IL-33 dysregulates regulatory T (Treg) cells and impairs established immunological tolerance in the lungs

Abstract

Background

Airway exposure to environmental antigens generally leads to immunological tolerance. A fundamental question remains: why is airway tolerance compromised in patients with allergic airway diseases? Interleukin (IL)-33 promotes innate and adaptive type 2 immunity and may provide the answer to this question.

Objective

The goal of this study was to investigate the roles IL-33 plays in altering regulatory T (Treg) cells in the lungs and in affecting previously established airway immunological tolerance.

Methods

We analyzed CD4⁺ forkhead box p3 (Foxp3)⁺ Treg cells that were isolated from the lungs of naïve BALB/c mice and those treated with IL-33. Airway tolerance and allergen-induced airway inflammation models in mice were used to investigate how IL-33 affects established immunological tolerance in vivo.

Results

CD4⁺Foxp3⁺ Treg cells in the lungs expressed IL-33 receptor ST2. When exposed to IL-33, Treg cells upregulated their expression of canonical Th2 transcription factor GATA3 as well as ST2, and produced type 2 cytokines. Treg cells lost their ability to suppress effector T cells in the presence of IL-33. Airway administration of IL-33 with an antigen impaired the immunological tolerance in the lungs that had been established by prior exposure to the antigen. Dysregulated Foxp3⁺ Treg cells with distinct characteristics of Th2 cells increased in the lungs of the mice undergoing IL-33-dependent allergen-driven airway inflammation.

Conclusions

IL-33 dysregulated lung Treg cells and impaired immunological tolerance to inhaled antigens. Established airway tolerance may not be sustained in the presence of an innate immunological stimulation, such as IL-33.

Graphical abstract

INTRODUCTION

Respiratory exposure to innocuous antigens generally results in immunologic tolerance.¹ On the other hand, patients with allergic airway diseases show chronic and persistent inflammatory responses to environmental antigens² that are mediated by Th2-type CD4⁺ T cells.^{3, 4} A novel subset of memory CD4⁺ cells that produces both Th2-type cytokines and interleukin (IL)-17 has also been identified in patients with severe asthma.⁵ Thus, a fundamental question remains as to why airway tolerance is compromised in patients with allergic airway diseases. Regulatory T (Treg) cells that develop in the peripheral mucosal tissues protect the individual from Th2-type inflammation in the gastrointestinal tract and lungs.⁶ Indeed, impaired development of antigen-specific forkhead box P3 (Foxp3)⁺ Treg cells is implicated in loss of immunological tolerance to inhaled allergens.^{7, 8} Other studies also suggest that Treg cells are involved in controlling the development and resolution of allergic diseases.^9–13 Accordingly, the balance between the activation of Th2-type immunity and the regulatory activity of Treg cells may play a pivotal role in the evolution of allergic airway diseases.

IL-33, which is produced by the airway epithelium and other cell types, is likely to be critical in mediating type 2 immune responses in patients with asthma.^14–17 Genome-wide association studies of patients with asthma have shown that polymorphisms in IL33 and IL1RL1 (i.e., ST2, the IL-33 receptor) are associated with an increased susceptibility to asthma.^18–21 IL-33 activates a variety of cell types that are implicated in allergic airway diseases, such as Th2-type CD4⁺ T cells, type 2 innate lymphoid cells (ILC2s), mast cells, and eosinophils.²² In the murine model of asthma, IL-33 induces Th2-type differentiation of naïve CD4⁺ T cells and promotes production of IL-5 and IL-13, hence amplifying airway hyperresponsiveness and eosinophilic airway inflammation.¹⁶ When added to resting Th2 cells together with signal transducer and activator of transcription 5 (STAT5)-activating cytokines, IL-33 enhances their expression of ST2.¹⁷ IL-33 also mediates development of highly pathogenic Th2-type T cells that produce a large quantity of IL-5.²³ However, little information is currently available regarding the effects of IL-33 on Treg cells.

While the immune suppressive function of Treg cells has been well established, recent studies have recognized that Treg cells are plastic and demonstrate tissue-specific alteration.^{24, 25} For example, Treg cells that express the canonical transcription factor Foxp3 have the propensity to co-express retinoic acid receptor-related orphan receptor-γt (RORγt) and differentiate into Th17-type cells in the inflamed intestine.^26–28 Similarly, Foxp3⁺ Treg cells that are recruited to a site of Th1-type inflammation express T-bet and produce interferon (IFN)-γ.²⁹ More recently, “Th2 cell-like” Treg cells have been identified in the intestine and secondary lymphoid organs in a mouse model of food allergy involving a gain-of-function IL-4Rα chain allele.⁹ In humans, Treg cells that express type 2 cytokines, such as IL-4 and IL-13, were detected in the skin of patients with systemic sclerosis.³⁰ Thus, Treg cells are likely altered when influenced by certain tissue microenvironments. However, our knowledge about Treg-cell plasticity in allergic airway diseases and their models and regulation of that plasticity is limited.

Accordingly, to fill these major gaps in our knowledge, we investigated the roles of IL-33 in controlling Treg cells. Our observations suggest that IL-33 alters lung Treg cells and impairs airway tolerance to airborne allergens. Hence, in addition to their established effects on Th2-type effector T cells and ILC2s, IL-33 may promote type 2 airway inflammation by modulating mucosal Treg cells.

MATERIALS AND METHODS

See the Methods section of this article’s Online Repository for more details.

Mice

BALB/c and BALB/c-Foxp3^EGFP (C.Cg-Foxp3^tm2Tch/J) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). BALB/c-Foxp3^EGFP, BALB/c-St2^−/−, and BALB/c-Foxp3^EGFPSt2^−/− mice were bred and maintained under specific pathogen-free conditions at the Mayo Clinic. Male and female mice ages 10 to 12 weeks were used in all of the experiments. All experiments were approved and conducted in accordance with the guidelines of the Mayo Clinic Institutional Animal Care and Use Committee, Mayo Clinic, Rochester, MN.

Fluorescence-activated cell sorting (FACS) analyses of lung and splenic Treg cells

Lungs were harvested from naïve Foxp3^EGFP reporter mice or Foxp3^EGFPSt2^−/− reporter mice that were euthanized by an overdose of pentobarbital. Alternatively, Foxp3^EGFP mice were exposed intranasally to phosphate-buffered saline (PBS) or IL-33 (100 ng/dose, eBioscience, San Diego, CA) on days 0, 2, and 4. Twelve to twenty-four hours after the last exposure, mice were euthanized, and lungs were collected. Single cell suspensions from lungs and spleens were preincubated with Fc-receptor blockers for 30 minutes at 4°C and stained with appropriate fluorochrome-conjugated antibodies.

