Coeliac disease (CD) is a chronic inflammatory disorder of the small intestine characterised by an immune response to ingested gliadin and related cereal proteins, affecting approximately 1 in 100 white individuals (1). Traditionally, CD was considered a disease of childhood, presenting with the classical symptoms of diarrhoea, malabsorbtion, and failure to thrive (2); however, the mean age of diagnosis has risen, with CD often presenting with more subtle and variable clinical symptoms (3).
In CD, autoimmunity arises via a complex interplay of environmental (gliadin) and genetic factors, with the strongest genetic association being with the human leukocyte antigen (HLA) genes (4). The majority of patients express the HLA-DQ2 molecule (5), with a minority expressing the HLA-DQ8 molecule (6). Continued gliadin ingestion results in the characteristic small intestinal histological profile of villous atrophy, crypt hyperplasia, and infiltration of mononuclear cells in the lamina propria (7). Serum immunoglobulin (Ig) A antibodies to tissue transglutaminase (tTG), identified as the major autoantigen in CD by Dieterich et al (8), are a sensitive and specific serological test for diagnosis of the disease (9).
tTG is a member of a family of related, calcium-dependent transglutaminase enzymes responsible for the posttranslational modification of proteins. Proteins can be cross-linked by tTG through the formation of an isopeptide bond between glutamine and lysine residues on each target protein (10). Under certain conditions, tTG will deamidate the target glutamine in the substrate protein, with the neutral glutamine being changed to a negatively charged glutamic acid residue (11). The deamidation reaction, catalysed by tTG, is of paramount importance in CD pathogenesis. The high proline and glutamine content of gliadin makes it an excellent substrate for deamidation by tTG, generating negative charges that strengthen binding at crucial anchor positions in the HLA-DQ2/DQ8 molecule (12). This facilitates presentation of deamidated gliadin peptides to CD4+ T cells in the intestinal lamina propria, and amplifies the immune response seen in CD.
The transamidation and deamidation reactions of tTG are mediated by a papain-like catalytic triad of amino acids, which is relatively conserved across all of the mammalian transglutaminases (13). In tTG, this comprises cysteine277, histidine335, and aspartic acid358, contained in the catalytic core domain. Previous studies by our group (14) and others (15) have demonstrated that CD IgA anti-tTG autoantibodies target the enzyme's catalytic core, representing a highly focused and specific autoantibody response. We have shown, using wild type and a site-directed mutagenic tTG lacking the entire Cys-His-Asp catalytic triad, that the antibody binding of adult CD anti-tTG antibodies was focused on this region, and its mutation caused a highly significant drop in antibody binding.
The aim of the present study was to map the evolution of this highly specific autoantibody response by comparing the specificity of paediatric and adult anti-tTG responses. This should shed light on the nature of the antibody response at the onset of disease: that is, does the paediatric anti-tTG response also focus on the catalytic triad of tTG, or is this a phenomenon that evolves during disease progression? The contribution of individual amino acids of the catalytic triad to CD autoantibody epitopes in both adults and children was also assessed.
In the present study, a series of novel tTG active-site mutants were generated, containing single and double amino acid substitutions of the catalytic triad, as well as the previously described triple mutant (14). The mutated tTG proteins were then applied as antigens in an enzyme-linked immunosorbent assay (ELISA) system with serum from adult and paediatric patients with CD, with reactivity of each mutant tTG then compared to the wild-type tTG protein. An IgG anti-tTG response also occurs in CD, but is less disease specific than the IgA response (16). To further examine whether the targeting of the tTG active site in CD is an isotype-specific IgA phenomenon, or whether there are shared IgA/IgG tTG epitopes, the epitope specificity of the IgG anti-tTG response in CD was also assessed.
