Insight into the genetics and immunologic mechanisms CLINICAL GENOMICS. Celiac Disease Genetics: Current Concepts and Practical Applications

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2005;3: CLINICAL GENOMICS Celiac Disease Genetics: Current Concepts and Practical Applications LUDVIG M. SOLLID and BENEDICTE A. LIE Institute of Immunology, University of Oslo, Rikshospitalet University Hospital, Oslo, Norway Celiac disease is a multifactorial disease with complex genetics. Both HLA and non-hla genes contribute to the genetic component, but recent findings suggest that the importance of non-hla genes might have been overestimated. No susceptibility genes other than HLA-DQ have yet been identified in celiac disease. In contrast to the meager knowledge regarding non-hla genes, we have acquired a detailed understanding about which HLA genes are predisposing for disease, and how they are involved in the pathogenesis. This knowledge might pave the road for novel treatments for the disease. The role of HLA as a necessary, but not sufficient, genetic factor can moreover be used for diagnostic purposes to exclude a celiac disease diagnosis. The applicability of HLA genotyping is particularly useful for excluding celiac disease in family members or risk groups with fairly unbiased distribution of HLA alleles (ie, patients with Turner syndrome and patients with Down syndrome) and in patients with a clinical suspicion of celiac disease. Insight into the genetics and immunologic mechanisms of celiac disease has made huge progress in recent years. Celiac disease is perhaps the best understood of the HLA-associated disorders. We even have an understanding at an atomic level of how the predisposing HLA-DQ2 molecule is involved in disease development. This article will briefly review the current state of play in celiac disease genetics and pathogenesis, discuss the practical application of HLA typing in the clinic, and offer some future prospects of novel therapies and disease prevention. Molecular Basis of Celiac Disease Celiac Disease Is a Multifactorial Disorder With Complex Genetics The celiac enteropathy develops as a result of an interplay between genetic and environmental factors (Figure 1). Gluten (consisting of the gliadin and glutenin subcomponents) is clearly a critical environmental component, and both HLA and non-hla genes are thought to be predisposing genetic factors. The importance of genetic factors is illustrated by a high degree of familial clustering, with about 10% of first-degree relatives being affected, 1 and by the high concordance rate ( 75%) among monozygotic twins. 2 By using the method of Risch to assess the importance of HLA linked and unlinked determinants in disease, the importance of non- HLA genes has been estimated to be greater than that of HLA genes. 1,3 The method does not take into account the higher degree of shared environment among family members. This might lead to an overestimation of the effect of non-hla genes. Moreover, the findings of higher disease prevalence and thereby lower familiality further reduce the estimate of the contribution of non- HLA genes. Petronzelli et al, 3 by using the method of Risch in a meta-analysis, estimated the genetic effect attributable to HLA to be 36%. Repeating their analysis but using a population prevalence of 1:100 rather than 3:1000, the estimate of the HLA effect is 53%. These estimates are uncertain because the modeling of gene interaction is hypothetical. However, taken together with the results from genome-wide linkage scans, which generally report clear linkage to HLA and very weak linkage to other regions, the results suggest that HLA is not only the single most important genetic factor, but also that the effect of HLA alone might supersede the combined effects of the non-hla genes. HLA as a Necessary but not Sufficient Genetic Factor Overwhelming evidence pinpoints HLA-DQ as the chief locus mediating the HLA linked effect in celiac disease. The great majority of the patients carry a variant Abbreviations used in this paper: HLA, human leukocyte antigen; IELs, intraepithelial lymphocytes; IFN-, interferon- ; IL, interleukin; MHC, major histocompatibility complex; TG2, transglutaminase 2 or tissue transglutaminase; TNF, tumor necrosis factor by the American Gastroenterological Association /05/$30.00 PII: /S (05)00532-X

2 844 SOLLID AND LIE CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 9 Figure 1. Schematic depiction of the multifactorial etiology of celiac disease. Environmental components (including gluten) and several genetic factors (including HLA) are involved in the development of celiac disease. There are interactions between genetic and environmental factors (eg, HLA and gluten), and there might also be interactions between the various predisposing genetic factors (ie, epistasis) and environmental risk factors. of DQ2 (DQA1*05/DQB1*02), and a minority of the patients carry DQ8 (DQA1*03/DQB1*0302) (Figure 2). The celiac disease associated DQ2 heterodimer can be encoded in cis by DQ alleles of the DR3-DQ2 haplotype or in trans by the DR5-DQ7 and DR7-DQ2 haplotypes (ie, DR5/DR7 heterozygotes; of note, DQ2 of the DR7- DQ2 haplotype [DQA1*0201/DQB1*0202] has a different DQA1 allele than that of the DR3-DQ2 haplotype) (Figure 2). The fact that the same DQ2 molecule (DQA1*05/DQB1*02) is expressed by celiac disease patients in cis or trans configuration pinpoints DQ2 as the culprit molecule. Compared with other HLA-associated diseases, the evidence for the primary HLA association is very strong. Individuals who carry a double dose of DQB1*02 (ie, DR3-DQ2 homozygous or DR3-DQ2/ DR7-DQ2 heterozygous) have a particularly increased risk for celiac disease. Those celiac disease patients who are neither DQ2 (DQA1*05/DQB1*02) nor DQ8 (DQA1*03/DQB1*0302) carry DQB1*02 together with DQA1 alleles other than DQA1*05, or they carry DQA1*05 together with DQB1 alleles other than DQB1*02. In other words, they carry genes coding for only one of the chains of the DQ2 (DQA1*05/DQB1*02) encoded heterodimer. 