Genome-wide epigenomic analyses in patients with nociceptive and neuropathic chronic pain subtypes reveals alterations in methylation of genes involved in the neuro-musculoskeletal system

Texte intégral

(1)Article. Genome-wide epigenomic analyses in patients with nociceptive and neuropathic chronic pain subtypes reveals alterations in methylation of genes involved in the neuro-musculoskeletal system STENZ, Ludwig, et al.. Reference STENZ, Ludwig, et al. Genome-wide epigenomic analyses in patients with nociceptive and neuropathic chronic pain subtypes reveals alterations in methylation of genes involved in the neuro-musculoskeletal system. Journal of Pain, 2021. DOI : 10.1016/j.jpain.2021.09.001 PMID : 34547430. Available at: http://archive-ouverte.unige.ch/unige:155120 Disclaimer: layout of this document may differ from the published version..

(2) Journal Pre-proof. Genome-wide epigenomic analyses in patients with nociceptive and neuropathic chronic pain subtypes reveals alterations in methylation of genes involved in the neuro-musculoskeletal system Ludwig Stenz , Joane Le Carré , François Luthi , Philippe Vuistiner , Cyrille Burrus , Ariane Paoloni-Giacobino , Bertrand Léger PII: DOI: Reference:. S1526-5900(21)00332-1 https://doi.org/10.1016/j.jpain.2021.09.001 YJPAI 4015. To appear in:. Journal of Pain. Received date: Revised date: Accepted date:. 3 April 2021 25 August 2021 11 September 2021. Please cite this article as: Ludwig Stenz , Joane Le Carré , François Luthi , Philippe Vuistiner , Cyrille Burrus , Ariane Paoloni-Giacobino , Bertrand Léger , Genome-wide epigenomic analyses in patients with nociceptive and neuropathic chronic pain subtypes reveals alterations in methylation of genes involved in the neuro-musculoskeletal system, Journal of Pain (2021), doi: https://doi.org/10.1016/j.jpain.2021.09.001. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © Published by Elsevier Inc. on behalf of United States Association for the Study of Pain, Inc..

(3) Genome-wide epigenomic analyses in patients with nociceptive and neuropathic chronic pain subtypes reveals alterations in methylation of genes involved in the neuro-musculoskeletal system Ludwig Stenz1, Joane Le Carré2,3, François Luthi2,4,5, Philippe Vuistiner2,3, Cyrille Burrus 2,4, Ariane Paoloni-Giacobino1, #, Bertrand Léger2,3, #,*.. 1. Department of Genetic Medicine and Development, Geneva University, Medicine Faculty, Geneva,. Switzerland 2. Institute for Research in Rehabilitation, Clinique romande de réadaptation, Sion, Switzerland. 3. Department of Medical Research, Clinique romande de réadaptation, Sion, Switzerland. 4. Department of Musculoskeletal Rehabilitation, Clinique romande de réadaptation, Sion, Switzerland. 5. Department of Physical Medicine and Rehabilitation, Orthopaedic Hospital, Lausanne University. Hospital, Switzerland.. * Corresponding author: Bertrand Léger PhD Department of medical research Avenue Grand Champsec 90 1951 Sion Switzerland Phone +41 27 603 2075 Fax. +41 27 603 2072. e-mail : Bertrand.Leger@crr-suva.ch #. contributed equally. Disclosures: This study was supported by a Grant from the State of Valais. The funder had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare that they have no competing interests. Running title : Epigenomic analysis in patients with chronic pain 1.

(4) Abstract. Nociceptive pain involves the activation of nociceptors without damage to the nervous system, whereas neuropathic pain is related to an alteration in the central or peripheral nervous system. Chronic pain itself and the transition from acute to chronic pain may be epigenetically controlled. In this cross-sectional study, a genome-wide DNA methylation analysis was performed using the blood DNA reduced representation bisulfite sequencing (RRBS) technique. Three prospective cohorts including 20 healthy controls (CTL), 18 patients with chronic nociceptive pain (NOCI), and 19 patients with chronic neuropathic pain (NEURO) were compared at both the single CpG and differentially methylated region (DMR) levels. Genes with DMRs seen in the NOCI and NEURO groups belonged to the neuro-musculoskeletal system and differed between NOCI and NEURO patients. Our results demonstrate that the epigenetic disturbances accompanying nociceptive pain are very different from those accompanying neuropathic pain. In the former, among others, the epigenetic disturbance observed would affect the function of the opioid analgesic system, whereas in the latter it would affect that of the GABAergic reward system. This study presents biological findings that help to characterize NOCI- and NEURO-affected pathways and opens the possibility of developing epigenetic diagnostic assays.. Perspectives. Our results help to explain the various biological pathways modifications underlying the different clinical manifestations of nociceptive and neuropathic pains. Furthermore, the new targets identified in our study might help to discover more specific treatments for nociceptive or neuropathic pains.. Data access RRBS data are available from the Gene Expression Omnibus of the National Center for Biotechnology Information under accession number GSE166255.. Keywords: DNA methylation, chronic pain, biomarkers, neuropathic pain, nociceptive pain.. 2.

(5) INTRODUCTION. Chronic pain conditions. 40. demonstrate complex interactions between biological, psychological,. environmental, and social factors, and their etiology and physiopathology remain obscure 19. In 2019, the International Association for the Study of Pain (IASP) defined two main types of pain, nociceptive and neuropathic, and established various criteria for their description. Concerning chronic nociceptive pain (NOCI), the trigger is an activation of the nociceptors, either at the surface of the body (i.e., skin) or at a deeper level (i.e., muscle, bones, joints, or tendons), without damage to the somatosensory nervous system. Patients experiencing NOCI report pain localized in the area of the injury or of musculoskeletal origin. 44. . In chronic neuropathic pain (NEURO), the trigger is an alteration in the. central or peripheral nervous system, such as trauma or infection. Patients experiencing NEURO 21. report sensations of burning, numbness, tingling. , mechanical hypersensitivity. 40. or allodynia or. hyperalgesia 42.. These two types of pain, in particular that which occurs due to an inflammatory response, induce 33. transcriptional and translational changes in neurons at the central level. that induce pain. hypersensitization by altering the pain threshold 50.. The epigenome is the interface between signals from the environment and genetic modifications that affect gene expression. Epigenetic factors modify the DNA structure through the (de)methylation of subsets of CpG islands, named differentially methylated regions (DMRs). 39. . Methylation in the. promoter region generally induces a decrease in the expression of the corresponding gene 4. Although epigenetic modifications are needed for normal development, they may also be responsible for some disease states. Various environmental factors, such as diet, pollutants, and psychological stressors, have been associated with epigenetic changes 2.. Epigenetic factors might be involved in the transition from acute to chronic pain and chronic pain maintenance 7. A few studies have confirmed the epigenetic control of the two main types of chronic pain, NOCI and NEURO, in humans. Concerning NOCI, two genome-wide methylation analyses have. 3.

