Coeliac disease and gluten sensitivity
Riccardo Troncone, Department of Pediatrics and European Laboratory for the Investigation of Food-Induced Diseases, University Federico II, via S Pansini 5, I-80131 Naples, Italy.
(fax: +39 081 5469811; e-mail: firstname.lastname@example.org).
Abstract. Troncone R, Jabri B (University Federico II, Naples, Italy; University of Chicago, Chicago, IL, USA). Coeliac disease and gluten sensitivity (Symposium). J Intern Med 2011; 269: 582–590.
Coeliac disease (CD) is a systemic immune-mediated disorder elicited by gluten in genetically susceptible individuals. The common factor for all patients with CD is the presence of a variable combination of gluten-dependent clinical manifestations, specific autoantibodies (anti-tissue transglutaminase/anti-endomysium), HLA-DQ2 and/or DQ8 haplotypes and different degrees of enteropathy. Recently, gluten sensitivity has received much interest, although the limits and possible overlap between gluten sensitivity and CD remain poorly defined. At present, a number of morphological, functional and immunological disorders that are lacking one or more of the key CD criteria (enteropathy, associated HLA haplotypes and presence of anti-transglutaminase two antibodies) but respond to gluten exclusion are included under the umbrella of gluten sensitivity. The possible immunological mechanisms underlying these conditions are discussed. Emphasis is given to specific autoantibodies as markers of the coeliac spectrum and to the hypothesis that innate epithelial stress can exist independently from adaptive intestinal immunity in gluten sensitivity.
In recent years, the view of coeliac disease (CD) has undergone a profound revision. A wide spectrum of clinical and histological presentations has been recognized, and significant progress has been made in the understanding of the genetic and immunological features of this condition . CD is now considered, more than a just a gluten-sensitive enteropathy, to be a systemic immune-mediated disorder elicited by gluten and related prolamines in genetically susceptible individuals. The common denominator for all patients with CD is the presence of a variable combination of gluten-dependent clinical manifestations, specific autoantibodies [anti-tissue transglutaminase (TG)2)/anti-endomysium (EMA) antibodies], HLA-DQ2 and/or DQ8 haplotypes and different degrees of enteropathy, ranging from lymphocytic infiltration of the epithelium to complete villous atrophy .
However, gluten may induce other pathological conditions, such as wheat allergy , which is an immunoglobulin (Ig)E-mediated disease also well characterized from the immunological and clinical point of view but completely unrelated to CD. More recently, attention has been given to another entity, gluten sensitivity (GS), for which the limits and possible overlap with CD are still poorly defined . Currently, a number of morphological, functional and immunological disorders have been considered under the umbrella of GS that miss one or more of the key CD criteria (enteropathy, associated HLA haplotypes and presence of anti-TG2 antibodies), but respond to gluten exclusion.
Here, we will first briefly review the main features that are now considered to characterize CD. We will then describe in more detail the characteristics of those subjects who, although having the immunological (autoantibodies) and genetic (HLA-DQ2 and DQ8) signs of CD, lack the enteropathy (potential CD, dermatitis herpetiformis and gluten ataxia). Finally, we will discuss disorders that are included within the term GS.
Several classifications of CD have been used in the past, most importantly with the distinction between symptomatic, asymptomatic, latent and potential CD. The most common gastrointestinal and extraintestinal presentations are discussed later. Silent CD is defined as the presence of positive CD-specific antibodies and HLA haplotype and small bowel biopsy findings compatible with CD, but without sufficient symptoms and signs to warrant clinical suspicion of CD. With the recognition of forms of CD characterized by a low-grade enteropathy, other terms have been proposed. Latent CD defines a condition in which subjects are without enteropathy but have had a gluten-dependent enteropathy at some time during their lives . Patients may or may not have symptoms. Potential CD is defined by the presence of specific CD antibodies and compatible HLA haplotypes, but with no histological abnormalities in duodenal biopsies . Patients may or may not have symptoms and may or may not develop a gluten-dependent enteropathy at a later time. This condition will be discussed in detail in the next section.
Coeliac disease is triggered by the ingestion of wheat gluten and related prolamines from rye and barley . Most studies have demonstrated that oats is not harmful; however, a few patients have oats prolamine-reactive mucosal T cells that can cause mucosal inflammation . CD is relatively common (1% prevalence of biopsy-proven disease) in populations of mainly Caucasian descent . Environmental factors might affect the risk of developing CD or the timing of its presentation. Feeding patterns in the first year of life  and possibly viral infections  may contribute to the development of the disease.
