- Email: [email protected]
Effects of Topical Steroids on Tight Junction Proteins and Spongiosis in Esophageal Epithelia of Patients With Eosinophilic Esophagitis
Clinical Gastroenterology and Hepatology 2014;-:-–-
Effects of Topical Steroids on Tight Junction Proteins and Spongiosis in Esophageal Epithelia of Patients With Eosinophilic Esophagitis David A. Katzka,* Ravikanth Tadi,* Thomas C. Smyrk,*,‡ Eesha Katarya,* Anamay Sharma,* Deborah M. Geno,* Michael Camilleri,* Prasad G. Iyer,* Jeffrey A. Alexander,* and Navtej S. Buttar* *Division of Gastroenterology and Hepatology, and ‡Division of Pathology, Mayo Clinic, Rochester, Minnesota BACKGROUND & AIMS:
The allergic response associated with eosinophilic esophagitis (EoE) occurs when food antigens permeate tight junction–mediated epithelial dilated intercellular spaces. We assessed whether levels of tight junction proteins correlate with the dilation of intercellular spaces (spongiosis) and the effects of topical steroids on these parameters.
METHODS:
We assessed esophageal biopsy samples from 10 patients with active EoE treated with topical fluticasone, 10 untreated patients, and 10 patients without esophageal disease (controls) for degree of spongiosis. Immunohistochemical assays were used to determine the levels of the tight junction proteins filaggrin, zonula occludens (ZO)-1, ZO-2, ZO-3, and claudin-1. Histology and immunohistochemistry results were assessed blindly, with levels of tight junction proteins and degree of spongiosis rated on scales of 0 to 3.
RESULTS:
The mean degrees of spongiosis in untreated and treated patients with EoE were 1.3 and 0.4, respectively (P [ .016). Esophageal epithelia did not stain significantly for ZO-1 or ZO-2. Filaggrin was observed in a predominant cytoplasmic pattern, compared with the cytoplasmic and membranous patterns of ZO-3 and claudin-1. In biopsy specimens from patients with active EoE, the mean staining intensities for filaggrin, ZO-3, and claudin-1 were 1.6, 1.4, and 0.7, respectively. In biopsy specimens from patients treated with fluticasone, levels of filaggrin, ZO-3, and claudin-1 were 2.8 (P [ .002 compared with untreated patients), 1.7 (P [ .46 compared with untreated patients), and 1.3 (P [ .25 compared with untreated patients), respectively. The correlation between the level of filaggrin and the degree of spongiosis was r [ 0.23, and between ZO-3 staining and the degree of spongiosis was r [ .016 (P [ .001 for filaggrin vs ZO-3 staining).
CONCLUSIONS:
Filaggrin, ZO-3, and claudin-1 (but not ZO-1 or ZO-2) are detected in the esophageal mucosa of patients with EoE treated with steroids and individuals without esophageal disease. Without treatment, spongiosis increases, corresponding with reduced levels of filaggrin, ZO-3, and claudin-1. Loss of tight junction regulators and dilation of intercellular spaces appear to be involved in the pathophysiology of EoE and could be targets for treatment.
Keywords: Esophagus; Inflammation; Therapy; Allergy.
osinophilic esophagitis (EoE) is an allergymediated disease in which food antigens in contact with esophageal mucosa generate a T-helper cell type 2 (Th-2) cascade leading to tissue eosinophilia, inflammation, and fibrosis.1 The recognition of these antigens is believed to be mediated by dendritic cells that reside in the esophageal mucosa. The precise mechanism by which this interaction occurs is unknown. For antigen penetration of the deeper layers of the mucosa to be achieved, paracellular transport must occur. This largely is regulated through tight junction regulators and proteins that can lead to widening of
E
the gap between epithelial cells. This is referred to histologically as dilated intercellular spaces or spongiosis. Spongiosis has been well documented as a common histopathologic finding in patients with EoE. It is not clear, however, how well spongiosis responds to steroid
Abbreviations used in this paper: EoE, eosinophilic esophagitis; eos, eosinophils; HPF, high-power field; ZO, zonula occludens. © 2014 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2014.02.039
2
Katzka et al
Clinical Gastroenterology and Hepatology Vol.
treatment. Furthermore, there is little information on tight junction proteins in esophageal epithelium and their role in dilated intercellular spaces. In this study, we investigated the following: (1) which tight junction proteins are present in esophageal epithelia, (2) if staining density corresponds inversely to the presence of dilation of intercellular spaces in esophageal epithelium as measured by the histologic appearance of spongiosis, and (3) if with a steroid-induced reduction in the intercellular space (manifest by spongiosis) there is an accompanying increase in staining for tight junction protein.
Methods Twenty patients with EoE were included in this study. The diagnosis of EoE was established by consensus guidelines.1 In addition, all patients had at least one atopic characteristic (asthma, rhinitis, oral allergy syndrome, or atopic dermatitis). Ten of these patients had untreated EoE with more than 15 peak eosinophils (eos) per high-power field (HPF), a history of dysphagia, consistent endoscopic findings, and a lack of histologic response to an 8-week course of proton pump inhibitors or a negative ambulatory esophageal pH monitoring study. Biopsy specimens were analyzed from another 10 patients with proven pretreatment active EoE who were treated with 880 ug oral fluticasone twice daily for 2 months. Esophageal biopsy specimens showing a reduction of esophageal eosinophilia to fewer than 5 eos per HPF obtained at the end of treatment were used for analysis. Esophageal biopsy specimens from 10 control patients also were analyzed. These patients underwent endoscopy for clinical reasons unrelated to esophageal disease and all had a normal esophageal endoscopic appearance and were histologically normal. None of these patients had a history of gastroesophageal reflux disease. Neither patients nor controls were on acid-suppressing therapy at the time of the study. Formalin-fixed, paraffin-embedded esophageal biopsy specimens were cut into 3-mm sections and the unstained sections were treated with a progressively concentrated xylene wash, and then incubated in a citrate buffer within a water bath heated to 85 F. After completion of this step to expose antigen-binding sites for one tight junction protein of interest (claudin-1, zonula occludens [ZO]-1, ZO-2, ZO-3, and filaggrin), the optimal dilution of the primary antibody for that protein was added to each
-,
No.
