Proton Pump Inhibitors Decrease Eotaxin-3 Expression in the Proximal Esophagus of Children with Esophageal Eosinophilia

PLOS ONE, Jul 2014

Objective Besides reducing gastric acid secretion, proton pump inhibitors (PPIs) suppress Th2-cytokine-stimulated expression of an eosinophil chemoattractant (eotaxin-3) by esophageal epithelial cells through acid-independent, anti-inflammatory mechanisms. To explore acid-inhibitory and acid-independent, anti-inflammatory PPI effects in reducing esophageal eosinophilia, we studied eotaxin-3 expression by the proximal and distal esophagus of children with esophageal eosinophilia before and after PPI therapy. In vitro, we studied acid and bile salt effects on IL-13-stimulated eotaxin-3 expression by esophageal epithelial cells. Design Among 264 children with esophageal eosinophilia seen at a tertiary pediatric hospital from 2008 through 2012, we identified 10 with esophageal biopsies before and after PPI treatment alone. We correlated epithelial cell eotaxin-3 immunostaining with eosinophil numbers in those biopsies. In vitro, we measured eotaxin-3 protein secretion by esophageal squamous cells stimulated with IL-13 and exposed to acid and/or bile salt media, with or without omeprazole. Results There was strong correlation between peak eosinophil numbers and peak eotaxin-3-positive epithelial cell numbers in esophageal biopsies. Eotaxin-3 expression decreased significantly with PPIs only in the proximal esophagus. In esophageal cells, exposure to acid-bile salt medium significantly suppressed IL-13-induced eotaxin-3 secretion; omeprazole added to the acid-bile salt medium further suppressed that eotaxin-3 secretion, but not as profoundly as at pH-neutral conditions. Conclusion In children with esophageal eosinophilia, PPIs significantly decrease eotaxin-3 expression in the proximal but not the distal esophagus. In esophageal squamous cells, acid and bile salts decrease Th2 cytokine-stimulated eotaxin-3 secretion profoundly, possibly explaining the disparate PPI effects on the proximal and distal esophagus. In the distal esophagus, where acid reflux is greatest, a PPI-induced reduction in acid reflux (an effect that could increase eotaxin-3 secretion induced by Th2 cytokines) might mask the acid-independent, anti-inflammatory PPI effect of decreasing cytokine-stimulated eotaxin-3 secretion.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0101391&type=printable

Proton Pump Inhibitors Decrease Eotaxin-3 Expression in the Proximal Esophagus of Children with Esophageal Eosinophilia

et al. (2014) Proton Pump Inhibitors Decrease Eotaxin-3 Expression in the Proximal Esophagus of Children with Esophageal Eosinophilia. PLoS ONE 9(7): e101391. doi:10.1371/journal.pone.0101391 Proton Pump Inhibitors Decrease Eotaxin-3 Expression in the Proximal Esophagus of Children with Esophageal Eosinophilia Jason Y. Park 0 Xi Zhang 0 Nathalie Nguyen 0 Rhonda F. Souza 0 Stuart J. Spechler 0 Edaire Cheng 0 Claudia Daniela Andl, Vanderbilt University, United States of America 0 1 Esophageal Diseases Center, Department of Pathology, Children's Medical Center, Eugene McDermott Center for Human Growth and Development, and the University of Texas Southwestern Medical Center , Dallas, Texas , United States of America, 2 Esophageal Diseases Center, Department of Internal Medicine, VA North Texas Health Care System, and the University of Texas Southwestern Medical Center , Dallas, Texas , United States of America, 3 Department of Pediatrics, Children's Medical Center, and the University of Texas Southwestern Medical Center , Dallas, Texas , United States of America, 4 Esophageal Diseases Center, Department of Internal Medicine, VA North Texas Health Care System, Harold C. Simmons Comprehensive Cancer Center, and the University of Texas Southwestern Medical Center , Dallas, Texas , United States of America Objective: Besides reducing gastric acid secretion, proton pump inhibitors (PPIs) suppress Th2-cytokine-stimulated expression of an eosinophil chemoattractant (eotaxin-3) by esophageal epithelial cells through acid-independent, antiinflammatory mechanisms. To explore acid-inhibitory and acid-independent, anti-inflammatory PPI effects in reducing esophageal eosinophilia, we studied eotaxin-3 expression by the proximal and distal esophagus of children with esophageal eosinophilia before and after PPI therapy. In vitro, we studied acid and bile salt effects on IL-13-stimulated eotaxin-3 expression by esophageal epithelial cells. Design: Among 264 children with esophageal eosinophilia seen at a tertiary pediatric hospital from 2008 through 2012, we identified 10 with esophageal biopsies before and after PPI treatment alone. We correlated epithelial cell eotaxin-3 immunostaining with eosinophil numbers in those biopsies. In vitro, we measured eotaxin-3 protein secretion by esophageal squamous cells stimulated with IL-13 and exposed to acid and/or bile salt media, with or without omeprazole. Results: There was strong correlation between peak eosinophil numbers and peak eotaxin-3-positive epithelial cell numbers in esophageal biopsies. Eotaxin-3 expression decreased significantly with PPIs only in the proximal esophagus. In esophageal cells, exposure to acid-bile salt medium significantly suppressed IL-13-induced eotaxin-3 secretion; omeprazole added to the acid-bile salt medium further suppressed that eotaxin-3 secretion, but not as profoundly as at pH-neutral conditions. Conclusion: In children with esophageal eosinophilia, PPIs significantly decrease eotaxin-3 expression in the proximal but not the distal esophagus. In esophageal squamous cells, acid and bile salts decrease Th2 cytokine-stimulated eotaxin-3 secretion profoundly, possibly explaining the disparate PPI effects on the proximal and distal esophagus. In the distal esophagus, where acid reflux is greatest, a PPI-induced reduction in acid reflux (an effect that could increase eotaxin-3 secretion induced by Th2 cytokines) might mask the acid-independent, anti-inflammatory PPI effect of decreasing cytokinestimulated eotaxin-3 secretion. - Funding: This work was supported by the Office of Medical Research, Departments of Veterans Affairs (RFS, SJS), the National Institutes of Health (R01-DK63621 to RFS and SJS, R01-CA134571 to RFS and SJS, K12-HD068369 to EC, K08-DK099383 to EC), NASPGHAN Foundation/AstraZeneca Award (EC), AGA Research Scholar Award (EC), and AAP Resident Research Grant (NN). