The microbiota-NK cell axis in colorectal cancer: a spatial and subset-centric framework.

The gut microbiota and natural killer (NK) cells jointly shape colorectal cancer (CRC) progression and immunotherapy response, yet existing reviews largely adopt a linear, pathogen-centric view that overlooks NK-cell heterogeneity and spatial tumor microenvironment architecture. We propose an integrated framework centered on three progressive pillars: (i) the microbiota appears to selectively shape three well-characterized NK-cell states (activated, dysfunctional, and memory-like), with a fourth regulatory-like state that remains mechanistically uncharacterized through metabolically and genetically mediated mechanisms; (ii) emerging spatial evidence suggests beneficial commensals and activated NK cells preferentially localize to perivascular niches, whereas pathogenic bacteria and dysfunctional NK cells are enriched within the tumor parenchyma, though multicenter validation is needed; and (iii) these insights support two testable patient-stratification strategies for future clinical testing, an "enrichment" approach for tumors with pre-existing activated NK signatures, and a "remodeling" approach for Fusobacterium nucleatum (F. nucleatum)-dominant immunosuppressive microenvironments, particularly in refractory microsatellite-stable disease. This framework seeks to shift the field toward a subset-centric, spatially resolved, and clinically testable model of the microbiota-NK axis, designed to generate hypotheses and guide prospective cohort studies rather than replace current standards.