Genetic Origin of IBD

Inflammatory Bowel Diseases (IBD) are diseases of unknown etiology. For this reason, it is necessary in clinical practice to exclude several alternative disorders before retaining the diagnosis of Crohńs Disease (CD) or Ulcerative Colitis (UC). 1 This lack of knowledge limits not only the diagnostic procedures but also the development of specific treatments. Indeed, without knowledge on causative factors, IBD are treated today by nonspecific anti-inflammatory or immunosuppressive agents. The main goal of geneticists is to try to fill this gap and to provide etiological data to the scientific community. From this point of view, recently obtained important results are considered to be a breakthrough toward a better understanding of IBD mechanisms. These advances allow us to formulate pathophysiologic hypotheses, which are now under investigation using functional models. As early as 1934, CD was recognized as a familial disorder and this observation was further confirmed by many groups in the middle of the 20th century. The proportion of familial aggregations was reported in the range of 8% to 10% for CD and 6% to 8% for UC (for review see 2). Interestingly, percentages of familial aggregations as high as 30% were shown in pediatric practice, suggesting more important familial factors in case of early age at onset. Familial aggregations often suggest that genetic factors play a role in a disease. However, environmental risk factors shared by family members (for example infectious agents), cannot be discarded. To resolve this problem, we recently analyzed the sibships with multiple affected siblings. They observed that the birth order of affected sibs is significantly different from a random distribution with a clustering of affected sibs. 3 This observation clearly demonstrates that environmental factors play a role in familial aggregations of the disease. Furthermore, it suggests that such a familial risk factor probably occurs during childhood and may thus contribute to the excess of familial aggregations in children. Twin studies tried to distinguish between the environmental and genetic components involved in IBD. Recently, Halfvarson et al extended the result previously obtained by Tysk on a Swedish twin registry with a mean observation time of about 30 years in healthy twins. 4 The reported concordance rates were 18.8% and 62.5% respectively for UC and CD monozygotic twins. These values demonstrate that IBD result from the interplay between genetic and environmental factors. In addition, they show that the genetic component is more important in CD than in UC. Interestingly, using the Vienna classification, a very good phenotype concordance was observed in CD monozygotic twins suggesting that the CD susceptibility gene is heterogeneous from one family to another one. 4 This observation confirms several reports of relatively good phenotypic concordance between family members. 5,–7 The analysis of the disease segregation in the pedigrees allowed inferring a genetic model of inheritance. Such segregation analyses were performed in the early 90's and argued for at least one gene with a strong contribution in disease development (also called major gene[s] ) in CD and UC. 8,–10 The mutated gene predisposing to CD was expected to have a frequency of about 1% in the general population and a recessive mode of inheritance. At the opposite, the UC mutated gene was calculated to be less frequent and with dominant properties. We know today that this point of view is too simple even it if contains a part of the truth as demonstrated by “recessive like” properties observed for CARD15 (see below). Indeed, further genome wide scans have demonstrated that it is unlikely to discover IBD genes with an attributable relative risk higher than 2. This finding clearly indicates that IBD are not Mendelian traits. At the contrary, they appear to result from the complex interplay between environmental risk factors and a relatively large number of genes, each of them having a limited individual impact (Fig. 1). Unifying model for diseases associated to CARD15. In non-mutated patients, CARD15 is activated by the muramyl dipeptide, a component of the bacterial cell wall. In Blau Syndrome, the mutation induces a self-activation of the protein occurring without any exogenous activator. In Crohn's Disease, it is speculated that a loss of inhibition can induce an activation of the CARD15 pathway. We have to keep in mind this first conclusion when considering the data provided by genetic laboratories. The discovery of one gene is not sufficient to resolve the problem of the genetic predisposition to IBD. As a result, the understanding of the disease mechanisms cannot be inferred by a single gene. Furthermore, etiological studies have to integrate genetic and epidemiological data if we want to clearly understand IBD and propose specific therapeutic approaches. On these bases, several groups tried to discover the IBD genes. Considering the phenotype corresponding to UC and CD, researchers started studying genes involved in the immune system such as HLA genes, TNF or its receptors, interleukins, ICAM1, vitamin D receptor, interferon, etc. Alternatively, genes playing a role in epithelial functions (ie, mucins), or cancer predisposition (ie, hMLH1) were also proposed as candidate genes. To date, the most striking reported association is between severe UC and the HLA DRB1*0103 allele. 