Abstract
Genome-wide association studies (GWASs) have provided many insights into cancer genetics. However, the molecular mechanisms of many susceptibility SNPs defined by GWASs in cancer heritability and in promoting cancer risk remain elusive. New research strategies, including functional evaluations, are warranted to systematically explore truly causal genetic variants. In this study, we developed an integrative functional genomics methodology to identify cancer susceptibility SNPs in transcription factor-binding sites across the whole genome. Employing integration of functional genomic data from c-Myc cistromics, 1000 Genomes, and the TRANSFAC matrix, we successfully annotated 12 SNPs present in the c-Myc cistrome with properties consistent with modulating c-Myc binding affinity in hepatocellular carcinoma (HCC). After genotyping these 12 SNPs in 1,806 HBV-related HCC case subjects and 1,708 control subjects, we identified a HCC susceptibility SNP, rs157224G>T, in Chinese populations (T allele: odds ratio = 1.64, 95% confidence interval = 1.32-2.02; p = 5.2 × 10(-6)). This polymorphism leads to HCC predisposition through modifying c-Myc-mediated transcriptional regulation of EPB41, with the risk rs157224T allele showing significantly decreased gene expression. Based on cell proliferation, wound healing, and transwell assays as well as the mouse xenograft model, we identify EPB41 as a HCC susceptibility gene in vitro and in vivo. Consistent with this notion, we note that EPB41 expression is significantly decreased in HCC tissue specimens, especially in portal vein metastasis or intrahepatic metastasis, compared to normal tissues. Our results highlight the involvement of regulatory genetic variants in HCC and provide pathogenic insights of this malignancy via a genome-wide approach.