Tissue-specific inducible IL-33 expression elicits features of eosinophilic esophagitis

Abstract

Background:

IL-33 is a type 2 inflammatory cytokine that is elevated in the esophageal epithelium of EoE subjects. We previously developed a mouse model of EoE dependent on constitutive overexpression of IL-33 from the esophageal epithelium (EoE33).

Objective:

Our objective was to develop an inducible, IL-33-dependent model of EoE and examine induction of EoE-associated pathology.

Methods:

We utilized a tetracycline-inducible system to express IL-33 in the esophagus by generating two transgenic mice. The first (iSophagus) expresses a reverse tetracycline transactivator (rtTA) from the esophageal epithelium. The second (TRE33) features a tetracycline-response element driving expression of IL-33. When crossed, these mice generate an inducible model of EoE (iEoE33). Mice were administered doxycycline-infused chow for up to 2 weeks. Cytokines were assessed by ELISA or bead-based multiplex. T cells were assessed by flow cytometry. Pathology was assessed by histology and immunohistochemistry for IL-33, eosinophil peroxidase, CD4, and Ki-67. iEoE33 was treated with steroids and crossed with IL-13^−/− mice. For detailed Methods, please see the Methods section in this article’s Online Repository at www.jacionline.org.

Results:

Doxycycline-treated iEoE33 mice demonstrated expression of IL-33 in the esophageal epithelium, and esophageal pathology including eosinophilia, CD4⁺ cell infiltrate, basal zone hyperplasia, and dilated intercellular spaces. These findings became pronounced on day 7 of induction, were accompanied by weight loss and esophageal thickening, and were steroid responsive and IL-13 dependent.

Conclusion:

Inducible IL-33 expression in the esophageal epithelium elicited features pathognomonic of EoE. iEoE33 enables investigation of EoE disease mechanisms as well as initiation, progression, and resolution.

Capsule Summary:

Inducible expression of IL-33 in the mouse esophagus elicits clinical and pathologic features of EoE “iEoE33”. iEoE33 pathology is steroid responsive and IL-13 dependent.

Graphical Abstract

Introduction

Eosinophilic esophagitis (EoE) is a chronic inflammatory disease of the esophagus that has markedly increased since the late 1990s. It is the leading cause of food impactions in adults¹ and a common cause of feeding disturbance in children.² EoE is linked to immune responses to food and/or environmental antigens and is defined by the histologic appearance of eosinophils in the esophagus. It is a disease characterized by increased levels of type 2 cytokines in the esophageal mucosa, which are likely critical for promoting cellular responses and driving pathological and functional changes in the esophagus.

IL-33 is an epithelium-derived alarmin associated with type 2 immune responses and is increased in esophageal biopsies from EoE subjects.^3–5 IL-33 release is associated with epithelial damage and stress.⁶ Notably, environmental factors such as sodium dodecyl sulfate can induce IL-33 expression, possibly through oxidative stress pathways.^{7, 8, 9–11} IL-33 signals through the ST2 (IL-1RL1) receptor and can promote IL-13 production by ST2 expressing cells including eosinophils, ILC2s, mast cells and Th2 cells.^12–14 We have recently shown that constitutive overexpression of a secreted and active form of IL-33 from the esophageal epithelium in “EoE33” mice promoted a robust phenotype pathognomonic of EoE.¹⁵ Observed features included eosinophilic, Th2 cell, and mast cell inflammation in the esophagus accompanied by increased levels of IL-13. Tissue remodeling changes included epithelial hyperplasia and dilated intercellular spaces (DIS). Notably, EoE33 mice mounted an adaptive immune response against wheat, a component of their chow. The EoE pathologies occurred independent of eosinophils but were dependent on IL-13 and were steroid responsive.

Little is known regarding how EoE is initiated and progresses to fibrostenosis. Symptoms display age-dependent heterogeneity and range in severity from mild dysphagia to food impaction. The EoE33 model showed esophageal pathology comparable to human EoE, however, IL-33 was constitutively overexpressed. To enable investigation of disease initiation, progression, and resolution, we modified the EoE33 model so that IL-33 expression is turned on and off. Specifically, we have developed tetracycline-inducible¹⁶ esophageal-specific IL-33 transgenic mice (i.e., iEoE33). Herein, iEoE33 mice reveal the kinetics and pathogenesis of EoE.