Sorting of lung Treg cells and culture

The entire population of CD4⁺Foxp3eGFP⁺ Treg cells or ST2⁻ and ST2⁺ subpopulations of CD4⁺Foxp3eGFP⁺ Treg cells were isolated by sorting. Briefly, CD4⁺ T cells were enriched from the single-cell suspensions of lungs and spleen by EasySep CD4⁺ cell isolation kit (StemCell Technologies, Vancouver, BC, Canada). After staining, Treg cells were sorted using a FACSAria™ flow cytometer (BD Biosciences, San Jose, CA). To examine cytokine protein production, sorted lung Treg cells were cultured in round-bottomed 96-well plates at 4x10⁵ cells/ml for up to 96 hours and stimulated with 10 ng/ml PMA plus 500 ng/ml ionomycin in complete culture medium for 6 to 8 hours.

Gene expression analyses

Cells were lysed in Trizol (Invitrogen, Carlsbad, CA), and total RNA was extracted and purified using the PureLink® RNA Mini Kit (ThermoFisher Scientific, Waltham, MA) followed by real-time quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) analyses. The relative expression level was determined via the 2-ΔΔCT method.³¹ CD4⁺Foxp3eGFP⁺ Treg cells or CD4⁺Foxp3eGFP⁻ T cells that were isolated from the lungs of naïve non-treated Foxp3^EGFP reporter mice by FACS sorting were used as controls throughout the study.

Analysis of Treg cell function

Treg (CD4⁺Foxp3eGFP⁺) cells and T responder (CD4⁺Foxp3eGFP⁻) cells were isolated from spleens of naïve BALB/c Foxp3^EGFP reporter mice by sorting as described above. T responder cells were labeled with Violet Cell Tracer dye (ThermoFisher Scientific, Life Technologies, Waltham, MA). T responder cells were incubated in 48-well plates at 1×10⁶ cells/ml in the presence of anti-CD3/anti-CD28 Dynabeads (20 μl of microbeads in 1.0 ml, ThermoFisher Scientific) with or without Treg cells at a ratio of 1:1. IL-33 at 40 ng/ml was added to some cultures. After a 4-day incubation, the proliferation of T responder cells was analyzed by flow cytometry.

Airway tolerance model

Endotoxin-free ovalbumin (OVA) was prepared as previously described.³² Airway immunologic tolerance to OVA was induced by using a similar protocol as previously described.^{7, 33} Briefly, endotoxin-free OVA (100μg/dose) was intranasally administrated to naïve BALB/c mice from day -12 to day -7. To verify the state of immunologic tolerance, OVA (10 μg/dose, Sigma-Aldrich, St. Louis, Mo) plus alum were injected intraperitoneally on day 7 and day 14. Control mice received sterile PBS plus alum. Peripheral blood was collected on day 28 to examine the plasma levels of IgE antibodies to OVA. On days 29, 30, and 31, mice were challenged intranasally with endotoxin free-OVA (100 μg/dose). On day 32, bronchoalveolar lavage (BAL) fluids and lungs were collected for analyses.

Allergen-induced airway inflammation model

To examine the effects of exposure to natural allergens, mice were exposed to extracts of natural allergens as described previously.³² Briefly, naïve wild-type (WT) BALB/c mice, BALB/c-St2^−/− mice, or BALB/c-Foxp3^EGFP reporter mice were exposed intranasally to a mixture of allergen extracts [10 μg each of Alternaria alternata, Aspergillus fumigatus, and house dust mite (HDM) (all from Greer Laboratories, Lenoir, NC)] plus 10 μg endotoxin-free OVA in 50 μl PBS, three times a week, for up to 4 weeks.

ELISA for cytokine and antibody levels

The levels of IL-4, IL-5, IL-13, IL-17, and IFN-γ in the supernatants of Treg cells and BAL fluids were measured using ELISA kits (R&D Systems, Minneapolis, MN). The plasma levels of OVA-specific IgE were measured using sandwich ELISA as previously described.³⁴

Statistical analyses

Data are shown as the mean ± standard error of the mean (SEM) for the numbers of mice or experiments indicated. All statistical analyses were assessed with paired or unpaired Student’s t-test as appropriate, or with one-way analysis of variance followed by Bonferroni multiple comparisons test if comparing more than two groups. The statistical analyses were performed using Prism (GraphPad Software, La Jolla, CA). Statistical significance was defined as p < 0.05.

RESULTS

CD4⁺Foxp3⁺ Treg cells in the lungs expressed IL-33 receptor ST2

Recent studies show that a significant proportion of intestinal Foxp3⁺ Treg cells co-express the canonical Th2 transcription factor GATA3,^35–38 which is known to upregulate expression of IL-33 receptor ST2 in Th2-type CD4⁺ T cells.¹⁷ To examine whether ST2 is expressed in Treg cells in the lungs, we analyzed CD4⁺Foxp3⁺ Treg cells in naïve BALB/c Foxp3^EGFP reporter mice. Foxp3^EGFPSt2^−/− mice were used as a control. ST2 expression was minimal on CD4⁺Foxp3⁻ conventional T cells in both spleen and lungs of naïve mice (Figure 1A). In contrast, a fraction of CD4⁺Foxp3⁺ Treg cells in the lungs, but not those in spleen, expressed ST2. As expected, no ST2 expression was detectable in lung Foxp3⁺ Treg cells from Foxp3^EGFPSt2^−/− mice. The histograms gated on Foxp3⁻ conventional T cells and Foxp3⁺ Treg cells revealed that the ST2 expression level was clearly higher in lung Foxp3⁺ Treg cells as compared to splenic Treg cells (Figure 1B); some expression of ST2 was observed in conventional CD4⁺ T cells (i.e., Foxp3⁻ cells) in the lungs as well. The quantitative analyses showed that proportions of ST2⁺ cells in both Treg cell and conventional CD4⁺ T cell populations were significantly higher in the lungs as compared to spleen (Figure 1C, p<0.01).

Figure 1. CD4⁺Foxp3⁺ Treg cells in the lungs expressed IL-33 receptor ST2.

(A) ST2 expression on CD4⁺Foxp3⁺ Treg cells was examined in spleen and lungs of BALB/c-Foxp3^EGFP reporter mice. Foxp3^EGFPSt2^−/− mice were used as a control. (B) ST2 expression in CD4⁺Foxp3eGFP⁺- and CD4⁺Foxp3eGFP⁻-cell populations is shown as histograms. In (A) and (B), data are representative of three independent experiments. (C) Frequencies of ST2⁺ cells within the CD4⁺Foxp3eGFP⁺ and CD4⁺Foxp3eGFP⁻ populations in spleen and lungs are shown. Data are shown as the mean ± SEM from four mice. ** p<0.01, compared to spleen. (D) ST2⁺CD4⁺Foxp3eGFP⁺ and ST2⁻CD4⁺Foxp3eGFP⁺ cells were isolated by sorting from the lungs of naïve Foxp3^EGFP mice and cultured with medium or IL-33 for 24 hours. St2 mRNA expression was examined by real-time qRT-PCR and normalized to its expression in CD4⁺Foxp3eGFP⁻ cells isolated from naïve Foxp3^EGFP reporter mice. Data are shown as the mean ± SEM from three mice. *p<0.05, **p<0.01 between the groups indicated by horizontal lines.

