Context.—The data on status of apoptosis in patients with celiac disease are conflicting. Furthermore, complex interaction between intrinsic and common apoptotic pathways, apoptotic inhibitors, and epithelial cell proliferation is largely unclear for patients with celiac disease.
Objectives.—To determine the role of apoptosis and epithelial cell regeneration in celiac disease.
Design.—Twenty-five treatment-naïve patients with celiac disease and 6 patients with functional dyspepsia, as controls, were included and duodenal biopsy specimens from all were subjected to immunohistochemistry with markers of intrinsic apoptotic pathway (AIF, H2AX, p53), common pathway (CC3, M30), apoptotic inhibitors (XIAP, Bcl2), and epithelial proliferation (Ki-67). Apoptotic and proliferation indices were calculated.
Results.—Expression of end-apoptotic products, that is, H2AX in the cell nuclei (P = .01) and M30 in the cell cytoplasm (P < .01), was significantly upregulated in celiac disease in comparison to controls. Cleaved caspase-3 was also upregulated in villous cytoplasm in celiac disease. Apoptotic inhibitor Bcl2 was significantly down-regulated in celiac disease in comparison to controls. In addition, Ki-67 proliferation index was upregulated both in the crypts and villous mucosal epithelium in comparison to the crypts of the controls.
Conclusions.—Treatment-naïve patients with celiac disease have significantly higher level of apoptosis that involves both the common and intrinsic apoptotic pathways. Increased apoptosis and unequaled cell regeneration in crypts probably results in villous atrophy. Down-regulation of apoptotic inhibitors in treatment-naïve celiac disease imparts an additional pro-apoptotic effect.
Villous atrophy is the histomorphologic hallmark of celiac disease, which possibly occurs owing to increased cell destruction/enhanced apoptosis and/or a defect in epithelial cell regeneration.1–5 While a few studies have suggested a role for apoptosis in villous atrophy in patients with celiac disease, others do not confirm this finding.6–14 In such a study, Moss et al6 had reported presence of fragmented DNA in the intestinal mucosa of treatment-naïve patients with celiac disease, which disappeared after administration of gluten-free diet. Most other studies in this respect were limited to the identification of the role of FAS and FAS-L–mediated apoptosis.10–14 Studies by Croft and Cotton8 (1973), Hall et al9 (1994), Nagata and Golstein7 (1995), Lin et al10 (1998), Palejwala and Watson11 (2000), Maiuri et al14 (2001), Ciccocioppo et al13 (2002), and Miura et al12 (2005) have identified the role of extrinsic apoptotic pathway activation in the mucosa of patients with celiac disease. On the other hand, apoptosis can be induced either by the extrinsic or the intrinsic apoptotic pathways. While the extrinsic apoptotic pathway is initiated by the activation of Fas (CD95) and tumor necrosis factor α receptors and their respective ligands, the intrinsic pathway is initiated by apoptotic stresses, such as p53, and ultimately results in disruption of the mitochondrial membrane and leakage of cytochrome-c into cell cytoplasm. Subsequently, the apoptotic process proceeds through the activation of a battery of caspases in the common apoptosis pathway. Finally, the insoluble nuclear (H2AX) and/or cytoplasmic (M30/dCK18) apoptotic degradation products are formed. Moreover, these pathways are regulated by various external inhibitory signals, such as B-cell lymphoma 2 protein (Bcl-2), Bcl-xl, and A1/Bfl-1, which inhibit the intrinsic pathway, or the X-linked inhibitor of apoptotic protein (XIAP), which inhibits the caspases.7–9
As has been highlighted earlier, while the extrinsic apoptotic pathway plays a role in villous atrophy in patients with celiac disease, studies on the intrinsic and common apoptotic pathways in patients with celiac disease are sparse. Moreover, most of the published literature has relied on the terminal deoxynucleotidyl transferase–deoxyuridine triphosphate nick-end labeling (TUNEL) method for identifying the apoptotic cells by immunofluorescence technique.10,12,14 Considerable expertise is needed to apply and interpret this technique, especially for intestinal lesions. Studies using monoclonal antibodies against apoptotic markers are only a few. Ehrmann et al15 have studied apoptotic markers in jejunal biopsy specimens of active and potential celiac disease and have suggested the roles of Fas and/or Fas-L, Bcl2, and tissue transglutaminase in villous atrophy. In another study, Augustin et al16 could not demonstrate higher apoptotic activity in the mucosal epithelium of patients with celiac disease, when using the TUNEL technique, and immunohistochemical staining for M30 and Ki-67.