Results and Discussion

Tetracycline-inducible esophageal-specific expression of IL-33 was achieved by generating and crossing two transgenic mice. The gene construct for the first mouse utilizes the Epstein-Barr virus–derived ED-L2 promoter to drive expression of a reverse tetracycline transactivator (rtTA) (Figure 1A). The ED-L2 promoter confers specificity to the squamous epithelium and the rtTA enables tetracycline-inducible expression when paired with a tetracycline response element (TRE) in the presence of tetracycline or the tetracycline analog, doxycycline (dox). We call the resulting mouse “iSophagus”. A similar transgenic mouse has previously been reported.¹⁷ The gene construct for the second mouse utilizes the TRE promoter driving expression of a secreted and active IL-33 transgene (Figure 1B). We call the resulting mouse “TRE33”. The secreted and active (sa) IL-33 construct was previously utilized in conjunction with the ED-L2 promoter for constitutive expression in the esophagus.¹⁵

Figure 1: iEoE33 mouse model.

(A) The Epstein-Barr virus-derived promoter, ED-L2, was used to drive expression of reverse tetracycline transactivator (rtTA) from the esophageal epithelium. (B) A tetracycline response element (TRE) was used to drive expression of secreted and active IL-33 (sa-IL-33). (C) Immunocytochemical staining of IL-33 (red) on transfected 293T cells with expression of a reverse tetracycline transactivator (pTet) and either saIL-33 (pTRE-sa-IL-33) or empty vector (pTRE) ± doxycycline (dox). (D) ELISA of eosinophil (eos)-derived IL-13 following in vitro culture with conditioned media (CM) from cell cultures shown in panel C (1–4). (E) An inducible model of EoE (iEoE33) was generated by crossing iSophagus (ED-L2-rtTA) with TRE33 (TRE-sa-IL-33) mice. In the presence of dox, expression of secreted and active IL-33 is induced in the esophageal epithelium. (F) IL-33 expression by ELISA at day 5 of induction with different concentrations of doxycycline. (G) Representative IL-33 immunohistochemistry staining (brown) of esophageal cross sections from wild-type (WT) and iEoE33 mice ± 2 weeks dox treatment. (H) IL-33 ELISA of whole tissue homogenates from WT and iEoE33 mice. (I) Esophageal diameter measurements from WT and iEoE33 mice. (J) Change in body weight measurements from WT and iEoE33 mice. n = 3–20 mice per group. Data are shown as mean ± SEM. Two-way ANOVA with a Tukey’s multiple comparison test (D, H, I, J). Student’s t-test (F) *p<0.05, **p<0.01, ****p<0.0001 compared to all groups. Scale bar = 200 μm.

To examine the functionality of the TRE33 transgene in vitro, we transfected 293T cells. In the presence of a transactivator (pTet) plus dox, TRE33 transfected (pTRE-sa-IL-33) 293T cells produced and released IL-33 as demonstrated by immunohistochemical staining (Figure 1C). The biological activity of released IL-33 was confirmed by examining IL-13 production by eosinophils cultured in the presence of conditioned media (CM) representing each condition (Figure 1D). iSophagus and TRE33 mice were generated following a standard transgenic protocol.¹⁶ iEoE33 mice were generated by crossing iSophagus with TRE33 mice (Figure 1E). To examine IL-33 expression and titratability in iEoE33 mice, we assessed whole esophageal homogenates after 5 days of induction with varied concentrations of doxycycline (Figure 1F). iEoE33 mice (8 wks of age) administered dox for two weeks (625mg/kg in food ad libitum) exhibited epithelial IL-33 staining in the esophagus (Figure 1G). Induced iEoE33 mice also showed elevated IL-33 in whole esophageal homogenates (Figure 1H) as well as in the tongue and forestomach, consistent with the expression pattern of the ED-L2 promoter.¹⁵ Marked gross esophageal thickening and weight loss were also observed (Figure 1I, J).