(6) been performed in patients with chronic musculoskeletal pain showing DMRs at sites enriched in neurological pathways. 32. and in the immune and GABAergic signaling systems. 36. . In NEURO patients,. epigenetic mechanisms have been shown to control the altered expression of neuronal ion channels and receptors, as well as glial cell and macrophage pro-inflammatory cytokine and chemokine production 38. Methylation patterns associated with chronic neuropathic pain occurring in chemotherapy or diabetes have also been described. Peripheral neuropathy induced in some breast cancer patients by chemotherapy treatment was associated with differential methylations in the hypoxia-inducible factor I (HIF-I)- signalling pathway 25. A genome-wide methylation study of diabetic mice with neuropathic pain allowed identifying gene pathways contributing to diabetes or pain 10. However, no genome-wide methylation analysis has been performed to date in patients with NEURO.. A systematic characterization of the DNA methylation profiles of the two main types of chronic pain, NOCI and NEURO, is needed in humans. Furthermore, it would be interesting to compare the epigenetic signatures of patients with NOCI and NEURO to better understand their molecular identities. Therefore, by recruiting relatively large cohorts of patients, we aimed to compare blood DNA methylation levels in NOCI and NEURO patients with healthy controls (CTL). The main hypothesis guiding our investigations was that DNA methylation levels of patients NOCI and NEURO differs both from controls and also from each other. To test this hypothesis we performed a genome wide blood DNA methylation analysis of patients with different origin of pain and healthy controls.. METHODS. Study population. Patients were recruited from the Musculoskeletal Rehabilitation Department of the Clinique romande de réadaptation between August 2015 and January 2018. Patients came to our service to complete their reeducation program after a trauma because of a need for intensification of therapies that was not possible on an ambulatory basis, or when the evolution needed the evaluation of multidisciplinary 4.

(7) specialists. Patients were addressed to our center by their physician or by the insurance in charge after their injury. In our department, patients mainly suffer from musculoskeletal deficiency after fractures, ligamentous injuries or soft tissue contusions. Patients older than 18 years old, suffering from chronic post-traumatic pain with neuropathic or nociceptive characteristics were included. Exclusion criteria were: history of diabetes, blood or infectious disorders (hepatitis, thalassemia, HIV or systemic infections), widespread pain syndrome and severe psychiatric comorbidity.. Patients were also. excluded if they suffered from major health conditions such as malignancy or chronic inflammatory disease (e.g., rheumatoid arthritis, systemic lupus erythematosus). Upon admission, a senior clinician assessed whether the patients met the inclusion criteria. Based on a thorough medical examination and the results of the DN4 questionnaire. 6. patients were classified into one of the two groups of chronic. pain (NEURO or NOCI) according to the recommendations of the IASP. Categorization as NEURO patients was done after a specialized neurological evaluation including ENMG and/or imaging to confirm the plausibility of the peripheral lesion causing the neuropathic pain. 42. . Twenty healthy. volunteers matched for sex, age, and BMI who reported no pain and taking no medication were recruited as the control group. An approximate number of patients was estimated from a previous study on DNA methylation profile in pain patients 7. In order to minimize potential sources of bias, patients were matched on the basis of their age, sex, BMI, time since injury, location of the lesion, pain medication, pain levels, anxiety and depression levels. Psychological status and pain level data were obtained using validated questionnaires, while clinical information and medication status were collected from medical records. After obtaining informed consent, all volunteers were scheduled for blood sampling, as previously described. 15. . The study protocol was approved by the local ethics. committee (CCVEM 034/12) and was conducted according to the recommendations of the Declaration of Helsinki.. Questionnaires The DN4 is a validated questionnaire that has been developed to diagnose polyneuropathy 6. DN4 is a clinician-administered questionnaire composed of 7 items related to pain symptoms, and 3 items linked with neuropathic pain examination. Healthcare professionals assess whether there is a reduced. 5.

(8) sensation to touch or a pinprick (hypoesthesia) and whether light brushing increases or causes pain (allodynia). The Brief Pain Inventory (BPI) is a self-administered questionnaire that allows patients to rate the severity of their pain and the degree to which their pain interferes with the common components of feeling and function. It has been broadly used in studies on pain 13. The Hospital Anxiety and Depression Scale (HADS) was used to detect states of anxiety and depression. It is a self-administered questionnaire that has been specifically developed in a hospital medical outpatient clinical setting 54.. DNA extraction and quality control. DNA was extracted from blood samples in three cohorts composed of 20 CTL, 20 patients with NOCI, and 20 patients with NEURO using the DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany) with minor modifications. Briefly, 40 µL of proteinase K and 200 µL of PBS were added to 200 µL of whole blood. RNA was then removed following the addition of 800 µg of RNase A. After incubation for 2 min, 400 µL of Buffer AL was added to the RNA-free mix and incubated for 15 min at 56 °C. Before loading the solution onto the spin column, 400 µL of EtOH was added to the mixture. The next steps of the protocol were followed according to the manufacturer’s instructions. The elution step was performed twice with 200 µL in two separate microtubes to increase DNA yield. DNA concentrations were measured using the Qubit® dsDNA BR Assay Kit (Thermo Fisher Scientific, Waltham, USA). DNA quality (relative to fragmentation levels) was assessed using the Fragment Analyzer TM and the DNF-488 High Sensitivity Genomic DNA Analysis Kit (Agilent, Santa Clara, United States). Three samples did not pass QC (pics at 12.3, 12.4 and 11.1 kb). Fifty-seven samples passed QC (size of DNA presenting a pic at 17 ± 2 kb); consequently 20 CTL, 18 patients with NOCI, and 19 patients with NEURO were included in the analysis.. Preparation of RRBS libraries and sequencing. 6.

(9) DNA was analyzed for cytosine methylation on a genome-wide scale using reduced representation bisulfite sequencing (RRBS). RRBS experiments were conducted using DNA methylation profiling (RRBS Service, Diagenode, Cat. G02020000). Libraries were generated and sequenced in paired-end mode of 50 bp using a HiSeq4000 instrument (Illumina Inc.) using one full lane for five multiplexed samples.. Analyses of RRBS data were performed using a bioinformatics pipeline written in Bash, Python, and R developed by Matthias Beyens (BISC Global). We implemented the pipeline directly in a highperformance computing cluster (HPC) named “Baobab” that belongs to the University of Geneva. We performed three comparisons: CTL vs. NOCI, CTL vs. NEURO, and NOCI vs. NEURO. The pipeline performed the bsseq-based detection of DMRs and the Fisher test of methylation levels at CpG sites on each chromosome separately. For each comparison, the different tables containing the results obtained on each chromosome in R were merged before the execution of the downstream steps (overlapping, consistency, and Wilcoxon testing, explained afterwards).. Read trimming and mapping. We trimmed the reads using TrimGalore (version 0.6.0) with the following options: --non-directional, --paired, --rrbs, --phred33, and with cutadapt (version 2.3) using the following options: --quality 30, -length 35. Trimmed reads were mapped against the human bisulfite genome (assembly GRCH38) using Bismark (version 0.22.1) running with Bowtie 2 (version 2.3.5.1) and with the specified options: --q, --phred33, --score-min L,0,-0.2, --ignore-quals, --no-mixed, --no-discordant, --devtail, --maxins 1000, option-non-directional. Alignment was performed on all strands (OT, OB, CTOT, and CTOB). We used the three-letter aligner Bismark to avoid overestimation of methylation levels appropriate for the wildcard aligner 5.. Methylation extraction. Methylation calls for every analyzed cytosine were performed using the Bismark methylation extractor algorithm with options --paired-end, --gzip, --bedGraph, --multicore 10, --cytosine_report, --ignore 3, 7.

