Wheat prolamins, gliadins, and glutenins can induce various gastrointestinal complaints in humans. In genetically susceptible people with human leukocyte antigens (HLA) DQ2 or DQ8, gliadin and related prolamins in rye and barley also can trigger celiac disease, an autoimmune disorder characterized by the immune reaction against gliadin peptides and a self-protein, type-2 transglutaminase (TG2). Patients with celiac disease develop severe villous atrophy in their small intestine, and also may suffer from extraintestinal disease, typically involving organs rich in TG2 (eg, liver, heart, skin, brain) (1). Patients with celiac disease elaborate autoantibodies against TG2, which can be detected from the blood and which also deposit on TG2 in diseased tissues (2). The production of TG2-specific antibodies is gluten dependent and stops after a sufficiently long time on a gluten-free diet (3), but the exact mechanisms of the autoantibody production are not fully elucidated. The most commonly suggested model is the hapten-carrier model, where gliadin-specific T lymphocytes provide help for the antibody production of both gliadin and TG2-specific B lymphocytes (3,4), but molecular mimicry also was implicated (5–11). However, neither of these hypotheses has been proven experimentally.
TG2-specific antibodies are highly specific for celiac disease and their production is restricted to individuals carrying HLA-DQ2 or HLA-DQ8. Antibodies against gliadin, in contrast, can be present in various disorders with intestinal inflammation and damage (eg, food allergy, Crohn disease) and also in healthy people without celiac-type genetic HLA background (3). In addition, sensitivity of the gliadin antibody test can vary from 30% to 100%, and did not prove to be satisfactory in clinical practice (12). For these reasons, gliadin antibody determinations are no longer recommended by current guidelines in the diagnostic workup of celiac disease (13,14), and were replaced by more sensitive and specific tests that detect transglutaminase autoantibodies by enzyme-linked immunosorbent assay (ELISA) or by the binding of antibodies to tissue sections containing TG2 in endomysial or reticulin structures (EMA test). These transglutaminase-based tests have 90% to 98% sensitivity and 95% to 98% specificity for untreated celiac disease (12).
Gliadin is a complex mixture of cereal storage proteins, and elicits an extended immune response toward several epitopes in patients with celiac disease (15). Most important, T and B cell epitopes are generated by an ordered deamidation of certain glutamine residues by the enzymatic function of transglutaminase or by the slightly acidic environment in the gut (16–18). These modified deamidated gliadin peptides (DGPs) were shown to have higher affinity to HLA-DQ2 and HLA-DQ8 molecules of antigen-presenting cells, and thereby induce enhanced T cell response. The study of B cell epitopes showed that deamidated hexapeptide motifs QPEQPF and QEQPFP are recognized with higher efficacy by celiac antibodies than corresponding native gliadin peptides (17). In the study by Schwertz et al (19), binding assay with the combined use of PLQPEQPFP and PEQLPQFEE peptides had 94% diagnostic efficiency compared with the 81% diagnostic efficiency of conventional gliadin antibody measurements. Also, clinical studies suggest the superior diagnostic performance of newly developed commercial ELISA tests using deamidated gliadin peptide antigens (20–22). Given that these tests seem to work as well as transglutaminase-based serological tests, the aim of our study was to explore whether structural similarities exist between these gliadin-derived antigens and transglutaminase, and whether deamidated gliadin peptides could mimic transglutaminase epitopes.
PATIENTS AND METHODS
Serum samples were available from 74 untreated celiac patients with EMA positivity and Marsh grade III villous atrophy (median age, 7 years; range, 1.6–39.7 years) and from 65 EMA-negative nonceliac controls with normal small bowel villous architecture (median age, 13 years; range, 0.9–42.4 years). The controls were investigated for various gastrointestinal complaints; the clinical diagnoses were Crohn disease (n = 7), food allergy (n = 7), gastroesophageal reflux (n = 8), chronic nonspecific diarrhea (n = 18), and functional disorders (n = 25). All of the subjects included had normal serum total IgA levels.
