Investigating the Role of ARID1A Inactivation in Colon Cancer Pathogenesis

Abstract

The genomics-era search for cancer-causing mutations has revealed epigenomic regulators, in particular SWI/SNF chromatin remodeling complexes, as major targets of mutation. Genes encoding SWI/SNF subunits are collectively mutated in 20% of all human cancers, showing a broad pattern of mutations across solid epithelial, brain, and hematological malignancies. SWI/SNF complexes are evolutionarily conserved regulators of transcription with ATP-dependent chromatin remodeling activity. However, the precise mechanisms by which they function in mammalian cells have been unclear, as have their roles in malignancy. Here, we investigate the tumor suppressor role of ARID1A, the SWI/SNF subunit that is most frequently mutated in human cancer. We demonstrate that inducible ARID1A inactivation in mice drives the formation of invasive colon adenocarcinoma with remarkable resemblance to the corresponding human cancer. Tumor formation does not require cooperating mutations in genes associated with human colon cancer. Tumors also do not show aberrant Wnt signaling, an initiating event in genetic models of colon cancer pathogenesis. Rather, ARID1A inactivation antagonizes tumorigenesis driven by aberrant Wnt signaling. Our investigation reveals that ARID1A targets SWI/SNF complexes to enhancers, where they function in control of enhancer activity. Upon ARID1A inactivation, SWI/SNF binding is lost at the majority of enhancers, which subsequently lose activity, showing reduced levels of H3K27acetylation and downregulation of nearest genes. Residual complexes containing ARID1B preserve SWI/SNF function at a subset of enhancers, but defects in SWI/SNF targeting and control of enhancer activity cause extensive dysregulation of gene expression. These results implicate enhancer-mediated gene regulation as a principal tumor suppressor function of ARID1A in the colon epithelium, with broad relevance to other SWI/SNF-mutant cancers. ARID1B has been identified as a synthetic lethal vulnerability in ARID1A-mutant human cancers; these results suggest that defective SWI/SNF control of enhancer activity drives tumorigenesis and also confers vulnerabilities that might be targeted for therapy. These findings represent an advance in colon cancer modeling, establishing a novel pathway to colon tumorigenesis and a new mouse model that recapitulates features of the human disease – aggressive local tissue invasion, long latency, exclusive origin to the colon rather than the small intestine – that are absent in current, widely utilized models.