Broad Transcriptional Response of the Human Esophageal Epithelium to Proton Pump Inhibitors

Abstract

Background

Proton pump inhibitors (PPIs) have been recognized as a primary treatment of eosinophilic esophagitis (EoE), an allergic inflammatory disease of the esophageal mucosa. The mechanisms underlying esophageal epithelial responses to PPIs remain poorly understood.

Objective

We hypothesized that PPIs can counteract interleukin 13 (IL-13)–mediated esophageal epithelial responses that are germane for EoE pathogenesis.

Methods

Transcriptional responses of human esophageal cells to IL-13 and the PPIs omeprazole and esomeprazole were assessed by RT-PCR and RNA sequencing. Cytokine secretion was measured by multiplex and ELISA.

Results

Human esophageal epithelial cells robustly responded to PPI stimulation by inducing a set of 479 core genes common between omeprazole and esomeprazole treatments. The transcriptional response to PPIs was partially mediated through the AHR signaling pathway, as the AHR antagonist GNF-351 modified approximately 200 genes, particularly those enriched in metabolic processes and regulation of cell death. PPI treatment reversed ~20% of the IL-13 transcriptome. Functional analysis of the PPI-responsive, upregulated genes revealed enrichment in metabolic and oxidation processes, and the unfolded protein response. In contrast, downregulated genes were overrepresented in functional terms related to cell division and cytoskeletal organization, that were also enriched for the genes in the EoE transcriptome reversed by PPIs. Furthermore, PPI treatment decreased the IL-13—induced proliferative response of esophageal epithelial cells.

Conclusions

These results demonstrate broad effects of PPIs on esophageal epithelium, including their ability to curtail transcriptomic processes involved in cellular proliferation and IL-13–induced responses, and highlight to the importance of AHR signaling in mediating these responses.

Introduction

Proton pump inhibitors (PPIs), such as omeprazole and esomeprazole, are membrane-permeable, weak bases that covalently bind to cysteine residues of gastric H+/K+−ATPase, leading to its inactivation and the subsequent inhibition of acid secretion from parietal cells. As acid suppressants, PPIs have been historically used for the treatment of acid-related diseases, mainly gastroesophageal reflux disease (GERD), dyspepsia, and peptic ulcer disease.¹ At the same time, the anti-inflammatory activities of PPIs have been attributed to mechanisms largely independent of their acid-suppressive effect on parietal cells but rather driven by a direct effect of PPIs on epithelial, endothelial, and immune cells, including mast cells.^2-4

PPIs are extensively used for treatment of eosinophilic esophagitis (EoE), a chronic allergic inflammation of the esophagus characterized by eosinophilic infiltration in the esophageal mucosa. Though initially used to distinguish GERD from EoE, PPIs recently have become a first-line therapy for EoE, as 10-50% of cases respond to this single treatment, yet the mechanism remains largely unclear.^{5, 6} EoE pathogenesis is driven by a transcriptional response of the esophageal tissue to the pro-atopy cytokine interleukin 13 (IL-13), which is largely dependent on the signal transducer and activator of transcription 6 (STAT6) transcription factor. Accordingly, blocking antibodies against IL-13, as well as inhibition of the STAT6 signaling pathway, result in decreased expression of IL-13–dependent genes and subsequent reduction of EoE responses in pre-clinical and early clinical studies.^{7, 8} The beneficial effects of PPIs in EoE have been mainly attributed to their ability to block expression and secretion of the key eosinophil chemoattractant eotaxin-3 from epithelial cells by inhibiting STAT6 phosphorylation and it’s binding to the eotaxin-3 promoter. Though this is interesting, it seems implausible that this alone can account for the broad effects of PPIs in EoE, including restoration of epithelial barrier function and differentiation.^9-12

Collectively, these findings highlight the critical need for elucidating the molecular mechanisms that govern esophageal epithelial responses to PPI in allergic inflammation and specifically in EoE. Herein, we tested the hypothesis that PPIs counteract IL-13–mediated esophageal epithelial responses that are germane for EoE pathogenesis. By performing gene expression analysis, we show that omeprazole and esomeprazole have more broad effects on the esophageal epithelium than previously appreciated and their effects were mainly elicited on pathways not attributed to IL-13. We provide evidence that these effects are mediated through the aryl hydrocarbon receptor (AHR) signaling pathway, at least in part. Collectively, our results suggest that the beneficial effect of PPIs in the treatment of EoE is likely driven by partial reversal of the IL-13–mediated, disease-associated esophageal transcriptome and by IL-13–independent effects on cellular proliferation and metabolism that are largely AHR dependent.