Monoclonal mouse antibodies CUB7402 and TG100 were obtained from NeoMarkers (Fremont, CA), and polyclonal goat anti-TG2 antibodies were obtained from Upstate (Charlottesville, VA). Further noncommercial monoclonal TG2-specific mouse antibodies (MAbs) were gifts from the sources listed in Table 1 and acknowledged here. All of the used MAbs were experimentally generated in mice, and none of the clones originated from celiac patients.
Expression and Purification of Recombinant TG2 Proteins
Full-length human recombinant N-terminally His-tagged TG2 (amino acids 1–687) was expressed in Rosetta 2 cells (Novagen, Darmstadt, Germany). pGEX-2T-TG2 DNA (24) was amplified with specific primers (primer 1, 5′-gac gac gac aag atg aga att cag acc atg gcc gag gag ctg g-3′ and primer 2, 5′-gag gag aag ccc ggt tga att cgg tta ggc ggg gcc aat gat gac-3′), and subcloned into pET-30 Ek/LIC Vector to obtain full-length human recombinant N-terminally His-tagged TG2 (amino acids 1–687). Domain deletion mutant TG2s were created by the same technique with a template pcDNA 3.1/myc-His mammalian expression vector (Invitrogen, Carlsbad, CA) containing the following combinations of TG2 structural domains: A/I+II+III (amino acids 1–584), B/I+II+IV (amino acids 1–471 and 585–687), C/II+III+IV (amino acids 139–687), D/I+III+IV (amino acids 1–138 and 472–687), and E/His-tag only, without the sequence of the enzyme. The templates were gifts from Prof Soichi Kojima (Molecular Cellular Pathology Research Unit, RIKEN, Wako, Saitama, Japan). TG2 mutant 1 to 648 was generated using primer 3, 5′-gac gac gac aag atg aga att cag acc atg gcc gag gag ctg g-3′ and primer 4, 5′-gag gag aag ccc ggt tga att cgg tta aac ttc ctc ccc tgc c-3′. The sequence of the mutants was confirmed by DNA sequencing using the PRISM 3100-Avant Genetic Analyzer (Applied Biosystems, Foster City, CA).
Rosetta 2 cells were transformed with the expression vectors and grown in lysogeny broth at 37°C to an OD600 0.6 to 0.8. To induce the expression of His-tagged proteins, the cultures were grown for 5 hours at 20°C in the presence of 0.3-mmol/L isopropyl β-D-thiogalactoside, and cells were harvested by centrifugation at 4°C. After cell lysis by sonification, His-tagged proteins were purified by Ni Sepharose High Performance column (GE Healthcare Bio-Sciences AB, Uppsala, Sweden), eluted with buffer containing 250 mmol/L imidazole, and concentrated with Amicon Centricon-YM 50 MW (Millipore, Billerica, MA). Then the buffer was exchanged to 20 mmol/L Tris-HCl, pH 7.2, with 150 mmol/L NaCl, 1 mmol/L dithiothreitol, 1 mmol/L ethylenediaminetetraacetic acid, and 10% by volume glycerol.
DGP-reactive antibodies were measured from patient serum samples and antibody preparations by Quanta Lite Celiac DGP Screen test kit (Inova Diagnostics, San Diego, CA) according to the manufacturer's instructions. This kit contains ELISA plate coated with synthetic peptides corresponding to deamidated gliadin peptides and a combined horseradish–peroxidase conjugate that recognizes both IgA- and IgG-class human antibodies. The peptide sequences in the kit are proprietary. The results were expressed as percentage values of a positive control sample, with values ≥20 U being positive.
The same plates also were probed with mouse monoclonal TG2-specific antibodies using the buffers provided in the kit. The mouse antibodies were recognized by horseradish and peroxidase–conjugated anti-mouse antibodies (DAKO, Glostrup, Denmark) diluted 1:4000 in the sample diluent or in tris-buffered saline with 0.1% Tween-20 and 3,3′-5,5′-tetramethyl-benzidine (Sigma Chemical, St Louis, MO). In some experiments, the DGP-ELISA plates were preincubated with 8 mol/L urea, 6 mol/L guanidine, 4 mol/L potassium thiocyanate, 200 mmol/L dithiothreitol, 100 mmol/L N-ethyl-maleimide, 100 mmol/L iodoacetamide, or 100 mmol/L cysteamine before the addition of antibodies.