(10) --ignore_r2 3, --ignore_3prime_r2 28. The “ignore” options corrected for the bias in methylation values found at the extremities of the reads.. C/T SNP filtration. Using a BASH script, we removed the methylation levels determined using the RRBS analysis that were located within known single nucleotide polymorphisms (SNPs) involving cytosine to thymine changes (C/T-SNP, 1000 genome project phase 3, ftp://ftp.ensembl.org/).. Identification of differentially methylated CpG. We used the fisherTests in the Bsseq package to test for differences in methylation between two groups for each of the three comparisons for all CpGs obtained after the SNP filtering process 22. We only kept CpGs with a p-value resisting Bonferroni correction for multiple testing (p<1.7e-9) and with a finite value for the logarithm in base 2 of the odds ratio, “log2(OR)”. Infinite values of log2(OR) are due to division by zero and were not considered for further analysis.. Identification of differentially methylated regions (DMRs). The BSmooth open source package was used to identify DMRs in R, taking biological variation into account. 22. . The quantile-based cutoff of the t-statistic was set at 0.025 and 0.975, and methylation. changes had to be present in at least three samples of the group with an absolute change in methylation of ≥0.25. The bsseq::dmrFinder function identified the DMRs reported based on the meanDiff and areaStat.. Identification of overlap between DMRs and CpG. This step validated DMRs and CpGs using two different approaches, resulting in the concordance of evidence. The GenomicRanges package was used for overlapping purposes. 29. . Files containing the. results of the Fisher exact test performed on individual CpGs and files containing the annotated DMRs were converted into GRanges objects using the “makeGRangesFromDataFrame” function in R. The. 8.

(11) “subsetByOverlaps” function in R identified DMRs overlapping with CpGs and CpGs overlapping with DMRs. Finally, we selected CpGs and DMRs that demonstrated consistent changes.. Annotation of DMRs and CpGs. DMRs were annotated depending on their overlap notably with the official gene symbols, as well as with the following genes structures: “1to5kb” (5-kb region upstream of promoters), “promoters,” “5UTRs,” “3UTR,” “exons,” “introns,” and “intronexonboundaries” (200 bp up/downstream of any boundary between an exon and intron). CpGs were annotated depending on their distance from a CpG island (CGI) as “islands” (overlapping), “shores” (2 kb of distance from one end), “shelf” (2–4 kb distance), and “inter” (>4 kb) using the Bioconductor package, “annotatr” 9. We manually annotated the CpGs that overlapped with DMRs.. Consistency. We defined consistency as methylation changes observed between significant CpGs and overlapping DMRs varying in the same direction according to the two groups compared. This means that if a CpG is hypermethylated in one group, its associated overlapping DMR should be hypermethylated in the same group and vice versa. Consistency was assessed using logical tests in Excel.. Functional analysis. Functional networks were detected using STRING (https://string-db.org/) following online submission of the lists of genes that were either hyper- or hypomethylated in each of the three comparisons performed 47. The reported gene functions were derived from the available literature.. Statistical analyses. The differences in socio-demographic and psychological data between the groups were evaluated by the Wilcoxon-Mann–Whitney rank-sum test.. 9.

(12) No corrections for potential batch effects were performed as all the samples were treated as a single batch for all experimental steps, including bisulfite treatment, RRBS libraries preparation and sequencingWe also performed non-parametric Wilcoxon tests between the compared groups in R using methylation percentages recorded at the single cytosine level taking into account coverage. We extracted the methylated (Cmeth) and unmethylated (Cunmeth), as well as the methylation levels from all samples, for all selected sites from the Bismark SNP-filtered coverage files. The percentage of methylation was equal to Cmeth/(Cmeth+Cunmeth) × 100, and coverage was equal to (Cmeth+Cunmeth)/number of samples.. RESULTS. Patient characteristics. There was no significant difference in clinical or psychological features (pain severity, anxiety and depression) between the two groups of patients, except for the results of the DN4 questionnaire, which was designed to discriminate neuropathic pain from other types (Table 1). As opioids were previously demonstrated to potentially influence methylation levels. 18. , we carefully monitored. medication intake for the entire cohort. Interestingly, there was no significant difference in medication use between the two groups of patients for the different classes of medications investigated (Supplementary Table 1).. Global DNA methylation changes at DMR and CpG sites. Through the RRBS analysis, we tested for differences in the distribution of methylation and unmethylation signals using the Fisher exact test for approximately 6 million CpGs in each comparison (CTL vs. NOCI, CTL vs. NEURO, NOCI vs. NEURO). We found 56,572 CpGs for the CTL vs. NOCI, 35,774 for the CTL vs. NEURO, and 70,730 for the NOCI vs. NEURO comparison 10.

(13) were below the genome-wide significant p-value threshold of 1.7e-9 and produced finite values for the log2(OR) parameter. The analysis of DMRs showed 2,066 significant DMRs in the CTL vs. NOCI, 1,098 in the CTL vs. NEURO, and 765 in the NOCI vs. NEURO comparison (Figure 1A).. Based on the overlap and consistency criteria, as well as the Wilcoxon test results (see methods), 26 DMRs were finally identified in the CTL vs. NOCI, 8 DMRs in CTL vs. Neuro and 20 DMR in NOCI vs. NEURO.. The genomic distribution of DMRs in the three groups was analyzed. Approximately 78–84% of DMRs were located in intergenic or intragenic regions, 13–17% in the promoter region, and 3–5% in the exon region (Supplementary Figure 1). The distribution of DMRs between chromosomes is shown in Supplementary Figure 2.. Genome-wide blood DNA methylation changes in NOCI and NEURO patients. When NOCI and NEURO patients were compared to CTL, more DMRs were identified in the NOCI group than in the NEURO group. Specifically, 39 CpGs overlapping 26 DMRs in the NOCI group compared with eight CpGs overlapping eight DMRs in the NEURO group were identified. This represented 4.9 and 3.3 more CpGs and DMRs, respectively in the NOCI group than in the NEURO group (Figure 1B). Notably, a Venn diagram of the selected DMRs (Figure 1 C) showed no overlap in the CTL vs. NOCI and CTL vs. NEURO comparisons.. However, in the two comparisons involving the NOCI group, five common genes with DMRs were found: RAB10, PNPLA6, LRRC59, BMP1, and P3H3 (Figure 1C).. The boxplots of the methylation levels at the single CpG site resolution of these five genes show the statistical significance of the comparisons between the NOCI and NEURO groups. PNPLA6 had the most statistically significant result (Wilcoxon p-value = 1.3 × 10-4) and the highest coverage (27-fold) at the single CpG site resolution (Figure 1D).. 11.