Materials and Methods

Cell culture and treatment

The esophageal hTERT-immortalized human epithelial cell EPC2 line was a kind gift from Dr. Anil Rustgi (University of Pennsylvania). Primary human esophageal epithelial cells and fibroblasts were isolated essentially as described. ¹³ Briefly, esophageal biopsies were mechanically dispersed and treated with collagenase and dispase following treatment with trypsin/EDTA. Cells were plated on the feeder layer of irradiated NIH 3T3 cells in keratinocyte serum-free media (KSFM, Life Technologies). Primary esophageal fibroblasts which occasionally appeared in the culture were separated from the epithelial cells by differential trypsinization and transferred to DMEM 10% FCS medium. Clinical characteristics of the patients whose biopsies were used for generating primary cells for the study are summarized in Table 1. For air liquid interface (ALI) culture 150 x 10⁵ cells were seeded and grown to confluence while fully submerged in low-calcium (CaCl2 0.09 mM) KSFM on 0.4-mm pore-size permeable supports (Corning Incorporated). Confluent monolayers were then switched to high-calcium (CaCl₂ 1.8 mM) KSFM for an additional 5 days. To induce epithelial differentiation, the culture medium was removed from the inner chamber of the permeable support to expose the cells to the air interface. Barrier formation was assessed by measuring trans-epithelial electrical resistance (TEER) with EVOM epithelial voltohmmeter with “chopstick” electrodes (World Precision Instruments). For the monolayer cultures epithelial cells were seeded at a density of 2.5 x 10⁵ cells/well in a 24-well plate in KSFM. Fibroblasts were seeded at 2 x 10⁵ cells/well in a 48 well plate in 350 μL of DMEM 10% FCS. The next day, the medium was replenished. After 24 hours, stimulants were added in a total of 350 - 500 μL of fresh medium for 24 hours. IL-13 (Peprotech, 200-13) was added to the final concentration of 100 ng/mL unless otherwise indicated. Omeprazole (TOCRIS, No.2583) and esomeprazole (Sigma, E7906) were dissolved in DMSO to the stock concentration of 100 mM, aliquoted into 10-μL amounts and stored at −80°C. Proton pump inhibitors (PPIs) were thawed once and used at a working concentration of 100 μM; cells were pre-treated with the PPIs for 1 hour prior to adding IL-13 to the medium. GNF-351 (MedChemExpress, HY-102023) was dissolved in DMSO to a 20-mM stock concentration. The stock was aliquoted in 10- μl doses and stored at −80C. Cells were pre-treated with GNF-351 at 2 μM final concentration for 30 min prior to adding PPIs. Ki67 immunohistochemistry was performed by the Pathology Research Core at CCHMC using CONFIRM anti-Ki-67 (30-9) rabbit monoclonal antibody (Roche, 790-4286).

Primary cells used in the study

Culture #	Primary cells	PPI-confirmed EoE	History of PPI	Peak distal eosinophils	Distal esophagus pathology
1	Epithelial	Yes	Yes	16	Eosinophilic Esophagitis
2		No	Yes	26	Eosinophilic Esophagitis
3		ND	Yes	0	Normal
4		Yes	Yes	32	Eosinophilic Esophagitis
5		No	ND	27	Eosinophilic Esophagitis
6		Yes	Yes	0	Normal
7		ND	Yes	20	Eosinophilic Esophagitis
8		ND	Yes	9	Abnormal
9		Yes	Yes	12	Abnormal
1	Fibroblasts	ND	Yes	0	Normal
2		ND	Yes	25	Eosinophilic Esophagitis

ND not determined

RNA extraction and cDNA synthesis

Cells were lysed in Tripure Isolation Reagent at 700 μL/well (Sigma, 11667165001) and stored at −20°C before RNA isolation. Chloroform was added at a volume of 140 μL per sample, and tubes were shaken vigorously for 15 seconds and allowed to stand for 2–15 minutes at room temperature. Samples were centrifuged at 12,000 × g for 15 min at 4°C, and 350 μL of the upper aqueous phase was collected and mixed at a 1:1 ratio with 100% ethanol. Further isolation was performed using the Quick-RNA MicroPrep Kit (Zymo Research, R1051) per the manufacturer’s instructions. The RNA concentration was measured by Nanodrop, and the RNA integrity number (RIN) was determined by the Gene Expression Core at Cincinnati Children’s Hospital Medical Center (CCHMC) using the Agilent 2100 Bioanalyzer. cDNA synthesis was performed according to the protocol of the ProtoScript II Reverse Transcriptase kit (NEB, M0368).

