Mucosal Penetration and Clearance of Gluten and Milk Antigens in Eosinophilic Oesophagitis

Anupama Ravi; Eric V. Marietta; Jeffrey A. Alexander; Kathryn Peterson; Crystal Lavey; Debra M. Geno; Joseph A. Murray; David A. Katzka


Aliment Pharmacol Ther. 2021;53(3):410-417. 

In This Article


Adult patients with EoE on a diet containing wheat and dairy (except where noted on diet elimination therapy—please see Table S1) undergoing clinically indicated standard upper endoscopy and biopsy participated in this study. Patients were being clinically assessed for histologic response to a 2-month course of proton pump inhibitor or topical steroid or for symptomatic recurrence. It was not known prior to endoscopy whether patients would have active or inactive disease. Wheat and dairy were chosen as they contain the two most common antigens that trigger EoE and not for their specific role in disease activity in individual patients. Dairy was also chosen as it allowed the study of two antigens with one food exposure (casein and whey) with different molecular characteristics. Three groups of patients were studied: (a) Patients with active EoE defined by consensus criteria with >15 eos/HPF. (b) Patients with inactive EoE defined as inactive by consensus criteria with ≤15 eos/HPF. (c) Control patients without oesophageal symptoms including dysphagia, heartburn or chest pain and with a normal appearing oesophagus on clinically indicated endoscopy and normal histology on biopsy. EoE patients were asked to eliminate wheat and/or dairy from their diet at varying intervals, 12, 24, 48, 72 or 96 hours, prior to their scheduled upper endoscopy. These time intervals were randomly chosen as there is no prior data in this area to guide experimental design. Repeat clinically indicated endoscopy was performed to assess histologic response to pharmacologic therapy. No patients were being treated with food elimination diet as a primary therapy. Patients received careful instruction on diet elimination from an EoE research coordinator and nurse experienced in diet elimination therapy instruction. This consisted of a 30-minute session of instruction during which time handouts are given on foods that contain wheat and dairy, cookbooks and specific recipes excluding dairy and wheat, a list of cookbooks and of companies that manufacture wheat and dairy free products and a list of multiple websites of organisations, restaurants and magazines that provide information on these diets. Non-EoE/control patients did not have dietary modifications but for one patient avoiding wheat. The patients used in this study were novel patients without participation in prior antigen exposure studies Table S2.[6]

Ethics Approval

The study was approved by Mayo Clinic Institutional Review Board and patients signed informed consent to have three additional oesophageal biopsies performed for research purposes and restrict their diet as discussed. All authors had access to the study data and reviewed and approved the final manuscript.

Collections of Biopsies and Histologic Analysis

During oesophagoscopy with clinical biopsies, an additional three biopsies were taken 10 cm above the gastro-oesophageal junction in all patients. Routine biopsy specimens were stained with haematoxylin and eosin and read by Mayo GI pathologists with expertise in EoE. The area of greatest eosinophil density was first located by low powered review of all tissue submitted from a given location. The area of greatest eosinophil density on oesophageal biopsies was used for analysis in each individual patient regardless of oesophageal location. Eosinophils were then counted using a 40× objective, a field diameter of 0.625 mm and a field area of 0.307 mm2. The peak eosinophil count/high power field (eos/HPF) was reported. From the area of greatest eosinophil density under low powered review, five random fields were chosen. Peak eosinophil counts from these five fields were then used to calculate a mean eosinophil count.

Measurement of Gliadin, Casein and Whey in Oesophageal Mucosa

Oesophageal biopsies taken during endoscopy were immediately placed into Optimal Cutting Temperature compound (VSR cat# 4692300) and snap frozen on dry ice. Cryosectioning of the frozen oesophageal tissue was done for immunofluorescent staining to determine the distribution/localisation of different cells and dietary proteins. For gliadin, anti-gliadin (Biorbyt catalog # orb157160) was used for immunofluorescence. Immunogen was KLH-conjugated synthetic peptide derived from aa241–286 of wheat gliadin. For casein, anti-casein antibody (Abcam catalog # ab91167) was utilised. Alpha-casein was detected. Immunogen was KLH-conjugated synthetic peptide selected from the central region of human casein. For whey, anti-whey antibody (Abcam catalog# ab112972) was utilised. Alpha lactalbumin was detected. Immunogen was whole protein bovine alpha lactalbumin. Additionally, IgG4(Abcam), and AF594 antirabbit (Jackson Immunoresearch Laboratories) were utilised for immunofluorescence (Figure 2). Intensity of staining was performed with a 0–3 grading system (Figure 2) of multiple biopsy sections by three investigators (AR, EVM, DAK) blind to the group studied. Nucleic acid was stained with DAPI (4',6-Diamidino-2-phenylindole; Sigma Aldrich). At least two high power fields were evaluated for each stain per tissue specimen. We did not process all specimens concurrently. Casein and whey staining were largely performed in a few batches. Gliadin staining was performed in an ongoing basis to further determine the interval of gluten dietary elimination.

Statistical Analysis

To further characterise the active EoE, inactive EoE and non-EoE control patients, summary statistics were utilised. A non-parametric Mann–Whitney test was conducted to compare active EoE, inactive EoE, and non-EoE/control patient for each of the following: presence or absence of antigen staining on oesophageal biopsies (gliadin, casein, whey), presence after selected duration of antigen elimination prior to upper endoscopy (12, 24, 48, 72 or 96 hours), and presence of IgG4 staining.

All analyses were performed using GraphPad Prism software. A P < 0.05 was considered statistically significant.