In silico and functional analysis of genomic aberrations associated with pancreatic and ovarian cancers

Abstract

High grade serous (HGS) ovarian cancer and pancreatic ductal adenocarcinoma (PDAC) are poor prognosis neoplasms that are characterised by high levels of genomic instability, DNA damage repair (DDR) defects, and drug resistance which ultimately results in poor survival rates. This study aims to use markers of genomic instability, in the form of copy number variations (CNVs) and gene expression changes, to identify DDR gene aberrations which contribute to poor patient prognosis and drug resistance in HGS ovarian cancer and PDAC, and which may be exploited for therapeutic benefit. Bioinformatic analyses of HGS ovarian cancer patient datasets (TCGA, Genome Institute of Singapore) were used to determine whether gene pairs co-amplified together, where one of the partners is a DDR gene, associated with poor patient prognosis. RECQL, a DNA helicase, was found to be co-amplified with 6 genes from the 3q12 locus, CPOX, MINA, TMEM45A, TOMM70A, CLDND1 and PTPLB, and all of these co-amplifications were found to significantly associate with reduced PFS in ovarian cancer patients. Functional validation, using siRNA to inhibit gene expression, was carried out to assess the role of these genes in regulating survival. Using a cell line (PEO4) model of chemotherapy resistant HGS ovarian cancer, no induction of apoptosis or changes in long term survival was observed upon silencing of any of the genes. In comparison, gene silencing of RECQL and PTPLB was found to induce an apoptotic response in PDAC cell lines (Panc1 and MiaPaCa2). Long term clonogenic assays also revealed that TMEM45A, TOMM70A and CLDND1 are regulators of Panc-1 cell survival. Together this indicates that in PDAC, the survival advantages of RECQL co-amplification with the 3q12 chromosomal region may be related to the effects of more than one gene. However, the PTPLB:RECQL co-amplification may be a driver of PDAC cell survival. Bioinformatic analyses were also carried out on HGS ovarian cancer and PDAC patient datasets (TGCA and ICGC) as well as cell line datasets (Sanger Institute) aimed at determining single genes within DDR pathways whose aberrations in copy number and gene expression impacts upon patient survival or drug response to cisplatin or gemcitabine. 6 genes, TSTA3, RECQL4, ESRP1, NBN, SUMO3 and EP300, were identified that affected one of the above parameters, of which, EP300 functionally validated as a potential therapeutic target. SiRNA-mediated silencing of this was found to induce apoptosis in cell line models of HGS ovarian cancer (SKOV3) and PDAC (Panc-1). In addition, loss of EP300 increased the apoptotic response of SKOV3 cells to treatment with cisplatin, gemcitabine, doxorubicin, paclitaxel and the DNA-PKcs inhibitor NU7441. In Panc1 cells, only response to gemcitabine and paclitaxel was significantly increased with EP300 loss. Mechanistic studies to define how EP300 regulates SKOV3 cell survival found that EP300 gene silencing induced dysregulated DNA damage recognition, and cell cycle arrest at the G2/M phase. RPPA proteomic analysis identified Wee1, Plk1, cyclin B, MAPK1 and histone 3 as potential mediators of the cell cycle arrest observed. Many changes to apoptotic signalling were also observed after loss of EP300, although this counter-intuitively, increased the levels of anti-apoptotic proteins, such as Bcl-xl, Bcl-2 and pS136 Bad. Also, pro-survival Akt activity, determined as phosphorylation at S473, was reduced by EP300 gene silencing. This indicates a novel link between p300 and Akt-mediated cell survival of HGS ovarian cancer. Overall, this study has combined in silico analyses of patient and cell line datasets with functional in vitro validation to identify RECQL and EP300 as potential therapeutic targets for HGS ovarian cancer and PDAC. Future work to validate these in additional cell line models, including those with defined genomic backgrounds, will further progress these as novel therapeutic strategies.