ELISA and multiplex

Cells were treated with IL-13 at 100 ng/mL for 24 hours. Prior to collecting supernatants, NaCl at a final concentration of 500 mM was added to each well, and the plate was rotated for 15-30 min at room temperature. Supernatants were centrifuged at 12,000 × g at 4°C for 15 min before use. ELISA was performed with the kits for human CCL26/Eotaxin-3 (R&D systems, DY346) per the manufacturer’s instructions. Multiplex analysis was performed by Eve Technologies using the Human Cytokine Array/Chemokine Array 65-Plex Panel (HD65). Experiments were performed with 1-2 independent cultures of EPC2 cells and 2 primary esophageal epithelial cells in duplicates. Only cytokines detected at the concentration over 5 pg/ml for at least one treatment were used for the analysis. T-test was performed to assess significant changes in the secretion by comparing individual treatments (IL-13, omeprazole) to the untreated cells and combined treatment (IL-13 + omeprazole) to IL-13 alone. P < 0.05 was considered significant.

3’ RNA sequencing and data processing

Submerged monolayer culture of the EPC2 cells seeded at density of 2.5 x 10⁵ cells/well in a 24-well plate was used for the RNA sequencing. RNA sequencing was performed with high-quality RNA (RIN > 8) using the QuantSeq 3’ mRNA-Seq Library Prep Kit FWD for Illumina (Lexogen, 015.96). Libraries were subjected to quality control and concentration measurements at the Gene Expression Core at CCHMC. Libraries were diluted to 5 nM final concentrations and sequenced on a HiSeq 4000 Illumina sequencing machine at the Genomics & Cell Characterization Core Facility at the University of Oregon with 100-150–bp length single reads. Data analysis and visualization were performed using the CLC Genomics Workbench v12.0.3 (Qiagen). Briefly, raw data sequencing files were imported into the software, sequences were trimmed from the adapters, and alignment was performed using the HG38 human genome. Total read counts for the samples were between 1.3 to 6.8 million per sample, and 85% to 95% of the reads were successfully mapped to the forward DNA strand. Of these reads, 63% to 79% were mapped to the protein-coding regions. Differentially expressed genes were defined by fold change and statistical filtering and clustered, as indicated in the figure legends. For some heat maps, Cluster 3.0 was applied for clustering genes using Euclidean distance and average linking parameters, and Java TreeView was used for visualization of heat maps (http://bonsai.hgc.jp/~mdehoon/software/cluster/software.htm). Gene ontology (GO) enrichment analysis, which uses statistical methods to determine functional pathways and cellular processes associated with a given set of genes, was performed with the ToppGene suite ¹⁴ (https://toppgene.cchmc.org/). Unless otherwise indicated, differentially expressed genes were used as input for GO analysis. Venny (https://bioinfogp.cnb.csic.es/tools/venny/) was used to intersect gene lists. As indicated, expression data were intersected with the EoE transcriptome, the list of 1607 significantly dysregulated transcripts identified by comparing gene expression in the biopsies of 10 patients with active EoE, 9 of which were irresponsive to the PPI treatment, to normal controls by the RNA sequencing analysis. ¹⁵

Results

Transcriptional response of submerged EPC2 cells to IL-13 and PPIs

We aimed to investigate the ability of PPIs to reverse IL-13–mediated transcriptional responses in EoE. To this end, we examined the immortalized esophageal epithelial cell line EPC2 that has been widely used to model epithelial properties of the homeostatic and diseased human esophagus.^{13, 16}

We analyzed global gene expression profiles by RNA sequencing of submerged monolayer culture of the EPC2 cells treated with IL-13, omeprazole and esomeprazole either alone or in combinations. Principal component analysis separated the gene expression profiles of each group (Figure 1A). Collectively, 709 differentially expressed genes (DEG) were identified by comparing between treatment groups (ANOVA, P < 0.05; fold change > 4; with Benjamini-Hochberg false discovery rate (FDR) (Figure 1B). Several characteristics of the cellular response to these stimuli became apparent from this analysis. First, EPC2 cells responded to IL-13 by upregulating and downregulating a number of genes, many of which were commonly observed in IL-13–mediated responses, including those in the EoE transcriptome (Supplemental Figure 1 and Supplemental Table 1, see bolded genes in UT vs IL13 sheet). Among the most highly upregulated genes were CCL26, TNFAIP6, NTRK1, SERPINB4, CAPN14, and ANO1, which are typically associated with IL-13 responsiveness of the esophageal epithelium in active EoE, as the expression of these genes is STAT6 dependent.^{8, 17} Second, the degree of transcriptional response was surprisingly stronger for PPIs than IL-13, as evidenced by a larger number of dysregulated genes in PPI-treated cells compared with IL-13–treated cells (see also Supplemental Figure 2). Notably, despite substantial overlap in the gene expression signature following omeprazole and esomeprazole stimulation, the latter induced a stronger response as evident by the scale of changes in gene expression (increased intensity of yellow and blue colors). Indeed, pairwise analysis revealed that 188 genes were altered in the cells treated with IL-13, whereas 573 genes and 1,564 genes were dysregulated in response to omeprazole and esomeprazole, respectively (FDR P < 0.05, 2-fold change; Supplemental Table 1). Overall, 479 genes were regulated by both omeprazole and esomeprazole, demonstrating remarkable similarity in their regulation (Figure 1C). Functional analysis of the upregulated genes shared between PPIs revealed enrichment of gene ontology (GO) terms associated with metabolic and oxidation processes, as well as the unfolded protein response and lipid metabolism. In contrast, the downregulated genes were overrepresented in GO categories related to cell division and cytoskeletal organization (Figure 1C, Biological processes).

