Evidence for a Causal Role for Escherichia coli Strains Identified as Adherent-Invasive (AIEC) in Intestinal Inflammation

Enrichment of adherent-invasive Escherichia coli (AIEC) has been consistently detected in subsets of inflammatory bowel disease (IBD) patients. Although some AIEC strains cause colitis in animal models, these studies did not systematically compare AIEC with non-AIEC strains, and causal links between AIEC and disease are still disputed. Specifically, it remains unclear whether AIEC shows enhanced pathogenicity compared to that of commensal E. coli found in the same ecological microhabitat and if the in vitro phenotypes used to classify strains as AIEC are pathologically relevant. Here, we utilized in vitro phenotyping and a murine model of intestinal inflammation to systematically compare strains identified as AIEC with those identified as non-AIEC and relate AIEC phenotypes to pathogenicity. Strains identified as AIEC caused, on average, more severe intestinal inflammation. Intracellular survival/replication phenotypes routinely used to classify AIEC positively correlated with disease, while adherence to epithelial cells and tumor necrosis factor alpha production by macrophages did not. This knowledge was then applied to design and test a strategy to prevent inflammation by selecting E. coli strains that adhered to epithelial cells but poorly survived/replicated intracellularly. Two E. coli strains that ameliorated AIEC-mediated disease were subsequently identified. In summary, our results show a relationship between intracellular survival/replication in E. coli and pathology in murine colitis, suggesting that strains possessing these phenotypes might not only become enriched in human IBD but also contribute to disease. We provide new evidence that specific AIEC phenotypes are pathologically relevant and proof of principle that such mechanistic information can be therapeutically exploited to alleviate intestinal inflammation. IMPORTANCE Inflammatory bowel disease (IBD) is associated with an altered gut microbiota composition, including expansion of Proteobacteria. Many species in this phylum are thought to contribute to disease under certain conditions, including adherent-invasive Escherichia coli (AIEC) strains, which are enriched in some patients. However, whether this bloom contributes to disease or is just a response to IBD-associated physiological changes is unknown. Although assigning causality is challenging, appropriate animal models can test the hypothesis that AIEC strains have an enhanced ability to cause colitis in comparison to other gut commensal E. coli strains and to identify bacterial traits contributing to virulence. We observed that AIEC strains are generally more pathogenic than commensal E. coli and that bacterial intracellular survival/replication phenotypes contributed to disease. We also found that E. coli strains lacking primary virulence traits can prevent inflammation. Our findings provide critical information on E. coli pathogenicity that may inform development of IBD diagnostic tools and therapies.