4 The fact that almost all celiac disease patients are either DQ2 (DQA1*05/DQB1*02)or DQ8 (DQA1*03/DQB1*0302) indicates that HLA is a necessary but not sufficient factor for celiac disease development. HLA Molecules: The Carte du Jour for T Cells The role of the HLA molecules is to bind and present peptide fragments to T cells. The HLA class II molecules (of the DR, DQ, and DP series) present peptides to CD4 T-helper cells, whereas the HLA class I molecules (of the A, B, and C series) present peptides to CD8 cytotoxic T cells. T cells recognize peptides in the context of HLA molecules; hence HLA molecules represent the menu list of antigenic peptides available for T-cell recognition. There is tremendous polymorphism in the HLA system. Polymorphic HLA variants exhibit different form and chemistry of their peptide-binding sites. Different HLA molecules thus present different sets of peptides. Gluten-reactive T cells of the celiac intestinal mucosa uniquely recognize gluten peptides in the context of DQ2 and DQ8, suggesting that these HLA molecules predispose to celiac disease by preferentially presenting gluten peptides to CD4 T cells. 5,6 The x-ray crystal structure of DQ2 complexed with gliadin peptides provides an atomic explanation of why DQ2 is capable of binding certain gluten peptides with high affinity. 7 The reason why other HLA molecules do not predispose to celiac disease is most probably because they are unable to efficiently present to T cells those gluten peptides present in the gut microenvironment. CD4 T Cells as Director of the Immunologic Orchestra Responding to Gluten The gluten-reactive T-helper cells become activated upon recognition of gluten peptides and produce many cytokines of which interferon- (IFN- ) is domi- Figure 2. HLA association in celiac disease. The great majority of the patients carry the DQA1*05 and DQB1*02 alleles located in cis on the DR3-DQ2 haplotype or in trans on the DR5-DQ7 and DR7-DQ2 haplotypes. A minority of the patients carry DQA1*03 and DQB1*0302 encoded by the DR4-DQ8 haplotype. The DQ chains encoded by DQA1*0501 and DQA1*0505 differ by one residue in the leader peptide, and the DQ chains encoded by DQB1*0201 and DQB1*0202 differ by one residue in the membrane proximal domain. It is unlikely that these differences have any functional consequence. Notably, the DR3 allele of the DR3-DQ2 is also termed DR17, and DR5 includes the variants DR11 and DR12.

3 September 2005 CELIAC DISEASE 845 nant. 8 Gluten might also stimulate the innate part the immune system. This immune stimulation and the activation of the CD4 gluten-reactive T cells probably kick off an array of inflammatory reactions that finally result in the lesion typical of celiac disease. The role of HLA as a necessary but not sufficient contributing factor for celiac disease development suggests that activation of gluten-reactive, DQ-restricted CD4 T cells within the intestinal mucosa must somehow control all parts of the immune response that leads to disease development. Gluten T-Cell Epitopes Several distinct gluten T-cell epitopes exist. A common feature among these epitopes is the presence of multiple proline and glutamine residues, which gives rise to unique structural and functional properties. The peptides are exceptionally resistant to proteolysis by gastric, pancreatic, and intestinal digestive proteases because of their high proline content. 9,10 As a result, a high intestinal concentration of potentially immunoreactive peptides is maintained when following a gluten-containing diet. Selected glutamine residues in these gluten peptides become, as a function of the spacing between the glutamine and proline residues, converted to glutamate (they become deamidated) by tissue transglutaminase (transglutaminase 2 or TG2), leading to enhanced peptide immunogenicity Typically, gluten peptides bind to the DQ2 and DQ8 molecules so that glutamate residues created by deamidation are accommodated in pockets of the binding site that have a preference for negatively charged side chains. 7,14,15 A 33-mer peptide fragment of -gliadin, which is a product of normal gastrointestinal digestion and which is extremely resistant to further proteolytic degradation, has been found to be particularly antigenic. 10,16 These peptide fragments contain 6 partly overlapping copies of 3 different DQ2-restricted T-cell epitopes. Gluten Stimulation of the Innate Immunity In addition to activating the adaptive immunity, gluten also seems to stimulate the innate branch of the immune system. Part of this response involves an increased expression of zonulin, which results in increased epithelial permeability. 17 It is still not fully clear which part of gluten stimulates the innate immune system. Peptide of a particular -gliadin induces rapid activation of factors in the innate immune system in biopsies of treated celiac disease patients. 18 At least some of these innate effects of gluten seem to be relayed through the cytokine interleukin (IL)-15. The expression of IL-15 increases after in vitro challenge of celiac biopsies with the peptide, as well as with peptic/ tryptic digests of gluten and some other (but not all) gluten peptides. 18,19 In active celiac disease, there is increased expression of IL-15, both in the lamina propria and in the epithelium. 20,21 Enterocyte bound IL-15 can activate and expand intraepithelial lymphocytes (IELs). 