11 Using an alternative approach, some geneticists screened the whole genome to get IBD genes. This strategy does not require any hypothesis on gene function and may allow discovering unexpected or even unknown genes as was the case for CARD. 15 Using this positional cloning approach, several groups in the word delineated regions on chromosomes 1, 5, 6p, 12, 14, 16p, 16q and 19 that certainly contain IBD loci. 12,–18 However, in most cases, the relevant genes are still unknown and the above mentioned regions of linkage have to be investigated further. On chromosomes 5 and 16, researchers were lucky enough to find an association between CD and genetic polymorphisms. Retrospectively, and considering the weight of genes in CD predisposition when compared with UC, it is not surprising that CD susceptibility genes were recognized first. On chromosome 5q, several genetic polymorphisms were associated with CD and more recently with UC. 19,20 Unfortunately, it was not possible to discriminate which of these polymorphisms carries the biologic effect. However, because the associated polymorphisms are localised in a cytokine cluster, a role in the regulation of one or more of these genes is expected. On chromosome arm 16q, genetic variations on NOD2 gene were reported by two independent groups and further confirmed by many additional investigators. 21,–27 Altogether, these data allow to firmly conclude that NOD2 (further re-called CARD15 by the HUGO international nomenclature committee) is one of the CD genes. Conversely, this gene has no role in UC. CARD15 is mutated in about 50% of CD patients and 20% of healthy controls. 28 Thus, mutations within the gene are neither necessary nor sufficient to develop the disease. This observation is in complete accordance with the retained complex genetic model for CD. For example, it has been calculated that CARD15 represent no more than 20% of CD genetic susceptibility. Thus the discovery of many additional IBD genes is still required before to fully understand the disease mechanisms. Interestingly, the proportion of CD patients carrying mutations on their two chromosomes (17%) is much higher than expected by chance. 28 This mutation dose-effect is reminiscent of the recessive model proposed by segregation analyses. Genetic studies performed on CARD15 were not limited by the demonstration that CARD15 is a CD predisposing gene. They also provided biologic hypotheses, which were in part tested in functional models. The sequence of the gene is similar as the sequence of the coded protein. By comparison with known sequences, Ogura et al recognized three different regions in the protein. 29 The C terminal part of the protein contains a Leucine Rich Repeat (LRR) domain, which is known to play a role in protein/protein interactions. In the middle part of the protein, a Nucleotide Binding Domain (NBD) seems to be involved in protein self-oligomerization. And the last part, the N terminal part of the protein contains two Caspase Recruitment Domains (CARDs) known to play a role in apoptosis and NF-kB activations pathways. Such a structure is reminiscent of very old proteins involved in host/pathogen interactions. For example, CARD15 resembles to the N gene involved in the resistance to tobacco mosaic virus. In tobacco, in case of virus infection, the activation of the N gene induces an apoptosis of the infected leaves and as a result protects the rest of the plant organism. This early observation indicated that CARD15 is certainly involved in innate immunity. The expression profile of the gene given by Northern blot analyses confirmed this point of view. The first experiments showed that CARD15 is mainly expressed at the basal level in phagocytes including monocytes, macrophages, dendritic cells, and polymorphonuclear cells. 30,31 More recently, Card15 was also detected in epithelial cells in vitro and in vivo. 31,–35 However, its expression is very low at basal level but increases in inflammatory conditions. These observations confirmed that CARD15 is involved in innate immunity. Lesage et al analyzed the mutational spectrum of the gene. They found at least 60 genetic polymorphisms within the gene including 30 of them with a putative functional effect. 28 The mutations are localized all along the gene excepting the CARD domains were no mutation was observed in the first domain and only 3 in the second domain. Considering that the CARDs are certainly the effectors of the protein, this first observation suggests that mutated proteins are deregulated rather than structurally inactive. The observation of many mutations, in the regulatory parts of the protein confirms this point of view and suggests that CD results of a lack of Card15 regulation. An excess of CD patients carrying two mutations (instead of a single gene variation) is widely recognized (see above) and it has been calculated that the risk to develop the disease in double dose mutation carriers is as high as 10-fold higher than the risk for people carrying only one mutation. 21,22,28 This dosage-effect is consistent with a loss of function model where the molecular defect of the gene lowers its efficiency. In summary, genetic data allow to infer that CD results, in some patients, of the deregulation of the Card15 protein. The mutated protein is expected to exhibit a loss of function of its regulatory part in presence of bacterial components present in the gut lumen. The above inferences, based on genetic data are mainly speculative but they have the advantage to provide hypotheses for further functional experiments. These functional studies were developed by several groups and are important advances toward the understanding of CD mechanisms. The major advance is certainly the discovery of the bacterial activator of Card15. 