Serum samples from 93 patients with CD were used to assess tTG autoantibody binding. The adult group consisted of 30 patients (male:female [M:F] 1:2.8; age 22–85 years, median 59), whereas the paediatric group consisted of 63 patients (M:F 1:1.7; aged 2–15 years, median 8). The diagnosis of CD was based upon positive IgA endomysial antibodies and anti-tTG serology, reduction in levels of these autoantibodies with gluten withdrawal, and duodenal histology where available. Both CD study populations were further subdivided by disease activity status. Individuals with partially treated CD were defined as patients with low to medium levels of IgA anti-tTG (20–50 arbitrary units [AU], Celikey ELISA system, Pharmacia Diagnostics, Uppsala, Sweden), in which serial tTG measurement (Celikey; Pharmacia Diagnostics) has demonstrated falling titres of IgA anti-tTG levels postcommencement of a gluten-free diet. In the paediatric CD group, 23 patients had untreated CD, 15 patients had partially treated CD, and 25 patients had treated CD. In the adult CD group, 5 were treated, 13 were partially treated, and 12 were untreated. Cutoff points for IgA and IgG anti-tTG ELISA positivity (mean + 2 standard deviation) were established by assaying serum from groups of normal adult (n = 30, M:F 1:2, age 24–82 years, median 52) and paediatric (n = 57, M:F 1:1, age 1–15 years, median 8) individuals with negative IgA anti-tTG serology, as measured by the Celikey ELISA system. All of the adult noncoeliacs had normal intestinal histology. All of the serum samples were obtained from the immunology laboratory, St James’ Hospital. Ethical approval for the present study was granted from St James’ Hospital and Our Lady's Children's Hospital, Dublin.
Polymerase Chain Reaction, Site-directed Mutagenesis and Cloning
The wild-type tTG cDNA was cloned into pGEX-4T-1 vector, as described previously (14). Selected residues of the catalytic triad Cys-His-Asp were replaced with alanine residues via site-directed mutagenesis, using the Quickchange system (Stratagene, La Jolla, CA) (Table 1), as per the method of Byrne et al (14). Escherichia coli BL-21 were then transformed with the appropriate sequence-containing plasmid and stored on bacterial cryobeads at −70°C until needed.
Protein Production and Purification
Cultures were grown in nutrient broth to OD600 ∼0.5 and protein expression was induced by addition of 1 μmol/L isopropyl β-D-1-thiogalactopyranoside followed by incubation overnight at room temperature with 150-rpm agitation. The bacteria were pelleted and resuspended in 20 mmol/L Tris, 150 mmol/L NaCl, 1 mmol/L EDTA, 1 mM dithiothreitol, 15% glycerol, pH 8.0, and pelleted at 3500g for 15 minutes. Pellets were stored at −80°C before purification. For purification of the recombinant tTGs, pellets were resuspended in 50-mL CelLytic reagent (Sigma-Aldrich, St Louis, MO) containing 1-mL protease inhibitor cocktail (Sigma-Aldrich), 10-mg lysozyme (Sigma-Aldrich), and 10-μL benzonase endonuclease (Sigma-Aldrich). Lysate was incubated at room temperature for 20 minutes with mixing and subsequently pelleted at 15,000g for 15 minutes at 4°C. Two millilitres of glutathione sepharose (GE Healthcare, Little Chalfont, UK) was added to the resulting supernatant, and incubated for 1 hour at 4°C, with mixing. The sepharose was washed 4 times with phosphate buffered saline (PBS), and the glutathione-S-transferase (GST)-fusion protein eluted from the beads with 1-mL elution solution (100 mmol/L glutathione, 20 mmol/L Tris, pH 8.0). Recombinant tTG was mixed with 50% glycerol to prevent freezing, and stored at −20°C. Purity of protein batches was routinely checked by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were quantified using the Bradford assay (Sigma-Aldrich).
