Functional characterisation of driver events in ovarian clear cell carcinoma

Abstract

Ovarian clear cell carcinoma (OCCC) is a distinct subtype of epithelial ovarian cancer (EOC) characterised by glycogen accumulation. Frequently arising within endometriotic cysts, it can be conceived as an ectopic endometrial cancer. Putative driver genomic events include HNF1B overexpression and inactivating ARID1A mutations. Importantly, these can also be found in a significant proportion of adjacent non-malignant endometriotic lesions and, therefore, are likely early events in OCCC pathogenesis. I hypothesised that the study of the functional consequences of these driver genomic events and metabolic perturbations would provide insights into potential therapeutic targets in this difficult to treat cancer. Gene expression arrays in normal mouse uterus, embryonic fibroblasts and human immortalised ovarian surface epithelium cells revealed that a core ARID1A-driven transcriptional programme, conserved across normal tissues and species, centred on regulation of mitosis and cell cycle progression-related genes, and involving potentially targetable kinases, exists. Despite this, the effect of ARID1A knockdown on proliferation in human cell lines and mouse cells and tissues was found to be context and tissue specific. Interestingly, in vivo knockout in the uterine epithelium of Arid1afl/fl mice was accompanied by a dramatic increase in proliferation, in support of its suggested driver-event role in uterinederived cancers. HNF1B overexpression has been previously reported to affect proliferation and metabolism in a variety of cell lines but studies in well characterised OCCC cells are lacking. Here, HNF1B was found to consistently drive proliferation in a panel of bona fide OCCC cell lines. Pathway analysis of HNF1B-regulated genes suggested that HNF1B is involved in interactions with the tumour microenvironment. Indeed, I observed that HNF1B negatively regulates migration and invasion. Additionally, I found that HNF1B overexpression drives glycogen accumulation and that its knockdown reverses the Warburg effect. These results point at trade-offs among proliferation, metabolism and metastatic capability and suggest that HNF1B overexpression may be one of the reasons that, in marked contrast to high-grade serous EOC, OCCC frequently presents as early stage disease. Little is known about the functional consequences of glycogen accumulation in OCCC. I report that OCCC cell lines display increased expression of glycogen metabolism enzymes and that inhibiting the rate limiting phosphorylase (PYGL) and synthase enzymes markedly decreased proliferation, even in the presence of plentiful extracellular glucose. This observation suggests a role for glycogen beyond that of a glucose store. Assays performed to elucidate how PYGL knockdown affects proliferation suggest that this may be through G2/M phase arrest, possibly caused by inhibition of lipid breakdown or altered PKA signalling. Furthermore, preliminary evidence suggests that the effects of PYGL knockdown on proliferation are limited to malignant cells only. x In conclusion, this project studied the functional consequences of three driver events in OCCC: ARID1A mutations, HNF1B overexpression and glycogen accumulation. Targeting ARID1Aregulated kinases and glycogen metabolism and perturbing HNF1B function require further investigation as potential therapeutic strategies in OCCC.