Postdoctoral Fellow, National Taiwan University Hospital
Pancreatic cancer (PC)-induced cachexia is a paraneoplastic syndrome featuring refractory weight loss and skeletal muscle wasting. It is highly prevalent and account for poor prognosis of PC patients. Cachexia-inducing PC is characterized by an inflammatory tumor microenvironment (TME), harboring heterogeneous cell types. However, their identities and cell crosstalk have not been determined. To determine the specific cell populations involved in the PC-induced cachexia, we collected 6 primary PC tumor samples, four with and two without cachexia, for droplet-based single-cell RNA sequencing (scSeq). We obtained a total of 42592 cells consisting of 28267 cells (66.4%) from the cachexia-inducing tumor, and 14325 cells (33.6%) from the non-cachexia counterpart. With cell clustering and annotation, we identified heterogeneous TME cells, including acinar cells, ductal cells, endothelial cells, cancer-associated fibroblasts (CAFs), immune cells, islet cells (majorly β cells) and malignant cells. After subclustering malignant cells, immune cells, CAFs, endothelial cells into diverse cells subsets, we identified 14 unique subsets enriched in cachexia-inducing PC tumor versus non-cachectic counterparts. We further leveraged spatial transcriptomics integrated with RCTD deconvolution to define the spatial distribution of these cell subsets in cachexia-inducing PC tumor. Our spatial images demonstrated interactions among 6 unique subsets, including non-classical monocytes, 3 malignant cell clones, 1 unique CAF, and endothelial subset. The spatial results were consistent to cell-cell interaction analysis with scSeq data, where the ligand-receptor pairs among these subsets were identified. Combined with enriched function and pathway analysis, we clarified how these TME cells coordinates cachexigenic functions through cell-cell interaction.