(14) Functional analyses of DMR genes in NOCI and NEURO patients shows an enrichment in neuro-musculoskeletal system-related genes. The direction of the methylation changes and the biological role of the genes with DMRs identified in the CTL vs. NOCI, CTL vs. NEURO, and NOCI vs. NEURO comparisons are shown in Figure 2. For the first functional analysis, DMR genes were segregated according to the tissue or system to which they belong, and five main families were identified: the nervous system (central and peripheral); skeletal muscle, bone, and connective tissue; and the inflammatory and immune systems. A few genes had two different tissue or system affiliations, which is also taken into account in Figure 2. Finally, networks were identified in each group comparison, as shown in Figure 2.. CTL vs. NOCI. In the CTL vs. NOCI comparison, robust enrichment was observed in the genes with DMRs associated with the neuro-musculoskeletal system (12/26 genes with DMRs, 48%). Ten genes with DMRs were found to be related to the nervous system, two to the skeletal muscle, and three to the bone. Eight genes with DMRs were related to the immune system or inflammation.. Functional analyses of genes with DMRs resulted in the identification of significant enrichment for the protein domain family PF12414 (false discovery rate (FDR) =0.001) in RBFOX-1 and RBFOX-3. RBFOX-1 mediates the neuron-specific splicing of the calcitonin gene, allowing for the neuronal production of calcitonin gene-related peptide (CGRP) 53. Interestingly, RBFOX3 was also associated with CNP in the STRING cluster (CL:619, FDR 0.02), which is involved in neuronal adhesion (Figure 2, upper panel).. CTL vs. NEURO. In the comparison between the CTL and NEURO groups, a large prevalence of genes with DMR associated with the neuro-musculoskeletal system was observed (6/8 DMR genes, 75%). Five genes. 12.

(15) with DMRs were related to the nervous system, one was related to the skeletal muscle, and one was related to the bone. One gene with DMRs was associated with the immune system.. Functional analyses of genes with DMRs resulted in the identification of two independent significant enrichments. The first enrichment (FDR 0.02) was observed in the histone acetylation pathway (HAS3214847) in the ELP2 and KANSL1 genes. A second enrichment (FDR 0.01) was associated with the calcium homeostasis STRING cluster (CL:9220) in the STOM and PVALB genes (Figure 2, middle panel).. NOCI vs. NEURO. In the comparison between the NOCI and NEURO groups, a robust enrichment in genes with DMRs associated with the neuro-musculoskeletal systems was observed (9/20 genes with DMRs, 45%). Six genes with DMRs were related to the nervous system, two were related to the skeletal muscle, and four were related to the bone. No genes with DMRs were associated with inflammation.. Functional analyses identified three independent enrichments. The first enrichment (FDR 003) is the “vesicle-mediated transport” (GO:0006892) involving the Golgi-related protein TGOLN2 in the RAB10 genes. The second enrichment (FDR 0.03) is the GO term “negative regulation of steroid” (GO:0010894) in the GFI1 and DKK3 genes. The third enrichment is “hematopoietic stem cell differentiation” (CL:17858) in the DIDO1 and C5orf56 genes (Figure 2, lower panel).. Supplementary Table 2 presents a list of the genes described in Figure 2, organized by selected tissues, cell types, and systems.. DISCUSSION. The results of this study, keeping in mind the relatively limited size of the cohorts, suggest the existence of NOCI and NEURO related epigenetic signatures. The two patient cohorts in this study. 13.

(16) were relatively small compared with those in previous chronic pain epigenome-wide studies. 8, 32. ;. however, the different types of pain, i.e., NOCI and NEURO, make the interpretation of the results particularly interesting.. Our epigenome-wide study was performed on whole blood because of the impossibility of obtaining human peripheral or central nervous system samples in living individuals. Previous studies have shown a correlation between brain tissue and blood methylation patterns. Davies et al. 14 reported that some inter-individual variations were reflected between brain and blood, although between-tissue variations could exceed it. Strong inter-individual correlation between brain and blood were observed in humans. 46. and rats 35. It is generally admitted that blood DNA methylation can be used as a proxy. for brain and/or other tissue-specific alterations. This justifies the extrapolation of data obtained from the whole blood to different human tissues, and blood was used as a proxy for the nervous system.. NOCI and NEURO differ in their etiology and consequences but have a common main symptom, pain, although it has different characteristics.At the statistical stringency used in this study, the epigenetic signatures in NOCI and NEURO groups differ from each other with no overlap.. Specific epigenetic effects can be induced either by the cause or secondary effects of chronic pain.. When interpreting the results, it should be kept in mind that false positives may still be present among the statistically significant genes with DMRs, but maximum care was taken to minimize their possibility. The criteria for selection could be classified as follows: a strong statistical significance of the comparisons, the biological relevance of the findings, and the presence of significantly enriched pathways. The statistical significance refers to the distribution of the C methylation and T unmethylation signals at the tested CpGs using the Fisher test; second, the methylation levels at the CpG (non-parametric Wilcoxon tests) and third, converging identifications with two independent approaches (a smoothing algorithm for DMRs and CpGs tested individually).. Top statistically significant candidates. 14.

(17) The five DMRs present not only in the NOCI vs. CTL comparison, but also in the NOCI vs. NEURO comparison may be considered top candidates. The relevance of the five top candidates was validated based on the observation that the direction of the methylation changes were the same in the two comparisons. RAB10, BMP1, and LRRC59 were hypomethylated in the NOCI group in both comparisons (NOCI vs. CTL and NOCI vs. NEURO), whereas PNPLA6 and P3H3 were hypermethylated.. Biological relevance. In the NOCI and NEURO groups, the function of the four genes might be associated with the causes or effects of pain: RAB10, RBFOX-1/CGRP, MAGI2, and PVALB.. RAB10, a member of the small GTPase family, was reported to increase the cell surface expression of the -opioid receptor (DOPr) in the nervous system and was therefore proposed to increase DOPragonist-mediated analgesia. 16. . RAB10 hypomethylation in patients with NOCI would most likely. increase its expression. Upregulation of the opioid system was observed under other conditions. Inflammation was reported to increase mu opioid receptor (MOR) expression at both the mRNA 41, 55 and protein 48 levels. This phenomenon may represent a compensatory reaction that reduces pain. The hypomethylation of RBFOX-1 in NOCI patients would increase the production of CGRP. 53. , an. important specific mediator of nociceptive pain, the involvement of which in neuropathic pain has not been demonstrated 43.. GABAergic signaling plays a key role in the initiation and/or maintenance of neuropathic pain. 30. .. MAGI2 (also called S-SCAM) is a synaptic scaffolding protein involved in GABAergic synapse transmission. Overexpression of MAGI2 in mice induced a loss of GABAergic synapses 45. PVALB is a calcium-binding protein expressed in GABAergic interneurons. Activating PVALB-expressing neurons in the dorsal horn neurons has been reported to decrease neuropathic pain and mechanical hypersensitivity 40 whereas activating PVALB-expressing neurons in the thalamic reticular nucleus has been reported to increase pain sensitivity 31 .. 15.

