Background: IgA anti-tissue transglutaminase-2 antibody (anti-TG2Ab) deposits in intestinal and extraintestinal organs have been used to link the respective pathological changes in these organs with celiac disease (CeD). Aims: To know if parts of intestine other than the duodenum, such as esophagus, stomach, and colon, have any pathology related to potential CeD or have mucosal IgA anti-TG2 Ab deposits. Settings and Design: A prospective case–control study conducted from April 2018 to December 2019. Materials and Methods: Nine patients with potential CeD and 27 age- and gender-matched patients with irritable bowel syndrome were recruited as cases and controls, respectively. Mucosal biopsies were collected from esophagus, stomach, duodenum, and rectosigmoid regions, histological changes were evaluated, and IgA anti-TG2 Ab deposits were analyzed in these regions by two-color immunohistochemical staining. Statistics: Data were analyzed using statistical software Stata 14.0. Results: No distinct difference in mucosal lymphocytosis were identified between biopsies of patients with potential CeD and controls at the following sites: esophagus (11.1% vs 0%, P = 0.079), stomach (14.3% vs 7.7%, P = 0.590), and rectum (20% vs 0%, P = 0.067). Co-localized IgA anti-TG2Ab deposits were observed more in potential CeD than in controls at esophagus 22.2% (2/9) vs 0%, P = 0.012; stomach 66.7% (6/9) vs 11.5% (3/26), P < 0.001; and duodenum 66.7% (6/9) vs 0%, P < 0.001 but not at rectum 0% (0/4) vs 0% (0/25). Conclusion: Although histological changes are not distinct, a subset of subjects with potential CeD has pan-intestinal involvement other than in the duodenum.
Keywords: Celiac disease, histology, IgA antitissue transglutaminase, pathogenesis, potential celiac disease
How to cite this article:Celiac disease (CeD) is one of the most common causes of enteropathy, and global seroprevalence and global prevalence of biopsy-confirmed CeD are 1.4% and 0.7%, respectively.[1],[2] Potential CeD is a subgroup of patients characterized by a positive celiac-specific serology but they lack features of advanced grade of villous abnormalities. With an increase in awareness and better screening methods, the prevalence of potential CeD has risen and it comprises one fifth of patients with CeD.[3–7] The grade of villous abnormalities from mild to advanced is a continuum of the disease process and thus patient may be detected to have early enteropathy, especially among those who are screened for CeD. Almost a third of patients with potential CeD progress to develop advanced grade of villous abnormalities on follow-up, another one third may remain as potential CeD over a follow-up of 9 years, and in rest, one third become seronegative.[4]
With growing knowledge, it is now apparent that the mucosal pathology in CeD is not limited only to the small intestinal mucosa but it also affects many other organs, including liver, skin, brain, and also non-small intestinal part of gastrointestinal tract, including esophagus, stomach, and large intestine.[8–10]
A prospective observational study was conducted to explore the involvement of esophagus, stomach, and rectosigmoid mucosa in addition to that of small intestine in patients with CeD via showing lymphocytosis and deposition of IgA anti-tissue transglutaminase-2 antibody (anti-TG2Ab) deposits at these sites. Originally, 54 patients suspected to have CeD and had positive anti-TG2Ab for CeD were recruited for this study. Of them, 42 patients with villous abnormalities of modified Marsh (mMarsh) grade 2 or more were included as cases and 12 had villous abnormalities of mMarsh grade 0–1, hence, were labeled as potential CeD and excluded.[11] The extra-small intestinal involvement of gastrointestinal tract, including esophagus, stomach, and the colon in patients with biopsy-confirmed CeD were published in 2021. As the involvement of extra-small intestinal organ involvement in a subset of patients with CeD was shown earlier, this study wanted to know if extra-small intestinal part of gastrointestinal tract, such as esophagus, stomach, and colon are affected even in those without any villous atrophy, i.e., potential CeD.
Materials and MethodsThis study was planned as a prospective case–control study, and during 21 months of total duration of this study, a total 12 consenting subjects who fulfilled the standard accepted definition for potential CeD were acquired. This study did not calculate the sample size of the target group as that was beyond the scope in this study. Out of the 12 patients with potential CeD, three patients on repeat serological investigation had normal IgA anti-TG2Ab titre; hence, they were excluded and finally nine patients with potential CeD were included in this study as cases.