ELISA was performed for recombinant wild-type and mutant tTGs. Certified 96-well Maxisorp plates (Nunc, Roskilde, Denmark) were coated with 0.3-μg recombinant protein per well in coating buffer (50 mmol/L Tris, 150 mmol/L NaCl, 5 mmol/L CaCl2, pH 7.5) overnight at 4°C. Wells were blocked with 5% casein in PBS for 1 hour, and washed 4 times with PBS containing 0.01% Tween between each step. Control antibodies CUB 7402 (Abcam, Cambridge, UK), polyclonal anti-tTG (Abcam), 5G7G6 (mouse monoclonal anti-tTG, a gift from Dr Fernando Chirdo), and anti-GST (Abcam) were used as primary antibodies to characterise each recombinant tTG protein. CUB 7402 was diluted 1:5000 in PBS plus 0.01% Tween, followed by horseradish peroxidase (HRP)–conjugated rabbit anti-mouse (Dako, Copenhagen, Denmark) diluted 1:2500 in PBS plus 0.01% Tween. Polyclonal anti-tTG was diluted 1:1000 in PBS plus 0.01% Tween, followed by HRP-conjugated swine anti-rabbit (Dako) diluted 1:2500 in PBS plus 0.01% Tween. 5G7G6 was diluted 1:5000 in PBS plus 0.01% Tween, followed by HRP-conjugated rabbit anti-mouse (Dako) diluted 1:2500 in PBS plus 0.01% Tween. Rabbit anti-GST was diluted 1:5000, followed by HRP-conjugated swine anti-rabbit (Dako) diluted 1:2500.
To generate AU for the reporting of the anti-tTG ELISA results, a standard curve was constructed for both the IgA and IgG assays, using 4 pooled CD serum samples with strong IgA anti-tTG positivity as measured by the Celikey ELISA system. For the IgA anti-tTG ELISA, the range of dilutions was 1:320 to 1:20,480, whereas for the IgG anti-tTG ELISA, the range was 1:80 to 1:2560. The highest point of each standard curve was then assigned a value of 100 AU. For IgA assays, human serum was diluted 1:100 in PBS plus 0.01% Tween, followed by HRP-conjugated polyclonal rabbit antihuman IgA (Dako) diluted 1:2500 in PBS plus 0.01% Tween. For IgG assays, human serum was diluted 1:1000 in PBS plus 0.01% Tween, followed by HRP-conjugated polyclonal rabbit antihuman IgG (Dako) diluted 1:2000 in PBS plus 0.01% Tween. All of the antibody incubations were performed for 1 hour at room temperature. ELISAs were developed using the 3,3′,5,5′-tetramethylbenzidene liquid substrate system (Sigma-Aldrich), and OD450 was measured with a Biotek ELx800 spectrophotometer (Biotek, Winooski, VT).
Differences between groups were evaluated for statistical significance using the Mann-Whitney test. The Wilcoxon signed-rank test was used to evaluate differences in individual patient binding to the mutant tTG proteins. The level of significance was set at 0.05. The GraphPad Prism 5 program (GraphPad Software, San Diego, CA) was used for the statistical analysis.
Polymerase Chain Reaction, Site-directed Mutagenesis and Cloning
A total of 5 recombinant human tTG proteins were generated. Required amino acid substitutions in the catalytic triad of amino acids Cys277, His335, and Asp358 were performed by site-directed mutagenesis, using various permutations of alanine screening. Table 1 lists the recombinant tTG proteins, with mutations introduced: CHDΔtTG, for example, had all 3 residues of the catalytic replaced with alanine residues, whereas HDΔtTG had His335 and Asp358 replaced. DNA sequencing (MWG Eurofins Operon Ltd, Ebersberg, Germany) confirmed the sequence identity of each recombinant protein, and successful introduction of mutations in the catalytic triad.
Protein Expression, Purification, and Quantification
All of the recombinant human tTGs were expressed as GST-fusion proteins, and purified in E coli BL-21. Because the GST tag is 26 kDa and tTG is 77 kDa, a band of 103 kDa was visible upon SDS-PAGE. SDS-PAGE analysis showed all of the recombinant proteins to be >95% pure (Fig. 1). A standard Bradford assay was used to calculate yields of the recombinant proteins, which were, on average, 1 mg/L of bacterial culture (data not shown).
