Celiac disease (CD) is triggered by ingested gluten in genetically predisposed individuals who carry the HLA-DQ2 or DQ8 haplotype (1). Gliadin is an alcohol-soluble fraction of wheat gluten that contains a high proportion of glutamines and prolines (2) and patients with CD often have circulating IgA and IgG antibodies against gliadin (3). More frequently, patients with active CD have autoantibodies against endomysium (EMA) (4) and its main endomysial antigen tissue transglutaminase (tTG) (5).
tTG belongs to a group of calcium-dependent enzymes that introduce a post-translational formation of isopeptide bonds between glutamine and lysine residues (6). Gliadins can function as substrates for tTG and tTG molecules themselves, as several other proteins may serve as substrates during the cross-linking process. As a result, supramolecular protein aggregates containing tTG alone or in combination with other proteins are formed and may act as stabilizing factors during the deposition of extracellular matrix in tissue (7). Tissue damage results in an increased extracellular matrix deposition of tTG, and increased levels of tTG have been demonstrated in damaged intestinal mucosa from CD patients (8). However, the physiological function of tTG is still a matter of debate and the pathophysiological role of tTG antibodies in CD remains unclear.
Studies using fragments from cDNA gene products indicate that autoantibodies from CD patients bind to a conformational region located in the core domain of tTG (9). It is possible that these autoantibodies are directed against a heterogeneity of epitopes (10,11). Furthermore, in vitro experiments have shown that autoantibodies from CD patients are capable of interfering with the enzymatic activity of tTG, suggesting that tTG antibodies may be involved in the pathogenesis of CD (12).
For the detection of tTG antibodies a number of different immunoassays have been described (13-19). With the ELISA technique, autoantibodies bind to tTG that is fixed to a solid support and the amount of bound antibodies is quantified. With the radioligand binding assay (RBA), autoantibodies initially react with radiolabeled tTG in a liquid phase and radioactive tTG bound to antibodies is precipitated by beads conjugated with antihuman immunoglobulin antibodies or protein A. Differences in the performance between solid phase ELISA and liquid phase RBA for the detection of autoantibodies have been reported, especially concerning the diagnostic validity of IgG anti-tTG (15,20,21). In theory, there is a risk of native epitopes become hidden, and novel epitopes might be created during the in vitro processing of the antigen used for an immunoassay. From a clinical point of view this technical drawback may imply a decreased sensitivity and specificity dependent on how well the CD relevant epitopes are preserved in both the ELISA and the RBA formats.
The aim of the present study was to investigate how the exposure of human recombinant tTG influenced the binding of IgA and IgG autoantibodies from children with CD.
MATERIALS AND METHODS
Included in this study were 138 serum samples obtained from children consecutively investigated for CD with small intestinal biopsy during the years 1993-2002. CD diagnosis was confirmed according to the revised ESPGHAN criteria (22) at the Department of Pediatrics, University Hospital MAS in Malmö, Sweden. In 2003 all biopsy sections were reexamined blinded by one pathologist according to a modification of Marsh criteria (23). In 15 cases the first and the second evaluation was discordant and these children were therefore excluded from the present study. The remaining 73 children (48 female 25 male; median age, 5.3 years; age range, 0.7-15.3) were diagnosed as having active CD; 17 of them were younger than 2 years of age and two had selective IgA deficiency. A total of 50 children (21 female, 29 male; median age, 3.2 years; age range, 0.9-17.3) were considered to have other disorders than CD. In the disease control group food intolerance or cow's milk protein allergy was diagnosed in 18 children, two had a Giardia lamblia infection, six had failure to thrive and 24 children had transient gastrointestinal symptoms. Eighteen of the control children were younger than 2 years of age. Included were also 80 children (53 female, 27 male; median age, 8.0 years; age range, 1.8-18.9 years) with diagnosed CD and treated with gluten-free diet for median 5.1 years (range, 0.5-16.5 years). The Ethics Committee of Medical Faculty, Lund University approved the intended use of all patient samples and the reexamination of biopsy sections.
