Abstract 4661: Deciphering the causes of the COSMIC mutational signature 17 by combining pan-cancer data with experimental mouse models

Gastric and esophageal adenocarcinomas (G/EAC) exhibit a characteristic pattern of somatic variants, known as signature 17. Signature 17 associates with high mutation burden, elevated neoantigen levels and a potential sensitivity to cell cycle inhibitors. Signature 17 correlates with poor survival, and is increased in cancer patients with a history of the common pre-malignant lesion called Barrett’s esophagus. We recently showed that oxidative damage during DNA replication is a likely source of signature 17; however this hypothesis remains to be validated. Intriguingly, we observe signature 17 arising in clonally expanded primary mouse cells (Hupki MEF) and in mouse models of liver cancer. Here, we aim to identify the mechanisms underlying signature 17 by mining TCGA and ICGC data, as well as generating genome-wide molecular profiles of relevant experimental cell systems. Mutational signatures were extracted from exome and whole-genome sequencing (WGS) data of over 460 G/EAC tumors. Pathway analysis of RNA-seq data from matching samples was performed using NIH DAVID. Exposure to signature 17 was next assessed with regards to clinico-pathologic data. Lastly, we compared the asymmetry of signature 17 mutations around replication origins and early/late replicating regions in WGS data from MEF cells and G/EAC cancers. Comparing the transcriptomes of signature 17-high and -low tumors implicated cell proliferation, regulation of apoptosis, DNA damage response and repair, oxidoreductase activity and gastric acid secretion as relevant pathways. We found positive associations between signature 17 intensity and Barrett’s esophagus history, poorer survival, and microsatellite instability. MEF WGS data revealed signature 17 enrichment in late-replicating genomic regions and on the lagging strand, analogous to observations in human tumors. Importantly, in all cultured MEF clones as well as in most EAC genomes, signature 17 co-occurred with signature 18 which is likely linked to oxidative damage of DNA, supporting the hypothesis that inflammation-driven oxidative DNA damage leads to these human cancers. The replication strand and timing asymmetries were significant in both signature 17 and signature 18 in EAC genomes, while the Hupki MEF clones exhibited significant strand asymmetry in signature 17 but not in signature 18, and showed significant correlation with replication timing in signature 18. Collectively, our results elucidate new features of a modifiable mutagenic activity underlying signature 17 in cancers of the esophagus, stomach, and in model systems. We have also identified novel clinical and molecular associations with signature 17. Investigations using the reported innovative cell-based models will help to identify the precise mechanistic basis of signature 17, with potential implications for primary and secondary cancer prevention and personalized therapy. Citation Format: Marketa Tomkova, Claire Renard, Lara Urban, Souad Kolli, Maude Ardin, Manuraj Pandey, Maria Zhivagui, Hana Huskova, Magali Olivier, Hiroyuki Marusawa, Benjamin Schuster-Bockler, Jiri Zavadil. Deciphering the causes of the COSMIC mutational signature 17 by combining pan-cancer data with experimental mouse models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4661.