The University of Manchester Manchester, United Kingdom
Disclosure: Disclosure information not submitted.
Chuan Fu Yap1, Paul Martin2, Darren Plant1, John Bowes1, Kazuyoshi Ishigaki3, Saori Sakaue4, Alex Macgregor5, Suzanne Verstappen1, Anne Barton1, Soumya Raychaudhuri6 and Sebastien Viatte1, 1The University of Manchester, Manchester, United Kingdom, 2The University of Manchester, Oberhaching, Germany, 3Riken, Bunkyo-ku, Tokyo, Japan, 4Broad Institute of Harvard and MIT, Cambridge, MA, 5The University of East Anglia, Norwich, United Kingdom, 6Brigham and Women's Hospital, Boston, MA
Background/Purpose: Rheumatoid arthritis (RA) genetic susceptibility has been well studied with five amino acid positions within the HLA explaining most of the association. Although genome wide association studies have also identified risk loci outside the HLA, no association with genetic polymorphisms within the T-cell receptor (TCR) regions has ever been reported below genome-wide significance in large studies. TCR polymorphisms have only an effect in the context of specific HLA alleles. We therefore decided to look for interaction effects between the TCR and HLA region using genotype data from the United Kingdom population, in a first instance.
Methods: RA patients recruited to the Biologics in Rheumatoid Arthritis Genetics and Genomics Study Syndicate (BRAGGSS) and to The Norfolk Arthritis Register (NOAR) were used as cases and individuals recruited to the Understanding Society Study as controls. Genotype data were generated using Illumina HumanCoreExome Arrays Standard quality control (QC) was applied prior to full genotype and HLA imputation on the Michigan Imputation Server. Following imputation low minor allele frequency (0.01) markers were removed. TCR polymorphisms were extracted from imputed genotype data for the alpha, beta, gamma and delta region by selecting all single nucleotide polymorphisms (SNP) within the earliest start position and the last end position for those genes. The interaction effects between TCR and 4-digit HLA alleles were studied using standard logistic regression, whilst the omnibus test was applied for interaction effects between TCR and amino acid positions; sex and the first ten principal components were used as covariates to adjust for the potential effect of sex and population stratification. Permutations were performed 100,000 times for the top amino acid interaction.
Results: In total, 3943 RA patients and 9191 healthy controls were analysed. Following QC, 5133 TCR SNPs, 119 HLA alleles and 339 amino acid positions were available for analysis. The strongest interaction effect for RA susceptibility was found between rs1076861 and HLA-A*26:01 (OR = 2.52, 95% CI = 1.64 – 3.89; p-value = 2.6e-5) as well as between rs7150263 and HLA-A amino acid position 9 (p-value = 3.9e-8). For HLA-A position 9, phenylalanine gives the strongest interaction association with rs7150263 (0.78, 95% CI = 0.72 – 0.85; p-value = 2.4e-8), followed by tyrosine (OR = 1.26, 95% CI = 1.13 – 1.41; p-value = 4.23e-5), threonine (OR = 1.32, 95% CI 1.13 – 1.53; p-value = 3.5e-4), and serine (OR = 1.05, 95% CI 0.9 – 1.2; p-value = 0.04). rs1076861 is an intergenic SNP located between TRDV3 and TRAJ61, and rs7150263 is also an intergenic SNP located between TRAV7 and TRAV8. Amino acid position 9 of HLA-A is located within the peptide binding groove.
Conclusion: This study has identified statistical interactions between TCR polymorphisms and HLA-A amino acids. Replication will be performed within a large transethnic dataset including > 200,000 samples (RA international genetics consortium).
Disclosures: C. Yap, None; P. Martin, None; D. Plant, None; J. Bowes, None; K. Ishigaki, None; S. Sakaue, None; A. Macgregor, None; S. Verstappen, None; A. Barton, Galapagos, Bristol-Myers Squibb(BMS), Roche Chugai; S. Raychaudhuri, Mestag, Inc, Rheos Medicines, Janssen, Pfizer, Biogen; S. Viatte, Bristol-Myers Squibb(BMS).
