Microflora in Crohn’s disease: the emergence of adherent and invasive Escherichia coli

Crohn’s disease (CD) is a chronic, relapsing, transmural, inflammatory bowel disease (IBD) with symptoms including diarrhea, abdominal pain, weight loss and perianal fistula or fissures. Although the exact etiology is still unknown, it is generally considered a consequence of uncontrolled intestinal inflammation in response to a combination of elusive environmental, enteric microbial and immuno-regulatory factors in genetically susceptible individuals [1]. A diverse spectrum of pathogenic bacteria, including Listeria, Streptococcus, Enterococcus, Enterobacteriaceae, Bacteroides, Clostridium, Yersinia and Mycobacterium avium subsp. paratuberculosis (MAP) [1–8], have been implicated in the development of inflammation, but no single agent has emerged as a consistent cause. In the absence of a defined cause, therapy is symptomatic and supportive and achieves remission rather than cure.

The experimental literature now provides a clear indication that the resident enteric microflora, rather than typically recognized pathogenic bacteria, is involved in the pathogenesis of IBD in rodent models. For example, Enterobacteriaceae, particularly Escherichia coli, are the dominant microflora in the inflamed intestines of IL-2^-/- mice, and the ceca of HLA-B27 β₂-microglobulin transgenic rats [9,10]. Resident E. coli and Entero-coccus faecalis have also been shown to induce colitis and cecal inflammation in monoinfected gnotobiotic IL-10^-/- mice [11]. Similarly, evidence implicating the resident enteric microflora in the pathogenesis of spontaneous IBD in people is provided by the clinical responses of CD to fecal stream diversion [12] and anti-microbials [2,13], the increased circulating and intraluminal humoral and T-cell responses to enteric commensal bacteria, and imbalances in the ratio of beneficial to harmful bacteria (termed ‘dysbiosis’) observed in CD [13–18]. Complementing this is the discovery of mutations in genes regulating microflora-sensing ability (NOD2/CARD15 and Toll-like receptor [TLR]4) of patients with CD, which provides mechanisms to explain individual susceptibility to the resident microflora [19,20] that, in the presence of the enteric microflora and appropriate triggers, may lead to upregulated mucosal cytokine production, delayed bacterial clearance and increased bacterial translocation, thereby promoting and perpetuating intestinal inflammation [1,21,22].

Culture-independent analysis of CD mucosa reveals a selective increase in E. coli & depletion in Clostridiales

Advances in molecular microbiology have led to a new awareness of the diversity and complexity of the enteric flora. Culture--independent analyses of 16S rDNA libraries reveal that only 30% of the fecal flora appears cultivable, and there is significant variation in the flora in different gastrointestinal segments and luminal contents versus the mucosa of healthy individuals [23–25]. Analysis of 16S rDNA libraries constructed from endoscopic biopsies of the ileum from patients with Crohn’s ileitis (a disease phenotype that occurs in approximately 70% of individuals with CD), demonstrates selective enrichment of ileal mucosa in E. coli and relative depletion in Clostridiales compared with both patients with CD restricted to the colon and to healthy individuals [26]. PCR of ileal DNA yielded no evidence of MAP, Listeria or Shigella and confirmed the increase in E. coli [26]. Similarly, 16S rDNA library-based evaluation of surgically resected CD ileum shows an increase in Proteobacteria (this phylum includes Enterobacteriaceae), decrease in Firmicutes (this phylum contains Clostridiales) and the absence of DNA from MAP in CD mucosa [27]. The relative decrease in Clostridiales in CD ileal mucosa echoes findings in studies of feces and colonic mucosa from patients with CD, and suggests that this clostridial dys-biosis may be a feature of CD in general, rather than a marker of CD ileitis [28–30].