-
slide (Abcam, Cambridge, MA) (Supplementary Table 1). Because there are little prior data studying these proteins in esophageal tissue, checkerboard titrations were used to determine the optimal dilutions of the primary antibody and other reagents. Skipping the primary antibody was used as a negative control and skin was used as a positive control. Specimens were left overnight at 4 C to maximize binding of the antibody. After a secondary antibody was added to enhance specificity, each tissue was treated with the ABC reagent kit (Vector Laboratories, Burlingame, CA) to attach an enzymatic substrate to each site. Samples subsequently underwent a colorimetric peroxidase reaction under timed conditions. Once microscopy was used to optimize the level of histologic staining, the reaction was quenched. Each slide then was washed and processed with a progressively dilute xylene for preservation of immunostaining. Finally, each sample was visualized under a microscope and a semiquantitative scale was applied to determine the location and concentration within each tissue specimen. Grading of tight junction staining was performed with a scale of 0 to 3þ based on the control patient biopsy specimens. On routine histologic analysis, the degree of spongiosis was graded on a scale of 0 to 3. All slides (both EoE and controls) were read blindly by one of the investigators (T.C.S.). Grading of spongiosis and tight junction staining was determined by calculating the average finding over all the biopsy specimens in each paraffin block to account for possible patchiness in the distribution of findings.
Statistical Analysis Analysis of variance with Tukey’s test were used to compare results from the 3 groups. The Spearman coefficient was used to determine correlation of spongiosis to tight junction staining. Approval was obtained for this study through the Mayo Clinic Institutional Review Board.
Results The clinical and histologic profiles of the 20 patients used in this study are shown in Table 1. Nineteen of 20 patients included in this study had typical endoscopic features of EoE including rings, linear furrowing, white plaques, or mucosal friability. In the treated group, 6 patients had no eos present and 8 patients had fewer
Table 1. Patient Group Characteristics EoE groups Untreated Fluticasone
Age, y (range)
Male
Peak eos/HPF (range)
Dysphagia
Characteristic EGD EoE, findings/mean EREF score19
40 (9–51) 39 (7–53)
9 9
57 (20–120) 37 (15–100)
10 10
10/3.2 9/3.5
EREF, Endoscopic Reference Score.
-
2014
Tight Junctions in Eosinophilic Esophagitis
3
Figure 1. Positive staining for (A) filaggrin, (B) ZO-3, and (C) claudin-1 in control patients. Demonstration of cytoplasmic (D) and membranous (E) staining. (F) Tight junctions can be seen with positive staining in higher magnification (arrow).
than 4 eos/HPF. In the untreated group, 2 patients had more than 100 eos/HPF. The most common atopic manifestations included rhinitis, asthma, oral allergy syndrome, and food allergies. Only 2 patients had atopic dermatitis.
Control Biopsy Specimens The mean staining intensity in control biopsy specimens for filaggrin, ZO-3, and claudin-1 (Figure 1) were 2.6, 2.1, and 1.3, respectively. There was no significant staining noted for ZO-1 or ZO-2. IgG staining alone was negative for all controls. For filaggrin, approximately 50% to 80% of the mucosa stained with filaggrin was located on the luminal side of the mucosa. Staining was noted mostly to be cytoplasmic. For ZO-3, most of the staining occurred on the basilar side of the mucosa (Figure 1C), but in 3 biopsy specimens luminal staining could be seen additionally on the luminal side and in 1 biopsy the entire mucosa was positive for ZO-3 staining. Staining was noted to be nuclear, cytoplasmic, and/or
membranous, and in some biopsy specimens positive staining of tight junctions could be seen. For claudin-1, most of the staining was also in the basal layer, estimated to range from 5% to 50%. Staining was mostly membranous, but there was some weak cytoplasmic staining. In some biopsy specimens tight junction staining also was seen.
Eosinophilic Esophagitis Biopsy Specimens In patients with active EoE, esophageal epithelia positively stained for claudin-1, ZO-3, and filaggrin (Figure 2). ZO-1 and ZO-2 were not detected. The mean staining grade intensity in order of increasing staining intensity in untreated patients was 0.7 for Claudin-1, 1.4 for ZO-3, and 1.6 for filaggrin. All untreated patients had positive staining for filaggrin and 6 of 10 patients had grade 1. Only 2 patients had a maximal filaggrin staining intensity score of 3. In the 7 untreated patients with positive filaggrin staining, 5 patients showed predominantly cytoplasmic staining although overlap was
Figure 2. Tight junction staining in patients with (A–C) untreated and (D–F) treated EoE. (A) Grades 0 to 1 filaggrin with spongiosis. (B) Grade 1 ZO-3 staining with spongiosis. (C) Grade 1 claudin-1 with spongiosis. (D) Grade 3 filaggrin staining with predominantly cytoplasmic staining. (E) Grade 1 ZO-3, membranous pattern. (F) Grade 3 claudin-1 with a membranous pattern. Original magnification, 10.
4
Katzka et al
common. For ZO-3, only 1 patient had grade 3 staining whereas 2 patients had no detectable staining. After fluticasone treatment, the mean staining intensity was 1.3 for claudin-1, 1.7 for ZO-3, and 2.6 for filaggrin. The increase in filaggrin staining was increased significantly (P ¼ .046), with 8 of 10 post-treatment biopsy specimens now showing a maximal staining score of 3. Similar to controls, the most intense staining occurred toward the luminal side of the esophageal epithelium. There was both robust cytoplasmic and membranous staining present.