Figure 1. Transcriptional signature of EPC2 cells in response to IL-13 and proton pump inhibitors (PPIs).

A, Principal component analysis (PCA) of RNA sequencing results from EPC2 cells, n = 4 for each stimulation is shown. B, The heat map shows clustering of differentially expressed genes (ANOVA, FDR P < 0.05, fold change 4). The IL-13 response represents genes dysregulated by IL-13; examples of genes commonly associated with IL-13 signature are indicated. Note the stronger response to esomeprazole (ESO) than omeprazole (OME) as judged by the scale of changes in the expression level of genes (compare intensities of yellow and blue colors in OME vs. ESO samples). C, The Venn diagram shows overlap of the genes dysregulated by omeprazole and esomeprazole. The heat map shows the expression level change for the 479 genes common to OME and ESO compared to untreated cells. The bar graphs show biological processes enriched for upregulated and downregulated genes defined by gene ontology (GO) analysis (FDR P < 0.05). For B and C, yellow and blue colors represent upregulated and downregulated genes, respectively. UT, untreated; OME, omeprazole; ESO, esomeprazole.

Transcriptional response of submerged primary cells to IL-13 and PPIs

To further assess epithelial responses to IL-13 in the absence and presence of the PPIs, the expression of several IL-13- and PPI-responsive genes in 4 primary epithelial cell lines were examined. Similar to EPC2 cells, primary cells robustly respond to PPIs, by increasing expression of CYP1A1, HMOX1, and MT1H with a stronger response following esomeprazole stimulation (Figure 2A). While following stimulation with IL-13 expression of several IL-13—inducible genes including NTRK1, SERPINB3 and TNFAIP6 was modestly decreased by esomeprazole in combination with IL-13 compared to IL-13 alone, PPIs did not prevent robust upregulation of these genes (Figure 2B).

Figure 2. Transcriptional response of primary epithelial cells to IL-13 and PPIs.

A,B, The expression level of the indicated genes was assessed by RT-PCR and normalized to the housekeeping gene GAPDH. Primary cell lines were stimulated with the PPIs alone (A) or in combination with IL-13 (B). Combined data for 4 independent primary cell lines in duplicates are shown as box-and-whiskers plot, **** P < 0.0001, *** P < 0.001, ** P < 0.01, ns – not significant, ANOVA with Holm-Sidak correction. For box-and-whisker plots, the box represents 50th percentile of the data, whiskers show minimum and maximum values, and the line in the box represents the median. UT, untreated; OME, omeprazole; ESO, esomeprazole.

To evaluate the cellular specificity of the PPI response, the expression of PPI-inducible genes in primary esophageal epithelial cells and primary esophageal fibroblasts was examined. Unlike epithelial cells, fibroblasts did not induce CYP1A1 and HMOX1; whereas, expression of MT1B and MT1H was upregulated, albeit variably (Supplemental Figure 3A). Similar to the epithelial response, induction of IL-13 target genes CCL26 and periostin (POSTN) was not primarily affected by the PPI treatment (Supplemental Figure 3B). Collectively, these results suggest cellular specificity of the transcriptional responses to PPIs.

Transcriptional response of differentiated epithelial cells to PPIs

Epithelial cells grown at the ALI represent a commonly used in vitro model for studying esophageal epithelial differentiation. ¹⁶ In this model, epithelial stratification and differentiation are induced following exposure of the monolayer cells grown on the membrane to the air. This results in formation of a multi-layered epithelial culture characterized by increased trans-epithelial electrical resistance (TEER). In the presence of IL-13, cellular proliferation is increased ¹⁸, the epithelial barrier is damaged and TEER is decreased. We utilized this system to test the response to the PPIs in the EPC2 and primary esophageal epithelial cells. Upon exposure of cells to the ALI (ALI D1), cells were treated for 3 days with IL-13, followed by co-treatment with IL-13 and PPIs for an additional 48 hours (Figure 3A). As expected, IL-13 treatment substantially decreased TEER compared to the untreated samples, which was not reversed by the PPIs (Figure 3A), indicating that the PPIs do not reverse IL-13—mediated loss of the epithelial barrier integrity. However, epithelial cells grown at the ALI robustly responded to the PPI stimulation as was evident by the transcriptional induction of the PPI-inducible genes CYP1A1, MT1B and MT1H (Figure 3B). Given that PPI treatment led to the down regulation of the genes related to cell proliferation (Figure 1C), we assessed the proliferative potential of IL-13 in the presence and absence of the PPIs (Figure 3C). During the last 48 hours of the culture cells were treated cells with IL-13 in the presence or absence of the PPIs and proliferation was assessed by immunostaining with Ki67, a nuclear protein that is associated with cellular proliferation. ¹⁹ IL-13 increased proliferation of the basal cells in the culture ¹⁸; whereas omeprazole and esomeprazole significantly diminished this effect (Figure 3C). Collectively, these results suggest that esophageal epithelial cells respond to PPIs independent of their differentiation state and support the anti-proliferative potential of the PPIs in the context of the IL-13 response.

