An exploratory prognostic risk model for colon cancer developed from methylation-associated genes identified in circulating tumor cells

Colon cancer (CRC) prognosis is closely associated with metastasis, with circulating tumor cells (CTCs) displaying DNA methylation abnormalities linked to metastatic behavior. This study aims to develop a risk model for CRC by utilizing methylation-associated genes identified from CTCs to predict prognosis and identify candidate compounds for pharmacologic investigation. RNA-Seq data, clinical information, and DNA methylation levels were sourced from GEO and TCGA databases. Differential gene expression analysis and intersections led to the identification of methylation-driven differentially expressed genes (MD-DEGs). Prognosis-related MD-DEGs (PMD-DEGs) were identified and employed to construct a CRC risk model, validated through risk assessments, survival analyses, receiver operating characteristic (ROC) curve evaluations, independent prognostic analyses, and an external cohort (GSE39582). Differences in drug sensitivity between high- and low-risk patients were also assessed. RT-qPCR, western blotting, and immunohistochemical analyses validated the PMD-DEGs. A total of 69 MD-DEGs were identified in the CRC cohort, with a risk model constructed incorporating five PMD-DEGs: IDUA, ELF5, ZBTB18, NOVA1, and TCEAL3. In the training cohort, the model stratified CRC patients by prognosis, although external validation revealed only limited discriminative capability. Computational drug-sensitivity analysis revealed predicted associations between these five PMD-DEGs and the sensitivity to 13 anticancer compounds. Additionally, RT-qPCR, western blotting, and immunohistochemistry confirmed significant differences in theexpression of these genes between CRC tissues and para-cancer normal (PCN) samples, and highlightedcandidate compounds warranting future experimental investigation.