Spongiosis In biopsy specimens from the 10 untreated patients with EoE, the mean degree of spongiosis was 1.3 (Figure 3D). In patients with EoE who had received treatment with fluticasone, the mean spongiosis score was significantly less (0.4; P ¼ .016). Within individual patients, spongiosis was detected in 8 of 10 untreated patients, whereas in patients who had received fluticasone treatment, only 3 patients were found to have spongiosis on biopsy. In these treated patients, no patients with absent eosinophilia (5 patients) had spongiosis. In the 10 patients with active EoE, the 2 patients without spongiosis had the lowest levels of esophageal eosinophilia (25 and 22 eos/HPF). In the 8 patients with
Clinical Gastroenterology and Hepatology Vol.
-,
No.
-
grades 1 to 3 spongiosis, there was no clear correlation, however, between the level of esophageal eosinophilia and the grade of spongiosis (data not shown). Of the 6 patients with treated EoE and no esophageal eosinophilia, none of these patients had detectable spongiosis. Only 1 control patient showed spongiosis.
Correlation to Tight Junction Proteins The correlation of the degree of spongiosis to the staining intensity of filaggrin and ZO-3 from individual biopsy specimens was calculated (Figure 4). There was a significant inverse correlation of filaggrin to spongiosis (r ¼ -0.68; P ¼ .001). For ZO-3, there was also an inverse correlation to spongiosis (r ¼ -0.53; P ¼ .016). There was no significant correlation between the degree of spongiosis and staining of claudin-1.
Discussion Tight junction proteins are essential to all cells that interconnect cell membranes to regulate paracellular flow of ions, compounds, and cells through the intercellular channels. They exist either as transmembrane proteins (eg, claudins) or as cytoplasmic proteins that connect the cell cytoskeleton to the membrane proteins (eg, zonulin). When there is dysfunction of these
Figure 3. Comparison of staining intensity in active, inactive, and control esophageal biopsy specimens for (A) filaggrin, (B) ZO-3, and (C) claudin-1. (D) Spongiosis in patient groups.
-
2014
Tight Junctions in Eosinophilic Esophagitis
5
Figure 4. Correlation of degree of spongiosis with intensity of (A) filaggrin and (B) ZO-3 staining in biopsy specimens of active and EoE.
proteins, the range of disease can vary2 from alterations in the blood brain barrier, suppression of tumorigenesis,3 to enhanced transmembranes, as occurs with gluten in celiac disease.4,5 In this study, in biopsy specimens from patients with EoE, we show the following: (1) tight junction proteins filaggrin, claudin-1, and ZO-3, but not ZO-2 or ZO-3, are present in esophageal epithelium; and (2) topical steroids are associated with less spongiosis, a histologic sign of dilated intercellular spaces, and greater staining for tight junction proteins, which correlates inversely to spongiosis. We show that there are specific populations of tight junction proteins in esophageal epithelium with different levels of staining. In our study, the most ubiquitous protein found was filaggrin, followed by ZO-3 and claudin-1. It is known that the relative proportion of these proteins found will vary with the tissue studied. In the esophagus in 2 previous reports, there was conflicting evidence on the presence of the tight junction, occludin, in esophageal epithelium.6 On the other hand, claudin -1, similar to our study, also has been shown in normal esophageal epithelium, whereas claudin-2 has not.7 Our data are novel in testing more comprehensively for tight junction presence in esophageal epithelium. Our data also may be important in specifically showing that filaggrin is decreased in untreated patients with EoE and is restored after treatment with steroids. As importantly, the restoration of filaggrin correlates to a reduction in histologic spongiosis, that is, dilation of the intercellular spaces. These data fit well with 2 key studies that described normalization of the esophageal transcriptome with steroid therapy8 (and therefore the potential to restore synthesis of tight junction proteins), and genetic studies in children with EoE showed abnormalities of the gene that encodes for filaggrin as a potential mechanism of disease.9 The question most relevant to EoE is how do these abnormalities of tight junctions in patients with EoE promote disease? Although increased antigenic penetration of the mucosa seems likely, there also are data showing the important role of tight junctions in allowing leukocyte migration.10 Unfortunately, further work is needed in identifying the precise correlation of change in
tight junction presence to dilation of intercellular spaces in EoE. Whether this specifically occurs in EoE, particularly with dendritic cell migration and antigen recognition, is unclear. Another important question is whether the findings described here are unique to EoE or whether they are a nonspecific consequence of esophageal inflammation from a variety of causes. Specifically, in patients with gastroesophageal reflux disease, clear changes in tight junction proteins in response to acid exposure and/or injury have been described. For example, in patients with erosive or nonerosive reflux disease, an inverse correlation between claudin-1 and mucosal impedance has been reported.11 In contrast, in an epithelial cell layer derived from bronchial cells, acid or bile exposure led to a decrease in impedance but a negative correlation with claudin-4.12,13 In a rat model of reflux, abnormalities in claudins 1, 3, and 4 have been reported.14,15 Thus, one major potential difference in the mechanism of change in mucosal impedance and dilation of intercellular spaces in gastroesophageal reflux and EoE might be the tight junction pathway through which this injury occurs. This study had some limitations. First, the analysis before and after fluticasone treatment did not use the same patients. On the other hand, all patients chosen met strict guideline criteria for EoE. Furthermore, a different treated group was chosen in which there was a complete histologic response (<5 eos/HPF) to ensure that interpretation of tight junction staining was not performed and therefore equivocally interpreted in the setting of a partial response. In other words, this analysis allowed us to determine reliably the bearing of the esophageal eosinophilic component on intercellular spacing and tight junction protein regulation. Often this is not easily achieved in all treated patients. Second, is that this was a relatively small number of patients studied, although large enough to at least make some statistically significant comparisons and qualitatively evaluate for the presence of various tight junction proteins in esophageal epithelia. Third, is that we cannot determine if these results are unique to EoE or relevant to other diseases that manifest with esophageal dilated intercellular spaces, such as in gastroesophageal reflux. Nonetheless,
6
Katzka et al
we conducted this study on patients off proton pump inhibitors to attempt to isolate the effect of steroid therapy. Fourth, a limitation of this study was the initial inaccuracy of light microscopy to assess the degree of dilated intercellular spaces rather than using electron microscopy.16 On the other hand, more recent studies11 have shown that routine H&E staining can be accurate in measuring the diameter of intercellular spaces in esophageal epithelium, even in a head-to-head comparison with electron microscopy.17 This most likely is owing to increased recognition of this pattern by expert gastrointestinal pathologists. A fifth limitation of this study was that the assessment of tight junction proteins was semiquantitative because tissue was not available for measurement using quantitative polymerase chain reaction. Future studies would benefit additionally from using this assessment of barrier protein presence. Finally, this study examined only one aspect of intercellular space regulation, that is, tight junctions. Recent literature also has reported desmosomal dysfunction in patients with EoE.18 In conclusion, our data show the existence and relative density of specific tight junction proteins in human esophageal epithelium. We further show that the staining density of these proteins in patients with EoE may be altered by the administration of steroid therapy. Finally, the density of 2 of these proteins, filaggrin and ZO-3, correlates inversely to spongiosis, that is, dilated intercellular spaces, when the latter is corrected with steroid therapy.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2014.02.039.
