Comprehensive transcriptomic analysis of high-grade serous tubo-ovarian carcinoma using single-cell RNA-sequencing.

High-grade serous tubo-ovarian cancer (HGSTOC) affects worldwide 239,000 women each year. HGSTOC is typically characterised by a high recurrence rate with poor long-term survival as it often becomes resistant to most treatment options — a phenomenon that has been attributed to its extensive inter- and intratumoural heterogeneity. Beside the very pronounced patterns of chromosomal instability in HGSTOC cancer cells, various other cellular phenotypes involved in immune activation, hypoxia and extracellular matrix remodelling determine a tumour microenvironment that favours disease progression and metastases. The last decade targeted therapy approaches against several cellular components of the tumour ecosystem have tried to overcome the shortcomings of standard therapeutic regimens but response rates are variable across patients, implying the importance of finding the (stromal) subpopulations that make the tumour more or less susceptible for it. In this thesis project, we applied single-cell RNA-sequencing (scRNA-seq) to provide an comprehensive view on the heterogeneous populations of stromal and tumour cells present in HGSTOC in first-line setting, revealing small populations of cells whose transcriptional programs were overlooked using bulk RNA-sequencing. In a first part, we established a high-resolution overview of the tumour microenvironment of HGSTOC, providing 43 new potential targets for therapy and identifying 6 cellular subtypes of prognostic relevance (chapter 3). Amongst these prognostic cell types, we found a pivotal role for the fibroblast subtypes and the transforming growth factor β-pathway. We also noticed a distinct distribution of these phenotypes across the previously established HGSTOC molecular subtypes. After correction for the 6 prognostic cell types, none of the molecular subtypes remained prognostic, confirming that our stromal phenotypes explain the association of the molecular subtypes with outcome. Evaluation of the unique and shared heterogeneity between ovarian cancer and other cancers, including lung, colorectal and breast cancer (chapter 4), allowed us to observe a strong enrichment of fibroblasts, in particular cancer-associated fibroblasts, and cells belonging to the myeloid population with immunosuppressive properties in HGSTOC, as compared to other cancers. Cellular subpopulation involved in the adaptive immune system (T cells and B cells) were scarce in ovarian cancer. In a second part, we leveraged scRNA-seq to map the dynamics of the tumour microenvironment underlying early response to standard frontline chemotherapy (paclitaxel-carboplatin) and bevacizumab, an anti-angiogenic agent, in HGSTOC (chapter 5). The addition of bevacizumab to standard chemotherapeutic treatment expedited vessel normalisation and immunostimulatory reprogramming, by reducing the number of tumour tip cells and hypoxia, as well as, the reduction of regulatory T cells and the stimulation of dendritic cell maturation. However, bevacizumab exposure was also associated with increased influx of PDGFC-expressing tumour-associated macrophages with the ability to establish cross-family receptor interactions with the VEGFR and bypass VEGFA-dependent tumour angiogenesis, providing a novel mechanism of acquired resistance to bevacizumab in HGSTOC. Although these scRNAseq studies comprise only a dozen of patients, we here provide initial evidence of novel biomarkers predicting prognosis and therapy response in HGSOC, to form the basis for future experiments and guide new trial design, expediting the development of more effective anti-cancer drugs. Only at this unprecedented resolution, we will be able to detangle the heterogeneity of the tumour microenvironment in each cancer type and, in extension, in each individual patient, to eventually evolve to more personalised and highly accurate cancer treatment.

Jaar van publicatie:2021

Toegankelijkheid:Embargoed