Development and evaluation of methodologies for analysis of CTC and ctDNA in patients with ovarian carcinoma

About 70% of patients with ovarian cancer (OC) are diagnosed at an advanced stage (III-IV), with associated poor prognosis, even after chemotherapeutic interventions (mostly platinum taxane-based), leading to poor survival rates. The current biomarkers available in the clinic are not enough for efficient prognostication and surveillance of OC. Hence, more accurate and reliable biomarkers of patient disease status are urgently required. Currently, histopathological evaluation of tumour biopsies is the gold standard in clinical practice with the purpose of evaluating specific biomarkers to predict therapeutic response. However, this is invasive, and associated complications may occur. Liquid biopsy is a minimally invasive test and has the advantage of allowing real-time monitoring of treatment response and comprehensive serial/repetitive phenotypic and genotypic profiling of the primary, metastatic and recurrent tumours. Circulating tumour cells (CTC) and circulating tumour DNA (ctDNA) represent a new generation of biomarkers that can be used for predicting therapy responsiveness and longitudinal monitoring of OC patients undergoing therapy. This thesis describes a series of investigations in OC which include, a methodological study to improve phenotypic characterization of detected CTCs, genomic validation of putative CTCs, and the evaluation of the clinical validity of ctDNA as biomarker of response to neoadjuvant chemotherapy (NACT). This thesis consists of six chapters: a general introduction to OC (Chapter 1); a thorough review of the literature regarding CTCs and ctDNA in OC (Chapter 2); three results chapters (Chapters 3 - 5); and a final chapter that comprises of a general discussion of the main findings, the limitations of the study and future directions (Chapter 6). The first chapter (Chapter 1) of the thesis includes a review of the literature, which consist of the characterisation and clinical landscape of OC, while briefly introducing the topic of liquid biopsy. This is then followed by a comprehensive review (Chapter 2) that highlights the progress in the field of liquid biopsy for OC from 2011 to 2019, focusing specifically on CTCs and ctDNA as potential biomarkers for the management of the disease. Based on the limitations identified in this review (in Chapter 2), a multi-marker antibody staining protocol was developed to target epithelial (cytokeratin (CK) and EpCAM), mesenchymal (vimentin), and OC specific (PAX8) markers for detection of CTCs (Chapter 3). CD45/16 and CD31 were used for the exclusion of white blood and vascular endothelial cells, respectively. PD-L1 was used for CTC characterisation. Optimisation of the CTC markers were done using OC cell lines, SKOV-3 and OVCA432. CTCs were enriched using the ParsortixTM (Angle plc.) system, and further detected using the multifluorescent markers from 16 OC patients. Patients were found to have heterogeneous CTCs, with 113/157 (72%) cells positive for CK/EpCAM (epithelial marker), 58/157 (37%) positive for vimentin (mesenchymal marker), and 17/157 (11%) positive for both (hybrid). Positivity for PD-L1 was significantly associated with the hybrid phenotype when compared with the epithelial (p = 0.0009) and mesenchymal (p = 0.007) expressing CTCs. In Chapter 4, we performed copy number alteration (CNA) profiling on putative CTCs that were previously identified in Chapter 3. CNA were detected in both CK/EpCAM and vimentin positive CTCs. However, a proportion of cells expressing CK/EpCAM, PD-L1 and CD31 were found not to carry CNAs. Further characterisation of CTCs in combination with clinical outcomes is needed to provide insight into the role that epithelial to mesenchymal plasticity and vasculogenic mimicry plays in OC and its relationship with PD-L1 expression. In addition, ctDNA was analysed in 58 patients with high grade serous ovarian cancer (HGSOC) enrolled in a neoadjuvant chemotherapy (NACT) plus immune checkpoint inhibitor clinical trial (iPRIME, ACTRN12618000109202) (Chapter 5). Cell free DNA (cfDNA) was sequenced using the Oncomine™ Breast cfDNA Research Assay V2, and selected mutations were validated using droplet digital PCR (ddPCR). For ctDNA, the frequencies of genes affected by non-synonymous somatic mutations included TP53 (67%), KRAS (11%), PIK3CA (2%) and SF3B1 (2%). A total of 41 gene variants were identified in either pre- or post-NACT samples; 36 (87.8%) of them had variants in TP53 alone, and 3 (7.4%) in KRAS and 1 (2.4%) each for PIK3CA and SF3B1. Undetectable levels of ctDNA post, but not prior to NACT was associated with chemotherapy response score 3 (CRS3) (p = 0.04) when compared with CRS1/2. Comparison of mutant copies per mL of plasma quantified by NGS and ddPCR demonstrated a strong correlation (Pearson’s r = 0.929; p \u3c 0.0001) between the two platforms. Longitudinal monitoring of ctDNA specific mutant gene variants using ddPCR, showed a predictive lead time, and anticipates disease progression earlier (7-12 weeks) compared to monitoring patients by analysis of serum levels of CA-125. These results suggest that ctDNA could be used as a biomarker to determine patients’ clinical outcomes and serve as a potential endpoint for clinical trials. Finally, Chapter 6 provides a general discussion of the studies covered in this thesis, highlighting potential contributions to the growing field of liquid biopsy and its application for the clinical management of OC patients. The limitations inherent to both CTCs and ctDNA analysis, and suggestions for improvements before its successful implementation in routine clinical practice, are comprehensively discussed