Figure 3. Response of differentiated epithelial cells to IL-13 and PPIs.

A, Schematics above the graph outlines the experimental approach, dotted arrows indicate duration of treatment. Cultures were treated with IL-13 at 20 ng/ml and proton pump inhibitors at 100 μM final concentration. The graph shows trans-epithelial electrical resistance (TEER) of the epithelial cells grown at the air liquid interface culture. Untreated cells show the highest TEER compared to other cultures. * P < 0.05, t-test, with Holm-Sidak correction; ns, not significant. B, The expression level of the indicated genes was assessed by RT-PCR and normalized to the housekeeping gene GAPDH. Combined results for 4 independent primary cell lines and EPC2 cells performed in duplicates are shown as box-and-whiskers plot, **** P < 0.0001, ** P < 0.01, * P < 0.05, Kruskal-Wallis test. For box-and-whisker plots, the box represents 50th percentile of the data, whiskers show minimum and maximum values, and the line in the box represents the median. C, Proliferative response to IL-13 in the absence and presence of PPIs was assessed by staining of the Ki67 marker. IL-13 was used at 100 ng/ml and cells were pre-treated with the PPIs for 1 h prior to stimulating with IL-13. Schematics outline of the experimental approach, representative images of the stained sections are shown above the graph. Arrows point to the brown nuclei of Ki67 positive cells. The graph shows the number of the Ki67 positive cells in 10 high power fields for 3 cultures per condition. **** P < 0.0001, *** P < 0.001, ANOVA with Holm-Sidak correction. UT, untreated; PPI, proton pump inhibitors, OME, omeprazole; ESO, esomeprazole.

Contribution of the AHR signaling pathway to transcriptional response to PPIs

AHR expression is elevated in the esophageal biopsies of EoE patients compared to controls and in IL-13—treated EPC2 cells grown at ALI compared with untreated cultures (Supplemental Figure 4). ^{15, 16} Omeprazole and esomeprazole have been shown to exert responses through activating AHR, leading to the strong induction of several cytochrome P450 (CYP) isoenzymes, including CYP1A1.^{20, 21} Accordingly, EPC2 cells robustly increased expression of CYP1A1 in response to omeprazole and esomeprazole (~ 20 fold), and this induction was blunted by the cell-permeable, high-affinity antagonist of AHR signaling GNF-351 (Figure 4A, upper graph, GNF).²² Notably, basal expression level of CYP1A1 was also decreased in GNF-351–treated compared to untreated cells, indicating active engagement of the AHR signaling pathway under these conditions. In contrast, induction of MT1H was not affected by GNF-351 (Figure 4A, lower graph).

Figure 4. Contribution of AHR signaling pathway to transcriptional response to PPIs.

A, The expression level of CYP1A1 and MT1H normalized to the housekeeping gene GAPDH was assessed by RT-PCR; the combined data for 4 independent experiments are shown as box-and-whiskers plot, **** P < 0.0001, ns – not significant, ANOVA with Holm-Sidak correction. On the plot the box represents 50th percentile of the data, whiskers show minimum and maximum values, and the line in the box represents the median. B, The heat map shows clustering of differentially expressed genes (ANOVA, FDR P < 0.05, fold change 4). Note the stronger inhibitory effect of GNF-351 on esomeprazole than omeprazole (compare intensities of yellow and blue colors of OME vs. OME-GNF and ESO vs. ESO-GNF) C, The Venn diagram shows the overlap of genes dysregulated by the indicated stimuli (P < 0.05, 1.5-fold change). D, The heat map shows clustering of 187 common genes from C. For B, D yellow and blue colors represent upregulated and downregulated genes, respectively. E, Shown are the representative biological processes significantly enriched for the 187 common genes from C defined by gene ontology (GO) analysis (FDR P < 0.05). UT, untreated; OME, omeprazole; ESO, esomeprazole; GNF, GNF-351.