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. In eosinophilic esophagitis (EoE), food allergens trigger a Thelper 2 (Th2) immune response with production of Th2 cytokines such as interleukin (IL)-13 and IL-4 [1,2]. These cytokines can stimulate the esophagus to express eotaxin-3, a potent eosinophil chemoattractant thought to play a key role in causing esophageal eosinophilia in EoE [35]. Esophageal eosinophilia underlies the esophageal dysfunction and tissue remodeling responsible for dysphagia and other symptoms that can seriously impair quality of life for EoE patients [68]. Consequently, a major goal of EoE treatment is to reduce esophageal eosinophil numbers [6]. Esophageal eosinophilia also can be a manifestation of gastroesophageal reflux disease (GERD), which can cause symptoms and endoscopic abnormalities similar to those of EoE [9]. To distinguish these two disorders in patients with esophageal eosinophilia, authorities have recommended a trial of proton pump inhibitor (PPI) therapy with the assumption that gastric acid inhibition is the only important effect of PPIs and, therefore, only an acid-peptic disorder like GERD can respond to PPIs. However, a number of recent observations have challenged this assumption. PPIs have been found to have anti-inflammatory actions unrelated to their inhibitory effects on gastric acid secretion [10]. For example, PPIs inhibit cytokine production by human endothelial and tracheal epithelial cells [11,12]. In esophageal epithelial cells in culture, we have reported that PPIs block the secretion of IL-8 and the secretion of eotaxin-3 stimulated by Th2 cytokines [5,13,14]. These acid-independent, anti-inflammatory effects of PPIs conceivably could contribute to resolution of esophageal eosinophilia in both GERD and EoE. Another serious challenge to the assumption that PPI responsiveness distinguishes GERD from EoE is the recent identification of patients with PPI-responsive esophageal eosinophilia (PPIREE). These patients have typical EoE symptoms and esophageal eosinophilia, both of which improve with PPIs even though they have no evidence of GERD by endoscopy or esophageal pH monitoring. It is not clear whether these patients respond to the acid-inhibitory effects of PPIs because they have occult GERD not detected by endoscopy and pH monitoring, or whether they respond to acid-independent, anti-inflammatory effects of PPIs because they have an immune/antigen-mediated esophageal disease (EoE or some EoE-like disorder). The purpose of this study was to explore the contributions of acid-inhibitory and acid-independent, anti-inflammatory effects of PPIs on esophageal eosinophilia in children. We reasoned that, if PPIs reduce esophageal eosinophilia by reducing acid reflux, then those acid-inhibitory effects should manifest most prominently in the distal esophagus where acid reflux exposure is greatest. On the other hand, if PPIs reduce esophageal eosinophilia through acidindependent, anti-inflammatory effects on eotaxin-3, then those effects should manifest more equally throughout the esophagus. Therefore, we studied eotaxin-3 expression by esophageal epithelial cells in biopsy specimens of the proximal, mid, and distal esophagus of children with esophageal eosinophilia before and after PPI therapy, and correlated those findings with the number of intraepithelial eosinophils in those same biopsy specimens. In vitro, we studied acid and bile salt effects on Th2cytokine-stimulated eotaxin-3 expression by esophageal epithelial cells, with or without omeprazole treatment. Materials and Methods Ethics Statement This study was approved by the University of Texas Southwestern Medical Center Institutional Review Board (STU 032013037). Patient clinical medical records and patient archived specimens used in this study were anonymized and de-identified prior to analysis, and the need for informed consent was waived. Subjects This study was approved by the University of Texas Southwestern Medical Center Institutional Review Board (STU 032013037). We reviewed the electronic medical records of patients evaluated in the Pediatric Gastroenterology, Hepatology, and Nutrition program at Childrens Medical Center between 2008 and 2012 to identify children who had an endoscopy with esophageal biopsies showing $15 eosinophils per high power field (eos/hpf) and who had follow-up endoscopy with esophageal biopsies after treatment with PPIs alone for $8 weeks. Patients were excluded if, prior to the index endoscopy showing esophageal eosinophilia, they had received treatment with PPIs, swallowed topical steroids, systemic steroids, elimination diet, or elemental diet. Patients also were excluded if they received any concurrent treatment for EoE in addition to PPIs between the index and follow-up endoscopy. We recorded pertinent clinical data including demographics, presenting symptoms, coexisting atopic disease, and endoscopic findings. Cases were classified as PPI responders if the post-treatment biopsies had a peak eosinophil count ,15 eos/ hpf, and as PPI non-responders if the post-treatment biopsies had a peak eosinophil count $15 eos/hpf. Tissue Specimens Archived, formalin-fixed, paraffin-embedded (FFPE) tissue blocks of patient biopsy specimens were obtained from the Department of Pathology at Childrens Medical Center. Histopathology Hematoxylin and eosin (H&E) stained slides were reviewed. For biopsy specimens taken at each level of the esophagus (proximal, mid, and distal), peak eosinophil counts in the squamous epithelium were determined by counting the number of eosinophils in the high power field with the greatest density of eosinophils. A BX41 microscope with an UPlanFL N 40x/0.75 objective lens with a FN22 eyepiece was used (Olympus America, Center Valley, PA); the calculated area per hpf is 0.237 mm2. Histopathological findings such as basal cell hyperplasia (.25% of thickness of epithelium), spongiosis, elongation