Celiac disease: a model disease for gene–environment interaction

Abstract

Celiac sprue remains a model autoimmune disease for dissection of genetic and environmental influences on disease progression. The 2010 Congress of Autoimmunity included several key sessions devoted to genetics and environment. Several papers from these symposia were selected for in-depth discussion and publication. This issue is devoted to this theme. The goal is not to discuss genetic and environmental interactions, but rather to focus on key elements of diagnosis, the inflammatory response and the mechanisms of autoimmunity.

In 1888, Samuel Gee presented the first clinical picture of celiac disease (CD, also known as gluten enteropathy or celiac sprue).¹ At the time, he could not imagine that this disease might affect up to 1–2% of the population.² Today, CD is common worldwide. There are some areas of very high CD prevalence, such as Saharawi in North Africa, where CD occurs in 5.6% of the population.³ In some regions, including China, studies of CD prevalence are still under way. Nevertheless, it is clear that there are substantial differences in CD incidence between world regions, depending on genetic, environmental and social factors.

By a general definition, CD is an intestinal inflammatory disorder that occurs in genetically predisposed individuals as a result of intolerance to the wheat protein gluten and related proteins (prolamines) from barley and rye. Histologically, the small-bowel mucosa in CD patients has villous atrophy (loss of villi), crypt hyperplasia and lymphocyte infiltration. For a better understanding of histological damage induced by CD, a system for staging mucosal changes (from I to IIIABC) was proposed by Marsh.⁴ In this system, Marsh I represents lymphocytic enteritis, Marsh II lymphocytic enteritis with crypt hyperplasia, and Marsh IIIABC partial (A), subtotal (B) and total (C) villous atrophy. These changes are accompanied by a gradual increase in the number of T cells and activation of immunoregulatory counteractions in the affected mucosa.^{5, 6}

The damage to the intestinal mucosa in celiac sprue is due to an immunological reaction against gluten and related prolamines; there is also a very strong autoimmune component, in which tissue (or type 2) transglutaminase (tTG) has a central role. The mechanisms of autoimmunity development against tTG are not fully known, although several hypotheses have been proposed.⁷ By contrast, immune reactions to gluten (or its component gliadin) have been studied rather extensively studied. We know that in the small-intestine mucosa, tTG deamidates glutamine residues in gliadin-derived peptides that have been incompletely digested in the upper gastrointestinal tract owing to their high proline content. This increases the ability of these peptides to bind human histocompatibility leukocyte antigen (HLA)-DQ2 or HLA-DQ8 molecules. The HLA-DQ molecule and gluten peptide complexes are preferentially recognized by CD4⁺ T lymphocytes in the intestinal mucosa. Here, antigen-presenting cells are the limiting factor that defines the extent of the CD4⁺ T-cell response. Individuals who are homozygous for HLA-DQ2 or -DQ8 are more prone to develop CD than those who are heterozygous for either of these genes. In HLA-DQ2-positive individuals, gluten-derived peptides bind to negatively charged residues in the pockets of the peptide-binding groove but to positively charged pockets in HLA-DQ8-positive individuals. Importantly, HLA-DQ2 seems to have a unique ability to accommodate proline residues in the P1 pocket. These distinct HLA-DQ2 and HLA-DQ8 binding characteristics result in distinct T-cell responses to different antigen epitopes, revealing why HLA-DQ2 is more strongly associated with CD than with HLA-DQ8.^{2, 5}

CD patients, with rare exceptions, have HLA-DQ2 or/and HLA-DQ8 alleles. However, these HLA types are also common in individuals without CD. We therefore propose the involvement of non-HLA genes and environmental factors in the development of CD. Genome-wide association studies have revealed about 40 non-HLA loci, from which one may infer that nearly 50% of genetic susceptibility to CD is known.⁸ Many of these genes encode immunologically relevant proteins with important influence on innate and/or adaptive immunity pathways in CD.⁹ It has been proposed that some immunologically relevant CD-associated alleles—including the SH2B3 rs3184504*A allele, which plays a protective role against bacterial infection—may have undergone positive selection rather recently.⁸ This finding, along with those in other recent studies of the role of infections¹⁰ and intestinal indigenous microflora composition^{11, 12} in CD, show that immune reactions to gluten are strongly influenced by changes in the intestinal microbiome.

As a result of these recent findings, much attention has been paid to the immunobiology of intestinal mucosa in normal and diseased states and the ability of upper gastrointestinal infections to trigger CD.¹³ IFN-α (Ref. 14) and IL-15 (Ref. 15) have been shown to be central players in CD pathogenesis, specifically, because of their significance as markers of the innate immune response to intracellular pathogens such as double-stranded RNA viruses. Notably, infections with rotavirus, a double-stranded RNA virus, have been shown to promote the development of CD in children.¹⁶ Nevertheless, it remains unclear how IFN-α and IL-15 production is maintained in CD without infection and how gluten and other prolamines might be related to the production of these and other cytokines (e.g., IFN-γ, IL-18 and IL-21) implicated in CD development. One explanation may be a hypothesized increased intestinal permeability in individuals susceptible to CD. According to Lammers et al.,¹⁷ gliadin may increase the MyD88-dependent release of zonulin, a protein that induces intercellular tight junction disassembly. This effect is due to gliadin binding to CXCR3, a chemokine receptor located in intestinal epithelium.¹⁷ The release of zonulin is an early change that precedes the gliadin-induced immune events in CD and is related to changes in interactions of molecules (e.g., tight junction protein 1, claudin and occludin) in intestinal mucosa tight junctions.¹⁸ Nevertheless, in genetically predisposed individuals, gliadin may stimulate other CXCR3-expressing cells, including T cells and natural killer (NK) cells. This may explain why NK cells are activated mainly in the first stages of CD development. Additionally, this interaction may explain the initial steps of extraintestinal manifestations of autoimmune processes in CD.

