Abstract Epithelial ovarian cancer (OC) is a heterogeneous disease that stratifies into different histologic subtypes including high-grade serous (HGSOC), clear cell (CCOC), endometrioid (EnOC) and mucinous (MOC) ovarian cancer. HGSOC is the most common, and each histotype is characterized by largely distinct germline genetics, somatic alterations and clinical biomarkers. Recently, whole genome sequencing (WGS) studies have catalogued genome-wide somatic variation for most OC histotypes. These data demonstrate that OC histotypes harbor thousands of noncoding somatic mutations and our next major challenge is to distinguish the few important noncoding somatic drivers from the thousands of passenger mutations. We hypothesized that driver noncoding somatic mutations impact disease development and progression through altering the sequence of regulatory elements (REs), such as enhancers and promoters, resulting in perturbed expression of genes involved in neoplastic development of OC. To systematically address this hypothesis, we established genome-wide H3K27ac epigenomic profiles, annotating active REs for the different ovarian cancer histotypes using chromatin immunoprecipitation sequencing (ChIP-seq) in 20 fresh frozen primary OC tissue samples—five tumors for each major histotype. In parallel, we performed transcriptional profiling using RNA sequencing (RNA-seq). Together, these two datasets enabled us to evaluate epigenetic alterations and the transcriptome. We identified histotype-specific active REs, and common active REs across all histotypes. First, we used the RNA-seq data to assess the effect of the histotype-specific REs in gene expression, and to find target genes of cis-REs and novel histotype-specific biomarkers. Next, we integrated these unique profiles with WGS data from 232 OCs (169 HGSOCs, 35 CCOCs and 28 EnOCs). The number of somatic single nucleotide variants per sample range from 481 to 40,764 (mean=7199, sd = 5751). Of these, 9.4% were noncoding and overlapped active REs in OC. Using a Poisson binomial distribution, we tested the significance of the observed number of mutated samples for any given active RE. We identified several frequently mutated REs in HGSOC, CCOC and EnOC, including promoters (POLR3E and CHCHD6 for HGSOC, CA5BP1 and C16orf87 for EnOC, DKC1 and SLIT3 for CCOC+EnOC), enhancers and super enhancers (KLF6 for HGSOC). We developed a novel approach to identifying target genes associated with a collection of frequently mutated enhancers by correlating enhancer activity (H3K27ac ChIP-Seq signal) vs. gene expression (RNA-seq signal), identifying HOXD cluster genes and C19orf44 as targets of mutated REs in OC. In conclusion, we have used a multi-omics method to identify functional, driver noncoding somatic mutations for ovarian cancer based on their interaction with disease-specific regulatory elements. This represents a powerful way to distinguish important noncoding somatic drivers from a much larger number of passenger mutations that accumulate during tumor development. Citation Format: Rosario I. Corona, Ji-Heui Seo, Dennis J. Hazelett, Xianzhi Lin, Paulette Y. Mhawech-Fauceglia, Jenny Lester, Sohrab P. Shah, David G. Huntsman, Beth Y. Karlan, Alexander Gusev, Benjamin P. Berman, Matthew L. Freedman, Simon A. Gayther, Kate Lawrenson. IDENTIFYING FUNCTIONAL NONCODING SOMATIC MUTATIONS IN OVARIAN CANCER [abstract]. In: Proceedings of the 12th Biennial Ovarian Cancer Research Symposium; Sep 13-15, 2018; Seattle, WA. Philadelphia (PA): AACR; Clin Cancer Res 2019;25(22 Suppl):Abstract nr AP11.
Read full abstract