Clinical features of CD vary considerably . Intestinal symptoms are common particularly in children diagnosed within the first 2 years of life; failure to thrive, chronic diarrhoea, vomiting, abdominal distension, muscle wasting, anorexia and irritability are present in most cases. The most frequent extraintestinal manifestation of CD is iron-deficiency anaemia that is unresponsive to iron therapy. Osteoporosis may be present. Other extraintestinal manifestations include short stature, endocrinopathies, arthritis and arthralgia, epilepsy with bilateral occipital calcifications, peripheral neuropathies, cardiomyopathy, elevation of transaminase levels, dental enamel hypoplasia, aphthous stomatitis and alopecia. The mechanisms responsible for the severity and the variety of clinical presentations remain obscure. Nutritional deficiencies or a dysfunctional immune (possibly autoimmune) response have been advocated. Silent CD is being increasingly recognized, mainly in asymptomatic first-degree relatives of patients with CD investigated during screening studies. Some diseases, many with an autoimmune pathogenesis, are found with a higher than normal frequency in patients with CD . Amongst these conditions are type 1 diabetes, autoimmune thyroid diseases, Addison’s disease, Sjögren’s syndrome, autoimmune cholangitis, autoimmune hepatitis and primary biliary cirrhosis. Other associated conditions include selective IgA deficiency, Down’s syndrome, Turner syndrome and Williams syndrome.
[ Coeliac disease with villous atrophy is at the end of a spectrum, being characterized by gluten-induced both innate and adaptive immunity. Our working hypothesis suggests gluten-induced epithelial distress playing a role in some forms of gluten sensitivity (e.g. IBS like). On the other hand milder forms of coeliac disease (e.g. potential coeliac disease) show gluten-specific T cell immunity, and yet lack important aspects of innate immunity. ]
A genetic predisposition is suggested by familial aggregation of CD and more than 70% concordance in monozygotic twins. . More than 90% of patients with CD express the HLA-DQ2.5 heterodimer, and almost all others express HLA-DQ8 . By contrast, only 30–40% of control patients express these CD-associated MHC class II molecules. CD is a T-cell-mediated chronic inflammatory disorder with an autoimmune component (for review see [1, 2]). Altered processing by intraluminal enzymes, changes in intestinal permeability and activation of innate immunity mechanisms precede the activation of the adaptive immune response. Immunodominant epitopes from gliadin are highly resistant to intraluminal and mucosal digestion; incomplete degradation favours the immunostimulatory and toxic effects of these sequences. Some gliadin peptides (p31–43) that are not recognized by T cells are able to activate innate immunity in antigen-presenting cells and intestinal epithelial cells; in particular, they induce interleukin (IL)-15. Others activate lamina propria T cells presented by HLA-DQ2 or DQ8 molecules [11, 12]. Gliadin-specific T-cell responses have been found to be enhanced by the action of tissue TG [13, 14]; the enzyme converts particular glutamine residues into glutamic acid, which results in higher affinity of these gliadin peptides for HLA-DQ2 or DQ8. The pattern of cytokines produced by gluten-specific T cells following gliadin activation is clearly dominated by interferon (IFN)γ (Th1 skewed)  and IL-21 . In addition, the innate cytokines IFNα , IL-15  and IL-18  are also produced. A complex remodelling of the mucosa takes place downstream of T-cell activation, involving increased levels of metalloproteinases and growth factors, which leads to the classical ‘flat mucosa’ of CD. An increased density of CD8+ cytotoxic intraepithelial lymphocytes is a hallmark of CD . IL-15 is implicated in the expression of natural killer receptors CD94  and NKG2D [22, 23], as well as in epithelial expression of stress molecules , thus enhancing cytotoxicity, cell apoptosis and villous atrophy. The most evident expression of autoimmunity is the presence of serum antibodies to tissue TG . However, the mechanisms leading to autoimmunity are largely unknown, as are their pathogenetic significance.
