Abstract 242: Functional Characterization of a Cis-eQTL Locus for Plasma Cholesterol Using CRISPR/Cas Genome Editing in Human Pluripotent Stem Cells

Abstract

Genome-wide association studies (GWASs) have discovered many novel genetic loci linked to serum lipid levels, yet pinpointing the causal variants remains a major challenge. Expression quantitative trait locus (eQTL) analysis of liver and adipose indicates that many lipid-associated variants influence gene expression in a cis-regulatory manner. To identify causal variants at 57 lipid-associated eQTL loci, we performed a massively parallel reporter assay (MPRA) in which the genomic region surrounding each candidate SNP was coupled to a reporter gene with a unique barcode identifier. This construct pool was transfected into 3T3-L1 adipocytes, and the copy number of each expressed barcode determined by RNAseq. Variants were prioritized according to allele-specific regulatory activity. The top-ranked variant, rs2277862, is associated with total cholesterol and is flanked by the ERGIC3, CPNE1, and CEP250 genes, none of which have previous connections to lipid metabolism. We hypothesized that rs2277862 is causal for the eQTL at the 20q11 locus, and that it lies within a transcriptional regulatory site to influence local gene expression. We identified human pluripotent stem cell (hPSC) lines with different genotypes at rs2277862 and, using CRISPR/Cas genome-editing technology, either deleted the putative regulatory site encompassing rs2277862 or knocked in the alternate SNP allele. Analysis of undifferentiated rs2277862 homozygous major deletion mutants revealed diminished expression of ERGIC3, CPNE1, and CEP250; conversely, heterozygous deletion mutants had increased expression of the same three genes. Our results suggest that in hPSCs the major and minor alleles of rs2277862 have opposing effects on gene expression. In ongoing work, we are validating these findings in differentiated adipocytes and seek to identify the factor(s) that binds the putative regulatory site to mediate allele-specific gene expression. We anticipate that this work will offer fresh insight into the mechanisms by which GWAS-implicated SNPs influence expression of causal genes for lipid metabolism and, more broadly, establish a novel methodology for finding causal variants for any phenotype of interest.