Celikey based on human recombinant tTG was used for quantification of IgA anti-tTG, and a research kit with the same antigen was used for detection of IgG anti-tTG (Pharmacia Diagnostics, Freiburg, Germany). The relative amount of tTG antibodies was expressed as UE/mL calculated from standard curves constructed to contain 0, 3, 7, 16, 40, and 100 UE/mL of respective IgA and IgG anti-tTG. Further details are provided elsewhere (24). The cutoff level chosen for the present study was 4 UE/mL for both IgA and IgG anti-tTG according to receiver operating characteristic curves (25).
Human tTG C was synthesized in the presence of 20 μCi 35S-methionine (Perkin Elmer Life Sciences, Inc., Boston, MA) by in vitro transcription and translation as previously described (26,27). The labeled proteins were stored at −80°C and used within 2 weeks, and the RBA was run as described previously (27) with several modifications. The IgA anti-tTG antigen/antibody complexes were isolated with 10% goat antihuman IgA Agarose (Sigma, St. Louis, MO), and the IgG anti-tTG antigen/antibody complexes was separated with 10% rabbit antihuman IgG Sepharose 4B (research coupling, Pharmacia Diagnostics AB, Uppsala, Sweden) or 30% rec-protein A Sepharose (PAS) conjugate 4B (Zymed Laboratories, Inc., San Francisco, CA). The relative amounts of IgA and IgG anti-tTG (UR/mL for RBA and UP/mL for PAS) were calculated from a six-point calibrator curve using the same standard material as used in the ELISA kits (Pharmacia Diagnostics). High binding samples were tested in further dilutions and the relative amount of 100 UR/mL corresponded with 15 UE/mL and 4 UE/mL for IgA and IgG anti-tTG, respectively. Cutoff level was set at 50 UR/mL for IgA-RBA and 50 UP/mL for PAS-RBA, respectively, which correspond with the 99th percentile of healthy controls (28). For IgG-RBA the control group was not available for this study and therefore the cutoff was approximated at 50 UR/mL. Intra-assay and interassay coefficients of variation were 10.6% and 16.5%, respectively.
EMA were analyzed at the Clinical Microbiology and Immunology, Lund University Hospital, Sweden, by an indirect immunofluorescence method (4). Patient serum was diluted 1:10 in phosphate-buffered saline/Tween (Euroimmun, Lubeck, Germany) and applied to reaction fields of reagent tray containing tissue slides of primate intestine, liver and esophagus (Euroimmun). EMA were detected with fluorescein isothiocyanate conjugated goat antihuman IgA antibodies. Results were expressed as the highest dilution factor giving a positive fluorescence pattern in microscope. All sera manifesting fluorescence titer ≥1:10 were considered to be positive.
Differences in tTG antibody levels between independent groups were calculated using the Kruskal-Wallis test and Dunn's multiple comparison test. Correlations between the assessed anti-tTG levels were evaluated using Spearman rank correlation (r). P values <0.05 were considered significant.
tTG Antibodies in Untreated CD Children and Controls
Both IgA assays (IgA-ELISA and IgA-RBA) detected elevated IgA anti-tTG levels in 65 of 73 (89%) untreated CD children; of this group, one child had partial IgA deficiency (total IgA of 0.06 g/L) (Fig. 1). A high positive correlation between the two IgA-assays (r = 0.94, P < 0.0001) was observed for the untreated CD children (Fig. 2). IgA-ELISA and IgA-RBA detected elevated antibody levels in 2 of 50 (4%) and 3 of 50 (6%) disease controls, respectively. The concordance in positive and negative results between the two IgA-assays was 99% for the group of untreated CD children and disease controls. EMA was present in 62 of 73 (85%) of the CD children and in 2 of 50 (4%) controls. These two control subjects had normal small intestinal mucosa.
All three IgG assays (IgG-ELISA, IgG-RBA and PAS-RBA) detected elevated IgG anti-tTG levels in 40 of 73 (55%) untreated CD children; of this group, two children had IgA deficiency (Fig. 1). Moreover, 6 of 73 (8%) untreated CD children with elevated IgG anti-tTG levels were detected with IgG-ELISA and PAS-RBA. Antibody positivity detected with PAS-RBA alone was observed in 20 of 73 (27%) untreated CD children. A high positive correlation was obtained between IgG-ELISA and IgG-RBA (r = 0.91, P <0.0001) for the untreated CD children and a similar correlation between IgG-ELISA and PAS-RBA (r = 0.81, P <0.0001) was observed (Fig. 2).