References 1. Liacouras CA, Furuta GT, Hirano I, et al. Eosinophilic esophagitis: updated consensus recommendations for children and adults. J Allergy Clin Immunol 2011;128:3–20. 2. Odenwald MA, Turner JR. Intestinal permeability defects: is it time to treat. Clin Gastroenterol Hepatol 2013;111:1075–1083. 3. Coisne C, Engelhardt B. Tight junctions in brain barriers during central nervous system inflammation. Antioxidants Redox Signaling 2011;15:285–303. 4. Fasano A, Not T, Wang W, et al. Zonulin, a newly discovered modulator of intestinal permeability, and its expression in celiac disease. Lancet 2000;355:1518–1519. 5. Drago S, El Asmar R, Di Pierro M, et al. Gliadin, zonulin and gut permeability: effects on celiac and non-celiac intestinal mucosa and intestinal cell lines. Scand J Gastroenterol 2006;41: 408–419.
Clinical Gastroenterology and Hepatology Vol.
-,
No.
-
6. Kimura Y, Shiozaki H, Hirao M, et al. Expression of occludin tight-junction-associated protein, in human digestive tract. Am J Pathol 1997;151:45–54. 7. Rendon-Huerta E, Valenzano MC, Mullin JM, et al. Comparison of three integral tight junction barrier proteins in Barrett’s epithelium versus normal esophageal epithelium. Am J Gastroenterol 2003;98:1901–1902. 8. Blanchard C, Mingler MK, Vicario M, et al. IL-13 involvement in eosinophilic esophagitis: transcriptome analysis and reversibility with glucocorticoids. J Allergy Clin Immunol 2007;120: 1292–1300. 9. Blanchard C, Wang N, Stringer KF, et al. Eotaxin-3 and a uniquely conserved gene-expression profile in eosinophilic esophagitis. J Clin Invest 2006;116:536–547. 10. Steed E, Balda MS, Matter K. Dynamics and functions of tight junctions. Trends Cell Biol 2009;20:141–149. 11. Zhong C, Duan L, Wang K, et al. Esophageal intraluminal baseline impedance is associated with severity of acid reflux and epithelial structural abnormalities in patients with gastroesophageal reflux disease. J Gastroenterol 2013;48: 601–610. 12. Oshima T, Koseki J, Chen X, et al. Acid modulates the squamous epithelial barrier function by modulating the localization of claudins in the superficial layers. Lab Invest 2012;92:22–31. 13. Chen X, Oshima T, Shan J, et al. Bile salts disrupt human esophageal squamous epithelial barrier function by modulating tight junction proteins. Am J Physiol Gastrointest Liver Physiol 2012;303:G199–G208. 14. Asaoka D, Miwa H, Hirai S, et al. Altered localization and expression of tight-junction proteins in a rat model with chronic acid reflux esophagitis. J Gastroenterol 2005;40:781–790. 15. Oguro M, Koike M, Ueno T, et al. Dissociation and dispersion of claudin-3 from the tight junction could be one of the most sensitive indicators of reflux esophagitis in a rat model of the disease. J Gastroenterol 2011;46:629–638. 16. Tobey NA, Carson JL, Alkiek RA, et al. Dilated intercellular spaces: a morphological feature of acid reflux–damaged human esophageal epithelium. Gastroenterology 1996;111: 1200–1205. 17. Ribolsi M, Perrone G, Caviglia R, et al. Intercellular space diameters of the oesophageal epithelium in NERD patients: head to head comparison between light and electron microscopy analysis. Dig Liver Dis 2009;41:9–14. 18. Sherrill JD, Kc K, Wu D, et al. Desmoglein-1 regulates esophageal epithelial barrier function and immune responses in eosinophilic esophagitis. Mucosal Immunol 2013 Nov 13. Epub ahead of print. 19. Hirano I, Moy N, Heckman MG, et al. Endoscopic assessment of the oesophageal features of eosinophilic esophagitis: validation of a novel classification and grading system. Gut 2013;62: 489–495. Reprint requests Address requests for reprints to: David A. Katzka, MD, Mayo Clinic, 200 First Avenue SW, Rochester, Minnesota 55905. e-mail: [email protected]; fax: (507) 538-5820. Conflicts of interest The authors disclose no conflicts.
-
2014
Tight Junctions in Eosinophilic Esophagitis
Supplementary Table 1. Tight Junction Antibody Dilutions S no. Antibody Manufacturer Catalogue no. Dilution used 1 2 3 4
Filaggrin Claudin-1 ZO-3 ZO-1
S, sample.
