Single-cell RNA sequencing reveals spatiotemporal heterogeneity and malignant progression in pancreatic neuroendocrine tumor

Aims: Using Single-cell RNA sequencing (scRNA-seq), we explored the spatiotemporal heterogeneity of pancreatic neuroendocrine tumors (pNETs) and the underlying mechanism for malignant progression. Methods: scRNA-seq was conducted on three tumor tissues (two primary tissues from different sites, one liver metastatic lesion), one normal liver tissue, and peripheral blood mononuclear cells from one patient with a metastatic G2 pNET, followed by bioinformatics analysis and validation in a pNETs cohort. Results: The transcriptome data of 24.544 cells were obtained. We identified subpopulations of functional heterogeneity within malignant cells, immune cells, and fibroblasts. There were intra-and inter-heterogeneities of cell subpopulations for malignant cells, macrophages, T cells, and fibroblasts among all tumor sites. Cell trajectory analysis revealed several hallmarks of carcinogenesis, including the hypoxia pathway, metabolism reprogramming, and aggressive proliferation, which were activated at different stages of tumor progression. Evolutionary analysis based on mitochondrial mutations defined two dominant clones with metastatic capacity. Finally, we developed a gene signature (PCSK1 and SMOC1) defining the metastatic potential of the tumor and its prognostic value was validated in a cohort of thirty G1/G2 patients underwent surgical resection. Conclusions: Our scRNA-seq analysis revealed intra-and intertumor heterogeneities in cell populations, transcriptional states, and intercellular communications among primary and metastatic lesions of pNETs. The single-cell level characterization of the spatiotemporal dynamics of malignant cell progression provided new insights into the search for potential novel prognostic biomarkers of pNETs.

Automated summary

Utilizing single-cell RNA sequencing technology, this study uncovered both intra- and inter-tumor heterogeneities in pancreatic neuroendocrine tumors, identifying activation of cancer progression hallmarks at different stages, defining two dominant clones with metastatic potential, and developing a gene signature for predicting tumor metastatic potential validated in patient cohorts.