Autoantibodies against MHC class I polypeptide-related sequence A are associated with increased risk of concomitant autoimmune diseases in celiac patients

Celiac disease (CD) was previously considered a relatively rare pathology that appeared only in childhood, but is now recognized as being a very common disease that can be diagnosed at any age [1, 2]. Its most typical characteristics are a strong genetic association with the human leukocyte antigen (HLA) alleles DQ2 and DQ8 [3, 4] and its triggering by an environmental factor, the intake of gluten. The soluble fraction of gliadin has been identified as the cause of this intolerance, but many other gluten proteins can be toxic in CD [5]. These proteins induce an inflammatory process in the intestine of susceptible people, but inflammation regresses after elimination of gluten-containing foods from the diet, leading to the recovery of the structure and function of the mucosa [3].

Autoantibodies, especially those directed against the tissue transglutaminase (tTG) enzyme, commonly appear in CD [6]. These antibodies are very important in the diagnosis of CD, but their role in the pathogenesis of the disease remains controversial [7]. Several studies have suggested that these antibodies are directly involved in the pathogenesis of CD. Zanoni et al. demonstrated the role of anti-TG2 antibodies in gut mucosal damage in patients with CD [8]. Their study showed that these antibodies are able to recognize an epitope that is common to TG2 and the Toll-like receptor 4. This interaction leads to the activation of the Toll-like receptor 4 pathway, an important initiator of innate immunity. Other studies have shown that anti-tTG antibodies may play an important role in epithelial cell proliferation [9] and interfere with intestinal epithelial cell adhesion [10]. Additionally, these autoantibodies disturb angiogenesis and modulate vascular permeability in vitro[11, 12]. Taken together, these observations suggest that future approaches to the study of CD should take the role of humoral immunity into account.

Other mechanisms, such as the MHC class I polypeptide-related sequence A (MICA)- Natural killer group 2, member D (NKG2D) interaction, are directly involved in the pathogenesis of the disease [13]. MICA and MICB show homology with classical HLA-class I, but have no role in antigen presentation. MICA and MICB are cell surface glycoproteins that are constitutively expressed in enterocytes [14]. These proteins are ligands for the killer cell lectin-like receptor subfamily K member 1, also known as NKG2D, which is an activating receptor that is mainly expressed in natural killer, CD8+ and γδT cells [15]. The MICA-NKG2D interaction in natural killer cells induces their cytolytic capacity, while in CD8+ T cells it acts as a co-stimulatory signal and complements antigen recognition by the T cell receptor [16]. MICA is strongly expressed in the enterocytes of patients with CD in response to the indirect toxic effect of gluten [17]. Furthermore, MICA binds the NKG2D receptor expressed on CD8+ intraepithelial lymphocytes and activates these T cells. This activation causes damage to the enterocytes, and could be the initiating event that ultimately leads to villous atrophy.

The tissue damage and increased expression of MICA may also induce the development of antibodies against this molecule. In fact, anti-MICA autoantibodies have been described in early-onset systemic lupus erythematosus (SLE) [18]. Moreover, these antibodies have been implicated in organ rejection in patients with renal [19‐21] and cardiac transplants [22, 23].

To test the hypothesis that the CD-related changes in gut mucosa may be associated with the development of antibodies against MICA, we investigated the presence of these antibodies in sera obtained from patients with active CD. We also considered the possibility that these antibodies play a role in the development of the additional autoimmune diseases (ADs) customarily associated with CD [24].

The mechanisms leading to systemic autoimmune aggression in patients with CD remain unknown. NKG2D and their ligands could play a significant role in the development of autoimmunity. In fact, these molecules are involved in the pathogenesis of some ADs that are commonly associated with CD, such as type 1 diabetes and rheumatoid arthritis. For example, blockage of NKG2D in the pre-diabetic stage in non-obese diabetic mice prevents the development of diabetes [33]. In rheumatoid arthritis, MICA and MICB are aberrantly expressed in pathological tissue from affected joints and could be involved in the continuation of the autoreactive process [34].

