Abstract
Recurrent MSI events are tumor type specific and show a distinct distribution compared with SNVs.
Major finding: Recurrent MSI events are tumor type specific and show a distinct distribution compared with SNVs.
Concept: The genome-wide localization and functional consequences of MSI in cancer genomes were profiled.
Impact: This analysis may identify driver genes that are functionally significant in tumorigenesis.
Inactivation of DNA mismatch repair (MMR) genes prevents efficient repair of replication errors introduced by DNA slippage at microsatellite tandem repeats and is associated with microsatellite instability (MSI) and increased mutation incidence. MSI is frequently observed in a subset of colorectal and endometrial cancers, but the distribution and functional consequences of MSI in cancer genomes have not been comprehensively characterized and it is unclear whether the genes targeted by MSI in these tumor types are similar. To determine the genome-wide MSI landscape, Kim and colleagues performed analysis of exome and whole-genome sequencing data of colorectal and endometrial cancer samples, including 57 microsatellite-unstable tumors. MSI events were largely correlated with silencing of the MMR gene MutL homolog 1 (MLH1), underscoring its role in inducing MSI, and were enriched in noncoding sequences, in particular 3′ UTRs. Recurrent MSI events in coding regions were characterized by an increased frequency of frameshift mutations compared with in-frame changes in microsatellite length and were accompanied by decreased transcript expression of affected genes, suggesting that inactivation of these genes may promote tumorigenesis. Genes frequently targeted by recurrent frameshift MSI events also exhibited significant tumor type specificity, including colorectal cancer–specific MSI in TGFβ receptor 2 (TGFBR2) and endometrial cancer–specific MSI in Janus kinase 1 (JAK1). Furthermore, comparison of the distribution of MSI with single-nucleotide variants (SNV) by whole-genome sequencing identified overrepresentation of MSI in intronic sequences and regions associated with histone marks of open chromatin, as well as depletion of MSI around nucleosome-bound DNA, in contrast with somatic SNVs, which have been shown to be enriched in heterochromatic and intergenic regions. These findings provide insight into the association of MSI with genomic and epigenomic features and may define functionally relevant tumor type–specific driver genes.