Bioinformatic-guided functional determination of predictive markers for PARPi in ovarian cancer

Introduction Several clinical trials have shown promising anti-cancer activity of PARP inhibitors (PARPi) in ovarian cancer (OC) patients with superior benefit in BRCA-mutated and homologous recombination deficiency (HRD)-enriched tumors. However, “true” molecular predictors are scarce. Thus, we functionally test previously reported genetic aberrations of PARPi in a semi-throughput manner using a panel of OC cell lines.

Material and methods Large CRISPR-Cas9 library screens and the Broad Institute Cancer Cell Line Encyclopedia (CCLE) were analysed to rank literature identified genes on ‘cell fitness’ and initially test profiled PARPi in OC cell lines, respectively. Results on PARPi and platinum-based drugs were verified by MTT assays. A lentiviral negative selection CRISPR-Cas9 screen was established to identify PARPi-related ‘essential’ genes using a flow cytometry-based assay.

Results In line with the CCLE drug response data, selected OC cell lines displayed a wide spectrum of PARPi sensitivity. Interestingly, platinum-resistant OC cells showed cross-resistance to Olaparib and Niraparib in a cell-dependent manner without a correlation to genetic aberrations (e.g. BRCA2 mutation in PARPi resistant KURAMOCHI cell line). Our established competition assay showed that mutations in DNA-repair genes RPA3 and ERCC3 negatively impact ‘cell fitness’ while TP53 and AAVS1 were marginally affected.

Conclusion Response to PARPi is enhanced in carboplatin- and taxol-sensitive OC cells. In contrast, response to PARPi in carboplatin and taxol-resistant cells was genetic signature- and cell line-independent. Thus, we established a novel screening tool to rapidly and functionally test candidate genes. Results obtained herein will be further validated in patient derived material.

Publication History

Article published online:
07 October 2020