metabolism include influences on osteoblasts, osteoclasts, and osteocytes [bib_ref] Minireview: live and let die: molecular effects of glucocorticoids on bone cells, Hofbauer [/bib_ref]. Studies reported that glucocorticoids may change the ratio between receptor activator of NF-B ligand (RANKL) and osteoprotegerin (OPG) in stromal and osteoblastic cells, which results in the induction of osteoclastogenesis and increased bone resorption [bib_ref] Effects of dissociated glucocorticoids on OPG and RANKL in osteoblastic cells, Humphrey [/bib_ref] [bib_ref] Corticosteroid-induced osteoporosis: an update for dermatologists, Clarke [/bib_ref]. There is also evidence that GCs directly upregulate the expression of CSF-1, an essential cytokine for the survival of osteoclast precursors, leading to GC-induced born resorption [bib_ref] Bisphosphonates in the treatment of glucocorticoid-induced osteoporosis: pros, Rossini [/bib_ref].
The principal pathophysiological mechanism of GIO is decreased bone formation, owing to suppression of osteoblast differentiation and function, and promotion of osteoblast apoptosis [bib_ref] Prednisone affects inflammation, glucose tolerance, and bone, turnover within hours of treatment..., Kauh [/bib_ref]. Several molecular pathways underlying GCinduced reduction in differentiation potential have been identified, such as Wnt/ -catenin and bone morphogenetic protein 2 (BMP-2) signaling [bib_ref] Wnt/beta-catenin signaling mediates osteoblast differentiation triggered by peptideinduced alpha5beta1 integrin priming in..., Saidak [/bib_ref] [bib_ref] KMUP-1 promotes osteoblast differentiation through cAMP and cGMP pathways and signaling of..., Liou [/bib_ref]. The therapeutic GCs could suppress both Wnt/beta-catenin and BMP-2 signaling, which leads to a significant reduction in osteoblastogenesis and consequently to a loss of osteoblast-generated proteins [bib_ref] Glucocorticoid receptor signaling in bone cells, Moutsatsou [/bib_ref]. GCs have also been reported to induce osteoblast apoptosis via an increase in proapoptotic factors of the Bcl-2 family such as Bim [bib_ref] Glucocorticoid induced osteoblast apoptosis by increasing E4BP4 expression via up-regulation of Bim, Chen [/bib_ref] and a reduction in 1-integrin expression [bib_ref] Effect of dexamethasone on human osteoblasts in culture: involvement of 1 integrin..., Naves [/bib_ref] , which eventually leads to bone loss and an increase in fracture risk.
Glucocorticoids could also indirectly affect bone and mineral homeostasis through various ways involving decreasing calcium absorption from the intestine, increasing renal excretion of calcium [bib_ref] Short-term effects of glucocorticoid therapy on biochemical markers of bone metabolism in..., Kuroki [/bib_ref] , reducing growth hormone secretion [bib_ref] Glucocorticoids and the regulation of growth hormone secretion, Mazziotti [/bib_ref] , and decreasing gonadal secretion of estrogen and androgens [bib_ref] Glucocorticoid-induced osteoporosis: pathophysiology and therapy, Canalis [/bib_ref] [bib_ref] Steroids and osteoporosis: the quest for mechanisms, Manolagas [/bib_ref] , which may lead to osteoporosis and fractures. However, the main mechanisms underlying GIO are attributed to the direct influence of GCs on bone cells [bib_ref] Glucocorticoids and bone: local effects and systemic implications, Henneicke [/bib_ref].
## Calcineurin inhibitors: cyclosporine and tacrolimus.
The introduction of cyclosporine A (CsA) to immunosuppressive therapy in the early 1980s has significantly improved patient and allograft survival after organ transplantation [bib_ref] Cyclosporine, Kahan [/bib_ref]. CsA inhibits calcineurin and reduces T-cell function via suppression of regulatory genes expressing products such as interleukin-2, interleukin receptors, and the protooncogenes H-ras and c-myc [bib_ref] Post-transplantation bone disease: the role of immunosuppressive agents and the skeleton, Epstein [/bib_ref]. The risk of fractures increases significantly in many patients treated with CsA because of rapid bone loss after transplantation [bib_ref] Bone loss in long-term renal transplantation: histopathology and densitometry analysis, Cueto-Manzano [/bib_ref] [bib_ref] Osteoporosis after liver transplantation, Compston [/bib_ref] [bib_ref] Bone mineral density and fracture prevalence in long-term kidney graft recipients, Durieux [/bib_ref]. Previous studies have proved that the effect of CsA on bone mineral metabolism is dependent on both the dose and the duration of administration [bib_ref] Cyclosporin-a in vivo produces severe osteopenia in the rat: effect of dose..., Movsowitz [/bib_ref] [bib_ref] Cyclosporin A elicits dose-dependent biphasic effects on osteoblast differentiation and bone formation, Yeo [/bib_ref]. Studies in rats have indicated that CsA therapy at 7.5 mg/kg can damage the mechanical properties and thus may increase the fracture risk of lumbar spine [bib_ref] Effects of cyclosporina on rat skeletal biomechanical properties, Chen [/bib_ref]. Some in vivo study suggested that CsA may contribute to high-turnover bone loss, leading to an uncoupling of the dynamic bone remodeling cycle with resorption exceeding formation [bib_ref] Immunosuppressant use without bone loss-implications for bone loss after transplantation, Goodman [/bib_ref]. However, other scholars reported a lack of bone loss in 13 renal transplant patients receiving cyclosporine in a steroid-free regimen [bib_ref] Osteoporosis after organ transplantation, Ponticelli [/bib_ref]. More interestingly, CsA monotherapy in renal transplant patients can significantly increase lumbar spine BMD while recipients treated with both CsA and glucocorticoids present a significant decrease [bib_ref] Effects of three immunosuppressive regimens on vertebral bone density in renal transplant..., Aroldi [/bib_ref]. Up to now, the precise mechanisms involved in CsA induced bone loss are still not well defined.
Tacrolimus (FK506), another calcineurin inhibitor, inhibits T-cell activation and proliferation and cytokine gene expression and also causes trabecular bone loss in rats [bib_ref] Post-transplantation bone disease: the role of immunosuppressive agents and the skeleton, Epstein [/bib_ref]. The possible mechanism of FK506 on bone metabolism is mainly due to an excess of bone resorption over formation involving the imbalance of RANK/RANKL/OPG system [bib_ref] Impact of sirolimus, tacrolimus and mycophenolate mofetil on osteoclastogenesis-implications for post-transplantation bone..., Westenfeld [/bib_ref] or partly by inhibiting ERK1/2 pathway [bib_ref] Immunosuppressant FK506 decreases the intracellular magnesium in the human osteoblast cell by..., Jeon [/bib_ref]. Both liver [bib_ref] Bone loss after orthotopic liver transplantation: FK 506 versus cyclosporine, Park [/bib_ref] and cardiac [bib_ref] Rapid trabecular bone loss after cardiac transplantation using FK506 (tacrolimus)-based immunosuppression, Stempfle [/bib_ref] transplant recipients have been shown to sustain rapid bone loss with tacrolimus-based immunosuppression. However, it has been found that lesser reduction in bone mineral density occurs with FK506, compared with CsA in rats [bib_ref] Lesser reduction in bone mineral density by the immunosuppressant, FK506, compared with..., Inoue [/bib_ref]. Similarly, liver transplant treatment with FK506 showed a more favorable long-term effect on bone mass evolution than CsA therapy, probably due to the lower dose of glucocorticoids used in the FK506 [bib_ref] Bone mass and mineral metabolism in liver transplant patients treated with FK506..., Monegal [/bib_ref]. A recent study of renal transplant patients found that those with high FK506 blood concentrations were at greater risk of bone loss [bib_ref] Impact of tacrolimus on bone metabolism after kidney transplantation, Luo [/bib_ref].
## Mammalian target of rapamycin inhibitors (mtori):
Sirolimus and Everolimus. Sirolimus (SIR) and everolimus (EVE) are anti-mTOR drugs commonly used in transplant recipients. It is well known that mTORi present both antiproliferative and antiangiogenic activities; thus their interference with bone metabolism is unavoidable. Alvarez-Garcia et al. found that SIR slowed down growth and caused marked alterations in the growth plate of young rats [bib_ref] Rapamycin retards growth and causes marked alterations in the growth plate of..., Alvarez-Garcia [/bib_ref]. In a cross-section study of renal transplant recipients, SIRbased regimen was associated with lower serum markers of osteoclast activity and maturation in contrast to a CIbased immunosuppression [bib_ref] Impact of sirolimus, tacrolimus and mycophenolate mofetil on osteoclastogenesis-implications for post-transplantation bone..., Westenfeld [/bib_ref] , which was in agreement with the experiment data from other researchers [bib_ref] Bone metabolism in renal transplant patients treated with cyclosporine or sirolimus, Campistol [/bib_ref]. In a previous animal study, SIR was reported to be a bone sparing immunosuppressant compared with tacrolimus and cyclosporine after administration for 28 days in rats [bib_ref] Rapamycin: a bone sparing immunosuppressant?, Romero [/bib_ref]. mTOR/S6 K has been identified as an essential signaling pathway engaged in the stimulation of cell survival in osteoclasts [bib_ref] M-CSF, TNF and RANK ligand promote osteoclast survival by signaling through mTOR/S6..., Glantschnig [/bib_ref]. Kneissel et al. studied the impact of EVE on mouse and human osteoclasts in vitro and on bone in an ovariectomized rat model and found EVE can suppress cancellous bone loss, bone resorption, and cathepsin K expression by osteoclasts [bib_ref] Everolimus suppresses cancellous bone loss, bone resorption, and cathepsin K expression by..., Kneissel [/bib_ref]. Westenfeld et al. also believed that EVE may exert beneficial bone effects, potentially reducing bone resorption and contributing to a bone-protective effect [bib_ref] Impact of sirolimus, tacrolimus and mycophenolate mofetil on osteoclastogenesis-implications for post-transplantation bone..., Westenfeld [/bib_ref].
## Other immunosuppressive agents.
Immunosuppressive drugs such as MMF and azathioprine had no deleterious effect on bone in rat model [bib_ref] Mycophenolate mofetil: a promising new immunosuppressant that does not cause bone loss..., Dissanayake [/bib_ref] [bib_ref] Azathioprine alone is bone sparing and does not alter cyclosporin A-induced osteopenia..., Bryer [/bib_ref]. There is little information available regarding the skeleton effects of other immunosuppressant agents. [fig_ref] Table 1: Effects of immunosuppressive agents on skeleton [/fig_ref] shows the effects of immunosuppressive drugs on skeleton.
## Different types of transplantation
## Renal transplantation.
Patients with end-stage renal disease (ESRD) are usually suffering from renal osteodystrophy [bib_ref] Evidence for a pth-independent humoral mechanism in post-transplant hypophosphatemia and phosphaturia, Green [/bib_ref]. It has been reported that renal osteodystrophy persists after transplantation because of elevated levels of PTH and fibroblast growth factor 23 (FGF-23), though at lower levels compared to pretransplantation [bib_ref] Recovery of hyperphosphatoninism and renal phosphorus wasting one year after successful renal..., Evenepoel [/bib_ref]. Several factors have been demonstrated to be associated with osteopenia after renal transplantation including preexisting bone diseases, secondary hyperparathyroidism, and low bone mineral density due to advanced age and/or vitamin D deficiency, as well as posttransplant immunosuppressive therapy, steroid dose, and ongoing disorder in the phosphate-calcium-parathyroid hormone-vitamin D axis [bib_ref] CKDmineral and bone disorder management in kidney transplant recipients, Alshayeb [/bib_ref] [bib_ref] Bone disease after renal transplantation, Weisinger [/bib_ref]. The loss of bone mass is particularly prominent during the first 6 months after renal transplantation and bone mineral density (BMD) decreases to a mean of 5.5%-19.5% [bib_ref] Bone mineral density changes within six months of renal transplantation, Mikuls [/bib_ref] and 2.6%-8.2% between the 6th and 12th month [bib_ref] Early rapid loss followed by long-term consolidation characterizes the development of lumbar..., Brandenburg [/bib_ref]. After 1 year, it is only 0.4%-4.5% of the mean [bib_ref] Early rapid loss followed by long-term consolidation characterizes the development of lumbar..., Brandenburg [/bib_ref]. The prevalence of "osteopenia" and "osteoporosis" after renal transplant has been reported to range from 25% to 35% and 8% to 28%, respectively [bib_ref] Bone mineral density after simultaneous kidney-pancreas transplantation: four years follow-up of 57..., Pereira [/bib_ref] [bib_ref] Evolution of bone disease at 2 years after transplantation: a single-center study, Falkiewicz [/bib_ref] [bib_ref] The prevalence and risk factors of osteoporosis in Thai renal-transplant patients, Techawathanawanna [/bib_ref]. A cross-sectional analysis of 389 stable renal transplant recipients has showed that osteopenia or osteoporosis is more prevalent at the femoral neck than lumbar spine [bib_ref] Determinants of bone mineral density in stable kidney transplant recipients, Gupta [/bib_ref]. The main pathology of bone remodeling following transplantation is considered to include decreased bone formation and mineralization associated with persistent bone resorption [bib_ref] Bone and mineral disorders after kidney transplantation: therapeutic strategies, Molnar [/bib_ref].