(18) MAGI2 and PVALB hypomethylation in patients with NEURO would contribute to increased GABAergic signaling by increasing the expression of these genes in neurons. The observation that MAGI2 and PVALB are differentially methylated in the NEURO group is in line with the involvement of GABAergic signaling in neuropathic pain.. Significantly enriched pathways. The involvement of a gene in an enrichment pathway should validate the relevance of its association with either the NOCI or NEURO.. The first enriched pathway detected in the NOCI group (Figure 2) involved RBFOX1 and RBFOX3, both of which play an important role in the nervous system. It is known that RBFOX family proteins interact, forming homo- and hetero-oligomeric complexes 11. RBFOX1-3 knockout has been found to result in brain defects in cytoskeletal membranes and synaptic proteins 24. RBFOX1 is connected with brain-derived neurotrophic factor (BDNF), a potent modulator of brain synaptic plasticity stimulates the neuronal synthesis of CGRP involved in the transmission of nociceptive pain. 49. and. 43. . The. second enrichment pathway detected in the NOCI group involved RBFOX3 and CNP. The latter gene is expressed in myelin-forming oligodendrocytes, and a missense variant of CNP is the cause of hypomyelinating leukodystrophy in humans 1.. The four genes implicated in the two enrichments detected in the NEURO group were all associated with the neuro-musculoskeletal system. ELP2 is an inhibitor of osteoblast differentiation. 51. and. KANSL1, found to be genetically associated with intellectual disability, should play a role in neurons 26. . PVALB is associated with both muscles and neurons involved in muscle relaxation. 12. and. GABAergic interneurons 17. Finally, STOM encodes stomatin, which has been shown to be involved in sensory neuron mechano-transduction 34.. Three enrichment pathways were identified in the NOCI vs. NEURO comparison. The first involves TGOLN2 and RAB10, the latter of which was found to sustain DOP receptor action, as well as CCDC91, which is associated with hip bone geometry 23 and mechanical stress-induced ossification 37. 16.

(19) A second enrichment involves GFI1 and DKK3, with the latter implicated in muscle atrophy. 52. and. impaired osteogenesis 3.. It is noteworthy that PVALB and RAB10, two of the three genes with DMRs highlighted above for their particular functional relevance to pain are found in the enriched pathways, supporting the fact that they might be strong candidates for chronic pain regulation.. Differences between NOCI and NEURO signatures. Comparison of the NOCI and NEURO DMRs shows that the two types of pain have different epigenetic targets.. NOCI epigenetically affects the opioid receptor by possibly modulating RAB10 and CGRP via RBFOX 1 and 3, whereas NEURO epigenetically affects the GABAergic system by possibly modulating MAG12 and PVALB. This suggests that NOCI and NEURO epigenetically affect pain genes involved in different pathophysiological processes. These results are consistent with a recent study by our group that showed completely different miRNA signatures in NOCI and NEURO patients 15. .. Genes with DMRs connected to the inflammatory system were detected only in the NOCI group, suggesting the more specific involvement of the inflammatory system in NOCI. Genes connected to the musculoskeletal system, although predominant in the NOCI, were also present in the NEURO group. Even if NEURO patients were selected in a population of patients with peripheral nervous system injury they all had musculoskeletal trauma that may explain this observation.. LIMITATIONS Nociplastic combination may be part of the process and was not addressed. We cannot rule out some nociplastic contribution to the chronic pain after a long post-injury evolution. The totality of our NOCI population suffered of musculoskeletal injury and in some cases had undergone surgery following trauma. Patients with widespread pain conditions or other nociplastic pain features (mechanical or. 17.

(20) thermal allodynia and/or any other clinically evoked pain hypersensitivity phenomena) were not included. 20, 27. . Nevertheless, some nociplastic adaptations may have developed in combination with no-. ciceptive or neuropathic pain processes. Epigenetic analyses and data obtained from blood cells were used here as a proxy for other tissues. Hydroxymethylation of DNA cannot be discriminated from methylation using the RRBS technique. The size of the cohorts is limited, which may reduce the statistical power and do not allow, for example, discrimination of the results according to gender. Moreover, this cross-sectional study does not allow us to exclude the possibility that DNA methylation levels change over time. A longitudinal study, would be necessary to evaluate possible variations in DNA methylation linked to the process of pain chronicization. CONCLUSIONS. In conclusion, this epigenome-wide study identified CpG methylation signatures in the blood, discriminating nociceptive from neuropathic types of chronic pain. Despite a relatively low number of patients, our model is quite robust and allows to reach a good level of generalization as our cohorts were matched for several variables known to influence methylation process (age, sex, medication, psychological factors,..).Genes associated with these signatures appear to have essential functions in the neuromusculoskeletal system. Additional studies are required to confirm these results.. Data sharing statement RRBS data are available from the Gene Expression Omnibus of the National Center for Biotechnology Information under accession number GSE166255. Acknowledgements We thank the nursing teams of the Clinique romande de réadaptation, Sion, Switzerland, who performed the blood sampling and collected the clinical information of the patients.. 18.

(21) Funding This study was supported by the Department of education from the State of Valais, Switzerland.. 19.

(22) Figures legends Figure 1. Blood DNA methylation changes in chronic neuropathic pain (NEURO) and chronic nociceptive pain (NOCI) patients as compared with healthy controls (CTL). A. Analytical workflow of pairwise comparison of NEURO, NOCI, and CTL RRBS-derived datasets. The workflow appears in successive lanes and the three pairwise comparisons in columns. Lane 1: number of CpGs individually tested by the Fisher exact test. The test estimates associations between read-based methylated and unmethylated calls (C or T) with the patient’s status. Lane 2: p-value threshold selected for considering a genome-wide significant association. Lane 3: number of CpG sites passing the threshold, thus considered significant. Lane 4: number of significant CpGs with a finite value for log2(OR). Lane 5: number of DMRs identified using the bisulfite sequencing smoothing algorithm (BSmooth). Lanes 6a and 6b: number of DMRs and significant CpGs overlapping each other, and, in lanes 7a and 7b: considered consistent. Consistency means that the methylation changes observed between overlapping CpGs and DMRs vary in the same direction. Lane 8a: final selected differentially methylated CpGs between both tested groups according to the non-parametric Wilcoxon test. Lane 8b: DMRs overlapping with the final selected CpGs. B. Genes found to be differentially methylated based on the final overlapping DMRs and CpGs (lanes 8a and 8b of panel A). Horizontal bars represent DMRs and dots represent CpGs. Note that more than one CpG may overlap with the same DMR. Hypo- and hypermethylation refer to the mentioned group in each pairwise comparison. Bars and dots corresponding to hypomethylated DMRs and CpGs are white, whereas bars and dots corresponding to hypermethylated DMRs and CpGs are black. C. Venn diagram based on the final selected DMR genes, showing no overlap between the NOCI and NEURO groups but five DMR genes (RAB10, PNPLA6, LRRC59, BMP1 and P3H3) shared by the two comparison groups involving NOCI (CTL vs. NOCI, NOCI vs. NEURO). D. Boxplot of methylation levels at a single CpG site resolution in the five genes in panel C, with genomic locations in GRCH38 and RRBS coverage. Dots in boxplots represent individual methylation levels (19 NEURO, 18 NOCI, and 20 CTL). CTL are in light gray, NEURO in white, and NOCI in dark gray. Fisher test-derived statistical significance with (*) p-value < 0.05, (**) p-value < 0.01, (***) p-value < 0.005.. 20.