As this study was planned to explore the extra-small bowel mucosal pathologies in these potential CeD patients, this study collected multisite biopsies from esophagus, stomach, and rectum in addition to duodenal biopsies in each of these nine subjects and also collected multisite biopsies from 27 patients with irritable bowel syndrome (IBS) that were included as controls for this study. Like this study subjects with potential CeD, all controls underwent IgA anti-tTG Ab serological screening using ELISA kit from Innova diagnostics, USA (>4 U/ml- pathological), basic biochemical and hematological evaluation, and histological analyses. The control subjects were also not the first-degree relatives of subjects of potential CeD included in this cohort. In addition to histological changes, this study also explored if there is any IgA anti-TG2Ab deposits in these mucosal sites, as this antibody deposits have already been established as one of the early laboratory markers in CeD.[11]
Esophagogastroduodenoscopy and sigmoidoscopy were done for both cases and controls, and multiple mucosal biopsies were taken and the endoscopic features were noted. A minimum of 4–6 mucosal biopsy fragments were obtained from the second/third parts of the duodenum, including at least one fragment from the duodenal bulb and oriented in the endoscopy room.[12]At the time of the obtaining duodenal biopsies, 2–4 mucosal biopsies from three sites of the esophagus, including lower esophagus (5 cm above the gastroesophageal junction), mid esophagus, and upper esophagus (5 cm below the upper esophageal sphincter) were obtained. At the same session, multifocal biopsies were also obtained from the antrum, incisura, body, and both lesser and greater curvatures of the stomach as per the modified Sydney classification along with an additional biopsy from the gastric fundus.[13],[14] Biopsies from the different parts of the stomach were histologically reported separately as per the modified Sydney system.[13] Sigmoidoscopy examination was done and 2–4 mucosal biopsies were also obtained from the recto-sigmoid junction. Recto-sigmoid biopsies were reported as per the European consensus.[15] All the biopsies from different sites, as described earlier, were collected in pre-labeled containers containing 10% neutral-buffered formalin and sent to the pathology laboratory. All biopsies were processed as per standard procedure and reported by two expert GI pathologists who were blinded to the diagnosis.
Esophageal biopsies were evaluated for intraepithelial lymphocytes (IELs)/HPF (×40), intraepithelial eosinophils (Eos)/HPF, neutrophils/HPF, basal cell layer hyperplasia, papillary elongation, intercellular spongiosis, blood lakes, ulcer, and dysplasia. Histologically, reflux esophagitis was defined by the presence of basal zone hyperplasia in >15% of the epithelial thickness, papillary elongation, dilated intercellular spaces, Eos, IELs, and neutrophils, and/or presence of blood lakes and/or ulcers. All parameters were histologically sub graded as per the Esohisto project of Fiocca et al.[16] Gastric biopsies were evaluated for the extent of the mononuclear cell infiltrate, neutrophilic infiltrate (activity), mucosal atrophy, intestinal metaplasia, the extent of H. Pylori-like organisms on Warthin Starry stain, number of eosinophils/HPF, and the number of IELs/100 epithelial cells as per the modified Sydney protocol.[13] Duodenal biopsies were assessed for the number of IELs per 100 epithelial cells at villous tips, presence of crypt hyperplasia, villous flattening, mucosal inflammation, eosinophils/HPF, presence of luminal parasites, and the grade of villous abnormalities as per the mMarsh-grading system.[17] Rectosigmoid biopsies were evaluated for distortion of crypt architecture, neutrophilic cryptitis/crypt abscess or ulcers, lamina propria inflammation, mucosal eosinophilic cell infiltrate (number/HPF), IELs/100 epithelial cells, and basement membrane thickening.
