OP12. Multiple functional variants at the 3p21 locus contribute to ulcerative colitis: Results from a European consortium
I. Cleynen1, M. Artieda2, H. Verspaget3, M. Szczypiorska2, F. Seibold4, P. Lakatos5, T. Ahmad6, R. Weersma7, J. Mahachie8, I. Arijs1, S. Müller9, K. Parnell10, A. Tordai11, D.W. Hommes12, K. Van Steen8, C. Wijmenga13, P. Rutgeerts14, D. Lottaz9, S. Vermeire14
1Catholic University Leuven, Gastroenterology, Leuven, Belgium; 2Progenika Biopharma, S.A., Derio, Spain; 3Leiden University Medical Center, Department of Gastroenterology & Hepatology, Leiden, Netherlands; 4Spitalnetz Bern, Switzerland; 5Semmelweis University, 1st Department of Medicine, Budapest, Hungary; 6Royal Devon & Exeter Hospital, Gastroenterology RM B205, Exeter, United Kingdom; 7University Medical Center Groningen, Gastroenterology and Hepatology, Groningen, Netherlands; 8University of Liège, Belgium; 9University of Bern, Bern, Switzerland; 10Peninsula Medical School, United Kingdom; 11Hungarian National Blood Transfusion Service, Molecular Diagnostics, Hungary; 12UCLA, Division of Digestive Diseases, Los Angeles, United States; 13University Medical Center Groningen, Groningen, Netherlands; 14University Hospital Gasthuisberg, Department of Gastroenterology, Leuven, Belgium
Background: Intestinal proteases are involved in host defense to counteract bacterial colonization, and affect mucosal barrier integrity. In a recent systematic review of protease (inhibitor) genes in inflammatory bowel disease (IBD), the 5 top-ranked proteases in ulcerative colitis (UC) were located on 3p21: APEH, MST1, DAG1, USP4, and USP19. MST1 has repeatedly been associated with UC. We aimed to unravel the genetic make-up of this region in relation with UC.
Methods: 16 haplotype tagging SNPs (tSNPs) in these 5 genes were genotyped in 2112 UC patients and 1796 healthy controls (HC) (exploration: 721 UC and 542 HC; validation: 1391 UC and 1254 HC). The USP4-MST1 gene region was imputed using the 1000 genomes dataset as a reference (BEAGLE v3.2). For 23 patients, gene expression data from colonic biopsies was available. Statistical analysis was performed using SVS v7.5.2 (crude association analysis, additive genetic model), and plink v1.0.7 (meta-analysis, imputed dataset, set-based analysis, and eQTL analysis). A corrected p < 0.05 was considered statistically significant.
Results: 3 tSNPs were significantly associated with UC in the meta-analysis: rs11130213, a conserved transcription factor binding site (APEH, p = 4.93e−4, OR = 1.19); rs9822268 (APEH, p = 6.26e−4, OR = 1.19); and rs6766131 (DAG1, p = 8.23e−4, OR = 1.19). 58 out of 150 imputed markers were significantly associated with UC, 51 of which were highly correlated with rs11130213 (r2 > 0.8). Set-based analysis revealed 4 independent signals (r2 < 0.8), with the peak signals located around USP4 (rs34343820, rs17650792) and BSN, a gene positioned between DAG1 and APEH (rs9827708, rs12497288). eQTL analysis showed no significant impact on expression of USP4, MST1, APEH or DAG1. Among the 58 markers are multiple functional variants: 1 non-synonymous coding (rs34762726, BSN), and 5 synonymous coding SNPs of which rs1801143 is a putative eQTL for USP4 (eqtl.uchicago.edu). Two SNPs are located in the 3'-UTR and 1 in the 5'-UTR.
Conclusions: Multiple functional variants at 3p21 were associated with UC. The main genes implicated trough this study were USP4, a deubiquitination enzyme that plays a role in the regulation of quality control in the ER; and BSN, a presynaptic cytomatrix protein, essential for regulating neurotransmitter release. These genes warrant further follow-up trough functional studies in UC. Complementing genetic studies with functional studies will be indispensable to unravel the exact role for this region in UC pathogenesis.