We verified this observation by examining mRNA expression. Previous studies also showed that IL-33 together with STAT5-activating cytokines enhances the expression of ST2 in resting Th2 cells.¹⁷ We therefore sorted ST2⁻ Treg cell and ST2⁺ Treg cell populations from the lungs of naïve Foxp3^EGFP mice and cultured them with or without IL-33 for 24 hours in vitro. When cultured with medium alone, ST2 mRNA was highly expressed by ST2⁺ Treg cells as compared to ST2⁻ Treg cells or to naïve Foxp3⁻ CD4⁺T cells (Figure 1D, p<0.01). Incubation with IL-33 significantly increased ST2 mRNA expression in both ST2⁺ and ST2⁻ Treg cell populations, suggesting that the ST2 molecules on Treg cells respond to exogenous IL-33. Altogether, these findings suggest that a proportion of Treg cells in the lungs of naïve BALB/c mice express IL-33 receptor ST2 and respond to IL-33 in vitro.

IL-33 altered lung Treg cells, causing them to become “Th2 cell-like” cells

To elucidate the effects of IL-33 on lung Treg cells, we administered IL-33 intranasally without exogenous antigens to naïve Foxp3^EGFP mice three times over 5 days and isolated them by sorting (Figure 2A). In colon of naïve mice, GATA3 is expressed by an ST2⁺ population of Treg cells.³⁵ In the lungs, small but apparent expression of GATA3 protein was detectable in CD4⁺Fopx3⁺ Treg cells from naïve mice and those treated with PBS (Figure 2B). Administration of IL-33 significantly increased a total number of lung Foxp3⁺ Treg cells by approximately 4-fold (Figure 2B and 2C, p<0.01). Importantly, the expression level of GATA3 protein in Treg cells significantly increased in mice treated with IL-33 (p<0.01), resulting in an approximately 20-fold increase in the number of GATA3⁺ Treg cells as compared to the mice treated with PBS.

Figure 2. IL-33 increased “Th2 cell-like” Treg cells in the lungs in vivo.

(A) Experimental protocol. Naïve Foxp3eGFP⁺ reporter mice were treated with IL-33 (100 ng/dose) or PBS intranasally every other day for 4 days. Twenty-four hours after the last administration, lungs were collected. (B) Lung single cells were stained for Foxp3 and GATA3, and they were analyzed by gating on CD4⁺ T cells. Representative FACS scattergrams are shown. (C) Quantitative analyses of total Foxp3⁺ Treg cells and Foxp3⁺/GATA3⁺ double-positive Treg cells are shown. Data are shown as the mean ± SEM from three mice per group and are representative of two experiments. ** p<0.01 as compared to PBS. (D) CD4⁺Foxp3eGFP⁺ cells were isolated from the lungs of PBS- or IL-33-treated mice by sorting. Messenger RNA expression levels were analyzed by real-time qRT-PCR and were normalized to the levels of CD4⁺Foxp3eGFP⁺ cells from naïve Foxp3^EGFP reporter mice. Data are shown as the mean ± SEM from three mice and are representative of two independent experiments. * p<0.05 as compared to PBS-treated mice. (E) CD4⁺Foxp3eGFP⁺ cells were isolated by sorting and cultured with PMA plus ionomycin for 6 hours. The levels of cytokines in cell-free supernatants were analyzed by ELISA. Date are shown as the mean ± SEM of a pool of two independent experiments, each of which used three mice per group. (F) Lung single-cell suspensions from IL-33-treated Foxp3^EGFP mice (WT mice) and Foxp3^EGFPSt2^−/− mice (ST2^−/− mice) were cultured with PMA plus ionomycin for 6 hours in the presence of GolgiPlug™. Cells were stained for Foxp3 and GATA3, intracellular IL-5 and IL-13, and cell-surface ST2. Representative data from three mice per group are shown.

When we analyzed mRNA expression in isolated cells, no differences were observed in the expression levels of Gata3 and Foxp3 mRNA in CD4⁺Foxp3eGFP⁺ Treg cells from PBS- and IL-33-exposed mice (Figure 2D). In contrast, Il5 and Il13 mRNA were highly expressed by 100- to 1,000-fold in Treg cells from IL-33-exposed animals as compared to naïve or PBS-treated animals. When isolated CD4⁺Foxp3eGFP⁺ Treg cells were further stimulated with PMA plus ionomycin in vitro, significant amounts of IL-5 and IL-13 proteins, but not IL-4, IL-17, or IFN-γ proteins, were detected in the cell-free supernatants of Treg cells from IL-33-exposed mice (Figure 2E, p<0.01). None of these cytokines were detectable in Treg cells from PBS-exposed mice. In contrast, the level of active TGF-β appeared to decrease in IL-33-exposed mice compared to PBS-exposed mice (p=0.06).

To elucidate the source of IL-5 and IL-13 within the lung Treg-cell populations, lung single-cell suspensions from Foxp3^EGFP reporter mice that had been exposed to IL-33 were re-stimulated with PMA and ionomycin in vitro for 6 hours in the presence of GolgiPlug™. The cells were stained for Foxp3 and GATA3, intracellular IL-5 and IL-13, and cell-surface CD4 and ST2 and analyzed by flow cytometry. No or minimal expression of ST2, IL-5, or IL-13 was observed in the GATA3⁻ population of Foxp3⁺ Treg cells (Figure 2F). In contrast, the GATA3⁺ population of Treg cells clearly expressed ST2 on their cell surfaces and produced IL-5 and IL-13. Similar findings were observed in Treg cells from wild-type mice that had been exposed to IL-33 (Figure E1 of Online Repository). These findings suggest that airway exposure to IL-33 increased the numbers of lung GATA3⁺ Treg cells that share characteristics of Th2 cells, including ST2 expression and production of type 2 cytokines.