TG2 antibodies from patient serum samples were measured with the Celikey kit (Phadia, Freiburg, Germany); the cutoff value for positivity was ≥5 U/mL. TG2 antibodies from antibody preparations also were measured by coating the full-length and mutant recombinant TG2 proteins to Maxisorp ELISA plates (Nunc A/S, Roskilde, Denmark) in 20 mmol/L Tris-HCl, pH 7.2, with 150 mmol/L NaCl and 2 mmol/L CaCl2, running the ELISA as previously described (27). Antigen amounts in this ELISA were adjusted to obtain similar optical density values with the high positive (100 U) calibrator samples from the Celikey kit and the Celiac DGP Screen test. To obtain comparable results for the different antibody levels in the affinity purified antibody preparations, the results were uniformly calculated as percentages of a common positive calibrator sample, similarly as in the DGP-ELISA. The cutoff for positivity by this method was 7 U or higher.
Endomysial Antibody Testing
Endomysial binding of patient serum samples, TG2-specific MAbs, and purified antibodies was tested by using unfixed frozen sections from human umbilical cord and nonceliac human jejunum tissue biopsy samples with normal villous architecture, as described earlier (28). Initial sample dilution was 1:2.5 in phosphate buffered saline (PBS). Samples that were nonreactive in this dilution were labeled negative; samples with endomysial binding were titrated to the highest dilution at which the reaction was still visible.
Monoclonal TG2-specific antibodies were added to the commercial DGP-ELISA plates in the presence of increasing amounts of human recombinant TG2 constructs or celiac patient serum samples. The His-tagged construct devoid of the sequence of the enzyme was used as negative control. The OD450 signal of the blank was defined as 100% inhibition, and the signal without added TG2 or celiac serum was defined as 0% inhibition. Competition studies also were carried out using ELISA plates coated with recombinant human TG2 constructs and soluble DGP as competitor. Soluble DGP had >95% purity and was donated by Inova Diagnostics.
Affinity Purification of DGP-antibodies From Patient Serum Samples
The 300-μg DGP was coupled to 0.3 g of CH-Sepharose 4B (Pharmacia Fine Chemicals, Uppsala, Sweden) in 0.1 mol/L NaHCO3/NaH2CO3, pH 8.0) containing 0.5 mol/L NaCl. Excess reactive groups were blocked by 0.1 mol/L Tris-HCl/0.5 mol/L NaCl. Then the resin was equilibrated with PBS containing 0.05% Tween-20 (PBST) and incubated for 30 minutes with the patient sera diluted in PBST. After centrifugation, the supernatant was saved for analysis, and the resin was extensively washed with PBST supplemented with 0.35 mol/L NaCl. The peptide-bound antibodies were eluted with 0.1 mol/L glycine-HCl, pH 2.5, neutralized with 2 mol/L Tris, dialyzed against PBS, and concentrated on Microsep 30 K columns (Pall, East Hills, NY).
Clinical Performance of the DGP-ELISA Test
All 74 EMA-positive untreated celiac disease patients were positive for both TG2 antibodies and DGP antibodies as measured by the Celikey and the Celiac DGP Screen tests. Among the controls, 1 subject was positive with the Celikey kit and 1 other with the DGP kit. Thus, the sensitivity of both tests was 100% and the specificity was 98.5% for detecting untreated celiac disease. Individual TG2 and DGP antibody levels in celiac patients showed a positive correlation (r = 0.83).