21 It was recently demonstrated that IL-15 can induce killing of enterocytes via aberrant expression of the major histocompatibility complex (MHC) class I related molecule MIC on enterocytes and increased expression of the MIC receptor NKG2D on IELs. 19,22 Strikingly, the innate effect of gluten in biopsies is only seen in celiac disease patients, 18,19 questioning whether activation of the innate immunity somehow is linked with the activation of the adaptive immune system. Other Predisposing Genes Within and Outside HLA Gene Complex No genes, in addition to the DQA1 and DQB1 genes known to predispose to celiac disease, have yet been identified. Even though the role of genetic factors outside HLA-DQ might have been overestimated, it still seems likely that such factors exist. The difference in concordance rates among monozygotic twins ( 75%) and HLA identical sibs ( 30%), the fact that only a minor fraction of individuals who carry DQ2 (DQA1*05/ DQB1*02) ever develop celiac disease, and the gender bias with a female to male ratio of approximately 2:1 23,24 support this notion. The model of Risch estimating the contribution of HLA linked and unlinked genes in celiac disease includes all genes in the HLA region and not only the DQ genes as part of the HLA linked effect. In fact, it might well be that one or more genetic factors in addition to DQ are encoded in the HLA gene complex. The so-called classic HLA complex on chromosome 6p21 spans 3.6 Mb. 25 Lately, as a result of the findings of MHC-related genes outside these borders, the gene complex has been extended to cover 7.6 Mb. 26 About 28% of the 252 expressed transcripts from the extended HLA complex appear to be involved in the immune response. 26 There is a high degree of linkage disequilibrium (non-random association of alleles at neighboring loci) in the HLA complex. Interwoven function might have been an evolutionary force to keep genetic variants within the gene complex together, and this supports the possibility of finding other celiac genes within the HLA complex. Linkage disequilibrium, however, creates a problem for the identification of susceptibility genes. Polymorphisms localized in the vicinity of the risk DQ genes will demonstrate secondary associations caused by hitchhiking effects, unless careful precautions are made to control for this. Matching cases and controls only for DQ2 positivity (ie, DQA1*05 and DQB1*02 or only

4 846 SOLLID AND LIE CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 9 DQB1*02) is insufficient, because many patients are DR3-DQ2 homozygous or DR3-DQ2/DR7-DQ2 heterozygous. Many articles have inadequately adjusted for linkage disequilibrium and reported results that might well be false-positive findings. Linkage disequilibrium can be controlled for by restricting the analyses to certain DQA1-DQB1 haplotypes. Initially, Karell et al 27 recognized that not all DQA1*05-DQB1*02 haplotypes conferred the same risk for celiac disease. Subsequent studies appropriately controlling for linkage disequilibrium have corroborated that there is an increased risk associated with the B8-DR3-DQ2 haplotype The effects are small, however, and it will be extremely difficult to pinpoint which gene(s) is responsible for the effect by genetic tools. Some help might come from functional models. The genes of MIC (ie, MICA and MICB) and tumor necrosis factor (TNF) are all localized within the region associated with increased risk. These are especially interesting candidates because there is evidence for involvement of their gene products in the celiac disease pathogenesis. MIC molecules are central for the enterocyte killing in celiac disease, 19,22 and TNF is up-regulated in the gluten-induced immune response. 32 Polymorphisms of the TNF promoter (TNF-308A), initially reported by McManus et al 33 to be associated with celiac disease and later confirmed in a study rigorously controlling for linkage disequilibrium, 34,35 is a particularly attractive candidate. The problem is that even though this polymorphism is associated with increased TNF- expression 36 and celiac disease, it could still just be a marker for the true etiologic polymorphisms carried elsewhere on the B8-DR3-DQ2 haplotype, as a result of linkage disequilibrium. To find non-hla genes that predispose to celiac disease, several genome-wide screens have been performed Altogether, these studies have pointed out chromosomal regions that potentially could contain susceptibility gene(s). Several regions are believed to be false positive, and many true regions are likely to have been overlooked. With the exception of HLA genes, there is relatively little consensus between the results, pointing in the direction of modest effects conferred by each of the non-hla genes. Factors such as gene-gene interactions, genetic/allelic heterogeneity between populations, and limited sample sizes studied could also add to the complexity in uncovering the predisposing genes. The region that has most consistently been linked to celiac disease lies on the long arm of chromosome 5 (5q31-33). 40,41,46 In meta-analysis of results from several European populations, this region reached genome-wide significance, leaving little doubt that there exists a susceptibility gene here. 47 Although evidence for other regions is less consistent, there is accumulating support for susceptibility loci on chromosome 11q 38 and on chromosome 19p To date, 4 regions have received status as susceptibility regions, ie, CELIAC1 (HLA-DQ), CELIAC2 (5q31-33), CELIAC3 (2q33), and CELIAC4 (19p13.11). So far, aside from HLA, only for the CELIAC3 region has a likely candidate gene been postulated, namely the gene encoding the negative co-stimulatory molecule CTLA It is conceivable that non-hla genes involved in celiac disease constitute background genes that are tuning the immune response to gluten in such a way that celiac disease becomes the end result. Both susceptibility and protective genes might exist. In addition, there might be gene-gene interactions, epistasis, so that the effect of one polymorphism is influenced by polymorphisms of other genes. Epistasis will severely complicate the identification of celiac genes in out-bred human populations because simple genotype-phenotype correlations are less likely to exist. Genetic Diagnostic Tests The final diagnosis of celiac disease is based on the findings of typical histologic alterations of the small intestinal mucosa of patients eating gluten. A clinical response on a gluten exclusion diet is also considered to be of importance. Serologic tests, particularly measuring IgA antibodies to TG2, are very useful tools to predict the disease and to select patients for subsequent endoscopic examination. Recent recommendations also include HLA genotyping as a diagnostic adjunct. 