The incidence rate of fractures has been reported to vary between 7% and 11% in nondiabetic renal transplant recipients but is much higher in patients with diabetic nephropathy [bib_ref] Risk of hip fracture among dialysis and renal transplant recipients, Ball [/bib_ref]. The recipients experienced a 3.6-3.8-fold higher risk of fracture over general population, and it is 30% higher in the first 3 years after transplantation than patients undergoing dialysis [bib_ref] Risk of hip fracture among dialysis and renal transplant recipients, Ball [/bib_ref]. And the recipients fractured mainly in appendicular sites (hips, long bones, ankles, and feet) rather than axial sites (lumbar spine and ribs) [bib_ref] Long-term fracture risk following renal transplantation: a population-based study, Vautour [/bib_ref]. Sukumaran Nair et al. reported that the hip fracture incidence was almost 5 events per 1000 person-years in the first posttransplant year, then closer to 3 events per 1000 person-years in the 2 subsequent years [bib_ref] Temporal trends in the incidence, treatment and outcomes of hip fracture after..., Nair [/bib_ref]. Patel et al. studied 165 renal transplant patients and found that 27 patients had either vertebral deformities or a history of a low trauma fracture after transplantation [bib_ref] Prevalence and causes of low bone density and fractures in kidney transplant..., Patel [/bib_ref]. A systematic review of 262,678 recipients showed that the 5-year cumulative incidence of fracture varied ranging from 0.85% to 27% [bib_ref] Fracture risk in kidney transplant recipients: a systematic review, Naylor [/bib_ref]. Recently, Sukumaran Nair et al. reported that case-mix adjusted hip fracture rates after renal transplantation declined substantially since 1997, partly benefitting from the improvements in immunosuppressive therapy [bib_ref] Temporal trends in the incidence, treatment and outcomes of hip fracture after..., Nair [/bib_ref].
A longitudinal study displayed that persistently high PTH level (>130 ng/L) at 3 months after transplantation was a major independent risk factor for fracture [bib_ref] Persistent hyperparathyroidism is a major risk factor for fractures in the five..., Perrin [/bib_ref]. Nikkel et al. identified some other factors that were also associated with increased fracture risk, such as older age, white race, prior dialysis, and pretransplant fracture [bib_ref] Pancreas-kidney transplantation is associated with reduced fracture risk compared with kidney-alone transplantation..., Nikkel [/bib_ref].
## Cardiac transplantation.
Osteoporosis is very common among patients with end-stage heart disease ranging between 8% and 23% [bib_ref] Bone density in heart or lung transplant recipients-a longitudinal study, Wang [/bib_ref] [bib_ref] Osteoporosis is a prevalent finding in patients with solid organ failure awaiting..., Dolgos [/bib_ref]. Longitudinal studies have found that BMD may decrease by 10% to 20% before cardiac transplantation [bib_ref] Loss of vertebral bone density in heart transplant patients, Muchmore [/bib_ref]. Rapid bone loss occurs most often in the first year after transplantation. Lumbar spinal BMD declines by 3% to 10% and femoral neck BMD by 6% to 11% during the first year [bib_ref] Post-transplantation osteoporosis, Stein [/bib_ref] , which is considerable when compared with the 1.41% decrease at the lumbar spine and 0.35% at the femoral neck in healthy population [bib_ref] Timing and quantification of bone loss in cardiac transplant recipients, Van Cleemput [/bib_ref]. BMD begins to stabilize in the second year and may even increase after the third year [bib_ref] Post-transplantation osteoporosis, Stein [/bib_ref]. Bone loss occurring shortly after heart transplantation is probably due to immunosuppressive therapy with highdose glucocorticoids and calcineurin inhibitors, particularly cyclosporine [bib_ref] Osteoporosis after solid organ and bone marrow transplantation, Cohen [/bib_ref]. Moreover, vitamin D deficiency and testosterone deficiency (in men) have also been considered to be associated with more severe bone loss [bib_ref] Approach to the patient with transplantationrelated bone loss, Ebeling [/bib_ref].
The prevalence of fragility fractures is estimated between 22% and 44% among cardiac transplant recipients, and the incidence of fractures during the first three years after transplantation ranges from 18% to 35% [bib_ref] Frequency and predictors of osteoporotic fractures after cardiac or liver transplantation: a..., Leidig-Bruckner [/bib_ref] [bib_ref] Fracture after cardiac transplantation: a prospective longitudinal study, Shane [/bib_ref]. The relationship between BMD and fractures was reported as weak, because lots of patients underwent a fracture at a relatively normal BMD [bib_ref] Fracture after cardiac transplantation: a prospective longitudinal study, Shane [/bib_ref]. In some study, a low incidence of fracture was found even though bone loss was evident [bib_ref] Unexpected low incidence of vertebral fractures in heart transplant recipients: analysis of..., Kerschan-Schindl [/bib_ref].
In a cross-sectional study of 157 patients within less than 10 years from cardiac transplantation, the authors found fractures by spine X-ray in 40% of subjects, whereas osteoporosis by DXA was present only in 13% of lumbar spine and 25% of hip scans. While -score threshold was modified to −1.5 standard deviations, the prevalence of fractured increased significantly, reaching 60%. Thus they thought that the standard densitometric criteria were unreliable to identify bone fragility after cardiac transplantation, and a different threshold should be considered [bib_ref] Densitometric threshold and vertebral fractures in heart transplant patients, Carbonare [/bib_ref].
In a longitudinal study by Wang et al., multivariate analysis showed that pretransplant BMD at either site and/or a history of fracture were significant predictors for developing osteopenia or osteoporosis after cardiac transplantation [bib_ref] Bone density in heart or lung transplant recipients-a longitudinal study, Wang [/bib_ref]. Other authors found, besides lumbar spine BMD before transplantation, age was also an important predictor of fracture [bib_ref] Frequency and predictors of osteoporotic fractures after cardiac or liver transplantation: a..., Leidig-Bruckner [/bib_ref].
## Liver transplantation.
Bone disorders in patients with chronic liver disease are termed as hepatic osteodystrophy, which includes osteoporosis and, less frequently, osteomalacia. The prevalence of bone disease has been reported to range from 12% to 55% among patients with cirrhosis [bib_ref] Bone disorders in chronic liver disease, Collier [/bib_ref] [bib_ref] Bone disorders in patients with chronic liver disease awaiting liver transplantation, Loria [/bib_ref] [bib_ref] Pathogenesis of osteoporosis in liver cirrhosis, Giouleme [/bib_ref] [bib_ref] Bone mineral density among cirrhotic patients awaiting liver transplantation, Sokhi [/bib_ref]. In a study by Bai et al., the risk of low BMD was significantly correlated with the primary liver disease and elevation of serum iPTH level when using univariate analysis [bib_ref] Elevation of intact parathyroid hormone level is a risk factor for low..., Bai [/bib_ref]. Bone disorders after liver transplantation can be related to many risk factors, for example, preexisting hepatic osteodystrophy, malnutrition, immobility, hypogonadism, and immunosuppressant.
Following liver transplantation, a rapid decline in BMD mainly occurs in the first 3 to 6 months [bib_ref] Bone mineral density before and after OLT: long-term followup and predictive factors, Guichelaar [/bib_ref] [bib_ref] Osteoporosis in liver diseases and after liver transplantation, Hay [/bib_ref] and is associated with a significant increase in the incidence of osteoporotic fractures. One bone histology study has demonstrated that bone loss halts approximately 6 months after liver transplantation and tends to increase, especially in lumbar spine, which leads to the recovery of bone mass within 2 years after the surgery [bib_ref] Bone disease after liver transplantation: a long-term prospective study of bone mass..., Monegal [/bib_ref]. The raise in BMD is significantly higher among premenopausal than perimenopausal and postmenopausal women, probably owing to the protective estrogen effect on the skeleton [bib_ref] Decrease in bone mass in women after liver transplantation: associated factors, Baccaro [/bib_ref].
The fracture incidence after successful liver transplantation is relatively high and ranges between 10% and 43% [bib_ref] Frequency and predictors of osteoporotic fractures after cardiac or liver transplantation: a..., Leidig-Bruckner [/bib_ref] [bib_ref] Incidence of vertebral fractures in the first three months after orthotopic liver..., Ninkovic [/bib_ref]. The majority of fractures take place during the first 2 years, mostly in lumbar spine [bib_ref] Bone disorders in chronic liver disease, Collier [/bib_ref]. Preexistent low BMD before transplantation has been regarded as the major risk factor for posttransplantation fractures [bib_ref] Elevation of intact parathyroid hormone level is a risk factor for low..., Bai [/bib_ref]. Besides, some conditions such as pretransplant fractures as well as older age, the type of liver disease, and retransplantation also contribute to the development of fractures [bib_ref] Liver and bone, Guanabens [/bib_ref].
## Lung transplantation.
Osteoporosis is prevalent in patients who are candidates for lung transplantation, especially among those with chronic obstructive pulmonary disease (COPD). The prevalence of osteoporosis and osteopenia varied as 9-69% and 27-67% in COPD patients, respectively [bib_ref] Current status of research on osteoporosis in COPD: a systematic review, Graat-Verboom [/bib_ref]. Contributing factors include older age, female gender, low BMI, history of fractures, and tobacco exposure [bib_ref] Osteoporosis risk in patients with chronic obstructive pulmonary disease: the EOLO study, Maggi [/bib_ref]. Steroid consumption has been regarded as a major risk factor [bib_ref] Osteoporosis risk in patients with chronic obstructive pulmonary disease: the EOLO study, Maggi [/bib_ref] [bib_ref] Osteoporosis in patients referred for lung transplantation, Jastrzebski [/bib_ref]. Furthermore, raised high density lipoprotein cholesterol levels are also considered to be independently associated with osteoporosis [bib_ref] Elevated HDL cholesterol levels are associated with osteoporosis in lung transplant candidates..., Reed [/bib_ref].
Rates of bone loss at lumbar spine and femoral neck vary between 2% and 5% in the first year after lung transplantation, respectively [bib_ref] Bone density in heart or lung transplant recipients-a longitudinal study, Wang [/bib_ref]. Pretransplant BMD is an important predictor of subsequent osteopenia or osteoporosis after transplantation [bib_ref] Bone density in heart or lung transplant recipients-a longitudinal study, Wang [/bib_ref].
There are limited studies evaluating fracture risk following lung transplantation. In an earlier study, symptomatic fracture rated up to 15-50% and a loss of BMD of 4% to 12% occurred in the first year after lung transplantation [bib_ref] Osteoporosis in patients undergoing lung transplantation, Ferrari [/bib_ref]. In a very recent research, Hariman et al. retrospectively analyzed 210 medical records of patients receiving lung transplantation and found 17 patients (8.0%) had fractures after transplantation. The incidence of fracture in lung transplant patients in years 1 to 5 was 4.8%, 1.0%, 0.0%, 1.0%, and 1.9%, respectively [bib_ref] Incidence of fractures after cardiac and lung transplantation: a single center experience, Hariman [/bib_ref] , probably owning to the early initiation of antiresorptive therapy. Another study conducted by showed that antiresorptive therapy with pamidronate preserved or increased BMD prior to and following transplantation, resulting in only 4% of new fractures [bib_ref] Prevention of bone loss and fracture after lung transplantation, Cahill [/bib_ref].
