Identification and validation of biomarkers related to mismatch repair for prognosis prediction in glioma

Background Gliomas are aggressive brain tumors with a dismal prognosis, and their development, progression, and responsiveness to treatment are all impacted by mismatch repair (MMR) deficiency. This work aimed to construct a predictive risk model based on MMR-related genes and confirm its clinical applicability, with a focus on identifying novel therapeutic targets. Methods Clinical and mRNA expression data from The Cancer Genome Atlas (TCGA) and The Chinese Glioma Genome Atlas (CGGA) glioma patients were analyzed. MMR-related genes were sourced from the Molecular Signatures Database (MSigDB). A risk score model was created using multivariate Cox regression and LASSO analysis. Patients were categorized into high- and low-risk groups based on the median risk score. The model’s performance was assessed using ROC curves, AUC, and Kaplan-Meier survival analysis. Immune cell infiltration was quantified using “CIBERSORT” and “QUANTISEQ”. Immunotherapy potential was evaluated via TMB, TME, and TIDE scores. The half maximal inhibitory concentration (IC 50 ) analyses were performed. In vitro validation involved MCM8 overexpression/knockdown in U251/LN229 glioma cells, followed by TMZ sensitivity testing via CCK-8 assays. Results An eight-MMR-related gene prognostic model (HMGB1, MCM8, MUTYH, PMS1, RNASEH2B, RNASEH2C, RPA3, TP73) was constructed. The risk score was an independent prognostic factor, with high-risk patients showing significantly poorer overall survival. The validity of this model has been validated in the CGGA dataset. Significant differences in immune infiltration, TME, and TMB scores were observed between risk groups. Drug sensitivity analysis revealed distinct IC 50 profiles for chemotherapeutic agents between the groups. Most importantly, in vitro experiments demonstrated that MCM8 overexpression increased glioma cell sensitivity to TMZ, while MCM8 knockdown decreased it, identifying MCM8 as a potential target for TMZ-based personalized therapy. Conclusion This study establishes a robust MMR-related prognostic model for glioma that effectively stratifies patients, predicts survival, and reflects distinct immune microenvironments. Critically, we identify MCM8 as a novel and actionable biomarker that modulates TMZ sensitivity, offering a promising avenue for personalized treatment in glioma patients.