A multi-omic machine learning approach deconstructs the role of amino acid metabolism in the immune microenvironment and prognosis of colon adenocarcinoma.

Background: The clinical heterogeneity of colon adenocarcinoma (COAD) complicates patient prognosis and treatment. While metabolic reprogramming is a key driver of tumor progression, the role of histidine metabolism in the COAD immune microenvironment remains unclear. Methods: Through an integrated analysis of public single-cell and bulk transcriptomic data, we mapped the COAD cellular atlas and assessed histidine metabolism activity. A prognostic Histidine Metabolism-related Model (HRM) was constructed using an ensemble of 101 machine learning algorithms, and its biological underpinnings were explored. The function of the key gene, TRIP6, was validated via in vitro experiments. Results: Single-cell analysis identified epithelial cells as the hub of histidine metabolism, which remodels TME intercellular communication. The machine learning-derived HRM robustly stratified patient prognosis across a primary and two validation cohorts. High HRM scores correlated with an "infiltrated-exhausted" immune phenotype, characterized by high immune infiltration alongside elevated checkpoint expression. The key signature gene TRIP6 was identified as a driver of poor prognosis and an immunosuppressive state, and its silencing suppressed malignant phenotypes in vitro. Conclusion: Histidine metabolism is a critical regulator of the COAD immune microenvironment. Our prognostic model, the HRM, provides a clinically relevant tool for risk stratification, while its key mediator, TRIP6, represents a novel therapeutic target linking tumor metabolism to immune evasion.