A polymicrobial view of disease potential in Crohn's-associated adherent-invasive E. coli

ABSTRACT

The human gut is home to trillions of bacteria and provides the scaffold for one of the most complex microbial ecosystems in nature. Inflammatory bowel diseases, such as Crohn's disease, involve a compositional shift in the microbial constituents of this ecosystem with a marked expansion of Enterobacteriaceae, particularly Escherichia coli. Adherent-invasive E. coli (AIEC) strains are frequently isolated from the biopsies of Crohn's patients, where their ability to elicit inflammation suggests a possible role in Crohn's pathology. Here, we consider the origins of the AIEC pathovar and discuss how risk factors associated with Crohn's disease might influence AIEC colonization dynamics within the host to alter the overall disease potential of the microbial community.

Introduction

Crohn's disease (CD) is an inflammatory bowel condition with increasing incidence worldwide, most recently expanding into newly industrialized nations in South America, Asia, and the Middle East.¹ In Canada, CD incidence is among the highest in the world where about 20 per 100,000 individuals are diagnosed annually with the disease.² The protracted and often refractory course of CD constitutes a significant societal burden, both in direct medical costs to the health care system, and also in indirect costs that often exceed health spending.¹

Transmural inflammation and ulceration are common signatures of CD and about half of patients develop extraintestinal symptoms as well.² An abnormal immune response to commensal gut microbes is believed to be the driving force for CD-associated inflammation, yet the triggers of this aberrant immune response remain unclear.³ Hampering the understanding of CD etiology is a complex web of host and environmental factors that interact in indistinct ways to drive disease progression. Genome-wide association studies have established genetic links to CD; however, only ∼12% of CD patients have a family history of the disorder^4,5 and no single gene variant comes close to full penetrance in affected individuals. These findings stress the importance of non-genetic factors in disease onset and progression. It is tempting to speculate that the non-hereditary factors have propelled the high levels of CD in various countries, yet more studies are required to investigate this hypothesis.

CD is now considered a global disease. The sharp rise in CD in newly industrialized countries again emphasizes the involvement of a constellation of non-hereditary factors.¹ This may include the consumption of diets high in fat and low in fiber,⁶ smoking,⁷ and the use of certain medications including oral contraceptives, aspirin, and non-steroidal anti-inflammatory drugs.² The use of antibiotics, particularly during childhood, is linked to an increased risk of new onset CD.⁸ Despite the obvious beneficial role of antibiotics in combating bacterial infections, their indiscriminate activity also disturbs the balanced partnership between the human host and the gut microbiome, now recognized to preside over diverse states of health and disease in our bodies.⁹ Multiple studies have used metagenomics and 16s RNA profiling to examine the composition of gut bacteria during CD. This work as led to the broad conclusion that the microbiome of CD patients is compositionally different compared to healthy subjects.^10-13 Changes in microbial abundance, also known as dysbiosis, can markedly alter human immune responses, thus disturbing gut homeostasis and possibly leading to disease. Dysbiosis in CD is characterized by a loss of keystone species in the phyla Firmicutes and Bacteriodetes and the enrichment of Actinobacteria and Gammaproteobacteria.^14,15

Nearly 20 years ago, a newly described pathovar of the species Escherichia coli was isolated from CD patients in the laboratory of the late Dr. Darfeuille-Michaud, who made seminal contributions to AIEC research during her lifetime.¹⁶ These E. coli were referred to as adherent-invasive E. coli (AIEC), reflecting their ability to adhere to gut epithelial cells and their unusual ability to invade into mucosal epithelial cells. This moniker differentiates AIEC from other better-described E. coli pathovars like enteropathogenic E. coli, enterohemorrhagic E. coli, and enteroinvasive E. coli (EIEC).¹⁷ Despite the common ability of AIEC and EIEC to invade and replicate within intestinal epithelial cells, a closer scrutiny reveals a clear distinction between the two pathovars at the genomic level. In addition to the distinctive biochemical differences between the two pathotypes, AIEC lacks the typical invasins and pathogenicity islands found in EIEC. AIEC are now known to have an intracellular lifestyle where they can induce inflammatory pathways in host cells.^18,19 Numerous studies have confirmed that AIEC are enriched in humans with CD, where they are about six-times more likely to be isolated from ileal and colonic samples compared to healthy controls and represent the dominant bacterial species present.^16,20-25 Attention around the potential role of AIEC in the pathophysiology of CD is growing; however much remains to be learned about the host-pathogen interactions that govern AIEC infection biology.