of rete papillae (. two-thirds the thickness of the epithelium), and eosinophilic microabscess ($3 eosinophils clustered at the luminal surface or within a maximum distance of one squamous cell from the luminal surface) were recorded. Subepithelial fibrosis was evaluated in biopsy specimens with sufficient lamina propria for meaningful analysis. Immunohistochemistry Eotaxin-3 polyclonal goat anti-human IgG (Human CCL26/ Eotaxin-3 Affinity Purified Polyclonal Ab, Catalog Number AF653, R&D systems, Minneapolis, MN) was used for immunohistochemistry. This eotaxin-3 polyclonal antibody was raised against a peptide from amino acid residues 24 (Threonine) through 94 (Leucine) of eotaxin-3. FFPE tissues were sectioned at 4 mm and mounted onto charged microscope slides. The slides were baked at 60uC for 30 minutes and then stained on a Discovery XT automated immunohistochemistry platform (Ventana Medical Systems, Tucson, AZ). The eotaxin-3 antibody was used at a final concentration of 2 mg/ml. The secondary antibody was UltraMap anti-goat horseradish peroxidase used at a prediluted concentration as a chromogenic reporter molecule (Catalog Number 760-4648, Ventana). Negative controls included the replacement of eotaxin-3 antibody with either non-specific goat IgG or saline. Quantification of Eotaxin-3 Immunohistochemistry Immunostaining for eotaxin-3 was evaluated by a gastrointestinal pathologist (JYP). The peak number of eotaxin-3-positive squamous epithelial cells per hpf (400x) was counted in the area of greatest staining intensity. An eotaxin-3 positive squamous cell was identified by the presence of granular, perinuclear immunostaining. Eotaxin-3-positive eosinophils (identified by their distinctive morphological features) were not counted. Culture of Esophageal Squamous Cells We used two non-neoplastic, telomerase-immortalized, esophageal squamous cell lines (EoE1-T and EoE2-T) that were created from esophageal mucosal biopsy specimens from patients with EoE as previously described by our laboratory [5]. Briefly, the patients fulfilled the criteria for EoE suggested in the 2007 consensus recommendations [15]. They both had a history of dysphagia and heartburn that had responded only partially or not at all to PPIs, and had esophageal biopsy specimens showing $15 eos/hpf; symptoms subsequently improved dramatically with fluticasone treatment. We also used two previously characterized, telomerase-immortalized, non-neoplastic esophageal squamous cell lines established from patients with GERD [16]. One line (NES-B10T) was created from endoscopic biopsy specimens of squamous epithelium in the esophagus of a patient who had GERD associated with long-segment Barretts esophagus (.3 cm of specialized intestinal metaplasia). The other line (NES-G4T) was established from esophageal biopsy specimens from a GERD patient who had Los Angeles grade C reflux esophagitis without Barretts esophagus. Cells were maintained in monolayer culture at 37uC in humidified air with 5% CO2 in growth medium co-cultured with a fibroblast feeder layer as previously described [16]. For individual experiments, cells were equally seeded into collagen IV-coated wells (BD Biosciences, San Jose, CA) and maintained in growth medium. Cytokine Stimulation and Omeprazole Treatment of Esophageal Squamous Cells Cells were stimulated with IL-13 (R&D Systems) in concentrations of 1 or 100 ng/ml for 48 hours. For studies involving PPI effects, omeprazole (Sigma-Aldrich, St. Louis, MO) 5 mM was acid-activated in medium with pH 5.5 for 30 minutes [5,13,17]. Cells were then pre-treated for 2 hours with omeprazole in medium with pH 7.2 prior to the addition of IL-13. Omeprazole remained in the media throughout the period of cytokine stimulation. Acid and Bile Salt Exposure of Esophageal Squamous Cells Cells were cultured either in neutral medium (pH 7.2) alone or in neutral medium with periodic exposures to acidic media (pH levels ranging from 4.0 to 7.2) with and without bile salts. Stable cell number and viability were verified by cell counting for each of these experimental conditions. Cells exposed to pH 3.0 were not viable. The bile salt media contained a mixture of conjugated bile acids (glycocholic acid, glycodeoxycholic acid [both from Calbiochem, San Diego, CA], taurocholic acid, glycochenodeoxycholic acid, taurochenodeoxycholic acid, and taurodeoxycholic acid [Sigma-Aldrich] in a 20:6:3:15:3:1 molar concentration, total concentration 50 mM) designed to simulate the bile acid composition of gastric refluxate described in patients with GERD and Barretts esophagus [1820]. Cells were exposed to acid and/or bile salt media for 10 minutes, three times a day, for 2 days. These exposure durations were chosen to simulate typical episodes of gastroesophageal reflux [20]. For experiments evaluating IL-13 and omeprazole effects, the media that was replaced (after the 10minute acid and bile salt media exposure) contained IL-13 with or without omeprazole. Enzyme-Linked Immunosorbent Assays (ELISA) for Eotaxin-3 We performed ELISA on conditioned media after 48 hours to assess the production of eotaxin-3 by esophageal epithelial cells. Conditioned media from the cells were collected and centrifuged to remove cellular debris. Eotaxin-3 concentrations were determined using commercially-available ELISA kits (R&D Systems) per manufacturers instructions. The absorbance of each well was read at 450 nm and 540 nm using a DTX 880 Multimode plate reader (Beckman Coulter). Results were expressed as pg/ml of eotaxin-3. All assays were performed in duplicate. Statistical analysis Categorical data are expressed as frequency and/or percentage, and bivariate analysis was performed with Fischers exact test. Continuous data are expressed as mean 6 standard error of the mean (SEM), and bivariate analysis was evaluated with t-tests. Paired t-tests were used to evaluate before and after therapy results. Correlations between eosinophil counts and eotaxin-3positive-staining epithelial cell counts were calculated by Pearson correlation coefficient r (R). For in vitro studies, data are expressed as mean 6 SEM. Multivariate analysis was performed with oneway