The in-depth knowledge of microfloral composition provided by 16S sequence analysis can be used to guide in situ examination of the spatial distribution of the mucosal flora interacting most closely with the innate immune system. Fluorescence in situ hybridization (FISH) with labeled oligonucleo-tide probes targeted to specific 16S ribosomal RNA sequences (for example [31]) allows spatial localization of different microbial subsets within histological samples. FISH with a probe restricted to E. coli/Shigella reveals that the ileal mucosa of CD patients harbors significantly more E. coli than normal ileum, and mucosally invasive E. coli are only detected in CD mucosa [26]. In patients with Crohn’s ileitis, E. coli represent approximately 3.5% of the mucosal flora labeled with the panbacterial FISH probe EUB338 and approximately 40% of the Enterobacteriaceae recognized by the enterobacterial FISH probe 1531 [26]. This contrasts with culture-based studies reporting that E. coli represent more than 50% of the cultivable mucosal flora in 40–58% of patients with ileal CD [32,33]. However, this discrepancy is readily explained by the inability to culture the vast majority of enteric mucosal flora and supports the use of culture-independent analysis. Furthermore, the number of E. coli visualized in ileal biopsies by Baumgart et al. [26], but not bacterial colonization in general, correlates with histological and endoscopic disease activity, which suggests E. coli could be specifically involved in the inflammatory process [26]. These findings complement previous studies that have documented E. coli antigens and DNA in granulomas and peri-ulcerative lesions in CD [6,34], and provide a rational explanation for the higher prevalence of antibodies directed against E. coli outer membrane porin C (OmpC) and flagellin in patients with CD involving the ileum versus colon-restricted CD [16,35].

The association of adherent & invasive E. coli with CD

Culture-based analysis of E. coli isolated from the mucosa of patients with CD has revealed that many strains display pathogen-like behavior in cultured cells [26,36–38]. The best-characterized CD-associated E. coli strain, LF82 (isolated by Darfeuille-Michaud from the ileum of a patient from Lille, France), adheres to, invades, persists and replicates in cultured intestinal epithelial cells [36]. Invasion is associated with trigger-type endocytosis and utilizes host cell microfilaments and microtubules [39]. LF82 is also able to replicate in J477-A1 macrophages, promote the elaboration of disease-relevant cytokines, such as TNF-a and IL-8, in cultured cells and elicit granuloma formation in vitro [8,39–41]. LF82 is distinct from known pathogenic E. coli and is the prototypical strain of a putative new pathogroup of E. coli, adherent and invasive E. coli (AIEC).

Adherent and invasive E. coli have been isolated from the ileum of 36–38% of patients with Crohn’s ileitis compared with 6% from the ileum of healthy individuals [26,32]. Conversely, in colonic specimens, AIEC strains were found in only 3.7% of CD patients, 0% of ulcerative colitis (UC) patients and 1.9% of controls [8,32], suggesting that AIEC may be preferentially involved in ileal, rather than colonic inflammation. The predominance of AIEC in the inflamed ileum has recently been addressed by Barnich et al. who found that AIEC strains adhere to the brush border of primary ileal enterocytes isolated from CD patients but not controls without IBD [42]. AIEC adhesion is dependent on both bacterial surface expression of type I pili and on expression of carcinoembryonic antigen-related cell adhesion molecule (CEACAM)6 on the apical surface of ileal epithelial cells, an interaction that can be inhibited by mannose [42]. These in vitro findings are highly consistent with recent in situ observations that E. coli colonization correlates with disease activity [26]. It is intriguing to think that CEACAM6 may be the proverbial door through which AIEC gain entry into the mucosa, whereupon they are eradicated in nonsusceptible individuals, but evade ineffectual host defenses in susceptible individuals, resulting in chronic intestinal inflammation.

While AIEC are generally associated with ileal colonization, invasive AIEC-like E. coli strains have also been isolated from the colonic mucosa of patients with CD, UC and colon cancer [37,38,43]. However, CD-associated E. coli strains, especially the strains from inflamed CD tissue, have been shown to be much more invasive than UC-associated strains, and E. coli strains from inflamed CD tissue induced significantly higher TNF-a expression in macrophage cell lines than UC-associated E. coli strains [38]. CEACAM6 is expressed in the colonic mucosa of patients with CD and UC [8,42], and could be a receptor for colonic E. coli strains that display AIEC-like activity in cultured epithelial cells. However, this possibility has to be reconciled with the results of in situ analysis that show marked increases in mucosally adherent Bacteroides rather than E. coli in CD and UC colonic mucosa [44].

