In silico analysis of cell-cell communication in ovarian cancer single-cell expression data

High-grade serous ovarian cancer is the most frequent histotype of ovar- ian cancer, often diagnosed at late-stage in postmenopausal women. Un- derstanding the role of tumor microenviroment in the progression and drug resistance of this disease is crucial, and cell-cell communication tools based on single-cell RNA sequencing data are an emerging tech- nology to face this challenge. This thesis provides a general overview about the different classes of tools and their biological assumptions, followed by a case study. In particular, the CellChat R package is em- ployed to infer cell-cell communication of four high-grade serous ovarian cancer samples, the first of which derives from the primary tumor site at the time of diagnostic laparoscopic surgery, while the others refer to three metastatic peritoneal sites after the neoadjuvant chemotherapy. By comparing each metastatic sample to the primary one, the major signaling changes of cancer cells were identified. Regarding the primary site, cell-cell communication reveals MIF, the macrophage migration inhibitory factor, being differentially overexpressed in cancer cells and determining a strong autocrine signaling. On the other hand, two of the metastatic sites exhibit strong interactions between collagen and CD44 receptor on cancer cells, which also resulted to be highly expressed by fibroblasts.

A plethora of bioinformatics tools have emerged to decipher cell-cell communications by exploiting the potential of single-cell transcriptomics. In the present thesis project, CellChat is employed to analyze ligand-receptor interactions in 4 single-cell RNA datasets of ovarian cancer.