Investigation of the DGP Antigen by Nonceliac Monoclonal TG2-specific Antibodies
In contrast to conventional anti-gliadin antibody measurements, the DGP-ELISA worked in our patients with similar efficacy and in similar ways as TG2 antibody detection. We thus explored whether some antigens in the DGP test could be similar to TG2 epitopes and would be recognized also by other nonceliac TG2-specific antibodies. We tested 13 different monoclonal mouse TG2-specific antibodies in the DGP-ELISA test, and 3 of them gave clearly positive results (Fig. 1). A similar reactivity was not found with crude gliadin extract (data not shown). The 3 DGP-reactive TG2-MAbs were clone 925, clone 4E1, and clone 895, which were referred to as MAb1, MAb2, and MAb3 in this study.
The 13 TG2-MAbs have different binding sites as characterized by earlier studies (Table 1 and unpublished data), spanning all 4 domains of TG2. MAb1 has a linear binding site in the core domain, whereas the epitopes of MAb2 (amino acids 637–648) (25) and MAb3 (amino acids 649–687) are in the C-terminal domain of TG2. The epitope for MAb3 is sensitive to chaotropic agents and probably is conformational. MAb1 and MAb3 are able to bind to TG2 because it is exposed in the endomysium and to give EMA-type staining patterns on tissues. MAb2 is unable to bind to the endomysium, and only recognizes TG2 in ELISA.
The DGP antigens on the plate also were recognized by polyclonal goat anti-TG2 antibodies, but to a lesser extent than by MAbs 1 to 3. The used target antigens were synthetic peptides, thus contamination by natural transglutaminase was unlikely. In line with this, commercial TG2-MAb CUB7402, which binds to 447 to 453 amino acids, and TG100 antibody, which binds to 447 to 538 amino acids of TG2, were unable to recognize DGP peptides.
Competition of Transglutaminase and DGP for Antibody Binding
Next, we investigated whether the binding of the TG2-MAbs to DGP was specific and could be inhibited by transglutaminase or its fragments. Full-length recombinant human TG2 completely inhibited the binding of MAb1, MAb2, and MAb3 to DGP, and the inhibition was dose dependent (Fig. 2). TG2 constructs not containing the respective binding domains for the MAbs and the empty His-construct did not inhibit the binding, thus the binding was specific for the specific TG2-binding sequence of the MAbs. Although MAb2 did not recognize mutant TG2 1 to 648 in an earlier study in ELISA, its binding to DGP was effectively inhibited by this construct, but not by TG2 1 to 584, which contains only domains I+II+III (Fig. 3). This result confirms the binding site of MAb2 to be different from that of MAb3. Together these results indicate the presence of multiple epitopes in DGP that can resemble TG2. Competition studies also were conducted with MAb3 in the opposite direction, using TG2 immobilized to the ELISA plate and soluble DGP as the competitor. Binding of MAb3 to TG2 was inhibited by 40% at 40 μg/well DGP concentration. Because of the limited amount of available DGP, this experiment was not carried out with all 3 MAbs.
The binding of celiac patient samples to DGP could not be abolished by either recombinant TG2 or any of the MAbs, which indicates that they also contain additional antibodies to DGP epitopes that are different from TG2. Similarly, soluble DGP at concentrations up to 10 μg/wells did not interfere with celiac antibody binding to TG2.
Reaction of Affinity-purified DGP Patient Antibodies With Endomysium and Transglutaminase
Antibodies were affinity purified with DGP from serum samples of 4 untreated celiac patients and tested in ELISA and immunofluorescent studies on human umbilical cord and normal human jejunum. The efficacy of DGP antibody extraction from the original serum samples was approximately 90% in 3 of the 4 cases, whereas EMA titers remained unchanged and transglutaminase antibody concentrations only slightly decreased in the supernatant after extraction with DGP preparations (Table 2). The DGP-purified antibodies did not bind to endomysial structures. However, they reacted with recombinant human TG2 in ELISA and also gave reactivity with the Celikey antigen.
These results collectively show that certain DGP and TG2 epitopes can have a similar appearance, and moreover, that patients with celiac disease elaborate antibodies that can react both with DGP and TG2. This antibody population is a fraction of TG2 antibodies that does not react with endomysium.