Table 1. HLA, immunological, pathological and clinical features of gluten-related conditions
|Wheat allergy|| ||Variable||+||–||Mast cells eosinophils||+?||–||–||Allergic symptoms|
|Gluten-dependent pathology with anti-TG2||Potential CD||+||+||+||–||+||–||–||–|
|Gluten ataxia||80% +a||+||+ (TG6)||?||+||–||Variable||Neurological|
|Gluten sensitivity||IBS||Variable||Variable||–||Innate IEC response?||–||Stress? IL-15?||–||–|
|Autoimmune (T1D)||Variable||–||–||Signs of dysfunctional mucosalimmune response to gluten||–||Stress? IL-15?||–||Autoimmunity|
Gluten-dependent pathology characterized by the presence of anti-TG antibodies with low-grade or no enteropathy
Clinical features. The histological changes in the small intestinal mucosa in CD cover a wide spectrum ranging from infiltration of the epithelium by lymphocytes to complete villous atrophy. Potential CD is defined by the presence of specific CD antibodies and compatible HLA haplotypes, but without histological abnormalities in duodenal biopsies. In fact, a small percentage of subjects positive for CD antibodies [anti-gliadin antibody (AGA), EMA and anti-tTG] have a small intestinal mucosa without villous atrophy. Some of these patients have a Marsh 0 lesion (i.e. no intraepithelial infiltration). In a significant proportion of cases, these subjects belong to at-risk groups (family members or presence of autoimmune disorders associated with CD). The titre of antibodies is often lower than that found in untreated CD patients with villous atrophy . In some cases, fluctuating titres are noted, and it is not uncommon for antibodies to disappear from serum during follow-up . The number of such patients is increasing because of the raised awareness of CD and the more diffuse screening of ‘at-risk’ subjects. In our institution, potential patients with CD now account for nearly 20% of patients with positive serology who undergo a small intestinal biopsy (R. Troncone, personal communication). It is very unlikely that these are false-positive results, as both the serology and the genetics are consistent with CD. There is no agreement on how to treat these patients. If symptomatic, they are usually offered a trial with gluten-free diet. Otherwise, most physicians advise a very strict follow-up. In fact, the natural history of these patients is still unknown, as are the risks if they remain on a gluten-containing diet. We recently performed a prospective, 3-year cohort study to determine the natural history of potential CD in children . The study included 106 children with potential CD, based on serological analysis and normal jejunal architecture. All but two carried the HLA-DQ2 and/or DQ8 haplotype. In all children, growth, nutritional parameters, CD serology and autoimmunity were evaluated every 6 months. γδ intraepithelial-, CD3- and lamina propria CD25-positive cells were counted in biopsy specimens, and jejunal deposits of anti-TG2 IgA were detected. Biopsy analysis was repeated after 2 years on patients with persistent positive serology and/or symptoms. Potential CD was detected primarily in first-degree relatives and patients with autoimmune disorders (40.6%). Eighty-nine of the 106 patients entered the follow-up study, with normal daily consumption of gluten. During the follow-up, antibodies disappeared in 14.6% of subjects and fluctuated in 32.6%. Villous atrophy was observed in 12/39 (30.8%) patients who underwent a repeat biopsy. The only useful parameter for identifying those patients more at risk of developing severe damage of the intestinal mucosa was the presence of intestinal deposits of anti-TG2 IgA at the time of the first observation. Most children with potential CD remained healthy. It is known that undiagnosed CD may predispose to nutritional deficiencies, in most cases subtle, even in the absence of overt clinical symptoms. However, there were no significant nutritional or bone complications in these patients.
Immunological features. The presence of anti-gluten and anti-TG antibodies indicates that these patients, with potential CD in general, despite the absence of a clear enteropathy, have developed an anti-gluten adaptive immune response. Therefore, despite a normal architecture, the immunohistochemical analysis of jejunal mucosa from patients with potential CD shows signs of T-cell and B-cell activation. The absence of villous atrophy in patients with potential CD is associated with increased IL-10 expression and subsequently an increased IL-10/IFNγ ratio as compared to those with active CD. Furthermore, the expression of IL-21 appears to be downregulated in the biopsies from patients with potential CD, compared to controls (R. Troncone, personal communication). In addition, there is a preferential increase in TCR γδ IEL over TCRαβ IEL in patients with potential CD. Finally, intraepithelial cytotoxic TCRαβ lymphocytes in patients with potential CD, in contrast to those with active CD, continue to express inhibitory NK receptors and have low levels of activating NK receptors (B. Jabri, personal communication). Altogether these observations suggest that patients with potential CD, compared to those with active CD, develop an inflammatory anti-gluten CD4 T-cell immune response of lower magnitude and lack IELs with an activated killer phenotype.