Where does adherent-invasive Escherichia coli come from?

E. coli is a diverse bacterial species whose members range from seemingly innocuous commensal strains to quite dangerous human pathogens. Pathovar designations are used to classify E. coli into groups with unique molecular mechanisms that govern their pathogenic behavior.²⁶ The genetic determinants that help define E. coli pathovars (including serotype, toxins, and virulence factors) represent the basic tenets for their identification and facilitate the tracking of their evolutionary history.^27,28 While much is known about the evolution of many E. coli pathotypes, the origin of the AIEC group is less clear. One key challenge in defining the AIEC pathovar is that the genetic factors conferring the adherent-invasive phenotype are not fully defined. Consequently, the identification of AIEC is done based on a series of in vitro phenotype assays that are laborious, time-consuming, and somewhat non-standardized. Also, virulence determinants that define other E. coli pathovars at the genetic level (i.e. Shiga toxin, type III secretion systems) are not found in AIEC.¹⁷ While this fact can be used as exclusion criteria when attempting to classify isolates of E. coli from patients, a molecular genetic signature that distinguishes the AIEC pathovar remains elusive. In a recent study, comparative whole-genome analysis of 14 AIEC strains identified a potential subgroup within the B2 phylotype that appeared more similar due to three genetic insertions that differentiated them from commensal E. coli.²⁹ A separate study by a different group did not identify a readily distinguishable genomic signature among 11 different B2 phylotype AIEC strains although these strains were from a different geographic locale.²² As shown in different genomic studies, AIEC appear to be most closely related to extraintestinal E. coli strains such as UPEC, APEC, and ExPEC, which are also among the B2 clade.^17,29,30 Furthermore, many virulence factors were found to be shared between AIEC and UPEC, including genes that are required for iron acquisition and transport.¹⁷ Together, these findings suggest that AIEC do not arise by parallel evolution and clonal expansion as described for the notorious O157:H7 enterohemorrhagic E. coli.³¹

At this point, available evidence suggests that AIEC can evolve from diverse founder populations and use genetically distinct mechanisms to attain the ‘AIEC phenotype’. This is reminiscent of the genetic variability that exists among uropathogenic E. coli (UPEC), which also lacks a common genetic signature.³² UPEC are members of the so-called ‘extraintestinal pathogenic E. coli’ to which AIEC are also closely related.¹⁷ The selective drivers of AIEC's evolutionary trajectory remain obscure, yet they likely originate within the host. Given that individual host environments can drive adaptive bacterial evolution,³³ an important question is whether unique host environments (i.e. different CD patients; biogeography within a single individual; or combinatorial risk factors, for example) might select for AIEC emergence and outgrowth. Considering that research is now mobilizing towards AIEC as a new therapeutic frontier in CD,^34-39 efforts to understand the selective pressures driving AIEC evolution should be redoubled.

