The emergence and application of novel molecular techniques over the last decade has provided a needed catalyst to studies of the pathogenesis of the chronic inflammatory bowel diseases (IBD): Crohn’s disease (CD) and ulcerative colitis (UC). Successful development of genetically engineered models of intestinal inflammation has not only provided insight into the dysregulation of the mucosal immune system characteristic of IBD but has also emphasised the critical and complex role of the bacterial flora in establishing and maintaining chronic intestinal inflammation.¹ These advances in understanding pathophysiology in turn have already led to novel therapeutic approaches.^2,3

However, it is in studies of human genetics that landmark progress has been made, widely recognised not only within gastroenterology but also by investigators in all complex diseases.⁴ Genome wide scanning led initially to the identification of a number of susceptibility loci, the statistical evidence for which satisfy stringent criteria for definite linkage.⁵ The subsequent detection of the NOD2/CARD15 gene^6–8 within the IBD1 linkage interval and the association of mutations within this gene with susceptibility to CD is widely regarded as the most stringent proof of principle for hypothesis free genome scanning in complex diseases.

In the time that has elapsed since the discovery of NOD2/CARD15, the contribution of this gene in determining susceptibility and disease behaviour in IBD has received detailed examination. It is now clear that NOD2/CARD15 mutations are associated with susceptibility to CD but not UC.⁶ However, the contribution is subject to considerable ethnic and even regional variation. Whereas mutations may be carried by up to 50% of central Europeans with CD,⁹ these mutations are not present in Japanese¹⁰ or Afro-American¹¹ patients. Even within Europe, there is considerable regional variation^12–15 and reported population attributable risks vary from 7.1% to 32%. Furthermore, there is heterogeneity within CD, and genotype-phenotype relationships clearly exist. Independent data suggest that NOD2/CARD15 variants are associated with early onset disease,^9,11 involvement of the terminal ileum,^16,17 and fibrostenosing disease,¹⁸ all phenotypic characteristics initially implicated in Crohn’s initial descriptions of regional enteritis.¹⁹

The most intriguing question that remains to be answered concerns the mechanism whereby mutations in the NOD2/CARD15 gene predispose towards the chronic intestinal inflammation characteristic of CD. Studies with respect to protein structure, expression, and function promise to provide the answers but the critical questions remain unresolved. The NOD2/CARD15 gene, initially described by Ogura and colleagues,²⁰ encodes a 1040 amino acid protein, a member of the Apaf1/CARD family of cytosolic proteins, involved in apoptosis (programmed cell death). NOD2/CARD15 has sequence homology with other family members,²¹ notably NOD1/CARD4, which itself is not associated with CD susceptibility.²² The gene comprises two N terminal caspase activation and recruitment domains (CARD), a nucleotide binding domain, and a C terminal sequence of leucine rich repeats (LRR). The majority of CD associated mutations directly affect the LRR, which is a motif common to bacterial resistance R proteins in plants and mammals, notably the Toll-like receptor family,²³ enabling recognition of pathogen associated microbial patterns (PAMPs). Following PAMP recognition, the NOD2/CARD15 proteins dimerise and an interaction with the serine-threonine kinase RICK in the cytosol occurs, triggering downstream nuclear factor κB (NFκB) activation.²⁰ The original expression studies had suggested that NOD2/CARD15 was expressed only in monocytes, and implicated the protein as an intracellular regulator of NFκB activity, sensitive to bacterial lipopolysaccharide (LPS), complementary to NOD1/CARD4.²⁰

In recent months strong scientific evidence has emerged to complement initial observations. It is now clear that NOD2/CARD15 expression occurs not only in monocytes but may also be induced in dendritic cells²⁴ and intestinal epithelial cells.²⁵ Furthermore, independent data suggest that the minimal bacterial motif recognised by NOD2/CARD15 may not be LPS, as initially suggested, but muramyl dipeptide (MDP),^26,27 a component of both Gram negative and positive bacterial cell walls. It has been a consistent and unexplained observation, if somewhat counterintuitive, that both common and rare CD associated variants of the NOD2/CARD15 gene result in reduced NFκB activity, although these data are from transfection experiments of NOD2/CARD15 gene constructs in embryonic kidney cells.^27,28 Conversely, the uncontrolled mucosal inflammation of Crohn’s disease is characterised by upregulation of NFκB activation.²⁹

