Dermatitis herpetiformis: Novel advances and hypotheses

Abstract

Dermatitis herpetiformis (DH) is a gluten-sensitive autoimmune blistering disorder with a chronic-relapsing course. Very recently, several Authors reported atypical cases of patients with DH, suggesting that different clinical subsets may exist at least among different ethnicities and that the classical picture of DH probably need a significant revision. Moreover, different pathogenetic aspects of the disease are currently under investigation, including the role of epidermal transglutaminase, apoptosis and inflammatory cells in the occurrence of skin lesions, in order to explain why only a subgroup of celiac patients will develop DH. Finally, although gluten-free diet is still regarded as the only curative approach to the disease, it is very hard to comply with and even small amounts of gluten can re-activate the disease. Therefore, different therapeutical approaches for the spectrum DH/celiac disease are still under investigation. In the present paper, the most recent advances in DH will be discussed, and a novel interpretation of the disease based on the data emerging from the Literature will be proposed.

Key Words: Dermatitis herpetiformis; Celiac disease; Transglutaminase; Gluten-free diet; Direct immunofluorescence

INTRODUCTION

Dermatitis herpetiformis (DH) is a gluten-sensitive autoimmune blistering disorder with a chronic-relapsing course. The classic cutaneous manifestations are markedly pruritic, symmetrically distributed papulovesicles that usually affect the extensor surfaces of the limbs, the sacral region and the buttocks. DH is considered the specific cutaneous manifestation of celiac disease (CD), as both diseases occur in gluten-sensitive individuals, share the same HLA haplotypes (DQ2 and DQ8) and improve following the administration of gluten-free diet (GFD)[1-5].

The diagnosis of DH is established clinically, histologically, immunopathologically [direct immunofluorescence (DIF)] and serologically [IgA anti-tissue transglutaminase antibodies (anti-tTG), IgA anti-epidermal TG (eTG) and IgA endomisial autoantibodies][1].

Histology of skin lesions in DH is characterized by a subepidermal blister with accumulation of neutrophils and very few eosinophils in the papillary dermis. DIF, that is considered the gold standard for the diagnosis, reveals granular IgA deposits along the basement membrane zone, mainly distributed at the papillary tips. Recent studies demonstrated that such IgA are mainly directed against eTG[6]. Although the features reported above have been considered the hallmarks of DH, very recently several Authors, mainly from Japan, reported atypical cases of patients with DH, suggesting that different clinical subsets may exist at least among different ethnicities and that the classical picture of DH probably need a significant revision.

CLINICAL MANIFESTATIONS

Although the name of the disease reflects the typical clinical presentation association with celiac disease and immunopathological features of dermatitis herpetiformis, these occur later in the disease and are often immediately excoriated, resulting in erosions, crusted papules, or areas of postinflammatory dyschromia, or there may be erythema or severe pruritus alone, making the diagnosis often challenging[7].

By contrast, as previously mentioned, an increasing number of clinical studies have recently described uncommon presentations in patients with DH. Among them, purpuric lesions on hands and feet are being reported more frequently in children with DH from their first description in 1971[8,9]. Other atypical presentation of DH include palmo-plantar keratosis, wheals of chronic urticaria and lesions mimicking prurigo pigmentosa[10-12].

Such atypical presentations are more often reported in Japanese patients with DH, who seem also to have a decreased frequency of CD[2], that is in contrast with the data regarding Caucasian patients with DH, who show typical CD alterations at the small bowel biopsy (ranging from villous atrophy to augmented presence of intraepithelial lymphocytes) almost in all the cases. It is not clear whether the decreased frequency of CD in Japanese patients with DH is due to a less aggressive and specific diagnostic approach to the potential gut involvement, decreased exposure to gluten in the Japanese patients, or a true difference in the pathogenesis of the disease[2,13].

Interestingly, according to the latter hypothesis, Japanese patients show different findings also at the DIF, where an increased incidence of fibrillar pattern of IgA deposition is seen. Whether this feature could be considered a mere morphological variant or have implications for the pathogenesis of the disease is still under debate. However, some data suggested that DH patients with fibrillar IgA deposits show atypical clinical features such as the sparing of the predilection sites, the lack of enteropathy and no HLA-DQ2⁄DQ8 haplotype[2,14,15].

