Tumor-immune interactions that delineate therapy responses in ovarian cancer

High-grade serous ovarian cancer (HGSC) is an aggressive disease with platinum-based chemotherapy and surgery still being the cornerstones of the treatment. Despite successes in other cancer types, immune checkpoint therapies (ICB) have failed to improve clinical outcomes in unselected HGSC patient populations, and no predictive biomarkers have been validated to date. A deeper understanding of the tumor microenvironment (TME) and its dynamics during treatment pressure is required in order to develop new patient stratification strategies and targets for combinatorial treatments. This thesis utilized highly multiplexed tissue cyclic immunofluorescence (t-CycIF) as well as single-cell and bulk transcriptomics (scRNAseq, RNAseq), GeoMX spatial transcriptomics and ex vivo patient derived immunocompetent cultures (iPDCs) to map, at single-cell resolution, genotype and therapy conditioned tumor-immune-stroma interactions in HGSC. Study I profiled BRCA1/2 mutated versus homologous recombination proficient (HRwt) tumors. Study II analyzed paired chemo-naïve and post-neoadjuvant chemotherapy (NACT) samples from prospective clinical cohorts. Spatial analyses quantified direct cell-cell contacts, cellular microcommunities as well as interactions within these. Our results revealed distinct interplay of tumor, immune and stromal cells in BRCA1/2 mutated as compared to HRwt tumors. BRCA1/2 mutated tumors showed evidence of increased immunesurveillance against highly proliferating epithelial tumor cells, which was further associated with better patient prognosis. We additionally observed chemotherapy-induced remodeling of the tumor microenvironment with reduced fraction of proliferating epithelial tumor cells and increased influx of immune cells. The cytotoxic T-cell responses were however hampered by myeloid cell interactions at the tumor-stromal interface as well as in so called “myelonets” defined as networks of interconnected myeloid cells. We identified enhanced activity of the TIGIT-NECTIN2 signaling axis after NACT. Importantly, targeting TIGIT in patient derived iPDCs showed responses that correlated with source tumor characteristics, namely TIGIT and NECTIN2 expression. Together, the findings contribute to a better understanding of the dynamic, genotype and therapy conditioned immune ecology in HGSC and nominate a biomarker-guided therapeutic window after NACT. Tumors characterized by CD8+T-cell residency within myeloid networks and at the tumor stromal interface and a heightened TIGIT-NECTIN2 signaling are rational candidates for anti-TIGIT combination therapies.