IL-33 impaired Treg cell-mediated suppression in vitro

The observations described above led us to hypothesize that exposure to IL-33 might alter normal immunological functions of lung Treg cells, such as suppression of effector T cells. To test this hypothesis, we isolated Treg cells (CD4⁺Fopx3eGFP⁺) and T responder cells (CD4⁺Foxp3eGFP⁻) from spleens of naive Foxp3^EGFP mice by sorting. While we recognize that this experiment should ideally be performed by using lung T cells, the poor yield of lung Treg cells from naïve mice did not allow us to perform an in vitro culture experiment. T responder cells were labeled with a cell tracer dye to monitor proliferation, and they were stimulated with anti-CD3 and anti-CD28 microbeads. This strategy was used to rule out the potential effects of IL-33 on antigen-presenting cells (APCs), such as dendritic cells (DCs), that might have occurred if we had used an antigen and APCs. When stimulated with anti-CD3/anti-CD28 without Treg cells, T responder cells proliferated (Figure 3A and 3B); the responder-cell proliferation was not affected by IL-33. When Treg cells were added, proliferation of T responder cells was markedly reduced. Nonetheless, in the presence of IL-33, proliferation of T responder cells was restored, even in the presence of Treg cells. Quantitative analyses showed that the percentage of non-proliferated T responder cells significantly increased in the presence of Treg cells, and the number decreased to the level that is comparable to that without Treg cells in the presence of IL-33 (Figure 3C, p<0.01). These results suggest IL-33 impaired the suppressive ability of Foxp3⁺ Treg cells in vitro, while IL-33 showed minimal effects on proliferation of naïve T responder cells.

Figure 3. IL-33 impaired the suppressive ability of Treg cells.

(A) and (B) CD4⁺Foxp3⁺ Treg cells and CD4⁺Foxp3⁻ T responder cells were isolated by sorting from spleens of naïve Foxp3^EGFP mice. T responder cells were labeled with a tracer dye and stimulated with anti-CD3/anti-CD28 microbeads with or without Treg cells (at a 1:1 ratio) or IL-33 (40 ng/ml) for 4 days. Representative histograms are shown. (C) Frequencies of non-proliferating T responder cells are shown. Data (mean ± SEM) are a pool of four independent experiments. ** p<0.01, between the groups indicated by horizontal lines.

Culture with IL-33 induces “Th2 cell-like” Treg cells in vitro

The above observations led us to speculate that lung Treg cells might respond directly to IL-33, resulting in a “Th2 cell-like” phenotype. To test this, we isolated CD4⁺Foxp3eGFP⁺ Treg cells from the lungs of naïve non-treated Foxp3^EGFP reporter mice by sorting and cultured them in vitro with IL-33 for 96 hours. IL-2 was added to all the cultures to maintain cell viability, but no agonists for T cell receptor were included. No apparent differences were observed in the expression levels of Gata3 and Foxp3 mRNA in Treg cells cultured with or without IL-33 (Figure 4A). In contrast, Il5 and Il13 mRNA were highly expressed in Treg cells cultured with IL-33. When the cells were further stimulated with PMA plus ionomycin, significantly greater amounts of IL-5 and IL-13 proteins, but not IL-4, IL-17, or IFN-γ proteins, were detected in the cell-free supernatants of Treg cells that had been cultured with IL-33 (Figure 4B, p<0.01).

Figure 4. Culture with IL-33 induced “Th2 cell-like” Treg cells in vitro.

CD4⁺Foxp3eGFP⁺ Treg cells were isolated from the lungs of naïve Foxp3^EGFP mice by sorting and were cultured with IL-2 (20 ng/ml) with or without IL-33 (40 ng/ml) for 96 hours. (A) Messenger RNA expression levels were analyzed by real-time qRT-PCR and were normalized to the levels in CD4⁺Foxp3eGFP⁺ cells isolated from naïve non-treated Foxp3^EGFP mice. (B) Cells were stimulated with PMA plus ionomycin for 6 hours. The levels of cytokines in cell-free supernatants were analyzed by ELISA. Data (n=6, mean ± SEM) are a pool of two independent experiments. *p<0.05, **p<0.01, compared to IL-2 alone.

IL-33 broke down established airway tolerance in vivo

Because IL-33 altered lung Treg cells, causing them to become “Th2 cell-like” cells both in vitro and in vivo, and impaired their suppressive function, we hypothesized that IL-33 may impair established airway immunological tolerance that is known to be mediated by Treg cells.^{7, 8} To test this hypothesis, we used an established mouse model of airway tolerance, in which mice are exposed to an innocuous antigen through the airways prior to sensitization.^{7, 33} We exposed naïve BALB/c mice to OVA intranasally for 5 consecutive days; the state of airway tolerance was verified by attempting to sensitize these mice by i.p. injection of OVA plus alum and challenging them by i.n. administration of OVA 3 weeks later (Figure E2A of Online Repository). When mice were previously exposed intranasally to PBS (as a control of tolerance induction), they were sensitized by i.p. injection of OVA plus alum, resulting in increased serum levels of OVA-specific IgE and IgG1 antibodies (Figure E2B). These mice developed marked airway eosinophilia and increased BAL levels of IL-5 and IL-13 when they were challenged by i.n. administration of OVA (Figure E2C). In contrast, when mice were previously exposed intranasally to OVA to induce tolerance, the development of OVA-specific IgE and IgG1 in plasma and type 2 immune responses in the airways were severely impaired, consistent with previous observations.^{7, 33}

To investigate the effects of IL-33 on established airway immunological tolerance, we administered OVA with or without IL-33 intranasally to the mice that had been previously tolerized to OVA (Figure 5A). As described above, without airway tolerance induction (i.n. administration of PBS), mice that were sensitized by i.p. injection of OVA plus alum showed increased plasma levels of anti-OVA IgE, marked eosinophilic airway inflammation, and increased BAL levels of IL-5 and IL-13 when they were subsequently challenged by i.n. OVA (Figure 5B and 5C). In contrast, when mice were tolerized by prior i.n. administration of OVA, i.p. injection of OVA plus alum failed to sensitize those mice, and all the immune responses as described above were significantly reduced or abolished (p<0.05 or p<0.01). Nonetheless, when these previously tolerized mice were exposed to OVA with IL-33 intranasally, they developed anti-OVA IgE antibody in plasma; OVA alone showed minimal effects. When these mice were subsequently challenged by i.n. administration of OVA, they developed robust airway eosinophilia and increased BAL levels of IL-5 and IL-13, higher than those achieved by i.p. sensitization with OVA plus alum in non-tolerized mice.

Figure 5. IL-33 broke down established airway tolerance in vivo.

(A) Experimental protocol. To establish tolerance, naïve BALB/c mice were treated intranasally with endotoxin-free OVA. The state of tolerance was verified by sensitization with i.p. administration of OVA plus alum. Mice were then exposed intranasally to OVA with or without IL-33. The immune responses were examined by analyzing the plasma levels of anti-OVA IgE on day 28 and then challenging mice by i.n. administration of OVA on days 29, 30, and 31. BAL fluids and lungs were collected 24 hours after the last challenge. (B) Plasma levels of anti-OVA IgE are shown. (C) Cell differential counts and cytokine levels in BAL fluids are shown. Data are shown as the mean ± SEM of three to eight mice per group. *p<0.05, **p<0.01, between the groups indicated by horizontal lines.