Characterisation of Wild-type and Mutant tTGs
The ability of control antibodies to tTG to bind to each recombinant protein was assessed by ELISA. As a control for conformational integrity and coating efficiency of the wild-type and mutant proteins, both monoclonal murine (CUB 7402, 5G7G6) anti-tTG and polyclonal anti-tTG were assayed with each protein. Similar binding patterns were observed for each recombinant protein, with the exception being a decrease in 5G7G6 binding to CΔ tTG (Fig. 2).
CD serum IgA reactivity to wild-type tTG correlated strongly with IgA anti-tTG results from the Celikey tTG ELISA system for all samples tested (r = 0.785). There was a significant difference in recognition of the wild-type tTG protein between patients with CD and age-matched controls (P < 0.0001 for both paediatric and adult CD, Mann-Whitney test). The range of IgA anti-tTG values was reflective of treatment status, with untreated patients displaying the highest levels in both children and adults.
In paediatric CD, mean IgA anti-tTG levels to the wild-type tTG were 96.3 AU for untreated patients, 47.5 AU for partially treated patients, and 13.3 AU for treated patients (cutoff 3.9 AU). In paediatric CD, mean reductions in IgA autoantibody binding compared with the wild-type protein were 89% for CΔtTG (P < 0.001), 90% for CHΔtTG (P < 0.001), 87% for CHDΔtTG (P < 0.001), and 91% for HDΔtTG (P < 0.001) (Wilcoxon signed-rank test).
IgG autoantibodies to wild-type tTG were more frequently detectable in untreated CD, with the highest levels being observed in untreated patients younger than 4 years. Mean IgG anti-tTG levels to the wild-type tTG in paediatric CD were 61.2 AU for untreated patients, 4.4 AU for partially treated patients, and 7.6 AU for treated patients (cutoff 15.3 AU). In paediatric CD, there was a significant difference in IgG anti-tTG binding to wild-type tTG between paediatric patients with CD and controls for untreated patients only (P = 0.0003, Mann-Whitney test). Paediatric patients with CD with untreated disease had significantly higher levels of IgG anti-tTG compared with those who were partially treated (P = 0.0063, Mann-Whitney test), or treated (P = 0.0042, Mann-Whitney test).
The mean reduction in binding to mutant tTG proteins, compared with wild-type tTG, was 33% in paediatric CD (range −100% to 180%). The mean reduction in binding was 47% for CΔtTG (P < 0.001), 43% for CHΔtTG (P < 0.001), 32% for CHDΔtTG (P < 0.001), and 32% for HDΔtTG (P < 0.001) (Wilcoxon signed-rank test). When the paediatric CD group was further subdivided by treatment status, the mean reduction in binding to all mutant proteins was highest in patients with untreated disease. The mean reduction in binding to all 4 mutants, compared with the wild-type protein, was 67% in untreated CD, 47% in partially treated CD, and 8% in treated CD.
In adult CD the pattern of reduced binding to mutant tTG proteins was replicated. In the IgA assay, there was a mean reduction in binding to mutant tTG proteins of 76% (range −99% to 47%, (P < 0.001 for each mutant protein, Wilcoxon signed-rank test) (Fig. 3A and B). Mean IgG anti-tTG levels in adult CD were 7.7 AU for untreated patients, 8.0 AU for partially treated patients, and 4.9 AU for treated patients (cutoff 5.3 AU). In adult CD there was a clear distinction in IgG recognition of the wild-type protein between patients and controls for patients with untreated (P = 0.0006, Mann-Whitney test) and partially treated CD (P = 0.0002, Mann-Whitney test). Reduction of IgG anti-tTG binding to the mutant tTG proteins was also observed in adult CD, with a mean reduction of 22% for all mutated tTGs (Fig. 4A and B). In adults with CD, the highest level of reduction in IgG anti-tTG binding to the mutant tTG proteins was seen in partially treated disease. The mean reduction in binding to all 4 mutants, compared with the wild-type protein, was 20% in untreated adult CD, 29% in partially treated adult CD, and 11% in treated adult CD.
