Abstract

Plain Language Summary Background: Immunotherapy resistance remains a major clinical challenge in advanced microsatellite instability-high (MSI-H) colorectal cancer (CRC), and the underlying gene regulatory networks (GRNs) distinguishing PD-1-resistant from PD-1-sensitive tumors are poorly defined. Methods: We performed single-cell RNA sequencing on tumor samples from three PD-1-resistant and three PD-1-sensitive MSI-H metastatic CRC patients. Resistance-associated transcription factors were identified by integrating machine learning-based prognostic modeling with GRN reconstruction. Bioinformatic findings were functionally validated using dual-luciferase reporter assays, chromatin immunoprecipitation, and in vitro co-culture experiments. Furthermore, the therapeutic efficacy of co-targeting the ETV4-MC1R axis with anti-PD-1 treatment was evaluated in vivo . Results: Several transcriptional regulators were highly active in resistant tumors, among which the transcription factor ETV4 was markedly upregulated in malignant epithelial cells and played a central role. A prognostic model based on ETV4 downstream target genes effectively stratified CRC into subtypes with distinct outcomes. Transcriptome-wide correlation analysis revealed a strong positive association between ETV4 expression and macrophage markers. Mechanistically, we identified MC1R as a critical downstream target of ETV4 mediating resistance. Functional assays confirmed direct binding and activation of the MC1R promoter by ETV4. Co-culture experiments demonstrated that ETV4-high CRC cells promoted macrophage polarization toward an immunosuppressive M2 phenotype via the MC1R pathway. In vivo validation showed that combining ETV4-MC1R targeting with anti-PD-1 therapy significantly reduced tumor volume by 55.6% compared to anti-PD-1 monotherapy ( P < 0.05). Conclusion: This study reconstructs the regulatory net