Most risk variants for autoimmune diseases affect CD4+ T cell gene regulation. Additionally, human activated CD4+ T cells express the major risk genes: HLA class II genes. Hence, understanding the genetic regulatory effects acting at CD4+ T cell states is critical to define autoimmune mechanisms. Since allele specific expression is largely driven by cis genetic regulatory variation, it can be exploited to define regulatory effects in many cell states without the need of a large sample size. We hypothesized that studying allelic expression at multiple time-points upon T cell activation, may identify genes with context-dependent cis regulatory variation.
We queried CD4+ memory T cells with RNA-seq after non-antigenic stimulation at 8 time-points over 72 hours in 24 individuals. We looked for cell-state specific cis regulatory effects by identifying dynamic allelic expression in 191 genes (5% FDR), 36 of which are in autoimmune disease loci (e.g. UBASH3A, IL10). We discovered a dynamic cis regulatory effect for a major autoimmune disease gene: HLA-DQB1. Using a novel HLA-personalized genome pipeline, we found that HLA-DQB1 allelic profiles can be classified into three dynamic cis regulatory programs. We have shown the late activation allelic effects translate to the protein level with flow cytometry. Fine-mapping putative regulatory variants, and functional validations with EMSA and CRISPR/Cas9 nucleotide conversion uncovered a causal SNP (P=0.0003). This SNP drives expression changes in T cells but not B cells or macrophages. Our results define the dynamics of regulatory variation in multiple T cell states and offer potential for understanding autoimmune mechanisms.