PATHOGENESIS OF DERMATITIS HERPETIFORMIS

As in CD, the pathophysiology of DH involves a complex interplay between genetics, autoimmune factors, and environment. In fact, it is now clear that in genetic predisposed individuals with HLA-DQ2⁄DQ8 haplotype, the exposure to a well known environmental factor, namely gluten, is able to trigger the disease by the induction of an autoimmune response against specific autoantigens, that are tTG for CD and, probably, eTG for DH.

However, the exact mechanisms through which gluten sensitivity results into the development of the specific lesions of DH are not completely understood. As an example, it is not clear why only a small cohort of patients with CD will develop DH. A possible clue could be represented by eTG that, as previously mentioned, is currently considered the target antigen of DH. In particular, eTG has been shown to colocalize with IgA deposits in the skin and anti-eTG antibodies have been found in the sera of patients with DH[6].

Besides these data, by passive transfer of anti-eTG antibodies to SCID mice engrafted with normal human skin, very recently Zone et al[16] reproduced the immunopathological pattern of DH, confirming the source of IgA in DH skin and the potential role of eTG in the disease. Accordingly, they proposed that epitope spreading from tTG to eTG could determine IgA anti-eTG autoantibodies production in a subset of celiac patients who then develop DH.

Unfortunately, they failed to reproduce DH lesions with transfer of anti-eTG antibodies in their model, suggesting that other unidentified proinflammatory factors trigger the cutaneous eruption. Moreover, data from the Literature showed that not all anti-eTG positive celiac patients will develop DH, and at least a subgroup of celiac patients without DH show IgA deposits at the dermal papillae[17-19].

As a consequence, several questions arise, including the following: (1) are anti-eTG antibodies really causative or are they simply a consequence of DH (2) Which are the additional factors involved in the inflammation found in DH skin and absent in the skin of celiac patients without DH and (3) Why eTG deposits are found at the papillary tips in DH skin

Regarding the latter, while tTG is ubiquitariously expressed in many tissue and, in the skin, it is found in the basal keratinocytes and dermal capillaries[20], eTG is not ubiquitary but it is primarily seen in the granular layer of the epidermis, small intestine, brain and testis[21]. Thus, it is not clear why it is found in the papillary dermis, and some potential explanations have been made. Probably, the most convincing one is linked to an accelerated apoptosis in the skin of DH patients, that has been recently demonstrated by a study of our research group[22]. Although DH should not be considered an apoptotic disease, the apoptosis of DH keratinocytes could lead to the liberation of proteins and enzymes from such cells, including eTG, that would deposit in the dermis. Accordingly, it has been shown that TRAIL, a pro-apoptotic molecule, can induce the expression of the keratinocyte differentiation markers, including transglutaminases[23]. Thus, the accumulation within the papillary dermis of the trigger antigen, i.e., eTG, might induce the development of an autoimmune response leading to the skin lesions of DH.

However, several data are in contrast with such a hypothesis. First of all, eTG deposits are found also in the skin of patients with CD that do not develop DH[19]; secondly, since apoptosis in DH skin is mainly confined in the basal layer of the keratinocytes, it could be expected a deposition in the papillary dermis even of tTG - that is present at high concentrations in such cells-, but it is not the case according to the study by Sardy et al[6], lastly, as previously discussed, the pathogenetic role of anti-eTG antibodies have to be determined yet and other factors are probably involved in the development of DH.

Among them, considering that DH skin lesions are predominantly distributed at sites of constant minor trauma (extensor surfaces, elbows, knee, buttocks), some Authors suggested that the latter can stimulate keratinocytes to secrete several cytokines that, in turn, are able to induce the expression of critical adhesion molecules (including E-selectin) on dermal endothelial cells and of chemokines such as interleukin (IL)-8, which would predispose these areas to the development of inflammation[24,25].