Factors mediating AIEC virulence

Despite the genetic diversity of AIEC, several strains do share common virulence factors, albeit not unique to AIEC. One of the defining features of AIEC is their ability to adhere to intestinal epithelial cells, which is likely facilitated by several bacterial surface structures. For example, surface expression of chitin-binding domains by AIEC was found to mediate their adherence to the chitinase-like receptors on the intestinal epithelium.⁴⁰ Additionally, long polar fimbriae mediate AIEC attachment to Peyer's patches, allowing AIEC to localize to the terminal ileum from where they are often isolated.⁴¹ And finally, AIEC type I pili can bind carcinoembryonic antigen-related cell adhesion molecule 6 (CEACAM6) via FimH, the terminal subunit of these surface appendages.^42,43 Unlike commensal E. coli strains, FimH in AIEC has accumulated a small number of non-synonymous mutations that may facilitate binding to CEACAM6.⁴⁴ Notably, the abundance of CEACAM6 in the ileum of CD patients⁴³ makes FimH a plausible target for pathoadaptation during AIEC evolution, and is the focus of most of the anti-adhesive strategies in therapeutic development in this space. Other host surface receptors that were found to be upregulated in the ileum of CD patients include the endoplasmic reticulum stress response glycoprotein Gp96. This host protein may be clinically relevant because it has been shown to interact with AIEC OmpA, a member of a family of outer membrane porins in Gram-negative bacteria.⁴⁵ Given that both CEACAM6 and Gp96 are more abundant during inflammation, the expression of inflammation itself may promote the redistribution of AIEC to sites proximal to the epithelial surface, a process seen by the host as a danger signal that initiates even greater inflammatory responses.⁴⁶

Other virulence factors in AIEC may help dictate its distribution within a host. These include GipA, a putative transcription factor involved in AIEC transcytosis of Peyer's patches similar to its function in the enteric pathogen Salmonella enterica serovar Typhimurium.^47,48 Access to the gut epithelium may be facilitated by Vat-AIEC, a vacuolating toxin related to one found in avian pathogenic E.coli that has proteolytic activity towards mucin.⁴⁹ It is noteworthy that both Vat-AIEC and long polar fimbriae were upregulated in the presence of bile salts. Although little is known about the chemical cues that direct AIEC gene regulation in the host, this finding infers the presence of potential regulatory circuits that have evolved to detect and respond to chemical signals that are unique to the gut environment. AIEC are able to invade, persist, and in some cases replicate within intestinal epithelial cells.⁵⁰ Multiple AIEC strains encode an invasin, called IbeA that is important for invasion of Caco-2 and M-cells, as well as survival within macrophages⁵¹ but it should be noted that a full accounting of how AIEC invades into cells is ongoing. Although IbeA is the only invasin identified in AIEC, the IbeA mutant strain was still able to invade epithelial cells, thus suggesting the possible presence of other uncharacterized invasins. Furthermore, the genomes of multiple AIEC strains encode for type VI secretion system (T6SS) components.²⁹ The T6SS is a surface structure related to contractile phage tails and used by Gram-negative bacteria to translocate effector proteins into both bacterial and eukaryotic cells. Although originally defined for its role in inter-bacterial competition,⁵² the T6SS is now recognized to have wider target specificity. For example, in some bacteria the T6SS is involved in dampening host immunity,⁵³ promoting intracellular growth in macrophages,⁵⁴ and interacting with the host microtubule network to promote bacterial internalization into host cells.⁵⁵ It remains to be determined whether type VI secretion contributes to AIEC pathogenesis, but given the role mediated by this machinery in other species, investigation of it in AIEC is warranted.

Antimicrobial peptides are host defense peptides that are capable of killing microbes by disrupting their membrane integrity and are often enriched in the gut during inflammation.⁵⁶ Many enteric pathogens can evade killing by these bactericidal molecules using enzymatic modification systems that alter surface chemistry of lipopolysaccharide, and by outer membrane proteases that cleave dibasic sites within the core of these cationic peptides. AIEC strain NRG857c, one of two prototype strains that have been fully sequenced, has high levels of resistance towards several antimicrobial peptides that are common in the gut.⁵⁷ Resistance to this arm of innate host defense in this particular strain is mediated by a plasmid-based genomic island (PI-6) that encodes an outer membrane protease (ArlC) and a Mig-14 family protein (ArlA). Mig-14 homologs are involved in antimicrobial peptide resistance in multiple bacteria, yet their mode of action is not fully understood.