(23) Figure 2. Functional analysis of genes with DMRs in NOCI and NEURO patients Genes with DMRs identified in the comparison groups are shown: CTL vs. NOCI in the upper panel, CTL vs. NEURO in the middle panel, and NOCI vs. NEURO in the lower panel. For each comparison, genes with DMRs were segregated according to the direction of the methylation changes observed: hypomethylation in the left and hypermethylation in the right column. Genes that belong to a functional network as detected by STRING. 47. appear surrounded and linked together by a line, with 21.

(24) the corresponding description and code as well as the false discovery rate (FDR). If there is no line between two genes, the interaction between these genes is not detectable by the STRING program. This may either reflect a true absence of connection or insufficient knowledge. Functional networks may imply Homo sapiens (HSA) pathways either based on cellular processes (Kyoto Encyclopedia of Genes and Genomes, KEGG), or based on molecular reactions (Reactome), a gene ontology term (GO), functionally related protein family (INTERPRO), conserved protein domain (PF), and STRING cluster (CL). Genes with a functional link to the nervous system are shown in purple, those linked to the muscle in red, those linked to the bone and connective tissue (referred to as bone) in white, and those linked to the inflammatory system or immunity in green. For genes linked to two different systems, two different colors were used, one for the gene name and the other for the surrounding frame. Genes with no links to the above systems are shown as dashed squares.. 22.

(25) Supplementary Figure 1: Genomic distribution of the DMRs identified in the participants of this study described in Table 1 Supplementary Figure 2: Chromosomal distribution of the DMRs identified in the participants of this study described in Table 1 Tables legends Table 1 Clinical characteristics of the participants Supplementary Table 1: Analgesic medication in NOCI and NEURO patients (sorted by ATC code) Supplementary Table 2: Genes with DMRs organized by their tissue or system distribution 23.

(26) 24.

(27) References. 1. 2. 3.. 4. 5. 6. 7.. 8. 9. 10.. 11. 12. 13. 14.. 15. 16.. 17.. Al-Abdi L, Al Murshedi F, Elmanzalawy A, Al Habsi A, Helaby R, Ganesh A, Ibrahim N, Patel N, Alkuraya FS. CNP deficiency causes severe hypomyelinating leukodystrophy in humans. Hum Genet. 139:615-622, 2020 Alegria-Torres JA, Baccarelli A, Bollati V. Epigenetics and lifestyle. Epigenomics. 3:267-277, 2011 Aslan H, Ravid-Amir O, Clancy BM, Rezvankhah S, Pittman D, Pelled G, Turgeman G, Zilberman Y, Gazit Z, Hoffmann A, Gross G, Domany E, Gazit D. Advanced molecular profiling in vivo detects novel function of dickkopf-3 in the regulation of bone formation. J Bone Miner Res. 21:1935-1945, 2006 Bird A. DNA methylation patterns and epigenetic memory. Genes & Development 16:6-21, 2002 Bock C. Analysing and interpreting DNA methylation data. Nat Rev Genet. 13:705719, 2012 Bouhassira D, Attal N, Fermanian J, Alchaar H, Gautron M, Masquelier E, Rostaing S, Lanteri-Minet M, Collin E, Grisart J, Boureau F. Development and validation of the Neuropathic Pain Symptom Inventory. Pain. 108:248-257, 2004 Bruehl S, Gamazon ER, Van de Ven T, Buchheit T, Walsh CG, Mishra P, Ramanujan K, Shaw A. DNA methylation profiles are associated with complex regional pain syndrome following traumatic injury. Pain.10.1097/j.pain.0000000000001624, 2019 Burri A, Marinova Z, Robinson MD, Kuhnel B, Waldenberger M, Wahl S, Kunze S, Gieger C, Livshits G, Williams F. Are Epigenetic Factors Implicated in Chronic Widespread Pain? PLoS One. 11:e0165548, 2016 Cavalcante RG, Sartor MA. annotatr: genomic regions in context. Bioinformatics. 33:2381-2383, 2017 Chen W, Lan T, Sun Q, Zhang Y, Shen D, Hu T, Liu J, Chong Y, Wang P, Li Q, Cui W, Yang F. Whole Genomic DNA Methylation Profiling of CpG Sites in Promoter Regions of Dorsal Root Ganglion in Diabetic Neuropathic Pain Mice. J Mol Neurosci.10.1007/s12031-021-01847-1, 2021 Choi S, Kim YE, Kim JW, Cho N, Cheon S, Kim KK. Rbfox family proteins make the homo- and hetero-oligomeric complexes. Biochem Biophys Res Commun. 495:1022-1027, 2018 Chu WY, Liu LS, Chen L, Yang PH, Li YL, Wang YH, Wang JH, Wu P, Zhu X, Zhang JS. Rapid Muscle Relaxation in Siniperca chuatsi is Coordinated by Parvalbumin (PVALB) and MiR-181a. Curr Mol Med. 15:772-779, 2015 Cleeland CS, Ryan KM. Pain assessment: global use of the Brief Pain Inventory. Ann Acad Med Singap. 23:129-138, 1994 Davies MN, Volta M, Pidsley R, Lunnon K, Dixit A, Lovestone S, Coarfa C, Harris RA, Milosavljevic A, Troakes C, Al-Sarraj S, Dobson R, Schalkwyk LC, Mill J. Functional annotation of the human brain methylome identifies tissue-specific epigenetic variation across brain and blood. Genome Biol. 13:R43, 2012 Dayer CF, Luthi F, Le Carre J, Vuistiner P, Terrier P, Benaim C, Giacobino JP, Leger B. Differences in the miRNA signatures of chronic musculoskeletal pain patients from neuropathic or nociceptive origins. PLoS One. 14:e0219311, 2019 Degrandmaison J, Abdallah K, Blais V, Genier S, Lalumiere MP, Bergeron F, Cahill CM, Boulter J, Lavoie CL, Parent JL, Gendron L. In vivo mapping of a GPCR interactome using knockin mice. Proc Natl Acad Sci U S A. 117:13105-13116, 2020 Del Tongo C, Carretta D, Fulgenzi G, Catini C, Minciacchi D. Parvalbumin-positive GABAergic interneurons are increased in the dorsal hippocampus of the dystrophic mdx mouse. Acta Neuropathol. 118:803-812, 2009. 25.