Eosinophilic esophagitis was defined as the presence of eosinophils >20/HPF (×40) with or without the presence of superficial eosinophilic micro-abscess.[18–20] Lymphocytic esophagitis was defined as dense peripapillary spongiosis and significant lymphocyte infiltration in the esophageal stratified squamous epithelium (>50/HPF).[21] Chronic-active gastritis was defined as the presence of both mononuclear cell infiltrate and neutrophilic activity in the gastric biopsy. Lymphocytic gastritis was defined as infiltration of ≥25 IEL/100 epithelial cells.[13],[22] Eosinophilic gastritis was defined as infiltration of ≥20 eosinophils/HPF (40×) with evidence of epithelial eosinophil infiltrate or eosinophilic foveolar abscess, with or without foveolar destruction. Atrophic gastritis was defined as the reduction of the native gastric glands, irrespective of whether they have been replaced by fibrosis or metaplastic epithelium.[22] Eosinophilic duodenitis was defined by the presence of ≥20 eosinophils/HPF (40×). Greater than 20 IELs/100 epithelial cells and greater than 20 eosinophils/100 epithelial cells were taken as the cutoff for defining lymphocytic colitis and eosinophilic colitis, respectively.[16],[23],[24] The pathological cutoff considered in this study is as per large published cohorts and the routine practice, though, needless to say that the described pathological cutoffs vary across different published studies.[20],[25],[26]
Localization and relative quantification of the IgA anti-TG2Ab deposits were performed on the sections cut from formalin-fixed paraffin-embedded blocks, after performing dual Immunohistochemistry (IHC) technique using the peroxidase-labeled rabbit anti-human polyclonal anti-human IgG TG2 (1:400 dilution, ab421; ABCAM), developed with 3,3'-Diaminobenzidine chromogen; with subsequent staining with alkaline phosphatase-labeled monoclonal mouse anti-human IgA antibody (1:1000 dilution, MERCK), developed with Vector Blue chromogen, Vector Lab, USA. Thus, the anti-TG2Ab deposits were identified by the light brown color and IgA deposits as blue-stained areas in the sections examined. The co-localized IgA anti-TG2Ab deposits thus appeared as dark brownish-blue deposits, in comparison to the isolated anti-TG2 and anti-IgA deposits as described earlier. The co-localized deposits were analyzed in terms of area of positivity and stain intensity, as negative (0), mild (1), moderate (2), and strongly positive (3) using the previously standardized visual analog scale.[27] The final semi-quantitative H scores were calculated individually in different anatomical locations by adding up the distribution score and intensity scores as follows. The anatomical locations studied include epithelium of duodenal villous and crypt mucosa, capillary wall, subepithelial basement membrane, and muscularis mucosae; esophageal basal keratinocytes, spinocytes, and vascular endothelial cells; colonic surface and crypt epithelium, subepithelial basement membrane, the capillary wall in lamina propria, and muscularis mucosae. Cumulative scores were finally graded as follows: 0–3 = Negative (None + mild), 1 (Cumulative score: 4–6) = Positive (moderate), 2 (Cumulative score: 6–9) = Strong positive (strong) as per the previous standardized IHC analog scale.[27] The study was carried out after receiving ethical clearance (No.IECPG 165/19.04.2018) from the Institutional Ethics Committee. Written and informed consent was obtained from all study participants.
Statistical analysis
Data were analyzed using statistical software Stata 14.0. Quantitative variables were expressed as mean ± standard deviation (SD) and median (Min–Max) for normal and skewed distributions, respectively. Categorical variables were compared using the Chi-square test/the exact Fischer test. Continuous variables following a skewed distribution were compared by the Mann Whitney U/Wilcoxon rank sum test. Variables having repeated measures were compared using the Wilcoxon sign rank test. All tests were two-tailed and a P value < 0.05 was taken as statistically significant.
ResultsThis study had nine patients of potential CeD as cases, and 27 IBS patients as controls. There was no difference in mean age, gender, and baseline hemoglobin between the two groups [Table 1].
Table 1: Demographic profile and clinical characteristics of patients with potential celiac disease and controlsEsophageal mucosa in patients with potential CeD and controls [Table 2]
Table 2: Histological changes in esophageal mucosal biopsies taken from upper, middle, and lower esophageal segments, and IgA anti-TG2Ab deposits in patients with potential celiac disease compared to controlsThere was no difference in the number of IELs in mucosal biopsies obtained from upper, mid, and lower part of esophagus in patients with potential CeD and controls. Only one patient with potential CeD and none of the controls had lymphocytic esophagitis (IELs >50 at the lower and mid esophagus). Neither any patient with CeD or control had eosinophilic esophagitis nor there was any difference in mean eosinophil count in them. None of the other histological parameters between these two groups were different between patients with potential CeD and controls. [Table 2].
Two out of nine cases had mucosal IgA anti-TG2Ab deposits vis-a-vis none in controls (P = 0.012) and none of these cases patients had lymphocytic esophagitis [Table 2]. At 18 months of follow-up, one of these two patients was started on Gluten free diet (GFD) in view of the appearance of villous atrophy in the follow-up biopsies. Histological changes are shown in [Figure 1].