The research questions of the present study were centered on finding out whether the intrinsic and common pathways of apoptosis were active in celiac disease and on learning what happens to cell regeneration in patients with celiac disease. Therefore, we immunostained the duodenal biopsy specimens from patients with celiac disease and controls with markers of the intrinsic and common apoptotic pathways, along with markers of apoptotic inhibitors and Ki-67 cell proliferation, to understand if there was an imbalance between epithelial cell apoptosis and cell regeneration.
MATERIALS AND METHODS
Twenty five treatment-naïve patients with celiac disease were included in this study. The diagnosis of celiac disease was suspected by the presence of typical clinical manifestations (chronic diarrhea, anemia, and short stature) along with positive results for serum immunoglobulin A (IgA) anti–tissue transglutaminase antibody and presence of villous atrophy on histology. All these patients were given a gluten-free diet and diagnosis was confirmed when unequivocal responses to the same were obtained. Patients who were already following a gluten-free diet; patients with partially treated celiac disease, coexistent systemic diseases, human immunodeficiency virus seropositivity, or active or past history of tuberculosis; and patients unwilling to participate in the study were excluded.
Six patients with dyspepsia who were undergoing diagnostic esophagogastroduodenoscopy were recruited as controls. The endoscopic examination showed no abnormality and multiple duodenal biopsy samples were obtained from the junction of the second and third part of the duodenum. Results for IgA anti–tissue transglutaminase antibody were negative for all the controls. The hematologic and biochemical parameters were within the reference range and the duodenal biopsy specimens showed a normal morphology (Marsh grade 0).17 The mucosa of the esophagus, stomach, and duodenum was normal on endoscopic examination.
The study was approved by the institution's ethics committee and an informed consent was obtained from all patients or guardians of all the participants (if younger than 18 years).
The clinical setting for this study was the Tertiary Care Center at All India Institute of Medical Sciences in New Delhi, India.
Evaluation at Baseline
All the patients underwent a detailed evaluation at baseline. All the clinical manifestations were recorded along with relevant hematologic and biochemical tests. Serum anti-transglutaminase IgA levels were measured by enzyme-linked immunosorbent assay, with kits procured from The Binding Site Group Limited, Birmingham, United Kingdom, or INOVA Diagnostics, San Diego, California. A cutoff value of 4 or 20 U/mL, respectively, was used for the kits, as recommended by the manufacturers. Upper gastrointestinal endoscopy was done and the status of the duodenal folds was recorded (normal, attenuation, and scalloping of mucosal folds).
Duodenal Biopsy Protocol
For both patients and controls, upper gastrointestinal endoscopy was performed after proper written informed consent was obtained, and 4 to 5 duodenal mucosal fragments were sampled from the second and third part of the duodenum in each. The biopsy fragments were separated from the biopsy forceps by using a needle and were then embedded on a small square white filter paper by using a lighted hand-held lens for proper orientation. The fragments adhered to the filter paper owing to coagulation of plasma and were put into 10% neutral buffered formalin as soon as possible to avoid any fixation artifacts.
Histopathologic Examination of Duodenal Mucosal Biopsy Specimens
Mucosal biopsy specimens both from the patients with celiac disease and the controls were processed routinely in an automated tissue processor and the tissues were embedded in paraffin blocks. Hematoxylin-eosin–stained slides comprising at least 10 step-cuts per biopsy sample were prepared on a slide. Only well-oriented areas (at least 3–4 crypts arranged perpendicularly over the muscularis mucosae and away from submucosal Brunner glands) were analyzed. The modified Marsh grading system of Oberhuber and colleagues was used for grading the mucosal changes.17 All the duodenal biopsy specimens were evaluated by 2 pathologists, blinded to the clinical profile of the patients.
Markers for Apoptosis in Treatment-Naïve Patients With Celiac Disease and Controls
The biopsy specimens from patients with celiac disease and controls were subjected to immunohistochemical (IHC) staining for markers of the intrinsic apoptotic pathway (apoptosis-inducing factor [AIF], p53, and H2AX), markers of common pathway (cleaved caspase-3 [CC3] and M30 cytodeath), and markers of apoptotic inhibitors (XIAP and Bcl2).
Markers for Epithelial Cell Proliferation in Treatment-Naïve Patients With Celiac Disease and Controls
All the biopsy specimens from patients with celiac disease and controls were subjected to Ki-67 immunohistochemical staining to observe the epithelial regeneration activity.
