Massively parallel analysis of genotype-dependent enhancer activity among atopic dermatitis genetic risk variants

Background Atopic dermatitis (AD) is an inflammatory, pruritic disease of the skin with a complex etiology involving environmental and genetic factors. Numerous genetic risk loci for AD have been nominated through genome-wide association studies, with most associated variants residing in non-coding regions. However, further work is needed to understand how genetic variation contributes to disease-related alterations to gene expression. Objective Here, we use massively parallel reporter assays (MPRAs) to identify genetic risk variants with genotype-dependent regulatory activity within a set of 2,381 variants distributed among 49 independent atopic dermatitis risk loci. Methods We employed our MPRA library in HaCaT skin keratinocytes, TE-7 esophageal epithelial cells, and Jurkat T cells, enabling the identification of shared and cell-type specific variants with genotype-dependent enhancer activity. Stimulation of HaCaT and TE-7 cells with the cytokine IL-13 revealed 14 variants with stimulation-dependent allelic expression patterns, elucidating a likely role for genotypic variation in Th2-mediated inflammatory signaling in AD. Results In total, our approach discovered 96 allelic enhancer variants among all investigated cellular contexts, representing 32 independent AD risk loci. We discover a possible mechanistic role for AP-1 family transcription factors, as revealed by motif enrichment, genotype-dependent binding, and enhanced ChIP-seq signal at AD risk variants with allelic enhancer activity compared to variants with non-allelic enhancer activity. We assign allelic enhancer variants to probable target genes, including previously identified risk genes such as KIF3A and FLG. Conclusion Collectively, our systematic, genome-scale approach implicates causal genotype-dependent gene regulatory mechanisms for the majority of AD risk loci, providing a unique resource for the discovery of the genetic mechanisms underlying this chronic inflammatory disease. Overall design: We performed ChIP-seq of the transcription factors FOSL2 and JUNB in human HaCaT skin keratinocytes. Experiments were performed in three replicates for each factor.