Abcam Abcam Abcam Abcam
ab ab ab ab
81468 15098 64326 59720
1:250 1:100 1:100 1:100
6.e1
Effects of Topical Steroids on Tight Junction Proteins and Spongiosis in Esophageal Epithelia of Patients With Eosinophilic Esophagitis David A. Katzka,* Ravikanth Tadi,* Thomas C. Smyrk,*,‡ Eesha Katarya,* Anamay Sharma,* Deborah M. Geno,* Michael Camilleri,* Prasad G. Iyer,* Jeffrey A. Alexander,* and Navtej S. Buttar* *Division of Gastroenterology and Hepatology, and ‡Division of Pathology, Mayo Clinic, Rochester, Minnesota BACKGROUND & AIMS:
The allergic response associated with eosinophilic esophagitis (EoE) occurs when food antigens permeate tight junction–mediated epithelial dilated intercellular spaces. We assessed whether levels of tight junction proteins correlate with the dilation of intercellular spaces (spongiosis) and the effects of topical steroids on these parameters.
METHODS:
We assessed esophageal biopsy samples from 10 patients with active EoE treated with topical fluticasone, 10 untreated patients, and 10 patients without esophageal disease (controls) for degree of spongiosis. Immunohistochemical assays were used to determine the levels of the tight junction proteins filaggrin, zonula occludens (ZO)-1, ZO-2, ZO-3, and claudin-1. Histology and immunohistochemistry results were assessed blindly, with levels of tight junction proteins and degree of spongiosis rated on scales of 0 to 3.
RESULTS:
The mean degrees of spongiosis in untreated and treated patients with EoE were 1.3 and 0.4, respectively (P [ .016). Esophageal epithelia did not stain significantly for ZO-1 or ZO-2. Filaggrin was observed in a predominant cytoplasmic pattern, compared with the cytoplasmic and membranous patterns of ZO-3 and claudin-1. In biopsy specimens from patients with active EoE, the mean staining intensities for filaggrin, ZO-3, and claudin-1 were 1.6, 1.4, and 0.7, respectively. In biopsy specimens from patients treated with fluticasone, levels of filaggrin, ZO-3, and claudin-1 were 2.8 (P [ .002 compared with untreated patients), 1.7 (P [ .46 compared with untreated patients), and 1.3 (P [ .25 compared with untreated patients), respectively. The correlation between the level of filaggrin and the degree of spongiosis was r [ 0.23, and between ZO-3 staining and the degree of spongiosis was r [ .016 (P [ .001 for filaggrin vs ZO-3 staining).
CONCLUSIONS:
Filaggrin, ZO-3, and claudin-1 (but not ZO-1 or ZO-2) are detected in the esophageal mucosa of patients with EoE treated with steroids and individuals without esophageal disease. Without treatment, spongiosis increases, corresponding with reduced levels of filaggrin, ZO-3, and claudin-1. Loss of tight junction regulators and dilation of intercellular spaces appear to be involved in the pathophysiology of EoE and could be targets for treatment.
Keywords: Esophagus; Inflammation; Therapy; Allergy.
osinophilic esophagitis (EoE) is an allergymediated disease in which food antigens in contact with esophageal mucosa generate a T-helper cell type 2 (Th-2) cascade leading to tissue eosinophilia, inflammation, and fibrosis.1 The recognition of these antigens is believed to be mediated by dendritic cells that reside in the esophageal mucosa. The precise mechanism by which this interaction occurs is unknown. For antigen penetration of the deeper layers of the mucosa to be achieved, paracellular transport must occur. This largely is regulated through tight junction regulators and proteins that can lead to widening of
E
the gap between epithelial cells. This is referred to histologically as dilated intercellular spaces or spongiosis. Spongiosis has been well documented as a common histopathologic finding in patients with EoE. It is not clear, however, how well spongiosis responds to steroid
Abbreviations used in this paper: EoE, eosinophilic esophagitis; eos, eosinophils; HPF, high-power field; ZO, zonula occludens. © 2014 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2014.02.039
2
Katzka et al
Clinical Gastroenterology and Hepatology Vol.
treatment. Furthermore, there is little information on tight junction proteins in esophageal epithelium and their role in dilated intercellular spaces. In this study, we investigated the following: (1) which tight junction proteins are present in esophageal epithelia, (2) if staining density corresponds inversely to the presence of dilation of intercellular spaces in esophageal epithelium as measured by the histologic appearance of spongiosis, and (3) if with a steroid-induced reduction in the intercellular space (manifest by spongiosis) there is an accompanying increase in staining for tight junction protein.
Methods Twenty patients with EoE were included in this study. The diagnosis of EoE was established by consensus guidelines.1 In addition, all patients had at least one atopic characteristic (asthma, rhinitis, oral allergy syndrome, or atopic dermatitis). Ten of these patients had untreated EoE with more than 15 peak eosinophils (eos) per high-power field (HPF), a history of dysphagia, consistent endoscopic findings, and a lack of histologic response to an 8-week course of proton pump inhibitors or a negative ambulatory esophageal pH monitoring study. Biopsy specimens were analyzed from another 10 patients with proven pretreatment active EoE who were treated with 880 ug oral fluticasone twice daily for 2 months. Esophageal biopsy specimens showing a reduction of esophageal eosinophilia to fewer than 5 eos per HPF obtained at the end of treatment were used for analysis. Esophageal biopsy specimens from 10 control patients also were analyzed. These patients underwent endoscopy for clinical reasons unrelated to esophageal disease and all had a normal esophageal endoscopic appearance and were histologically normal. None of these patients had a history of gastroesophageal reflux disease. Neither patients nor controls were on acid-suppressing therapy at the time of the study. Formalin-fixed, paraffin-embedded esophageal biopsy specimens were cut into 3-mm sections and the unstained sections were treated with a progressively concentrated xylene wash, and then incubated in a citrate buffer within a water bath heated to 85 F. After completion of this step to expose antigen-binding sites for one tight junction protein of interest (claudin-1, zonula occludens [ZO]-1, ZO-2, ZO-3, and filaggrin), the optimal dilution of the primary antibody for that protein was added to each
-,
No.