56 In this respect it is useful to know what the performances of HLA genotype tests are, and in what situations these tests are useful. We have illustrated this by calculating the predictive information obtained from HLA genotyping in various settings in the Scandinavian population (Table 1). The performances will be slightly different in populations with different distributions of HLA alleles. We have considered 3 groups with variable prevalence of celiac disease: (1) the general population with an assumed prevalence of 1:100, (2) a risk group with a prevalence of 1:10, eg, patients with Down syndrome (5% 12%), and finally, (3) a high-prevalence group including patients with clinical symptoms or laboratory results compatible with celiac disease. Regardless of the settings, the specificity and positive predictive value of HLA testing are poor (Table 1). Likewise, the proportion of individuals with false-positive results is very large (Figure 3). The reason why HLA genotyping is almost worthless to predict celiac disease is the high frequencies

5 September 2005 CELIAC DISEASE 847 Table 1. Illustration of the Performance of HLA-DQ Genotyping as a Diagnostic Test for Celiac Disease in Scandinavia Prevalence a 1:100 General population 1:10 Risk group b 1:2 Selected patient group (clinical suspicion) HLA genotyping test c Sens (%) Spec (%) Predictive values (%) Predictive values (%) Predictive values (%) DQB1* PPV, 2.7 PPV, 22.8 PPV, 72.6 NPV, 99.9 NPV, 98.9 NPV, 90.6 DQ2 heterodimer d PPV, 3.4 PPV, 27.9 PPV, 77.7 NPV, 99.9 NPV, 98.7 NPV, 89.7 DQ2 heterodimer or DQ PPV, 1.7 PPV, 15.9 PPV, 63.1 NPV, 99.9 NPV, 99.5 NPV, 95.9 DQ2 heterodimer or DQ8 or one half of the DQ2 heterodimer PPV, 1.5 PPV, 13.9 PPV, 59.9 NPV, 100 NPV, 100 NPV, 100 Sens, Sensitivity, the fraction of celiac disease patients with a positive test result; Spec, Specificity, the fraction of non-celiac individuals with a negative result; PPV, positive predictive value, the probability that the patient has the disease, given a positive test; NPV, negative predictive value, the probability that the patient does not have the disease, given a negative test. a The prevalences in each group are rough estimates from the literature. b These calculations will apply for situations where the distribution of the HLA-DQ genotypes does not deviate from that of the general population (eg, Down syndrome patients). In risk groups where the distribution of HLA-DQ genotypes is biased (eg, patients with type I diabetes or IgA deficiency), the specificity and predictive values of the tests will be different. c HLA genotype frequencies of patients and controls based on typing of 225 Swedish-Norwegian celiac disease patients (probands of simplex families) diagnosed by ESPGHAN criteria (Louka et al, unpublished) and 361 blood donors from the Norwegian Bone Marrow Donor Registry. In populations with different frequencies of HLA alleles among controls and celiac patients, the performance of the HLA test will differ. d DQ2 heterodimer denotes DQA1*05/DQB1*02. of the DQ risk alleles among unaffected individuals. In Europe about 25% of the general population is positive for the DQ2 heterodimer (and approximately 50% if DQ8 is also included), although the vast majority of these people will never develop celiac disease. Even when the a priori chance of diagnosing celiac disease is 50%, positive HLA tests will not increase the likelihood for disease very much. By contrast, the negative predictive values obtained from HLA-DQ testing are impressive (Table 1), and the proportion of false negatives is small (Figure 3). Therefore, HLA testing is useful to rule out celiac disease. Repeated antibody testing has been recommended to follow high-risk individuals such as close family members and patients who are affected by diseases associated with celiac disease such as type 1 diabetes, Turner syndrome, and Down syndrome. A negative test for the HLA risk alleles renders celiac disease highly unlikely, and further serologic testing of these individuals is unwarranted. 56 It should be emphasized, however, that the use of HLA genotyping has limitations when the background frequencies of the HLA-DQ risk alleles in the test group are increased. This is clearly so for type 1 diabetes (associated with DQ8 and DQ2), IgA deficiency (associated with DQ2), Sjögren s syndrome (associated with DQ2), and Graves disease (associated with DQ2). In accordance with this, a recent Italian study found that the distribution of the DQA1 and DQB1 alleles did not discriminate between the type 1 diabetes patients with or without celiac disease and concluded that HLA typing is of limited use in this setting. 57 The family situation is a special case, and we argue that HLA testing is useful in this setting, because the presence of the risk factor is more directly linked with the disease risk. Several HLA alleles can be included in the genotyping, but notably, typing only for DQB1*02 will provide substantial information. Typing for the DQ2 heterodimer (ie, both DQA1*05 and DQB1*02) compared with typing for DQB1*02 does not provide anything extra for the exclusion of celiac disease except that it reduces the number of false positives (ie, more subjects to follow up). Some celiac disease patients are negative for DQ2 (10% 20%), and such individuals would be falsely acquitted of the diagnosis if typing only for DQB1*02 (or DQA1*05/ DQB1*02). Many of these individuals are positive for DQ8, 58 and for this reason it has been suggested to include typing for DQ8 (ie, DQB1*0302) in a test to increase the negative predictive values. 59 There also exist some few celiac disease patients who are negative for the DQ2 heterodimer and for DQ8, and these patients almost always carry one of the alleles encoding one half of the DQ2 heterodimer (ie, DQA1*05 or DQB1*02). 4 To really be sure to exclude celiac disease, the test subjects should be typed negative for DQA*05, DQB1*02, and DQB1*0302 (Table 1), but this will then be at the cost of including more false positives (Figure 3).