Epitope Studies and Molecular Modeling
To further explore the epitopes of DGP that may resemble TG2, we attempted to modify the conformation of DGP by treating the ELISA plate with denaturing and chaotropic agents and checking whether these modifications influence the binding of the antibodies. However, we did not find any difference in antibody binding after having incubated the DGP plate with 8 mol/L urea, 6 mol/L guanidine, 4 mol/L potassium thiocyanate, or high concentrations of various reducing and alkylating agents, whereas such treatments have reduced the binding of celiac antibodies or that of the conformation-sensitive MAb3 to TG2.
We also performed search and alignment with TG2 and published deamidated gliadin sequences (17–19) using the DNAstar and ClustalW platforms (http://www.ebi.ac.uk/clustalw), and Visual Molecular Dynamics modeling (29). Those peptides previously found to be antigenic were only aligned as discontinuous sequences with constituents located too far from each other on the 3-dimensional structure of TG2 to form a common epitope. Moreover, there was no linear homology at all for the 649 to 687 part of TG2. Interestingly, the known target epitope of MAb2 in human TG2 (637EIPDPVEAGEEV648 (25)) also differs from the gliadin sequences. MAb2 also is able to recognize guinea pig TG2, which contains an adjacent glutamic acid (E) and glutamine (Q) in the corresponding position, and this EQ motif commonly found in deamidated gliadin sequences may form an anchor point for this antibody. An exact evaluation of potentially homologue sequences could not be done because the primary amino acid sequences of the used DGP peptides were not available publicly. These findings suggest that the 3-dimensional appearance and not the primary sequences of the DGP peptides have importance in the similarity to TG2.
In this article we used several approaches to establish whether epitopes in type-2 transglutaminase and deamidated gliadin peptides are structurally related. Our results clearly show that deamidated gliadin epitopes are recognized by transglutaminase-specific monoclonal antibodies, and antibodies affinity purified with deamidated gliadin peptides do react with transglutaminase 2. These findings shed light on new relations between gliadin and transglutaminase and suggest the possibility of potential molecular mimicry in the induction of celiac autoantibodies targeted against TG2.
We found that a fraction of celiac patient antibodies indeed cross-reacts with both deamidated gliadin epitopes and TG2. These antibodies do not bind to endomysium, thus they represent an additional antibody population to conventional celiac antibodies, which bind both to endomysium and TG2 in ELISA. Patient antibodies are polyclonal, thus also may contain additional antibodies against deamidated gliadin peptides and TG2 that do not overlap. This explains why TG2 preparations or DGP could not completely inhibit the reactivity to each other in this study, or in the recently published work of Liu et al (30).
Endomysium is an anatomical entity, a connective tissue sheet surrounding smooth muscle cells, which can be visualized by silver staining. It consists of collagen fibers and also contains fibronectin and fibronectin-bound TG2, which can be removed from the tissues by chemical agents disrupting the fibronectin-TG2 binding (28). In early studies, the celiac autoantigen was shown to be associated with the surface of collagen fibers (31).
Although all of the endomysial antibodies in humans seem to be TG2-targeted antibodies (32), some TG2-specific antibodies, including MAb2 and some other published monoclonal anti-TG2 antibodies (33), cannot bind to endomysium, depending on whether their binding epitope is accessible when TG2 is complexed to fibronectin and other tissue proteins. The celiac antibodies targeting both deamidated gliadin peptides and TG2 also seem to represent such an antibody population. This finding may have importance in EMA-negative celiac disease (34), which is an important diagnostic challenge in adults but also in young children. Detection of celiac antibodies by deamidated gliadin peptides may increase diagnostic sensitivity in such cases (22,35).
