Perturbation of the insomniaWDR90GWAS locus pinpoints rs3752495 as a causal variant influencing distal expression of neighboring gene,PIG-Q

Abstract

ABSTRACTAlthough genome wide association studies (GWAS) have been crucial for the identification of loci associated with sleep traits and disorders, the method itself does not directly uncover the underlying causal variants and corresponding effector genes. The overwhelming majority of such variants reside in non-coding regions and are therefore presumed to impact the activity ofcis-regulatory elements, such as enhancers. Our previously reported ‘variant-to-gene mapping’ effort in human induced pluripotent stem cell (iPSC)-derived neural progenitor cells (NPCs), combined with validation in bothDrosophilaand zebrafish, implicatedPIG-Qas a functionally relevant gene at the insomnia ‘WDR90’ locus. However, importantly that effort did not characterize the corresponding underlying causal variant at this GWAS signal. Specifically, our genome-wide ATAC-seq and high-resolution promoter-focused Capture C datasets generated in this cell setting brought our attention to a shortlist of three tightly neighboring single nucleotide polymorphisms (SNPs) in strong linkage disequilibrium in a candidate intronic enhancer region ofWDR90that contacted the openPIG-Qpromoter. The objective of this study was to investigate the influence of the proxy SNPs collectively and then individually onPIG-Qmodulation and to pinpoint the causal “regulatory” variant among the three SNPs. Starting at a gross level perturbation, deletion of the entire region harboring all three SNPs in human iPSC-derived neural progenitor cells via CRISPR-Cas9 editing and subsequent RNA sequencing revealed expression changes in specificPIG-Qtranscripts. Results from more refined individual luciferase reporter assays for each of the three SNPs in iPSCs revealed that the intronic region with the rs3752495 risk allele induced a ∼2.5-fold increase in luciferase expression (n=10). Importantly, rs3752495 also exhibited an allele specific effect, with the risk allele increasing the luciferase expression by ∼2-fold compared to the non-risk allele. In conclusion, our variant-to-function approach and subsequentin vitrovalidation implicates rs3752495 as a causal insomnia risk variant embedded at theWDR90-PIG-Qlocus.