6 848 SOLLID AND LIE CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 9 Figure 3. HLA genotyping test in relation to celiac disease. The distribution of positive and negative test results among celiac disease patients and unaffected individuals from 4 different HLA genotyping tests: (A) scoring for presence of DQB1*02, (B) scoring for presence of the DQ2 heterodimer (DQA1*05 and DQB1*02), (C) scoring for presence of the DQ2 heterodimer or DQ8, and (D) scoring for presence of the DQ2 heterodimer, DQ8, or one of the alleles of the DQ2 heterodimer (DQA1*05 or DQB1*02). Informing an individual that she/he is the carrier of predisposing genes has an ethical aspect, and the possibility of a psychological impact of an HLA test in relation to celiac disease must be considered. Facing these challenges, the doctors should be careful to convey the message to their patients that the presence of celiac disease predisposing HLA alleles means little, whereas the absence of the same HLA alleles makes celiac disease an improbable disorder for that patient. Future Therapy and Disease Prevention Many patients cope with the gluten-free diet easily. Others find that the dietary restrictions are laborious, negatively influencing their quality of life. Compliance with the gluten-free diet is often incomplete among celiac disease patients, and there is a need for alternative treatments. Encouragingly, the new insight into the molecular mechanisms of celiac disease has uncovered novel targets for therapy. 60 The activation of the CD4 gluten-reactive T cells seems to be a critical checkpoint in the development of celiac disease. Interference with this step should thus be a way to control the disease. One possibility, which is basically an extension of today s treatment with a glutenfree diet, is to produce cereals with bread-making properties that are devoid of T-cell epitopes, either by breeding programs or transgenic technology. Blocking the binding sites of the DQ2 and DQ8 HLA molecules would prevent the presentation of diseaseinducing gluten peptides and thereby activation of T cells. This would be a treatment for which few side effects can be envisaged. The concept of HLA blockade is not new and was developed without much success for the treatment of type 1 diabetes and rheumatoid arthritis. The lack of success was partly due to difficulties in obtaining effective drug delivery. This should be less of a problem in celiac disease because the blocking compound can be administered locally in the affected organ before or in parallel with the antigen (ie, gluten).

7 September 2005 CELIAC DISEASE 849 Another possibility is enzyme supplementation with the aim to either destroy T-cell epitopes directly or to facilitate their gastrointestinal proteolysis. 9,10 Prolyl endopeptidases are particularly attractive enzymes because they will target the proline-rich regions of gluten that harbor the T-cell epitopes. TG2 is a target for intervention because of the critical role it plays in generating gluten T-cell epitopes. Analogs of gluten peptide containing warheads that function as irreversible inhibitors of human TG2 have been designed. 61 One possible problem with this approach is that TG2 inhibitors might have unacceptable side effects, because TG2 is involved in many different physiologic processes including programmed cell death. 62 Also, TG2 inhibitors might not be specific for TG2 but might affect the function of other transglutaminases, of which there is an entire family. Permanent elimination or silencing of gluten-reactive T cells is a potential way to treat celiac disease. This effect could be obtained by performing oral gluten challenge concomitantly with the administration of agents that alter the outcome of the T-cell activation. Antibodies to CD3 63 and CD154 (CD40L), 64 for example, can induce T-cell silencing. However, such antibodies produce unwanted side effects such as toxic cytokine syndrome (anti-cd3) and thromboembolic events (anti- CD154). Targeting gluten epitopes to dendritic cells that induce T-cell tolerance could also be a way to achieve this, but the fact that there exist multiple T-cell epitopes complicates this approach. Moreover, effective methods to target tolerogenic dendritic cells are not established yet. If successful, the T-cell elimination or the T-cell silencing treatment would probably have to be repeated over time, because new T cells are continuously generated in the thymus, and some of these might become gluten-reactive T cells in the gut. Control of celiac disease by disease prevention is the ultimate goal of the future. How this should be achieved is currently science fiction, but some concepts merit attention. The introduction of gluten in the infant feeding is relevant for subsequent development of celiac disease. This is illustrated by experiences from Sweden where there was an epidemic of celiac disease with a 3-fold increase in incidence between 1985 and 1995 among children younger than 2 years. 