## Management of bone disease before and after organ transplantation
## Pretransplantation assessment.
Since bone disease is very common in patients with end-stage organ disease, all patients awaiting transplantation should be given comprehensive evaluation. Bone densitometry of the lumbar spine and hip measured by DXA should be performed routinely. Spine radiographs should be measured to diagnose prevalent fractures. Any secondary causes of osteoporosis should be identified and treated. Some conditions such as hypogonadism, vitamin D deficiency, and secondary hyperparathyroidism must be corrected whenever possible. In addition, candidates for transplant should be encouraged to change detrimental lifestyle such as immobilization, smoking, and alcohol abuse. Furthermore, all patients should be administered the recommended daily allowance for calcium (1000-1500 mg/d) and vitamin D (400-800 IU/d).
## Posttransplantation management.
It is well known that both pretransplant bone disorders and immunosuppressive agents lead to rapid bone loss and increased fracture risk following transplantation. Therefore, prevention therapy should be adopted immediately after transplantation. Perhaps the low-dose use of GCs in a shortest period is of the greatest importance. Besides, at present, there are several kinds of drugs available for prevention and treatment of posttransplant osteoporosis.
## Active vitamin d metabolites.
Vitamin D deficiency is prevalent in pre-and posttransplant patients. In a BioMed Research International 5 cross-sectional study of 173 renal transplant recipients, moderate-to-severe vitamin D deficiency was observed in 29% of the patients, with a further 51% of insufficiency [bib_ref] Vitamin D status in kidney transplant patients: need for intensified routine supplementation, Ewers [/bib_ref]. The active vitamin D metabolites have been proved effective on the prevention and treatment of osteoporosis. The probable underlying mechanisms include that they may overcome GC-induced decreases in intestinal calcium absorption, reduce the potential for secondary hyperparathyroidism, promote the differentiation and maturation of osteoblast precursors, and potentiate the immunosuppressive action of CsA [bib_ref] Reduction of immunosuppressant therapy requirement in heart transplantation by calcitriol, Briffa [/bib_ref]. Torres et al. showed in a double-blind, randomized and controlled trial that renal transplant recipients who received intermittent calcitriol (0.5 g/48 h) for 3 months, plus oral calcium supplementation (0.5 g/day) during 1 year, had less bone loss than those with calcium alone [bib_ref] Treatment with intermittent calcitriol and calcium reduces bone loss after renal transplantation, Torres [/bib_ref]. Similarly, other authors observed that treatment with active vitamin D (0.25 g/d) and calcium could prevent bone loss at the lumbar spine and proximal femur during the first 6 months after renal transplantation [bib_ref] Treatment with vitamin D and calcium reduces bone loss after renal transplantation:..., De Sévaux [/bib_ref].
However, there are some contradictory results. Another double-randomized study of heart or lung transplant recipients who were given placebo or calcitriol (0.5-0.75 g/d) for either 1 or 2 years after transplantation found that bone loss at the proximal femur was significantly reduced compared with placebo group, but there was no significant difference in lumbar spine BMD between groups [bib_ref] Effect of calcitriol on bone loss after cardiac or lung transplantation, Sambrook [/bib_ref].
## Bisphosphonates.
Bisphosphonates are currently the most commonly used medications for the treatment of osteoporosis. They inhibit osteoclast-mediated bone resorption through several mechanisms as yet not fully elucidated. Oral or intravenous bisphosphonates have been regarded as the most promising approach for the management of posttransplant osteoporosis [bib_ref] Approach to the patient with transplantationrelated bone loss, Ebeling [/bib_ref]. Compared head-to-head with vitamin D sterols, bisphosphonates had greater efficacy on improving bone mineral density after transplantation [bib_ref] Interventions for preventing bone disease in kidney transplant recipients, Palmer [/bib_ref].
In a recent meta-analysis regarding the value of pamidronate to prevent bone loss of lumbar spine and femoral neck in the first year of renal transplantation, 6 randomized trials including 281 patients were identified. Among all 6 trials, administration of pamidronate was associated with significant reduction of bone loss in the lumbar spine compared with the control group. No major side effects were reported [bib_ref] Clinical efficacy and safety of pamidronate therapy on bone mass density in..., Wang [/bib_ref]. Similar results have been observed in a randomized controlled trial of patients with liver transplantation [bib_ref] Pamidronate in the prevention of bone loss after liver transplantation: a randomized..., Monegal [/bib_ref] [bib_ref] Better late than never? Experience with intravenous pamidronate treatment in patients with..., Dodidou [/bib_ref] and in studies of cardiac [bib_ref] Better late than never? Experience with intravenous pamidronate treatment in patients with..., Dodidou [/bib_ref] and lung [bib_ref] Prevention of bone loss and fracture after lung transplantation, Cahill [/bib_ref] transplant recipients independent of the time following transplantation.
A 1-year randomized, double-blind, placebo-controlled study of 129 renal transplant recipients with early stable renal function (≤28 days after transplantation, GFR ≥ 30 mL/min) was conducted by Smerud and his colleagues. They found that patients with intravenous 3 mg ibandronate every 3 months for 1 year as add-on to calcitriol and calcium had a significant improvement of BMD in total femur and ultradistal radius [bib_ref] A 1-year randomized, double-blind, placebo-controlled study of intravenous ibandronate on bone loss..., Smerud [/bib_ref]. In a randomized study of male cardiac transplant patients, researchers found that intravenous ibandronate not only preserved bone mass, but also reduced vertebral fracture risk and prevented uncoupling of bone formation and resorption [bib_ref] Ibandronate prevents bone loss and reduces vertebral fracture risk in male cardiac..., Fahrleitner-Pammer [/bib_ref] , which was in accordance with another study of liver transplant recipients [bib_ref] Ibandronate and calcitriol reduces fracture risk, reverses bone loss, and normalizes bone..., Wagner [/bib_ref].
Another intravenous bisphosphonate, zoledronic acid, has also been used in a randomized placebo-controlled trial of 20 patients after renal transplantation. It was administrated in a dose of 4 mg twice within 3 months after surgery. At the end of 6 months, it was shown that zoledronic acid improved the calcium content of cancellous bone and increased lumbar spine BMD [bib_ref] Zoledronic acid to prevent bone loss in the first 6 months after..., Haas [/bib_ref].
In a case-control study, short-term weekly use of alendronate (Fosamax), an oral bisphosphonate, could significantly improve both lumbar spine BMD and bone condition, regardless of effect of immunosuppressant after a long period after renal transplantation [bib_ref] Use of alendronate sodium (Fosamax) to ameliorate osteoporosis in renal transplant patients:..., Huang [/bib_ref]. In a long-term of lowdose weekly use of alendronate (35 mg/week) for long-term renal transplant recipients, authors found alendronate could suppress bone turnover but not fracture in kidney transplant recipients with hyperparathyroidism and chronic kidney disease [bib_ref] Oral alendronate can suppress bone turnover but not fracture in kidney transplantation..., Yamamoto [/bib_ref].
## Calcitonin.
Studies of calcitonin therapy in transplant patients with osteoporosis are scarce. Preliminary data from a single-center cross-sectional study has indicated that calcitonin may help to restore bone mass in kidney transplant recipients with osteoporosis [bib_ref] Osteoporosis after kidney transplantation: preliminary report from a single center, Wang [/bib_ref].
## Antibody to the receptor activator of nuclear factor-kb
Ligand. Denosumab is a fully human monoclonal antibody to the receptor activator of nuclear factor-kB (RANK) ligand that blocks binding to RANK, inhibiting bone resorption and increasing bone density. In a large clinical trial involving 7868 postmenopausal women aged 60 and 90, administration of denosumab twice yearly for 3 years was associated with a significant reduction in the risk of vertebral, nonvertebral, and hip fractures in women, as compared with placebo [bib_ref] Denosumab for prevention of fractures in postmenopausal women with osteoporosis, Cummings [/bib_ref]. In a dose-response-based meta-analysis including 142 randomized controlled trials, 3 years with denosumab resulted in bigger changes in BMD of lumber spine and total hip compared with 3 years of treatment with 10 mg/d oral alendronate, 5 mg/y i.v. zoledronic acid, 5 mg/d oral risedronate, 150 mg/mo oral ibandronate, 3 mg i.v. ibandronate every 3 months, 60 mg/d oral raloxifene, and 200 IU/d calcitonin [bib_ref] Time course of bone mineral density changes with denosumab compared with other..., Mandema [/bib_ref]. Given its unique actions, denosumab may be useful for the prevention and treatment of osteoporosis and fractures in organ transplant patients.
## 4.7.
Teriparatide. Teriparatide is a recombinant human parathormone (PTH 1-34) currently used among patients with osteoporosis who are at high risk for fracture [bib_ref] Teriparatide or alendronate in glucocorticoid-induced osteoporosis, Saag [/bib_ref].
Only limited data are available on teriparatide therapy for patients receiving organ transplantation. In a randomized double-blind trial of 26 kidney transplant patients, daily subcutaneous injection teriparatide (20 g) of 6 months was well tolerated but did not improve either BMD or bone turnover early after transplantation [bib_ref] Effect of teriparatide on early bone loss after kidney transplantation, Cejka [/bib_ref]. However, for the kidney transplant patients with severe hypocalcemia and low PTH, treatment with teriparatide showed good results without side effects [bib_ref] Teriparatide efficacy in the treatment of severe hypocalcemia after kidney transplantation in..., Nogueira [/bib_ref].
## Hormone replacement therapy (hrt).
Hormonal replacement therapy has been proposed to reduce the risk of osteoporosis in menopausal women. However, there is restricted literature regarding effect of HRT on transplant recipients. One study was reported by Isoniemi et al. in 33 postmenopausal female liver transplant recipients with 2-year follow-up [bib_ref] Transdermal oestrogen therapy protects postmenopausal liver transplant women from osteoporosis. A 2-year..., Isoniemi [/bib_ref]. The lumber spine BMD increased by 5.3 and 1.2%, respectively, in the first and second year, while the femoral neck BMD increased by 3.3 and 1.2%. Only 20% of the patients had osteoporosis after 2 years of treatment, whereas over 50% of the women had osteoporosis at baseline. Another 10-year-experience study showed, with the treatment of estradiol and progestin, total regression of climacteric symptoms was reported in 75% of kidney transplant patients without negative effect [bib_ref] Safety and efficacy of hormonal therapy in menopausal kidney-allograft recipients, Pietrzak [/bib_ref]. Further studies are needed to establish the efficacy and safety of HRT among transplant recipients.
# Conclusion
In recent years, more and more attention has been drawn to new therapeutic challenges in the long-term management of SOT recipients due to the greatly improved patient and graft survival. And posttransplant osteoporosis is one of them. Although the pathogenesis is not well established, preexisting bone disorders and transplant-specific therapies appear to play a key role. At present, there is no golden standard for the prevention and treatment of osteoporosis and fractures following organ transplantation. However, treatment with vitamin D and/or bisphosphonate in the first year following transplantation was associated with a reduction in the incidence of osteoporosis and fractures. Other medications including calcitonin, teriparatide, and HRT have also been evaluated in some cases. More research is needed to verify the effect of various therapeutic approaches and develop a clinical guideline of therapy strategy for bone disorders after organ transplantation.
[table] Table 1: Effects of immunosuppressive agents on skeleton. [/table]
|
Usefulness of the Geriatric Depression Scale 15-item version among very old people with and without cognitive impairment
Objectives: The aim of this population-based study was to investigate the usefulness of the Geriatric Depression Scale 15-item version (GDS-15) to assess depressive symptoms among very old people with differing levels of cognitive function. Methods: The 834 participants were aged 85 and over. Feasibility of GDS-15 was evaluated as the proportion of people who completed the scale. Concurrent criterion validity was evaluated by calculating correlations between GDS-15 and Philadelphia Geriatric Center Morale Scale (PGCMS). PGCMS measures psychological wellbeing which is closely related with depressive symptoms. Correlations were calculated within groups according to cognitive function assessed with Mini-Mental State Examination (MMSE); 0-4, 5-9, 10-14, 15-19, 20-24, 25-27, and 28-30, using Pearson's two-sided correlation and compared using Fisher r-to-z transformation. Internal consistency of the GDS-15 was evaluated by calculating Cronbach's in each group. Results: In total, 651 (78%) of the 834 participants completed the GDS-15. For the two MMSE-groups with scores of 510, the proportion who completed GDS-15 were 1% and 42%, respectively, compared to 65-95% in the MMSE-groups with scores of !10. Cronbach's in each MMSE-group ranged from 0.636 (MMSE 28-30) to 0.821 (MMSE 5-9). The level of correlation between GDS-15 and PGCMS did not significantly differ between MMSE-groups with scores of 5-27 compared to the MMSE-group with scores of 28-30. Conclusions: The GDS-15 seems to have an overall usefulness to assess depressive symptoms among very old people with an MMSE score of 10 or more. More studies are needed to strengthen the validity of GDS-15 among older people with MMSE scores of 10-14. For older people with MMSE scores lower than 10, there is a need to develop and validate other measurements.
