Abstract 4381: A multi-omic characterization of PARP enzymes in cancer to identify novel monoPARP drug targets

Abstract

Abstract The poly-ADP-ribose polymerases (PARPs) are a family of 17 enzymes with conserved catalytic domains. They regulate a wide variety of important cellular processes including cellular stress signaling pathways implicated in inflammation and cancer. Much of the PARP research has been dedicated to the four polyPARPs (PARP1, 2, 5a, and 5b) which transfer poly-ADP-ribose chains on their target proteins. In particular, the critical role of PARP1/2 in DNA damage response and repair has been studied extensively, leading to effective cancer therapy. However, the majority of PARPs are monoPARPs, which transfer a single ADP-ribose to their target proteins. Recently, several of these family members have emerged in the literature as playing cancer-specific roles. While focused studies of individual monoPARPs are ongoing, a broad integrated in silico survey of the complete PARP family has yet to be done. Thus, we set out to characterize the molecular features of PARPs and their role in human cancer by mining the deep collection of publicly available molecular data from primary cancer, normal tissue samples and cancer cell lines. We designed and executed in silico analyses with the data available in the largest cancer public datasets, The Cancer Genome Atlas (TCGA) and the Cancer Dependency Map (DepMap). We explored standard oncogene hypotheses for all the PARPs, including mutational hotspots, copy-number variations, tumor mRNA overexpression, survival associations to genomic or expression variation, and cancer cell line dependency. Notably, two of the monoPARPs, PARP7 and PARP10, were found to be frequently amplified in multiple cancer types. Further analyses were aimed to identify significant bivariate relationships between PARP molecular features (e.g., expression and methylation) and other cancer-related biomarkers, such as tumor mutation burden or microsatellite instability. For gene expression data, we specifically determined associations between PARP mRNA levels and inferred tumor immune cell types from bulk RNA-sequencing, which implicated five monoPARPs having expression variation strongly associated with the tumor immune cell contexture in ten or more cancers. We also explored patterns of gene co-expression among the PARPs themselves and against the full genome. Our results provide the first pan-cancer in silico characterization of the PARP family, revealing a broad molecular and potential mechanistic diversity among the PARPs across cancer. Notwithstanding the lack of traditional oncogenic features, such as mutational hotspots, in the PARPs, our analyses highlight several monoPARPs with potential oncogenic roles and further support our focus of targeting these in the clinic. Citation Format: Ryan P. Abo, Mario Niepel, Heike Keilhack. A multi-omic characterization of PARP enzymes in cancer to identify novel monoPARP drug targets [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4381.