Tissue Transglutaminase: Does the Key Fit the Celiac Lock?

Autoantibodies to tissue transglutaminase as predictors of celiac disease. Dieterich W, Laag E, Schopper H, Volta U, Ferguson A, Gillett H, Riecken EO, Schuppan D. Gastroenterology 1998;115:1317–21. The authors describe their experience with the latest addition to the serological diagnostic armamentarium for celiac disease. One hundred six blood samples obtained from patients with untreated celiac disease of varying clinical and histologic severity, 43 from patients with celiac disease who were consuming a gluten-free diet (GFD), and 114 from control subjects were tested. Immunoglobulin (Ig)A endomysial antibody (EmA) titers were measured by two reference laboratories using monkey esophagus as the substrate. The samples were diluted until immunofluorescence was undetectable, according to a semiquantitative scoring system of 0 to 5. The method for tissue transglutaminase (tTG) enzyme-linked immunosorbent assay (ELISA) is described. The authors discuss improvements to their original assay. They avoided using bovine serum albumin (BSA) in the blocking agent, because patients had previously been noted to have antibodies to BSA, which may have caused false-positive results. Calcium, added to the buffer, improved the sensitivity, because the anti-tTG antibodies in patients with celiac disease seemed to react more strongly with the calcium-activated form of tTG, because it is a calcium-dependent enzyme. Their cutoff value of 15 (calculated from optical density) excluded 95% of patients without celiac disease. Four with celiac disease had normal tTG values but weakly positive EmAs. Ten blood samples were negative for EmAs but had elevated tTG: These “false positives” declined to six when it was apparent that one patient had celiac disease with partial villous changes and three had celiac disease and were consuming a GFD. When patients with biopsy-proven celiac diseases with positive EmAs were taken into account, the sensitivity was 98.1% and specificity 94.7%. Six patients with elevated intraepithelial lymphocyte (IEL) counts and no morphologic changes had elevated tTG titers but low EmA titers. A positive correlation was found when EmA and tTG titers were compared, with tTG titers being significantly higher in patients with untreated celiac disease than in those consuming a GFD and in control subjects. Interestingly, the control group of 114 patients included 6 patients with elevated tTG and negative EmA, none of whom had biopsies. The authors discuss the utility of tTG titers compared with determination of IgA EmA. There was good correlation between the assays. There is good evidence that titers were significantly higher in untreated versus treated patients. There are discrepancies in results between the two assays, with EmA remaining the more sensitive and specific. Positive EmA and negative tTG results are discussed and the nonhuman nature of the substrates highlighted as a potential cause of discrepancy. Cross-linking of tTG with gliadin is also discussed as a potential source of false-positive result, in that sera may react with the neoepitopes rather than with the tTG itself. Of the 10 EmA-negative, tTG-positive patients, four were considered to have celiac disease. Biopsies were not performed in the other six patients, leaving some doubt about whether these were true false positives or merely extremely sensitive tests and predictive of future development of histologic change. It is suggested that there may be elevation of IgA anti-tTG despite low or absent IgA levels, and that IgG anti-tTG may be positive in such instances. However, the authors stress that IgG anti-tTG antibodies may be elevated in other diseases. Therefore, the utility for IgG tTG has yet to be established. Overall, the correlation between IgA EmA and tTG titers is good, but discrepancies occur, and more refinement may yet come with the advent of human tTG studies. Correlation with compliance with diet may prove its use as a means of following up and screening patients. Since the publication of the original article of Dieterich et al. (Nat Med 1997;3:797–801) describing tissue transglutaminase (tTG) as the putative autoantigen to EmA, there has been an explosion of interest in tTG and continued debate about the mechanisms contributing to the pathophysiology of the enteropathy that underpins celiac disease. Tissue transglutaminase is a ubiquitous enzyme present in many tissues and is calcium dependent, catalyzing the cross-linking of proteins. It is present in cell cytoplasm but is secreted in low amounts by mononuclear cells and endothelial cells and by fibroblasts during wound healing. It cross-links proteins thereby stabilizing the extracellular matrix in damaged tissue. Gliadin contains many glutamine residues and serves as a substrate for tTG. The tTG deamidates glutamine residues and introduces a covalent bond between lysine and glutamine. These neoepitopes of gliadin–gliadin, deamidated gliadin, and tTG–gliadin complexes (Nat Med 1998;4:713–7, J Immunol 1998;161:1585–8) serve as antigens and are recognized by HLA DQ2 on gliadin-specific T cells (Gut 1995;37:766–76) which initiates production of the CD4+ responses. This induces the inflammatory cascade and a reduction in active TGF-β occurs. This contributes to inflammation in the matrix, lamina propria, and epithelium and impairs the process of enterocyte differentiation (Nat Med 1998;4:666–7). EmA antibody is directed against tTG. Essentially, EmA and tTG antibodies can be viewed as two methods of looking at the same thing. Endomysial antibody is the current standard serology test in those more than 2 years of age with sensitivity and specificity usually over 90%. EmA negativity tends often to indicate that a biopsy is not required. Some have suggested (Gut 1998;42:594) that if we perform biopsies only in those positive for EmA, (ignoring positive anti-gliadin antibodies), then those with celiac disease who are negative for EmA may be missed. Reports of studies comparing the