Related Experiment Video
Updated: Aug 5, 2026
Molecular Profiling of the Invasive Tumor Microenvironment in a 3-Dimensional Model of Colorectal Cancer Cells and Ex vivo Fibroblasts
Published on: April 29, 2014
Identification of Methionine Metabolism-Driven Heterogeneous Subtypes in Colorectal Cancer and Their Associated
Junhao Zhou1,2, Wanhua Deng1,2, Manzhao Ouyang1,2,3
1Department of Gastrointestinal Surgery, The Eighth Affiliated Hospital, Southern Medical University (The First People's Hospital of Shunde, Foshan), Shunde, Foshan, China.
Aim:
Tumor heterogeneity, driven by metabolic reprogramming, challenges colorectal cancer (CRC) treatment. Methionine metabolism is crucial for tumor progression, but its role in CRC heterogeneity and the tumor immune microenvironment (TIME) requires systematic investigation.
Methods:
A systematic evaluation of 101 combinations of machine learning and statistical algorithms was conducted within a 10-fold cross-validation framework to develop and validate the optimal model, termed the methionine metabolism-related risk score (MMRS). The Cancer Genome Atlas-Colon Adenocarcinoma (TCGA-COAD) dataset served as the training cohort, while two independent Gene Expression Omnibus (GEO) cohorts (GSE39582 and GSE17536) were employed for external validation. Immune infiltration was assessed using the microenvironment cell populations-counter (MCP-counter) algorithm. Model discrimination was evaluated using time-dependent receiver operating characteristic (ROC) analysis, with the area under the curve (AUC) calculated at 1, 3, and 5 years across all cohorts.
Results:
Methionine metabolism-high (MMH) and metabolism-low (MML) subtypes were defined via unsupervised clustering. The MMH subtype exhibited significantly poorer overall survival and an immunosuppressive TIME. From subtype-associated differentially expressed genes (DEGs), 41 prognostic genes were identified. The optimal model (StepCox[both] + plsRcox) formed the MMRS, which effectively stratified patients into high- and low-risk groups with significantly different survival across all cohorts (all p < 0.05). Core signature genes (MID2, KIF7, GSR) were consistently selected. The high-risk group showed depleted antitumor immunity (e.g., fewer CD8+ T and NK cells) and a stroma-rich phenotype with enrichment of cancer-associated fibroblasts, underpinned by oxidative stress and alterations in energy metabolism pathways. Time-dependent ROC analysis confirmed the discriminative capacity of the MMRS, with 5-year AUCs of 0.803 (TCGA-COAD), 0.632 (GSE39582), and 0.608 (GSE17536), respectively, further supporting its prognostic accuracy across independent patient populations.
Conclusion:
Methionine metabolism heterogeneity is a key determinant of prognosis and immune contexture in CRC. The MMRS is a prognostic signature that effectively stratifies CRC patients by risk, though its clinical utility as a predictive biomarker for treatment selection requires prospective validation.
