Identifying Novel Candidate Biomarkers of Early Epithelial Ovarian Cancer

Ovarian cancer (OC) is the fifth leading cause of cancer deaths in women, and is the most lethal gynaecological cancer worldwide, with the most common subtype being epithelial OC (EOC). The asymptomatic nature and limited understanding of early disease hampers research into early-stage OC and culminates in diagnostic methods only capable of detecting malignant markers of OC. Therefore, there is an urgent need for biomarkers of early-stage OC to be identified.

The first aim of this thesis was to validate a unique mouse model of early OC development. The homozygous Fanconi anemia complementation group D2 knock-out mouse (Fancd2-/-) develops multiple ovarian tumor phenotypes in a sequential manner as they age. Our group previously identified by immunohistochemistry the purported initiating precursor cells, termed ‘sex cords’, that are hypothesised to progress into EOC in this model. To genetically validate this hypothesis, the sex cords, a late-stage tumor phenotype (tubulostromal adenoma), and Fancd2 wild-type (Fancd2+/+) granulosa cell and stromal controls, were isolated by laser-capture microdissection for downstream genetic investigation. Multiplexed gene expression analyses of ~ 90 selected genes were performed using the Genome Lab GeXP Genetic Analysis System (GeXP). Principal component analyses and unbiased hierarchical clustering of the resultant gene expression data determined that the gene signature profiles within sex cords and tubulostromal adenoma samples indeed clustered together genetically confirming the identity of a precursor lesion in this mouse model. Furthermore, key epithelial OC markers included in the GeXP assay (Pax8, Cdh1 and Muc16) were expressed in both the sex cords and tubulostromal adenoma demonstrating neoplastic transformation was already occurring in the precursor lesion.

Despite confirmation of the identity and developmental origin of the precursor lesion (sex cords), an in-depth understanding of the model was limited due to restricted numbers of genes able to be investigated using the GeXP assay, thus unique markers of the precursor lesion were not identified. Therefore, the second aim of this thesis was to comprehensively investigate the genetic signature of the Fancd2-/- model via total and small RNA sequencing of micro-dissected cells of the sex cords, tubulostromal adenomas and the equivalent Fancd2+/+ controls. Sequencing results of matched messenger (m) RNAs and micro (mi) RNA expression in micro-dissected tissue of the phenotypes of interest provided novel information about the development of early EOC and unique transcriptional markers, namely upregulated Brinp2, Slc6a7, Smpd3, Grin2b, Grik4, AW55198, Lrp11 and Mro mRNAs and downregulated miR-122/b-5p, miR-129-5p, miR-138-5p, miR-465a/b/-3p, miR-465c-3p/5p, miR-500-3p and miR-129b-3p miRNAs of the sex cord phenotype through differential gene expression analysis.

Whilst tissue markers of EOC are relevant as biomarkers of EOC, the location of the ovaries within the peritoneum prevents their diagnostic implementation without surgical intervention. To circumvent this limitation, liquid biopsy has become paramount for OC diagnosis. However, the detection of early biomarkers of EOC in humans is ultimately hampered by their low systemic concentration and the presence of circulating factors from all other organ systems. Therefore, to bypass detection limitations within systemic circulation, the next aim of this thesis was to leverage the established Fancd2-/- model of early EOC to provide proof of concept that secreted miRNA encapsulated into extracellular vesicles (EV) secreted from the precursor lesion (i.e., sex cords) could be detected in conditioned media from ovary explant culture. Performing RNA sequencing from EV-derived small RNAs isolated from conditioned culture medium resulted in the identification of five common upregulated secreted EV miRNAs in Fancd2-/- mouse ovaries that had sex cords or tubulostromal adenomas present, compared to their age-matched Fancd2+/+ counterparts. Additionally, the small RNA sequencing data of EVs was compared to the sequencing data of the micro-dissected sex cords and tubulostromal adenomas and these identical five miRNAs were also significantly upregulated as compared to the control group. This suggests that these five secreted EV miRNAs (miR-200a-3p, miR-200b-3p, miR-200c-3p, miR-429-3p and miR-375-3p) originate from the sex cord and tubulostromal adenoma cells and therefore provided proof-of-concept that small transformational changes such as the development of a precursor lesion can be detected within secreted EVs.

The final aim of this thesis was to determine whether the methodologies adopted for early candidate biomarker discovery in a murine model can be applied to human biopsy specimens and subsequently the relevance of the Fancd2-/- model for identifying novel biomarkers for human EOC. Two serous subtypes of human EOC, early low-grade serous borderline tumours (SBT; as an early phenotype) and high-grade serous ovarian cancer (HGSOC; as a late malignant phenotype) were selected for study. SBT and HGSOC specimens were prepared for matched micro-dissected neoplastic epithelial cells and explant cultures. Total and small RNA sequencing of micro-dissected tissue and small RNA sequencing of EVs from explant cultures were performed as per the mouse studies. This sequencing data resulted in the identification of SBT tissue-specific mRNA and miRNA markers as well as SBT-secreted miRNA markers. Remarkably, comparing the human and mouse sequencing data revealed an overlap in differentially expressed genes between the SBT, HGSOC, sex cords, and tubulostromal adenomas, pertaining to key epithelial markers of EOC (Muc16, Cdh1, Krt7, Krt19, Epcam) and the five secreted EV miRNAs firstly described in the murine studies. This provides strong evidence that these novel candidate biomarkers are relevant to early human EOC.

Overall, the results from this thesis confirmed the relevance of the Fancd2 model for human OC investigation and identified novel candidate biomarkers of early EOC that may be applied to both the mouse model and human ovarian cancer.