Another important factor could be represented by T cells. Accordingly, although the typical histopathological pattern of DH is characterized by neutrophilic accumulation at papillary tips, several reports showed that a predominantly lymphocytic dermal infiltrate may also be found in up to 40% of cases (and almost in all the patients in the initial phases of the disease)[26-28]. Moreover, activated CD4+ T cells with a cytokine expression pattern belonging to the Th2 phenotype have been documented in recent DH skin lesions as well as in the perilesional skin, suggesting their role in the early phases of DH skin inflammation[29,30]. In fact, the early recruitment of Th2 cells in the preferential sites of DH may allow the liberation of cytokines (such as IL-4 and IL-5) and chemotactic agents that are probably responsible for the accumulation of neutrophils and eosinophils in DH skin.

The fact that a previous study by Baker[31] failed to find gluten-specific CD4+ T cells from the skin of untreated DH patients, probably does not narrow the importance of such cells in the pathogenesis of DH, since further studies are required to confirm the absence of gluten-specific T cells in DH skin and, even in this case, some Authors suggested that CD4+ T cells in the gut could become sensitized to gluten and cause the activation of other cell types, such as neutrophils, that then migrate to the skin[3].

NEW THERAPEUTIC APPROACHES

Since DH is considered the specific cutaneous manifestation of CD, besides the symptomatic therapies often used in DH patients to control the skin flares at least in the first phases (i.e., dapsone, that should still be considered as the best option to clear DH skin lesions), GFD is still regarded as the only curative approach to the disease. However, since GFD is very hard to comply with and even small amounts of gluten can re-activate the diseases, different approaches are still under investigation.

Among them, the enzyme therapy could be considered a promising one. It is well known that gliadin peptides are highly resistant to digestive processing by pancreatic and brush border proteases[32]. Thus, the use of bacterial endopeptidases (including a supplementation to patients or pretreatment of whole gluten with them) has been proposed to promote complete digestion of cereal proteins and thus destroy gluten epitopes.

Other interesting alternatives to a GFD, although quite complicated, are represented by the use of engineered grains containing gliadin peptides able to inhibit the T cell response to gluten or different immunomodulatory strategies, including the use of tTG inhibitors, of DQ2 and DQ8 blocking peptides, as well as of inhibitors of the adaptive or the innate immunity (i.e., monoclonal antibodies against IL-15 and inhibitors of the neutrophil migration)[33,34].

Moreover, potential treatments for DH are those aimed to induce tolerance to gluten with bioengineered probiotic strains of microorganisms that could be administered orally or with intranasal administration of recombinant α-gliadin[3,35]. For example, strains of Lactococcus lactis that expressed a gluten-derived DQ8-restricted gliadin peptide[36] or secreted IL-10[37] were found to inhibit the activation of gluten-specific T cells or to treat mouse models of colitis, respectively.

A further alternative approach to a GFD is the correction of the intestinal barrier defect that, together with genetic predisposition and environmental triggers, is thought to be one of the most important factors in the development of autoimmunity. Recent evidence suggest that, in CD, the gluten-induced up-regulation of zonulin, a peptide involved in tight junction regulation, is responsible, at least in part, for the increase in gut permeability and the subsequent abnormal passage of gluten into the lamina propria[38]. Interestingly, the zonulin inhibitor AT1001, tested in recent clinical trials[39], was shown to be tolerated and to reduce gluten-induced intestinal barrier dysfunction, pro-inflammatory cytokine production, and gastrointestinal symptoms in celiac patients.

Finally, the results from recent reports lead to an interesting question: is a lifelong commitment to GFD always necessary in patients with DH Very recently Paek et al[40] reported that 12% of their DH patients experienced remission, defined as absence of skin lesions and symptoms of DH for more than 2 years while not adhering to a GFD, suggesting that clinicians should continually re-evaluate the need for medical therapy and a GFD for their patients with well-controlled DH, with the idea that DH might actually be in remission in some patients. However, this approach needs confirmation in larger cohorts of patients and, in particular, an accurate histopathologic examination of the intestinal mucosa and a serologic evaluation of anti-tTG antibodies should be performed after the reintroduction of gluten-containing foods into the diet, to ensure that the remission of DH was also associated with the remission of the intestinal disease[41].

