Identifying Novel Non-Coding Variants Regulating a Gene Causally Associated With Seronegative Arthritis in Adults and Children

Abstract

Autoimmune diseases are a family of conditions in which an individual’s immune system misrecognizes self-tissues as foreign, resulting in inflammatory attack that leads to tissue injury. One common manifestation of human autoimmunity is inflammatory arthritis. This family of diseases includes rheumatoid arthritis (RA) and juvenile idiopathic arthritis (JIA). Both RA and JIA include a subset of seronegative arthritis (lacking rheumatoid factor) that exhibits considerable genetic similarity across the age spectrum but has received less attention that the numerically more abundant seropositive subset. Genome wide association studies (GWAS) conducted in these conditions implicate not only the human leukocyte antigen (HLA) region and genes implicated across the autoimmune arthritides but also a single gene, ANKRD55 (Ankyrin Repeat Domain 55), that is associated with seronegative RA and JIA but not seropositive arthritis. In silico data suggest that this gene is expressed in T cells, including T regulatory cells, but its function is unknown. Through analysis of GWAS data, we find that the region around ANKRD55 contains no common coding variants, strongly suggesting that regulatory non-coding variation mediates the effect of this locus on disease risk. We sought to identify associated causal variants and the pathways they engage. Analyzing qPCR and Western blot data in sorted cell subsets, we confirmed that ANKRD55 is expressed in CD4 naïve T cell population. Analyzing publicly-available sequence data, we identified 6 common non-coding single nucleotide polymorphisms (SNPs) in linkage disequilibrium r^2>0.8 around the tagging GWAS SNP at ANKRD55. We then employed electrophoretic mobility shift and luciferase assays to screen these non-coding variants, identifying three SNPs that exhibited both allele-specific binding to protein in Jurkat cell nuclear extract and that modulated gene expression in transfected Jurkat cells. Employ these candidate functional SNPs (fSNPs) as “bait”, we then employed DNA pulldown methods combined with a bioinformatic prediction algorithm to identify GR-beta and TFII-I as lead candidate proteins modulating allele-specific disease transcription of ANKRD55. Together, these studies identify new non-coding variants that regulating a gene expressed in CD4+ T cells and is causally associated with seronegative arthritis in adults and children.