# Introduction
Depression is a common problem among older people causing emotional suffering and increased mortality as well as an increased risk of physical inactivity and disability [bib_ref] Origins of depression in later life, Blazer [/bib_ref]. The highest prevalence of depressive symptoms has been found among the very old, aged 80 years and over, in older people living in residential care facilities, and among people with dementia [bib_ref] Depression among the very old with dementia, Bergdahl [/bib_ref] [bib_ref] Depression among the oldest old: The Umea˚85þ study, Bergdahl [/bib_ref] [bib_ref] Origins of depression in later life, Blazer [/bib_ref]. Many older people with depression do not receive adequate treatment perhaps because depressive symptoms are not recognized [bib_ref] Depression among the oldest old: The Umea˚85þ study, Bergdahl [/bib_ref]. Thus, instruments to assess depressive symptoms are required for both clinical and research purposes not only to enable recognition and assessment of the severity of depression among older people, but also to facilitate evaluation of treatment effects.
The Geriatric Depression Scale (GDS) is an instrument that was developed, regarding both content and design, to assess depressive symptoms and screen for depression among older people [bib_ref] Development and validation of a geriatric depression screening scale: A preliminary report, Yesavage [/bib_ref]. Somatic symptoms such as weight loss, sleep disturbances, and pessimism about the future, are common symptoms of depression among younger people. However, these can be related to aging itself, and are not included in the GDS, which focuses instead on psychiatric symptoms. During development of the GDS, clinicians and researchers in the geriatric field were asked to suggest items that could separate those with and those without depression. From these suggestions, 30 items were chosen for inclusion in the scale. The GDS can be selfadministered or presented as an interview, and the questions have a yes/no format in order to be easy to understand for older people who may suffer from impaired cognitive function. Shorter versions have been suggested to reduce problems in completing the scale arising from fatigue or concentration difficulties. A 15-item version was presented by [bib_ref] Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version, Sheikh [/bib_ref] , based on the items that correlated best with depressive symptoms, and was equally successful as the 30-item version in differentiating between those with and without depression among people aged 55 years and over and living in the community [bib_ref] Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version, Sheikh [/bib_ref].
Cognitive impairments are common among very old people, and whether the psychometric properties of the GDS are influenced by this is an important inquiry. Studies have investigated validity among people with cognitive impairment or dementia and have found decreasing validity along with decreasing level of cognitive function [bib_ref] The geriatric depression scale and the Cornell scale for depression in dementia...., Kørner [/bib_ref] [bib_ref] Detecting depression in Alzheimer's disease: Evaluation of four different scales, Mu¨ller-Thomsen [/bib_ref] [bib_ref] Screening for depression and assessing change in severity of depression. Is the..., Smalbrugge [/bib_ref]. Cognitive impairment is defined in various ways in these studies, and it is difficult to interpret the level to which the GDS is feasible and valid since studies have used a variety of cognitive levels for inclusion and for subgroup analyses. However, the GDS (15-and 30item versions) seems to be valid for people with cognitive impairments down to scores of 15/30 on the Mini-Mental State Examination (MMSE) [bib_ref] Diagnostic accuracy of the original 30-item and shortened versions of the Geriatric..., Jongenelis [/bib_ref] [bib_ref] Validating the GDS depression screen in the nursing home, Mcgivney [/bib_ref] [bib_ref] Screening for depression and assessing change in severity of depression. Is the..., Smalbrugge [/bib_ref]. There is a lack of knowledge regarding the validity of the scale among people with even lower cognitive function specifically since this, to our knowledge, has only been evaluated in two studies which both used the 30-item version [bib_ref] Is the geriatric depression scale a reliable screening tool for depressive symptoms..., Debruyne [/bib_ref] [bib_ref] Measurement of overall quality of life in nursing homes through self-report: The..., Gerritsen [/bib_ref]. In addition, only one study has focused on very old people aged 85 years and over and that study included mainly people with no or mild cognitive impairment [bib_ref] Standardised assessment scales for elderly people. London, UK: The Royal College of..., Dall [/bib_ref]. The present study was performed in a large representative sample of very old people aged 85 years and over, including people with a wide variety of living situations, capacities in activities of daily living, and cognitive function. This enables a comparison of the usefulness of the GDS between people with differing levels of cognitive function.
## Aim
The aim of the present population-based study was to investigate the usefulness of the GDS 15-item version to assess depressive symptoms among very old people with differing levels of cognitive function.
# Methods
## Study design
The study was based on cross-sectional data conceived from the Umea˚85þ/GERDA (Gerontological Regional Database) study [bib_ref] Low blood pressure is associated with cognitive impairment in very old people, Molander [/bib_ref] [bib_ref] Health status in the oldest old. Age and sex differences in the..., Von Heideken Wa˚gert [/bib_ref]. GERDA is a collaborative population-based cohort study performed by Umea˚University, Sweden, and Å bo Akademi/Vaasa University, Finland. Data collection was carried out in the county of Va¨sterbotten in the urban municipality of Umea˚and in five rural municipalities during . Data collection was carried out in in two municipalities in Pohjanmaa, Finland. The study included every second person aged 85 years, and the total population of people aged 90 and 95 and over, registered in the National Tax Board in Sweden and the Finnish Population Register Centre in Finland. The study was approved by the Regional Ethical Review Board in Umea˚(x99-326 and x05-063M) and the Ethics Committee of Vaasa Central Hospital (registration number 05-87).
## Participants
The present study included 834 participants who were assessed for cognitive function. A flowchart of the inclusion is displayed in [fig_ref] Figure 1: Flow chart over the inclusion of participants [/fig_ref] , and baseline characteristics in [fig_ref] Table 1: Characteristics of the participants [/fig_ref]. A number of individuals (n ¼ 101) took part in both Swedish data collections. For those individuals, only the data from was used in the present study and the rational was that they would contribute with cross-sectional data from when they were older and thus were expected to have lower cognitive function which was of interest in the present study.
## Procedure
A letter with information about the study was sent to all participants and followed up with a phone call about one week later where informed consent to participation was obtained. For those living in institutions, staff was contacted and asked about the cognitive state of the individual. The next of kin was contacted for informed consent when appropriate due to cognitive impairment. A structured interview covering a variety of areas regarding health and sociodemographic data, including assessments, was carried out in the same order during one to three home visits [bib_ref] Health status in the oldest old. Age and sex differences in the..., Von Heideken Wa˚gert [/bib_ref]. The assessors were trained physicians, nurses, physical therapists or medical students who were unaware of the aim of the present study. Data were also collected from relatives, caregivers and medical charts when approval was given.
## Assessments
The 15-item Swedish version of the GDS (GDS-15) was used to assess depressive symptoms and was interviewadministered for all participants [bib_ref] Comparison of six depression rating scales in geriatric stroke patients, Agrell [/bib_ref] [bib_ref] Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version, Sheikh [/bib_ref]. The score ranges from 0 to 15 and a score of zero to four is considered to be within the normal range, five to nine indicates mild depression, and a score of 10 or more indicates moderate to severe depression [bib_ref] Short versions of the geriatric depression scale: A study of their validity..., Almeida [/bib_ref].
Cognitive state was assessed using the MMSE, which is a test of cognitive aspects of mental function, e.g. orientation, memory, ability to follow verbal and written commands. The MMSE has a score ranging between 0-30, where a score of 17 indicates severe cognitive impairment and 18-23 indicates mild cognitive impairment [bib_ref] The mini-mental state examination: A comprehensive review, Tombaugh [/bib_ref].
The Philadelphia Geriatric Center Morale Scale (PGCMS) was used to assess morale and was also interview-administered for all participants [bib_ref] The Philadelphia geriatric center morale scale: A revision, Lawton [/bib_ref]. High morale is described as a basic sense of satisfaction with oneself, a feeling that there is a fit between personal needs and what the environment offers, and a certain acceptance of what cannot be changed [bib_ref] The Philadelphia geriatric center morale scale: A revision, Lawton [/bib_ref]. It is suitable for use with older people living either in the community or in institutions, and the questions' yes/no format facilitates understanding for people with impaired cognitive function [bib_ref] The Philadelphia geriatric center morale scale: A revision, Lawton [/bib_ref] [bib_ref] Assess not assumemeasuring the morale of cognitively impaired elderly, Ryden [/bib_ref]. The 17-item version was used in the present study [bib_ref] The Philadelphia geriatric center morale scale: A revision, Lawton [/bib_ref] , where scores of 0-9 indicate low morale, 10-12 the middle range and 13-17 high morale, according to the administration and scoring instructions.
The Barthel Index (0-20) was used to assess activities of daily living, where higher scores indicate a greater degree of independence [bib_ref] The Barthel ADL Index: A reliability study, Collin [/bib_ref] [bib_ref] Functional evaluation: The Barthel index, Mahoney [/bib_ref]. Assessment using the Organic Brain Syndrome Scale contributed to the assessment of depression, dementia and delirium and to differentiate between those diagnoses [bib_ref] A comparsion between three psychogeriatric rating scales, Jensen [/bib_ref]. Information was collected regarding diagnoses and prescribed drugs from the participants, staff and/or medical records. Diagnosis of depression and dementia were based on earlier diagnosis according to medical charts, ongoing pharmacological treatment, and on assessments during the interviews. All information was reviewed by an experienced physician, and diagnoses of depression and dementia were set according to the DSM-IV criteria (American Psychiatric Association, 1994).
## Feasibility, internal consistency, and validity
The feasibility of GDS-15 was evaluated as the proportion of people who completed the scale. Internal consistency of the GDS-15 was evaluated using Cronbach's , a measure of the inter-item correlation of a scale. Concurrent criterion validity refers to investigating whether scores on an instrument agree with a measurement of the same theme when assessed at the same point of time. The concurrent criterion validity of the GDS-15 was evaluated against the PGCMS which measures morale. Morale is often used synonymously with psychological or subjective wellbeing, and the PGCMS has been recommended for use in assessing subjective wellbeing among older people by the British Geriatrics Society and the Royal College of Physicians, London [bib_ref] Standardised assessment scales for elderly people. London, UK: The Royal College of..., Dall [/bib_ref]. The rationale for using PGCMS when evaluating concurrent criterion validity was that depressive symptoms are closely related with poor psychological well-being among older people [bib_ref] Does the use of the Geriatric Depression Scale make redundant the need..., Coleman [/bib_ref] [bib_ref] Measurement of overall quality of life in nursing homes through self-report: The..., Gerritsen [/bib_ref] [bib_ref] Morale in the oldest old: The Umea˚85þ study, Von Heideken Wa˚gert [/bib_ref] [bib_ref] Depression is the predominant factor contributing to morale as measured by the..., Woo [/bib_ref]. Further, the scales are constructed in a similar way, both using the format of closed-ended questions where a yes or no answer indicates presence of symptoms.
## Statistical analyses
Differences between those assessed with the GDS-15 (n ¼ 693) and those who declined or were not able to answer the questions (n ¼ 141) were evaluated regarding age and MMSE scores using independent samples t-test, and regarding sex using chi-square test. The criterion for a complete assessment was no more than one missing answer for GDS-15 and PGCMS, respectively, i.e. answering 14 questions or more on the GDS-15, and 16 or more in the PGCMS. Missing answers were imputed with the score zero. A logistic regression was made to evaluate the impact of MMSE scores for completion (feasibility) of the GDS-15. Completion (yes/no) was the dependent variable and MMSE score was the independent variable. The sample was divided into seven groups according to MMSE, to compare the validity and the internal consistency of the GDS-15 among individuals with different levels of cognitive function. Each group represented five points on MMSE, except for the two groups with the highest scores which represented three points . The rationale for choosing a smaller interval in the groups with the highest scores on MMSE was that many participants had high scores compared to low scores. Correlations were calculated between the GDS-15 and the PGCMS (concurrent criterion validity), within each MMSE-group using two-sided Pearson's correlation presented with correlation coefficients including 95% confidence intervals (CI), and p-values. Fisher r-to-z transformation, a two-tailed test for independent samples, was used to compare the correlation values for each MMSE-group with the value for the group with the highest cognitive function (MMSE 28-30). Cronbach's (internal consistency) for the GDS-15 was calculated in each MMSE-group. All these analyses were also calculated using only individuals who completed all 15 questions in the GDS-15 and 17 in PGCMS, showing essentially the same results (data not shown).