65,66 The sharp rise and subsequent abrupt decrease in incidence were likely related to changes in infant feeding habits. An increased risk for celiac disease was associated with a higher proportion of infants being introduced to gluten when breast-feeding had been discontinued. 66,67 It is still unclear whether compliance with feeding habits to reduce the risk for celiac disease will lead to a lifetime reduced risk, and whether these experiences from Sweden are directly transferable to other societies. The pattern of breast-feeding, for instance, shows great variation between countries. The concept that the amount of gluten might matter has, however, received support from in vitro gluten stimulation of T cells. 68 It might be that permanently reduced gluten consumption will lead to a lower lifetime risk for celiac disease. If wheat grains with a low content of harmful sequences can be produced, and the production yield on a large scale is economically acceptable, this might be a viable option that can be achieved on a population basis. This scenario, however, warrants some words of caution. The change in infant feeding habits in Sweden, leading to the celiac disease epidemic, was caused in part by revised national health recommendations in The recommended introduction of gluten was changed from 4 to 6 months, with the result that more infants were introduced to gluten without ongoing breast-feeding. At the same time, the content of gluten in baby foods was increased. The revised recommendation was mainly motivated to avoid severe celiac disease among young children. In 1996 the national recommendation was revised again, now recommending gradual introduction of gluten from the age of 4 months, preferably while still breast-feeding. The incidence of celiac disease among children younger than the age of 2 years dropped dramatically at about the same time. The celiac epidemic in Sweden is probably more complex than just being the outcome of changed feeding habits as a result of changed national recommendations, but it illustrates the point that without fully sufficient insight, well-intended advice might have adverse and unintended consequences. This experience should serve as a booster for the motivation to obtain further insight into this complex disorder. References 1. Risch N. Assessing the role of HLA-linked and unlinked determinants of disease. Am J Hum Genet 1987;40: Greco L, Romino R, Coto I, et al. The first large population based twin study of coeliac disease. Gut 2002;50: Petronzelli F, Bonamico M, Ferrante P, et al. Genetic contribution of the HLA region to the familial clustering of coeliac disease. Ann Hum Genet 1997;61: Karell K, Louka AS, Moodie SJ, et al. HLA types in celiac disease patients not carrying the DQA1*05-DQB1*02 (DQ2) heterodimer: results from the European Genetics Cluster on Celiac Disease. Hum Immunol 2003;64: Lundin KEA, Scott H, Hansen T, et al. Gliadin-specific, HLA-DQ ( 1*0501, 1*0201) restricted T cells isolated from the small intestinal mucosa of celiac disease patients. J Exp Med 1993; 178: Lundin KEA, Scott H, Fausa O, et al. T cells from the small intestinal mucosa of a DR4, DQ7/DR4, DQ8 celiac disease pa-

8 850 SOLLID AND LIE CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 9 tient preferentially recognize gliadin when presented by DQ8. Hum Immunol 1994;41: Kim CY, Quarsten H, Bergseng E, et al. Structural basis for HLA-DQ2-mediated presentation of gluten epitopes in celiac disease. Proc Natl Acad Sci U S A 2004;101: Nilsen EM, Lundin KEA, Krajci P, et al. Gluten specific, HLA-DQ restricted T cells from coeliac mucosa produce cytokines with Th1 or Th0 profile dominated by interferon gamma. Gut 1995;37: Hausch F, Shan L, Santiago NA, et al. Intestinal digestive resistance of immunodominant gliadin peptides. Am J Physiol Gastrointest Liver Physiol 2002;283:G996 G Shan L, Molberg Ø, Parrot I, et al. Structural basis for gluten intolerance in celiac sprue. Science 2002;297: Molberg Ø, McAdam SN, Körner R, et al. Tissue transglutaminase selectively modifies gliadin peptides that are recognized by gutderived T cells. Nat Med 1998;4: Vader LW, de Ru A, van de Wal Y, et al. Specificity of tissue transglutaminase explains cereal toxicity in celiac disease. J Exp Med 2002;195: Fleckenstein B, Molberg Ø, Qiao SW, et al. Gliadin T cell epitope selection by tissue transglutaminase in celiac disease: role of enzyme specificity and ph influence on the transamidation versus deamidation process. J Biol Chem 2002;277: Johansen BH, Vartdal F, Eriksen JA, et al. Identification of a putative motif for binding of peptides to HLA-DQ2. Int Immunol 1996;8: van de Wal Y, Kooy YMC, Drijfhout JW, et al. Peptide binding characteristics of the coeliac disease-associated DQ( 1*0501, 1*0201) molecule. Immunogenetics 1996;44: Qiao SW, Bergseng E, Molberg Ø, et al. Antigen presentation to celiac lesion-derived T cells of a 33-mer gliadin peptide naturally formed by gastrointestinal digestion. J Immunol 2004;173: Clemente MG, De Virgiliis S, Kang JS, et al. Early effects of gliadin on enterocyte intracellular signalling involved in intestinal barrier function. Gut 2003;52: Maiuri L, Ciacci C, Ricciardelli I, et al. Association between innate response to gliadin and activation of pathogenic T cells in coeliac disease. Lancet 2003;362: Hüe S, Mention JJ, Monteiro RC, et al. A direct role for NKG2D/ MICA interaction in villous atrophy during celiac disease. Immunity 2004;21: Maiuri L, Ciacci C, Auricchio S, et al. Interleukin 15 mediates epithelial changes in celiac disease. Gastroenterology 2000; 119: Mention JJ, Ben Ahmed M, Begue B, et al. Interleukin 15: a key to disrupted intraepithelial lymphocyte homeostasis and lymphomagenesis in celiac disease. Gastroenterology 2003;125: Meresse B, Chen Z, Ciszewski C, et al. Coordinated induction by IL15 of a TCR-independent NKG2D signaling pathway converts CTL into lymphokine-activated killer cells in celiac disease. Immunity 2004;21: Ciacci C, Cirillo M, Sollazzo R, et al. Gender and clinical presentation in adult celiac disease. Scand J Gastroenterol 1995;30: Ivarsson A, Persson LA, Nystrom L, et al. The Swedish coeliac disease epidemic with a prevailing twofold higher risk in girls compared to boys may reflect gender specific risk factors. Eur J Epidemiol 2003;18: The MHC sequencing consortium. Complete sequence and gene map of a human major histocompatibility complex: the MHC sequencing consortium. Nature 1999;401: Horton R, Wilming L, Rand V, et al. Gene map of the extended human MHC. Nat Rev Genet 2004;5: Karell K, Holopainen P, Mustalahti K, et al. Not all HLA DR3 DQ2 haplotypes confer equal susceptibility to coeliac disease: transmission analysis in families. Scand J Gastroenterol 2002;37: Louka AS, Moodie SJ, Karell K, et al. A collaborative European search for non-dqa1*05-dqb1*02 celiac disease loci on HLA- DR3 haplotypes: analysis of transmission from homozygous parents. Hum Immunol 2003;64: Bolognesi E, Karell K, Percopo S, et al. Additional factor in some HLA DR3/DQ2 haplotypes confers a fourfold increased genetic risk of celiac disease. Tissue Antigens 2003;61: Louka AS, Lie BA, Talseth B, et al. Coeliac disease patients carry conserved HLA-DR3-DQ2 haplotypes revealed by association of TNF alleles. Immunogenetics 2003;55: Lie BA, Mora B, Boland A, et al. The 13th international histocompatibility work group for celiac disease joint report. In: Hansen JA, Dupont B, eds. HLA Immunobiology of the human MHC. Proceedings from the 13th International Histocompatibility Workshop. Seattle: IHWG Press, Nilsen EM, Jahnsen FL, Lundin KEA, et al. Gluten induces an intestinal cytokine response strongly dominated by interferon gamma in patients with celiac disease. Gastroenterology 1998; 115: McManus R, Wilson AG, Mansfield J, et al. TNF2, a polymorphism of the tumour necrosis- gene promoter, is a component of the celiac disease major histocompatibility complex haplotype. Eur J Immunol 1996;26: Louka AS, Lie BA, Talseth B, et al. Coeliac disease patients carry conserved HLA-DR3-DQ2 haplotypes revealed by association of TNF alleles. Immunogenetics 2003;55: Woolley N, Mustalahti K, Mäki M, et al. Cytokine gene polymorphisms and genetic association with coeliac disease in the Finnish population. Scand J Immunol 2005;61: Wilson AG, Symons JA, McDowell TL, et al. Effects of a polymorphism in the human tumor necrosis factor promoter on transcriptional activation. Proc Natl Acad Sci U S A 1997;94: Zhong F, McCombs CC, Olson JM, et al. An autosomal screen for genes that predispose to celiac disease in the western counties of Ireland. Nat Genet 1996;14: Greco L, Corazza G, Babron MC, et al. Genome search in celiac disease. Am J Hum Genet 1998;62: King AL, Yiannakou JY, Brett PM, et al. A genome-wide familybased linkage study of coeliac disease. Ann Hum Genet 2000; 64: Naluai TÅ, Nilsson S, Gudjónsdóttir AH, et al. Genome-wide linkage analysis of Scandinavian affected sib-pairs supports presence of susceptibility loci for celiac disease on chromosomes 5 and 11. Eur J Hum Genet 2001;9: Liu J, Juo SH, Holopainen P, et al. Genome-wide linkage analysis of celiac disease in Finnish families. Am J Hum Genet 2002;70: Neuhausen SL, Feolo M, Camp NJ, et al. Genome-wide linkage analysis for celiac disease in North American families. Am J Med Genet 2002;111: Woolley N, Holopainen P, Ollikainen V, et al. A new locus for coeliac disease mapped to chromosome 15 in a population isolate. Hum Genet 2002;111: Van Belzen MJ, Meijer JW, Sandkuijl LA, et al. A major non-hla locus in celiac disease maps to chromosome 19. Gastroenterology 2003;125: Rioux JD, Karinen H, Kocher K, et al. Genome-wide search and association studies in a Finnish celiac disease population: identification of a novel locus and replication of the HLA and CTLA4 loci. Am J Med Genet 2004;130A: Greco L, Babron MC, Corazza GR, et al. Existence of a genetic risk factor on chromosome 5q in Italian coeliac disease families. Ann Hum Genet 2001;65:35 41.