Mechanisms and unanswered questions. The mechanisms underlying the low-grade inflammatory pattern in potential CD remain unclear. There are two main nonmutually exclusive possible explanations for this pattern. First, epithelial cells in patients with potential CD are normal, i.e. they show no signs of distress and lack high levels of IL-15 expression and ligands for activating NK receptors (e.g. nonclassical MHC class I molecules HLA-E and MICA/B). The hypothesis that a lack of innate activation of epithelial cells explains the absence of villous atrophy is supported by studies in humanized HLA-DQ2 and DQ8 mice showing that adaptive anti-gluten immunity and the presence of anti-TG2 antibodies is insufficient to induce villous atrophy. It is also in accordance with our observations that patients with potential CD do not display signs of epithelial distress, as defined by the upregulation of IL-15 and heat shock proteins. The second possible explanation is that a low level of proinflammatory adaptive anti-gluten immunity may be responsible for the lack of villous atrophy observed in patients with potential CD. This could be observed either because the number of inflammatory anti-gluten CD4 T cells is insufficient to induce pathology or because of the presence of a protective regulatory IL-10 response. The former is supported by reports suggesting that there is an HLA gene dosage effect , i.e. the number of T-cell-stimulatory HLA-DQ/gluten peptide complexes is critical for disease development . The latter is supported by the well-known anti-inflammatory properties of IL-10; the level of this cytokine was found to be increased in patients with potential CD compared to those with active CD. The role of FOXP3 T cells is difficult to assess as inducible gluten-specific FOXP3 T cells cannot be distinguished from natural regulatory T cells with specificity for self that may infiltrate inflamed intestinal tissue in a nonspecific manner. Supporting the difficulty in interpreting these data, FOXP3-positive cells are at significantly lower level in patients with potential CD than in those with active CD with villous atrophy . Future studies will determine whether patients with low-grade enteropathy have a different genetic make-up (HLA and non-HLA) from those with active CD or whether they lack key environmental triggers in addition to gluten.
Gluten-dependent extraintestinal pathology with anti-TG antibodies
Dermatitis herpetiformis. Dermatitis herpetiformis (DH) is an example of a condition characterized by genetic (HLA association) and immunological (adaptive anti-gluten immunity and anti-TG antibodies) features typical of CD, but associated mainly with skin lesions and a variable degree of enteropathy. This gluten-dependent blistering skin disorder is characterized by skin IgA anti-TG deposits associated with an adaptive anti-gluten immune response. Epidermal TG3 and the closely related TG2 are considered to be autoantigens in DH because a majority of patients with this condition have IgA specific for TG2 and TG3 . The levels of circulating autoantibodies against TG2 and TG3 are correlated with each other and both appear related to the degree of enteropathy, suggesting that the gut is the site at which the autoimmune response occurs . Deposits of anti-TG2 are present at the jejunal level as in classical CD. DH is considered by many as an extraintestinal presentation of CD .
In a series of studies conducted in the UK over a period of 13 years, 101 of 215 subjects with idiopathic sporadic ataxia showed serological evidence of an abnormal immune response to gluten (for review see ). A response to treatment with a gluten-free diet was evident in some cases, although it appeared to depend on the duration of symptoms. Subsequently, other neurological conditions, in particular axonal peripheral neuropathy, were found to be related to a gluten-induced immune response. In patients with ataxia and other neurological conditions, the HLA distribution was slightly different from that seen in those with classical CD, with HLA-DQ2 (70%) and DQ8 (10%) accounting overall for only 80% and HLA-DQ1 in the remaining patients. The prevalence of AGA-positive patients amongst those with idiopathic sporadic ataxia was as high as 45%, compared with 10% in genetically confirmed ataxia. Less than 10% of patients with gluten ataxia presented gastrointestinal symptoms, and only a third showed evidence of enteropathy. Once anti-TG assessment became available, up to 38% of patients were found to be positive, often at lower titres than those seen in CD, with IgG class antibodies being more common than IgA. IgG/IgA antibodies against the TG isoenzyme TG6, primarily expressed in the brain, were present in more than half of patients with gluten ataxia. Overall, 85% of patients with ataxia and AGAs had antibodies to tissue TG2 and/or TG6. It is interesting that IgA deposition on jejunal TG2 in gut tissue from patients with gluten ataxia was associated with widespread deposition around vessels in their brain . To what extent this is responsible for the clinical symptoms remains uncertain.
Mechanisms and unanswered questions.
