Recent interest has focused on innate-type cytokines as promoters of type 2 immunity and targets for drug development in asthma. IL-33 induces production of IL-4 and/or IL-13, which is associated with STAT6-dependent responses in innate cells, including group 2 innate lymphoid cells (ILC2s), macrophages, and eosinophils. Our published data show that STAT6-immunomodulatory peptide (STAT6-IP), an immunomodulatory peptide designed to inhibit the STAT6 transcription factor, reduces induction of Th2 adaptive immunity in respiratory syncytial virus infection and asthma models. Nevertheless, the mechanism of STAT6-IP–dependent inhibition has remained obscure. In this study, we demonstrate that STAT6-IP reduced IL-33–induced type 2 innate lung inflammation. Specifically, our data show that STAT6-IP reduced recruitment and activation of eosinophils as well as polarization of alternatively activated macrophages. Decreases in these cells correlated with reduced levels of IL-5 and IL-13 as well as several type 2 chemokines in the bronchoalveolar lavage fluid. STAT6-IP effectively inhibited expansion of ILC2s as well as the number of IL-5– and IL-13–producing ILC2s. Our data suggest that STAT6-IP effectively disrupts IL-13–dependent positive feedback loops, initiated by ILC2 activation, to suppress IL-33–induced type 2 innate immunity in the murine lung.
Asthma, which ranks among the costliest of all chronic diseases, has been steadily increasing during the last two to three decades in both severity and occurrence (1). Affecting at least 300 million people worldwide, asthma is a chronic inflammatory disease of the airways characterized by intermittent airflow obstruction and airway hyperresponsiveness (AHR) (2).
Historically, symptoms of allergic asthma have been attributed to Th2 adaptive immunity where CD4+ Th2 cells orchestrate responses in both immune and structural cells of the lung, which result in the structural and functional changes that define the asthmatic lung (3, 4). IL-13 is considered the main driver of many of these responses, across the disease severity spectrum (2, 3, 5); both circulating and sputum eosinophils (Eos) as well as AHR are associated with greater mRNA levels of IL-13–responsive genes (5, 6).
This paradigm of disease orchestrated by Th2 cells has shifted recently, in part because of data from murine models showing that IL-13 produced by innate cells, and more specifically group 2 innate lymphoid cells (ILC2s), can induce allergic inflammation and/or AHR, even in the absence of Th2 adaptive immunity (7–12). In particular, ILC2-dependent production of IL-5 promotes recruitment of Eos into the lung (13, 14), and ILC2-derived IL-13 promotes mucus production and AHR and drives dendritic cell (DC)–dependent lymph node responses that culminate in Th2 differentiation and cytokine production (12, 14–16). Moreover, our new data implicate ILC2-dependent IL-13 production in Eos activation and transit into the airways, activities that are independent from IL-13–induced production of Eos chemokines (17).
There is now widespread research focusing on regulation of allergic airways disease and type 2 inflammation by innate cytokines, including IL-33 (17–21). Airway epithelial cells expressing IL-33 are more abundant in asthmatics compared with controls (20, 22), and polymorphisms in genes encoding IL-33 and its cognate receptor, ST2, are highly correlated with asthma (23, 24). In murine models of asthma induced by the house dust mite, type 2 inflammation is reduced in mice lacking ST2 or upon treatment with soluble ST2 (9, 25, 26).
IL-33 is released from structural cells in the lung in response to environmental triggers, such as allergens or respiratory viruses (18, 21, 22). Although ILC2s are often described as the main targets of IL-33 in the lung, other innate cells in the lung also respond to this cytokine, including Eos, DCs, and macrophages (Mϕs), among others (27). We (17) and others (15, 20, 28, 29) have shown that delivery of IL-33 to the lungs of naive mice induces cytokine and chemokine production, influx and activation of Eos, as well as Mϕ differentiation into alternatively activated Mϕs (AAMs). Consistent with the ability of IL-33 to induce production of IL-13 by these cells, many IL-33–induced responses in the lung are dependent upon IL-13, the IL-13 receptor, or its target, the STAT6 transcription factor (17, 20, 30). This does not exclude cooperative interactions between IL-33 and IL-13, however, as many cells that respond to IL-33, including ILC2s, also encode functional IL-13 receptors (11, 31). Synergy between IL-33 and IL-13 has been described in (bone marrow–derived) Mϕs (20), a response that likely also occurs in Mϕs in vivo.