ANOVA. Statistical significance was determined by P value # 0.05. Statistical analyses were performed with GraphPad Prism 6 (GraphPad Software, Inc, La Jolla, CA). Baseline Patient Characteristic and Clinicopathological Features We identified 264 patients with esophageal eosinophilia (Figure 1). Forty patients who had been treated with PPIs had pre- and post-treatment endoscopic examinations; 30 of those 40 were excluded because they had received concurrent treatment in addition to PPIs, leaving 10 study subjects for evaluation. Histological review of the post-treatment biopsy specimens identified 5 PPI responders (,15 eos/hpf) and 5 PPI nonresponders ($15 eos/hpf) (Table 1). There were no significant differences in baseline characteristics between the groups except for the symptoms of weight loss/poor weight gain and vomiting, which were more frequent in the PPI responders, and baseline peak eosinophil counts, which were lower in the PPI responders. For the total 10 study patients, the mean age was 11 years (range 416 years), 8 were male, and all were Caucasian. Five patients had a history of atopy (2 allergic rhinitis, 1 eczema, 1 asthma alone, 1 asthma and food allergies). Presenting symptoms included dysphagia (50%), weight loss/poor weight gain (40%), vomiting (40%), abdominal pain (30%), oral aversion (10%), and PPI Responders (N = 5) PPI Non-Responders (N = 5) Mean age (y 6 SEM) Weight loss/poor weight gain PPI dose (mg/kg/d 6 SEM) PPI duration (months 6 SEM) Peak eosinophils (eos/hpf 6 SEM) *Comparison between PPI responders and PPI non-responders using t-tests or Fischers exact test. #4 of 5 cases had sufficient lamina propria to evaluate for fibrosis. doi:10.1371/journal.pone.0101391.t001 food impaction (10%). The index endoscopy demonstrated furrows (80%), rings (30%), strictures (10%), white plaques (50%), and edema (50%). Histopathological features included basal cell hyperplasia (100%), spongiosis (100%), papillary elongation (100%), and eosinophilic microabscesses (90%). Nine patients had subepithelial lamina propria in their biopsy specimens, and fibrosis was seen in all 9; one patient did not have sufficient subepithelial lamina propria for evaluation of subepithelial fibrosis, but fibrosis was seen in the lamina propria present in the rete papillae. Pre- and Post-PPI Treatment Peak Eosinophil Counts and Histological Features The mean of the highest peak eosinophil count recorded at any level in the esophagus (proximal, mid or distal) for the total 10 patients did not decrease significantly with PPI treatment (Figure 2A). However, the mean of the highest peak eosinophil count fell significantly (from 33 to 7 eos/hpf, P = 0.049) (Figure 2B) in the PPI responders, and did not change significantly in the PPI non-responders (Figure 2C). Pre- and post-treatment biopsies from all three levels of the esophagus were available for most, but not all of the 10 patients. After PPI treatment, the mean peak eosinophil count decreased, but not significantly, in the proximal esophagus (from 40 to 16 eos/hpf, P = 0.28), changed little in the mid esophagus (30 vs. 27 eos/hpf, P = 0.80) and increased, but not significantly, in the distal esophagus (from 26 to 42 eos/hpf, P = 0.41) (Figure S1). Histological findings of basal cell hyperplasia, spongiosis, and papillary elongation were seen in all 10 cases prior to PPI treatment. These three features resolved in 2 of the 5 PPI responders, but in none of the non-responders (Table S1). Nine patients (4 PPI responders, 5 PPI non-responders) had eosinophilic microabscesses at baseline, and PPI treatment resulted in complete resolution of microabscesses in all of the 4 PPI responders and in 2 of the 5 PPI non-responders. Pre- and Post-PPI Treatment Eotaxin-3-Positive Esophageal Epithelial Cells We found variable intensities of eotaxin-3 immunostaining of squamous epithelial cells within and between biopsy specimens from patients with esophageal eosinophilia (Figure 3B); no eotaxin3 labeling was noted in squamous epithelial cells from a normal control esophagus (Figure 3A). High power magnification (1000x) revealed a granular, perinuclear pattern of cytoplasmic staining in the squamous cells (Figure 3C). Eosinophils also demonstrated immunostaining by eotaxin-3, but eosinophils were readily Figure 2. Highest peak eosinophil count at any esophageal level. Highest pre- and post-PPI treatment peak eosinophil counts in (A) all PPItreated cases, (B) PPI responders, and (C) PPI non-responders. Red bars represent means. doi:10.1371/journal.pone.0101391.g002 identified by their distinctive morphology and were excluded from quantification (Figure 3D). The mean of the highest peak number of eotaxin-3-positive squamous epithelial cells at any esophageal level (proximal, mid or distal) for the total 10 patients decreased significantly with PPI treatment (from 35 to 20 cells/hpf, P = 0.0106) (Figure 4A). The mean of the highest peak number of eotaxin-3-positive cells fell significantly in the PPI responders (from 26 to 7 cells/hpf, P = 0.0053) (Figure 4B), and decreased, but not significantly, in the PPI non-responders (Figure 4C). There was a strong correlation between the highest peak eosinophil count and the highest peak number of eotaxin-3-positive epithelial cells (R = 0.8332, P, 0.0001) (Figure 4D). Pre- and post-treatment biopsies from all three levels of the esophagus were available for most, but not all of the 10 patients. Interestingly, after PPI treatment, the mean peak number of eotaxin-3-positive epithelial cells decreased significantly in the proximal esophagus (from 28 to 6 cells/hpf, P = 0.049) (Figure 5A), decreased minimally in the mid esophagus (from 13 to10 cells/hpf, P = 0.72) (figure 5B), and increased, but not significantly, in the distal esophagus (from 14 to 24 cells/hpf, P = 0.41) (Figure 5C). In the proximal esophagus, the peak number of eotaxin-3-positive epithelial cells correlated strongly with the peak eosinophil count, both before and after treatment (Figure 5D). In the mid and distal esophagus, there were strong correlations between the peak number of eotaxin-3-positive epithelial cells and peak eosinophil count after treatment, but not before