The weakness of this study is that primary amino acid sequences of the used commercial deamidated gliadin peptides were unavailable; therefore we were unable to exactly identify the corresponding epitopes in TG2. Our results, however, indicate that several such epitopes do exist and 2 of them are located in the C-terminal domain of TG2, where other important celiac epitopes also were found (36). Homology search with published deamidated gliadin sequences did not provide good hits to this region, but it also may be possible that the corresponding regions are just mimotopes with similar 3-dimensional surface appearance despite different underlying primary amino acid sequences. Our experimental results with urea and other chemical agents support this possibility because we were not able to see any difference in the epitope recognition after attempts to modify the secondary and tertiary structure of the peptides. In contrast, such modifications in TG2 led to the loss of the corresponding epitopes. Several lines of evidence already have been collected that suggest celiac antibodies recognize conformational epitopes in TG2 and are probably composed of discontinuous sequences (25,36,37). Therefore, merely the shape of the binding sites and not the primary sequences seem to have importance in the similarity.
Short linear peptide sequences are usually poor targets for antibodies in ELISA, and need some support or structural motifs to stabilize their conformation (19). In the citrullinated peptides used in the diagnosis of rheumatoid arthritis, the desired conformation could be stabilized by producing cyclic derivatives of the antigen (38). In deamidated gliadin peptides, the exact way that this 3-dimensional shape of the antigen was industrially achieved is not known.
Molecular mimicry between an exogen (usually microbial) antigen and a self-antigen can initiate autoimmune processes (39), and after spreading of the antibody response to further self-epitopes, an autoimmune disease can be established (40). We show here for the first time that deamidated gliadin peptides share common epitopes with TG2 and can have the potential to trigger such events. In line with this, DGP-reactive antibodies were shown to appear early during the development of celiac autoimmunity in a prospectively followed cohort of young children, and in some cases, reactivity to DGP did indeed precede the detectability of conventional TG2 antibodies (29). In our study, the youngest patient in the affinity purification experiment had a relatively larger proportion of antibodies against the shared DGP and TG2 epitopes than older children (Table 2), also supporting this hypothesis. The initial triggering of DGP and TG2-reactive (but EMA negative) antibodies also may explain the higher frequency of EMA-negative celiac cases among children younger than 2 years old.
Recently, some rotavirus peptides also were shown to be similar to TG2, and celiac patients were found to have antibodies to rotavirus proteins (7) or to seroconvert to them before TG2 autoantibody production (41). It also is possible that 2 exogen triggers, rotavirus infection and gliadin intake, can cooperate in the induction of autoantibodies, as well as that shared DGP-TG2 epitopes can further trigger antibody production via hapten-carrier mechanisms.
Interestingly, the antibody response to deamidated gliadin peptides seems to be missing in subjects without celiac disease. If the formation of these peptides results from the normal deamidation action of TG2, then they would be expected to form in the intestine of normal subjects as well. Thus, further work is needed to elucidate why these peptides are antigenic only in subjects with celiac disease and whether they are being handled differently, prevented from penetration, or degraded before deamidation in healthy people.
In conclusion, our results indicate that certain deamidated gliadin–derived peptides have similar 3-dimensional appearance to epitopes in transglutaminase, and patients with celiac disease elaborate antibodies that can react both with these peptides and transglutaminase. This antibody population may have a role in the induction of celiac disease.
The authors thank Inova Diagnostics for providing synthetic deamidated gliadin peptides for this study.
1. Green PH, Jabri B. Coeliac disease. Lancet 2003; 362:383–391.
2. Korponay-Szabo IR, Halttunen T, Szalai Z, et al
. In vivo targeting of intestinal and extraintestinal transglutaminase 2 by coeliac autoantibodies. Gut 2004; 53:641–648.
3. Mäki M. Autoantibodies as markers of autoimmunity in coeliac disease pathogenesis. In: Gastrointestinal Immunology and Gluten-sensitive Diseases. Dublin: Oak Tree Press; 1992. pp. 246–252.
4. Sollid LM, Molberg O, McAdam S, et al
. Autoantibodies in coeliac disease: tissue transglutaminase—guilt by association? Gut 1997; 41:851–852.
5. Barbeau WE, Novascone MA, Elgert KD. Is celiac disease due to molecular mimicry between gliadin peptide-HLA class II molecule-T cell interactions and those of some unidentified superantigen? Mol Immunol 1997; 34:535–541.