The immunological background of extraintestinal manifestations, as well as of other disease associations in CD, is not well studied. The induction of autoimmune reactions to other tissues due to molecular mimicry or other gluten (gliadin)-related mechanisms such as the immunostimulatory effect of gliadin has been suggested. Immune reactions to type 3 tTG in dermatitis herpetiformis and type 6 tTG in cerebellar ataxia are excellent examples of such associations.¹⁹ Very important data recently obtained in diabetes studies show that wheat proteins are involved in the induction of pancreatic islet-related autoimmunity.²⁰

The discussion herein is relevant not only to CD but, more importantly, in a generic way to other autoimmune as well as autoinflammatory diseases. There is a great increase in data available from epidemiologic studies on prevalence and incidence of autoimmunity and this increase in epidemiologic information is providing key clues on pathogenesis. For one thing, it is now clear that genetic predisposition is a necessary first step to all autoimmune diseases. In addition, it is also clear that diseases tend to cluster. For example, one cluster would consist of patients with systemic lupus, rheumatoid arthritis, scleroderma and Sjogren's syndrome. In other words, if a patient has systemic lupus, there is a higher risk that members of the family will have rheumatoid arthritis, scleroderma or Sjogren's syndrome. However, within this cluster, are excluded multiple sclerosis and ulcerative colitis, for example. On the other hand, ulcerative colitis would fall in the cluster of inflammatory bowel disease which would include Crohn's disease and primary sclerosing cholangitis. An additional cluster are the autoimmune endocrine diseases. This clustering becomes an important feature for our understanding of immunopathology as it indicates not only the likelihood of common genes, but also the probability of common environmental factors.

These comments notwithstanding, it is also clear that the genetic predisposition may not be the same for different ethnic or racial populations. This has been known by rheumatologists for many years in studies of the significance of HLA-B27 in ankylosing spondylitis and its relevance for disease for populations in Europe versus Asia. It is now equally clear with such studies of the nonobese diabetic gene in patients with Crohn's disease in varied ethnic populations as well. Indeed, an enormous amount of money and effort has been spent studying genome-wide association as well as candidate genes in the many autoimmune diseases. This effort, although providing us with great springboards, has been largely disappointing. The disappointment arises because of the complexity of the analysis and the appreciation that many of these diseases have very complex multigenic origins. This is perhaps illustrated best by systemic lupus and the large number of lupus associating genes. At the bench it is even more noteworthy in study of the genetics of the first mouse model of autoimmunity, the New Zealand mouse. The New Zealand black mouse was discovered in 1959, making it the oldest animal model of autoimmunity, and now in its fifty-first year of study. Here again, despite the plethora of publications on New Zealand mice, we are not much further to defining a precise etiology to this disease than we were 50 years ago. However, this is not to say that such study has not been of value. Indeed, study of this mouse model and others have led to enormous advances in diagnostics, therapeutics and indeed even the very fundamental technology that has gone hand to glove with analysis of effector mechanisms. The next steps within the genetics are next generational sequencing. This would not have even been possible were it not for the improved cost effectiveness of such sequence analysis. When the human genome project was launched it was considered the most expensive biologic project in history. One can now do such analysis at a fraction of the cost and the authors herein predict that by the end of the decade, deep sequencing will be the norm and indeed the field of bioinformatics will become the dominant field of its nature in the genetics of autoimmunity. These comments become critical because celiac sprue is a very homogeneous disease. It is readily diagnosed. The epidemiology is clear. The agent appears to be known and as such we should be closer to better treatments and ideally even prevention.

In recent years, two central questions have been raised regarding CD. First, are we ready to diagnose CD via immune (anti-tTG, anti-deamidated gliadin antibodies) and immunogenetic analyses without confirming the presence of the disease with small-intestine biopsies? Second, do we have tools for immunological treatment of CD that avoids the need for a distressing lifelong gluten-free diet? Although substantial success has been achieved in both areas,^{2, 7} we should acknowledge that we are not ready to give final answers to these questions because of the variability of CD presentation. Nevertheless, there are a relatively large number of cases in which CD-specific antibodies are present in association with normal small-intestine architecture. In these cases, starting with a gluten-free diet or other treatment procedures (e.g., modulation of small-intestine permeability or induction of tolerance to gliadin) is not justified. However, we strongly believe that additional intensive studies will very soon bring us to the point where most cases of CD will be diagnosed and effectively treated without the need for small-intestine biopsy. Finally, we note that there were two dedicated issues devoted to the environmental influences of autoimmunity, and the reader is referred to this series of articles.^{21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62}