Background/Purpose: Genome wide association studies (GWASs) and dense genotyping of immune-related disease regions have identified 17 loci associated with juvenile idiopathic arthritis (JIA) (p<5×10^-8), 11 further loci with suggestive evidence for association (p<1×10^-6), and 2 additional independent loci remaining after stepwise regression in the STAT1/STAT4 region. Given that the majority of these genome-wide associations are located in non-coding regions of the genome, we investigated if specific TFs might bind DNA at multiple JIA-associated genetic loci.

Methods: We defined a locus as those variants that had disequilibrium at r²>0.8 with the most significant genotyped marker at the locus. We assembled a large collection of publically available datasets, including 1,613 chromatin immunoprecipitation with next generation sequencing (ChIP-Seq) datasets, and “Active Chromatin” maps generated from combinations of histone marks in 126 different cell and tissue types. We developed a simulation test to assess the statistical significance of the number of JIA loci that intersect with each ChIP-seq dataset. To generate a null distribution, 2,000 simulations were performed in which each JIA locus was randomly assigned to a genomic location, maintaining linkage disequilibrium and allele frequency structure, and the intersect with each dataset was calculated. Bonferroni corrected P-values (p_c) were estimated for each ChIP-seq dataset evaluated. We confined our attention to TF data sets that bound DNA in ≥3 JIA genetic loci and that had p_c<0.01. The Toppgene web portal was used to characterize functional annotations and protein interactions between TFs.

Results: Our procedure implicates several immune cell types in JIA pathogenesis, including multiple CD4+ cells (Th cells, Th17 cells, Tregs, and memory T cells), CD56+ NKT cells, and CD8+ T cells. 95 of the 1,613 ChIP-seq datasets had p_c<0.01. The top result involves STAT4 binding in IL-12 stimulated CD4+ Th1 cells (p_c<10^-12, 24-fold enrichment). Two Vitamin D receptor (VDR) ChIP-seq datasets are also highly significant (both p_c<10^-9), together comprising 7 of the 30 JIA loci. The 95 significant datasets collectively constitute 47 TFs (also including IRF3, STAT1, and RELA), which are significantly enriched for involvement in pathways that involve cytokine binding and activity (p_c<10^-4).

Conclusion: These data are consistent with common transcriptional control mechanisms operating across multiple JIA risk loci in a shared intracellular environment. Notably, most GWAS loci have small odds ratios (usually <1.2). If there are coordinated mechanisms across loci that alter disease risk with larger effect sizes, then shared gene regulatory mechanisms such as these are important in generating disease risk.

« Back to 2016 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/multiple-genetic-susceptibility-loci-in-juvenile-idiopathic-arthritis-are-bound-by-a-set-of-transcription-factors/