Metabolomics assisted by transcriptomics analysis to reveal metabolic characteristics and potential biomarkers of early-onset colorectal cancer.

Metabolic reprogramming plays a crucial role in the occurrence and progression of colorectal cancer (CRC). However, the relationship between metabolism and CRC development remains to be fully elucidated. We collected 45 colorectal tissue samples, including 15 cases of early-onset colorectal cancer (EOCRC), 15 cases of later-onset colorectal cancer (LOCRC), and 15 adjacent normal tissues for metabolomic analysis. A prediction model based on differentially expressed metabolites (DEMs) was established using the random forest algorithm with 10-fold cross-validation. Transcriptome analysis was performed on the GSE39084 dataset, which included 25 EOCRC and 35 LOCRC cases. Pathway-level integrative analyses combining metabolomics and transcriptomics were further conducted. Significant differences in metabolic profiles were observed among EOCRC, LOCRC, and adjacent normal tissues. A total of 141 DEMs were identified, showing significant enrichment in 20 metabolic pathways. The random forest model, consisting of 10 metabolites, achieved a cross-validated area under the curve (AUC) value of 0.916 for distinguishing EOCRC from other groups. Weighted Gene Co-expression Network Analysis (WGCNA) revealed a gene module strongly positively correlated with Lynch syndrome, microsatellite instability (MSI) status, and PIK3CA mutations. Pathway-level integrative analysis highlighted the Citrate cycle (TCA cycle) and Central carbon metabolism pathways as consistently enriched at both metabolic and transcriptional levels. Furthermore, four metabolism-related genes were identified as candidate genes warranting further investigation in EOCRC. The integration of metabolomics and transcriptomics in this study provides novel insights into the pathological alterations associated with EOCRC, facilitating the identification of candidate therapeutic targets.