We subsequently hypothesized that AHR signaling was important for PPI-mediated transcriptional responses in esophageal epithelial cells. To test this hypothesis, we pre-treated cells with GNF-351 followed by stimulation with PPIs and performed RNA sequencing. Unsupervised clustering of the affected genes (ANOVA, FDR P < 0.05, 4 fold change) revealed good separation of samples based on the treatment groups (Figure 4B). Comparing the range of effect of GNF-351 pre-treatment on PPI stimulation we found that GNF-351 treatment primarily affected the response to esomeprazole as was evident by the scale of the changes in gene expression (compare intensities of yellow and blue colors in ESO+GNF vs ESO and OME+GNF vs OME columns). As a positive control, expression of CYP1A1 and CYP1B1 genes was reversed by GNF-351 pre-treatment in both omeprazole- and esomeprazole-treated cells. Focusing on the response to esomeprazole, 187 genes were affected by GNF-351 in combination with esomeprazole compared to esomeprazole alone; the basal expression of these 187 genes was not altered by GNF-351 pre-treatment (Figure 4C). Notably, these genes were primarily upregulated by esomeprazole, and this induction was blocked by the AHR signaling antagonist GNF-351 (Supplemental Table 2). Gene ontology (GO) analysis revealed that AHR signaling is overrepresented by genes associated with biological processes related to metabolism, response to stimuli, and cell death (Figure 4D-E). Collectively, these findings demonstrate that PPI-mediated transcriptional responses in the esophageal epithelium are partially mediated by AHR signaling.

Combined effects of IL-13 and PPIs on cytokine secretion and gene expression

Previous research has primarily focused on the ability of PPIs to decrease expression of CCL26 (eotaxin-3) by decreasing the binding of the IL-13 signaling molecule STAT6 to the CCL26 promoter.¹⁰ Consistent with these observations, we observed a modest (~2 fold) decrease of the CCL26 protein level in IL-13–treated EPC2 cells and primary esophageal epithelial cells, but not in the primary esophageal fibroblasts following pre-treatment with PPIs (Figure 5A and Supplemental Figure 5). Notably, this effect was reversed by the AHR inhibitor GNF-351 (Figure 5B).

Figure 5. Effect of PPIs on IL-13–mediated transcription.

A, The relative secretion level of eotaxin-3 in EPC2 and primary esophageal epithelial cells is shown as box-and-whiskers plot and represents the combined data for n = 24 (EPC2 cells) and n = 11 (primary cells) independent experiments, **** P < 0.0001, *** P < 0.001, * P < 0.05, ANOVA with Holm-Sidak correction. On the plot the box represents 50th percentile of the data, whiskers show minimum and maximum values, and the line in the box represents the median. B, The relative secretion of eotaxin-3 was assessed by ELISA in EPC2 cells treated with IL-13 and PPIs in the presence of GNF-35, as indicated. Secretion was normalized to IL-13 treatment, combined data for 7 independent experiments are shown. **** P < 0.0001, ANOVA with Holm-Sidak correction. For box-and-whisker plots, the box represents 50th percentile of the data, whiskers show minimum and maximum values, and the line in the box represents the median. C, Cytokine produced by EPC2 and primary esophageal epithelial cells in response to IL-13 and/or esomeprazole were assessed by the multiplex array. Cytokines that were secreted above 5 pg/ml and significantly changed in at least one of the stimulations (IL-13, ESO vs UT, IL-13 + ESO vs IL-13, P < 0.05, see Supplemental Table 3) are indicated. D,E, Expression of the cytokines shared between EPC2 and primary cells is shown as log(10) fold change (FC) relative to untreated cells (average +/− SEM). Each dot represents an independent sample. F, The Venn diagram shows the intersection of genes dysregulated by either IL-13 alone or in combination with PPIs. The heat map shows the expression of the 86 shared genes (bolded in Venn diagram). The genes whose expression is reversed by PPIs are indicated by lines; examples of IL-13–upregulated genes reversed by PPIs are shown. Yellow and blue colors represent upregulated and downregulated genes, respectively. UT, untreated; OME, omeprazole; ESO, esomeprazole.

To further assess interaction between IL-13 and PPI responses, we performed a multiplex analysis of 65 cytokines in EPC2 and primary esophageal epithelial cell stimulated with IL-13 and esomeprazole either alone or in combination. Cytokines whose average secretion exceeded 5 pg/ml concentration for at least one stimulation were included in the analysis (Supplemental Table 3). We subsequently identified 12 cytokines that were significantly dysregulated in the epithelial cells in response to IL-13 and esomeprazole compared to untreated cells, or a combination of both compared to IL-13 alone (t-test, P <0.05). Six cytokines were shared between EPC2 and primary cells following this analysis (Figure 5C). Secretion of fractalkine, IL-8, IL-18 and VEGF-A was elevated by treatment with esomeprazole alone (Figure 5D). Moreover, eotaxin-3 and RANTES were upregulated by IL-13 and repressed by esomeprazole (Figure 5E). These results show that the response of the epithelial cells to PPIs includes increased secretion of cytokines, such as IL-8 and IL-18, and that PPIs can modify IL-13—mediated cytokine secretion.