For IHC, antigen retrieval was done by boiling the tissue sections in 0.01 M citrate buffer at pH 6. After blocking endogenous peroxidase activity and nonspecific antigen binding, the tissue sections were incubated with monoclonal antibodies against p53, AIF, XIAP, CC3 (Asp175), anti–histone H2AX, Bcl2, M30, and Ki-67 overnight in a moist chamber. Appropriate dilutions of these antibodies were used after standardization at our laboratory (Table 1). After washing in Tris buffer saline, the tissue sections were incubated with universal secondary antibody (REAL EnVision System, DAKO, Glostrup, Denmark). Immune detection was done by using 3,3′-diaminobenzidine as chromogen and hydrogen peroxide (H2O2) as substrate, followed by counterstaining with hematoxylin.
All the stained slides were analyzed for apoptotic index (AI) (number of positive cells present among 100 enterocytes examined), and at least 2000 duodenal enterocytes were assessed before recording the AI and stain intensity separately. The intensity of immunopositivity was graded as strongly positive (3+) when the stain intensity was equal to the stain intensity in positive control slides; faintly positive (1+) when the stain was just perceptible in the section examined; and moderately intense (2+) when stain intensity was in between the above 2 intensity grades. The IHC stains were standardized in appropriate tissue sections, and the standardization details are given in Table 1. Appropriate control sections were used for all the immunohistochemical markers (Figure 1, A through D). All the immunostains were analyzed by 2 pathologists with special interest in this field, and staining patterns were noted both in the villi and crypt nuclei as well as separately in the cytoplasm.
Proliferation index (PI) was calculated as the number of Ki-67–positive cells present among 100 enterocytes. Here also at least 2000 duodenal enterocytes were screened before calculating the PI.
Normally distributed continuous variables were expressed as mean (standard deviation). Categorical data were presented as proportions. Comparisons were done with the Mann-Whitney U test or t test for continuous variables and the χ2 test for discrete variables. A P value <.05 was taken as statistically significant. Data were analyzed with STATA software (version 11, StataCorp LP, College Station, Texas).
Baseline Characteristics of Patients Included
Mean age of the patients with celiac disease was 19.4 ± 8.3 years (14 males). The presenting manifestations were chronic diarrhea in 18 (72%), anemia in 20 (85%), and short stature in 15 (60%). The mean duration of symptoms was 6.3 ± 4.6 years. The mean baseline values were as follows: body weight, 32.6 ± 8.3 kg; body mass index, 15.8 ± 2.2 kg/m2; hemoglobin, 8.3 ± 1.8 g/dL; total protein, 7.7 ± 0.6 g/dL; albumin, 4.2 ± 0.4 g/dL; and calcium, 8.9 ± 0.6 mg/dL. Attenuated duodenal folds were present in 12 patients, scalloping of folds in 9, and normal duodenal folds in 4.
Baseline Characteristics of Controls
The mean age of controls was 20.7 ± 7.3 years (4 males). The mean baseline values were as follows: body weight, 35.2 ± 6 kg; body mass index, 17.5 ± 1.8 kg/m2; hemoglobin, 13.5 ± 0.6 g/dL; total protein, 8.1 ± 0.4 g/dL; albumin, 4.4 ± 0.3 g/dL; and calcium, 9.5 ± 0.4 mg/dL. The duodenal folds were normal in all.
Light Microscopic Features for Patients With Celiac Disease and Controls
Of the 25 patients with celiac disease, 22 (88%) had modified Marsh grade 3c villous abnormalities and the other 3 (12%) had grade 3b abnormalities. The duodenal mucosal biopsy specimens from all the 6 controls were within normal limits (Marsh grade 0).
Comparison of Apoptotic Markers in Duodenal Biopsy Specimens of Treatment-Naïve Patients With Celiac Disease and Controls
H2AX is the nuclear end product and M30 is the cytoplasmic end product of apoptosis, identifiable in a cell undergoing apoptosis. H2AX is the phosphorylated histone protein breakdown product, which was noted both in the enterocyte nuclei as well as in the cytoplasm of control duodenal biopsy specimens. The AI for H2AX was significantly higher in the villi and crypts of treatment-naïve patients with celiac disease than in those of controls (villi nucleus AI, P = .01; crypt nucleus AI, P = .01). The stain intensity was also higher in diseased mucosa than in the control biopsy samples (villous nuclear intensity, P = .008; crypt nuclear intensity, P = .005) (Table 2) (Figure 3E through 3H).
M30 antibody detects degraded fragments of cytokeratin 18 (dCK18) in the cytoplasm of a cell undergoing apoptosis and is considered the cytoplasmic end product of apoptosis. The cytoplasmic AI and stain intensity were significantly higher in the duodenal villi and crypts of patients with celiac disease (villi cytoplasmic AI, P < .001; cytoplasmic intensity, P < .001; crypt cytoplasmic AI, P < .001; crypt cytoplasmic intensity, P = .03) (Figure 2, A through D: M30 immunostain; Figure 2, E through H: H2AX immunostain) (Table 2).