A particular genetic background and/or environmental factors may predispose some patients with adaptive anti-gluten immunity and anti-TG antibodies to develop extraintestinal pathology. DH is associated with anti-TG3 antibodies and gluten ataxia with anti-TG6 antibodies. In line with this, the TG6 and TG3 isoenzymes are particularly expressed in the brain and skin, respectively. Of interest, both are able to specifically deamidate gliadin peptides, although their precise specificities are different . By contrast, only 10% of patients with gluten ataxia have antibodies against deamidated peptides. The mechanisms underlying the production of anti-TG3 and anti-TG6, like those of anti-TG2, are still poorly understood, but probably involve gluten-specific T cells. Furthermore, whether DH and gluten ataxia are induced by anti-TG antibody deposits, which in turn are dependent on the presence of anti-gluten T cells, remains to be determined. It is also unclear whether the site of gluten immunization is intestinal or extraintestinal in these pathologies. Because anti-TG antibodies are of the IgG isotype and, particularly in cases without enteropathy, no deamidated peptides are present in gluten ataxia, it is likely that immunization against gluten occurs outside the intestinal environment in these patients.
GS pathology without anti-TG antibodies or enteropathy
Gluten sensitivity has been defined by the presence of morphological, functional and immunological disorders that respond to gluten exclusion and yet lack the characteristic features that define CD. This definition encompasses a wide range of disorders that may have very different underlying mechanisms but have in common the regression of symptoms in response to a gluten-free diet in the absence of anti-TG antibodies and histological enteropathy. It is important to recognize that disorders included under the term GS are actually conditions for which a precise definition and even rudimentary knowledge of the underlying mechanisms is lacking. However, we will attempt to classify GS disorders and discuss whether they are linked to adaptive anti-gluten immunity or innate/stress responses to gluten. Attempting such a classification even though it is somewhat speculative may help to design studies to identify objective markers and better delineate these disorders.
Irritable bowel syndrome-related GS
Recently, the relationships between irritable bowel syndrome (IBS), CD and GS have received particular attention. Both IBS and CD present with clinical symptoms that are largely overlapping. It is likely that many individuals with unrecognized CD are labelled as IBS patients. In fact, recent evidence has suggested that CD, as diagnosed by positive serology and positive biopsy, is four times more prevalent in IBS than in non-IBS populations. Patients with diarrhoea-predominant IBS are certainly amongst those who should be actively screened for CD . More interestingly, CD may predispose to IBS. It is in fact known that a persistent low-grade inflammation may favour development of IBS. For instance, IBS may develop following a gastrointestinal infection . Hence, it is plausible that gliadin-induced mucosal inflammation may also have a predisposing effect on the insurgence of IBS. Another possibility is that within the group of patients with IBS, there are those who cannot be classified as patients with CD, but still present a dysfunctional immune response to gluten and symptomatic responses to gluten challenge or withdrawal. Arranz and Ferguson found that a pattern of intestinal antibodies (high intestinal AGA IgA and IgM antibody titres) was able to identify a group of patients with gluten-sensitive diarrhoea in the absence of enteropathy . At the time this study was conducted, anti-TG antibodies were not identified; thus, it remains to be shown that these patients have anti-gluten antibodies in the absence of anti-TG antibodies. In a later study by Wahnschaffe and colleagues , a subset of IBS patients was found to have anti-gliadin IgA and IgG but not anti-TG2 IgA antibodies. These authors also reported anti-gluten antibodies in the absence of anti-TG2 antibodies in patients with IBS. Thus, the question of whether a group of patients with gluten-sensitive symptoms and anti-gluten immunity in the absence of anti-TG2 antibodies can be recognized remains open. Induction of anti-gluten immunity in the absence of anti-TG2 antibodies may define a very early stage of CD, before TG2 becomes activated and an amplification of the anti-gluten immune response occurs.
The possibility that a subset of IBS patients may be characterized by an innate immune response to gliadin is strengthened by other observations suggesting that such an innate response to gliadin may be dissociated from the adaptive specific immune response: (i) rectal gluten challenge studies showed a response characterized by infiltration of T cells similar to that observed in CD patients in six of 13 siblings of children with CD . Interestingly, the positivity was not associated with the presence of the haplotypes present in the great majority of patients CD ; (ii) gluten peptides that are not recognized by T cells may activate epithelial cells; and (iii) epithelial stress, as assessed by the presence of IL-15 and heat shock protein, was found in the absence of anti-gluten immunity and anti-TG antibodies in family members of CD patients (B. Jabri, personal observations). In this form of IBS-related GS, innate and not adaptive anti-gluten immunity would play a role and epithelial cells would be the targets.
