Postbiotic-related host transcriptomic axes link tumour immune microenvironment, stage and survival in colorectal cancer
摘要
The tumour immune microenvironment shapes colorectal cancer progression and therapeutic vulnerability, but host metabolic programmes linked to immune-inflamed tumour states remain incompletely defined. Microbiome-derived and postbiotic-related metabolites influence epithelial, stromal and immune signalling, yet their downstream tumour transcriptomic readouts have rarely been assessed as host-side indicators of immune microenvironment state.
MethodsFour mechanistically defined postbiotic-related host axes were curated: short-chain fatty acid, tryptophan/AhR, polyamine and ferroptosis/redox. Axis scores were calculated as mean z-standardized expression of curated genes in TCGA colorectal cancers and evaluated in GSE39582. Associations with marker-based immune/stromal signatures, microsatellite instability, cytolytic activity, stage and overall survival were tested using Spearman correlations, Wilcoxon tests and Cox models with false-discovery-rate correction.
ResultsIn TCGA, the tryptophan/AhR axis showed the strongest coupling to marker-based immune and myeloid signatures, including macrophage-, Treg-, NK-cell-, CD8 T-cell-, Th1-like and cytolytic-activity programmes. Tryptophan/AhR, short-chain fatty acid and ferroptosis/redox axes were enriched in MSI-H and immune-high tumours, whereas the polyamine axis showed weaker immune associations. Short-chain fatty acid and ferroptosis/redox axes decreased recurrently in advanced-stage disease across cohorts. No universal prognostic axis emerged, although the short-chain fatty acid axis retained a stage-adjusted association with improved overall survival in GSE39582.
ConclusionsPostbiotic-related host transcriptomic axes provide interpretable tumour-level readouts connecting metabolic cues with immune microenvironment states in colorectal cancer. They should be considered hypothesis-generating host-side immune-context readouts rather than ready-to-use clinical signatures.