The diversity of virulence factors displayed by multiple AIEC strains, some of them shared with ExPEC, suggests that members of this pathovar have evolved different strategies to colonize their hosts. This is also in keeping with the genetic variability that exists within the AIEC. It remains to be determined whether the host environment might favor AIEC strains that harbor a specific set of virulence determinants. We believe that an important next step in our understanding of AIEC is to broaden the scope of genomic studies to both bacteria and host. This could reveal potential associations between bacterial genotypes and host genetic backgrounds from which the strains arose.

A polymicrobial view of AIEC pathogenesis: The enigmatic role of acute infectious gastroenteritis in the long-term risk of CD onset

Crohn's disease is more common in individuals following acute infectious gastroenteritis caused by Salmonella and other enteric pathogens, sometimes with onset times on the order of years after the infectious episode.^58-60 The mechanistic basis for this long-term risk association following an acute infectious stimulus is enigmatic. The interaction between AIEC and their host seem to bring about multiple states, ranging from commensalism to pathogenic; however, the triggers for transitioning between these states are not known. The pathogenic potential of a microbe is often influenced by synergistic and antagonistic interactions with other microbes that collectively influence the community's pathogenic potential, known as nososymbiocity.⁶¹ Thus, disruption of gut homeostasis may be a key trigger that increases the overall disease potential of the microbial community.⁶¹ The well known ability of enteric pathogens to induce gut inflammation and disrupt the microbial community therein led us to question whether acute infectious gastroenteritis could be a trigger for the emergence of keystone pathogens like AIEC.

To address this question, we leveraged a previous model for chronic AIEC colonization that we developed in a range of inbred mouse lines.⁶² This resulted in two new polymicrobial infection models to measure host comorbidities related to acute gastroenteritis in mice colonized with a clinical isolate of AIEC. In these models, mice are colonized sub-clinically with AIEC and later exposed to secondary infection stimuli with either S. Typhimurium or Citrobacter rodentium, modeling what might occur in healthy AIEC-colonized individuals exposed to acute infectious gastroenteritis. S. Typhimurium is a prominent enteric pathogen linked to food poisoning in humans that triggers a strong inflammation in the gut,⁶³ while C. rodentium is a murine pathogen that recapitulates the gastroenteritis mediated by enteropathogenic E. coli in humans.⁶⁴ Using these models, we showed that AIEC-colonized mice exposed to acute infectious gastroenteritis have a significantly worsened outcome compared to AIEC-naïve animals exposed to the same infection stimuli.⁶⁵ To our surprise, disease activity was not driven by the secondary pathogen but rather by AIEC in the post-infectious period when the host had largely constrained the secondary pathogen. Acute bacterial gastroenteritis induced an AIEC bloom in mucosal tissue that was associated with host damage. Importantly, rendering AIEC susceptible to host defense peptides, or quelling the inflammation induced by the secondary pathogen prevented AIEC blooms and mitigated the disease activity, indicating that AIEC is a tractable disease modifier following acute infectious gastroenteritis.

Other environmental factors can instigate gut inflammation that might be relevant to AIEC colonization dynamics. For example, dysbiosis triggered by antibiotics or a “Western diet” high in sugar and fat was shown to cause a reduction in short-chain fatty acid (SCFA) levels in the intestine.^66-68 SCFA can mediate anti-inflammatory functions in the gut via the activation of regulatory T cells,⁶⁹ and the reduction of oxygen availability.⁶⁹ Antibiotic treatment was shown to increase the inflammatory tone of the gut mucosa in mice, characterized by increased infiltration of neutrophils and inflammatory monocytes in the lumen. The converging view is that the induction of inflammation in mice with either antibiotics,⁶² Western diet,⁶⁶ or acute infectious gastroenteritis⁶⁵ causes expansion of resident AIEC that is linked to worsening levels of gut pathology, leading to the suggestion that inflammation provides AIEC with a selective advantage in the gut (Fig. 1).

Figure 1.