-
slide (Abcam, Cambridge, MA) (Supplementary Table 1). Because there are little prior data studying these proteins in esophageal tissue, checkerboard titrations were used to determine the optimal dilutions of the primary antibody and other reagents. Skipping the primary antibody was used as a negative control and skin was used as a positive control. Specimens were left overnight at 4 C to maximize binding of the antibody. After a secondary antibody was added to enhance specificity, each tissue was treated with the ABC reagent kit (Vector Laboratories, Burlingame, CA) to attach an enzymatic substrate to each site. Samples subsequently underwent a colorimetric peroxidase reaction under timed conditions. Once microscopy was used to optimize the level of histologic staining, the reaction was quenched. Each slide then was washed and processed with a progressively dilute xylene for preservation of immunostaining. Finally, each sample was visualized under a microscope and a semiquantitative scale was applied to determine the location and concentration within each tissue specimen. Grading of tight junction staining was performed with a scale of 0 to 3þ based on the control patient biopsy specimens. On routine histologic analysis, the degree of spongiosis was graded on a scale of 0 to 3. All slides (both EoE and controls) were read blindly by one of the investigators (T.C.S.). Grading of spongiosis and tight junction staining was determined by calculating the average finding over all the biopsy specimens in each paraffin block to account for possible patchiness in the distribution of findings.
Statistical Analysis Analysis of variance with Tukey’s test were used to compare results from the 3 groups. The Spearman coefficient was used to determine correlation of spongiosis to tight junction staining. Approval was obtained for this study through the Mayo Clinic Institutional Review Board.
Results The clinical and histologic profiles of the 20 patients used in this study are shown in Table 1. Nineteen of 20 patients included in this study had typical endoscopic features of EoE including rings, linear furrowing, white plaques, or mucosal friability. In the treated group, 6 patients had no eos present and 8 patients had fewer
Table 1. Patient Group Characteristics EoE groups Untreated Fluticasone
Age, y (range)
Male
Peak eos/HPF (range)
Dysphagia
Characteristic EGD EoE, findings/mean EREF score19
40 (9–51) 39 (7–53)
9 9
57 (20–120) 37 (15–100)
10 10
10/3.2 9/3.5
EREF, Endoscopic Reference Score.
-
2014
Tight Junctions in Eosinophilic Esophagitis
3
Figure 1. Positive staining for (A) filaggrin, (B) ZO-3, and (C) claudin-1 in control patients. Demonstration of cytoplasmic (D) and membranous (E) staining. (F) Tight junctions can be seen with positive staining in higher magnification (arrow).
than 4 eos/HPF. In the untreated group, 2 patients had more than 100 eos/HPF. The most common atopic manifestations included rhinitis, asthma, oral allergy syndrome, and food allergies. Only 2 patients had atopic dermatitis.
Control Biopsy Specimens The mean staining intensity in control biopsy specimens for filaggrin, ZO-3, and claudin-1 (Figure 1) were 2.6, 2.1, and 1.3, respectively. There was no significant staining noted for ZO-1 or ZO-2. IgG staining alone was negative for all controls. For filaggrin, approximately 50% to 80% of the mucosa stained with filaggrin was located on the luminal side of the mucosa. Staining was noted mostly to be cytoplasmic. For ZO-3, most of the staining occurred on the basilar side of the mucosa (Figure 1C), but in 3 biopsy specimens luminal staining could be seen additionally on the luminal side and in 1 biopsy the entire mucosa was positive for ZO-3 staining. Staining was noted to be nuclear, cytoplasmic, and/or
membranous, and in some biopsy specimens positive staining of tight junctions could be seen. For claudin-1, most of the staining was also in the basal layer, estimated to range from 5% to 50%. Staining was mostly membranous, but there was some weak cytoplasmic staining. In some biopsy specimens tight junction staining also was seen.
Eosinophilic Esophagitis Biopsy Specimens In patients with active EoE, esophageal epithelia positively stained for claudin-1, ZO-3, and filaggrin (Figure 2). ZO-1 and ZO-2 were not detected. The mean staining grade intensity in order of increasing staining intensity in untreated patients was 0.7 for Claudin-1, 1.4 for ZO-3, and 1.6 for filaggrin. All untreated patients had positive staining for filaggrin and 6 of 10 patients had grade 1. Only 2 patients had a maximal filaggrin staining intensity score of 3. In the 7 untreated patients with positive filaggrin staining, 5 patients showed predominantly cytoplasmic staining although overlap was
Figure 2. Tight junction staining in patients with (A–C) untreated and (D–F) treated EoE. (A) Grades 0 to 1 filaggrin with spongiosis. (B) Grade 1 ZO-3 staining with spongiosis. (C) Grade 1 claudin-1 with spongiosis. (D) Grade 3 filaggrin staining with predominantly cytoplasmic staining. (E) Grade 1 ZO-3, membranous pattern. (F) Grade 3 claudin-1 with a membranous pattern. Original magnification, 10.
4
Katzka et al
common. For ZO-3, only 1 patient had grade 3 staining whereas 2 patients had no detectable staining. After fluticasone treatment, the mean staining intensity was 1.3 for claudin-1, 1.7 for ZO-3, and 2.6 for filaggrin. The increase in filaggrin staining was increased significantly (P ¼ .046), with 8 of 10 post-treatment biopsy specimens now showing a maximal staining score of 3. Similar to controls, the most intense staining occurred toward the luminal side of the esophageal epithelium. There was both robust cytoplasmic and membranous staining present.