AIEC are associated with intestinal inflammation in dogs

Granulomatous colitis of Boxer dogs (GCB) is a severe disease of unknown etiology that typically affects Boxer dogs [45]. GCB has features in common with UC (macroscopic appearance, regional distribution and immunopathology), Whipple’s disease (PAS-positive macrophages) and CD (granulomatous inflammation, bacteria within macrophages and may respond to fluroquinolones) in humans but it is not identical to any of these disease syndromes [46–49]. 16S rDNA libraries generated from GCB dogs are dominated by sequences for Enterobacteriaceae, predominantly E. coli and Shigella and FISH probes recognize invasive E. coli within the colonic mucosa (in the lamina propria and macrophages) of Boxer dogs with GCB but not unaffected controls, and non-GCB colitis. E. coli strains isolated from affected Boxer dogs display an AIEC pathotype in cultured cells, and resemble LF82 in phylo-geny and virulence genes [45]. Subsequent studies of GCB-affected dogs have shown that antibiotic-induced clinical remission correlates with eradication of invasive E. coli. Thus, GCB is a curable enteropathy associated with mucosally invasive E. coli.

AIEC strains may share common pathoadaptive determinants of virulence that promote intestinal inflammation across species

Bacterial virulence is regulated typically by genes that promote the invasion or replication of organisms in target cells. In contrast to pathogenic E. coli, such as O157:H7, AIEC strains differ in serotype, overall genotype [26,45] and ribotype [50] and are not a single virulent clone. By use of PCR-based virulence screening and genome subtraction, we discovered that AIEC strains with unique multilocus sequence type sequences and distinct phylogenetic backgrounds (A, B1, B2 and D) harbor chromosomal and episomal elements that are homologous to those described in extraintestinal pathogenic E. coli (uropathogenic E. coli [UPEC] and avian pathogenic E. coli), and pathogenic Enterobacteriaceae, such as Salmonella and Yersinia [26,45], and support the possibility that AIEC strains share common pathoadaptive determinants of virulence. For example, nucleotide sequences resembling ratA in Salmonella typhimurium , which is associated with colonization of the cecum and Peyer’s patches [51], and another resembling a hemolysin coregulated protein (hcp) secreted by a type VI secretion system associated with virulence in Vibrio [52], are potentially disease-relevant genes present in AIEC from CD and GCB. A plasmid resembling Yersinia pestis pMT1 is present in strain LF82, but not other AIEC strains, which has important implications as LF82 is regarded as the prototype AIEC strain, and it may not be representative of AIEC strains in general [8,39].

While studies to date reinforce the high degree of diversity in E. coli as a species and its propensity for acquiring DNA from distantly related organisms, they provide only a glimpse of the potential genetic armory of AIEC. Genome sequencing will be required to resolve whether AIEC strains have acquired common virulence determinants in parallel, or have adopted the ‘mix-and-match’ approach described for UPEC [54], and to aid in the identification of additional candidate genes associated with virulence. It is also important to consider that differences in gene regulation, rather than presence or absence of a locus or molecular selection, may influence virulence of AIEC. This possibility is reinforced by recent studies of type I pili, flagella and OmpC in AIEC LF82 [8,53].

Conclusion

There is mounting evidence, derived from independent studies utilizing an array of culture-based and culture-independent methodologies that links E. coli to CD. The correlation between the number of E. coli visualized in ileal biopsies and disease activity, coupled with the ability of CD-associated E. coli to display pathogen-like adherent and invasive behavior in cultured cells, supports a role for E. coli in the inflammatory process. The remarkable similarity between E. coli strains isolated from Crohn’s ileitis and GCB raises the possibility that AIEC are a novel group of pathogenic E. coli that share common pathoadaptive determinants of virulence that promote intestinal inflammation. The isolation of AIEC from healthy and inflamed mucosa indicates that their mere presence alone is insufficient to cause disease. The increased numbers of mucosa-associated and invasive E. coli relative to other bacteria observed in patients with Crohn’s ileitis suggest that AIEC, with the appropriate stimulus, are most likely opportunistic pathogens that can exploit the mucosal environment of a Crohn’s-susceptible individual who has, for example, reduced defensin production and delayed bacterial clearance associated with functionally important polymorphisms in NOD2/CARD15 [1,22]. Alternatively, AIEC may permissively proliferate subsequent to a depletion of normal flora, such as Faecalibacterium, or alterations in bacterial products like butyrate, which is important for colonic health and ileal regeneration [55,56], or a combination of these possibilities. Resolving these issues will provide unique insights into the biology of CD and opportunities for therapeutic intervention.

Financial & competing interests disclosure

Belgin Dogan is supported by a grant from the Jill Roberts Center for Inflammatory Bowel Disease, Weill Medical College of Cornell University, NY 10021, USA. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.