In 2003, Hisamatsu et al provided perhaps the most elegant evidence to date, that CARD15/NOD2 may function as an antibacterial factor in CaCo2 intestinal epithelial cells.³⁰ Cells stably transfected with a wild-type CARD15/NOD2 gene construct were able to prevent invasion by Salmonella typhimurium. This protective effect was lost in cells transfected with gene constructs of mutant CARD15/NOD2. In the same issue of Gastroenterology, Rosenstiel et al also demonstrated that NOD2/CARD15 expression in intestinal epithelial cells might be upregulated by the proinflammatory cytokine tumour necrosis factor α (TNF-α).²⁵

Thus with NOD2/CARD15 identification, the emphasis in studies of IBD pathophysiology has shifted to investigations of the innate immune response. The story now develops further with data that suggest that NOD2/CARD15 may be expressed in the Paneth cells of the small intestine³¹, published in this issue of Gut[see page 1591]. Ogura et al have used the techniques of immunohistochemistry and reverse transcription-polymerase chain reaction to examine NOD2/CARD15 expression in the ileal or colonic tissue from IBD patients and controls. NOD2/CARD15 expression was found to be localised to Paneth cells, within the ileum, or metaplastic Paneth cells within the colon. Indeed, in a parallel paper in Gastroenterology, the same authors have extended these observations using in situ hybridisation and laser capture microdissection.³² They demonstrated that NOD2/CARD15 expression was enriched in crypts, compared with villi, and cells expressing NOD2/CARD15 also strongly expressed the proinflammatory cytokine TNF-α, itself a potent stimulus to NOD2/CARD15 expression. However, NOD2/CARD15 expression was not a feature of tissue macrophages in the intestine.

Paneth cells are specialised epithelial cells located mainly in the crypts of the small intestine, in close proximity to epithelial stem cells.³³ Paneth cells secrete antibacterial substances, initially located in granules within the cytosol, in response to prokaryotic rather than eukaryotic pathogens.³⁴ The main antimicrobial factors secreted by the Paneth cell include lysozyme, phospholipase A2, trypsin,³³ α-defensins,³⁴ and angiogenins.³⁵ In the current studies, NOD2/CARD15 expression was noted in close proximity to the secretory granules. Indeed, this close proximity prompts speculation that NOD2/CARD15 may be involved in degranulation and mediator released. It should be borne in mind that Paneth cell degranulation may be triggered not only by MDP but also by other bacterial components.³⁴ One may hypothesise that NOD2/CARD15 might therefore be one of a number of receptors involved in degranulation but functional data would not sustain the hypothesis that it is the sole regulator of this function.

There is increasing interest in the importance of members of the defensin family of molecules in regulating innate immune defences. The α-defensins, which are expressed in Paneth cell granules, are particularly pertinent to the present studies. These cationic cysteine rich peptides are synthesised and stored as precursor proteins,³⁶ and in the mouse release requires lysis of the prodefensin molecule by matrix metalloproteinase 7 (MMP7);³⁷ in humans, this lysis is thought to be mediated by a Paneth cell specific trypsin.³⁸ MMP7 deficient mice have decreased responses to bacterial infections although they do not exhibit chronic intestinal inflammation. It is intriguing to note recent provocative data which suggest that carriage of NOD2/CARD15 variants may be associated with reduced α-defensin release from Paneth cells in response to bacterial cell wall components (J Wehkamp, personal communication, Falk Symposium, Berlin, 2003). Could defective defensin release by the Paneth cell provide the missing link whereby reduced NOD2/CARD15 activity impair host defences to bacteria and underlie persistent intestinal inflammation? The interrelationship between NOD2/CARD15 genotype, NFκB activity, and Paneth cell secretions clearly bears detailed examination. It is worth mentioning in this context that the β-defensin 2 gene contains an NFκB binding site in the promoter region,³⁹ and although this defensin is not a component of Paneth cell granules, it is overexpressed in colonic CD.⁴⁰

In the present study in this issue of Gut,³¹ Ogura et al were unable to find consistent NOD2/CARD15 expression in colonic mucosa. Only one patient with colonic CD and concomitant Paneth cell metaplasia showed NOD2/CARD15 colonic expression.³² This predominant ileal localisation would explain the association of NOD2/CARD15 mutations with ileal disease. However, conflicting data have recently been published,⁴¹ and the issue remains to be resolved.

The functional role of the Paneth cell, initially identified by Joseph Paneth in Vienna in 1888, has remained unclear for more than 100 years. The recent data ever more strongly implicate the Paneth cell as a contributory factor in the innate immune response to bacteria. It is, of course, of great interest to speculate that this expression is critical in the pathogenesis of chronic CD and further data are eagerly awaited. Of particular interest will be the phenotypic and morphological characteristics of transgenic animals lacking the NOD2/CARD15 gene, and subsequent attempts to reintroduce NOD2/CARD15 protein into the Paneth cells of these animals.