Understand the different forms of IBS-related IBS would require in-depth clinical, biological and immunological phenotyping. For instance, the presence of anti-TG antibodies, signs of epithelial distress and T-cell activation before and after a gluten-free diet should be investigated.
The presence of gluten-sensitive extraintestinal manifestations in patients with anti-gluten antibodies has been suggested. For instance, 16% of patients with psoriasis have been found to present high levels of IgA and/or IgG antibodies to gliadin . This observation was considered a possible sign of GS and prompted the evaluation of the effect of a gluten-free diet on the manifestations of psoriasis. A significant improvement was observed in patients with elevated AGAs when they were put on a gluten-free diet, whereas lesions deteriorated when the ordinary diet was resumed in 18/30 patients . Similar observations have been made in patients with recurrent oral ulceration. Indeed, it is interesting that 85% of patients with oral ulceration who responded to gluten withdrawal had high serum levels of AGAs . The immunological basis for these pathologies is unknown. There are two key immunologically related issues that need to be addressed to better understand this disease entity. The first is related to the presence of anti-gluten antibodies in the absence of anti-TG antibodies. Possible explanations for this are that the site of immunization against gluten may be extraintestinal and/or that TG may not be activated. The second is related to what mediates extraintestinal GS. Are anti-gluten T cells and/or anti-gluten antibodies directly involved in the disease process? Do they have a direct role by inducing local inflammation? And if so, why do they induce inflammation only at certain sites?
Autoimmune (Type 1 diabetes) GS
Type 1 diabetes (T1D) is considered to be an organ-specific autoimmune disease mainly precipitated in genetically susceptible individuals in whom T lymphocytes infiltrate the islets of the pancreas and destroy the insulin-producing β-cell population . Environmental factors such as food antigens and viruses have been associated with T1D . There is much indirect evidence of a role for the intestinal immune system in humans: enhanced immune responsiveness to food antigens occurs in patients with T1D , and a close association between CD and T1D has been reported . In fact, a dysfunctional immune response to gliadin is likely to be a feature of T1D patients (who do not present anti-EMA and anti-tissue TG antibodies in their serum). Indeed, it has recently been reported that a subset of these patients react with lymphocyte infiltration to rectal instillation of gliadin , and it was demonstrated that jejunal biopsies from T1D patients mount an inflammatory response to gliadin in an organ culture system . However, we were unable to grow gliadin-specific T-cell lines from the jejunal mucosa of such patients, suggesting that mechanisms other than those of adaptive anti-gluten immunity were responsible for the inflammation elicited by gliadin. It is presently unclear whether in some of the patients gluten may play a causal role in the development of T1D. It is also possible that patients with T1D have nongluten-related intestinal inflammation and that the abnormal response to gluten is a by-product of mucosal inflammation. This, however, does not exclude the possibility that gluten may contribute to disease development by further amplifying intestinal inflammation. In agreement with a role for gluten in the development of T1D and autoimmune diseases in general, the sooner a gluten-free diet is started in patients with active CD the lower their risk of developing an autoimmune disorder . Conversely, intestinal inflammation as found in T1D  may favour development of CD. It is impossible to draw firm conclusion, given how little we know at present about the nature and cause of intestinal inflammation and its relationship to gluten in T1D and other autoimmune disorders. Depending on whether intestinal inflammation in T1D is influenced by gluten should determine whether T1D is included amongst GS disorders.
There is still a long way to go before all forms of GS can be accurately defined and their underlying mechanisms determined. Perhaps to progress, it is important to acknowledge how little we know and how confusing the terminology is. Indeed, in the term gluten sensitive is the word gluten and sensitive, and hence by definition classical CD, potential CD, wheat allergy and dermatitis herpetiformis should all be part of GS. The fact that these conditions all have specific terms, rather than simply GS, is because they are better defined than other pathological manifestations that have been discussed under the term GS in this review. Stringent clinical phenotyping to determine whether anti-TG antibodies can be pathological and mediate extraintestinal manifestations, and assess whether innate epithelial stress is a reality and can exist independently from adaptive intestinal immunity will help define gluten-induced or gluten-dependent pathologies. In future, GS may be replaced by a set of well-defined gluten pathologies.
Conflicts of interest
The authors have no conflicts of interest to declare.