Abstract
Bacteroides fragilis is a human commensal bacterium that colonizes the majority of adults. There are two subtypes of this bacterium, Nontoxigenic Bacteroides fragilis (NTBF) and Enterotoxigenic Bacteroides fragilis (ETBF), which differ based on the production of Bacteroides fragilis toxin (BFT) by ETBF. ETBF may asymptomatically colonize humans, but it may also cause diarrheal illness. Early microbiome studies of patients with colorectal cancer (CRC) revealed that 90% of patients were colonized with ETBF, suggesting that ETBF was involved in cancer formation or progression. Indeed, in vivo studies in genetically susceptible mice showed that colonization with ETBF could lead to tumor formation in the colon within 4 weeks. Tumor formation is dependent on a pro-inflammatory signaling cascade that begins with E-cadherin cleavage in colon epithelial cells (CECs). T helper 17 (Th17) cells producing IL-17A are recruited to the area and stimulate STAT3 and NFκB in CECs, leading to aberrant expression of cytokines and, eventually, the formation of tumors. The initial host response to BFT has not been fully characterized, so we focused on early changes in CECs that could promote inflammatory disease. To do this, we measured sphingolipids, a class of lipids with a sphingoid backbone that play a wide number of roles in cells, from growth and differentiation to cell death. We found that normal C57BL/6J mice colonized with ETBF for one week had increased levels of glucosylceramide in their distal colon. Glucosylceramide is a critically important lipid in the intestines because of its role in maintaining the epithelial cell barrier. In order to determine the purpose of BFT-induced glucosylceramide increases in the colon, we utilized colon organoids (colonoids) derived from the distal colons of C57BL/6J mice. Colonoids treated with BFT showed higher levels of glucosylceramide, consistent with our previous in vivo findings. We assessed the importance of glucosylceramide in CEC response to BFT by using pharmacological inhibitors of glucosylceramide synthase (GCS), the enzyme responsible for generating glucosylceramide, and glucocerebrosidase (GBA), the enzyme responsible for breaking down glucosylceramide. We found that inhibition of GCS caused colonoids to burst, a phenomenon that could be prevented by blocking GBA. The prevention of colonoid bursting was due to the stabilization of tight junction protein 1 (TJP1), an important mediator of tight junctions that regulates paracellular permeability. In addition, we found that glucosylceramide was released from CECs in extracellular vesicles (EVs), although the role of glucosylceramide in EVs is unclear at this point. Together, we have shown for the first time a novel mechanism that CECs use to protect the epithelial barrier from bacterial toxins.
