Celiac disease (CD) is defined as a small bowel enteropathy triggered by the ingestion of wheat gliadin and related prolamines in genetically susceptible individuals (1). Genetic susceptibility to CD is determined, in particular, by the presence of some alleles of the class II major histocompatibility complex such as DQA1*0501-DQB1*02 (DQ2) and to a lesser extent DQA1*0301-DQB1*0302 (DQ8) (2). According to the criteria of the European Society for Paediatric Gastroenterology, Hepatology, and Nutrition, the demonstration of enteropathy on a gluten-containing diet remains the main criterion for the diagnosis (3), and the presence of serum antiendomysium (EMA) and/or anti-tissue-transglutaminase (anti-TG2) antibodies, the immunohistochemical analysis, in particular an increased infiltration of γδ intraepithelial lymphocytes (γδ IEL) (4) and human lymphocyte antigen compatibility, play only an ancillary role.
It is now clear that CD encompasses a large spectrum of clinical manifestations that range from serious symptomatic forms to completely asymptomatic forms. At the same time, from a histological point of view, it goes from the typical severe lesions of duodenal mucosa (villous atrophy, hyperplasia of the crypts, increased infiltration of the intraepithelial lymphocytes) to forms characterized by minor degrees of enteropathy (5). In clinical practice, it is increasingly more common to find patients who have serum positivity for CD-related antibodies but whose duodenal mucosa shows a normal histological appearance (6). For these individuals, the diagnosis is difficult because even though several markers of gluten sensitivity have been proposed, such as the increased number of γδ IEL (7) or the count of intraepithelial lymphocytes at the villous tip (8), none of them is so specific as to identify with certainty those patients in whom clear intestinal mucosal damage will develop and for whom a gluten-free diet is necessary.
The presence of CD-specific circulating antibodies, produced against proteins of extracellular matrix, has been demonstrated for many years (9). Dieterich et al (10) have demonstrated that the autoantibodies are directed against tissue transglutaminase (TG2), which is now recognized as the main autoantigen of CD. Recent evidence suggests that anti-TG2 autoantibodies concentrate in the mucosa of small intestine before appearing in the peripheral blood (11). More recently, evidence in patients with CD has shown specific intestinal deposits of anti-TG2 IgA localized below the basement membrane, along the villous and the crypt and around mucosal vessels (12). In patients with no villous atrophy, these deposits would have the best sensitivity and specificity to predict the evolution to overt CD, if compared with all the other tests (13). If the intestinal mucosa is normal, their presence would be important to demonstrate a condition of gluten sensitivity.
The aim of this work was to look for the presence of anti-TG2 IgA deposited in the intestinal tract mucosa of children with normal duodenal villous architecture, to find out how they compare with other markers of gluten sensitivity.
PATIENTS AND METHODS
The study involved 57 children (27 boys and 30 girls, median age 7 years 1 month, range 9 months to 17 years 11 months) who underwent small intestinal biopsy for suspected CD at the Department of Pediatrics in Naples University Hospital Federico II from January 2000 to December 2005. The biopsy specimens showed normal villous architecture but also markers of gluten sensitivity. One group (group 1, 39 patients) showed positive serum EMA and/or high levels of anti-TG2 antibodies; also, in 29 patients a greater density of γδ IEL was noted. A second group (group 2, 18 patients) included seronegative patients with only a greater density of γδ IEL. Furthermore, studied 34 patients (21 boys and 13 girls, median age of 4 years 6 months, range 8 months to 17 years 10 months) who underwent small intestinal biopsy in the same time period and had normal villous architecture with normal values of anti-TG2 antibodies, absence of serum EMA, and no IEL infiltration with normal values of γδ IEL. Their final diagnoses were iron deficiency anemia, failure to thrive, esophagitis, recurrent abdominal pain, and inflammatory bowel disease.
EMA and TG2 Antibodies
Serum IgA EMA were detected by indirect immunofluorescence on 7-μm-thick frozen sections of human umbilical cord as the source of antigen. Sera were tested diluted 1:5 and 1:50 and incubated for 30 minutes at room temperature. After serum was removed by washing with phosphate-buffered saline (PBS) solution, fluorescein isothiocyanate–labeled rabbit antihuman IgA diluted 1:80 was added for 30 minutes. Samples were considered positive if a thin fluorescent network appeared around the smooth muscle fibers. Positive results were further quantified by titration.
