Unraveling Gut Inflammation

See allHide authors and affiliations

Science  25 Aug 2006:
Vol. 313, Issue 5790, pp. 1052-1054
DOI: 10.1126/science.1131997

Our intestinal mucosa, particularly that lining the large bowel and terminal ileum, is in constant contact with a resident microflora that is both luxuriant and complex: It consists of some 100 trillion discrete prokaryotic cells comprising an astonishing variety of Gram-positive and Gramnegative facultative aerobic and anaerobic organisms (1). Under normal circumstances, the presence of this vast biotic mass is useful and benign in that it helps prevent colonization of the bowel by pathogens and does not in itself evoke inflammatory immune responses (the normal mucosal immune system is tolerant of its many antigens). Nevertheless, this flora can also cause disease, because there is now good evidence that excessive mucosal immune responses to components of the microflora, either due to abnormal or impaired effector or regulatory (suppressor) cell activity of the host, is the prime cause of inflammatory bowel disease (IBD) (2). One way by which cells in the intestinal milieu may prevent this possible outcome is to apply a number of checks on the resident microflora that control its overall size and composition. On page 1126 of this issue, Cash et al. (3) characterize one such check—the ability of epithelial cells lining parts of the mucosa to produce RegIIIγ (regenerating islet 3 gamma), a C-type lectin with bactericidal properties. This substance may not only regulate the mix of intestinal organisms, but may also eliminate potential pathogens that cannot be controlled by the microflora alone.

Production of RegIIIγ was previously shown to be up-regulated in mouse intestinal cells after induction of mucosal inflammation and in patients with IBD (4). Cash et al. found that it is the product of Paneth cells, secretory epithelial cells at the base of the intestinal crypts in the terminal part of the small intestine (see the figure). However, these cells may not be the only site of production because RegIIIγ is also produced in the large bowel, which lacks Paneth cells (4). In addition, they found that RegIIIγ and its human homolog, HIP/PAP, bind to a carbohydrate component of peptidoglycan, a proteoglycan in the bacterial coat of most bacteria, particularly Grampositive bacteria. The authors observed that such binding leads to bacterial killing, thus making RegIIIγ one of a number of antibacterial (antimicrobial) substances produced by Paneth cells. This arsenal also includes lysozyme, secretory phospholipase A2, angiogenin 4, and α-defensins (5). That said, it may be that RegIIIγ fills a particular antibacterial niche because it targets intestinal Gram-positive bacteria such as Enterococcus fecalis rather than the more predominant intestinal Gramnegative bacteria.

To underscore the possible significance of RegIIIγ, Cash et al. indicate the importance of strict sequestration of resident microbial organisms, lest they cross into the host's interior milieu and cause inflammation and/or sepsis. They also point out that deficiencies in the production of antimicrobial peptides, the α-defensins, are associated with Crohn's disease, a form of IBD. Each of these points requires some discussion.

The authors are quite correct to point out that entry of bacteria into mucosa lying beneath the intestinal epithelium—the lamina propria of the gut wall—may lead to inflammation. This was shown in studies of transgenic mice that were genetically engineered to have faulty N-cadherin function, and thus created leaky “tight” junctions between epithelial cells in spatially defined areas of the intestine. This allowed bacteria in the mucosal microflora to breach the epithelial barrier and cause inflammation in affected areas, but not in areas where tight junctions were intact (6).

Despite these findings, it is now well established that resident flora do interact with antigen-presenting cells in the lamina propria and/or actually cross the epithelial barrier without causing inflammation. For instance, antigen-presenting dendritic cells in the lamina propria extend processes between epithelial cells and into the gut lumen. By this periscope-like activity, they take up bacteria or antigens associated with bacteria in the microbial flora (7, 8). In addition, it has been demonstrated that a commensal organism introduced into the gut lumen of a germ-free mouse gains rapid entry into the gut-associated tissue of the mouse by some as yet unidentified route (9). Whereas most bacteria thus entering the mucosa are killed by macrophages, some survive in dendritic cells for a limited time, where they are processed for stimulation of B lymphocytes. The B cells, in turn, produce noninflammatory immunoglobulin A (IgA) responses that can inhibit binding of microorganisms to ordinary (columnar) epithelial cells. The IgA responses also promote bacterial uptake across specialized cells (M cells) overlying Peyer's patches, the lymphoid follicles embedded in the mucosa. Finally, it has been observed that so-called probiotic organisms (commensal bacteria with anti-inflammatory properties) activate regulatory T lymphocytes when introduced into the lumen of mice that have the ability to inhibit subsequent induction of inflammatory responses (10). Evidently, these microorganisms gain access to mucosal lymphoid tissue (dendritic cells) or else the induction and expansion of the regulatory immune cells would not be possible.