A schematic diagram that summarizes some of the inflammation-mediated changes in the gut and their influence on AIEC colonization. Gut inflammation can be triggered by multiple factors like host genetics, antibiotic administration, diet, and acute gastroenteritis. The resulting proinflammatory environment mediates pronounced changes in the human gut, which includes hypoferremia, the generation of alternative electron acceptors, increased secretion of antimicrobial peptides, and overexpression of CEACAM6 surface receptors by intestinal epithelia. During their evolution, AIEC strains have gained traits that confer on them a competitive advantage in the inflamed gut. This includes the ability to utilize S- and N- oxides as alternative electron acceptors, an abundance of iron acquisition genes, resistance to antimicrobial peptides and the ability to bind to CEACAM6 through a modified FimH protein.

In addition to the overt pathological changes triggered by inflammatory reactions at mucosal sites, inflammation can change nutrient availability in the gut. For example, hypoferremia is a hallmark of gut inflammation, and acts as a host defense mechanism against invading pathogens by limiting iron availability.⁷⁰ Genes involved in iron utilization such as siderophores are enriched in AIEC compared to E. coli strains from other pathotypes.⁷¹ Siderophores are also commonly found in extraintestinal E. coli species such as UPEC where it is an important virulence factor. Since siderophores, like aerobactin, provide a strong fitness advantage in hypoferremic conditions,⁷² this may suggest that AIEC emerge under selection in an iron-poor environment. Similar to the closely related UPEC and APEC, aerobactin was found to be important for AIEC virulence.¹⁷ The production of reactive oxygen and nitrogen species by neutrophils is another important outcome of gut inflammation.⁶⁸ Upon release into the gut lumen, reactive oxygen and nitrogen species can react with organic sulphides and tertiary amines to generate S- and N- oxides that can be used as alternative electron acceptors by some facultative anaerobes. It is conceivable that nutritional changes associated with inflammation might create a selective niche for AIEC that brings about their pathogenic potential, or perhaps renders hosts more susceptible to de novo colonization through loss of some facet of colonization resistance, a host state that we are just beginning to understand with granularity.⁷³ More studies are required to investigate the metabolic behavior of AIEC during gut inflammation, which might provide insights to the evolutionary trajectories that shaped this pathovar.

Concluding remarks

While a microbial basis for Crohn's pathogenesis is well founded, it is currently unclear how microbes influence, and are influenced by, the inflammatory environment in the gut. The role that environmental risk factors play in disease expression is also not fully understood, however the interactions created along a genes-environment-microbe axis hold the key to unlock future preventions and therapies. Emerging data from multiple studies place microbes at the epicentre of CD pathogenesis, with members of the emergent AIEC pathovar being possible disease modifiers. Evidence that the host environment influences AIEC evolution, colonization, and disease potential makes it a fascinating case study in host-pathogen interactions. For example, if gut inflammation plays a fundamental role in evoking the pathogenic character of AIEC, then many questions emerge. How does host genotype affect AIEC behavior in vivo? Do anti-inflammatory medicines used in CD treatments have secondary effects on AIEC colonization dynamics that improve disease outcome? How does AIEC influence disease potential of the microbial community in which it resides? Are individuals with gut inflammation more susceptible to host-to-host transmission of AIEC? Phenotypic changes in the gut induced by inflammatory reactions might provide the selective niche for the evolution of AIEC strains from different phylogroups and explain the lack of a common ancestor. To mitigate the burden of CD, a primary goal of research centers on the interactions between genes, the environment, and intestinal microbes using robust preclinical models. More studies investigating AIEC pathogenesis in the context of host genetics and environmental risk factors linked to CD will be informative towards this goal. Finally, as evidence mounts against AIEC as a tractable disease modifier, understanding the provenance of AIEC, their movement through the environment, and their transmission dynamics from host-to-host is likely to yield major public health dividends.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

Research on AIEC in the Coombes laboratory is funded by grants from the Canadian Institutes of Health Research and Crohn's and Colitis Canada. A.O. is supported by a Canada Graduate Scholarship from the Canadian Institutes of Health Research. B.K.C. is the Canada Research Chair in Infectious Disease Pathogenesis.