To further explore the relationship between PPIs and IL-13–mediated responses in the esophageal epithelium, we assessed the effect of combined stimulation with IL-13 and PPIs on gene expression. By comparing the transcriptional profile of EPC2 cells treated with IL-13 alone or pre-treated with the PPIs prior to IL-13 stimulation, we identified 86 genes whose expression was affected by co-stimulation with the PPIs and IL-13 (27 + 51 + 8 genes shared with UT vs IL-13 circle in the Venn diagram; FDR P < 0.05, 2-fold change; Figure 5F). Expression of 32 of these genes was reversed by both PPIs compared to IL-13 treatment alone. This effect was more pronounced on the genes upregulated by IL-13, including those most highly upregulated in the EoE transcriptome (Figure 5F). Taken together, these results demonstrate that although PPIs can partially reverse the IL-13–mediated transcriptome, the effect on gene expression in the esophageal epithelium is largely IL-13 independent.

Functional enrichment analysis of the epithelial response to PPIs

We hypothesized that the biological processes critical for EoE pathogenesis may be inversely regulated by PPIs. To test this hypothesis, we intersected the PPI-mediated gene expression signature with the EoE transcriptome, the list of genes significantly transcriptionally dysregulated in the biopsies of patients with active EoE compared to unaffected controls. ¹⁵ This analysis revealed 327 shared genes, and the expression of ~70 of those genes with PPIs changed oppositely from that of the EoE transcriptome (Figure 6A). Notably, genes with this inverse expression were enriched for GO terms associated with the cell cycle and microtubule organization (Figure 6B). Collectively, these results suggest that PPIs regulate biological processes germane to EoE pathogenesis, including cell division and metabolism.

Figure 6. Functional enrichment analysis of transcriptional response to proton pump inhibitors.

A, The Venn diagram shows the intersection of genes dysregulated by PPIs and the EoE transcriptome.¹⁵ The shared genes (bolded) were used to generate a gene expression heat map using log(2) fold change compared to either untreated cells (for OME and ESO) or control patients (for EoE transcriptome). Genes whose expression with PPIs changed oppositely from that of the EoE transcriptome are indicated by lines. B, The bar graph shows the biological processes enriched for genes oppositely affected by PPIs (vs. EoE transcriptome) as defined by gene ontology (GO) analysis (FDR P < 0.05). Yellow and blue colors represent upregulated and downregulated genes, respectively. UT, untreated; OME, omeprazole; ESO, esomeprazole.

Discussion

Given the now widely accepted beneficial effects of PPIs for the treatment of EoE, understanding the mechanisms behind their effects is a timely scientific pursuit. Using human esophageal epithelial cells, we demonstrate robust responses to omeprazole and esomeprazole at the transcriptional level and that these responses are partially mediated by the AHR signaling pathway. We show that though IL-13–mediated gene expression and cytokine secretion are partially reversed by PPIs, the PPI-induced responses are largely separate from those induced by IL-13. This finding broadens the protective and therapeutic effects of PPIs compared with simply inhibiting the end-stage inflammatory responses triggered by IL-13. The results also suggest potential benefit of co-administering PPI and anti–type 2 (e.g., anti–IL-13 and anti–IL-4Rα) therapies. We demonstrate that the biological pathways regulated by PPIs are primarily associated with metabolic responses and the cell cycle. Importantly, the latter are enriched for genes from the EoE transcriptome that are reversed by PPIs. Potential anti-proliferative effect of the PPIs is further supported by the ability of the PPIs to decrease IL-13—mediated proliferation in vitro. Collectively, we provide evidence that the esophageal epithelial responses to PPIs are broader than previously reported and that PPIs regulate biological processes germane for EoE pathogenesis. While limited to the submerged epithelial culture, these findings represent the first step in explaining the emerging positive effects that PPI therapy has for EoE, uncovering their broad ability to mechanistically regulate homeostatic epithelial processes (metabolism and cell proliferation), likely providing protection from inflammatory insults, particularly those driven by IL-13 in the case of EoE.

