Enteric Microflora in IBD: Pathogens or Commensals?

Summary

Both pathogenic and normal enteric microflora can induce and perpetuate chronic intestinal inflammation with systemic manifestations in genetically susceptible hosts. At the present time, there is no convincing indication that the majority of cases of ulcerative colitis or Crohn's disease is caused by persistent infection by Mycobacterium paratuberculosis, measles, Listeria monocytogenes, or Helicobacter species, but this possibility remains a valid hypothesis. Transient infection with any of a number of pathogens including upper respiratory tract infections and common enteric pathogens could provide one of the environmental triggers that initiate or reactivate IBD, which is then perpetuated in susceptible hosts by resident (not pathogenic) commensal luminal bacteria. Recent results in animal models demonstrate the absence of colitis, gastritis, and arthritis in a sterile (germ-free) environment, showing the importance of resident bacteria as persistent antigenic stimuli in the genetically susceptible hosts. Furthermore, there is an indication that not all normal luminal bacteria have equal capacities to induce mucosal injury, since some species can induce inflammation (Bacteroides), some are neutral (E. coli) and others may be protective (Lactobacilli). These observations have important therapeutic implications, such that altering luminal bacterial components and thereby decreasing the persistent antigenic drive offer alternative or adjuvant approaches to ongoing efforts to block mucosal immune responses to these stimuli.

Microflora, Lumen, Inflammation, Crohn's disease

Recent data in animal models have led to the widespread belief that luminal microbial constituents provide the chronic antigenic stimulus that perpetuates idiopathic inflammatory bowel diseases (IBDs) (^1,2). The unresolved issue is whether chronic inflammation is the result of injury induced by a persistent pathogenic organism, an overwhelming exposure to resident normal luminal bacterial products because of enhanced mucosal permeability, or an aberrant host immune response to normal luminal components (Table 1). This conceptual review will briefly summarize evidence that enteric pathogens or resident flora have an etiologic or contributory role in clinical and experimental chronic intestinal inflammation. The reader is referred to a number of recent summaries on this subject that comprehensively reference this rapidly advancing area of research (^3,–6).

Table 1.

Etiologic theories of IBD¹

Specific Pathogen

In the accompanying hypothesis article, Blaser eloquently states the case for a persistent pathogen causing IBD, using the humbling lessons learned from the role of Helicobacter pylori in the spectrum of peptic ulcer disease, chronic gastritis, and gastric malignancy (⁷). As correctly pointed out by Blaser, failure to identify an etiologic agent does not negate a possible role for a pathogen. Indeed, there are intriguing similarities between peptic ulcer disease and idiopathic IBD, such as familial associations, transient responses to nonspecific therapy, spontaneous relapses, and associated neoplasia. However, at the present time, there is no consistent evidence that IBD in humans is caused by a persistent microbial pathogen, despite numerous attempts to invoke an etiologic agent (³).

Recent studies have implicated Mycobacterium paratuberculosis, measles virus and Listeria monocytogenes in Crohn's disease (Table 1) (^8,–10). Although the latter two associations have not been reproduced by other groups (^9,11), M. paratuberculosis continues to stir controversy. The strongest support for M. paratuberculosis comes from the clinical overlap of Crohn's disease with Johne's disease, a chronic, spontaneously relapsing granulomatous enterocolitis in ruminants (¹²) and from detection of this organism in clinical tissues by polymerase chain reaction (PCR) (¹³). However, recent PCR results in Crohn's disease biopsy specimens have either been negative (¹⁴) or nonspecific (¹⁵), and careful studies have failed to detect the organism by immunohisto-chemistry, to establish an epidemiologic link, or to demonstrate an immune response to this agent (reviewed in ref. 8). The lack of an easily detectable immune response is particularly damaging, because the only way an extremely paucibacillary infection can cause disease is by initiating an injurious immune response. Thus, current data do not suggest that Crohn's disease is caused by M. paratuberculosis, although it is quite possible that this organism is far more prevalent in the environment than previously suspected (^15,16) and that it may secondarily invade the ulcerated mucosa of patients with Crohn's disease.

Although most investigators have studied pathogenic organisms in Crohn's disease, there is some evidence that ulcerative colitis is a consequence of functionally abnormal luminal bacteria that can cause disease by secretion of enterotoxins that increase epithelial cell permeability (¹⁷), by elaboration of immunosuppressive proteins that dysregulate the mucosal immune system (¹⁸), or by impairing epithelial cell metabolism (¹⁹). The latter observation has led Roediger et al. to postulate that ulcerative colitis is the result of epithelial starvation, caused by defective short-chain fatty acid metabolism (¹⁹). The distal colonocyte is particularly dependent on butyrate as a primary energy source, so that blockade of butyrate metabolism by hydrogen sulfide, liberated by sulfate-reducing luminal bacteria, could have detrimental effects on mucosal permeability and healing. This theory has received independent support by the observation that hydrogen sulfide and sulfate-reducing bacterial concentrations are increased in patients with ulcerative colitis (²⁰).

