Pancreatic ductal adenocarcinoma (PDAC) has a unique tumor microenvironment composed of various cell populations such as cancer cells, cancer-associated fibroblasts (CAFs), immune cells, and endothelial cells. Recently, single-cell RNA-sequencing analysis (scRNA-seq) has systemically revealed the genomic profiles of these cell populations in PDAC. However, the direct comparison of cell population composition and genomic profile between primary tumors (at both early- and late-stage) and metastatic tumors of PDAC is still lacking. In this study, we combined and analyzed recent scRNA-seq datasets of transgenic KPC mouse models with autochthonous PDAC and matched liver metastasis, revealing the unique tumor ecosystem and cell composition of liver metastasis in contrast to primary PDAC. Metastatic PDAC tumors harbor distinct cancer cell subpopulations from primary tumors. Several unique markers, including HMGA1, were identified for metastasis-enriched cancer cell subpopulations. Furthermore, metastatic tumors reveal significantly enriched granulocytic myeloid-derived suppressor cells (G-MDSCs), mature neutrophils, and granulocyte-myeloid progenitors (GMPs). A common GMP population across primary tumors, liver metastases, and healthy bone marrow was identified as the putative cell origin of tumor-associated neutrophils/granulocytes.
Keywords: genetically engineered mouse models; liver metastasis; pancreatic cancer; single-cell RNA-sequencing analysis; tumor immunology; tumor microenvironment.