Spongiosis In biopsy specimens from the 10 untreated patients with EoE, the mean degree of spongiosis was 1.3 (Figure 3D). In patients with EoE who had received treatment with fluticasone, the mean spongiosis score was significantly less (0.4; P ¼ .016). Within individual patients, spongiosis was detected in 8 of 10 untreated patients, whereas in patients who had received fluticasone treatment, only 3 patients were found to have spongiosis on biopsy. In these treated patients, no patients with absent eosinophilia (5 patients) had spongiosis. In the 10 patients with active EoE, the 2 patients without spongiosis had the lowest levels of esophageal eosinophilia (25 and 22 eos/HPF). In the 8 patients with
Clinical Gastroenterology and Hepatology Vol.
-,
No.
-
grades 1 to 3 spongiosis, there was no clear correlation, however, between the level of esophageal eosinophilia and the grade of spongiosis (data not shown). Of the 6 patients with treated EoE and no esophageal eosinophilia, none of these patients had detectable spongiosis. Only 1 control patient showed spongiosis.
Correlation to Tight Junction Proteins The correlation of the degree of spongiosis to the staining intensity of filaggrin and ZO-3 from individual biopsy specimens was calculated (Figure 4). There was a significant inverse correlation of filaggrin to spongiosis (r ¼ -0.68; P ¼ .001). For ZO-3, there was also an inverse correlation to spongiosis (r ¼ -0.53; P ¼ .016). There was no significant correlation between the degree of spongiosis and staining of claudin-1.
Discussion Tight junction proteins are essential to all cells that interconnect cell membranes to regulate paracellular flow of ions, compounds, and cells through the intercellular channels. They exist either as transmembrane proteins (eg, claudins) or as cytoplasmic proteins that connect the cell cytoskeleton to the membrane proteins (eg, zonulin). When there is dysfunction of these
Figure 3. Comparison of staining intensity in active, inactive, and control esophageal biopsy specimens for (A) filaggrin, (B) ZO-3, and (C) claudin-1. (D) Spongiosis in patient groups.
-
2014
Tight Junctions in Eosinophilic Esophagitis
5
Figure 4. Correlation of degree of spongiosis with intensity of (A) filaggrin and (B) ZO-3 staining in biopsy specimens of active and EoE.
proteins, the range of disease can vary2 from alterations in the blood brain barrier, suppression of tumorigenesis,3 to enhanced transmembranes, as occurs with gluten in celiac disease.4,5 In this study, in biopsy specimens from patients with EoE, we show the following: (1) tight junction proteins filaggrin, claudin-1, and ZO-3, but not ZO-2 or ZO-3, are present in esophageal epithelium; and (2) topical steroids are associated with less spongiosis, a histologic sign of dilated intercellular spaces, and greater staining for tight junction proteins, which correlates inversely to spongiosis. We show that there are specific populations of tight junction proteins in esophageal epithelium with different levels of staining. In our study, the most ubiquitous protein found was filaggrin, followed by ZO-3 and claudin-1. It is known that the relative proportion of these proteins found will vary with the tissue studied. In the esophagus in 2 previous reports, there was conflicting evidence on the presence of the tight junction, occludin, in esophageal epithelium.6 On the other hand, claudin -1, similar to our study, also has been shown in normal esophageal epithelium, whereas claudin-2 has not.7 Our data are novel in testing more comprehensively for tight junction presence in esophageal epithelium. Our data also may be important in specifically showing that filaggrin is decreased in untreated patients with EoE and is restored after treatment with steroids. As importantly, the restoration of filaggrin correlates to a reduction in histologic spongiosis, that is, dilation of the intercellular spaces. These data fit well with 2 key studies that described normalization of the esophageal transcriptome with steroid therapy8 (and therefore the potential to restore synthesis of tight junction proteins), and genetic studies in children with EoE showed abnormalities of the gene that encodes for filaggrin as a potential mechanism of disease.9 The question most relevant to EoE is how do these abnormalities of tight junctions in patients with EoE promote disease? Although increased antigenic penetration of the mucosa seems likely, there also are data showing the important role of tight junctions in allowing leukocyte migration.10 Unfortunately, further work is needed in identifying the precise correlation of change in
tight junction presence to dilation of intercellular spaces in EoE. Whether this specifically occurs in EoE, particularly with dendritic cell migration and antigen recognition, is unclear. Another important question is whether the findings described here are unique to EoE or whether they are a nonspecific consequence of esophageal inflammation from a variety of causes. Specifically, in patients with gastroesophageal reflux disease, clear changes in tight junction proteins in response to acid exposure and/or injury have been described. For example, in patients with erosive or nonerosive reflux disease, an inverse correlation between claudin-1 and mucosal impedance has been reported.11 In contrast, in an epithelial cell layer derived from bronchial cells, acid or bile exposure led to a decrease in impedance but a negative correlation with claudin-4.12,13 In a rat model of reflux, abnormalities in claudins 1, 3, and 4 have been reported.14,15 Thus, one major potential difference in the mechanism of change in mucosal impedance and dilation of intercellular spaces in gastroesophageal reflux and EoE might be the tight junction pathway through which this injury occurs. This study had some limitations. First, the analysis before and after fluticasone treatment did not use the same patients. On the other hand, all patients chosen met strict guideline criteria for EoE. Furthermore, a different treated group was chosen in which there was a complete histologic response (<5 eos/HPF) to ensure that interpretation of tight junction staining was not performed and therefore equivocally interpreted in the setting of a partial response. In other words, this analysis allowed us to determine reliably the bearing of the esophageal eosinophilic component on intercellular spacing and tight junction protein regulation. Often this is not easily achieved in all treated patients. Second, is that this was a relatively small number of patients studied, although large enough to at least make some statistically significant comparisons and qualitatively evaluate for the presence of various tight junction proteins in esophageal epithelia. Third, is that we cannot determine if these results are unique to EoE or relevant to other diseases that manifest with esophageal dilated intercellular spaces, such as in gastroesophageal reflux. Nonetheless,
6
Katzka et al
we conducted this study on patients off proton pump inhibitors to attempt to isolate the effect of steroid therapy. Fourth, a limitation of this study was the initial inaccuracy of light microscopy to assess the degree of dilated intercellular spaces rather than using electron microscopy.16 On the other hand, more recent studies11 have shown that routine H&E staining can be accurate in measuring the diameter of intercellular spaces in esophageal epithelium, even in a head-to-head comparison with electron microscopy.17 This most likely is owing to increased recognition of this pattern by expert gastrointestinal pathologists. A fifth limitation of this study was that the assessment of tight junction proteins was semiquantitative because tissue was not available for measurement using quantitative polymerase chain reaction. Future studies would benefit additionally from using this assessment of barrier protein presence. Finally, this study examined only one aspect of intercellular space regulation, that is, tight junctions. Recent literature also has reported desmosomal dysfunction in patients with EoE.18 In conclusion, our data show the existence and relative density of specific tight junction proteins in human esophageal epithelium. We further show that the staining density of these proteins in patients with EoE may be altered by the administration of steroid therapy. Finally, the density of 2 of these proteins, filaggrin and ZO-3, correlates inversely to spongiosis, that is, dilated intercellular spaces, when the latter is corrected with steroid therapy.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2014.02.039.
