Molecular and Genetic Features Underlying Treatment Response in Relapsed Ovarian Cancer

Abstract

Most women with advanced epithelial ovarian cancers (EOC) will relapse. Currently, there are no predictive tests to guide clinicians when treating relapsed EOC. The aim of the work described in this thesis was to identify genomic markers that predict response to treatments commonly used in relapsed EOC: platinum chemotherapy, pegylated liposomal doxorubicin (PLD) and endocrine therapy. We identified women treated with platinum, PLD or endocrine therapy, performed molecular tests on primary tumour and associated these with response. In platinum-treated patients, we confirmed that homologous recombination (HR) deficiency, either HR pathway gene mutations or an overall ‘HRD score’, was associated with higher response rates. CCNE1 gain/amplification was associated with non-response to platinum (p=0.03). In PLD treated patients,BRCA mutations were associated with increased response (p=0.04). We examined HGSOC expression subtypes (C1, C2, C4, C5) in relation to platinum and PLD response, and although no significant associations were found there was an over-representation of the C4 subtype in platinum non-responders and an under-representation of the C2 subtype in PLD responders. Using RNA-Seq, we identified 41 differentially expressed genes (DEGs)between PLD-sensitive and -resistant, newly established HGSOC cell lines; 4 of these genes were also differentially expressed in tumours from PLD complete responders and non-responders. In patients treated with endocrine therapy, increasing ERa histoscores in HGSOC tumours may be associated with response (p=0.07). Using RNA-Seq, we identified oestrogen (E2)-regulated genes in ER+ HGSOC cell lines with differential E2growth response. We identified 27 DEGs between HGSOC tumours from endocrine therapy responders and non-responders, and found that most of these genes were E2-regulated. This E2-regulated signature may identify patients with ER+ HGSOC that respond to endocrine therapy. Taken together, we found that molecular markers in primary tumour could potentially predict treatment outcome in relapsed EOC. Our data provide a rationale for testing primary tumour for specific predictive molecular markers. If validated, these predictive markers could be used to personalise treatment of relapsed EOC.