treatment (Figure 5E-F). Acid and Bile Salt Effects and PPI Anti-Inflammatory Effects on IL-13-Induced Eotaxin-3 Protein Secretion by Esophageal Epithelial Cells Although PPI effects in reducing acid reflux are strongest in the distal esophagus, we found that PPIs decreased eotaxin-3 expression significantly only in the proximal esophagus. This suggested that acidic refluxate might influence eotaxin-3 expression by esophageal epithelial cells. To explore this possibility, we studied the effects of simulated gastroesophageal refluxate solutions (containing acid and/or bile salts) on eotaxin-3 secretion by esophageal epithelial cells with and without IL-13 stimulation. Without IL-13 stimulation, we found little basal secretion of eotaxin-3 by esophageal epithelial cells (Figures 6A, 7A, 8A). In EoE1-T cells, this low basal level of eotaxin-3 protein secretion was not affected significantly by exposure to acid and/or bile salts (Figure S2). Next, we explored the effects of acid and bile salts on IL-13-stimulated eotaxin-3 secretion by EoE1-T cells. As seen in our earlier studies, IL-13 (100 ng/ml) added to neutral control medium induced a significant and profound increase in eotaxin-3 protein secretion (Figure 6A, white bar) [5,13]. This IL-13 stimulated increase in eotaxin-3 secretion was attenuated significantly by exposure to acid alone (pH levels of 4.0 to 6.0), bile salts alone, or the combination of acid and bile salts) (Figure 6A, black and gray bars). We found no significant differences between acid alone, bile salts alone, and the combination (at each pH level studied) in suppressing IL-13-stimulated eotaxin-3 secretion. Therefore, we selected to use the combination of acid and bile salts, a more physiological representation of typical gastroesophageal reflux episodes, for subsequent experiments. In earlier studies, we showed that PPIs block Th2-cytokinestimulated eotaxin-3 secretion through mechanisms independent of effects on gastric acid secretion (i.e. acid-independent, Figure 4. Highest peak number of eotaxin-3-positive epithelial cells at any esophageal level and correlation with highest peak eosinophil count. Highest pre- and post-PPI treatment peak number of eotaxin-3-positive epithelial cells in (A) all PPI-treated cases, (B) PPI responders, and (C) PPI non-responders. Red bars represent means. (D) Correlation between highest peak eosinophil count and highest peak number of eotaxin-3-positive epithelial cells. doi:10.1371/journal.pone.0101391.g004 anti-inflammatory PPI effects) [5,13]. Now, we studied how acid and bile salts affect omeprazoles ability to block IL-13stimulated eotaxin-3 secretion by EoE1-T cells. As in our earlier studies, omeprazole significantly suppressed IL-13stimulated eotaxin-3 secretion at neutral conditions without bile salts (Figure 6B) [5,13]. As in the experiments shown in Figure 6A, exposure to acid-bile salt medium also significantly suppressed IL-13-stimulated eotaxin-3 secretion (Figure 6B). Omeprazole further suppressed eotaxin-3 secretion at each acid-bile salt exposure condition, but not as profoundly as at neutral conditions without bile salts (Figures 6B and C). We repeated these experiments in EoE2-T cells. In response to IL-13 in a dose of 100 ng/ml, however, the amount of eotaxin-3 protein secreted by EoE2-T cells is almost two orders of magnitude greater than that secreted by EoE1-T cells (Figure 7A). Unlike EoE1-T cells, acid and bile salts did not significantly affect eotaxin3 secretion by EoE2-T cells stimulated with100 ng/ml of IL-13 (Figure 7A). Also unlike EoE1-T cells, omeprazole suppressed that IL-13-stimulated eotaxin-3 secretion in EoE2-T cells exposed to acid and bile salts to a level comparable to that achieved at neutral control conditions (Figure 7B). In separate experiments, we confirmed that neither acid alone, bile salts alone, nor the combination suppressed the robust secretion of eotaxin-3 stimulated by IL-13 in a dose of 100 ng/ml (Figure S3). Therefore, we repeated our experiments using a lower dose of IL-13 (1 ng/ml) at which EoE2-T cells secrete eotaxin-3 in amounts comparable to EoE1-T cells. With this lower dose of IL-13, results in EoE2-T cells were similar to those in EoE1-T cells. Specifically, all acid-bile salt exposure conditions significantly suppressed IL-13-induced eotaxin-3 secretion (Figure 7C), and the magnitude of omeprazoleinduced suppression in cells exposed to acid and bile salts was less than that at neutral conditions without bile salts (Figure 7D). Finally, to determine whether these findings are specific to esophageal cells from patients with EoE, we examined acid-bile salt effects on IL-13-stimulated eotaxin-3 secretion in cell lines derived from patients with GERD (NES-G4T and NES-B10T). We used IL-13 in a dose of 100 ng/ml, since NES-G4T and NESB10T have previously demonstrated lower ranges of Th2 cytokinestimulated eotaxin-3 secretion [5]. In both GERD cell lines, acidbile salt exposures significantly suppressed IL-13-induced-eotaxin3 secretion (Figure 8A&C). The addition of omeprazole further suppressed eotaxin-3 secretion but, as in the EoE cell lines, the magnitude of that suppression was considerably less than that at neutral conditions without bile salts (Figure 8B&D). In esophageal biopsy specimens from children with esophageal eosinophilia, we have demonstrated a strong correlation between the peak number of eosinophils and the peak number of eotaxin-3immunostained epithelial cells. Although there may well be other factors contributing to esophageal eosinophilia, this strong correlation suggests that eotaxin-3 produced by epithelial cells plays a major role in attracting eosinophils to the esophagus. To our surprise, we found that PPI treatment significantly decreased eotaxin-3 expression (which correlated with eosinophil count) only in the proximal esophagus, even though PPI effects in reducing acid reflux are strongest in the distal esophagus. To explore possible mechanisms for this disparity, we studied acid and bile salt effects on IL-13-stimulated eotaxin-3 secretion in esophageal squamous cell lines, and found that exposures to acid alone, bile salts alone, and the combination of acid