The picture that emerges from these and other studies is that the epithelial barrier is actually quite porous and there is a constant interplay between resident microbial flora and mucosal immune elements and perhaps even limited entry of organism into the host mucosal system. However, why such interaction and entry do not provoke inflammation or is in fact anti-inflammatory whereas wholesale entry of organisms causes inflammation is not at all clear. One attractive possibility is that the development of inflammation depends on the degree to which bacterial entry is accompanied by robust cellular signaling via innate immune mechanisms, most notably signaling by bacterial components in the flora that function as ligands for Toll-like receptors (TLRs) expressed on immune cells. In the case of invading pathogens, such signaling leads to vigorous stimulation of antigen-presenting cells that initiate and sustain host-protective inflammatory responses. In contrast, in the case of commensals penetrating the epithelial barrier, such signaling may be weak and unable to incite inflammation, unless the organisms enter in great numbers. This explanation fits well with recent evidence that mutations in Card15 are a susceptibility factor in Crohn's disease (11). The Card15 gene encodes NOD2, an intracellular protein that recognizes a peptide component of peptidoglycan, the same bacterial component recognized by RegIIIγ. In mice, NOD2 functions mainly as a regulatory element that limits the TLR2 response to peptidoglycan (11). This strongly implies that Card15 mutations and consequent defective NOD2 function lead to overly robust responses to TLR2 ligands associated with commensal organisms that set the stage for the inflammation of Crohn's disease.

Antibacterial factors in mucosal homeostasis.

The secretion of Paneth cell factors (including RegIIIγ) into the intestinal crypts regulates the number and type of bacteria in the crypt space and gut lumen under normal conditions and during the introduction of a potential pathogen. Loss of Paneth cell function (including decreased α-defensin production) results in bacterial overgrowth. Inflammatory bowel disease (Crohn's Disease) involves a dysregulated mucosal immune response.


This conclusion brings us to the second point—the role of Paneth cell-derived antibacterial factors in the pathogenesis of IBD. Paneth cell production of α-defensin—5 is decreased in patients with Crohn's disease, whereas production of other antibacterial factors by these cells is not changed (12). Furthermore, the production of this defensin is particularly decreased in patients bearing a Card15 mutation, thus tying α-defensin production to a Crohn's disease susceptibility gene. Its relation to α-defensin production is probably due to NOD2 expression in both Paneth and dendritic cells; in the former, it is involved in inducing α-defensin production. Finally, in bacterial killing assays, extracts of intestinal tissue from patients with Crohn's disease and decreased α-defensin production, exhibited a somewhat lower capacity to kill Staphylococcus aureus and Escherichia coli in vitro. These and other data suggested that the α-defensin defect is not a secondary manifestation of the inflammation, and that abnormally low α-defensin production is a primary defect in Crohn's disease. This concept stands on its head the previously prevailing concept that IBD, including Crohn's disease, is due to a dysregulated and excessive response of mucosal immune cells to the resident microflora. Instead, it posits that the excessive response is, rather, a secondary event that follows from the primary alteration in the microflora related to the α-defensin deficiency.

Before we can accept this new theory of Crohn's disease pathogenesis, however, several other experimental and clinical observations have to be considered. The first is that because the mucosal immune system is in constant contact with the bacterial microflora in the normal state, it is not clear that bacterial overgrowth in the gut lumen or intestinal crypts would necessarily lead to mucosal inflammation. Second, whereas Crohn's disease of the ileum is associated with decreased α-defensin production, there is no evidence that a similar defect applies to the large intestine, where the disease occurs in many patients, even in those with NOD2 mutations. Third, and perhaps most important, there are several mouse models in which impaired Paneth cell function leads to abnormal secretion of all antibacterial factors, not only α-defensins. In one such model, mice have a mutation in the cystic fibrosis transmembrane conductance regulator gene and consequently manifest some of the features of cystic fibrosis, namely decreased fluid in secretions and insolubility of secreted mucus (13). As a result, such as mice have blocked intestinal crypts, resulting in severe intestinal bacterial overgrowth, a phenomenon also noted in patients with cystic fibrosis. Despite this defect, the mice do not develop chronic inflammation of the bowel wall that characterizes Crohn's disease; nor do patients with cystic fibrosis. Likewise, no bowel wall inflammation develops in other models of Paneth cell deficiency, including mice with matrilysin deficiency that cannot elaborate mature, active defensins (14).

The lack of support for the view that an α-defensin secretion abnormality is a primary factor in IBD pathogenesis raises the probability that the primary defect lies in the response of the mucosal immune system to antigens in the microflora rather than in the capacity of these antigens to evoke an inflammatory response. This doesn't mean that defensins and other antibacterial substances such as RegIIIγ play no role in IBD pathogenesis, because these substances can act as gate-keepers that to some extent control the ability of organisms in the microflora to move through the epithelial barrier and, in so doing, incite disease in an already dysregulated mucosal immune system.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
View Abstract

Stay Connected to Science

Navigate This Article