Serum levels of IgA anti-TG2 were determined by enzyme-linked immunosorbent assay by use of a kit based on human recombinant antigen (Eurospital Kit Eu-TG2 IgA-Trieste). Samples were diluted 1:26 before use and allowed to react in the wells for 60 minutes. Nonspecific antibodies were removed by washing. Horseradish peroxidase–labeled goat antihuman IgA was added and allowed to bind human antibodies in the well, the excess conjugate was washed away, and a chromogenic substrate was added. After an incubation period of 20 minutes, the optical density value was measured (nm 450) by SpectraCount (Packard Bioscience Company). Values were considered positive if they equaled or exceeded 7 UA/mL.
Genotyping in 70 of 91 patients was performed for HLA class DRB1 and DQB1 molecules. In 9 cases, DNA was extracted from biopsy specimens by use PrepMan Ultra Sample Preparation Reagent (Aplied Biosystems). A Dynal Allset+ SSP DR low-resolution kit, a Dynal Allset+ SSP DQ low-resolution kit, a Dynal Allset+ SSP DQB103, and a Dynal Allset+ SSP DQA1 were used for typing. Results were obtained after 2% agarose gel electrophoresis.
Duodenal Biopsy and Immunohistochemical Analysis
At least 3 small intestinal biopsy specimens were obtained. Two fragments were fixed in 10% formalin, embedded in paraffin wax, sectioned at a 5-μm thickness, and stained with hematoxylin and eosin. The remaining fragment was immediately embedded in optimal cutting temperature compound (OCT-BioOptica) and stored in liquid nitrogen. For the immunohistochemical study, cryostat sections were cut from the biopsy specimens at 4 μm and fixed in acetone for 10 minutes. After a 20-minute preincubation with normal rabbit serum (1:100, Dako), sections were covered for 1 hour with anti-CD3 (1:200, Dako) and anti-TCRγδ (1:80, Thema) monoclonal antibodies, followed by rabbit antimouse immunoglobulins for 30 minutes. Monoclonal antibodies were diluted in Tris pH 7.4, and all incubations were performed at room temperature in a humid chamber. As a negative control, primary antibody was replaced with mouse IgG2a/IgG1 (1:100, Dako). After washing with Tris pH 7.4, the sections were layered with monoclonal mouse peroxidase-antiperoxidase (1:100, Dako) for 30 minutes, and 2-amino-9-ethyl-carbazole (AEC) (Sigma) was used as peroxidase substrate. Finally, sections were counterstained with Meyer hematoxylin and mounted with Aquamount (BDH Prolabo Chemicals). The density of cells expressing CD3 and TCRγδ+ in the intraepithelial compartment was determined by counting the number of stained cells per milllimeter epithelium.
Double Immunofluorescence and Confocal Analysis
All 91 patients were investigated for TG2-related extracellular IgA deposits. Fixed in acetone, 5-μm frozen sections of each patient were examined by double immunofluorescence. After a 15-min preincubation with rabbit normal serum (1:100, Dako), the sections were covered with a monoclonal mouse antibody against guinea pig TG2 (CUB 7402) (1:200, NeoMarkers) for 1 hour at room temperature in a humid chamber. The sections were washed in PBS and incubated with a mixture of fluorescein isothiocyanate–labeled rabbit antibody against human IgA (1:100, Dako) to detect (in green) IgA deposits, and R-phycoerythrin–labeled rabbit antimouse antibody (1:40, Dako), to detect (in red) tissue transglutaminase2, for 30 minutes in the dark. Finally, the sections were washed several times in PBS and mounted by glycerol/PBS (1:10). The preparations were analysed with an Axioscope2 (Zeiss) microscope linked to an analysis image system (Siemens). The colocalization of IgA mucosal deposits and TG2 resulted in yellow at fluorescence microscope.
Where possible, specimens were also analysed by a confocal microscope (LSM510, Zeiss). With the support of LSM510 software (Zeiss) we further analyzed the colocalization images. A scatter plot of the images that shows a dot for each pixel in the corresponding images was produced. Position in the graph of each dot depends on the intensity of the fluorescence signal. Green dots will be scattered along the Ch1 channel, red dots will be scattered along the Ch2 channel, colocalizing dots will be scattered in the central region of the plot approximately along the diagonal of the plot. The region of interest was drawn in the central area of the scatter plot to highlight the colocalizing dots (white dots in the images).