9 September 2005 CELIAC DISEASE Babron MC, Nilsson S, Adamovic S, et al. Meta and pooled analysis of European coeliac disease data. Eur J Hum Genet 2003;11: Holopainen P, Arvas M, Sistonen P, et al. CD28/CTLA4 gene region on chromosome 2q33 confers genetic susceptibility to celiac disease: a linkage and family-based association study. Tissue Antigens 1999;53: Naluai ÅT, Nilsson S, Samuelsson L, et al. The CTLA4/CD28 gene region on chromosome 2q33 confers susceptibility to celiac disease in a way possibly distinct from that of type 1 diabetes and other chronic inflammatory disorders. Tissue Antigens 2000; 56: King AL, Moodie SJ, Fraser JS, et al. CTLA-4/CD28 gene region is associated with genetic susceptibility to coeliac disease in UK families. J Med Genet 2002;39: Popat S, Hearle N, Wixey J, et al. Analysis of the CTLA4 gene in Swedish coeliac disease patients. Scand J Gastroenterol 2002; 37: Mora B, Bonamico M, Indovina P, et al. CTLA-4 49 A/G dimorphism in Italian patients with celiac disease. Hum Immunol 2003;64: Amundsen SS, Naluai AT, Ascher H, et al. Genetic analysis of the CD28/CTLA4/ICOS (CELIAC3) region in coeliac disease. Tissue Antigens 2004;64: van Belzen MJ, Mulder CJ, Zhernakova A, et al. CTLA4 49 A/G and CT60 polymorphisms in Dutch coeliac disease patients. Eur J Hum Genet 2004;12: Hunt KA, McGovern DP, Kumar PJ, et al. A common CTLA4 haplotype associated with coeliac disease. Eur J Hum Genet 2005;13: Hill ID, Dirks MH, Liptak GS, et al. Guideline for the diagnosis and treatment of celiac disease in children: recommendations of the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr 2005;40: Contreas G, Valletta E, Ulmi D, et al. Screening of coeliac disease in north Italian children with type 1 diabetes: limited usefulness of HLA-DQ typing. Acta Paediatr 2004;93: Spurkland A, Sollid LM, Polanco I, et al. HLA-DR and -DQ genotypes of celiac disease patients serologically typed to be non- DR3 or non-dr5/7. Hum Immunol 1992;35: Kaukinen K, Partanen J, Mäki M, et al. HLA-DQ typing in the diagnosis of celiac disease. Am J Gastroenterol 2002;97: Sollid LM, Khosla C. Future therapeutic options for celiac disease. Nat Clin Pract Gastroenterol Hepatol 2005;2: Hausch F, Halttunen T, Mäki M, et al. Design, synthesis, and evaluation of gluten peptide analogs as selective inhibitors of human tissue transglutaminase. Chem Biol 2003;10: Aeschlimann D, Paulsson M. Transglutaminases: protein crosslinking enzymes in tissues and body fluids. Thromb Haemost 1994;71: Chatenoud L. CD3-specific antibody-induced active tolerance: from bench to bedside. Nat Rev Immunol 2003;3: Burkly LC. CD40 pathway blockade as an approach to immunotherapy. Adv Exp Med Biol 2001;489: Ascher H, Krantz I, Kristiansson B. Increasing incidence of coeliac disease in Sweden. Arch Dis Child 1991;66: Ivarsson A, Persson LA, Nystrom L, et al. Epidemic of coeliac disease in Swedish children. Acta Paediatr 2000;89: Ivarsson A, Hernell O, Stenlund H, et al. Breast-feeding protects against celiac disease. Am J Clin Nutr 2002;75: Vader W, Stepniak D, Kooy Y, et al. The HLA-DQ2 gene dose effect in celiac disease is directly related to the magnitude and breadth of gluten-specific T cell responses. Proc Natl Acad Sci U S A 2003;100: Address requests for reprints to: Ludvig M. Sollid, MD, Institute of Immunology, University of Oslo, Rikshospitalet University Hospital, N-0027 Oslo, Norway. l.m.sollid@medisin.uio.no; fax: Supported by grants from the Research Council of Norway, the European Community (BHM4-CT , QLK1-CT , and QLK1-CT ), the Norwegian Cancer Society, the Juvenile Diabetes Foundation International ( and ), the NIH (DK65965; together with C. Khosla), and EXTRA funds from the Norwegian Foundation for Health and Rehabilitation. The authors thank Anneli Ivarsson, Henry Ascher, Thore Egeland, and Lars Mørkrid for helpful discussions and Andrew S. Louka for critical reading of the manuscript.