(28) 18. 19.. 20. 21.. 22. 23.. 24. 25.. 26.. 27. 28. 29. 30.. Doehring A, Oertel BG, Sittl R, Lotsch J. Chronic opioid use is associated with increased DNA methylation correlating with increased clinical pain. Pain. 154:1523, 2013 Fillingim RB, Bruehl S, Dworkin RH, Dworkin SF, Loeser JD, Turk DC, WiderstromNoga E, Arnold L, Bennett R, Edwards RR, Freeman R, Gewandter J, Hertz S, Hochberg M, Krane E, Mantyh PW, Markman J, Neogi T, Ohrbach R, Paice JA, Porreca F, Rappaport BA, Smith SM, Smith TJ, Sullivan MD, Verne GN, Wasan AD, Wesselmann U. The ACTTION-American Pain Society Pain Taxonomy (AAPT): an evidence-based and multidimensional approach to classifying chronic pain conditions. J Pain. 15:241-249, 2014 Fitzcharles MA, Cohen SP, Clauw DJ, Littlejohn G, Usui C, Hauser W. Nociplastic pain: towards an understanding of prevalent pain conditions. Lancet. 397:20982110, 2021 Freynhagen R, Baron R, Tolle T, Stemmler E, Gockel U, Stevens M, Maier C. Screening of neuropathic pain components in patients with chronic back pain associated with nerve root compression: a prospective observational pilot study (MIPORT). Curr Med Res Opin. 22:529-537, 2006 Hansen KD, Langmead B, Irizarry RA. BSmooth: from whole genome bisulfite sequencing reads to differentially methylated regions. Genome biology. 13:R83, 2012 Hsu Y-H, Estrada K, Evangelou E, Ackert-Bicknell C, Akesson K, Beck T, Brown SJ, Capellini T, Carbone L, Cauley J, Cheung C-L, Cummings SR, Czerwinski S, Demissie S, Econs M, Evans D, Farber C, Gautvik K, Harris T, Kammerer C, Kemp J, Koller DL, Kung A, Lawlor D, Lee M, Lorentzon M, McGuigan F, Medina-Gomez C, Mitchell B, Newman A, Nielson C, Ohlsson C, Peacock M, Reppe S, Richards JB, Robbins J, Sigurdsson G, Spector TD, Stefansson K, Streeten E, Styrkarsdottir U, Tobias J, Trajanoska K, Uitterlinden A, Vandenput L, Wilson SG, Yerges-Armstrong L, Young M, Zillikens MC, Rivadeneira F, Kiel DP, Karasik D. Meta-Analysis of Genomewide Association Studies Reveals Genetic Variants for Hip Bone Geometry. Journal of Bone and Mineral Research. 34:e3698, 2019 Jacko M, Weyn-Vanhentenryck SM, Smerdon JW, Yan R, Feng H, Williams DJ, Pai J, Xu K, Wichterle H, Zhang C. Rbfox Splicing Factors Promote Neuronal Maturation and Axon Initial Segment Assembly. Neuron. 97:853-868 e856, 2018 Kober KM, Lee MC, Olshen A, Conley YP, Sirota M, Keiser M, Hammer MJ, Abrams G, Schumacher M, Levine JD, Miaskowski C. Differential methylation and expression of genes in the hypoxia-inducible factor 1 signaling pathway are associated with paclitaxel-induced peripheral neuropathy in breast cancer survivors and with preclinical models of chemotherapy-induced neuropathic pain. Mol Pain. 16:1744806920936502, 2020 Koolen DA, Kramer JM, Neveling K, Nillesen WM, Moore-Barton HL, Elmslie FV, Toutain A, Amiel J, Malan V, Tsai AC, Cheung SW, Gilissen C, Verwiel ET, Martens S, Feuth T, Bongers EM, de Vries P, Scheffer H, Vissers LE, de Brouwer AP, Brunner HG, Veltman JA, Schenck A, Yntema HG, de Vries BB. Mutations in the chromatin modifier gene KANSL1 cause the 17q21.31 microdeletion syndrome. Nat Genet. 44:639-641, 2012 Kosek E, Clauw D, Nijs J, Baron R, Gilron I, Harris RE, Mico JA, Rice AS, Sterling M. Chronic nociplastic pain affecting the musculoskeletal system: clinical criteria and grading system. Pain.10.1097/j.pain.0000000000002324, 2021 Krueger F, Andrews SR. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics. 27:1571-1572, 2011 Lawrence M, Huber W, Pagès H, Aboyoun P, Carlson M, Gentleman R, Morgan MT, Carey VJ. Software for Computing and Annotating Genomic Ranges. PLOS Computational Biology. 9:e1003118, 2013 Li C, Lei Y, Tian Y, Xu S, Shen X, Wu H, Bao S, Wang F. The etiological contribution of GABAergic plasticity to the pathogenesis of neuropathic pain. Mol Pain. 15:1744806919847366, 2019. 26.