The method of detecting IgA anti-TG2 deposits was validated by testing it on 12 untreated patients with CD and on 8 patients with CD in remission; 12 of 12 untreated patients with CD showed positive results (Fig. 1), and 8 of 8 patients in remission showed negative results for deposits of anti-TG2 IgA. Untreated patients presented a clear positivity of the double immunofluorescence with a pattern of intestinal deposits of anti-TG2 IgA characterized by a thick yellow band below the basement membrane, along the villous and the crypt, and around mucosal vessels. Confocal analysis confirmed colocalization of TG2 and IgA on all biopsy specimens from patients with untreated CD. The intensity of IgA deposition was graded as follows: way: +, low intensity of the signal; +++, strong intensity of the signal; patchy, difference in the intensity of the signal in different areas of the mucosa.
Statistical differences between study groups were evaluated by the Pearson χ2 test. Values of P < 0.05 were considered significant.
Deposits in Patients With Normal Mucosa but With Positive Serum EMA and/or High Levels of Anti-TG2 Antibodies
In 39 patients, the serum showed positivity for EMA, and/or high levels of anti-TG2 antibodies were investigated. In 33 of 39 patients (85%), subepithelial anti-TG2 IgA intestinal deposits were seen (Fig. 2). In contrast to patients with villous atrophy, in some cases we observed a patchy distribution and less thick subepithelial band. To confirm our results, some specimens that presented a patchy IgA deposit pattern were analyzed by confocal microscope to further establish the colocalization of IgA deposits with TG2. Confocal analysis confirmed all our previously described observations.
All of the patients were HLA DQ2/DQ8 positive, and 22 of 39 belonged to at-risk groups: 9 were first-degree relatives of patients with CD, 13 had type 1 diabetes mellitus, 11 had clinical features suggestive of CD (iron-deficiency anemia (n = 2), abdominal pain (n = 4), weight loss (n = 3), and short stature (n = 2), 1 of 39 had an autoimmune lymphoproliferative syndrome, and 5 of 39 were asymptomatic and had discovered their positivity during routine testing. Of interest, 4 of the 6 patients negative for TG2 deposits and who were positive initially for serum anti-TG2 antibodies, at the 1 year follow-up were negative for serum anti-TG2 antibodies as well, despite the gluten-containing diet.
Deposits in Patients With Only Greater Density of γδ IEL
In 18 patients, the serum showed an absence of EMA and normal levels of anti-TG2 antibodies, but density of γδ IEL higher than cutoff (3.4 cells/mm epithelium) were investigated. In this group, anti-TG2 IgA subepithelial deposits were noted in 12 of 18 patients (66%) (Fig. 2). The pattern is similar to that described for patients with serum positivity of CD-related autoantibodies and normal mucosa (Fig. 3). Even in this case, we performed confocal analysis on some specimens with patchy distribution, confirming our results.
In this subgroup of patients, 9 of 12 with deposits were investigated for HLA, and we found that 9 of 9 were positive for DQ2 and/or DQ8. The final diagnoses of all the patients are shown in Table 1. During the past months of follow-up in 1 of these 9 patients (with type 1 diabetes mellitus), anti-TG2 antibodies have appeared in the serum.
Deposits in Patients With Normal Mucosa and Negative for Markers of Gluten Sensitivity
We studied 34 patients with normal intestinal mucosa and no other markers of gluten sensitivity. In most control patients, no colocalization of IgA and TG2 was noted (Fig. 4). Only 3 of 34 patients (8.8%) showed faint intestinal deposits (Fig. 2); 2 of 3 were investigated for HLA and showed positive results for DQ2 or DQ8. The diagnosis in all cases was gastroesophageal reflux disease.