Because IL-33 can activate ILC2s and mast cells,^39–41 we questioned whether IL-33 alone without antigen is sufficient to break airway immunological tolerance. However, when mice were given IL-33 alone intranasally, the tolerized mice did not develop anti-OVA IgE, airway eosinophilia, or type 2 immune responses (Figure E3 of Online Repository), suggesting that both antigen and IL-33 are concomitantly necessary to impair established tolerance.

Histologic analyses showed that non-tolerized mice that we subsequently sensitized by i.p. OVA plus alum developed peribronchial infiltration of inflammatory cells and mucosal hyperplasia when they were challenged by i.n. OVA (Figure E4 of Online Repository, Group B). The tolerized mice that were subsequently sensitized by i.p. OVA plus alum (Group C) developed minimal pathological changes. In contrast, the tolerized mice that were subsequently exposed to OVA plus IL-33 intranasally (Group D) developed comparable airway pathological changes similar to non-tolerized mice.

A potential caveat of the design of the experiment as described above is that we verified the state of tolerance by i.p. injection of OVA plus alum; this process may favor break down of tolerance when mice are subsequently exposed to OVA with IL-33 intranasally. To examine this possibility, we tolerized mice by i.n. administration of OVA, and then those mice were subsequently exposed to OVA with IL-33 intranasally without prior i.p. administration of OVA plus alum (Figure 6A). As described above, the OVA-tolerized mice that were subsequently sensitized by i.p. injection of OVA plus alum showed minimal plasma levels of IgE antibody and type 2 inflammation in the airways (Figure 6B and 6C). When tolerized mice were subsequently exposed to OVA plus IL-33 intranasally and then challenged by OVA intranasally, they developed significant airway inflammation, consisting of neutrophils and eosinophils, as well as marked increase in BAL levels of IL-5 and IL-13 (Figure 6C). Interestingly, minimal increase in IgE antibody was observed in these mice (Figure 6B). Altogether, these findings suggest that inhalation of antigen together with IL-33 was sufficient to convert previously established airway tolerance to Th2-type immune responses in the airways.

Figure 6. Airway exposure of antigen with IL-33 was sufficient to break down established airway immunological tolerance without systemic administration of antigen with an adjuvant.

(A) Experimental protocol. To establish tolerance, naïve BALB/c mice were treated intranasally with endotoxin-free OVA. Mice were then exposed intranasally to OVA with IL-33 or sensitized by i.p. injection of OVA plus alum. The immune responses were examined by analyzing the plasma levels of anti-OVA IgE on day 28 and then challenging mice by i.n. administration of OVA on days 29, 30, and 31. BAL fluids and lungs were collected 24 hours after the last challenge. (B) Plasma levels of anti-OVA IgE are shown. (C) Cell differential counts and cytokine levels in BAL fluids are shown. Data are shown as the mean ± SEM of three to six mice per group and are representative of two independent experiments. *p<0.05, **p<0.01, between the groups indicated by horizontal lines.

Lung Treg cells demonstrated “Th2 cell-like” characteristics during allergen-induced airway inflammation

The experiments thus far described used exogenous IL-33 to drive Th2-type airway immune responses. A major feature of patients with asthma is chronic and persistent airway inflammation,^42–44 which is likely mediated by exposure to environmental allergens, microbes, or other factors. Therefore, to elucidate whether dysregulated “Th2 cell-like” Treg cells as described above develop in more physiological models, we exposed mice to a mixture of natural airborne allergen extracts, including A. alternata, A. fumigatus, and HDM, together with OVA as described previously³³ (Figure 7A); marked type 2 airway inflammation in this model was dependent on IL-33 and thymic stromal lymphopoietin (TSLP).³² A 2-week exposure of naïve BALB/c Foxp3^EGFP mice to these allergen extracts induced airway inflammation, which was characterized by increased numbers of eosinophils and other inflammatory cells in BAL (Figure 7B). FACS analyses of CD4⁺ T cells in the lungs showed that a proportion (~30%) of Foxp3eGFP⁺ Treg cells expressed ST2 (Figure 7C and 7D). The total number of Treg cells also increased significantly as compared to those exposed to PBS (p<0.01). To characterize these ST2⁺ Treg cells, we sorted them and analyzed them by real-time qRT-PCR. ST2⁺ Treg cells from mice exposed to allergens showed significant increase in Il5 and Il13 mRNA as compared to those from PBS-exposed mice (Figure 7E, p<0.05 and p<0.01). However, no differences were observed in the expression levels of Gata3 and Foxp3 mRNA. These findings suggest that, similar to the results in IL-33-exposed mice (Figure 2), “Th2 cell-like” Treg cells increased in the lungs during allergen-induced airway inflammation.

Figure 7. ST2⁺ Treg cells in the lungs produced Th2-type cytokines during allergic airway inflammation.

(A) Experimental protocol. Naïve BALB/c Foxp3^EGFP reporter mice were exposed to PBS or a cocktail of airborne allergens, including extracts of A. alternata, A. fumigatus, and HDM, and OVA every other day for 2 weeks. BAL fluids and lungs were collected 12 hours after the last exposure. (B) Cell differential counts in BAL fluids are shown. (C) Single-cell suspensions of lungs were stained for CD4 and ST2 and analyzed by FACS. Representative FACS scattergrams are shown. (D) Quantitative analyses of total Foxp3⁺ Treg cells and the ST2⁺ and ST2⁻ Treg-cell subpopulations are shown. (E) The Foxp3eGFP⁺ST2⁺ Treg-cell population was isolated by sorting, and mRNA expression was analyzed by real-time qRT-PCR. The data were normalized to the levels of total Foxp3eGFP⁺ Treg cells from naïve mice. In (B), (D) and (E), data are shown as the mean ± SEM of at least three mice per group and are representative of two independent experiments. *p<0.05, **p<0.01, between the groups indicated by horizontal lines.

To examine the roles of IL-33 in the development of these “Th2 cell-like” Treg cells in the allergen exposure model, we used St2^−/− mice. We exposed naïve WT BALB/c or St2^−/− mice to allergens and analyzed GATA3 protein expression to identify “Th2 cell-like” Treg cells (Figure 7A) because ST2 cannot be used as a marker in St2^−/− mice. In WT mice that were exposed to PBS, approximately 2% of Foxp3⁺ Treg cells were GATA3⁺ (Figure 8B). When exposed to allergens, the total number of Foxp3⁺ Treg cells as well as the proportion of GATA3⁺ cells among the Treg-cell population increased significantly (Figure 8B and 8C, p<0.05). Approximately 10-fold higher numbers of GATA3⁺ Treg cells were observed in mice exposed to allergens as compared to those exposed to PBS. Furthermore, the development of GATA3⁺ Treg cells was nearly eliminated in ST2^−/− mice (p<0.01), suggesting that development of “Th2 cell-like” Treg cells was largely dependent on the IL-33 pathway.

