Validation of gwas-identified variants for anti-tnf drug response in rheumatoid arthritis: a meta-analysis of three large cohorts

Background:The interplay between genetics and drug response in rheumatoid arthritis (RA) has shown that response to biologics varies between individuals and that a large proportion of patients show no clinical improvement (Plenge and Bridges, 2011). Despite the disappointing scenario, to date, only a few genetic markers have been consistently identified and we are far from being able to optimize drug dosing or prioritize drug combinations based on genetic findings.Objectives:With this background, we sought to validate the association of GWAS-identified variants for response to TNF inhibitors (TNFi) in a two-stage case control association study and to shed some light into the functional role of the most interesting markers.Methods:The discovery population consisted of 1361 RA patients ascertained through the REPAIR consortium and the DANBIO registry. RA patients fulfilled the 1987 revised American College of Rheumatology (ACR) and the ACR/EULAR 2010 classification criteria. The validation cohort included 706 Dutch RA patients from the DREAM registry. The study followed the Declaration of Helsinki and study participants gave their written informed consent to participate in the study, which was approved by the ethical review committee of participant institutions. Twenty-seven single-nucleotide polymorphisms (SNPs) were selected through a literature search of relevant GWAS. Linear regression analysis adjusted for age, sex and country of origin was used to determine the association between GWAS-identified SNPs and changes in DAS28 (ΔDAS28) after 3 or 6 months of treatment. The meta-analysis of both populations was performed using a fixed effect model. Correction for multiple testing was performed using the Bonferroni method but also considering the number of inheritance models tested (P=0.0009). To assess the role of the most interesting markers in modulating immune responses, stimulation experiments in whole blood, peripheral mononuclear cells (PBMCs) and monocyte-derived macrophages using a large number of pathogens and microbiome bacteria were performed in 408 subjects from the Human Functional Genomic Project cohort. We also evaluated the correlation of these SNPs with plasmatic levels of 108 inflammatory proteins, 7 serum steroid hormones and counts of 91 blood-derived immune cell populations.Results:The meta-analysis of the discovery cohort and DREAM registry including 2067 RA patients treated with TNFi revealed an overall association of the LINC02549rs7767069 SNP with a decreased drop in DAS28 that remained significant after correction for multiple testing (per-allele ORMeta=0.83, PMeta=0.000077; PHet=0.61). In addition, the meta-analysis of these large cohorts showed that each copy of the LARRC55rs717117G allele significantly decreased the drop in DAS28 in RF-positive patients (per-allele ORMeta=0.67, P=0.00058; PHet=0.06) whereas an opposite but not significant effect was found in RF-negative subjects (per-allele ORMeta=1.38, P=0.10; PInteraction=0.00028; PHet=0.45). Interestingly, the meta-analysis also showed potentially interesting but not statistically significant overall and RF-specific associations for the MAFBrs6071980 and CNTN5rs1813443 SNPs with ΔDAS28 (per-allele ORMeta_rs6071980=0.84, P=0.0059; PHet=0.63 and ORMeta_rs1813443_RF+=0.81, P=0.0059; PHet=0.69 and ORMeta_rs1813443_RF-=1.00, P=0.99; PHet=0.12; PInteraction=0.032). Although analysis of functional data is ongoing, so far, we found that carriers of the LARRC55rs717117G allele showed decreased levels of IL6 after stimulation of PBMCs with Borrelia burgdorferi and Escherichia Coli bacteria (P=0.00046 and 0.00044), which suggested a reduced IL6-mediated anti-inflammatory effect of this marker to worsen the response to TNFi.Conclusion:This study confirmed the influence of the LINC02549 and LARRC55 loci to determine the response to TNFi in RA patients and a weak effect of the MAFB and CNTN5 loci that needs to be further investigated.References:[1]Plenge RM et al 2011. Arthritis Rheum 63, 590-3.ACKNOWLEDGEMENTS:We thank all participants who have agreed to participate in this study. Authors also thank María Dolores Casares, Ángeles Molina, Carmen Oloriz for the collection of Spanish samples and Hans Jurgen Hoffmann, Marianne Thomsen, Vibeke Østergaard Thomsen, Malene Rohr Andersen, Lise Lotte B. Laursen, Helle Jørgensen, Ram Benny Christian Dessau, Niels Steen Krogh, Ulla Vogel, Paal Skytt Andersen, Ivan Brandslund, Steffen Bank, Frederik Trier Møller, Nikolai Toft and Niels Møller Andersen for the participation in collection and purification of Danish samples. We also thank the Danish Departments of Rheumatology for their implication in the collection of clinical data from RA patients included in the DANBIO cohort and the Danish Rheumatologic Biobank. Likewise, we would like to thank Teun van Herwaarden for steroid hormone measurements in serum samples from subjects ascertained through the HFGP initiative.Disclosure of Interests:None declared