iEoE33 mice fed dox chow showed histopathologic changes consistent with EoE (Figure 2A). Eosinophil peroxidase (EPX) immunohistochemistry (IHC) revealed extensive eosinophil infiltrate (EI), CD4 IHC showed CD4⁺ cell infiltrate, and Ki-67 IHC showed basal zone hyperplasia (BZH). iEoE33 mice were evaluated using the EoE Histologic Scoring System (EoEHSS) adapted for mice.¹⁵ EoEHSS scores were increased in dox-fed iEoE33 mice for EI, eosinophilic abscesses (EA), eosinophilic surface layering (SL), DIS, BZH, and lamina propria fibrosis (LPF) (Figure 2B). Type 2 inflammation is associated with EoE, and we observed increased Th2 cells (CD4⁺, ST2⁺, CD69⁺) by flow cytometry in dox-fed iEoE33 mice (Figure 2C). Increased IL-4, 5, 9, and 13 as well as TARC and eotaxin-1 were found in esophageal homogenates of dox-fed iEoE33 mice (Figure 2D). These data are consistent with our EoE33 (constitutive) model.¹⁵ However, unlike the constitutive model, mast cells were not increased (data not shown), suggesting this is a feature that takes longer than two weeks to develop. Together, these data suggest that induction of IL-33 expression from the esophageal epithelium in dox-fed iEoE33 mice can reproduce key features of EoE pathology.

Figure 2: Doxycycline-treated iEoE33 mice exhibited features pathognomonic of eosinophilic esophagitis.

(A) Representative hematoxylin & eosin (H&E), eosinophil peroxidase (EPX (red)), CD4 (brown), and Ki-67 (brown) staining of esophageal cross sections from wild-type (WT) and iEoE33 mice ± 2 weeks doxycycline (dox) treatment. (B) H&E-stained esophageal biopsies from WT - dox and iEoE33 + dox mice were assessed using the EoE Histologic Scoring System (EoEHSS). Mean EoEHSS component and composite scores are shown. EI, eosinophil infiltration, EA; eosinophil abscess, SL; eosinophil surface layering, DIS; dilated intercellular spaces; BZH; basal zone hyperplasia, and LPF; lamina propria fibrosis. H&E-stained esophageal biopsies from WT - dox and iEoE33 + dox mice were assessed for peak eosinophils per high-power field (eos/hpf) in the epithelium and stroma. (C) Flow cytometry plot and quantification of esophageal Th2-type CD4⁺ T cells in WT - dox and iEoE33 + dox mice. Frequencies shown as a percentage of live, CD45⁺, CD3⁺, CD4⁺ population. (D) Esophageal type 2 cytokine levels in WT - dox and iEoE33 + dox mice. n = 4–6 mice per group. Data are shown as mean ± SEM. Student’s t-test or Mann-Whitney test depending on data distribution. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Scale bar = 200 μm. Inset scale bar = 100 μm. ND – not detected.

By leveraging induced expression of IL-33, progression of esophageal pathology in iEoE33 mice was assessed at multiple time points following dox administration. Dox-treated iEoE33 mice demonstrated expression of secreted IL-33 in the esophageal epithelium by day 1 (Figure 3A). BZH and stromal eosinophilia were observed by day 5 (Figure 3A, B). CD4⁺ cell infiltration was accompanied by striking tissue pathology on day 7 including BZH, DIS, and EA and SL (Figure 3A, B). Interestingly, EA and SL appeared to peak on day 7. EoE-associated pathologies continued through day 14 and were accompanied by increased esophageal thickening (Figure 3C) and weight loss (Figure 3D). Esophageal homogenates showed detectable levels of IL-33 and IL-13 on day 1 that increased through day 14 (Figure 3E, F).

Figure 3: iEoE33 induction kinetics.

(A) Hematoxylin & eosin (H&E), IL-33, eosinophil peroxidase (EPX), CD4, and Ki-67 staining of esophageal cross sections from doxycycline (dox)-treated iEoE33 mice over time. (B) H&E-stained esophageal biopsies from iEoE33 + dox mice were assessed using the EoE Histologic Scoring System (EoEHSS) for grade and stage. Mean EoEHSS component scores and composite scores are shown. EI, eosinophil infiltration, EA; eosinophil abscess, SL; eosinophil surface layering, DIS; dilated intercellular spaces; BZH; basal zone hyperplasia, and LPF; lamina propria fibrosis. Interepithelial and stromal eosinophils/hpf, CD4+ cells/section, and Ki-67+ nuclei/mm basement membrane were counted. (C) Esophageal diameter measurements from WT - dox and iEoE33 + dox mice. (D) Change in body weight measurements from WT - dox and iEoE33 + dox mice, repeated measures ANOVA showed significant decreases in body weight in iEoE33 + dox mice; F(1;8) = 34.85; p=0.0004. (E) IL-33 ELISA and (F) IL-13 ELISA of whole tissue homogenates from WT - dox and iEoE33 + dox mice. n = 4–5 mice per group. Data are shown as mean ± SEM. Student’s t-test or Mann-Whitney test depending on data distribution. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Scale bar = 200 μm. ns – not significant. ND – not detected.

