Abstract
Background: Mismatch repair-proficient (pMMR) microsatellite stability (MSS) in colorectal cancer (CRC) indicates an unfavorable therapeutic response to immunotherapy with immune checkpoint inhibitors (ICIs). However, the molecular characteristics of CRC patients with pMMR MSS remain largely unknown. Methods: Heterogeneities between mismatch repair-deficient (dMMR) microsatellite instability (MSI) and pMMR MSS CRC patients were investigated at the single-cell level. Next, an MSS-related risk score was constructed by single-sample gene set enrichment analysis (ssGSEA). The differences in immune and functional characteristics between the high- and low-score groups were systematically analyzed. Results: Based on the single-cell RNA (scRNA) atlas, an MSS-specific cancer cell subpopulation was identified. By taking the intersection of the significant differentially expressed genes (DEGs) between different cancer cell subtypes of the single-cell training and validation cohorts, 29 MSS-specific cancer cell marker genes were screened out for the construction of the MSS-related risk score. This risk score signature could efficiently separate pMMR MSS CRC patients into two subtypes with significantly different immune characteristics. The interactions among the different cell types were stronger in the MSS group than in the MSI group, especially for the outgoing signals of the cancer cells. In addition, functional differences between the high- and low-score groups were preliminarily investigated. Conclusion: In this study, we constructed an effective risk model to classify pMMR MSS CRC patients into two completely different groups based on the specific genes identified by single-cell analysis to identify potential CRC patients sensitive to immunotherapy and screen effective synergistic targets.