Bidirectional Mendelian Randomisation Analysis Provides Evidence for the Causal Involvement of Dysregulation of CXCL9, CCL11 and CASP8 in the Pathogenesis of Ulcerative Colitis

Background and Aims Systemic inflammation is well recognised to be associated with ulcerative colitis [UC], but whether these effects are causal or consequential remains unclear. We aimed to define potential causal relationship of cytokine dysregulation with different tiers of evidence. Methods We first synthesised serum proteomic profiling data from two multicentred observational studies, in which a panel of systemic inflammatory proteins was analysed to examine their associations with UC risk. To further dissect observed associations, we then performed a bidirectional two-sample Mendelian randomisation [TSMR] analysis from both forward and reverse directions using five genome-wide association study [GWAS] summary level data for serum proteomic profiles and the largest GWAS of 28 738 European-ancestry individuals for UC risk. Results Pooled analysis of serum proteomic data identified 14 proteins to be associated with the risk of UC. Forward MR analysis using only cis-acting protein quantitative trait loci [cis-pQTLs] or trans-pQTLs further validated causal associations of two chemokines and the increased risk of UC: C-X-C motif chemokine ligand 9 [CXCL9] [OR 1.45, 95% CI 1.08, 1.95, p = 0.012] and C-C motif chemokine ligand 11 [CCL11] [OR 1.14, 95% CI 1.09, 1.18, p = 3.89 x 10(-10)]. Using both cis- and trans-acting pQTLs, an association of caspase-8 [CASP8] [OR 1.04, 95% CI 1.03, 1.05, p = 7.63 x 10(-19)] was additionally identified. Reverse MR did not find any influence of genetic predisposition to UC on any of these three inflammation proteins. Conclusion Pre-existing elevated levels of CXCL9, CCL11 and CASP8 may play a role in the pathogenesis of UC