We designed STAT6-immunomodulatory peptide (STAT6-IP), to act in a dominant negative manner to inhibit the native STAT6 protein (32). We have shown that STAT6-IP delivery to the lung reduces maladaptive type 2–biased airway inflammatory responses in both allergy and respiratory syncytial virus (RSV) models (32–35). Our most recent data demonstrate that STAT6-IP delivery at the time of priming, in both asthma and RSV infection models, diminishes the induction of Th2 adaptive immunity, possibly through STAT6-IP–dependent inhibition of AAM differentiation and/or DC migration to the lung-draining lymph nodes (35, 36).
The goal of our current study was to better understand mechanisms by which STAT6-IP interacts with innate cells to inhibit type 2 inflammatory responses induced by recombinant IL-33 delivery to the lung, focusing on early events that occur within days of IL-33 exposure. Our data demonstrate that STAT6-IP effectively reduced type 2 inflammatory responses induced by IL-33, including: 1) AAM polarization, 2) Eos recruitment and activation, 3) bronchoalveolar lavage fluid (BALF) Eos influx, 4) production of type 2 cytokines and chemokines, and 5) ILC2 expansion and cytokine production. STAT6-IP delivery to the lung also disrupted the ability of ex vivo–cultured lung explants to produce IL-13, even upon IL-33 restimulation. Taken together, our data demonstrate that STAT6-IP effectively targets innate immune cells to reduce type 2 inflammation by durably reducing IL-13 production that drives responses in effector cells, including Eos and AAMs.
Materials and Methods
Six- to 8-wk-old male wild-type BALB/c mice were bred in-house and kept under pathogen-free conditions in cages supplemented with water and irradiated food.
STAT6-IP and STAT6-control peptide (STAT6-CP) were synthesized by JPT Peptide Technologies (Berlin, Germany) as described previously (32). Peptides were diluted to a final concentration of 10 µg/µl in normal saline (SAL) and stored at −80°C.
IL-33–induced airway inflammation
After brief isoflurane anesthesia, mice were treated intranasally (i.n.) with SAL as a negative control or with IL-33 (1.50 µg) in a volume of 30 µl daily for each of 3 consecutive days. A volume of 30 µl of STAT6-IP (100 µg), STAT6-CP (100 µg), or SAL was given i.n. at 24 h and again 1 h before each of the three IL-33 applications. IL-33 was purchased from eBioscience (San Diego, CA). Following the last treatment, mice were allowed to rest for 3 d, after which they were sacrificed by a lethal dose of sodium pentobarbital. Experiments were repeated at least twice, and the number of mice per group and the number of times each experiment was repeated are presented in the figure legends.
Lungs were perfused with 10 ml of PBS and slowly inflated with 1 ml of 10% formalin. Entire lungs were carefully detached, placed in histology cassettes, and immersed for 72 h in 10% formalin. Tissues were processed with the Leica ASP300S tissue processor and embedded in paraffin. Sections of 0.4 μm were cut and mounted on glass slides. Adjacent lung sections were stained with a solution of 1% Harris hematoxylin and 5% eosin-phloxine (H&E) to assess inflammation, or with 1% periodic acid–Schiff (PAS) reagent to assess mucus production. Several photomicrographs of the left lobe were taken with a ×20 objective on the Zeiss Axio Imager 2 microscope and stitched together to obtain an image of the entire left lobe. The magnified areas displayed in Fig. 7 were obtained from the original images. Scale bars are 100 and 200 µm as indicated in the figure. All images were visualized, captured, and analyzed with Zen software (Zeiss). A blinded observer scored the images based on inflammation level or abundance of PAS+ airways. For semiquantitative analysis of mucus levels, 10 airways, including all positive airways, were selected and scored as follows: 0, none; 1, few; 2, ≤50%; 3, >50%; 4, 100%. The average mucus score per mouse was then calculated. For semiquantitative analysis of immune cell infiltration, 5–10 airways per mouse were scored as follows: 0, none; 1, modest; 2, abundant. Images representative of the average score of the treatment groups were selected.