In this study we applied to small intestinal biopsy specimens obtained from children the method introduced by Korponay-Szabò et al (12) to detect intestinal deposits of anti-TG2 IgA. The method is based on the colocalization in immunofluorescence of antibodies to TG2 and antibodies to IgA. In all untreated patients with CD with severe villous atrophy, we confirmed the intestinal deposition of anti-TG2 IgA. The deposits were thick, with a clear localization, confirmed by confocal microscopy, of anti-TG2 IgA below the basement membrane, along the villous and the crypt, and around mucosal vessels. The presence of intestinal deposits does not necessarily mean that anti-TG2 IgA is locally produced, but evidence from studies based on organ culture (14) and on antibody libraries from the gut (11) are conclusive in this regard.
We subsequently investigated the small intestine of children with serum CD-associated autoantibodies but normal mucosa. These are children whose presence is increasingly recognized; in fact, in our institution during the past 3 years, they have represented approximately 10% of the small intestinal biopsies performed (Tosco et al, manuscript in preparation). These patients are also defined as patients with potential CD, inasmuch as the progression to a more severe histological picture has been reported in most cases (15). As expected, our results showed that 85% of such patients (33 of 39) turned out to be positive for intestinal deposition of anti-TG2 IgA. In contrast to patients with CD with villous atrophy, in this subgroup of children the positivity of intestinal deposits is less clear, with tracts with evident deposits and tracts in which these deposits are absent. This different distribution could be due to a lower titer of anti-TG2 antibodies; serum anti-TG2 IgA antibodies titers in these potential patients with CD are in fact lower than those in patients with villous atrophy. An alternative explanation could reside in a lower affinity of these antibodies for the intestinal transglutaminase.
There is little doubt that these patients represent a challenge to diagnosis. All of them had HLA alleles consistent with the diagnosis of CD. Most of them belonged to at-risk groups, such as first-degree relatives of patients with CD and patients with type 1 diabetes mellitus, and presented no symptoms. In fact, they also pose a problem regarding the prescription of a gluten-free diet. In this context it would be important to identify markers predictive of the evolution toward intestinal damage, and it will be interesting to determine with long-term follow-up studies whether patients who present intestinal deposits are the same as those who will experience intestinal damage. However, preliminary data from follow-up studies we are conducting on patients with potential CD (16) show that some of them have fluctuating antibody titers of serum anti-TG2 antibodies, sometimes becoming even negative. Of interest, the disappearance of serum IgA anti-TG2 has recently been reported in children carrying a genetic risk for CD (17). Also, in these latter cases it would be interesting to understand what happens in the intestinal tract at the same moment.
We also investigated children who were negative for serum CD-associated autoantibodies but with an increased number of γδ IEL. We found even in these patients a 66% (12/18) positivity for deposits. The determination of γδ IEL is considered highly sensitive and specific for CD (4–18); increased density has been recently been shown to be indicative of early developing CD (13). Of interest, most patients with high numbers of γδ IEL and HLA DQ2 or DQ8 have been found to have intestinal deposits of IgA anti-TG2 (19). By contrast, apparently false-positive increased γδ IEL have been reported, in most cases being associated with autoimmunity (20).
In these patients, the relation of these deposits to CD is difficult to establish. Recent studies have shown the high predictive value of anti-TG2 IgA deposits of forthcoming villous atrophy (13). However, it is still possible that intestinal anti-TG2 is an expression of an autoimmunity state. Some serologically negative individuals with anti-TG2 intestinal deposits are in fact patients with autoimmune diseases, such as type 1 diabetes mellitus. In this context we have recently reported the high prevalence of anti-TG2 intestinal deposits in patients with type 1 diabetes mellitus. The observation that in these cases the use of V chains is not restricted to the VH5 family (21) because it occurs in CD (22) seems to suggest that these anti-TG2 antibodies are not related to dietary gluten.
In conclusion, our studies confirmed the presence of anti-TG2 deposits in the small intestine of untreated patients with CD, irrespective of the severity of mucosal damage. Furthermore, we have also shown the presence of such deposits in serologically negative individuals. Only long-term prospective studies will tell to what extent anti-TG2 intestinal deposits are a marker of gluten sensitivity and a predictor of forthcoming villous atrophy.
1. Green PH, Jabri B. Coeliac disease. Lancet 2003; 362:1418–1419.
2. Margaritte-Jeannin P, Babron MC, Bourgey M, et al
. HLA-DQ relative risks for coeliac disease in European populations: a study of the European Genetics Cluster on Coeliac Disease. Tissue Antigens 2004; 63:562–567.