Despite their primary use for acid-related disorders, the beneficial effects of PPIs extend beyond their anti-secretory properties. For example, PPIs can protect mice against TLR-dependent and independent acute systemic inflammation by inhibiting TNF-α and IL-1β production by macrophages.²³ PPIs are also capable of effectively scavenging reactive oxygen species, thereby protecting DNA from damage during oxidative stress.^{24, 25} Treatment with esomeprazole has been effective in preventing fibrosis in lung and ocular tissues by downregulating TGF-β, fibronectin, and matrix metalloproteases.^{2, 26} In addition, anti-inflammatory properties have been attributed to the ability of PPIs to block the production of IL-6, IL-8, and TNF-μ.^{26, 27} Consistent with previous finding, PPIs efficiently counteracted IL-13–mediated secretion of eotaxin-3 in EPC2 and primary esophageal epithelial cells, but not in primary fibroblasts. Herein, we extended this finding by suggesting that the beneficial effect of PPIs greatly extend past inhibition of eotaxin-3 alone. For example, we observed substantial inhibition of another eosinophil and T cell chemoattractant, RANTES, whose expression is also increased in the esophageal tissue of patients with EoE.²⁸ At the same time, our data show that esomeprazole increased secretion of several cytokines, including IL-18, which on the one hand has been implicated in promoting EoE ²⁹, but on the other hand had a protective role in the allergic asthma. ³⁰ The consequences of the IL-18 induction in response to PPIs might depend on the context of the environmental cues at the time of the intervention. Though PPIs are considered beneficial for the treatment of EoE, epidemiologic studies have linked early life exposure to PPIs as a substantial risk factor for EoE and food allergy, although the operational mechanisms are unclear.^{31, 32} Our findings, while limited to the epithelial responses in the cell culture, have potential implications for understanding these mechanisms. The demonstration that PPIs have broad transcriptional consequences on esophageal epithelial cells highlights their potential ability to modify processes (e.g., metabolism and proliferation) that are likely germane for response to subsequent inflammatory triggers associated with the development of chronic allergic responses. Transcriptional response of other cell types involved in the propagation of the allergic inflammation, including fibroblasts and mast cells are likely contributing to overall clinical benefits of the PPIs.

Signaling and metabolic pathways of PPIs differ depending on their chemical structure. For example, omeprazole and esomeprazole signal through AHR, which integrates multiple environmental signals.²¹ Subsequently, activation of AHR leads to induction of the family of CYP450 enzymes that in turn metabolize PPIs.²⁰ The increased expression of CYP1A1 and CYP1B1 enzymes in the EPC2 and primary esophageal epithelial cells in response to PPIs and the efficient blocking of this induction by the AHR antagonist GNF-351 indeed indicate that the omeprazole- and esomeprazole-induced response is mediated by the AHR signaling pathway. It remains to be investigated whether AHR-dependent responses are operational in the patient’s response to PPIs, increased expression of the AHR in the biopsies of patients with active EoE compared to controls support relevance of this mechanism. In contrast to omeprazole and esomeprazole, the PPI rabeprazol is not a ligand for AHR nor a primary target of CYP450 enzymes.²¹ This suggests that mechanisms other than AHR signaling are likely contributing to the effects of PPIs. For example, blocking of eotaxin-3 secretion in IL-13–stimulated human sinonasal epithelial cells by PPIs is linked to inhibition of the non-gastric H⁺/K⁺ ATPase ATP12A.³³ In mast cells, omeprazole likely blocks a vacuolar-type H⁺ ATPase (V-ATPase) ⁴, whereas in melanocytes, omeprazole reduces melanogenesis by inhibiting a copper-transporting P-type ATPase.³⁴ Our data show that the AHR pathway partially contributes to the transcriptional response to PPIs, though the contribution of AHR and other signaling pathways to the beneficial effects of the PPIs in EoE, including on the EoE transcriptome warrants further investigation.

Basal zone hyperplasia (BZH) is a major histopathologic characteristic of EoE.³⁵ Though the molecular mechanisms behind BZH are not fully understood, IL-13 signaling is likely involved.^{18, 36} Our finding that PPIs can inversely regulate genes related to cell cycle in the EoE transcriptome and to counteract IL-13—mediated proliferation in culture expands PPI responses in EoE beyond simply regulating eosinophilic infiltration.⁶ Moreover, the contribution of other biological processes directly regulated by PPIs, such as lipid metabolism and unfolded protein responses, are likely critical for understanding the beneficial role of PPIs in the treatment of EoE. Notably, metallothionein transcripts, which are some of the most highly upregulated genes induced by PPIs, encode for proteins that have immunomodulatory activities.³⁷ In summary, our results demonstrate that the esophageal epithelium is a critical target for PPIs in EoE and that PPIs regulate biological processes germane to EoE pathogenesis, including cell proliferation and metabolism. These findings call attention to further understanding the pathways that are regulated by PPIs, as they likely underlie those involved in EoE pathogenesis, including AHR signaling.