Additional analyses were performed to evaluate the impact of age on the correlation analyses, by analyzing the age groups separately (85, 90, and 95 and over, respectively), and likewise to evaluate impact of sex by analyzing women/men separately. In the additional analyses, individuals were divided into two groups according to cognitive function; MMSE 10-24 and 25-30, to avoid risk of low power in the analyses because there were so few individuals in some MMSEgroups. Fisher r-to-z transformation test was used to compare correlation levels for sex and age, respectively, within each of the two MMSE-groups. For age, correlation values were compared to that of the youngest age group, 85-year-olds.
Analyses were performed using the SPSS software, version 17.0 (SPSS Inc., Chicago, IL), and a p-value of less than 0.05 was considered to indicate statistical significance.
# Results
The characteristics of the 834 participants are presented in groups according to cognitive function [fig_ref] Table 1: Characteristics of the participants [/fig_ref]. The mean score on the MMSE for the whole sample was 19.8 (SD ¼ 8.0, range 0-30). Mean age was 90.2 years, 584 participants (70%) were women, and 344 (41%) lived in institutional care. Six hundred and six participants (73%) lived in Sweden, 578 (69%) lived in urban areas, and 682 (82%) lived alone. Three hundred and twenty-eight participants (39%) had a dementia disorder, 293 (35%) a diagnosis of depression, 180 (22%) had previous stroke, and mean number of drugs for regular use was 6.9 (range 0-29).
## Feasibility
In total, 651 (78%) of the 834 participants completed the GDS-15 [fig_ref] Figure 1: Flow chart over the inclusion of participants [/fig_ref] , [fig_ref] Table 1: Characteristics of the participants [/fig_ref]. The mean score for GDS-15 was 3.7 (SD ¼ 2.6, range 0-14). Of the remaining 183 participants; 141 declined or were not able to answer the questions, and 42 answered less than 14 questions of the GDS-15. Those who declined or were unable to answer the GDS-15 (n ¼ 141) were older (mean age 91.5 vs. 89.9, p ¼ 0.001), more likely to be women (83% vs. 67%, p 5 0.001), and had lower cognitive function (MMSE mean score 8.1 vs. 22.2, p 5 0.001), than those who answered GDS-15 (n ¼ 693).
With increasing cognitive impairment fewer people completed the GDS-15 interview [fig_ref] Figure 2: Proportion of people for whom the GDS-15 version was completed, in relation... [/fig_ref] , odds ratio ¼ 1.22 (CI ¼ 1.18-1.26, p 5 0.001). For the two MMSE-groups with scores of 5 10, the proportion who completed GDS-15 were 1% and 42%, respectively, compared to 65-95% in the MMSE-groups with scores of !10 [fig_ref] Table 1: Characteristics of the participants [/fig_ref]. Of the 651 who completed the GDS-15, 573 (88%) participants answered all 15 questions and 78 (12%) answered 14 [fig_ref] Figure 1: Flow chart over the inclusion of participants [/fig_ref]. Among those participants who completed 14 questions, 25 (32%) did not answer question 10 'more problems with memory than most', 13 (17%) question 15 'most people better off', seven (9%) question 11 'wonderful to be alive', and seven (9%) did not answer question 9 'prefers to stay in'.
## Internal consistency
The Cronbach's for the GDS-15 among groups of people with MMSE scores of five or more ranged from 0.636 (MMSE 28-30) to 0.821 (MMSE 5-9) [fig_ref] Table 1: Characteristics of the participants [/fig_ref].
## Concurrent criterion validity
The correlations between the GDS-15 and the PGCMS were statistically significant among groups of people with MMSE scores of five or more and the coefficients ranged from À0.585 (MMSE 28-30) to À0.726 (MMSE 10-14) (Table 2, [fig_ref] Figure 3: Correlation between the GDS-15 and the PGCMS among people with differing levels... [/fig_ref]. The correlations between the GDS-15 and PGCMS did not differ between the groups of people with MMSE scores of 5-9, compared to the group of people with MMSE scores of 28-30 (data not shown).
## Additional analyses
There were no significant differences in correlation levels between the age groups, or between males and females (data not shown).
# Discussion
The results from this population-based study indicate the overall usefulness of the GDS-15 to assess depressive symptoms among very old people with MMSE scores of 10 or more. Almost two thirds of the participants with MMSE scores of 10-14 were able to complete the assessment of GDS-15, compared to the two groups with MMSE scores of less than 10 where the proportion that completed GDS-15 were only 1% and 42%, respectively. For participants in groups with MMSE scores of five or more, the internal consistency of the GDS-15 seems comparable and the correlations between the GDS-15 and PGCMS did not differ from participants with the highest cognitive function (MMSE 28-30). The correlations did not differ between women and men, or between the age groups (85, 90, and 95 and over), respectively.
The results from the present study, indicating the usefulness of the GDS-15 among people with a level of Notes: MMSE (scores 0-30 points, higher score ¼ better cognitive function), GDS-15 (scores 0-15 points, higher score ¼ more symptoms of depression), PGCMS (scores 0-17 points, higher score ¼ higher morale). Correlation between variables calculated with two-sided Pearson's correlation, presented with correlation coefficient (r) and p-value. A p-value of less than 0.05 was regarded as statistically significant. *n includes assessments where GDS-15 and PGCMS was completed, i.e. at least 14/15 and 16/17 questions, respectively, were answered. cognitive impairment down to MMSE scores of 10, seem to be in accordance with those Gerritsen et al presented when comparing GDS-30 with a self-rating scale for depression (Depression List) and one for psychological wellbeing (PGCMS) among nursing home residents [bib_ref] Measurement of overall quality of life in nursing homes through self-report: The..., Gerritsen [/bib_ref]. Gerritsen et al found acceptable internal consistency and relatively strong correlations between GDS-30 and the self-rating scales among people with MMSE scores of 5 or more. The decline in completion rate with decreasing cognitive function was also found in the study by showed that 72% of people with MMSE scores of 5-12 completed the GDS, compared to 100% of people with MMSE scores 22-30 [bib_ref] Measurement of overall quality of life in nursing homes through self-report: The..., Gerritsen [/bib_ref]. The higher completion rate in that study, compared to the present population-based study, might be due to a possible selection of the sample in the low cognition groups because the GDS-30 was not offered to all participants both for practical considerations and the frailty of the residents. The results from the present study showed overall stronger correlations than the results presented in a study by Debruyne et al, who compared GDS-30 scores with results from another scale assessing depressive symptoms (Cornell Scale for Depression in Dementia) among old people with varying levels of cognitive functions including those with severe cognitive impairment [bib_ref] Is the geriatric depression scale a reliable screening tool for depressive symptoms..., Debruyne [/bib_ref]. One explanation for the low correlations in that study may be the administration of the Cornell scale, where the participants were interviewed together with a caregiver. Among people with MMSE-scores of five or more, self-reports of depressive symptoms and well-being seems more valid than ratings or observations by staff or family carers [bib_ref] Measurement of overall quality of life in nursing homes through self-report: The..., Gerritsen [/bib_ref].
It seemed that two items in particular were more difficult to answer in the present study. These were items 10 and 15, which are both questions which include a comparison to other people. Sutcliffe et al suggested that GDS-15 could be shortened for people living in residential care facilities because they found increased internal consistency when three items (item 9, 10, and 15) were removed [bib_ref] A new version of the geriatric depression scale for nursing and residential..., Sutcliffe [/bib_ref]. This was based on difficulties experienced with these specific questions, perhaps because of reluctance among old people to make assumptions about other people's life situations (item 10 and 15), or because some people simply never go outside (item 9) [bib_ref] A new version of the geriatric depression scale for nursing and residential..., Sutcliffe [/bib_ref].
# Methodological considerations
The large sample size in the present population-based study made it possible to divide the sample into groups according to level of cognitive function and to compare groups with severe cognitive impairments with a group with high cognitive function. The sample comprised very old people and with a wide variety of health and living conditions, reflecting the heterogeneity existing in this group. Unfortunately, the reliability of the GDS-15 was not evaluated but all participants were assessed by a trained investigator following the same procedure. No consensus was found in the literature on how to handle missing answers on the GDS-15; some allow and some do not allow missing answers [bib_ref] Validating the GDS depression screen in the nursing home, Mcgivney [/bib_ref] [bib_ref] A new version of the geriatric depression scale for nursing and residential..., Sutcliffe [/bib_ref]. One missing answer was allowed in the present study with regard to the population studied, and it was considered unlikely to have a significant impact on the results, as was confirmed in the analyses.
The PGCMS seems to be an appropriate choice when evaluating concurrent criterion validity of the GDS-15 since depressive symptoms are closely related with poor psychological well-being among older people [bib_ref] Does the use of the Geriatric Depression Scale make redundant the need..., Coleman [/bib_ref] [bib_ref] Measurement of overall quality of life in nursing homes through self-report: The..., Gerritsen [/bib_ref] [bib_ref] Morale in the oldest old: The Umea˚85þ study, Von Heideken Wa˚gert [/bib_ref] [bib_ref] Depression is the predominant factor contributing to morale as measured by the..., Woo [/bib_ref]. The correlations between GDS-15 and PGCMS and the internal consistency of the GDS-15 in our study indicate that answers are not given randomly, which could be expected from people with cognitive impairments or dementia. In order to further investigate the usefulness of the GDS-15, future studies should evaluate the validity of GDS-15 against a depression diagnosis or another scale of depressive symptoms among people with MMSE scores of 10-14. Unfortunately, this was not possible in the present study since the depression diagnosis was partly based on the score from the GDS-15.
The instructions for the GDS-15 calls for the depressive symptoms to be rated according to how they have been during the preceding week, which is a task that requires recalling the preceding week. This may be difficult for individuals with severe cognitive impairment both to understand and answer and it is likely that, among those with cognitive impairment in the present study, the answers are probably mainly based on their feelings at the time of the interview rather than the preceding week. However, there is support in the literature that a large proportion of people with cognitive levels down to MMSE scores of 10 can answer questions about their quality of life in a valid way [bib_ref] Use of the QOL-AD for measuring quality of life in people with..., Hoe [/bib_ref] [bib_ref] Not knowing where I am doesn't mean I don't know what I..., Mozley [/bib_ref] , which strengthens the proposed conclusions that the answers to GDS-15 were not given randomly in this group of old people. Further, the GDS-15 seems suitable because of the yes/ no format of the questions compared to questionnaires that use Likert scale alternatives. In addition, administrating the GDS-15 as an interview facilitates the completion of the questions for people with impairments concerning, e.g. vision.
# Conclusions
In conclusion, since almost two thirds of the participants with MMSE scores of 10-14 were able to complete the scale, and that the internal consistency of the GDS-15 and the level of agreement with PGCMS did not differ from that for people with higher cognitive function, the GDS-15 seems to have an overall usefulness for assessing depressive symptoms in clinical and research purposes among very old people with an MMSE score of 10 or more. More studies are needed to strengthen the validity of GDS-15 among older people with MMSE scores of 10-14, by evaluating GDS-15 against a depression diagnosis or another scale of depressive symptoms. For older people with MMSE scores lower than 10, and those over 10 not able to complete the GDS-15, there is a need to develop and validate other measurements to assess depressive symptoms.