Together, these reports illustrate the complexities of colonic microecology, because subtle alterations in bacterial function can have profound implications for mucosal barrier function and immune responses, yet routine bacterial cultures would demonstrate only “normal” organisms. Clearly, we cannot exclude the possibility that an occult pathogen or functional abnormality of commensal bacteria could cause Crohn's disease or ulcerative colitis, and it is quite likely that transient pathogenic infections trigger relapses and perhaps initiate disease by breaking the mucosal barrier or activating the mucosal immune response in genetically susceptible hosts. This hypothesis is further supported by the observation that H. hepaticus colonization is widespread in murine colonies (²¹), raising the possibility that opportunistic infection or aberrant immune responses to this organism may be a factor in “spontaneous” colitis in genetically engineered mice (R. Sellon, S. Tonkonogy, and R. B. Sartor, unpublished observations).

Resident Luminal Bacteria

An alternative mechanism by which enteric microbial agents could stimulate chronic enterocolitis and related systemic inflammation is through activation of mucosal lamina propria macrophages and T cells by luminal bacterial constituents leaking across the epithelial barrier (Fig. 1). The colon and distal ileum contain high concentrations of predominantly anaerobic bacteria (10^11-12 and 10^8-9 colonies/g, respectively) and bacterial components such as peptidoglycan-polysaccharide (PG-PS), lipopolysaccharide (endotoxin), and formylated chemotactic peptides, which can activate intestinal macrophages, T lymphocytes, eosinophils, mesenchymal cells, endothelial cells, and even epithelial cells to secrete cytokines, eicosanoids, reactive oxygen metabolites, nitric oxide, and proteases, thereby inducing local damage (^1,3,22). This injury further enhances mucosal permeability, establishing a self-sustaining cycle of mucosal inflammation, uptake of luminal bacterial products, and translocation of viable bacteria. Systemic distribution of these bacterial components and locally produced inflammatory mediators through mesenteric lymphatics and the portal vein lead to extraintestinal manifestations of IBD and experimental colitis (^23,24). Thus, normal luminal bacterial components can lead to chronic intestinal inflammation in a susceptible host after initial mucosal permeability abnormalities from either environmental triggers or from an intrinsic defect in barrier function.

Fig. 1.

Hypothetical framework for the mucosal uptake and systemic distribution of luminal bacterial constituents that induce local and systemic inflammation. (Used with permission by publishers of ref. 24.)

Extensive evidence in clinical and experimental intestinal inflammation supports this hypothesis (Table 2). The colon is preferentially involved in mice with systemic immunoregulatory defects due to targeted deletion of interleukin-2 (IL-2), IL-10, or T-cell receptor α or β, or due to expression of human HLA-B₂₇/β2 microglobulin (^25,–28). The lack of inflammation in relatively sterile organs probably reflects the lack of chronic antigenic drive in these immunologically mediated inflammatory models (²). Furthermore, colonic inflammation does not develop in these models in the absence of luminal bacteria (^25,26,29,30). Antibiotic trials indicate that anaerobic bacteria have a dominant role in Crohn's disease (³¹) and experimental colitis, whereas aerobic flora may be more active in ulcerative colitis (³²). Reconstitution studies show that Bacteroides vulgatus is uniquely capable of inducing colitis in HLA-B₂₇ transgenic rats (²⁹) and guinea pigs fed carrageenan (³³), and establish the important concept that all resident luminal bacteria do not have equal capacities to induce or perpetuate intestinal inflammation. This point has considerable clinical relevance, and is in keeping with the ability of metronidazole to attenuate Crohn's disease by chronically suppressing fecal Bacteroides species without substantially affecting total luminal bacterial concentrations (³⁴). Very recent preliminary studies take advantage of the selective nature of bacterial stimulation by inhibiting recurrent inflammation in interleukin-10 (IL-10) knockout mice (³⁵) and refractory pouchitis after ileal pouch-anal anastomosis (M. Campieri, unpublished data) with Lactobacillus colonization.

Table 2.