Moreover, the need for a GFD should be questioned even in Japanese patients who, as previously discussed, seem to have a decreased frequency of CD and lack the typical HLA haplotypes associated to the development of both DH and CD[11-13].

FUTURE PERSPECTIVES

As previously mentioned, recent reports from the Literature suggest that the features as well as the diagnostic criteria of DH should be carefully revised. In particular, the sensitivity and the specificity of the DIF, that is considered the gold standard for the diagnosis of DH, should be re-evaluated. In fact, recent reports showed that skin biopsies from patients with CD but without DH can show granular deposits of IgA at the dermal-epidermal junction[19]. Accordingly, a clinical trial investigating a large series of CD patients for the findings at DIF (both on DH “predilection” and “non-predilection” sites) should be performed.

A second important point about DH is its relationship with CD and with the HLA DQ2/DQ8 haplotype, that are currently considered two of the main features of the disease. However, as reported above, Japanese DH patients seem not to show those features in all the cases[2]; moreover, patients with “paraneoplastic DH” with all the classic clinical, histopathological and immunopathological findings of DH, but without intestinal involvement and no response to a GFD were recently reported[42,43]. Do these patients really have DH Or should CD and response to gluten be considered mandatory features and, therefore, patients without intestinal involvement should be classified as having a skin disease different from DH-although sharing common clinical, histopathological and immunopathological findings To address these questions and understand whether both CD-associated and non CD-associated forms of DH do exist, several studies mainly focusing on HLA and non-HLA genetic background and on potential specific serologic markers of DH but not of CD should be performed.

Focusing on the CD-associated forms, according to Fry et al[44], a very interesting issue still remains: why do some people with CD develop DH and other do not An answer to this interesting question is unlikely to be found only investigating the humoral immunologic response against eTG; studies analyzing the different T cell populations, including Th2 and Th17 lymphocytes specific for the several antigens involved in the CD/DH spectrum (eTG, tTG, gliadin), would be probably required. Moreover, the use of experimental models of DH would help to increase the knowledge of the disease. Recently, besides the mouse model by Zone et al[16] described previously, another interesting model was developed[45]. It was a gluten-sensitized non-obese diabetic DQ8+ mouse that satisfied various requirements for an animal model of DH: gluten sensitivity, Major Histocompatibility Complex II dependence for gluten sensitivity, and replication of many histopathological and immunological features of the lesional skin of DH. There were, however, three main differences between that model and the majority of DH patients (the absence of CD4+ T cells in the lesional dermis, the absence of entheropathy and the lack of colocalization of eTG with the IgA deposition on the perilesional areas), suggesting that other models closer to the disease should be developed. However, this model could also suggest the existence of a subtype of human DH overlapping its features.

Accordingly, also this model rises the problem about the presence of enteropathy in patients with DH and, therefore, the question of whether enteropathy is a crucial element of DH pathogenesis or a gluten-specific immune response without gut involvement is sufficient for the occurrence of the disease. Accordingly, as a significant minority of DH patients (10%) does not have detectable enteropathy, but do have an immune response to gluten, it might be possible that the villous destruction is an independent consequence of the inflammatory response toward gluten in the intestine of DH patients[3].

CONCLUSION

Finally, it could be hypothesised the existence of patients with non CD-associated DH, showing all the clinical features of DH but without a gluten-specific serologic immune response (Table 1). These subsets would have a great impact on patients with DH, since if a subgroup of them did not have CD, they would be managed even without a GFD that, as reported above, is difficult to comply with. However, further studies are required to confirm such hypothesis and, to date, GFD still remains the best therapeutic option for patients with DH.

Table 1 Hypothetical spectrum of dermatitis herpetiformis.

Classical CD-related DH	Latent/potential CD-related DH	Non CD-related DH
Histopathological features typical of CD	Absence of villous atrophy1	Absence of histopathological features of CD
Presence of gluten specific immunological response	Presence of gluten specific immunological response	Absence of gluten specific immunological response[42,43]

¹Documented by experienced gastroenterologists and pathologists after repeated bowel biopsies. CD: Celiac disease; DH: Dermatitis herpetiformis.