Figure 1: (a) Control esophageal biopsies show only occasional intraepithelial lymphocytes (×100) (b) The esophageal biopsy of a patient with potential CeD shows increased intraepithelial lymphocytes (arrows) (×100) (c) Esophageal biopsy of controls does not show IgA anti-TG2Ab deposits (×100) (d) Biopsy in patients with potential CeD shows dark bluish-brown IgA anti-TG2Ab deposits in the basal keratinocytes (arrows) and in capillary wall in papillae (Inset, arrow) (×100) (e) Control gastric biopsies do not show increased intraepithelial lymphocytes (×200) (f) Biopsies of patients with potential CeD show increased IELs (arrows) (×100) (g) IgA anti-TG2Ab deposits were not seen in gastric biopsy from controls (×100 arrow, IgA-positive plasma cells) (h) In potential CeD, deposits were identified in mucosal capillary wall (arrows, Inset) (×100) (i) The duodenal biopsy of controls did not show an increase in intraepithelial lymphocytes (×200) (j) The duodenal biopsy of patients with potential CeD shows an increase in IELs at the villous tips (arrow) (×200) (k) IgA anti-TG2Ab deposits were not identified in the control biopsy (×100, the arrow indicated IgA-positive plasma cells) (l) In duodenal biopsies of patients with potential CeD, IgA anti-TG2Ab deposits were identified in mucosal capillaries, muscularis mucosae (arrows), and epithelium (arrow, inset) (×100)Stomach in patients of potential CeD and controls [Table 3]
Table 3: Histological changes in the mucosal biopsies taken from the gastric antrum, lesser curvature, greater curvature, and fundus, and IgA anti-TG2Ab antibody deposits in patients with potential celiac disease in comparison to controlsNeither any patient with potential CeD or control had any difference in the mean number of IELs in any of the stomach biopsies sampled according to the modified Sydney protocol nor was there any significant difference in the number of patients having lymphocytic gastritis between potential CeD and controls (14.3% [1/7] vs 7.7% [2/26], P = 0.590). There was no difference in the number of mucosal eosinophilic cell infiltrates between the two groups. A higher number of subjects with potential CeD had IgA anti-TG2Ab mucosal deposits as compared with controls (66.7% [6/9] vs 11.5% [3/23], P = 0.001). Two patients with potential CeD who had esophageal IgA anti-TG2Ab deposits also had mucosal deposits of IgA anti-TG2Ab in the stomach also.
Duodenum in patients with potential CeD and controls [Table 4]
Table 4: Histological changes in the mucosal biopsies taken from duodenum and rectum and IgA anti-TG2Ab deposits in patients with potential celiac disease in comparison to controlsMucosal IELs per 100 epithelial cells (31 vs 12, P = 0.026) as well as mucosal eosinophil densities/HPF (5.5 vs 2, P = 0.009) were significantly higher in mucosal biopsies of subjects with potential CeD compared to controls. IgA anti-TG2Ab deposits were also observed in 4 of 7 subjects (57.1%) compared with none of 25 controls (P < 0.001) [Table 3].
Rectum in patients with potential CeD and controls
There was no difference in the number of IELs or mucosal eosinophil cell infiltrate in the rectal mucosa of patients with potential CeD and controls, and none of them had IgA anti-TG2Ab deposits in the rectum [Table 4].
Concordance of mucosal histological changes and IgA anti-TG2Ab deposits
There was concordance between duodenal and gastric mucosal IgA anti-TG2Ab deposits as all six patients with duodenal IgA anti-TG2Ab deposits also had gastric IgA anti-TG2Ab deposits. Furthermore, two of these six patients with potential CeD also had esophageal deposits. Three of this study patients of potential CeD did not have any deposits at any of the sites and two of the patients had deposits at all three sites.
Follow-up of patients with potential CeD included in this study
After 18 months of follow-up, three patients with potential CeD who did not have IgA anti-TG2Ab deposits at any mucosal site continued to be clinically qualified for potential CeD, whereas two out of six patients (two out of six were lost to follow-up; therefore, this study had follow-up in four patients only) in whom IgA anti-TG2 deposits were identified in baseline biopsies developed CeD and were started on GFD (one developed villous atrophy and another patient developed features of malabsorption, i.e., anemia and did not consent for biopsy).
DiscussionIn this prospective pilot case–control study, presence of deposits of IgA anti-TG2Ab in the mucosa of the duodenum, stomach, and esophagus in a subset of patients with potential CeD compared to controls was observed. This study highlights involvement of extra-intestinal organs in patients with potential CeD by demonstration of IgA anti-TG2Ab deposition in a subset of them. Importantly, patients with potential CeD who had deposition of IgA anti-TG2Ab in their biopsies progressed to develop CeD after an interval of 18 months, suggesting that IgA anti-TG2Ab mucosal deposits can be an early marker of potential CeD who could develop CeD in near future.
