AIM: To indirectly determine if tissue transglutaminase (tTG)-specific T cells play a crucial role in the propagation of celiac disease.

METHODS: Anti-deamidated gliadin peptide (DGP) and anti-tTG IgA and IgG were measured in the sera of celiac patients (both untreated and treated). The correlations were determined by Spearman’s rank correlation test.

RESULTS: In celiac patients, we found a very significant correlation between the production of DGP IgA and IgG (r = 0.75), indicating a simultaneous and ongoing production of these two isotypes reminiscent of oral vaccination studies. However, there was far less association between the production of tTG IgA and tTG IgG in celiac patients (r = 0.52). While tTG IgA was significantly correlated with DGP IgA (r = 0.80) and DGP IgG (r = 0.67), there was a weak correlation between production of anti-tTG IgG and the production of anti-DGP IgA (r = 0.38) and anti-DGP IgG (r = 0.43).

CONCLUSION: These data demonstrate that the production of anti-tTG IgA is directly correlated to the production of anti-DGP IgG and IgA, whereas anti-tTG IgG is only weakly correlated. This result therefore supports the hapten-carrier theory that in well-established celiac patients anti-tTG IgA is produced by a set of B cells that are reacting against the complex of tTG-DGP in the absence of a tTG-specific T cell.

Celiac disease is a gluten-sensitive disease that afflicts primarily the small bowel, resulting in the shortening of villi, increased numbers of intraepithelial lymphocytes, and crypt hyperplasia[12]. One unique feature of celiac disease that is utilized as a diagnostic and screening tool is the production of IgA specific for tissue transglutaminase (tTG) that circulates in the blood[3]. It is unclear though, as to why these antibodies are generated when a celiac patient eats gluten. One association between tTG and gliadin is that intestinal T cells from celiac patients respond to specific gliadin peptides that have been deamidated, a process that is mediated by tTG binding to gliadin peptides[4]. Anti-tTG IgA is tightly associated with the development of enteropathy and brings into question whether it is a cause or consequence of enteropathy[56]. It is especially perplexing because many celiac patients will produce IgA against whole gliadin, a storage protein of gluten, yet this production has a much lower specificity for celiac disease than the anti-tTG IgA ELISA assay[7].

One theory for the origin of anti-tTG IgA was proposed in 1997[8]. This proposal was based on a hapten-carrier model wherein gliadin-specific T cells contribute to the stimulation of B cells that are specific for tTG. This would be achieved by the tTG-specific B cells internalizing complexes of tTG and gliadin peptides and later presenting gliadin epitopes to the gliadin-specific T cells. In this manner, gliadin-specific T cells would contribute to the tTG-specific B cells producing antibodies against tTG. At that time, this proposal was supported by a lack of evidence for tTG-specific T cells. It is notable that to this date, there is still no evidence that a tTG-specific T cell exists in the small intestine of celiac patients. Of course, it is difficult to prove the absence of a cell type. However, novel ELISAs have been recently developed that can detect antibodies against deamidated gliadin peptides (DGP) in celiac patients[910]. This allows us to look at this process in an indirect manner, by determining the correlation among the production of antibody isotypes against DGP and tTG.

The production of IgA and IgG specific for DGP and tTG was evaluated in celiac patients and plotted such that a direct comparison was made between the production of IgG versus IgA for each antigen group and each patient group (Figure 1). There was a significantly stronger correlation between the production of IgA and IgG specific for DGP (r = 0.75) in celiac patients than those specific for tTG (r = 0.52). When untreated celiac patients (gluten-containing diet; GCD) were separated from treated celiac patients (GFD), the correlation coefficients in comparing anti-DGP IgG and IgA were 0.78 for GCD and 0.58 for GFD, whereas a significantly lower correlation was found for comparing anti-tTG IgG and IgA (r = 0.60 for GCD and r = 0.44 for GFD).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Effect of diet upon isotype correlations. The titers of IgG and IgA against DGP and tTG were evaluated and plotted against each other for celiac patients. For mixed (treated and untreated) celiac patients, the Spearman’s rank correlation coefficients were r = 0.75 for DGP (A) and r = 0.52 for tTG (B). For untreated celiac patients, r = 0.78 for DGP (C) and r = 0.60 for tTG (D). For treated celiac patients, r = 0.58 for DGP (E) and r = 0.44 for tTG (F).