Figure 8. IL-33 signaling was required for the increase in GATA3⁺ Treg cells in the lungs.

(A) Experimental protocol. Naïve wild-type BALB/c mice (BALB/c) or BALB/c St2^−/− mice (ST2^−/−) were exposed to PBS or a cocktail of airborne allergens as described in Figure 7 every other day for 4 weeks. Lungs were collected 24 hours after the last exposure. (B) Single-cell suspensions of lungs were stained for Foxp3 and GATA3 and analyzed by FACS. Representative FACS scattergrams are shown. (C) Quantitative analyses of total Foxp3⁺ Treg cells and the GATA3⁺ population of Treg cells are shown. Date are shown as the mean ± SEM from a pool of two experiments, each of which used three mice per group. *p<0.05, between the groups indicated by horizontal lines.

Finally, we verified the capacity of natural allergens to break established airway tolerance. By using a model analogous to that in Figure 6, we tolerized mice by i.n. administration of OVA, and those mice were subsequently exposed to OVA together with a mixture of allergen extracts (Figure E5A of Online Repository). The OVA-tolerized mice that were subsequently sensitized by i.p. injection of OVA plus alum showed minimal plasma levels of IgE antibody and type 2 inflammation in the airways (Figure E5B and E5C of Online Repository). However, when tolerized mice were subsequently exposed to OVA plus a mixture of allergen extracts intranasally, they developed a robust anti-OVA IgE antibody response. When challenged by OVA intranasally, they developed significant eosinophilic airway inflammation and a marked (significant?) increase in BAL levels of IL-5 and IL-13.

DISCUSSION

Previous studies show that lung-resident macrophages induce Foxp3⁺ Treg cells in the lungs by producing TGF-β and retinoic acid, which promotes airway immunological tolerance to environmental antigens.⁸ Exposure to protease activities in allergens or infection with rhinovirus impairs development of these Treg cells by promoting expression of OX40L by DCs and by creating a lung environment rich in pro-inflammatory mediators, such as TSLP, IL-33, IL-1, IL-6, and TNF-α.^{7, 8, 45} These findings suggest the importance of the balance between TGF-β and pro-inflammatory cytokines/molecules in the development of Treg cells in the lungs. Our findings in this study add to this knowledge by demonstrating that previously established Treg cells and immunological tolerance in the lungs were not sustained in the presence of an innate cytokine, IL-33. Dysregulated Treg cells that share characteristics of Th2 cells also increased in the lungs of mice with allergen-driven IL-33-dependent airway inflammation.

We found that isolated lung Treg cells can be activated directly by IL-33 and enhance their expression of IL-33 receptor ST2 and type 2 cytokines, suggesting self-perpetuating effects of IL-33 on lung Treg cells (Figure 1 and 4). Furthermore, airway exposure to IL-33 without antigens increased Foxp3⁺ Treg cells, in particular those co-expressing canonical Th2 transcription factor GATA3 (Figure 2). While no changes in Gata3 mRNA expression levels were observed, GATA3 can be regulated by posttranscriptional mechanisms, such as phosphorylation of p38⁴⁶ or ERK,⁴⁷ and ubiquitination.⁴⁷ Furthermore, the levels of GATA3 protein are regulated directly by mRNA stability and magnitude of translation.^{48, 49} Therefore, IL-33 may increase the levels of GATA3 protein and/or activate GATA3, resulting in enhanced transcription of type 2 cytokine genes.

The effects of IL-33 on Treg cells have been reported previously in other non-lymphoid organs. For example, Foxp3⁺ Treg cells that express ST2 and GATA3 reside in the intestines of naïve mice, and culture of these cells with IL-33 increases their expression of ST2.³⁵ IL-33 also plays a key role in the accumulation and maintenance of Treg cells in intestinal tissues during chronic colitis caused by infection with Helicobacter hepaticus³⁵ and in injured skeletal muscle tissues.⁵⁰ In visceral adipose tissues, Treg cells expand in response to IL-33 and prevent obesity-associated inflammation.⁵¹ Thus, IL-33 appears to play a major role in expression of ST2 by Treg cells and subsequent recruitment and/or expansion of them in non-lymphoid organs. Our findings of lung Treg cells in mice that were exposed to exogenous IL-33 (Figure 2) or to natural airborne allergens (Figure 8) are consistent with these previous observations.

The biological functions of these IL-33-responsive tissue Treg cells are an active area of investigation. For example, IL-33 promotes repair of chemically-injured skeletal muscle injury, and the process appears to be mediated by a Treg-cell subset producing amphiregulin.^{50, 52} Similarly, in the lungs of mice infected with the influenza virus, amphiregulin derived from ST2-positive IL-33-responsive Treg cells limits lung pathology and improves oxygen exchange.⁵³ Treg cells also inhibit colonic inflammation in an IL-33-dependent manner by as yet unknown mechanisms.³⁵ While detailed information as to whether IL-33-responsive Treg cells produce type 2 cytokines has not been available from these prior studies, we found that GATA3⁺ST2⁺ Treg cells expressed Il5 and Il13 mRNA and produced IL-5 and IL-13 proteins when mice were exposed to IL-33 through the airways (Figure 2). Comparable IL-5/IL-13-competent Treg cells were detected in the lungs of mice exposed to natural airborne allergens (Figures 7 and 8). GATA3 has been shown to be involved in upregulation of a unique set of genes, including Il5, Il13, Il1rl1 (i.e. ST2), and Areg (i.e., amphiregulin) in both ILC2s and Th2-type CD4⁺ T cells;⁵⁴ therefore, expression of amphiregulin (as shown in previous studies) and IL-5/IL-13 (as demonstrated in this study) by tissue-resident GATA3⁺ST2⁺ Treg cells may not be totally unpredictable.

“Th2 cell-like” Treg cells have been reported recently in mice and humans. In a mouse model of food allergy involving a gain-of-function IL-4Rα chain allele, oral administration of an antigen and staphylococcus enterotoxin B induces Treg cells that produce IL-4;⁹ mast cell-derived IL-4 contributes to development of such “Th2 cell-like” Treg cells. Treg cells that produce IL-5 and IL-13 were also identified in the skin of patients with systemic sclerosis.^{9, 30} Thus, Treg cells may be more plastic than previously predicted,²⁵ and their functions and phenotypes can be altered depending on their microenvironment. This may explain the apparent paradox in previous clinical studies. Indeed, large numbers of Treg cells are found at the sites of ongoing inflammation.^55–60 However, in chronic inflammation, whether these Treg cells are able to carry out their suppressive function, whether they are overwhelmed by increased numbers of effector T cells, or even whether they contribute to inflammation remained unclear.