3. Walker-Smith JA, Guandalini S, Schmitz J, et al
. Revised criteria for the diagnosis of celiac disease. Arch Dis Child 1990; 65:909–911.
4. Iltanen S, Holm K, Ashorn M, et al
. Changing jejunal gamma/delta+ T cells receptor (TCR)-bearing intraepithelial lymphocyte density in celiac disease. Clin Exp Immunol 1999; 117:51–55.
5. Marsh MN. Gluten, major histocompatibility complex, and small intestine: a molecular and immunobiologic approach to the spectrum of gluten sensitivity. Gastroenterology 1992; 102:330–354.
6. Paparo F, Petrone E, Tosco A, et al
. Clinical, HLA and small bowel immunohistochemical features of children with positive serum antiendomysium antibodies and architecturally normal small intestinal mucosa. Am J Gastroenterol 2005; 100:2294–2298.
7. Maki M, Holm K, Collin P, et al
. Increase in gamma/delta T cell receptor bearing lymphocytes in normal small bowel mucosa in latent celiac disease. Gut 1991; 32:1412–1414.
8. Jarvinen TT, Collin P, Rasmussen M, et al
. Villous tip intraepithelial lymphocytes as markers of early-stage celiac disease. Scand J Gastroenterol 2004; 39:428–433.
9. Maki M. The humoral immune system in celiac disease. Bailliere Clin Gastroenterol 1995; 9:231–249.
10. Dieterich W, Ehnis T, Bauer M, et al
. Identification of tissue transglutaminase as the autoantigen of celiac disease. Nature Med 1997; 3:797–801.
11. Marzari R, Sblattero D, Florian F, et al
. Molecular dissection of the tissue transglutaminase autoantibody response in celiac disease. J Immunol 2001; 166:4170–4176.
12. Korponay-Szabò IR, Halttunen T, Szalai Z, et al
. In vivo targeting of intestinal and extraintestinal transglutaminase 2 by celiac autoantibodies. Gut 2004; 53:641–648.
13. Salmi TT, Collin P, Jarvinen O, et al
. Immunoglobulin A autoantibodies against transglutaminase 2 in the small intestinal mucosa predict forthcoming celiac disease. Aliment Pharmacol Ther 2006; 24:541–552.
14. Picarelli A, Maiuri L, Frate A, et al
. Production of antiendomysial antibodies after in-vitro gliadin challenge of small intestine biopsy samples from patients with coeliac disease. Lancet 1996; 348:1065–1067.
15. Collin P, Helin H, Maki M, et al
. Follow-up of patients positive in reticulin and gliadin antibody tests with normal small-bowel biopsy findings. Scand J Gastroenterol 1993; 28:595–598.
16. Salvati VM, Tosco A, D'Adamo G, et al
. A 2-year follow-up study in children with positive coeliac disease-specific serum antibodies and architecturally normal small intestinal mucosa. Dig Liv Dis 2006; 38:A111.
17. Simell S, Hoppu S, Hekkala A, et al
. Fate of five celiac disease-associated antibodies during normal diet in genetically at-risk children observed from birth in a natural history study. Am J Gastroenterol 2007; 102:2026–2035.
18. Camarero C, Eiras P, Asensio A, et al
. Intraepithelial lymphocytes and coeliac disease: permanent changes in CD3−/CD7+ and T cell receptor gammadelta subsets studied by flow cytometry. Acta Paediatr 2000; 89:285–290.
19. Kaukinen K, Peraaho M, Collin P, et al
. Small-bowel mucosal transglutaminase 2-specific IgA deposits in coeliac disease without villous atrophy: a prospective and randomized clinical study. Scand J Gastroenterol 2005; 40:564–572.
20. Iltanen S, Holm K, Partanen J, et al
. Increased density of jejunal gamma/delta+ T cells in patients having normal mucosa-marker of operative autoimmune mechanisms? Autoimmunity 1999; 29:179–187.
21. Florian F, Vecchiet M, Gazzin S, et al
. Production of antitransglutaminase autoantibodies in the jejunum of patients with type1 diabetes. J Pediatr Gastroenterol Nutr 2005; 40:617.
22. Sblattero D, Florian F, Azzoni E, et al
. One-step cloning of anti tissue transglutaminase scFv from subjects with celiac disease. J Autoimmunity 2004; 22:65–72.