Targeting m6A-SCG2-TAMs axis overcomes 5-FU resistance in colorectal cancer via a multi-omics model.

Chemoresistance to 5-fluorouracil (5-FU) remains a critical barrier in colorectal cancer (CRC) management. This study integrates multi-omics data from 26,192 human samples to elucidate the RNA methylation-mediated crosstalk between tumor-associated macrophages (TAMs) and tumor cells. Machine learning models were able to effectively stratify patients by risk and identified a core signature of six genes (including SCG2), whose expression patterns were associated with poor prognosis and chemotherapy resistance-related phenotypes. Mechanistically, 5-FU elevates m6A modification in TAMs, polarizing them toward the M2 phenotype. SCG2 mRNA methylation promotes TNF-α ubiquitination, reducing its levels and thereby sustaining NF-κB activation in tumor cells to drive PCD resistance. The core candidate genes (CCGs) model effectively predicts survival outcomes. Targeting SCG2 represents a novel strategy to reverse chemoresistance by disrupting TAMs-tumor crosstalk, offering actionable targets for personalized therapy optimization.