An outer membrane vesicle‐associated two‐track model for enterotoxigenic Bacteroides fragilis oncogenesis

Enterotoxigenic Bacteroides fragilis (ETBF) is a pro-carcinogenic pathobiont thought to drive oncogenesis solely through E-cadherin cleavage by its metalloprotease toxin, Bacteroides fragilis toxin (BFT). This canonical framework leaves three paradoxes unresolved: the cell type selectivity of BFT, the cytosolic engagement of NOD1 by an extracellularly secreted toxin, and the divergent epidemiology of BFT isoforms. The recent identification of claudin-4 as the obligate BFT receptor resolves the first but sharpens the others. Here we propose an outer membrane vesicle (OMV)-associated two-track model. In Track 1, the surface program, claudin-4-dependent docking drives E-cadherin ectodomain cleavage, junctional disruption and β-catenin/Th17-driven inflammation. In Track 2, the proposed intracellular program, OMV-mediated endocytosis delivers OMV-anchored BFT-1 into the endocytic compartment, where endosomal escape would permit cytosolic NOD1 engagement and activation of a GAK-NUMB-NOTCH1-HEY1 axis linked to cancer stem cell expansion and chemoresistance-an axis so far demonstrated in breast cancer cells, not in colonic epithelium. We further hypothesize an intrinsic isoform bias: BFT-2 toward surface cleavage, BFT-1 toward intracellular NOD1 engagement. This spatially resolved model unifies recent mechanistic advances with unexplained epidemiological patterns.