We found anti-MICA antibodies to be present in half of the patients with untreated CD but in only 3.5% of healthy controls. Moreover, anti-MICA antibodies were present in most patients diagnosed with CD plus other ADs, but were rare in patients solely with ADs. Furthermore, anti-MICA antibodies disappeared from most patients after a year on a GFD, similar to what occurs with anti-tTG antibodies. These results imply that, generally, the relationship between anti-MICA and associated ADs is directly linked to CD rather than being a phenomenon specifically associated with autoimmunity. However, as previously mentioned, early-onset SLE may be an exception to this pattern, since these antibodies have also been described in this disease [18].

Previous studies of CD show that MICA protein is overexpressed in enterocytes obtained from the damaged mucosa of patients [13, 17]. A high level of MICA expression seems to promote mucosal damage by infiltrating intraepithelial CD8+ T lymphocytes, which express the MICA receptor, NKG2D. In fact, other authors have suggested that overexpression of MICA could be an initiating event in the progression of mucosal damage [35‐37]. In the early phase of CD, some gluten-derived peptides, which have been characterized as non-immunogenic, are directly involved in increasing expression of MICA and IL-15. These ‘toxic’ peptides activate the innate immunity through an unknown pathway, leading to the destruction of the gut mucosa [35]. As has been demonstrated in other inflammatory processes, massive cellular destruction may be involved in the development of autoantibodies [38]. In CD, an initially higher level of expression of MICA due to gluten-derived peptides followed by the destruction of enterocytes by intraepithelial lymphocytes could trigger the subsequent development of anti-MICA antibodies in susceptible individuals.

The fact that fewer than half of patients with CD have anti-MICA prevents the use of this antibody as a diagnostic marker. However, the greater prevalence of these autoantibodies in patients with additional ADs observed in our study implies that anti-MICA antibodies might be useful for predicting the risk of AD development. ADs are diseases that are frequently associated with CD. Several studies have found a close association between a long period of gluten exposure [15, 39] and a common genetic background, although others found no relationship between ADs and prolonged gluten intake in CD [40]. We found that ADs in our group of patients were more frequent in adults than in the young, but this could be related to the established effect of age on the prevalence of these diseases, rather than due to the duration of gluten exposure. Moreover, the association of anti-MICA autoantibodies with the presence of additional ADs is substantially modified by age. Surprisingly, although ADs are more frequent in adults with CD, the risk associated with anti-MICA antibodies of developing concomitant AD is clearly greater in childhood, so the determination of these autoantibodies may be very useful in clinical practice to establish the risk in children. In our study, only one pediatric patient with ADs was negative for anti-MICA autoantibodies. In children, CD and other ADs are usually more aggressive processes than in adults. In CD, this could result in increased expression of MIC molecules in younger patients, which would favor the generation of autoantibodies. The presence of anti-MICA autoantibodies could augment tissue destruction and lead to the early development of ADs. As mentioned previously, Dai et al. also demonstrated the presence of anti-MICA autoantibodies in 27 patients with juvenile-onset SLE [18]. Our group of 428 patients with AD included 92 diagnosed with SLE. Six of these patients were classified as juvenile-onset SLE and all had anti-MICA (data not shown). Although these antibodies have not been directly implicated in the disease, they might be an additional risk factor in the development of SLE and an interesting biomarker in the diagnostic of this disease in childhood.

We have shown here that the development of anti-MICA antibodies is associated with a gluten-containing diet in patients with CD. In addition, anti-MICA antibodies are apparently related to the development of ADs, especially in younger patients although the number of individuals in this study group was small. The determination of these autoantibodies might be less suitable than others for the diagnosis of the disease, but it could be very useful in clinical practice for predicting the development of associated ADs in patients with CD.

The mechanism by which these antibodies could influence the development of autoimmunity might be related to direct tissue damage, the activation of cellular immunity, or another mechanism similar to that demonstrated for anti-tTG in other studies [8‐12]. Further work is needed to establish the role and significance of these autoantibodies in CD and, possibly, in other ADs such as SLE, especially in patients with a young age of onset.