Markers of Intrinsic Pathway of Apoptosis
Apoptosis-inducing factor is one of the important inducers of the intrinsic apoptotic pathway. Immunostaining for AIF showed a granular positivity pattern in the cell cytoplasm as well as focal nuclear positivity. Although we observed an increase in the pattern of cytoplasmic expression of AIF in the diseased mucosa compared to the controls, this difference was not statistically significant (nuclear intensity, P = .40 and cytoplasmic intensity, P = .10) (Figure 3, A through D).
p53, though classically a nuclear protein, showed nuclear as well as occasional cytoplasmic positivity. There was no significant difference in the nuclear p53 AI or intensity in either villi or crypts between biopsy specimens of treatment-naïve patients with celiac disease and those of controls (Table 2) (Figure 3E through 3H).
Markers of Common Apoptotic Pathway
Immunostaining with CC3 was identified either as dotlike cytoplasmic positivity of the apoptotic bodies or as focal nuclear positivity in the enterocytes. No significant difference in the AI was noted between the biopsy specimens of patients with treatment-naïve celiac disease and controls (Table 2). The intensity of CC3 staining was significantly higher in the villous cytoplasm of patients with celiac disease than in that of the controls (villous cytoplasmic intensity was 2.05 ± 1.39 in celiac disease versus 0.83 ± 1.32 in control biopsy specimens; P = .04); however, the stain intensity in crypts did not differ between patients with celiac disease and controls (P = .25) (Figure 4, A through D; Table 2).
Markers for Apoptosis Inhibitors
There was no significant difference in XIAP expression either in the villi or crypts in biopsy specimens from patients with treatment-naïve celiac disease or controls (Figure 4, E and F; Table 2). Bcl2 expression was taken as positive when expression was noted in the enterocyte cytoplasm or in the nuclear membrane. There was significant reduction of Bcl2 expression in the crypts (cytoplasmic distribution, P = .006; intensity, P = .002) as well as in the villi (villous distribution, P = .001; intensity, P = .001) in comparison to its expression in control biopsy specimens (Figure 4, G and H; Table 2).
Markers of Mucosal Epithelial Proliferation Index
Ki-67 protein expression was taken as positive when expressed in the duodenal enterocyte nuclei. Ki-67 PI was significantly higher in both the duodenal villi and crypts for biopsy specimens of treatment-naïve patients with celiac disease than for those of controls (villi: 51.4 ± 28.3 versus 10 ± 9.3, P = .004; crypts: 88.2 ± 23.3 versus 86 ± 7.6, P = .05) (Table 2). Whereas in control duodenal biopsy specimens the Ki-67 PI was restricted mostly to the crypt base, with occasional Ki-67–positive proliferating cells in duodenal villi, in active celiac disease the Ki-67 PI was higher both in the mucosal crypts and in the flattened mucous epithelium. The crypts showed nuclear positivity along their whole length in comparison to the basal positivity in control biopsy specimens, with many identifiable Ki-67–positive nuclei in the mucous epithelium above (Figure 5, A and B; Table 2).
In the present study apoptotic indices, as identified by immunohistochemical staining in the duodenal biopsy samples, were significantly higher for treatment-naïve patients with celiac disease than in controls. There was upregulation of cytoplasmic CC3 and M30 expression in the common apoptotic pathway, along with upregulation of H2AX nuclear apoptotic end product, which is the end product of the intrinsic apoptotic pathway. Among the other markers of intrinsic apoptotic pathway, nuclear p53 protein expression was significantly higher in the diseased mucosa than in the control biopsy specimens; cytoplasmic p53 expression was significantly higher in the former. A few authors, including Fontoura et al,18 have shown cytoplasmic p53 expression by demonstrating cytoplasmic covalent p53 linking with 5.8S ribosomal RNA component. Feng et al19 also demonstrated p53 protein expression in the distal transitional rectal mucosa within 4 cm of the margin in rectal carcinoma, in the cytoplasm of goblet cells. Whenever the nuclear DNA is injured, p53 causes cell cycle arrest through the upregulation of BAX and release of cytochrome-c and AIFs. However, in the absence of differential expression of nuclear p53 between diseased mucosa and controls, we refrain from reporting sole cytoplasmic p53 positivity in our cases. The AIF enters the enterocyte nucleus and causes phosphorylation of histone protein H2AX. H2AX is the final nuclear apoptosis degradation product of the intrinsic apoptotic pathway.20 In the present study, H2AX AI was significantly increased in both the duodenal villi and crypts of patients with celiac disease compared to controls. However, AIF expression was not significantly different in diseased mucosa and in controls. After being released, the cytochrome-c activates the caspases by sequential activation of procaspases 9 and 3, resulting in formation of cleaved CC3.21,22 The CC3 AI in the present study was higher in the diseased duodenal villi than in the controls. This cleaved CC3 finally cleaves cytoplasmic CK18, resulting in formation of the cytoplasmic end-apoptotic product M30 through the common apoptotic pathway. M30 was significantly overexpressed in both villi and crypts of the duodenal biopsy specimens from treatment-naïve patients with celiac disease in comparison to the controls. As we did not study the extrinsic apoptotic pathway in this study, an additive role of the extrinsic pathway could not be ruled out for our patients. However, keeping in mind the results of the previously published studies,8–14 we can conclude that, along with the extrinsic pathway, the common as well as the intrinsic apoptotic pathways are active in treatment-naïve patients with celiac disease.