# Declaration of interest
[fig] Figure 1: Flow chart over the inclusion of participants. [/fig]
[fig] Figure 3: Correlation between the GDS-15 and the PGCMS among people with differing levels of cognitive function, measured using the MMSE. Correlations that were significant(Table 2) are presented. Bars represent 95% CIs for r. [/fig]
[fig] Figure 2: Proportion of people for whom the GDS-15 version was completed, in relation to level of cognitive function, measured with the MMSE. [/fig]
[table] Table 1: Characteristics of the participants (n ¼ 834) and internal consistency (Cronbach's ) of the GDS-15. [/table]
[table] Table 2: Correlation between the GDS-15 and the PGCMS among people with differing levels of cognitive function. [/table]
|
Surveillance of medically‐attended influenza in elderly patients from Romania—data from three consecutive influenza seasons (2015/16, 2016/17, and 2017/18)
Background:Influenza is an acute infection affecting all age groups; however, elderly patients are at an increased risk. We aim to describe the clinical characteristics and the circulation of influenza virus types in elderly patients admitted for severe acute respiratory infection (SARI) to a tertiary care hospital in Bucharest, Romania, part of the I-MOVE+ hospital network.Methods:We conducted an active surveillance study at the National Institute for Infectious Diseases "Prof. Dr Matei Balș," Bucharest, Romania, during three consecutive influenza seasons: 2015/16, 2016/17, and 2017/18. All patients aged 65 and older admitted to our hospital for SARI were tested for influenza by PCR. Results: A total of 349 eligible patients were tested during the study period, and 149 (42.7%) were confirmed with influenza. Most patients, 321 (92.5%) presented at least one underlying condition at the time of hospital admission, the most frequent being cardiovascular disease, 270 (78.3%). The main influenza viral subtype circulating in 2015/16 was A(H1N1)pdm09, followed by A(H3N2) in 2016/17 and B influenza in 2017/18. Case fatality was highest in the 2015/16 season (3.7%), 0% in 2016/17, and 1.0% in 2017/18. Vaccination coverage in elderly patients with SARI from our study population was 22 (6.3%) over the three seasons.Conclusions:Our study has highlighted a high burden of comorbidities in elderly patients presenting with SARI during winter season in Romania. The influenza vaccine coverage rate needs to be substantially increased in the elderly population, through targeted interventions.K E Y W O R D Scase fatality, comorbidities, elderly, influenza, subtype | 531 PIȚIGOI et al.
## | backg rou n d
Influenza is an acute infection affecting all age groups; however, elderly patients are at an increased risk, particularly due to a clustering of comorbidities which puts them at risk of severe influenza, complicated influenza, or influenza-related decompensation of underlying conditions.On a global level, in 2017, influenza has been estimated to be responsible for 54 481 000 episodes of medically-diagnosed lower respiratory tract infections leading to 9 459 000 hospitalizations, and specifically in Romania, with a population in 2017 of 19.6 million people, 174 000 episodes of lower respiratory tract infections with 65 000 hospitalizations, as calculated by the Global Burden of Disease Study 2017.Continued surveillance of influenza is warranted in each country, in order to better inform public health policies and to fill the existing information gaps, particularly for specific influenza risk groups.
In Romania, influenza surveillance is performed at national level by two surveillance systems, one for severe acute respiratory infections (SARI: testing one patient hospitalized with SARI per week from 20 hospitals throughout six counties, during the influenza surveillance season), and one for influenza-like illness (ILI: testing all patients with ILI attending 192 sentinel general practitioners from 16 counties every Tuesday).The National Institute for Infectious Diseases "Prof. Dr Matei Balș," a tertiary care hospital in Bucharest, Romania, was part of the I-MOVE+ hospital network (http://www.i-movep lus.eu/) from 2015 to 2018 and implemented a protocol for systematic screening for influenza in elderly patients admitted to the hospital for SARI.Here, we aim to describe the clinical features of the elderly patients admitted with SARI in our hospital and the circulation of influenza virus types in these patients during three consecutive influenza seasons in Bucharest, Romania, in order to characterize the epidemiology of influenza in this particular patient population, at high risk for influenza-related morbidity.
## | me thods
An active epidemiologic surveillance study was implemented at the National Institute for Infectious Diseases "Prof. Dr Matei Balș,"
Bucharest, Romania, during three consecutive influenza seasons: 2015/16, 2016/17, and 2017/18. The study consisted of systematic daily screening of all consecutive admissions in patients aged 65 and older admitted to our hospital with an acute (onset <7 days) illness meeting the following SARI case definition: one or more general signs or symptoms, defined as: fever/feverishness, malaise, headache, myalgia, and altered clinical state (asthenia, anorexia, confusion, or weight loss) associated with one or more respiratory signs or symptoms defined as: cough, odynophagia, and dyspnea.Patients were excluded from the study if they met any of the following exclusion criteria: had a contraindication for influenza vaccine, had SARI onset ≥48 hours after hospital admission, were unwilling to participate or unable to communicate and give consent (either by the patient or by her/his legal representative), were institutionalized, had a respiratory specimen taken ≥8 days after SARI onset, and had tested positive for any influenza virus in the current season before the onset of symptoms leading to the current hospitalization. No information was collected on the number of exclusions overall and by each criterion, and therefore, this information is not available and will not be reported in Results.
For each consenting patient, a standardized medical questionnaire was filled out by study investigators through review of medical (hospital or general practitioner) records and patient/relative interview, collecting demographic variables, a complete medical history, influenza vaccination status in the respective season, and data related to the current SARI episode's onset, characteristics and outcome, as previously described.MGB probes were designed for both B virus lineages that can be detected and discriminated simultaneously, as only one of the two probes will give a fluorescent signal. 8 All real-time RT-PCR reactions were performed using commercial kits-SuperScript ® One-Step qRT-PCR System (Invitrogen). We report descriptive data as number and percentage for categorical variables, and as median and interquartile range (IQR) or range for non-parametric continuous variables. Statistical associations were tested using the chi-squared test for categorical variables and Mann-Whitney's U test for continuous non-parametric variables. Two-tailed P values <.05 were interpreted as statistically significant. IBM SPSS Statistics for Windows, version 20 (IBM Corp.) was used for the statistical analysis.
## | re sults
A total of 349 eligible patients were tested in this study, ranging between 53 and 191 by season included in the study. The baseline characteristics overall and by each of the three influenza seasons are presented in . The median (IQR) age was 74 (68, 80) years, and 43.6% of patients were men.
Most patients (92.5%) had at least one underlying condition at the time of hospital admission, the most frequent being cardiovascular disease, present in 78.3% of cases, followed by diabetes (32.5%), obesity (29.2%), and chronic lung disease (22.6%). Other types of comorbidities were less frequent, being reported in less than 15% of patients During the study period, 20 (13.4%) strains of influenza virus were tested for antiviral resistance; all strains showed susceptibility to oseltamivir.
Among the SARI criteria, fever (94.6%), malaise (90.6%), and myalgia (71.7%) were the most frequently encountered signs or symptoms in patients testing positive for influenza, and specifically fever (OR = 3.0) and odynophagia (OR = 1.8) were significantly predictive for testing positive for influenza . The distribution of comorbidities was comparable between patients with and without influenza . We observed no significant differences in clinical signs or symptoms, or in the distribution of comorbidities between patients with influenza A and influenza B .
Overall, 65.6% of the patients included in the study received antiviral treatment with oseltamivir during hospital admission, and the proportion was higher in patients with laboratory-confirmed influenza (81.2%). The clinical course of disease was generally favorable. However, eight patients died during hospital admission, all having at least one underlying condition; of them, four had been confirmed with influenza, and four were negative for influenza. All four deaths that occurred in patients with laboratory-confirmed influenza were caused by influenza A [three A(H1N1)pdm09 and one A(H3N2)]; none of these four patients had been vaccinated against influenza, three were females, all were aged 76 years old, and had a range of 1-3 comorbidities, as follows: three had diabetes, two had chronic lung disease, one had cardiovascular disease and one had cancer. The overall calculated case fatality in this study was 2.3%: 2.7% in patients with laboratory-confirmed influenza and 2.0% in the patients testing negative . Case fatality was higher in the 2015/16 influenza season than in the subsequent seasons (3.7%, 0.0%, 1.0%) Death during hospitalization g 4 (4.3%, 95% CI: 1.2%-10.6%) 0 (0.0%, 95% CI: 0.0%-6.4%) 4 (2.7%, 95% CI: 0.7%-6.7%) P = .297
[formula] 2 (0-6) 2 (0-4) 2 (0-5) 2 (0-6) Frailty index, median (IQR) a 0.2 (0.1, 0.3) 0.2 (0.1, 0.3) 0.2 (0.1, 0.3) 0.2 (0.1, 0.3) Cardiovascular disease b 149 [/formula]
Note: All data is presented as number (percentage), unless otherwise specified. Statistical analysis was performed with the two-tailed chi-squared test or Mann-Whitney's U test. Frailty index was calculated by dividing the number of active comorbidities by the total number of comorbidities assessed for each patient.
Abbreviations: 95% CI, 95% confidence interval; N/A, not applicable; OR, odds ratio; SARI, severe acute respiratory infection. a Missing data for 1 patient. b Missing data for 2 patients. sensitivity and specificity for diagnosing influenza.Falsey et al also suggested that in elderly patients, the threshold for defining fever could be as low as 37.3°C.Overall, influenza A(H1N1)pdm09 was the main viral subtype circulating in 2015/16, followed by A(H3N2) in 2016/17 and B (supposedly mainly B/Yamagata) in 2017/18 in our study. These data are in line with that reported on a national level by the Romanian National Center for Surveillance and Control of Transmissible Diseases, with certain particularities, discussed below.
For the 2015/16 season, influenza A was the main circulating type in Romania in all age groups, accounting for 90.6% of all influenza cases reported and analyzed on a national level, 17 compared to 97.1% in our study of elderly patients, and the main subtype was A(H1N1)pdm09, accounting for 93.5% of all subtyped A strains nationally, 17 compared with 89.4% in our study. The A(H3N2) subtype was relatively less frequently identified at the national level (6.5% 17 compared with 10.6% in our study), while the circulation of B strains, reported at 9.4% in the country, 17 was almost negligible in our study (2.9%) for the 2015/16 season. These slight differences might arise from a different sampling frame (one tertiary care hospital analyzing hospitalized SARI in our study, compared to an ILI-based national surveillance system 17 in the country as a whole), but they might also reflect to a certain degree a particularity of the patients admitted in our hospital that is a referral hospital for the south of Romania.For the 2016/17 season, the data from our current study confirmed to some extent our previous report regarding the exclusive co-circulation of A(H3N2) and B/Victoria.However, in our study of elderly patients the season was dominated by the circulation of A viruses (90.0% compared with the national estimate of 65.7% in all age groups 20 and compared to our previous report of 33.9% in all age groups with an emphasis on children, suggesting that A viruses predominated in elderly patients in 2016/17; A(H3N2) accounted for 100% of all circulating A viruses in 2016/17 in the current study and in our previous report.Data from Bulgaria and Poland also report the predominance of influenza A virus (97.5% 21 and 95.5% 22 of all influenza cases, respectively), and specifically the A(H3N2) subtype in the 2016/17 season, but A(H1N1)pdm09 was also present in these two countries, to a lower extent.Our data are also in agreement with a recent meta-regression analysis reporting that type B influenza viruses are somewhat less frequent in elderly patients compared to the other age groups.For the 2017/18 season, the predominance of B viruses was evident in our SARI study, accounting for 86.9% of all influenza cases, but the rate was higher than the one reported for ILI-based surveillance at the national level (66.5%).and that the overall severity of influenza was high that season.Furthermore, based on the same national report, the general fatality rate for SARI in 2015/16 was reported at 20.7% and at 49.0%
for SARI confirmed as influenza,and another study has reported a case fatality proportion of 39.8% (95% CI: 29.5%-50.8%) for that same season for SARI confirmed as influenza.Differences in the lower case fatality reported in our study compared to the national F I G U R E 2 Influenza vaccine coverage in Romania in the general population and in elderly patients, 2007/08 to 2018/19 estimate could potentially be due to selection testing bias of severe cases in the conventional surveillance system, leading to an artificial increase in the reported fatality ratio. These differences may also be partly explained by the specific profile of our institute, which is a major reference center for infectious diseases in Romania, and is well equipped to manage influenza cases in a timely manner, and to institute specific antiviral treatment promptly after hospital admission of SARI cases. In our study, 65.6% of patients admitted with SARI received antiviral treatment. In SARI cases confirmed as influenza, the percentage was higher, 81.2%, whereas in patients who tested negative for influenza, 54.0% received an antiviral, but treatment was stopped after influenza was ruled out and an alternative diagnosis was established.