and bile salts have suppressive effects on IL-13-induced eotaxin-3 secretion. Acid-independent, antiinflammatory effects of omeprazole seem to further suppress that eotaxin-3 secretion, but the magnitude of that suppression is attenuated under acid-bile salt exposure conditions. Our in vitro observations provide a possible explanation for the disparate PPI effects that we observed on eotaxin-3 expression in the proximal and distal esophagus of our patients with esophageal eosinophilia. In patients with Th2-cytokine driven esophageal eosinophilia who have little or no reflux, epithelial cell expression of eotaxin-3 will be high in both the proximal and distal esophagus (Figure 9A). When those patients are given PPIs, epithelial cells in both the proximal and distal esophagus can respond to the acidindependent, anti-inflammatory effects of the PPI, which profoundly diminish eotaxin-3 expression in both locations (Figure 9B). For patients who have Th2-cytokine driven esophageal eosinophilia and gastroesophageal reflux, in contrast, eotaxin3 expression in the distal esophagus might be suppressed by reflux, while eotaxin-3 expression remains high in the proximal esophagus where reflux exposure is minimal (Figure 9C). When those patients are treated with PPIs, the proximal esophagus responds to the acid-independent, anti-inflammatory PPI effects in suppressing eotaxin-3, and the decreased expression of this eosinophil chemoattractant results in a commensurate reduction in proximal esophageal eosinophilia (Figure 9D). In the distal esophagus, however, refluxed acid already might be suppressing eotaxin-3 expression (Figure 9C). By decreasing that acid reflux, PPI treatment will free those distal esophageal cells from acid-induced suppression of eotaxin-3. Although the anti-inflammatory effects of PPIs might decrease eotaxin-3 expression, those effects would be masked in the distal esophagus by the increase in eotaxin-3 expression that results from the suppression of acid reflux. Thus, PPI treatment might have little net effect on eotaxin-3 expression in the distal esophagus. Even patients who have esophageal eosinophilia without clinically apparent GERD might have disparate PPI effects on the proximal and distal esophagus. Esophageal pH monitoring studies have documented that the distal esophagus can be exposed to acid (pH ,4) for 4% to 5% of a 24-hour monitoring period in normal adults, and for 8% to 13% of the monitoring period in normal newborns and infants [20,21]. In contrast, the proximal esophagus normally is exposed to acid for ,1% of the day [22]. In the distal esophagus, conceivably, even the normal amount of reflux might be sufficient to suppress eotaxin-3 secretion. In this and earlier studies, we have observed great variability in the amount of eotaxin-3 secretion induced by Th2 cytokine stimulation in esophageal squamous cells from different patients. In response to IL-13 in a dose of 100 ng/ml, for example, the amount of eotaxin-3 protein secreted by EoE2-T cells is almost two orders of magnitude greater than that secreted by EoE1-T cells. At that high level of cytokine-stimulated eotaxin-3 secretion, we could not demonstrate its suppression by acid and bile salts and, even though omeprazole did suppress eotaxin-3 secretion significantly, the levels of secretion remained profoundly elevated nevertheless. When we repeated our experiments using a lower dose of IL-13 (1 ng/ml) at which EoE2-T cells secrete eotaxin-3 in amounts comparable to EoE1-T cells, the two cell lines responded similarly to acid, bile salts and omeprazole. These findings suggest that, for patients with very high cytokine-stimulated levels of eotaxin-3 secretion, the effects of acid, bile and PPIs might not be clinically apparent. Individual differences in esophageal sensitivity to Th2 cytokine-stimulated eotaxin-3 secretion might explain why some patients respond to PPI treatment and others do not. In a study published in abstract form, Molina-Infante et al. observed that PPI treatment caused transcriptional downregulation of eotaxin-3 mRNA in both the proximal and distal esophagus of adults with PPI-REE [23]. Unlike our study, which used immunohistochemical staining to demonstrate the epithelial cells responsible for eotaxin-3 protein production, those investigators evaluated eotaxin-3 mRNA levels in esophageal biopsy specimens containing numerous non-epithelial cells, including eosinophils, which also can express eotaxin-3. Since, by definition, eosinophils decrease with PPI treatment in PPI-REE, it is possible that the observed reductions in post-treatment eotaxin-3 mRNA levels might have been the result rather than the cause of the decreased eosinophils, and may not reflect what is occurring in the epithelial cells. It is also possible that acid and bile might affect the translation or post-translational degradation of eotaxin-3 rather than its transcription. Further studies are needed to establish the mechanisms responsible for acid, bile salt, and PPI effects on eotaxin-3 expression. For our experiments delineating the effects of pH and bile on eotaxin-3 secretion, we exposed esophageal squamous cells to a medium containing a mixture of acid and bile salts simulating the composition of gastric refluxate described in patients with GERD [1820]. We felt that exposure to acid-bile salt medium at various pH levels would provide more physiologically relevant data on the effects of acidic reflux than exposure to various strengths of hydrochloric acid alone. For esophageal pH monitoring studies, acid reflux typically is defined by an esophageal pH drop below 4. However, weakly acidic gastroesophageal reflux (with pH levels between 4 and 7) has been shown to occur frequently in patients with GERD who take PPIs [24]. Therefore, we exposed our cells to pH levels ranging from 4.0 to 7.2 to simulate the effects of gastroesophageal reflux in patients on and off PPIs. We did not expose esophageal squamous cells to medium with pH levels ,4, because such strongly acidic medium kills these cells in culture. We identified 264 children with esophageal eosinophilia who were treated at our medical center from 2008 to 2012, but we found