Background:Over the past decade, genome wide association studies (GWAS) have identified the JAZF1 locus as a risk locus for several autoimmune diseases, including rheumatoid arthritis (RA)1. However, the exact causal variants in the JAZF1 locus and their underlying regulatory events contributing to RA are still not known. Here, we focus on the effect of these variants on gene expression in synovial fibroblasts (SF).Objectives:To characterize the functional consequences of RA-causal variants in the JAZF1 locus in SF.Methods:Genetic fine-mapping of RA loci was conducted by computing sets of credible variants driving GWAS signals. These credible variant sets were integrated with DNA architecture (ChIP-seq), 3D chromatin interactions (3C, HiC and capture HiC), DNA accessibility (ATAC-seq) and gene expression (RNA-seq and CAGE-seq) datasets to select putative RA-causal variants in SF. Selected variants in the JAZF1 locus were tested for regulatory function by luciferase reporter assays and electrophoretic mobility shift assays (EMSA) in the fibrosarcoma cell line HT1080. The JASPAR2020 database was used to identify putative transcription factors (TF) binding to the selected variants. The expression of HOTTIP was measured by quantitative PCR in hand SF (n=23). Genotyping was done by pyrosequencing.Results:Genetic fine mapping revealed 47 variants in the JAZF1 locus. Integration of these variants with the chromatin datasets prioritized rs2158624, rs57585717 and rs186735625 as the top candidates (posterior probability of causality >0.1) in the JAZF1 locus. We found that rs2158624 and rs186735625 are located in the vicinity of enhancer elements in SF as determined by ATAC-seq. In addition, the region of rs2158624 exhibited strong chromatin interactions with the genomic region of HOTTIP and HOXA13. Both these transcripts were previously shown to be specifically expressed in SF isolated from hands and feet2. Based on this, we selected rs2158624 as the most promising candidate in the JAZF1 locus. We found that the rs2158624-C allele (risk) is associated with lower expression of HOTTIP, but not HOXA13, in hand SF compared to the rs2158624-T allele (non-risk) (p=0.02). Luciferase assays in HT1080 cells demonstrated enhancer activity with both the rs2158624-C allele (p=0.006) and T allele (p=0.04), with no significant difference in enhancer activity between the rs2158624-C and T allele. EMSAs identified stronger specific binding of HT1080-cell nuclear extract for the rs2158624-T allele than for the C allele (risk). Based on the JASPAR2020 database, we identified NFAT5 as a potential TF that can bind to rs2158624 and may regulate the expression of HOTTIP.Conclusion:We were able to substantially narrow down the potential functional variants in the JAZF1 locus using our data integration approach and functional assays. We suggest that the risk allele of rs2158624 influences the binding of TFs controlling the expression of the long non-coding RNA HOTTIP in SF, which might confer specific risk to develop RA in hands.References:[1]Okada Y et al. Genetic of rheumatoid arthritis contributes to biology and drug discovery. Nature 2014;506:376.[2]Frank-Bertoncelj M et al. Epigenetically-driven anatomical diversity of synovial fibroblasts guides joint-specific fibroblast functions. Nat Commun 2017;8:14852.Disclosure of Interests:Miranda Houtman: None declared, Xiangyu Ge: None declared, Amanda McGovern: None declared, Kerstin Klein: None declared, Gisela Orozco: None declared, Mojca Frank Bertoncelj: None declared, Miriam Marks: None declared, Oliver Distler Speakers bureau: Bayer, Boehringer Ingelheim, iQone, Medscape, MSD, Novartis, Pfizer and Roche, Consultant of: Abbvie, Acceleron Pharma, Amgen, AnaMar, Arxx Therapeutics, Bayer, Baecon Discovery, Boehringer, CSL Behring, ChemomAb, Corbus Pharmaceuticals, Galapagos NV, GSK, Glenmark Pharmaceuticals, Horizon Pharmaceuticals, Inventiva, Italfarmaco, iQvia, Kymera, Lilly, Medac, Medscape, Mitsubishi Tanabe Pharma, MSD, Pfizer, Roche, Roivant Sciences, Sanofi and UCB, Grant/research support from: Kymera Therapeutics and Mitsubishi Tanabe, Paul Martin: None declared, Stephen Eyre: None declared, Caroline Ospelt: None declared
An expanded analysis framework for multivariate GWAS connects inflammatory biomarkers to functional variants and disease