In the present study, we mapped the evolution of the autoantibody response to tTG, demonstrating specific targeting of the core region early in the disease course, and through the generation of unique active-site mutants of tTG, investigated the importance of single residues of the tTG catalytic triad for autoantibody binding. In the mean, the active-site mutations in tTG reduce binding of IgA anti-tTG by 82% and IgG anti-tTG binding by 28% for the paediatric and adult CD study groups (Figs. 3 and 4). Replacement of even a single amino acid of the catalytic triad with an alanine residue (CΔ tTG) caused a marked reduction in CD autoantibody binding. The diminished recognition of all of the mutant tTGs by CD autoantibodies indicates the importance of the core region for CD antigenicity, with an intact catalytic triad of amino acids, cysteine277, histidine335, and aspartic acid358 required for autoantibody binding. This is as one would expect for a B-cell epitope, with amino acids distant in the linear protein sequence being in close proximity upon protein folding (17). There is a possibility that the introduction of mutations in the core region of tTG could have affected protein folding, and hence conformation-dependent autoantibody binding. A similar pattern of recognition by control monoclonal and polyclonal anti-tTG antibodies was observed for the wild-type and mutant tTG proteins, suggesting correct protein folding. The similar recognition of all of the proteins by the monoclonal anti-tTG CUB 7402 may be especially pertinent because the epitope recognised by this antibody (amino acids 447–478) is in the core region of tTG.
Production of anti-tTG in CD occurs at local level in the mucosal-associated lymphoid tissue, a system that is under constant antigenic challenge and stimulation, and has evolved to preferentially produce IgA to neutralise pathogens and build a tolerance to dietary antigens. Multiple rounds of somatic hypermutation and variable diversity joining region recombination occur at mucosal level, the result being affinity matured and highly specific IgA antibodies (18). It is possible that the generation of highly specific IgA anti-tTG antibodies is a result of affinity maturation and epitope focusing over a prolonged disease duration (19). To examine this hypothesis, we compared the anti-tTG specificity of both adult and paediatric patients with CD, with the paediatric population representing an earlier stage of disease evolution and duration. Many subtle immunological differences, which may have relevance in the aetiology and pathogenesis of CD, have been noted between adults and children. In studies investigating the T-cell responses to gliadin peptides, Vader et al (20) demonstrated that T cells from children with CD target a more diverse set of epitopes when compared with adult patients. Furthermore, paediatric T cells reacted to nondeamidated gliadin peptides more frequently than adult T cells, suggesting that the childhood T-cell response changes over time, focusing on immunodominant epitopes. A more rapid response to gluten withdrawal has also been noted in children, with the possibility of regaining tolerance to gluten (21). In the present study, we demonstrated a highly focused and specific anti-tTG response in paediatric CD, indicating that the core region of tTG is targeted from early on in disease progression.
The demonstration of autoantibodies directed against enzymes is a recurring feature of autoimmune disease (22–25); however, the targeting of a single region of tTG by CD autoantibodies is an interesting immunological phenomenon. Generally speaking, autoimmune responses directed against self-proteins, particularly enzymes, are of the IgG class and involve the recognition of multiple antigenic sites. Multiple IgG autoantibody epitopes have been described for proteinase 3 in Wegener granulomatosis (26) and for GAD65 in type 1 diabetes mellitus (27). Interestingly, the antigenic regions on these 2 proteins include their active site. Our results indicate that the IgG anti-tTG response is specifically dependent on an intact tTG active site, similar to the IgA anti-tTG response. This would concur with reports by Sblattero et al (15), who have demonstrated a similar binding pattern for both IgA and IgG anti-tTG to truncated fragments of recombinant human tTG. We have also observed higher levels of IgG anti-tTG in untreated paediatric patients with CD younger than 4 years, possibly reflective of an evolving mucosal immune system, which is not yet completely proficient at preferentially class switching to IgA. Additionally, a significant IgG anti-tTG response was detected in 15% of the paediatric control group (data not shown). These individuals were negative for IgA anti-tTG, and showed similar recognition of the wild-type and mutant tTG proteins, implying structural integrity and correct folding of the mutant tTG proteins. It is possible that these children were suffering from an infectious disease because this has been implicated for anti-tTG positivity in noncoeliac individuals (28).