Evidence that resident luminal bacterial stimulate experimental and human chronic intestinal inflammation (modified from reference 23)

Normal luminal bacteria probably induce and perpetuate intestinal inflammation by initiating an immune response. Chronic intestinal inflammation is immune mediated, as demonstrated by lack of inflammation in Tcell-deficient hosts and transfer of disease with CD4⁺ T lymphocytes (²). In preliminary studies, Cong et al. (³⁶) transferred colitis from C₃H/HeJBir mice to immunodeficient recipients with CD4⁺ lymphocytes activated in vitro by novel colonic bacteria, but not by equal numbers of unstimulated T cells. Both IBD patients and rodents with colitis have humoral and cellular immune responses to resident enteric bacteria (^3,36,–40). Provocative studies by Duchmann et al. (^38,39) suggest that clinical and experimental intestinal inflammation is associated with loss of tolerance to luminal bacteria, which would result in chronic stimulation of a potentially destructive immune response.

Factors that determine host susceptibility to luminal bacterial-induced injury include mucosal barrier function, epithelial restitution, immune responsiveness, and genetic predisposition (¹). Focal disruption of tight junctions in intestinal crypts by inactivating N-cadherin leads to focal inflammation (⁴¹). Similarly, deletion of intestinal trefoil factor inhibits mucosal healing and intensifies injury (⁴²). Targeted disruption of immunosuppressive cytokines, such as IL-10, leads to colitis in the presence of luminal bacteria, presumably due to unopposed activation of macrophages and TH₁ lymphocytes (⁴³). Finally, non-major histocompatibility complex (MHC)-regulated genes determine aggressiveness and chronicity of experimental colitis in response to bacterial antigen stimulation, as demonstrated by chronic granulomatous enterocolitis with fibrosis and systemic inflammation in susceptible inbred Lewis rats injected subserously with purified PG-PS polymers, but self-limited intestinal inflammation without systemic manifestations in MHC identical Fischer rats (⁴⁴).

Conclusions

These results clearly indicate that both pathogenic and normal enteric microflora can induce and perpetuate chronic intestinal inflammation in genetically susceptible hosts. At the present time, there is no indication that ulcerative colitis or Crohn's disease is caused by a persistent pathogen, but this possibility remains a valid hypothesis (⁷). It is likely that transient infection could provide one of the environmental triggers that initiate or reactivate IBD, which then is perpetuated in a susceptible host by resident (nonpathogenic) commensal luminal bacteria. The boundaries between pathogenic and commensal bacteria become obscure when the possibility that susceptible hosts can mount a pathogenic immune response to resident bacteria is considered. Thus, all bacteria could be considered to be potential pathogens. The intriguing question of why only a small subset of people exposed to the phlogistic anaerobic bacteria and bacterial antigens and cell wall polymers of the distal colonic microenvironment develop chronic intestinal inflammation probably holds the answer to understanding the pathogenesis of IBD. We suggest that the normal host does not respond to their native (autochthanous) flora because of efficient antigen exclusion by the intact mucosal barrier and a state of tolerance actively maintained by immunosuppressive cytokines (IL-10, IL-4, and transforming growth factor β), T lymphocytes, prostaglandins, and neuropeptides (Fig. 2). Thus, homeostasis is a state of “controlled intestinal inflammation.” However, this delicate balance can be perturbed by genetic defects in barrier function, mucosal healing and immunoregulation, and by environmental “triggers” that break the mucosal barrier, initiate nonspecific inflammation, or alter bacterial composition. In the absence of these genetic abnormalities, the normal (resistant) host has a remarkable capacity to rapidly heal mucosal injury without resident damage, despite the presence of proinflammatory bacterial products. Thus, chronic intestinal inflammation depends on the interaction of host genetic susceptibility, environmental triggers (including enteric infections), and the chronic antigenic drive of anaerobic luminal bacteria. This hypothesis has profound clinical implications, because, in theory, intestinal inflammation could be prevented and treated by elimination of the dominant antigens chronically stimulating the mucosal immune response. This approach should be aggressively investigated in parallel with ongoing progress in blocking the mucosal immune response to these stimuli.

Fig. 2.

Balance of proinflammatory luminal bacterial components and host protective forces, modified by environmental triggers and host genetic susceptibility factors. [Adapted from R. B. Sartor, Canadian Journal of Gastroenterology 4:271, 1990; used with permission.]

Acknowledgment

I thank Susie May for her diligent secretarial support, and the numerous collaborators who helped generate these concepts. Original investigations were supported by National Institutes of Health grants DK 40249, DK 34987, DK 47700, and DK 43735, and by the Crohn's and Colitis Foundation of America.