Abstract
Colon cancer (CRC) demonstrates significant heterogeneity, and identifying effective biomarkers can advance the development of precision therapies. Emerging evidence implicates SUMOylation-regulated genes as pivotal regulators of cancer-associated pathways, yet their prognostic potential and therapeutic implications in CRC remain unexplored. A comprehensive analysis of SUMOylation-regulated gene expression, clinical and prognostic value in CRC was performed using transcriptomic data from TCGA-COAD and GEO datasets. We identified 46 differentially expressed SUMOylation-regulated genes (33 upregulated, 13 downregulated) in CRC tumors versus normal tissues. Unsupervised clustering based on 216 SUMOylation-related genes stratified CRC patients into two distinct subtypes: SUMO Cluster 1 (aggressive phenotype, poor prognosis) and SUMO Cluster 2 (favorable prognosis). Cluster 1 exhibited advanced tumor stages (N-stage, p < 0.05) and may present an immunosuppressive microenvironment marked by reduced HLA/immune checkpoint gene expression, while Cluster 2 showed enhanced anti-tumor immunity (activated dendritic cells, γδ T cells). A five-gene SUMOylation-based prognostic signature (MC1R, LRRC4C, SAGE1, GJB6, HOXC5) was developed, and patients were divided into high Riskscore and low Riskscore groups with significant survival differences (log-rank p < 0.05). The nomogram integrating risk score, age, and stage demonstrated robust predictive accuracy (C-index = 0.763, AUC = 0.728-0.785). Nomoscore-high patients exhibited resistance to AMG.706 and ABT.888, suggesting therapeutic vulnerabilities. These findings highlight SUMOylation plays a critical role in CRC heterogeneity, immune modulation, and prognosis, offering a novel biomarker system for risk stratification and personalized therapy.