Glucocorticoids and anti-IL-4rα¹⁸ (shared IL-4 and IL-13 receptor) are effective treatment options for EoE.¹⁹ We evaluated iEoE33 steroid responsiveness and IL-13 dependency by treating iEoE33 mice with dexamethasone and crossing iEoE33 with IL-13^−/− mice, respectively. iEoE33 mice treated with dexamethasone for two weeks concurrent with dox administration showed normal esophageal histology compared to saline-treated, dox-fed iEoE33 mice (Figure 4A, B). Esophageal thickening was reduced, and IL-13 levels were comparable to WT mice (Figure 4C, D). Similarly, crossing iEoE33 with IL-13^−/− mice also resulted in reduced pathology, although mild BZH and stromal eosinophils were observed in the absence of IL-13 (Figure 5).

Figure 4: Induced iEoE33 mice were responsive to steroid treatment.

(A) Representative hematoxylin & eosin (H&E), eosinophil peroxidase (EPX (red)), CD4 (brown), and Ki-67 (brown) staining of esophageal cross sections from doxycycline (dox) treated wild-type (WT) and iEoE33 mice that were concurrently treated with saline (sal) or dexamethasone (dex) for 2 weeks. (B) H&E-stained esophageal biopsies from dox-treated WT and iEoE33 mice also treated with sal or dex were assessed using the EoE Histologic Scoring System (EoEHSS). Esophageal cross sections were scored for both degree (grade) and extent (stage) of pathology. Mean EoEHSS component scores and composite scores are shown. EI, eosinophil infiltration, EA; eosinophil abscess, SL; eosinophil surface layering, DIS; dilated intercellular spaces; BZH; basal zone hyperplasia, and LPF; lamina propria fibrosis. H&E-stained esophageal cross sections were assessed for peak eosinophils per high-power field (eos/hpf) in the epithelium and stroma. (C) Esophageal diameter measurements from dox-treated WT and iEoE33 mice also treated with sal or dex. (D) IL-13 ELISA from whole esophagus homogenates. n = 3 – 8 mice per group. Data are shown as mean ± SEM. Two-way ANOVA with a Tukey’s multiple comparison test. **p<0.01 compared to all groups. Scale bar = 200 μm. ND – not detected.

Figure 5: iEoE33 pathologies were dependent on IL-13.

(A) Representative hematoxylin & eosin (H&E), eosinophil peroxidase [EPX (red)], CD4 (brown), and Ki-67 (brown) staining of esophageal cross sections from 2 weeks doxycycline (dox)-treated wild-type (WT), IL-13^−/−, iEoE33, and iEoE33 x IL-13^−/− mice. (B) H&E-stained esophageal biopsies from dox-treated WT, IL-13^−/−, iEoE33, and iEoE33 x IL-13^−/− mice were assessed using the EoE Histologic Scoring System (EoEHSS). Esophageal cross sections were scored for both degree (grade) and extent (stage) of pathology. Mean EoEHSS component scores and composite scores are shown. EI, eosinophil infiltration, EA; eosinophil abscess, SL; eosinophil surface layering, DIS; dilated intercellular spaces; BZH; basal zone hyperplasia, and LPF; lamina propria fibrosis. H&E-stained esophageal cross sections were assessed for peak eosinophils per high-power field (eos/hpf) in the epithelium and stroma. (C) Esophageal diameter measurements from dox-treated WT, IL-13^−/−, iEoE33, and iEoE33 x IL-13^−/− mice. (D) IL-13 ELISA from whole esophagus homogenates. n = 3 – 9 mice per group. Data are shown as mean ± SEM. Two-way ANOVA with a Tukey’s multiple comparison test. ***p<0.001, ****p<0.0001 compared to all groups. Scale bar = 200 μm. ND – not detected.