only 10 who qualified for our study. There are several reasons for the large number of patient exclusions. Endoscopy in pediatric patients usually involves general anesthesia with its attendant inconvenience, expense and risks, including recent concerns regarding neurotoxic effects of anesthesia on the developing brain [25]. Most of our patients with esophageal eosinophilia were seen prior to publication of the 2011 updated consensus statement recognizing the entity of PPI-REE [6]. Therefore, our children evaluated for EoE-like symptoms usually were given an empiric trial of PPIs (without a prior endoscopy) to exclude GERD, as recommended in the 2007 consensus guidelines [15]. Patients whose symptoms resolved on PPIs usually were assumed to have GERD, and were continued on PPIs without endoscopic evaluation. Therefore, the large majority of our index endoscopies showing esophageal eosinophilia were performed for patients whose symptoms had not responded to PPIs. Since these patients had already received PPIs, they were not eligible for our study. Thus, we could identify only 40 potential subjects who had esophageal biopsies taken both before and after PPI treatment, and we excluded 30 of those because they had received other, concomitant treatments during their PPI therapy (steroids and/or diet). Consequently, in clinical practice it is difficult to find pediatric patients with esophageal eosinophilia who have had endoscopic examinations before and after a course of PPI therapy alone. Five (50%) of our 10 study patients with esophageal eosinophilia had PPI-REE, a frequency within the range reported by others [2631]. Compared to the PPI non-responders, our patients with PPI-REE had more frequent symptoms of weight loss/poor weight gain and vomiting, and their baseline peak eosinophil counts were somewhat lower. No other clinical features distinguished our PPI responders from non-responders. Our patient numbers are too small for meaningful multivariable analysis, but it seems unlikely that the minor clinical differences between the groups that we observed would have clinical utility for identifying patients with esophageal eosinophilia who are likely to respond to PPIs. Thus, our findings in children are similar to those in adults with esophageal eosinophilia reported by Dellon et al., who found that no clinical or endoscopic features independently distinguished PPI responders from non-responders. [32] Our study has a number of limitations in addition to its small sample size. It suffers from limitations inherent in a retrospective investigation conducted at a single medical center. We cannot assess for patient compliance with PPI therapy, for concomitant use of unprescribed therapies, and for seasonal and environmental influences. Also, our patients did not have esophageal pHimpedance monitoring studies. With our findings that differences in acid exposure might underlie the different effects of PPIs on eotaxin-3 expression in the proximal and distal esophagus, it will be important for future studies to measure the pH, frequency, duration and proximal extent of gastroesophageal reflux to elucidate the role of acid exposure in modifying PPI-responsiveness. In conclusion, in this study designed to explore the contributions of the acid-inhibitory and the acid-independent, anti-inflammatory effects of PPIs on esophageal eosinophilia, we have found that PPIs decrease eotaxin-3 expression significantly in the proximal but not the distal esophagus. This finding was contrary to our expectation that PPI effects in reducing acid reflux would preferentially benefit the distal esophagus. Our in vitro finding that exposure to acid and bile inhibits Th2-cytokine-stimulated eotaxin-3 secretion by esophageal epithelial cells might account for the surprising results. In the distal esophagus, exposure to acid reflux might suppress eotaxin-3 expression, rendering it less sensitive to the acid-independent, anti-inflammatory effect of PPIs. In the proximal esophagus, those acid-independent, anti-inflammatory suppressive effects of PPIs on eotaxin-3 are manifest most prominently. Further studies are needed to substantiate and expand upon our findings. However, since it appears that PPIs can have disparate effects on eotaxin-3 expression in the proximal and distal esophagus, clinicians should consider taking biopsy samples from both the proximal and distal esophagus when assessing the effects of PPIs on esophageal eosinophilia. Supporting Information Pre- and Post-PPI Treatment Histological Findings. Conceived and designed the experiments: JYP XZ RFS SJS EC. Performed the experiments: JYP XZ NN EC. Analyzed the data: JYP XZ NN RFS SJS EC. Contributed reagents/materials/analysis tools: JYP XZ NN RFS SJS EC. Wrote the paper: JYP XZ NN RFS SJS EC. 1. Straumann A , Kristl J , Conus S , Vassina E , Spichtin HP , et al. ( 2005 ) Cytokine expression in healthy and inflamed mucosa: probing the role of eosinophils in the digestive tract . Inflamm Bowel Dis 11 : 720 - 726 . 2. Straumann A , Bauer M , Fischer B , Blaser K , Simon HU ( 2001 ) Idiopathic eosinophilic esophagitis is associated with a T(H)2-type allergic inflammatory response . J Allergy Clin Immunol 108 : 954 - 961 . 3. Blanchard C , Mingler MK , Vicario M , Abonia JP , Wu YY , et al. ( 2007 ) IL-13 involvement in eosinophilic esophagitis: transcriptome analysis and reversibility with glucocorticoids . J Allergy Clin Immunol 120 : 1292 - 1300 . 4. Blanchard C , Wang N , Stringer KF , Mishra A , Fulkerson PC , et al. ( 2006 ) Eotaxin-3 and a uniquely conserved gene-expression profile in eosinophilic esophagitis . J Clin Invest 116 : 536 - 547 . 5. Cheng E , Zhang X , Huo X , Yu C , Zhang Q , et al. ( 2013 ) Omeprazole blocks eotaxin-3 expression by oesophageal squamous cells from patients with eosinophilic oesophagitis and GORD. Gut 62 : 824 - 832 . 6. Liacouras CA , Furuta GT , Hirano I , Atkins D , Attwood SE , et al. ( 2011 ) Eosinophilic esophagitis: updated consensus recommendations for children and adults . J Allergy Clin Immunol 128 : 3 - 20 e26; quiz 21 - 22 . 7. DeBrosse CW , Franciosi JP , King EC , Butz BK , Greenberg AB , et al. ( 2011 ) Long-term outcomes in pediatric-onset esophageal eosinophilia . J Allergy Clin Immunol 128 : 132 - 138 . 