In Romania, influenza vaccination is provided each year by health authorities, through the general practitioners, to priority risk groups as defined by the World Health Organization (WHO). However, vaccine coverage remains generally low for most risk groups, and for elderly patients in particular, with a marked decrease in the immediate post-2009-pandemic period (49.4% to 19.1% for elderly patients in 2008/09 vs 2010/11), as also reported for other countries,and a slowly increasing trend in the past 3 years . In our study population, the influenza vaccination uptake was low over the three seasons. In the 2015/16 season, the vaccination uptake was 7.9% (95% CI: 4.5%-12.7%) in our study population, and the national reported rate for elderly patients was 10.3%.For the 2016/17 season, the influenza vaccine coverage in our study was 3.8% (95% CI: 0.5%-13.0%), and the national reported coverage was 8.2% in elderly patients. 20 For the 2017/18 season, the vaccine coverage in our study (4.8%, 95% CI: 1.6%-10.8%) was lower that the coverage in elderly patients on a national level, reported at 16.3%.Given the overall low number of vaccinated patients included in our study, we cannot conclude whether there was a protective effect of vaccination against hospital admission for SARI. Stronger data from Australia have shown that an influenza vaccine coverage of 80.2% in elderly patients was able to prevent 49.5% of hospital admissions due to influenza in the 2015 influenza season.We recorded a total number of seven cases of influenza inPublished studies have shown that standard-dose influenza vaccines may induce lower hemagglutination-inhibition HAI titers and seroprotection rates in elderly patients.In Romania, high-dose influenza vaccines have so far not been available and therefore, the data we are reporting here refer to elderly patients who had received standard-dose inactivated trivalent vaccines.
Our study's main strength is that it employed systematic screening of all elderly patients admitted for SARI to one tertiary care hospital in Romania, applying the same standardized methodology throughout three consecutive influenza seasons, thus allowing an analysis of the differences from season to season in terms of clinical characteristics, viral type/subtype circulation, case fatality, and use of healthcare resources.
This study also had a number of limitations. The series is relatively small, 349 patients from one single tertiary care center over the course of three consecutive influenza seasons. The overall low vaccination uptake in our study population and in the country pre- Continued surveillance of influenza is needed in order to inform the best local practices. In elderly patients from our setting, given the high burden of comorbidities that we have characterized, interdisciplinary management and good control of underlying diseases are important, and should be coupled with targeted interventions to substantially increase vaccine uptake.
## | con clus ions
|
Renal colic caused by mycotic iliac artery aneurysm
A 33-year-old female presented with acute colicky left loin-to-groin pain and microscopic haematuria, with a background of 6 months of muscle and joint pains and diplopia. A CT kidneys/ureters/bladder demonstrated fat stranding surrounding the left ureter, as it passed over the left common iliac vessels. Arterial and delayed phase imaging revealed an obstructed ureter secondary to a left common iliac artery aneurysm, later found to be mycotic. No previous descriptions of a mycotic aneurysm presenting as renal colic have been found in the literature. The diagnosis and management of infective endocarditis and mycotic aneurysm are discussed, with a review of the literature. This serves as a good example of a common presenting complaint occurring secondary to a rare and serious pathology. Case report: Renal colic caused by mycotic iliac artery aneurysm BJR|case reports 3 of 3 birpublications.org/bjrcr BJR Case Rep;2:20150155
## Summary
A 33-year-old female presented with acute colicky left lointo-groin pain and microscopic haematuria, with a background of 6 months of muscle and joint pains and diplopia. A CT kidneys/ureters/bladder (KUB) demonstrated fat stranding surrounding the left ureter, as it passed over the left common iliac vessels. Arterial and delayed phase imaging revealed an obstructed ureter secondary to a left common iliac artery aneurysm, later found to be mycotic. No previous descriptions of a mycotic aneurysm causing this common presenting complaint have been found in the literature. The management of infective endocarditis (IE) and mycotic aneurysm is discussed, with a review of the literature. This serves as a good example of a common presenting complaint occurring secondary to a rare and serious pathology.
## Clinical presentation
A 33-year-old female presented to the emergency department of a district general hospital with acute onset of left loin-to-groin pain and microscopic haematuria. 1 year prior to this presentation, the patient had undergone bilateral breast augmentation with subpectoral silicone breast implants. A blood culture was performed prior to discharge for isolated pyrexia, which was positive for Staphylococcus aureus.
Over the 6 months prior to her acute presentation, she had been seen in the rheumatology and neurology outpatient departments for muscle and joint aches and an isolated episode of visual field defect/diplopia. On examination, there were no signs of peritonism. Her blood tests revealed significantly elevated C-reactive protein (CRP) levels and neutrophilia. The pain resolved with diclofenac and opiates.
## Differential diagnosis
The provisional diagnosis of renal colic was made and a CT KUB was arranged by the emergency department to examine for radiopaque renal tract calculus. Given the elevated CRP level, pyelonephritis was also considered. With a background of neurological and rheumatological symptoms, a vasculitis such as systemic lupus erythematosus could be considered, although there was no malar rash or photosensitivity.
## Imaging
Low-dose, unenhanced CT KUB revealed [fig_ref] Figure 1: CT kidneys/ureters/bladder performed on admission to the emergency department with left renal... [/fig_ref] fat stranding and inflammation around the left ureter, as it crossed the left common iliac vessels, which were indistinct. No radiopaque calculus or other intraluminal obstruction could be seen within the ureter. The proximal ureter was distended, indicating an obstruction, but no hydronephrosis was evident. A delayed phase CT scan demonstrated an obstructed ureter at the level of the iliac vessels and the subsequent arterial phase [fig_ref] Figure 2: Arterial phase contrast CT [/fig_ref] demonstrated a 3-cm common iliac artery aneurysm causing ureteric stricturing, as well as multiple aneurysms in the right profunda femoris. An intracranial CT angiogram demonstrated a right middle cerebral artery aneurysm [fig_ref] Figure 3: Three-dimensional arterial phase CT scan showing stent graft of the left common... [/fig_ref]. An MRI of the brain revealed a tiny foci of ischaemia affecting the left parietal and occipital lobes bilaterally, suggestive of septic emboli. Transoesophageal echocardiography (TOE) performed to examine for endocarditis revealed aortic valve vegitations.
## Treatment
Endocarditis organism eradication was initiated with long-term intravenous antibiotics. The patient underwent metallic aortic valve replacement for regurgitation and was therefore placed on wafarin, despite having multiple mycotic aneurysms elsewhere.
Owing to the risk of fistulation between the ureter and the artery, the left ureter was stented (retrograde double-J stent; Well Lead, Guangzhou, China) and a covered stent was placed across the left common iliac artery aneurysm, with an amplatzer vascular plug (St. Jude Medical, Plymouth, MN) deployed in the internal iliac artery owing to the aneurysm's proximity to the iliac bifurcation and the risk of type II endoleak. Angiography of the medial sacral/L4 lumbar arteries demonstrated a type II endoleak, which was therefore coiled.
The right profunda femoris artery aneurysm was injected twice with 500 units of human thrombin via a microcatheter. The patient presented 1 year later with right profunda femoris artery rupture. This was embolized with Onyx (Covidien, Plymouth, MN). Two 5 Â 5.5 mm Vortex microcoils (Boston Scientific, Natick, MA) were then deployed proximal to the Onyx cast within the main profunda femoris artery [fig_ref] Figure 4: Flouroscopic image of ruptured right profunda femoris artery aneurysm after embolization with... [/fig_ref]. The patient was kept under continuous surveillance for a year and remained stable during this period.
# Discussion
Infective endocarditis IE is defined as an inflammation of the endocardium, most commonly affecting the valves. The aetiology is usually bacterial infection with S. aureus and streptococcus (80%). Staging, grading and classification is performed according to the modified Duke criteria. Imaging traditionally involves TOE. [bib_ref] Pathology and pathogenesis of infective endocarditis in native heart valves, Thiene [/bib_ref] [bib_ref] ACC/AHA 2006 guidelines for the management of patients with valvular heart disease, Bonow [/bib_ref] [bib_ref] Multislice computed tomography in infective endocarditis: comparison with transesophageal echocardiography and intraoperative..., Feuchtner [/bib_ref] The natural history of IE affecting the prosthetic valves includes abscesses, fistulas and valve dehiscence, resulting in paravalvular regurgitation. [bib_ref] Quantitative histological examination of mechanical heart valves, Lepidi [/bib_ref] Septic emboli and haematogenous seeding to remote sites are frequent (mostly neurological symptoms). In cases with severe or delayed diagnosis, the disease carries a high mortality rate of up to 40%.
## Mycotic aneurysms
A mycotic aneurysm is defined as an aneurysm arising from an infection of the arterial wall, usually bacterial. They rapidly grow as a focal saccular aneurysm arising eccentrically from the arterial wall. Mycotic aneurysms only represent 0.7-2.6% of all aortic aneurysms. Mycotic aneurysms are most common in the aorta, less commonly occurring elsewhere. [bib_ref] Symptomatic peripheral mycotic aneurysms due to infective endocarditis: a contemporary profile, González [/bib_ref] Pathologically, mycotic aneurysms are either primary, arising from a distant, unknown or remote source of infection, or secondary, arising from a specific source of infection. Bacterial aortitis is most commonly caused by salmonella or S. aureus.
On imaging, periarterial soft tissue stranding, oedema and fluid are seen on the CT images. Bacterial aortitis rarely calcifies, with the exception of syphilitic aortitis, which shows curvilinear calcifications. The infection may spread to the adjacent vertebral bodies or the psoas muscle if the aneurysm is aortic. On MRI, T 1 weighted images show periaortic low signal intensity in the absence of gadolinium (Gd), with subsequent aortic and periaortic enhancement following Gd, especially evident on fat-suppressed images. On T 2 weighted imaging, perianeurysmal high signal intensity is seen on fat-suppressed T 2 weighted images. A contrast-enhanced MR angiogram shows a saccular aneurysm(s) arising from the aortic wall. A nuclear medicine scan shows an increased uptake of labelled leukocytes at the site of the aneurysm. [bib_ref] Symptomatic peripheral mycotic aneurysms due to infective endocarditis: a contemporary profile, González [/bib_ref] Surgical resection/grafting following antibiotic therapy is the definitive treatment. Open surgical repair of infected arterial aneurysms carries a significant risk of mortality and morbidity. In situ graft revascularization is viable in afebrile patients or patients who have good response to preoperative antibiotic therapy. [bib_ref] In situ versus extraanatomic reconstruction for primary infected infrarenal abdominal aortic aneurysms, Lee [/bib_ref] [bib_ref] In situ replacement for mycotic aneurysms on the thoracic and abdominal aorta..., Uchida [/bib_ref] Extra-anatomic bypass grafting for infected infrarenal abdominal aneurysm has a similar long-term survival rate and should be considered in patients who are unsuitable for in situ graft revascularization; however, the postoperative complication rate is higher. [bib_ref] In situ versus extraanatomic reconstruction for primary infected infrarenal abdominal aortic aneurysms, Lee [/bib_ref] [bib_ref] In situ replacement for mycotic aneurysms on the thoracic and abdominal aorta..., Uchida [/bib_ref] [bib_ref] Single-center experience with open surgical treatment of 36 infected aneurysms of the..., Weis-Müller [/bib_ref] In situ grafts often employ a bactericidal coating (rifampicin bonded) with an omental pedicle graft as a cover; the more recent techniques include the use of cryopreserved arterial homograft. [bib_ref] In situ replacement for mycotic aneurysms on the thoracic and abdominal aorta..., Uchida [/bib_ref] [bib_ref] Infected abdominal aortic aneurysm treated by in situ replacement with cryopreserved arterial..., Kerzmann [/bib_ref] Endovascular treatment remains a controversial treatment option, but is considered if aneurysmectomy is not feasible in high-risk surgical candidates, although, even in this cohort, it is crucial that the infection is treated adequately prior to stent graft placement. [bib_ref] Outcome after endovascular stent graft treatment for mycotic aortic aneurysm: a systematic..., Kan [/bib_ref] [bib_ref] Stent-graft treatment of infected aortic and arterial aneurysms, Lee [/bib_ref] Stent graft placement provides valuable treatment of arterial fistulation, which was the feared complication in this case (fistulation between the common iliac artery and the ureter). [bib_ref] Endovascular stent graft repair for infected thoracic aortic pseudoaneurysms-a durable option?, Ting [/bib_ref] [bib_ref] Is endovascular repair of mycotic aortic aneurysms a durable treatment option?, Clough [/bib_ref] The natural history of mycotic aneurysms are always fatal if left untreated, with acute rupture/haemorrhage seen in 75% of cases. [bib_ref] Is endovascular repair of mycotic aortic aneurysms a durable treatment option?, Clough [/bib_ref]
## Learning points
1. Multiple mycotic aneurysms, presenting as renal colic, is rare and, to our knowledge, not previously described. 2. A high index of suspicion is essential for the diagnosis of this rare condition, as septic emboli can cause devastating sequelae and all untreated infected aneurysms eventually rupture. 3. Although aneurysmectomy and antibiotics is the treatment of choice, complicating factors such as multiple large mycotic aneurysms, acute risk of utereralarterial fistulation and warfarinization after valve replacement prohibited this definitive treatment and necessitated immediate aneurysm exclusion by endovascular treatment.