utility of tTG are starting to come thick and fast. By validating the antibody tests on patients (in control groups) who have had biopsy and in whom serum is available, we can gain some appreciation of the correlation between negativity and a normal mucosa. We know from recent studies that tTG levels are high in active celiac disease and that the levels decrease toward the normal range when patients maintain a strict GFD (Gastroenterology 1999;116:G3835–6). Patients with biopsy-proven changes usually have positive EmA, and there is mostly good correlation between the two antibody tests. What should we do with patients with positive anti-tTG antibodies but negative IgA EmA? Do they have any mucosal anomalies? Studies are limited in number. The Berlin group reported 10 patients who were EmA negative but tTG positive. Four patients had celiac disease but had suboptimal GFD compliance. The remaining six did not undergo biopsy. Troncone recently reported (J Pediatr 1999;134:166–71) that during gluten challenge, IgA to EmA was a more sensitive predictor of mucosal relapse than tTG, which remained in the normal range in 50%. It appears that tTG's positivity in celiac disease relates to the presence of the DQ2 haplotype. This does not fully explain the situation, because up to 30% of the general population also have DQ2, and it is suggested that other factors, genetic or otherwise, must be present to account for this condition. The debate continues about whether pathogenesis of celiac disease is due to a humoral or cellular mechanism. Currently the strongest argument is that the immunopathology is caused by CD4+ lymphocytes, activated by gluten, driving a T-cell–mediated response as would be seen in graft-versus-host disease. Specific DQ2-restricted lymphocytes have been cloned, which are shown to release proinflammatory agents, strengthening the argument for T cells as a driving factor. It is a matter of debate whether the antibodies play a role in the pathogenesis. Data relating disease-specific IgA antibodies to the development of the mucosal lesion continue to be presented (Gastroenterology 1999;116:566–72), yet celiac disease occurs in patients with IgA deficiency. Indeed, those with IgA deficiency have a 10-fold increase in incidence compared with IgA-sufficient patients. This suggests that it is not a humoral-driven condition, but is cell-mediated and in such cases, patients may have an IgG antibody response to tTG, anti-gliadin antibodies, or EmA. With the advent of small intestinal biopsy came the understanding that many disease processes could have similar mucosal findings. For definitive diagnosis of celiac disease, gastroenterologists relied on removal of gluten, subsequent demonstration of histologic improvement, and then relapse again after gluten rechallenge. The revised ESPGAN criteria (Arch Dis Child 1990:65;909–11) attempts to make follow-up simpler but still leaves a group of patients who may have positive serology but none of the current criteria (elevated IELs or other more subtle abnormalities). Are they within the celiac spectrum? The discovery of celiac-related antibodies, particularly the tissue-related ones (anti-reticulin, anti-jejunal, and anti-endomysial) has led to widespread usage of serology. With the advent of these highly specific and sensitive screening tests, we continue to recognize that more subtle histologic findings can be made (Scand J Gastroenterol 1998;33:944–9). Many now believe that a high IEL count is sufficient in the face of serologic evidence to diagnose celiac disease, particularly when symptoms are present. It is the asymptomatic group that causes the dilemma. Do we treat or not? Some suggest gluten loading and rebiopsy. Some suggest a GFD and rebiopsy. We suspect most gastroenterologists would adopt a wait-and-see approach, but there is strong evidence that many patients who are seropositive but asymptomatic eventually experience development of celiac disease (Scand J Gastroenterol 1993;28:595–8). It follows therefore that the more we screen, the more we will find cases that do not fit the standard criteria for diagnosis of celiac disease without symptoms, and the more debate there will be about whether to advise a GFD. There is a powerful argument for a GFD when the risks of potential complications of untreated celiac disease are appreciated. The natural history of minimal-change mucosal damage and its correlation with serology urgently needs to be clarified. Is it ethical not to treat? With the discovery and now subsequent validation by many investigators that tTG is indeed the antigen to which EmA is directed, the race is now on to refine this exciting new prospect into a usable and reliable means of looking at large groups of at-risk patients. The exact role of tTG in the pathogenesis of celiac disease must be clarified. It is clear from the reports of Dieterich et al. (and Sulkanen in the same issue) that EmA still has the edge on tTG for sensitivity and specificity, but an interesting group will be those patients positive for tTG but negative for EmA. The natural history of this group will be fascinating. Kits are becoming available for tTG, and it will be interesting to assess their reliability. Standardization is still a problem, as it has been with anti-gliadin antibodies and EmA. With the lack of standardization in testing there is still tremendous variability in results from different groups and the physician must know the sensitivity and specificity of the test the laboratory is using. Will tTG replace EmA? Not just yet, even with human tTG giving reportedly better results. Screening strategies must be evolved. It may be practical to screen with tTG as a first-line attempt at diagnosis and then to determine EmA in those positive for tTG. Local policies will have to be established as the test becomes available, but the take-home message is, as always, that any patient with suspected celiac disease should undergo biopsy irrespective of the serology results.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.