Colorectal cancer (CRC) is highly metastatic, yet the interaction between tumor cells and microenvironment components remains unclear, as does its impact on patient outcomes. Single-cell data (GSE225857 and GSE166555) were collected for transformed and primary CRC to identify differing cell types and their subpopulations. Prognostic signature genes for CRC were identified through Cox analysis of tumor cell subpopulations and cancer-associated fibroblasts (CAFs) interaction genes, leading to the development of a prognostic model. In the single-cell dataset of metastatic CRC patients in GSE225857 and nonmetastatic CRC patients in GSE166555, eight tumor cell subpopulations were identified, in which T0 was significantly enriched in metabolism-associated pathways. The CAFs subpopulation CAF0 and T0 had extensive cell communication and were approachable in three-dimensional space. A prognostic signature predicting CRC patient survival was developed and validated based on these signature genes of CAF0 cells. This prognostic signature serves as an independent and effective factor for prognosis. In this study, we identified characteristic cell subpopulations in metastatic and primary CRC. Based on the reciprocal genes between them, we constructed a prognostic model for CRC. Our findings provide a scientific basis for understanding the metastatic mechanisms of CRC.
Tan et al. (Thu,) studied this question.