Comprehensive characterization and effective combinatorial targeting of high-grade serous ovarian cancer via single-cell analysis

Ovarian cancer kills more than 40 000 women in Europe and more than 150 000 women globally each year. High-grade serous ovarian cancer (HGSOC) is the most common and most difficult to treat subtype of the disease. The high-grade serous tumors are highly heterogeneous, therefore, though most of the patients respond well to surgery and chemotherapy initially, more than half experience relapse. HERCULES is a research project funded by the EU H2020 program with the target to characterize comprehensively high grade serous ovarian cancers to find novel therapeutic strategies to fight them. The aim of my PhD thesis was to validate biomarkers identified within HERCULES project in a retrospective case study of patients affected by HGSOC, studying tumor heterogeneity and evaluating the effects of pre-analytical variables, in particular fixation, in the validation process. High grade serous ovarian cancer samples were characterized validating selected biomarkers at both RNA and protein level. Molecular analysis and in situ analysis were performed on multiple tissue biopsies in order to detect spatial heterogeneity and, moreover, biomechanical proprieties of fixed tumor tissues were measured. Lastly, the reliability of molecular analyses on archive tissues were assessed, determining the effect of formalin and Bouin’s fixation at RNA level and evaluating their impact on gene expression using different platforms. Our results showed that detail morphological and immunophenotypical analyses, at the level of the entire tissue slide and not of TMA spots (Tissue micro array), of HGSOC tumors are paramount for the differential diagnosis as well as for both prognostication and therapy. In this view, biomechanical properties, by AFM can support the morphological findings. Among the immunohistochemical markers, Ki67 and BRCA1 have been shown their predictive value for response to first line chemotherapy and overall survival in HGSOC patients. Furthermore, neo adjuvant chemotherapy seems to have a detrimental effect on patient in our cohort. At the RNA level, cyclin C and HLA-B biomarkers showed their prognostic value indicating longer overall survival, while AKTs isoforms have shown a different impact on patients’ outcome. Regarding the pre-analytical variables, fixation confirmed to have a deep impact on molecular analyses, especially in RNA expression. Tissues with highly fragmented RNA such those fixed in Bouin’s can lead to analytical bias in both ddPCR, RT-qPCR, Nanostring and RNAscope technologies. A careful selection of samples with proper nucleic acids quality and integrity is of paramount importance before starting any molecular analysis. Also in that case, to minimize the effect of sample to sample variability a proper sample size should be used. Bouin’s fixed samples because of their high level of nucleic acids fragmentation are not recommended for mRNA expression analyses, especially for low expressed targets. Contrarily, miRNAs, giving their length, are more resistant to fixation procedures and can be used for RNA expression analyses in both formalin and Bouin’s tissues after a proper method of normalization