We have generated an inducible esophageal epithelium-specific IL-33 transgenic mouse that models EoE. The possibility of on/off expression of IL-33 in the esophagus of iEoE33 mice complements our previously published constitutive (chronic) model and provides an alternative to acute studies utilizing i.p. administration of recombinant IL-33.⁴ Moreover, the iSophagus and TRE33 mice used to create iEoE33 provide valuable and versatile tools for research. For example, iSophagus may be used to drive inducible expression of other mediators of interest (e.g., TSLP) from the esophageal epithelium. TRE33 may be used to investigate the activities of IL-33 in numerous conditions (e.g., asthma or atopic dermatitis) by crossing with alternative tissue- or cell-specific transactivator mice. While our knowledge of disease induction timing in humans is limited to activation of adaptive immunity upon food trigger reintroduction, it is interesting to note that, similar to humans²⁰, iEoE33 mice develop intraepithelial eosinophilia by day 7.

Limitations of the iEoE33 model include IL-33 expression in the tongue and forestomach. These tissues include squamous epithelium where the ED-L2 promoter is active. The mouse forestomach is prone to human pathologies unique to the distal esophagus such as intestinal metaplasia (aka Barrett’s esophagus) as demonstrated in the ED-L2-promoter-driven IL-1β transgenic mice.²¹ Thus, observed EoE-like changes in the forestomach may represent certain human EoE features more prevalent in the distal esophagus.²² In addition, the ED-L2 promoter is more active in the parabasal and suprabasal epithelial cell layers of the mouse esophagus as suggested by transgenic IL-33 localization (Figure 2A), while esophageal IL-33 is increased in the basal zone in patients with active EoE.⁵ Crossing a basal epithelium (e.g., Krt5) promoter-driven rtTA mouse with TRE33 could address this limitation.²³

Future directions include kinetic studies examining the roles of various cell types and mediators (e.g., eosinophils, ILC2s, mast cells, T cells, IL-13) in the initiation, progression, and reversibility of EoE. Our findings suggest a pivotal role for IL-13 in mediating the immunopathology of EoE (Figure 5). Additional studies are required to define the cause and relevant sources of increased IL-13 expression, which may vary over time. The constitutive EoE model (EoE33) develops sensitivity to wheat in mouse chow. We did not observe this during the two-week induction protocol in iEoE33. However, future investigations include longer duration and intermittent dox exposure to examine the development of food sensitization and its potential to drive EoE pathogenesis. Notably, the IL-33 expression may be fine-tuned by titrating the dose of doxycycline (Figure 1F).^{24, 25} We speculate the early histopathologic changes in iEoE33 mice are driven by IL-33 activation of innate immune cells (e.g., ILC2s and mast cells), which, in turn, promote antigen sensitization and chronic adaptive immune responses perpetuated by local T resident effector memory cells.

In summary, dox-treated iEoE33 mice demonstrate robust EoE-like pathology that is responsive to steroids and dependent on IL-13. Induction kinetics revealed BZH and stromal eosinophilia beginning at day 5 followed by EA and SL at day 7 along with a CD4⁺ cell infiltrate. iEoE33 and its constituents iSophagus and TRE33 are powerful new tools for dissecting the disease mechanisms of EoE as well as other esophageal and IL-33-associated diseases.

Key Messages:

• iEoE33 is a novel and robust inducible mouse model of eosinophilic esophagitis (EoE) that enables mechanistic studies and enhanced understanding of disease induction, progression, and resolution.

• iEoE33 mice exhibit hallmark disease features of EoE including esophageal type 2 inflammation and tissue remodeling.

• iEoE33 pathology is steroid responsive and IL-13 dependent.

Abbreviations:

BZH

basal zone hyperplasia

CM

conditioned media

DIS

dilated intercellular spaces

dex

dexamethasone

dox

doxycycline

EA

eosinophilic abscess

EI

eosinophilic inflammation

EoE

eosinophilic esophagitis

eos

eosinophils

eos/hpf

eosinophils per high-power field

EoEHSS

Eosinophilic Esophagitis Histologic Scoring System

EPX

eosinophil peroxidase

H&E

hematoxylin and eosin

i.p.

intraperitoneal

LPF

lamina propria fibrosis

ND

not detected

p

plasmid

p(A)

poly(A)

pTet

transactivator plasmid

pTRE

tetracycline response element plasmid

rtTA

reverse tetracycline transactivator

sa

secreted and active

sal

saline

SEM

standard error of the mean

SL

eosinophil surface layering

Tet

tetracycline

TRE

tetracycline response element

WT

wild type