In the last decade, the epidemiological and cohort studies from Europe have shown that approximately 20% of all patients with CeD have potential CeD.[3],[6],[28] In the index study, 18% (9/51) of patients were classified as potential CeD (based on the data from previous study).[11] Similar prevalence of potential CeD has been shown in larger studies by Volta et al.[3](20%, 77/787) and Biagi et al.[28] (18%, 42/187) in European populations.
Approximately, 66% of patients with potential CeD had mucosal IgA anti-TG2Ab deposits at the duodenum, which is in concordance with a large follow-up study by Tosco et al.,[6] who also demonstrated that these immune complex deposits are early predictor of villous atrophy. In a meta-analysis of 6 studies (only one study included adult patients), IgA anti-TG2Ab deposits were identified in 68–100% of patients with potential CeD.[29] In a 9-year follow-up study on potential CeD, IgA anti-TG2Ab deposits were more frequent in those who progressed later to develop villous atrophy.[4] On further follow-up of the cohort by the same group at 12 years, 43% of potential CeD children progressed to have villous atrophy, and patients who had IgA anti-TG2Ab deposits at baseline had twice the chance of progression to villous flattening than those without IgA anti-TG2Ab deposits at baseline.[5]
Although this study has shown that a subset of patients with CeD have involvement of extra-small intestinal involvement of esophagus, stomach, and the colon, this study also observed the similar trend, albeit less in number with subjects with potential CeD. This study had already shown the correlation between mucosal lymphocytosis and IgA anti-TG2Ab mucosal deposits in CeD patients.[11] The IgA anti-TG2Ab deposits and mucosal intraepithelial lymphocytosis (IELosis) represent the histological characteristics of a potential CeD. Possible immune tolerance at this stage of the disease may be responsible for the maintained villous architecture, except for IELosis in a some cases.[7] This study also had some of the controls having IgA anti-TG2Ab deposits. One of the plausible explanations could be that few of the IBS controls may be either seronegative CD (5–7% in literature) or having non-celiac gluten sensitivity (0.6–10%). They have been shown to exhibit IgA anti-TG2Ab deposition in mucosal biopsies in published studies.[30] Hence, although this study investigated them to the best, such incipient overlap could not be completely ruled out.
In addition to the findings of extra-duodenal mucosal IELosis in CeD and potential CeD, one should also keep in mind the other relatively more common small bowel inflammatory disorders associated with IELosis at these sites, such as Crohn's disease, primary motility disorders of esophagus, esophageal candidiasis, and gastroesophageal reflux disease as causes for esophageal IELosis[21],[31]; or, at the stomach in Helicobacter Pylori gastritis, microscopic colitis, medication-related (non-steroidal anti-inflammatory drugs and angiotensin II receptor blockers)[32]; and in the colon in microscopic colitis or inflammatory bowel disease.[33]
There were several limitations of this study including small number of subjects and lack of follow-up in all subjects. As a comparator, this study could not include biopsies from healthy individuals because of ethical consideration and hence this study had to include patients with IBS as control. Patients with IBS are known to have histological changes that could have influenced the results.[34] The limitations of multi-color IHC must also be keept in mind, which is expensive, time consuming, and expertise dependent. Furthermore, there is a possibility of inter-individual difference in interpreting stain intensities on IHC.[35],[36] Larger studies applying two-color IHC technique to validate the observation of this study is needed.
In addition, experience on potential CeD is less and not fully standardized. As with other extra-duodenal involvement in CeD, GFD seems effective in potential CeD patients and reduction of IELosis as well as IgA anti-TG2Ab deposits with GFD in extra-duodenal sites in patients with CeD has been demonstrated in the preceding study.[11]
ConclusionsIgA anti-TG2Ab deposits can be used as a marker of disease progression in potential CeD. Additionally, as observed in patients with CeD, a subset of subjects with potential CeD have involvement of part of gastrointestinal tract other than small intestine. Further multicentre studies including large cohorts of potential CeD need to be conducted for the confirmation of the findings of this study.
Acknowledgments
We sincerely appreciate the support of all the patient participants of this study. We acknowledge the support of the Research Section for the smooth conduct of this study. We also acknowledge Department of Biotechnology for providing grants for establishing Consortium on Celiac disease and national Biorepository of Celiac disease and supporting Dr. Alka Singh for her research fellowship.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
Correspondence Address:
Govind Makharia
Department of Gastroenterology and Human Nutrition, All India Institute of Medical Sciences, New Delhi
India
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/ijpm.ijpm_354_22
留言 (0)