In the present study, apoptotic inhibitor Bcl2 was significantly underexpressed in the diseased mucosa compared to normal duodenal biopsy specimens. Bcl2, in addition to being an antiapoptotic factor, has also antiautophagic activities.23,24 In celiac disease, cytokine-induced apoptotic activities increase owing to gluten-induced autoimmune reactions, as shown in this study. Simultaneously, owing to stresses in the local microenvironment, macroautophagy and chaperone-mediated autophagy destroy the cells' internal structures with the activity of lysosomes. Bcl2 is also known to have an antagonistic activity against one of the autophagy-inducing proteins, beclin-1, thus preventing unwanted self-destruction of cell organelles.23,24 In celiac disease, Bcl2 expression is significantly down-regulated, possibly preventing both the apoptotic and autophagy activities in the already-diseased mucosa.23,24 However, another apoptotic inhibitor, XIAP, was not differentially expressed in diseased mucosa and controls. Despite the presence of XIAP, the apoptotic process was seen to continue, possibly because of some unidentified inactivating phosphorylation process. Overall, our results may indicate that accelerated apoptosis, in combination with inhibition of antiapoptotic signals, cause villous atrophy in patients with celiac disease.
The role of apoptosis in celiac disease is controversial, however, and this point has been addressed in various publications.6,12,14–16 Van Der Woude et al25 could not demonstrate enhanced apoptosis with immunohistochemical stains for active caspase-3, CK18, and apoptosis-related proteins Fas, iNOS, Bcl-2, Bcl-xl, and Bax in patients with celiac disease and thus the authors concluded that the loss of villi in active celiac disease may be the result of other unknown mechanisms. However, our results show the role of apoptosis in villous atrophy in celiac disease. In a study by Maiuri et al,26 interleukin 15 was identified as a factor modulating epithelial FAS, transferrin receptor, and Ki-67 proliferation index, while interleukin 10 has been shown to diminish epithelial cell apoptosis owing to interferon ∞, tumor necrosis factor, and anti-Fas monoclonal antibody.27–29 Interleukin15 is believed to mediate the apoptotic changes in duodenal mucosa in celiac disease.26
What Is the Role of Cell Regeneration in Celiac Disease?
Normal regenerative activity is seen mostly in the duodenal crypts. In the present study, there was significant increase of Ki-67 PI in both the duodenal villi and crypts in treatment-naïve patients with celiac disease as compared to the controls (Table 2). These data show that apoptosis and cellular regeneration go hand in hand in the diseased villi. A logical proposition may be that the regenerating activity in the crypts is not sufficient to replenish the high cellular turnover in the villous mucosal lining cells.16,17,26 The findings presented in this article are exciting as they raise the possibility that apoptotic activity may be the principal factor responsible for duodenal villous atrophy, though we cannot rule out the additional role of stress-induced autophagy in our patients. Future therapeutic approaches may be aimed at reducing the increased apoptotic activity in celiac disease. Studying the apoptotic end-product levels during follow-up of patients given gluten-free diets may prove beneficial.
In conclusion, apoptotic activity is increased in the intestinal mucosa of treatment-naïve patients with celiac disease. Along with the extrinsic apoptotic pathway, the intrinsic as well as the final common apoptotic pathways take part in executing the apoptotic activity. The apoptotic inhibitors are down-regulated in diseased mucosa, thus ensuring efficient apoptotic activities. The increased level of apoptosis, along with the inefficient regenerative activity in crypts, most probably results in villous atrophy in celiac disease.
The authors have no relevant financial interest in the products or companies described in this article.