Identification of a signature gene set for oxaliplatin sensitivity prediction in colorectal cancer.

Background: Colorectal cancer (CRC) is a leading cause of cancer-related mortality worldwide. Oxaliplatin-based chemotherapy is a cornerstone of treatment for many CRC patients; however, the development of chemoresistance severely limits its therapeutic efficacy and remains a major clinical challenge. The identification of robust biomarkers to predict oxaliplatin sensitivity is therefore critical for personalizing treatment and improving patient outcomes. Methods: Here, we conducted a comprehensive analysis of large-scale genomic datasets, including from the Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA). Machine learning algorithms to these datasets was applied to identify genes associated with oxaliplatin response. The prognostic value of the candidate genes was evaluated using progression-free survival (PFS) analysis. The predictive robustness of the identified gene set was further validated in external datasets from GEO and the Genomics of Drug Sensitivity in Cancer (GDSC) database using multivariate logistic regression analysis. Finally, we experimentally assessed the functional role of these genes by examining their expression in oxaliplatin-resistant cell lines and by performing gene knockdown experiments in colorectal cancer cells to measure subsequent changes in oxaliplatin sensitivity. Results: Our integrated bioinformatics approach identified 14 genes potentially linked to oxaliplatin sensitivity. Subsequent PFS analysis narrowed this set to four key genes (AXDND1, BAMBI, MAPK8IP2, and BMP7) that were significantly associated with patient survival following oxaliplatin-based therapy. External validation confirmed that different combinations of these four genes consistently and robustly predicted oxaliplatin sensitivity. Furthermore, the expression levels of these four genes were significantly altered in both independent validation datasets and in established oxaliplatin-resistant cell lines. Functional studies demonstrated that silencing these genes directly influenced oxaliplatin cytotoxicity: knockdown of MAPK8IP2 significantly enhanced oxaliplatin-induced cell death, whereas knockdown of BAMBI and BMP7 significantly reduced cellular sensitivity to the drug. Conclusion: Our study identifies a novel four-gene signature that is strongly associated with oxaliplatin sensitivity in colorectal cancer. These findings support developing prognostic biomarkers to optimize oxaliplatin use, aiming to identify sensitive patients and avoid treatment for those with resistance.