6. Kagnoff MF, Austin RK, Hubert JJ, et al
. Possible role for a human adenovirus in the pathogenesis of celiac disease. J Exp Med 1984; 160:1544–1577.
7. Zanoni G, Navone R, Lunardi C, et al
. In celiac disease, a subset of autoantibodies against transglutaminase binds toll-like receptor 4 and induces activation of monocytes. PLoS Med 2006; 3:e358.
8. Tuckova L, Tlaskalova-Hogenova H, Farre MA, et al
. Molecular mimicry as a possible cause of autoimmune reactions in celiac disease? Antibodies to gliadin cross-react with epitopes on enterocytes. Clin Immunol Immunopathol 1995; 74:170–176.
9. Natter S, Granditsch G, Reichel GL, et al
. IgA cross-reactivity between a nuclear autoantigen and wheat proteins suggests molecular mimicry as a possible pathomechanism in celiac disease. Eur J Immunol 2001; 31:918–928.
10. Hadjivassiliou M, Boscolo S, Davies-Jones GA, et al
. The humoral response in the pathogenesis of gluten ataxia. Neurology 2002; 58:1221–1226.
11. Alaedini A, Okamoto H, Briani C, et al
. Immune cross-reactivity in celiac disease: anti-gliadin antibodies bind to neuronal synapsin I. J Immunol 2007; 178:6590–6595.
12. Rostom A, Dube C, Cranney A, et al
. The diagnostic accuracy of serologic tests for celiac disease: a systematic review. Gastroenterology 2005; 128(Suppl 1):S38–46.
13. Rostom A, Murray JA, Kagnoff MF. American Gastroenterological Association (AGA) Institute technical review on the diagnosis and management of celiac disease. Gastroenterology 2006; 131:1981–2002.
14. Hill ID, Dirks MH, Liptak GS, et al
. Guideline for the diagnosis and treatment of celiac disease in children: recommendations of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition. J Pediatr Gastroenterol Nutr 2005; 40:1–19.
15. Vader W, Kooy Y, Van Veelen P, et al
. The gluten response in children with celiac disease is directed toward multiple gliadin and glutenin peptides. Gastroenterology 2002; 122:1729–1737.
16. Van de Wal Y, Kooy Y, van Veelen P, et al
. Selective deamidation by tissue transglutaminase strongly enhances gliadin-specific T cell reactivity. J Immunol 1998; 161:1585–1588.
17. Osman AA, Gunnel T, Dietl A, et al
. B cell epitopes of gliadin. Clin Exp Immunol 2000; 121:248–254.
18. Aleanzi M, Demonte AM, Esper C, et al
. Celiac disease: antibody recognition against native and selectively deamidated gliadin peptides. Clin Chem 2001; 47:2023–2028.
19. Schwertz E, Kahlenberg F, Sack U, et al
. Serologic assay based on gliadin-related nonapeptides as a highly sensitive and specific diagnostic aid in celiac disease. Clin Chem 2004; 50:2370–2375.
20. Sugai E, Vazquez H, Nachman F, et al
. Accuracy of testing for antibodies to synthetic gliadin-related peptides in celiac disease. Clin Gastroenterol Hepatol 2006; 4:1112–1117.
21. Prince HE. Evaluation of the INOVA diagnostics enzyme-linked immunosorbent assay kits for measuring serum immunoglobulin G (IgG) and IgA to deamidated gliadin peptides. Clin Vaccine Immunol 2006; 13:150–151.
22. Kaukinen K, Collin P, Laurila K, et al. Resurrection of gliadin antibodies in coeliac disease. Deamidated gliadin peptide antibody test provides additional diagnostic benefit. Scand J Gastroenterol
, June 19, 2007; 1–6. [Epub ahead of print].
23. Trejo-Skalli AV, Velasco PT, Murthy SN, et al
. Association of a transglutaminase-related antigen with intermediate filaments. Proc Natl Acad Sci USA 1995; 92:8940–8944.