36,37 This activator is the muramyl dipeptide (MDP), a component of the peptidoglycan, which is a located in the bacterial wall of most bacteria. Today, CARD15/NOD2 together with CARD4/NOD1 is regarded as the intra-cytoplasmic counterpart of the Toll Like Receptor (TLR) pathway involved in pathogen associated molecular pathway (PAMP) recognition. 38,39 Even if the exact interplay between these two redundant pathways and the importance of each one in case of bacterial infection is still to be defined, this result is really exciting. The second major finding is that Card15 is able to activate the NF-kB pathway via a Rick/Rip2 interaction. 40 Because NF-kB is a key pathway in inflammatory response, Card15 can be considered as a pro-inflammatory molecule. Interestingly, by its CARDs, Card15 is expected to play a role not only in NF-kB activation but also in apoptosis pathways. However, little is known on the pro-apoptotic function of the protein and this lack of data makes difficult to definitively conclude on the exact functions of the gene. A crucial point is the understanding of the link between the known function of the protein (a pro-inflammatory role in presence of peptidoglycan) and CD. First of all, we have to keep in mind that the first experiments available today are rather rough, mainly based on transient transfections in embryonic cell lines and focusing on NF-kB activation. Many additional data on more sophisticated cell and animal models are required, and by consequence, our understanding of the disease mechanisms is mainly speculative. The current model proposed for CD is a defect of function model supported by the first experimental data consisting in transcient transfection experiments. 22 In this model, the mutated protein is unable to activate the NF-kB pathway in presence of MDP. 36 By extension, it is postulated that CD result of a lack of response to MDP from the bacterial flora present in the gut. However, it has been noted that such a system is unexpected for an inflammatory condition. In fact, it is difficult to conciliate the observed loss of pro-inflammatory function at the molecular level and the observed inflammation at the histologic and clinical level. Several hypotheses can be formulated to resolve this conflict. First, it has been suggested that the defect in the innate immunity may trigger alternative pathways of the adaptive immune system. Such a deviation from the innate to the adaptive immunity may be deleterious, especially in the digestive tract which is defined by an immune tolerance. According to different authors, several ways of activation of the adaptive immune system have been proposed and are discussed elsewhere. 41 Second, it has been shown that in some NF-kB knock out models, mice can develop a colitis. 42 Such a finding suggest that under certain circumstances (for example in case of chronic inflammation of the digestive tract), the NF-kB pathway can trigger anti-inflammatory properties. According to this hypothesis, it is thus possible to postulate that a loss of NF-kB activation observed for mutated Card15 proteins can contribute to an inflammation. Unfortunately, the excess of activated NF-kB reported in CD lesions do not support this hypothesis. Third, it is also possible to speculate that the observed defect in protein activation by MDP is not the relevant mechanism for CD. According to the mutational spectrum of the gene, a deregulation of the protein function instead of a complete loss of function is expected (see above). Thus, instead of an activation deficiency, a defect in protein inhibition can also be examined. Little is known about protein inhibitors but we can easily speculate that they can exist. In theory, CD can thus result of a defect of inhibition (by a yet unknown component) instead of a defect of activation (by MDP) (Fig. 1). This model is certainly highly speculative today but it has the advantage to conciliate genetic and biologic data. In addition it is supported by the comparison between CD and Blau Syndrome (BS). BS is a rare Mendelian dominant disorder characterized by a granulomatous inflammation in skin, eyes and joints but not in the gut. Miceli-Richard et al and Chamaillard et al were able to show that affected patients carry gain of function mutations in the NBD region of the protein, resulting in an excess of NF-kB activation, in the absence of any activator. Data are thus compatible with a self-activation of the protein (Fig. 1). The clinical and histologic resemblances between BS and CD argues for a common unified mechanism for these two inflammatory disorders. It can thus be postulated that BS and CD associated mutations finally result in an activation of NF-kB. From this point of view, the model of a loss of inhibition in CD seems the most pertinent (fig1). Interestingly, it is to note that according to this hypothesis, Card15 inhibitors have to be searched in bacterial components of the gut. Genetic studies performed in the last few years changed our knowledge on IBD. In fact, it may be seen as a paradox that these studies prompted researchers to focus on infectious agents which are environmental risk factors. However, this is the perfect illustration that IBD are complex genetic disorder and that the cause of the disease will be fully understood only by the integration of epidemiological and genetic data. The authors thank the fondation Jean Dausset and the Institut National pour la Santé et la Recherche Médicale (INSERM) for financial support.