Identifying the mechanisms driving pancreatic ductal adenocarcinoma stem cell characteristics using single-cell RNA-sequencing

Pancreatic ductal adenocarcinoma (PDAC) is among the deadliest human malignancies. Surgery, the only curative treatment, is precluded by the late stage at diagnosis in 80% of cases. Recurrence after surgery is common, and the disease does not respond well to chemotherapy and radiotherapy. Resistance to therapy and recurrence are thought to be driven by pancreatic cancer stem cells (PCSCs), a subset of cells with self-renewal and differentiation capacities. Annihilating these cells is therefore of paramount importance for treating PDAC. Identifying these cells, however, has proven challenging. Currently, there is no gene signature able to identify PCSCs. In this thesis, I employed single-cell RNA-sequencing to integrate multiple approaches towards the identification of PCSCs (experimentally-derived markers, bioinformatics-based gene sets, and computational tools to infer developmental potential from expression data), to uncover the genes and processes characterizing PCSCs, and to assess the effects of I-BRD9, an inhibitor of BRD9, a bromodomain-containing protein involved in chromatin remodelling upon the PCSCs, using two single-cell RNA-sequencing PDAC datasets. The results evidenced cell cycle abnormalities as crucial to cancer stemness in PDAC, with multiple lines of evidence converging towards the identification of clusters whose markers significantly overlapped with cell cycle-related stemness-associated gene sets as PCSCs. Traditionally-derived PCSC markers were found to be largely of low reliability. A transitional cell population distinct from both PCSCs and the bulk of the cells, and one with an advanced stage of differentiation which however regained partial stemness-like characteristics, were identified as highly drug-resistant, suggesting that greater than previously believed PDAC cell heterogeneity, not merely PCSCs, is involved in chemoresistance. I-BRD9 achieves a ~6-fold reduction of PCSCs in one dataset, likely mediated by the demonstrated downregulation of key G2/M DNA replication checkpoint-linked genes such as TOP2A and CDK1, but the effects are partially reversed by the addition of PDAC drug gemcitabine