In this study, in the lung tissues of mice that were exposed to IL-33 or natural airborne allergens, the total number of Treg cells increased (Figures 2 and 7). Nonetheless, these Treg cells, in particular those that co-expressed ST2 and GATA3, produced IL-5 and IL-13, suggesting that they may contribute to inflammation. Furthermore, in the presence of IL-33, Treg cells lost their ability to suppress the proliferation of effector T cells and were altered to produce type 2 cytokines (Figures 3 and 4). In the intestinal mucosa, Treg cells showed the propensity to express RORγt and differentiate into “Th17-type” cells in inflammatory conditions.^{26, 27} Our study suggests that similar alteration in the functions and phenotypes of Treg cells may occur in the airway mucosa during type 2 airway inflammation. Further research is needed to examine which environmental cues help sustain suppressive function in Treg cells or, alternatively, drive them to acquire new functions, such as the production of type 2 cytokines. In addition, future clinical studies are necessary to address not only the quantities of Treg cells, but also their functions.

Another major finding in this manuscript is that the mice, which had been previously tolerized to OVA and were unable to be sensitized by systemic administration of the antigen plus an adjuvant alum, were sensitized to the same antigen when it was administered into the airways together with IL-33 (Figures 5 and 6). These observations suggest a potent adjuvant-like activity of IL-33 that not only induces type 2 immune responses in naïve mice,¹⁶ but also breaks down established airway immunological tolerance. While isolated Treg cells can be activated directly by IL-33 (Figure 1), we were unable to address directly whether such potent immunostimulatory activity of IL-33 is mediated by its actions on Treg cells, effector T cells, or both. It is also possible that IL-33 activates APCs, such as DCs,^{61, 62} and may indirectly affect the differentiation and functions of Treg cells and effector T cells. Future experiments using conditional knockout mice (e.g., St2^fl/fl mice) will be necessary to address this remaining question once these mice become more widely available.

Nonetheless, our observations suggest instability of established airway immunological tolerance that can be broken down by potent pro-inflammatory cytokines, such as IL-33. Recent clinical studies indicate several immunological and biomedical pathways that define asthma phenotypes;⁶³ onset of allergic or Th2-type asthma can occur in childhood or the adult years. Our study provides an answer to why established immunologic tolerance or once well-controlled immunity in the lungs may become compromised later in life in some patients. In addition, IL-33 may contribute to immunopathology of persistent asthma by activating ILC2s and airway epithelial cells.⁶⁴ Thus, further studies to elucidate the immunological mechanisms of IL-33 production and its biological effects on immune cells and immune systems may provide a better understanding of the pathogenesis of certain airway diseases and development of new prevention and treatment strategies for them.

METHODS

Fluorescence-activated cell sorting (FACS) analyses of lung and splenic Treg cells

Lungs were harvested from naïve Foxp3^EGFP reporter mice or Foxp3^EGFPSt2^−/− reporter mice that were euthanized by an overdose of pentobarbital. Alternatively, Foxp3^EGFP mice were exposed intranasally to phosphate-buffered saline (PBS) or IL-33 (100 ng/dose, eBioscience, San Diego, CA) on days 0, 2, and 4. Twelve to twenty-four hours after the last exposure, mice were euthanized, and lungs were collected. Minced spleen or lungs were incubated with 0.08 Wünsch units/ml of liberase (Roche, Mannheim, Germany) and 26 μg/ml of DNase I (StemCell Technologies, Vancouver, BC, Canada) in RPMI 1640 medium (Gibco, Rockville, MD) for 1 hour at 37°C. Red blood cells were lysed with ammonium chloride/potassium lysing buffer, and single cells were washed and resuspended in PBS contained 1% of bovine serum albumin (BSA). In some experiments, lung single-cell suspensions were cultured in 24-well plates at 5×10⁶ cells/ml and stimulated for 6 hours with 10 ng/ml phorbol 12-myristate 13-acetate (PMA) plus 500 ng/ml ionomycin in the presence of GolgiPlug™ (BD Biosciences, San Jose, CA) in RPMI 1640 medium with 10% fetal bovine serum (FBS, Hyclone, Logan, UT), 50 μM 2-ME, 2mM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (complete medium).

For staining, cells were preincubated with Fc-receptor blockers for 30 minutes at 4°C and stained with combinations of PerCP Cy™5.5 -conjugated anti-CD4 (clone RM4-5, BD Biosciences) or Brilliant Violet 785™-conjugated anti-CD4 (clone RM4-5, Biolegend, San Diego, CA) and biotin-conjugated anti-ST2 (DJ8, MD Biosciences, St. Paul, MN), followed by streptavidin-allophycocyanin. In some experiments, FITC-conjugated anti-ST2 (DJ8, MD Biosciences) was used. To exclude dead cells from the analyses, 7-aminoactinomycin D (7-AAD, BD Biosciences) or Fixable Viability Dye eFluor® 506 (eBioscience) was added. In some experiments, cells were fixed, permeabilized, and stained with biotin-conjugated anti-Foxp3 antibody (clone FJK-16s, eBioscience), followed by streptavidin-PerCP Cy™5.5 (BD Biosciences), PE-conjugated anti-GATA3 antibody (clone L50–823, BD Biosciences), allophycocyanin-conjugated anti-IL-5 (clone TRFK5, BD Biosciences), and eFluor 660-conjugated anti-IL-13 (clone eBio13A, eBioscience). After washing, the cells were resuspended in PBS containing 1% BSA, fixed with 1% paraformaldehyde, and analyzed by BD LSRII™ flow cytometer and FlowJo (version 9.8, Tree Star) software (BD Biosciences Immunocytometry Systems). The lymphocytic cell population was gated by using a forward and side scatter plot, and 7-AAD⁺ or eFluor® 506⁺ dead cells were excluded from the analyses.

Sorting of lung Treg cells and culture

The entire population of CD4⁺Foxp3eGFP⁺ Treg cells or ST2⁻ and ST2⁺ subpopulations of CD4⁺Foxp3eGFP⁺ Treg cells were isolated by sorting. Briefly, CD4⁺ T cells were enriched from the single-cell suspensions of lungs and spleen by EasySep CD4⁺ cell isolation kit (StemCell Technologies). The cells were stained as described above, and Treg cells were sorted using a FACSAria™ flow cytometer (BD Biosciences). To examine cytokine protein production, sorted lung Treg cells were cultured in round-bottomed 96-well plates at 4x10⁵ cells/ml and stimulated with 10 ng/ml PMA plus 500 ng/ml ionomycin in complete culture medium for 6 to 8 hours. Cell-free supernatants were harvested for quantitation of cytokine levels by enzyme-linked immunosorbent assay (ELISA). To examine St2 mRNA expression, sorted ST2⁺ and ST2⁻ Treg cells were then cultured with IL-33, 40 ng/ml, under the stimulation of anti-CD3/anti-CD28 dynabeads (ThermoFisher, Life Technologies, Waltham, MA) for 24 hours. Cells were lysed for mRNA expression analyses as described below. Alternatively, sorted lung Treg cells were cultured with 20 ng/ml IL-2 (R&D Systems) with or without 40 ng/ml IL-33 for 96 hours. Cells were lysed immediately for mRNA analysis (see below) or stimulated with PMA plus ionomycin for 6 hours for analyses of cytokine proteins in cell-free supernatants.