8. Straumann A , Spichtin HP , Grize L , Bucher KA , Beglinger C , et al. ( 2003 ) Natural history of primary eosinophilic esophagitis: a follow-up of 30 adult patients for up to 11.5 years . Gastroenterology 125 : 1660 - 1669 . 9. Spechler SJ , Genta RM , Souza RF ( 2007 ) Thoughts on the complex relationship between gastroesophageal reflux disease and eosinophilic esophagitis . Am J Gastroenterol 102 : 1301 - 1306 . 10. Kedika RR , Souza RF , Spechler SJ ( 2009 ) Potential anti-inflammatory effects of proton pump inhibitors: a review and discussion of the clinical implications . Dig Dis Sci 54 : 2312 - 2317 . 11. Sasaki T , Yamaya M , Yasuda H , Inoue D , Yamada M , et al. ( 2005 ) The proton pump inhibitor lansoprazole inhibits rhinovirus infection in cultured human tracheal epithelial cells . Eur J Pharmacol 509 : 201 - 210 . 12. Handa O , Yoshida N , Fujita N , Tanaka Y , Ueda M , et al. ( 2006 ) Molecular mechanisms involved in anti-inflammatory effects of proton pump inhibitors . Inflamm Res 55 : 476 - 480 . 13. Zhang X , Cheng E , Huo X , Yu C , Zhang Q , et al. ( 2012 ) Omeprazole blocks STAT6 binding to the eotaxin-3 promoter in eosinophilic esophagitis cells . PLoS One 7 : e50037 . 14. Huo X , Zhang X , Yu C , Zhang Q , Cheng E , et al. ( 2013 ) In oesophageal squamous cells exposed to acidic bile salt medium, omeprazole inhibits IL-8 expression through effects on nuclear factor-kappaB and activator protein-1. Gut 63 : 1042 - 1052 . 15. Furuta GT , Liacouras CA , Collins MH , Gupta SK , Justinich C , et al. ( 2007 ) Eosinophilic esophagitis in children and adults: a systematic review and consensus recommendations for diagnosis and treatment . Gastroenterology 133 : 1342 - 1363 . 16. Zhang HY , Zhang X , Chen X , Thomas D , Hormi-Carver K , et al. ( 2008 ) Differences in activity and phosphorylation of MAPK enzymes in esophageal squamous cells of GERD patients with and without Barrett's esophagus . Am J Physiol Gastrointest Liver Physiol 295 : G470 - 478 . 17. Cortes JR , Rivas MD , Molina-Infante J , Gonzalez-Nunez MA , Perez GM , et al. ( 2009 ) Omeprazole inhibits IL-4 and IL-13 signaling signal transducer and activator of transcription 6 activation and reduces lung inflammation in murine asthma . J Allergy Clin Immunol 124 : 607 - 610 , 610 e601 . 18. Kauer WK , Peters JH , DeMeester TR , Feussner H , Ireland AP , et al. ( 1997 ) Composition and concentration of bile acid reflux into the esophagus of patients with gastroesophageal reflux disease . Surgery 122 : 874 - 881 . 19. Liu T , Zhang X , So CK , Wang S , Wang P , et al. ( 2007 ) Regulation of Cdx2 expression by promoter methylation, and effects of Cdx2 transfection on morphology and gene expression of human esophageal epithelial cells . Carcinogenesis 28 : 488 - 496 . 20. Richter JE , Bradley LA , DeMeester TR , Wu WC ( 1992 ) Normal 24-hr ambulatory esophageal pH values . Influence of study center, pH electrode , age, and gender. Dig Dis Sci 37 : 849 - 856 . 21. Vandenplas Y , Goyvaerts H , Helven R , Sacre L ( 1991 ) Gastroesophageal reflux, as measured by 24-hour pH monitoring, in 509 healthy infants screened for risk of sudden infant death syndrome . Pediatrics 88 : 834 - 840 . 22. Hirano I , Richter JE ( 2007 ) ACG practice guidelines: esophageal reflux testing . Am J Gastroenterol 102 : 668 - 685 . 23. Molina-Infante J , Rivas MD , Rodriguez GV , Hernandez-Alonso M , Due nasSadornil C , et al. ( 2013 ) SU1828 Remission in Proton Pump InhibitorsResponsive Esophageal Eosinophilia Correlates With Downregulation of Eotaxin-3 and TH2 Cytokines, Similarly to Eosinophilic Esophagitis After Steroids [abstract] . Gastroenterology 144 ( 1 Suppl): S484 -. 24. Frazzoni M , Conigliaro R , Melotti G ( 2011 ) Weakly acidic refluxes have a major role in the pathogenesis of proton pump inhibitor-resistant reflux oesophagitis . Aliment Pharmacol Ther 33 : 601 - 606 . 25. Olsen EA , Brambrink AM ( 2013 ) Anesthesia for the young child undergoing ambulatory procedures: current concerns regarding harm to the developing brain . Curr Opin Anaesthesiol 26 : 677 - 684 . 26. Dranove JE , Horn DS , Davis MA , Kernek KM , Gupta SK ( 2009 ) Predictors of response to proton pump inhibitor therapy among children with significant esophageal eosinophilia . J Pediatr 154 : 96 - 100 . 27. Sayej WN , Patel R , Baker RD , Tron E , Baker SS ( 2009 ) Treatment with highdose proton pump inhibitors helps distinguish eosinophilic esophagitis from noneosinophilic esophagitis . J Pediatr Gastroenterol Nutr 49 : 393 - 399 . 28. Peterson KA , Thomas KL , Hilden K , Emerson LL , Wills JC , et al. ( 2010 ) Comparison of esomeprazole to aerosolized, swallowed fluticasone for eosinophilic esophagitis . Dig Dis Sci 55 : 1313 - 1319 . 29. Molina-Infante J , Ferrando-Lamana L , Ripoll C , Hernandez-Alonso M , Mateos JM , et al. ( 2011 ) Esophageal eosinophilic infiltration responds to proton pump inhibition in most adults . Clin Gastroenterol Hepatol 9 : 110 - 117 . 30. Schroeder S , Capocelli KE , Masterson JC , Harris R , Protheroe C , et al. ( 2013 ) Effect of proton pump inhibitor on esophageal eosinophilia . J Pediatr Gastroenterol Nutr 56 : 166 - 172 . 31. Moawad FJ , Veerappan GR , Dias JA , Baker TP , Maydonovitch CL , et al. ( 2013 ) Randomized controlled trial comparing aerosolized swallowed fluticasone to esomeprazole for esophageal eosinophilia . Am J Gastroenterol 108 : 366 - 372 . 32. Dellon ES , Speck O , Woodward K , Gebhart JH , Madanick RD , et al. ( 2013 ) Clinical and Endoscopic Characteristics do Not Reliably Differentiate PPIResponsive Esophageal Eosinophilia and Eosinophilic Esophagitis in Patients Undergoing Upper Endoscopy: A Prospective Cohort Study . Am J Gastroenterol 108 : 1854 - 1860 .


This is a preview of a remote PDF: http://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0101391&type=printable

Jason Y. Park, Xi Zhang, Nathalie Nguyen, Rhonda F. Souza, Stuart J. Spechler, Edaire Cheng. Proton Pump Inhibitors Decrease Eotaxin-3 Expression in the Proximal Esophagus of Children with Esophageal Eosinophilia, PLOS ONE, 2014, DOI: 10.1371/journal.pone.0101391