24. Jeon JH, Choi KH, Cho SY, et al
. Transglutaminase 2 inhibits Rb binding of human papillomavirus E7 by incorporating polyamine. EMBO J 2003; 22:5273–5282.
25. Di Niro R, Ferrara F, Not T, et al
. Characterizing monoclonal antibody epitopes by filtered gene fragment phage display. Biochem J 2005; 388:889–894.
26. Mohan K, Pinto D, Issekutz TB. Identification of tissue transglutaminase as a novel molecule involved in human CD8+ T cell transendothelial migration. J Immunol 2003;171:3179–86. 26a. Ambrus A, Banyai I, Weiss MS, et al. Calcium binding of transglutaminases: a 43Ca NMR study combined with surface polarity analysis. J Biomol Struct Dyn 2001;19:59–74.
27. Sulkanen S, Halttunen T, Laurila K, et al
. Tissue transglutaminase autoantibody enzyme-linked immunosorbent assay in detecting celiac disease. Gastroenterology 1998; 115:1322–1328.
28. Korponay-Szabo IR, Sulkanen S, Halttunen T, et al
. Tissue transglutaminase is the target in both rodent and primate tissues for celiac disease–specific autoantibodies. J Pediatr Gastroenterol Nutr 2000; 31:520–527.
29. Humphrey W, Dalke A, Schulten K. VMD: visual molecular dynamics. J Mol Graph 1996; 14:27–28.
30. Liu E, Li M, Emery L, et al
. Natural history of antibodies to deamidated gliadin peptides and transglutaminase in early childhood celiac disease. J Pediatr Gastroenterol Nutr 2007; 45:293–300.
31. Karpati S, Meurer M, Stolz W, et al
. Ultrastructural binding sites of endomysium antibodies from sera of patients with dermatitis herpetiformis and coeliac disease. Gut 1992; 33:191–193.
32. Korponay-Szabo IR, Laurila K, Szondy Z, et al
. Missing endomysial and reticulin binding of coeliac antibodies in transglutaminase 2 knockout tissues. Gut 2003; 52:199–204.
33. Lock RJ, Gilmour JE, Unsworth DJ. Anti-tissue transglutaminase, anti-endomysium, and anti–R1-reticulin autoantibodies—the antibody trinity of coeliac disease. Clin Exp Immunol 1999; 116:258–262.
34. Salmi TT, Collin P, Korponay-Szabo IR, et al
. Endomysial antibody-negative coeliac disease: clinical characteristics and intestinal autoantibody deposits. Gut 2006; 55:1746–1753.
35. Agardh D. Antibodies against synthetic deamidated gliadin peptides and tissue transglutaminase for the identification of childhood celiac disease. Clin Gastroenterol Hepatol 2007; 5:1276–1281.
36. Sblattero D, Florian F, Azzoni E, et al
. The analysis of the fine specificity of celiac disease antibodies using tissue transglutaminase fragments. Eur J Biochem 2002; 269:5175–5181.
37. Seissler J, Wohlrab U, Wuensche C, et al
. Autoantibodies from patients with coeliac disease recognize distinct functional domains of the autoantigen tissue transglutaminase. Clin Exp Immunol 2001; 125:216–221.
38. Schellekens GA, Visser H, de Jong BA, et al
. The diagnostic properties of rheumatoid arthritis antibodies recognizing a cyclic citrullinated peptide. Arthritis Rheum 2000; 43:155–163.
39. Oldstone MB. Molecular mimicry, microbial infection, and autoimmune disease: evolution of the concept. Curr Top Microbiol Immunol 2005; 296:1–17.
40. Farris AD, Keech CL, Gordon TP, et al
. Epitope mimics and determinant spreading: pathways to autoimmunity. Cell Mol Life Sci 2000; 57:569–578.
41. Stene LC, Honeyman MC, Hoffenberg EJ, et al
. Rotavirus infection frequency and risk of celiac disease autoimmunity in early childhood: a longitudinal study. Am J Gastroenterol 2006; 101:2333–2340.
© 2008 Lippincott Williams & Wilkins, Inc.