Comparisons were also made between the production of anti-tTG IgA and the production of DGP IgA and IgG in celiac patients (Figure 2). Anti-tTG IgA was highly correlated with the production of both anti-DGP IgA (r = 0.80) and DGP IgG (r = 0.67) which was similar to a previous finding[9].

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Comparing anti-tTG IgA and IgG production with anti-DGP IgA and IgG. The titers of anti-tTG IgA (A-B) and anti-tTG IgG (C-D) were compared with the titers of anti-DGP IgA (A and C) as well as anti-DGP IgG (B and D) in all treated and untreated celiac patients. Spearman’s rank correlation coefficients were 0.80 (A), 0.67 (B), 0.38 (C), and 0.43 (D).

Finally, comparisons were made between the production of anti-tTG IgG and the production of IgA and IgG specific for DGP. In contrast to anti-tTG IgA which was strongly correlated with DGP antibodies, anti-tTG IgG was weakly correlated with the production of anti-DGP IgA (r = 0.38) and anti-DGP IgG (r = 0.43).

The data presented in this manuscript support the theory that the generation of anti-tTG IgA is directly linked to the B cell immune response against DGP, possibly even the T-cell immune response to DGP as well. The reduced correlation in celiac patients between the production of anti-tTG IgG and anti-tTG IgA (r = 0.52) as compared to the production of anti-DGP IgG and anti-DGP IgA (r = 0.75) also demonstrates that there is a fundamental difference between the generation of antibody isotypes against the two antigens in celiac patients. Another difference between the production of IgG and IgA against DGP and tTG is that dietary gliadin mainly affects the production of both IgG and IgA against DGP, but not against both tTG IgG and IgA.

The lack of correlation between the production of anti-tTG IgG and anti-DGP IgG and IgA and anti-tTG IgA (Figures 1 and 2) therefore raises several questions. If the inflammatory T cells that are specific for deamidated gliadin are providing help to the B cells that are specific for a tTG/gliadin complex, why are not more tTG-specific B cells isotype class switching to IgG? One explanation is that the tTG/gliadin-specific B cell group in well-established celiac patients is fully committed to being IgA-positive. A memory B cell could be one such type of long lived IgA-positive B cell, and in this way, require minimal help from a bystander T cell response in order to be fully activated. It is notable that treated celiac patients will relapse with a gluten challenge, even after years of adhering to a GFD, indicating that there is a strong memory component of T cells, B cells, or both in celiac patients[13–15].

Also of interest are studies that demonstrated significant variability in the production of anti-tTG IgA in children. One study found that only six out of 14 celiac patients less than 2 years of age had anti-endomysial antibodies[16]. Another group reported markedly fluctuating levels of transglutaminase antibodies in children[17]. These data would indicate that anti-tTG antibodies do not develop immediately at the time of initial exposure to dietary gliadin, but instead develop after 1-2 (or more) years of continued gliadin exposure.

Our data, as well as the data from others, therefore support the hapten-carrier theory that a B cell that is specific for a tTG/gliadin complex is present in well-established celiac patients and is helped by gliadin-specific T cells. However, our data are also compatible with the model based on molecular mimicry between tTG and gliadin[918]. Indeed, it is our belief that the hapten-carrier model and molecular mimicry model are not exclusive. A potential “combined” model would be that a catalyst-like IgA+ memory B cell specific for regions that are shared between tTG and gliadin exists long term in well-established celiac patients. With the consumption of gliadin, these B cells would internalize tTG/gliadin complexes, become activated with minimal T cell help, and then present gliadin peptides to gliadin-specific T cells. This would result in the amplification of both deamidated gliadin specific T- and B-cell responses, as well as the production of anti-tTG IgA antibodies.