IgG-ELISA and IgG-RBA detected 2 of 50 (4%) disease controls with elevated IgG anti-tTG levels and PAS-RBA detected 3 of 50 (6%) control children. The concordance in positive and negative results among the untreated CD children and disease controls was 95% when IgG-ELISA and IgG-RBA were compared, 83% when IgG-ELISA and PAS-RBA were compared and 75% when IgG-RBA and PAS-RBA were compared. Moreover, the concordance between IgA-RBA and PAS-RBA was 99%.
The distribution of antibody levels of the 26 of 73 (36%) untreated CD children with elevated IgG anti-tTG levels according to PAS-RBA while also being seronegative with IgG-RBA is shown in Fig. 3. For these children a positive correlation between IgG anti-tTG levels measured with PAS-RBA and the IgA anti-tTG levels was found (IgA-RBA: r = 0.76, P < 0.0001; IgA-ELISA: r = 0.56, P < 0.0001), whereas no correlation was observed between the antibody levels measured with IgG-ELISA or IgG-RBA.
tTG Antibodies in Treated CD Children
None of the treated children had positive EMA titers. A total of 24 of 80 treated children had either elevated IgA or IgG anti-tTG levels and for 8 of 24 children an intestinal biopsy was available; of this group, all showed normal villous architecture and a normal intraepithelial lymphocyte count. IgA-RBA detected elevated IgA anti-tTG levels in 21 of 80 (26%) of the treated CD children who had been on a gluten-free diet for a median period of 4.5 years (range, 0.5-14.8 years), whereas IgA-ELISA detected 3 of 80 (4%) of the treated children (Fig. 1). Furthermore, 6 of 21 children with positive IgA anti-tTG levels according to IgA-RBA had also elevated IgG anti-tTG levels with IgG-ELISA or IgG-RBA. Elevated IgG anti-tTG levels were detected with IgG-RBA in 3 of 80 treated children, and two of them were also detected with IgG-ELISA and PAS-RBA.
Few studies have compared ELISA and RBA when both assays are based on human recombinant tTG. The present study investigated how the exposure of human recombinant tTG influenced the binding of IgA and IgG autoantibodies from untreated and treated children with CD. The approach was to use the most common methods for assessment of CD-associated autoantibodies: an ELISA method where human recombinant tTG is fixed to a solid support, a RBA method where human recombinant tTG is kept in a liquid phase and the measurement of EMA where tTG is attached to the extracellular matrix in tissue.
The binding of IgA anti-tTG levels did not appear to be qualitatively different whether human recombinant tTG was kept in a liquid phase or fixed to a solid support. The same untreated CD children were detected and the correlation between the measured antibody levels was high. A similar agreement between soluble and fixed tTG regarding the binding of IgG anti-tTG was observed when IgG-ELISA and IgG-RBA were compared. IgG-ELISA detected more untreated CD children than IgG-RBA, suggesting that the chosen threshold value was not optimal for the latter assay. The results from the present study showed nonetheless that far from all untreated IgA-sufficient CD children had increased IgG anti-tTG levels, whereas these antibodies were highly elevated in the untreated CD children with IgA-deficiency. These results are in line with other reports indicating a limited diagnostic value of the IgG anti-tTG beyond the identification of IgA deficient CD patients (15,20,24).
It has been suggested that IgG1-EMA or IgG anti-tTG may be used as alternative markers in subgroups of IgA-EMA negative patients without IgA deficiency (29,30). In this study, none of the IgA sufficient untreated CD children without EMA had detectable IgG anti-tTG levels. The remaining children who lacked both EMA and IgA anti-tTG were diagnosed at an age of 2 years or younger. However, all these children had IgA anti-gliadin antibodies (data not shown) and this is in agreement with previous data showing that the diagnostic sensitivity for CD increases when anti-gliadin antibodies are combined with EMA during infancy (3).