There is a possibility that some enzymes, by their very nature, are predisposed to becoming targets of autoimmunity. Their active site may normally be sequestered from the immune system during inactivity, or new intermediates formed during catalysis can induce autoimmunity via neo-epitope formation. Certainly, a strong argument can be made for both of these processes occurring with tTG. It is known that under inactive conditions, the active site of tTG is buried deep within a narrow cleft between domains 3 and 4 of the protein (10) and that calcium activation of tTG causes a conformational change that exposes the active site (29). tTG/gliadin complexes occur, with tTG capable of cross-linking gliadin to itself at the active site. This could potentially freeze the enzyme in its “open” configuration, with exposure of the active site, and possible neo-epitope formation (30). The recently described increased sensitivity of CD IgA anti-tTG detection when tTG is in its “open” conformation supports this notion (31).
The pathological role of anti-tTG antibodies in CD remains controversial. Conflicting reports exist as to the degree, if any, of inhibition of the enzymatic function of tTG by autoantibodies, and of the biological relevance of demonstrable inhibition (32–34). In the present study, we confirmed the targeting of the tTG active site by CD autoantibodies, and one would expect some interference with enzymatic function, if only prevention of substrate access to the active site. We demonstrated significant inhibition of tTG function by CD anti-tTG IgA (35).
To conclude, we have demonstrated the presence of a highly conformational epitope in the core region of tTG that is specifically targeted by CD autoantibodies from early on in the disease course. We believe that this is a result of intramolecular epitope spreading from gliadin to tTG, possibly as a result of the formation of gliadin/tTG complexes at the active site of tTG. The potential contribution to CD pathology and the effects on tTG function of these autoantibodies requires further elucidation.
1. Dube C, Rostom A, Sy R, et al. The prevalence of celiac disease in average-risk and at-risk western European populations: a systematic review. Gastroenterology
2. Maki M, Collin P. Coeliac disease. Lancet
3. Ravikumara M, Tuthill DP, Jenkins HR. The changing clinical presentation of coeliac disease. Arch Dis Child
4. Bourgey M, Calcagno G, Tinto N, et al. HLA related genetic risk for coeliac disease. Gut
5. Kagnoff MF. Celiac disease: pathogenesis of a model immunogenetic disease. J Clin Invest
6. Karell K, Louka AS, Moodie SJ, et al. HLA types in celiac disease patients not carrying the DQA1*05-DQB1*02 (DQ2) heterodimer: results from the European Genetics Cluster on Celiac Disease. Hum Immunol
7. Marsh MN. Gluten, major histocompatibility complex, and the small intestine. A molecular and immunobiologic approach to the spectrum of gluten sensitivity (’celiac sprue’). Gastroenterology
8. Dieterich W, Ehnis T, Bauer M, et al. Identification of tissue transglutaminase as the autoantigen of celiac disease. Nat Med
9. Rostom A, Dube C, Cranney A, et al. The diagnostic accuracy of serologic tests for celiac disease: a systematic review. Gastroenterology
10. Griffin M, Casadio R, Bergamini CM. Transglutaminases: nature's biological glues. Biochem J
11. Fleckenstein B, Molberg O, Qiao SW, et al. Gliadin T cell epitope selection by tissue transglutaminase in celiac disease. Role of enzyme specificity and pH influence on the transamidation versus deamidation process. J Biol Chem