[fig] Figure 1: CT kidneys/ureters/bladder performed on admission to the emergency department with left renal colic. The indistinct left iliac vessels and ureter are noteworthy. [/fig]
[fig] Figure 2: Arterial phase contrast CT (maximum intensity projection) showing left common iliac and right profunda femoris artery aneurysms. [/fig]
[fig] Figure 3: Three-dimensional arterial phase CT scan showing stent graft of the left common iliac artery aneurysm, left ureteral stent and right profunda femoris artery (prior to thrombin injection, rupture or embolization coils). [/fig]
[fig] Figure 4: Flouroscopic image of ruptured right profunda femoris artery aneurysm after embolization with platinum coils. [/fig]
|
Hepatic sarcoidosis resembling primary sclerosing cholangitis
## Summary
We report the case of a 29-year-old man who presented with progressive weight loss, night sweats, abdominal pain and pruritus who was found to have obstructive jaundice and cholestatic pattern of liver injury on laboratory workup. Though findings on magnetic resonance cholangiopancreatography were initially concerning primary sclerosing cholangitis, he was ultimately diagnosed with biliary sarcoidosis after a liver biopsy. This case brings attention to the rare phenomenon of hepatic sarcoidosis causing hyperbilirubinemia and highlights the importance of reaching the correct diagnosis early, as the patient's symptoms improved after initiation of steroids.
# Background
Sarcoidosis is a multisystemic granulomatous disease of unknown origin that affects individuals worldwide, with highest incidence and prevalence in African Americans.It is characterised by the presence of non-caseating granulomas on tissue biopsy and can involve any organ, including the liver. Hepatic sarcoidosis was found in 11.5% of patients with sarcoidosis and clinical manifestation vary significantly; most of the patients are asymptomatic but 5%-15% develop signs and symptoms of cholestasis, portal hypertension, cirrhosis and Budd-Chiari syndrome.Cholestatic liver disease in hepatic sarcoidosis could be intrahepatic or extrahepatic. Intrahepatic sarcoidosis can resemble primary sclerosing cholangitis (PSC) or in rare circumstances these two entities can coexist.We present a case report of a patient with radiological findings 'typical' for PSC and liver biopsy showed non-caseating granulomas consistent with hepatic sarcoidosis.
## Case presentation
A 29-year-old man with no significant medical history presented to the emergency department as referred by his general practitioner for scleral icterus. His symptoms began 6 months ago with shortness of breath, dry cough, nausea, vomiting, epigastric abdominal pain, fatigue and night sweats along with unintentional weight loss of 60 lbs. For the past 3 weeks he experienced intense body itching that is worse at night and unalleviated with over-the-counter antihistamines. He also noticed yellow discolouration of his eyes 3 days ago. He does not take any prescription medications or herbal supplements. He denies tobacco, alcohol, or recreational drug use. Family history is insignificant.
On review of vital signs, the patient was afebrile, normotensive with blood pressure 132/82 mm Hg and heart rate of 87 bpm. Physical examination was pertinent for icteric sclera, but otherwise unremarkable.
## Investigations
Initial laboratory analysis was significant for alanine aminotransferase of 98 U/L, aspartate aminotransferase of 124 U/L and alkaline phosphatase (ALP) of 790 U/L demonstrating a mixed pattern cholestatic and hepatocellular liver injury. Furthermore, total bilirubin was 11.1 mg/dL with a direct bilirubin of 8.1 mg/dL. The rest of the initial laboratory workup can be found in table.
Chest X-ray demonstrated bilateral hilar lymphadenopathy. A right upper quadrant abdominal ultrasound showed no stones within the gallbladder, a normal common bile duct measuring 0.2 cm and no intrahepatic biliary ductal dilatation.
Further testing revealed absence of autoantibodies, including negative antinuclear antibody, PR3-ANCA, anti-liver-kidney-microsomal antibody, anti-mitochondrial antibody (AMA) and smooth muscle antibody. The patient had a normal IgG4 level. Eventually, an ACE level returned elevated at 104 U/L, and soluble interleukin-2 receptor levels elevated to 2576 U/mL. He underwent Magnetic resonance cholangiopancreatography (MRCP) which showed peripheral intrahepatic biliary ductal dilatation consistent with changes of primary sclerosing cholangitis (figure 1). A liver biopsy performed by the interventional radiology team showed non-caseating granulomas. A CT of the chest without contrast confirmed bilateral hilar and mediastinal adenopathy, but showed no parenchymal lung disease. Given the diagnostic uncertainty, he underwent a colonoscopy, which was unremarkable.
## Differential diagnosis
The differential is broad in a patient presenting with jaundice and pruritus whose laboratory workup show conjugated hyperbilirubinemia. A useful branch point is determining if the obstruction is extrahepatic or intrahepatic. A history that is inconsistent with an acute presentation and an unremarkable right upper quadrant ultrasound narrows the differential to intrahepatic causes of cholestasis. With findings of ductal dilatation and Case report non-necrotising granulomas on biopsy, the leading differential for our patient included primary sclerosing cholangitis, primary biliary cholangitis, and biliary sarcoidosis.
## Treatment
Patient found mild symptomatic relief with cholestyramine and ursodiol. After a negative fungal workup, he was discharged home on budesonide 3 mg two times a day.
## Outcome and follow-up
On outpatient follow-up, he reported doing well since discharge. His jaundice and pruritus have improved. Three months after starting steroids, his transaminases also normalised. Interestingly, ALP level remained elevated.
# Discussion
Though systemic sarcoidosis can frequently involve the liver, the granulomatous infiltration is only clinically significant in 5%-15% of cases.Biliary sarcoidosis leading to symptoms of cholestasis have been reported in literature in a dozen of cases; among many of these an alternative diagnosis was first considered, PSC being the top contender. Overall, patients' presentations ranged from minimal symptoms to biliary sepsis,however, jaundice and elevated ALP levels are common findings. Non-invasive (ie, MRI) and biliary tree imaging (ie, MRCP) demonstrate stricture or 'beading' of intrahepatic biliary ducts, and less commonly, extrahepatic involvement.Interestingly, most reported cases proceed to Endoscopic retrograde cholangiopancreatography (ERCP) after a cross-sectional imaging modality. While more invasive, ERCP has the ability to obtain brush cytology in cases where cholangiocarcinoma is high on the differential. As a sum, these cases demonstrate that the correct diagnosis can save the patient from invasive procedures, emotional stress in cases of misdiagnosed cholangiocarcinoma,and even surgery, such as liver transplantation and exploratory laparotomies.The leading mimicker PSC is a chronic, progressive disease that insidiously destroys the biliary ducts and eventually leads to cirrhosis and liver transplantation. It is more common in men than women and has a strong association with inflammatory bowel disease. Anti-neutrophil cytoplasmic antibodies, and more recently, the subtype PR3-ANCA, are specific but poorly sensitive (38.5%) for PSC.Unlike sarcoidosis, PSC does not have systemic manifestations, but instead it is associated with other autoimmune diseases such as hypothyroidism, type 1 diabetes mellitus and autoimmune hepatitis.Interestingly, granulomas can be found in 5% of liver biopsies in patients with confirmed PSC.Of note, at least two cases of coexisting diagnosis of PSC and sarcoidosis have been described. In one, a patient with ulcerative colitis was diagnosed with PSC, and 1 year later developed pulmonary sarcoidosis.Another case featured a patient with 6-year history of sarcoidosis who was found to have PSC and succumbed to cholangiocarcinoma.Both authors propose shared pathogenesis in immunological pathways. On liver biopsy, neither patient had granulomas. A brief overview of the similarities and differences of PSC and hepatic sarcoidosis have been summarised in table 3.
Other granulomatous diseases of the liver exist, such as primary biliary cholangitis. The two can be differentiated based on the characteristics of the granuloma-in PBC, granulomas tend to be portal based and poorly formed whereas in sarcoidosis granulomas are well formed and associated with presence of giant cells.Additionally, AMA is 90% sensitive and specific to PBC. The role of ACE in diagnosing sarcoidosis is limited due to its high levels in other granulomatous diseases and its low sensitivity, around 40%-60%.Biopsy is the gold standard for establishing the diagnosis, but a biopsy taken at the time of extensive cholestatic injury may only show nonspecific changes, and occasionally a second biopsy after the initial insult has resolved is needed for diagnosis. Oral glucocorticoids are the most common first-line treatment. Response to steroid therapy varies; most patients with biliary sarcoidosis in case reports have resolution of symptoms and improvement in liver function tests with a single short course. More extensive pulmonary involvement appears to predict persistent symptoms and liver enzyme elevations after steroids.There have not been trials to study treatment in biliary sarcoidosis. In a retrospective study of 27 patients with hepatic sarcoidosis, 83% responded to medical treatment with either oral glucocorticoids, anti-metabolites or biological agents. In this group, 40% (9 out of 22) of patients who initially presented with elevated ALP had levels that were lowered but not normalised, similar to that in our patient.In a similar study, 345 patients with sarcoidosis at a single US centre were screened for hepatic involvement-19 patients had abnormal liver tests, liver imaging or abnormal liver at autopsy (3 patients).Of the patients who received oral glucocorticoids, 27% of patients had lower but not normalised ALP at the end of follow-up. Lack of normalisation of the ALP was not correlated with worse outcome (ie, development of cirrhosis) in this study.Both studies were conducted to identify hepatic involvement retrospectively in patients with sarcoidosis. None of Incidence of sarcoidosis: 1-40 in 100 000 with symptomatic liver involvement in 5%-30% (up to 80% have asymptomatic involvement). Women>men. 3× higher incidence in AA.
## Clinical presentation
Abdominal pain, pruritus, diarrhoea, jaundice, fatigue and fever, but many patients are asymptomatic and diagnosed on incidental laboratory work.
Jaundice, nausea, vomiting, abdominal pain and hepatosplenomegaly.
Serological markers Elevated ALP, elevated aminotransferase levels but less than three times upper limit. Bilirubin is normal in 60% of patients. Positive ANCA (84%), anticardiolipin antibodies (66%) and ANA (53%).
Elevated ALP, ALT and AST. Negative ANA, AMA and ASMA. Normal IgG, normal or elevated ACE. Normal or elevated calcium and vitamin D.
Imaging features Multifocal intrahepatic and/or extrahepatic biliary strictures. 'Beading appearance'.
No imaging findings are sensitive or specific for diagnosis. May show diffuse heterogeneity of the hepatic parenchymal on cross-sectional imaging.
## Histology
Cholangitis, periductal fibrosis, ductular reaction and ductopenia. 'Onion skin' appearance (specific but not sensitive).
Non-caseating granulomas with macrophages with surrounding inflammatory cells.
Natural history Progresses to end-stage liver disease requiring liver transplant. At increased risk of developing cholangiocarcinoma.
3%-18% develop portal hypertension.
## Treatment
No targeted medical therapy. Liver transplantation only curative treatment.
Corticosteroids normalise liver tests and reduce granulomas, but they may not prevent progression.
AA, African americans; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AMA, Antimitochondrial antibody ; ANA, Antinuclear antibody; ANCA, Antineutrophil cytoplasmic antibodies; ASMA, Anti-smooth muscle antibody ; AST, aspartate aminotransferase; IBD, Inflammatory bowel disease; PSC, primary sclerosing cholangitis.
the patients in the Ungprasert study presented with jaundice or pruritus, and only a quarter in Sedki's group experienced cholestatic symptoms. Consequently, the conclusions drawn from these patients may not apply to patients whose disease primarily involves the biliary tree. An early response to steroids and the lack of extensive systemic involvement seems to predict a better outcome, however, it is not known what factors predict a favourable response to steroids. More research is needed to delineate the optimal treatment and natural history of sarcoidosis involving the hepatobiliary system.
## Learning points
► Systemic sarcoidosis can frequently involve the liver and granulomatous infiltration is only clinically significant in 5%-15% of cases. ► A multisystemic approach is required to establish the diagnosis of hepatic sarcoidosis that may resemble primary sclerosing cholangitis on imaging studies. ► While glucocorticoids are the most used first-line therapy, there are no randomised controlled trials for their use in hepatic sarcoidosis. Levels of alkaline phosphatase may remain elevated even after treatment with glucocorticoids. The significance of this finding is unclear. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.