References 1. Liacouras CA, Furuta GT, Hirano I, et al. Eosinophilic esophagitis: updated consensus recommendations for children and adults. J Allergy Clin Immunol 2011;128:3–20. 2. Odenwald MA, Turner JR. Intestinal permeability defects: is it time to treat. Clin Gastroenterol Hepatol 2013;111:1075–1083. 3. Coisne C, Engelhardt B. Tight junctions in brain barriers during central nervous system inflammation. Antioxidants Redox Signaling 2011;15:285–303. 4. Fasano A, Not T, Wang W, et al. Zonulin, a newly discovered modulator of intestinal permeability, and its expression in celiac disease. Lancet 2000;355:1518–1519. 5. Drago S, El Asmar R, Di Pierro M, et al. Gliadin, zonulin and gut permeability: effects on celiac and non-celiac intestinal mucosa and intestinal cell lines. Scand J Gastroenterol 2006;41: 408–419.
Clinical Gastroenterology and Hepatology Vol.
-,
No.
-
6. Kimura Y, Shiozaki H, Hirao M, et al. Expression of occludin tight-junction-associated protein, in human digestive tract. Am J Pathol 1997;151:45–54. 7. Rendon-Huerta E, Valenzano MC, Mullin JM, et al. Comparison of three integral tight junction barrier proteins in Barrett’s epithelium versus normal esophageal epithelium. Am J Gastroenterol 2003;98:1901–1902. 8. Blanchard C, Mingler MK, Vicario M, et al. IL-13 involvement in eosinophilic esophagitis: transcriptome analysis and reversibility with glucocorticoids. J Allergy Clin Immunol 2007;120: 1292–1300. 9. Blanchard C, Wang N, Stringer KF, et al. Eotaxin-3 and a uniquely conserved gene-expression profile in eosinophilic esophagitis. J Clin Invest 2006;116:536–547. 10. Steed E, Balda MS, Matter K. Dynamics and functions of tight junctions. Trends Cell Biol 2009;20:141–149. 11. Zhong C, Duan L, Wang K, et al. Esophageal intraluminal baseline impedance is associated with severity of acid reflux and epithelial structural abnormalities in patients with gastroesophageal reflux disease. J Gastroenterol 2013;48: 601–610. 12. Oshima T, Koseki J, Chen X, et al. Acid modulates the squamous epithelial barrier function by modulating the localization of claudins in the superficial layers. Lab Invest 2012;92:22–31. 13. Chen X, Oshima T, Shan J, et al. Bile salts disrupt human esophageal squamous epithelial barrier function by modulating tight junction proteins. Am J Physiol Gastrointest Liver Physiol 2012;303:G199–G208. 14. Asaoka D, Miwa H, Hirai S, et al. Altered localization and expression of tight-junction proteins in a rat model with chronic acid reflux esophagitis. J Gastroenterol 2005;40:781–790. 15. Oguro M, Koike M, Ueno T, et al. Dissociation and dispersion of claudin-3 from the tight junction could be one of the most sensitive indicators of reflux esophagitis in a rat model of the disease. J Gastroenterol 2011;46:629–638. 16. Tobey NA, Carson JL, Alkiek RA, et al. Dilated intercellular spaces: a morphological feature of acid reflux–damaged human esophageal epithelium. Gastroenterology 1996;111: 1200–1205. 17. Ribolsi M, Perrone G, Caviglia R, et al. Intercellular space diameters of the oesophageal epithelium in NERD patients: head to head comparison between light and electron microscopy analysis. Dig Liver Dis 2009;41:9–14. 18. Sherrill JD, Kc K, Wu D, et al. Desmoglein-1 regulates esophageal epithelial barrier function and immune responses in eosinophilic esophagitis. Mucosal Immunol 2013 Nov 13. Epub ahead of print. 19. Hirano I, Moy N, Heckman MG, et al. Endoscopic assessment of the oesophageal features of eosinophilic esophagitis: validation of a novel classification and grading system. Gut 2013;62: 489–495. Reprint requests Address requests for reprints to: David A. Katzka, MD, Mayo Clinic, 200 First Avenue SW, Rochester, Minnesota 55905. e-mail: [email protected]; fax: (507) 538-5820. Conflicts of interest The authors disclose no conflicts.
-
2014
Tight Junctions in Eosinophilic Esophagitis
Supplementary Table 1. Tight Junction Antibody Dilutions S no. Antibody Manufacturer Catalogue no. Dilution used 1 2 3 4
Filaggrin Claudin-1 ZO-3 ZO-1
S, sample.
Abcam Abcam Abcam Abcam
ab ab ab ab
81468 15098 64326 59720
1:250 1:100 1:100 1:100
6.e1