Gene expression analyses

Cells were lysed in Trizol (Invitrogen, Carlsbad, CA), and total RNA was extracted and purified using the PureLink® RNA Mini Kit (ThermoFisher Scientific, Waltham, MA) followed by real-time quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) analyses. Complementary DNA was generated with iScript Advance cDNA synthesis kit (Bio-Rad, Hercules, CA). In some experiments, cDNA amplification was performed using an Ovation PicoSL WTA kit (NuGen Technologies Inc., San Carlos, CA). Real-time PCRs were performed on ABI RT-PCR StepOne Plus System by the use of TaqMan probes for Foxp3, Gata3, Il1r1, Il5, and Il13 and normalized to the hypoxanthine phosphoribosyltransferase gene (HPRT) (all from ThermoFisher Scientific). The relative expression level was determined via the 2-ΔΔCT method.³¹ CD4⁺Foxp3eGFP⁺ Treg cells or CD4⁺Foxp3eGFP⁻ T cells that were isolated from the lungs of naïve non-treated Foxp3^EGFP reporter mice by FACS sorting were used as controls throughout the study.

Airway tolerance model

Endotoxin-free ovalbumin (OVA) (<0.5 EU/mg, verified by a Limulus amebocyte lysate assay; Wako Chemicals USA, Inc., Richmond, VA) was prepared using specific pathogen-free chicken eggs (Charles River Laboratories, Wilmington, MA) under sterile conditions as previously described.³² Airway immunologic tolerance to OVA was induced by using a similar protocol as previously described.^{7, 33} Briefly, endotoxin-free OVA (100μg/dose) was intranasally administrated to naïve BALB/c mice from day -12 to day -7. To verify the state of immunologic tolerance, OVA (10 μg/dose, Sigma-Aldrich, St. Louis, Mo) plus alum were injected intraperitoneally on day 7 and day 14. Control mice received sterile PBS plus alum. Peripheral blood was collected on day 28 to examine the plasma levels of IgE antibodies to OVA. On days 29, 30, and 31, mice were challenged intranasally with endotoxin free-OVA (100 μg/dose) and euthanized on day 32 by an overdose of pentobarbital. Bronchoalveolar lavage (BAL) fluids and lungs were collected for analyses. To examine the effects of airway administration of IL-33 on immune tolerance, mice were exposed to i.n. endotoxin-free OVA (100 μg/dose) alone, IL-33 alone (100 ng/dose, eBioscience), or OVA plus IL-33 on days 14 and 21. In some experiments, OVA alone, OVA plus IL-33, or OVA plus a mixture of allergen extracts (50 μg each of A. alternata, A. fumigatus, and HDM) were administered intranasally on days 0 and 7 without i.p. administration of OVA plus alum.

Cell numbers in the BAL fluid were counted with a hemocytometer, and differentials were determined in cytospin preparations stained with Wright-Giemsa. More than 200 cells were analyzed using conventional morphologic criteria. The lungs were fixed with 10% formaldehyde and embedded in paraffin. Sections of fixed lung tissues were stained with hematoxylin and eosin (H&E) stain and periodic acid-Schiff (PAS) stain for histology analysis.

Allergen-induced airway inflammation model

To examine the effects of exposure to natural allergens, mice were exposed to extracts of natural allergens as described previously.³² Briefly, extracts of natural allergens, including Alternaria alternata, Aspergillus fumigatus, and house dust mite (HDM, Dermatophagoides pteronyssinus), were purchased from Greer Laboratories (Lenoir, NC). These extracts contained undetectable levels of endotoxin (<10 ng/mg extract, <0.50 ng/dose). Naïve wild-type (WT) BALB/c mice, BALB/c-St2^−/− mice, or BALB/c-Foxp3^EGFP reporter mice were exposed intranasally to a mixture of allergen extracts (10 μg each of A. alternata, A. fumigatus, and HDM plus 10 μg endotoxin-free OVA in 50 μl PBS), three times a week, for up to 4 weeks under isoflurane inhalation anesthesia. Control mice received PBS alone. Mice were euthanized by an overdose of pentobarbital 12 hours after last exposure (2-week model) or 24 hours after last exposure (4-week model). Lungs were harvested, and single-cell suspensions were prepared and stained as described previously.

ELISA for cytokine and antibody levels

The levels of IL-4, IL-5, IL-13, IL-17, and IFN-γ in the supernatants of Treg cells and BAL fluids were measured using ELISA kits (R&D Systems, Minneapolis, MN) following the protocol recommended by the manufacturer. The levels of transforming growth factor (TGF)-β were measured using the Legend Max Free Active TGF-β1 ELISA kit (Biolegend). The plasma levels of OVA-specific IgE were measured using sandwich ELISA as previously described.³⁴

Key Messages.

• IL-33 altered Treg cells to express GATA3 and Th2-type cytokines and impaired their suppressive function.

• Established immunologic tolerance in the airway was broken down by IL-33.

• “Th2 cell-like” Treg cells were found in allergen-driven airway inflammation.

Abbreviations

7-AAD

7-aminoactinomycin D

APCs

antigen presenting cells

BAL

bronchoalveolar lavage

BSA

bovine serum albumin

DC

dendritic cell

eGFP

enhanced green fluorescent protein

ELISA

enzyme-linked immunosorbent assay

FACS

fluorescence-activated cell sorting

FBS

fetal bovine serum

Foxp3

forkhead box P3

H&E

hematoxylin and eosin

HDM

house dust mite

IFN

interferon

IL

interleukin

ILC2s

type 2 innate lymphoid cells

i.n

intranasal

i.p

intraperitoneal

OVA

ovalbumin

PAS

periodic acid-Schiff

PBS

phosphate-buffered saline

PMA

phorbol 12-myristate 13-acetate

qRT-PCR

quantitative reverse transcription polymerase chain reaction

RORγt

retinoic acid receptor-related orphan receptor-γt

SEM

standard error of the mean

STAT5

signal transducer and activator of transcription 5

TGF

transforming growth factor

TNF

tumor necrosis factor

Treg

regulatory T

TSLP

thymic stromal lymphopoietin

WT

wild-type