Phenotypic and transcriptional characterization of myeloid cells in high grade serous ovarian carcinoma

Mononuclear cells are a very heterogeneous population of immune cells, with monocytes and macrophages showing high plasticity in response to different environmental signals. Moreover, dendritic cells subsets appear in different abundances and perform diverse functional roles according to the environment they serve. <br> High grade serous ovarian carcinoma (HGSOC) is one of the most lethal gynecological types of cancer, with low overall and 5-year survival rate. Although it is a very immunogenic neoplasm with high infiltration of lymphocytes and mononuclear cells, currently there is no approved immunotherapy in Europe. <br> In this dissertation, I systematically explored the mononuclear cell compartment of the ascites of HGSOC patients to identify and characterize the tumor associated mononuclear subsets and how disease progression regulates their transcriptional status. Single cell RNA Sequencing revealed 6 main subsets of myeloid cells in the tumor microenvironment of ovarian cancer, with their abundances being regulated by disease progression, with corresponding functional alteration of their immune responses. Findings were validated in a larger patient cohort, with in depth transcriptional characterization on population level, which revealed a type I interferon response in relapse patients. Furthermore, network modeling of tumor associated mononuclear cells led to identification of progression specific gene regulatory modules, with cytokines and immunoregulatory molecules indicated as the most important components. These gene modules were further processed through a computational pipeline for prediction of candidate target genes. A 7-gene signature was defined, which correlates inversely with patients' overall survival of HGSOC, lung squamous cell carcinoma, pancreatic adenocarcinoma and cervical squamous cell carcinoma. Lastly, in an independent study, transcriptional profiling of peritoneal tumor associated macrophages of ovarian cancer mouse model showed that cholesterol metabolism switches TAM polarization towards a protumoral phenotype. <br> Overall, applying cutting-edge transcriptomic approaches in combination with advanced computational and wet lab methods, uncovers the cellular identities of tumor associated mononuclear cells, refines progression specific gene regulatory networks and predicts candidate genes with therapeutic potential.