A striking difference, however, was noted between the RBA method assessed with protein A and the two other IgG-specific assays for the measurement of IgG anti-tTG. A subgroup of untreated CD children without IgG anti-tTG according to the IgG-RBA method was detected with PAS-RBA and antibody levels correlated with the IgA anti-tTG levels. Obviously, the difference was not dependent on differences in the exposure of tTG. The polyclonal antibody preparation used for the IgG-RBA was specific for human IgG and bound all subclasses equally well (data not shown). In contrast, protein A has the ability to bind several immunoglobulin classes other than IgG, including IgA, while IgG3 is not bound at all. Furthermore, polar binding interactions between protein A and VH3 encoded antibodies have been shown to occur (31). An explanation to the results from the present study could be that in some individuals with untreated CD complexes between tTG and IgA bind to protein A more avidly than free IgG antibodies present in the serum.
Despite the possibility that PAS-RBA may not exclusively measure IgG antibodies against tTG, the method effectively detected the same CD children as both the two other IgA-specific assays and also the untreated CD children with IgA deficiency. This has similarly been reported by others (21) and, moreover, supports the proposed view from a previous screening study that measurement of IgG anti-tTG with PAS-RBA increases the reliability of the IgA anti-tTG test (32). The present study, however, might contain a bias because children with dubious biopsy results were omitted; therefore it is possible that the clinical value of the PAS-RBA reported here is overestimated.
Serum levels of IgA anti-tTG have been reported to disappear even after a short period on a gluten-free diet (24). Furthermore, another report has shown that seroconversion to normal antibody levels precede mucosal normalization in patients with normal serum IgA levels (33). It is therefore notable that 26% of the treated CD children included in the present study had increased IgA anti-tTG levels assessed with the IgA-RBA method despite the fact that several of them had a normal intestinal mucosa. None of these children had EMA and only 4% of them were detected with IgA-ELISA. Whether the observed differences between the ELISA and RBA formats depend on certain epitopes of tTG being hidden when the molecule is immobilized or RBA detecting lower levels of antibodies remains to be investigated. The clinical consequence of this finding also needs to be clarified if sustained levels of IgA autoantibodies indicate either a dietary transgression with an early immune response or the presence of a low response sustaining after mucosal recovery.
In summary, no qualitative difference between ELISA and RBA in detecting IgA or IgG anti-tTG from untreated CD children was demonstrated. However, discrepancies in the binding of IgA anti-tTG from a subgroup of treated CD children indicated that alterations of tTG might occur on fixation of the antigen. Further studies are needed to establish how the fixation affects tTG and whether the PAS-RBA method alone could be used for the diagnosis of CD. Finally, we conclude that in screening of childhood CD IgA anti-tTG detection can be performed either by ELISA or RBA because the two assays are interchangeable.
1. Sollid L. Coeliac disease: dissecting a complex inflammatory disorder. Nat Rev Immunol
2. Wieser H. Relation between gliadin structure and coeliac toxicity. Acta Paediatr
3. Burgin-Wolff A, Gaze H, Hadziselimovic F, et al. Antigliadin and antiendomysium antibody determination for coeliac disease. Arch Dis Child
4. Chorzelski TP, Sulej J, Tchorzewska H, et al. IgA class endomysium antibodies in dermatitis herpetiformis and coeliac disease. Ann N Y Acad Sci
5. Dieterich W, Ehnis T, Bauer M, et al. Identification of tissue transglutaminase
as the autoantigen of celiac disease. Nat Med
6. Mariani P, Carsughi F, Spinozzi F, et al. Ligand-induced conformational changes in tissue transglutaminase
: Monte Carlo analysis of small-angle scattering data. Biophys J
7. Griffin M, Casadio R, Bergamini CM. Transglutaminases: nature's biological glues. Biochem J