(29) 31.. 32.. 33. 34. 35.. 36.. 37.. 38. 39. 40.. 41.. 42.. 43. 44. 45. 46.. 47.. Liu J, Zhang MQ, Wu X, Lazarus M, Cherasse Y, Yuan MY, Huang ZL, Li RX. Activation of Parvalbumin Neurons in the Rostro-Dorsal Sector of the Thalamic Reticular Nucleus Promotes Sensitivity to Pain in Mice. Neuroscience. 366:113123, 2017 Livshits G, Malkin I, Freidin MB, Xia Y, Gao F, Wang J, Spector TD, MacGregor A, Bell JT, Williams FMK. Genome-wide methylation analysis of a large population sample shows neurological pathways involvement in chronic widespread musculoskeletal pain. Pain. 158:1053-1062, 2017 Lutz BM, Bekker A, Tao YX. Noncoding RNAs: new players in chronic pain. Anesthesiology. 121:409-417, 2014 Martinez-Salgado C, Benckendorff AG, Chiang LY, Wang R, Milenkovic N, Wetzel C, Hu J, Stucky CL, Parra MG, Mohandas N, Lewin GR. Stomatin and sensory neuron mechanotransduction. J Neurophysiol. 98:3802-3808, 2007 Massart R, Dymov S, Millecamps M, Suderman M, Gregoire S, Koenigs K, Alvarado S, Tajerian M, Stone LS, Szyf M. Overlapping signatures of chronic pain in the DNA methylation landscape of prefrontal cortex and peripheral T cells. Sci Rep. 6:19615, 2016 Montesino-Goicolea S, Sinha P, Huo Z, Rani A, Foster TC, Cruz-Almeida Y. Enrichment of genomic pathways based on differential DNA methylation profiles associated with chronic musculoskeletal pain in older adults: An exploratory study. Mol Pain. 16:1744806920966902, 2020 Nakajima H, Watanabe S, Honjoh K, Okawa A, Matsumoto M, Matsumine A. Expression Analysis of Susceptibility Genes for Ossification of the Posterior Longitudinal Ligament of the Cervical Spine in Human OPLL-related Tissues and a Spinal Hyperostotic Mouse (ttw/ttw). Spine. 45, 2020 Penas C, Navarro X. Epigenetic Modifications Associated to Neuroinflammation and Neuropathic Pain After Neural Trauma. Front Cell Neurosci. 12:158, 2018 Perez MF, Lehner B. Intergenerational and transgenerational epigenetic inheritance in animals. Nature Cell Biology. 21:143-151, 2019 Petitjean H, Pawlowski SA, Fraine SL, Sharif B, Hamad D, Fatima T, Berg J, Brown CM, Jan LY, Ribeiro-da-Silva A, Braz JM, Basbaum AI, Sharif-Naeini R. Dorsal Horn Parvalbumin Neurons Are Gate-Keepers of Touch-Evoked Pain after Nerve Injury. Cell Rep. 13:1246-1257, 2015 Puehler W, Zollner C, Brack A, Shaqura MA, Krause H, Schafer M, Stein C. Rapid upregulation of mu opioid receptor mRNA in dorsal root ganglia in response to peripheral inflammation depends on neuronal conduction. Neuroscience. 129:473479, 2004 Scholz J, Finnerup NB, Attal N, Aziz Q, Baron R, Bennett MI, Benoliel R, Cohen M, Cruccu G, Davis KD, Evers S, First M, Giamberardino MA, Hansson P, Kaasa S, Korwisi B, Kosek E, Lavand'homme P, Nicholas M, Nurmikko T, Perrot S, Raja SN, Rice ASC, Rowbotham MC, Schug S, Simpson DM, Smith BH, Svensson P, Vlaeyen JWS, Wang SJ, Barke A, Rief W, Treede RD, Classification Committee of the Neuropathic Pain Special Interest G. The IASP classification of chronic pain for ICD-11: chronic neuropathic pain. Pain. 160:53-59, 2019 Schou WS, Ashina S, Amin FM, Goadsby PJ, Ashina M. Calcitonin gene-related peptide and pain: a systematic review. J Headache Pain. 18:34, 2017 Schug SA, Lavand'homme P, Barke A, Korwisi B, Rief W, Treede RD, Pain ITftCoC. The IASP classification of chronic pain for ICD-11: chronic postsurgical or posttraumatic pain. Pain. 160:45-52, 2019 Shin SM, Skaar S, Danielson E, Lee SH. Aberrant expression of S-SCAM causes the loss of GABAergic synapses in hippocampal neurons. Sci Rep. 10:83, 2020 Stenz L, Zewdie S, Laforge-Escarra T, Prados J, La Harpe R, Dayer A, PaoloniGiacobino A, Perroud N, Aubry JM. BDNF promoter I methylation correlates between post-mortem human peripheral and brain tissues. Neurosci Res. 91:1-7, 2015 Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, Simonovic M, Doncheva NT, Morris JH, Bork P, Jensen LJ, Mering CV. STRING v11: protein27.

(30) 48. 49.. 50. 51.. 52. 53. 54. 55.. protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 47:D607d613, 2019 Tekieh E, Zaringhalam J, Manaheji H, Maghsoudi N, Alani B, Zardooz H. Increased serum IL-6 level time-dependently regulates hyperalgesia and spinal mu opioid receptor expression during CFA-induced arthritis. EXCLI J. 10:23-33, 2011 Tomassoni-Ardori F, Fulgenzi G, Becker J, Barrick C, Palko ME, Kuhn S, Koparde V, Cam M, Yanpallewar S, Oberdoerffer S, Tessarollo L. Rbfox1 up-regulation impairs BDNF-dependent hippocampal LTP by dysregulating TrkB isoform expression levels. Elife. 8, 2019 Woolf CJ. Central sensitization: implications for the diagnosis and treatment of pain. Pain. 152:S2-15, 2011 Xu CP, Sun HT, Yang YJ, Cui Z, Wang J, Yu B, Wang FZ, Yang QP, Qi Y. ELP2 negatively regulates osteoblastic differentiation impaired by tumor necrosis factor α in MC3T3-E1 cells through STAT3 activation. J Cell Physiol. 234:1807518085, 2019 Yin J, Yang L, Xie Y, Liu Y, Li S, Yang W, Xu B, Ji H, Ding L, Wang K, Li G, Chen L, Hu P. Dkk3 dependent transcriptional regulation controls age related skeletal muscle atrophy. Nat Commun. 9:1752, 2018 Zhou HL, Baraniak AP, Lou H. Role for Fox-1/Fox-2 in mediating the neuronal pathway of calcitonin/calcitonin gene-related peptide alternative RNA processing. Mol Cell Biol. 27:830-841, 2007 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand. 67:361-370, 1983 Zollner C, Shaqura MA, Bopaiah CP, Mousa S, Stein C, Schafer M. Painful inflammation-induced increase in mu-opioid receptor binding and G-protein coupling in primary afferent neurons. Mol Pharmacol. 64:202-210, 2003. Table 1 Clinical characteristics of the participants. 28.

(31) Neuropathic. Nociceptive. Controls. (N=19). (N=18). (N=20). Age. 44 (14). 43 (10). 40 (10). 0.54. Sex F. 4 (21%). 3 (17%). 7 (35%). 0.40. BMI. 27 (3). 27 (5). 25 (3). 0.08. 488 (90-1349). 452 (213-1061). -. 0.62. 19 (100%). 0 (0%). -. <10-6***. Characteristics. Time since injury (days) DN4 ≥ 4. p-value. 1.00. Pain location - Upper limb. 7 (37%). 6 (33%). -. - Lower limb. 7 (37%). 7 (39%). -. - Rachis. 4 (21%). 5 (28%). -. - Multiple pain locations. 1 (5%). 0 (0%). -. 15 (79%). 14 (78%). -. 1.00. BPI pain severity (0-10). 4.5 (2.1). 4.5 (0.1). -. 0.96. HADs Anxiety (0-21). 9.9 (4.4). 10.2 (3.6). -. 0.84. HADs Depression (0-21). 7.4 (3.9). 8.1 (4.1). -. 0.63. Medication (patients taking any medication). The values are expressed as mean (standard deviation), median (range) or number (%); *p≤0.01; ***p<0.001. BMI, body mass index; DN4, Douleur Neuropathique en 4 Questions; BPI, Brief Pain Inventory; HADs, Hospital Anxiety and Depression Scale; TSK, Tampa Scale of Kinesiophobia; PCS, Pain Catastrophizing Scale; CIRS, Cumulative Illness Rating Scale.. 29.

(32)