12. Vader LW, de Ru A, van der Wal Y, et al. Specificity of tissue transglutaminase explains cereal toxicity in celiac disease. J Exp Med
13. Lorand L, Graham RM. Transglutaminases: crosslinking enzymes with pleiotropic functions. Nat Rev Mol Cell Biol
14. Byrne G, Ryan F, Jackson J, et al. Mutagenesis of the catalytic triad of tissue transglutaminase abrogates coeliac disease serum IgA autoantibody binding. Gut
15. 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
16. Picarelli A, di Tola M, Sabbatella L, et al. Identification of a new coeliac disease subgroup: antiendomysial and anti-transglutaminase antibodies of IgG class in the absence of selective IgA deficiency. J Intern Med
17. Rubinstein ND, Mayrose I, Halperin D, et al. Computational characterization of B-cell epitopes. Mol Immunol
18. Cerutti A. The regulation of IgA class switching. Nat Rev Immunol
19. Cucnik S, Kveder T, Krizaj I, et al. High avidity anti-beta 2-glycoprotein I antibodies in patients with antiphospholipid syndrome. Ann Rheum Dis
20. 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
21. Matysiak-Budnik T, Malamut G, de Serre NP, et al. Long-term follow-up of 61 coeliac patients diagnosed in childhood: evolution toward latency is possible on a normal diet. Gut
22. Barera G, Bazzigaluppi E, Viscardi M, et al. Macroamylasemia attributable to gluten-related amylase autoantibodies: a case report. Pediatrics
23. Bogdanos DP, Dalekos GN. Enzymes as target antigens of liver-specific autoimmunity: the case of cytochromes P450 s. Curr Med Chem
24. McLachlan SM, Rapoport B. Autoimmune response to the thyroid in humans: thyroid peroxidase--the common autoantigenic denominator. Int Rev Immunol
25. Auger I, Balandraud N, Rak J, et al. New autoantigens in rheumatoid arthritis (RA): screening 8268 protein arrays with sera from patients with RA. Ann Rheum Dis
26. Williams RC Jr, Staud R, Malone CC, et al. Epitopes on proteinase-3 recognized by antibodies from patients with Wegener's granulomatosis. J Immunol
27. Fenalti G, Rowley MJ. GAD65 as a prototypic autoantigen. J Autoimmun
28. Ferrara F, Quaglia S, Caputo I, et al. Anti-transglutaminase antibodies in non-coeliac children suffering from infectious diseases. Clin Exp Immunol
29. Pinkas DM, Strop P, Brunger AT, et al. Transglutaminase 2 undergoes a large conformational change upon activation. PLoS Biol
30. Fleckenstein B, Qiao SW, Larsen MR, et al. Molecular characterization of covalent complexes between tissue transglutaminase and gliadin peptides. J Biol Chem
31. Lindfors K, Koskinen O, Kurppa K, et al. Serodiagnostic assays for celiac disease based on the open or closed conformation of the autoantigen, transglutaminase 2. J Clin Immunol
32. Dieterich W, Trapp D, Esslinger B, et al. Autoantibodies of patients with coeliac disease are insufficient to block tissue transglutaminase activity. Gut
33. Kiraly R, Vecsei Z, Demenyi T, et al. Coeliac autoantibodies can enhance transamidating and inhibit GTPase activity of tissue transglutaminase: dependence on reaction environment and enzyme fitness. J Autoimmun
34. Esposito C, Paparo F, Caputo I, et al. Anti-tissue transglutaminase antibodies from coeliac patients inhibit transglutaminase activity both in vitro and in situ. Gut
35. Byrne G, Feighery C, Jackson J, et al. Coeliac disease autoantibodies mediate significant inhibition of tissue transglutaminase. Clin Immunol