8. Hansson T, Ulfgren AK, Lindroos E, et al. Transforming growth factor-beta (TGF-beta) and tissue transglutaminase
expression in the small intestine in children with coeliac disease. Scand J Immunol
9. 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
10. Marzari R, Sblattero D, Florian F, et al. Molecular dissection of the tissue transglutaminase
autoantibody response in celiac disease. J Immunol
11. 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
12. 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
13. Sulkanen S, Halttunen T, Laurila K, et al. Tissue transglutaminase
autoantibody enzyme-linked immunosorbent assay in detecting celiac disease. Gastroenterology
14. Seissler J, Boms S, Wohlrab U, et al. Antibodies to human recombinant tissue transglutaminase
measured by radioligand assay: evidence for high diagnostic sensitivity for celiac disease. Horm Metab Res
15. Sblattero D, Berti I, Trevisiol C, et al. Human recombinant tissue transglutaminase
ELISA: an innovative diagnostic assay for celiac disease. Am J Gastroenterol
16. Hansson T, Dahlbom I, Hall J, et al. Antibody reactivity against human and guinea pig tissue transglutaminase
in children with celiac disease. J Pediatr Gastroenterol Nutr
17. Agardh D, Borulf S, Lernmark A, et al. Tissue transglutaminase
immunoglobulin isotypes in children with untreated and treated celiac disease. J Pediatr Gastroenterol Nutr
18. Williams A, Annis P, Lock R, et al. Evaluation of a high-throughput second antibody radiobinding assay of measuring IgA antibodies to human tissue transglutaminase
. J Immunol Meth
19. Bonamico M, Tiberti C, Picarelli A, et al. Radioimmunoassay to detect antitransglutaminase autoantibodies is the most sensitive and specific screening method for celiac disease. Am J Gastroenterol
20. Van Meensel B, Hiele M, Hoffman I, et al. Diagnostic accuracy of ten second-generation (human) tissue transglutaminase
antibody assays in celiac disease. Clin Chem
21. Bazzigaluppi E, Lampasona V, Barera G, et al. Comparison of tissue transglutaminase
-specific antibody assays with established antibody measurements for coeliac disease. J Autoimmun
22. Walker-Smith J, Guandalini S, Schmitz J, et al. Revised criteria for diagnosis of coeliac disease: Report of Working Group of European Society of Paediatric Gastroenterology and Nutrition (ESPGAN). Arch Dis Child
23. Marsh M. Gluten major histocompability complex and small intestine. Gastroenterology
24. Hansson T, Dahlbom I, Rogberg S, et al. Recombinant human tissue transglutaminase
for diagnosis and follow-up of childhood coeliac disease. Pediatr Res
25. Dahlbom I, Olsson M, Forooz N, et al. Immunoglobulin G (IgG) anti-tissue transglutaminase
antibodies used as markers for IgA-deficient celiac disease patients. Clin Diagn Lab Immunol
26. Grubin C, Daniels T, Toivola B, et al. A novel radioligand binding assay to determine diagnostic accuracy of isoform-specific glutamic acid decarboxylase antibodies in childhood IDDM. Diabetologia
27. Falorni A, Ortqvist E, Persson B, et al. Radioimmunoassays for glutamic acid decarboxylase (GAD65) and GAD65 autoantibodies using 35S or 3H recombinant human ligands. J Immunol Methods
28. Agardh D, Nilsson A, Tuomi T, et al. Prediction of silent celiac disease at diagnosis of childhood Type 1 diabetes by tissue transglutaminase
autoantibodies and HLA. Pediatric Diabetes
29. Cataldo F, Lio D, Marino V, et al. IgG(1) antiendomysium and IgG antitissue transglutaminase (anti-tTG) antibodies in coeliac patients with selective IgA deficiency. Working Groups on Celiac Disease of SIGEP and Club del Tenue. Gut
30. 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
31. Graille M, Stura EA, Corper AL, et al. Crystal structure of a Staphylococcus aureus protein A domain complexed with the Fab fragment of a human IgM antibody: structural basis for recognition of B-cell receptors and superantigen activity. Proc Natl Acad Sci U S A
32. Agardh D, Carlsson A, Lynch K, et al. Using radioligand binding assays to measure tissue transglutaminase
antibodies in young children. Acta Paediatr
33. Kaukinen K, Sulkanen S, Maki M, et al. IgA-class transglutaminase antibodies in evaluating the efficacy of gluten-free diet in coeliac disease. Eur J Gastroenterol Hepatol