BALF acquisition and cell counting
Mice were euthanized and a vertical incision was made at the neck region to access the trachea, which was cannulated with a 20G metal catheter and secured with surgical thread to prevent backflow. BALF was collected by lavaging the lungs twice each with 1 ml of ice-cold PBS. The first lavage was centrifuged at 500 rcf for 5 min to separate the cells from the supernatant, which was stored at −80°C for further analysis (see below). The remaining cells from the second lavage were combined with the cells from the first lavage and the supernatant was discarded. The total recovered number of live and dead cells was then quantified using a hemocytometer and trypan blue staining. Afterward, 20,000 cells per sample were spun (Cytospin) onto glass slides and air dried overnight. Slides were stained with Diff-Quik (Siemens Healthcare Diagnostics, Princeton, NJ) for manual differential cell counting, which relies on morphologic criteria to quantify the relative proportion of BALF Mφs, Eos, neutrophils, and lymphocytes. The absolute number of each cell type in the BALF was quantified by multiplying the proportion of each cell type by the total cell count.
Lung explant culture
Lung explant cultures were prepared by finely mincing the upper right lobe of the lung for culture in a total of 1 ml of RPMI 1640 media supplemented with 10% FBS, 5% penicillin/streptomycin, 1 mM sodium pyruvate, 1 mM nonessential amino acids, and 55 µM 2-ME, containing either control SAL or IL-33 (20 ng/ml) at 37°C and 5% CO2 (20). After 48 h, supernatants were collected, and IL-13 production was quantified by ELISA.
ELISA and Quansys assays
IL-13 protein was quantified in duplicates from lung explant cultures following the Ready-SET-Go! IL-13 ELISA kit instructions from Thermo Fisher Scientific (Carlsbad, CA). Samples from mice treated with either IL-33 alone or with IL-33 + STAT6-CP were diluted 1:20 to guarantee that all values fell within the assay detection limits (7–500 pg/ml), whereas all other samples were prepared neat. CCL11, CXCL1, CCL22, CCL17, IL-4, IL-5, IL-13, IL-17, and IFN-γ protein levels in the BALF were quantified using the Q-Plex multiplex ELISA arrays (PBL Assays, Quansys Biosciences, Logan, UT) following the manufacturer’s instructions. All samples were run neat. Samples in which cytokine/chemokine levels were undetectable were assigned a value corresponding to half the lower detection limit of the employed kit. For CCL22 analyzed by Quansys (Fig. 3), IL-33 and STAT6-CP levels were above the limit of detection, and so were assigned the maximum.
Real-time quantitative PCR
The quantitative PCR (qPCR) assay was conducted following methodology previously described by Zhao et al. (17), and in accordance to the Minimum Information for Publication of Quantitative Real-Time PCR Experiments guidelines (37). Briefly, the inferior right lung lobe was flash-frozen in liquid nitrogen immediately after dissection and stored at −80°C. Total RNA was isolated via the phenol-chloroform method by means of TRIzol reagent (Invitrogen, Carlsbad, CA), and RNA purity and concentration were evaluated by spectrophotometry using NanoDrop 2000. RNA integrity was visualized on a 1% agarose gel comprising 0.12% (v/v) NaClO and 0.5 µg/ml ethidium bromide. After treating 1 µg of RNA with dsDNase (Thermo Scientific, Carlsbad, CA), cDNA was generated with the Thermo Scientific Maxima cDNA synthesis kit (no. M1669) using random hexamer primers. qPCR reactions were performed in duplicate with the StepOnePlus real-time PCR system (Applied Biosystems, Carlsbad, CA). Purity of each set of primers was validated with a “no-template” control for every experiment. Murine primer sequences for the genes of interest are presented in Supplemental Table I. Levels of mRNA were then calculated with the ΔΔCt method normalized to the reference gene β-actin and presented relative to levels in SAL-treated control mice.
Lung digestion and flow cytometry
Lung tissue was finely minced and enzymatically digested with a mixture of DNase I (200 µg/ml; Sigma-Aldrich, St. Louis, MO), Liberase TM (100 µg/ml; Roche, Indianapolis, IN), hyaluronidase 1a (1 mg/ml; Life Technologies, Carlsbad, CA), and collagenase XI (250 µg/ml; Life Technologies, Carlsbad, CA) in RPMI 1640 for 30 min at 37°C and 5% CO2 as described previously (38). Subsequently, cells were washed with RPMI 1640 media containing 1% penicillin/streptomycin and 5% FBS to maintain cell viability. RBCs were lysed with sterile, filtered ammonium-chloride-potassium buffer. Remaining viable cells were recovered via filtration through a 0.7-µm strainer and counted using trypan blue exclusion. Samples were diluted to 1.25 × 106 cells for ILC2 staining and 1 × 106 cells for Mϕ and Eos staining. For ILC2 staining, total lung cells were stimulated with PMA (50 ng/ml) and ionomycin (0.5 µg/ml) in the presence of GolgiStop (0.665 µl/ml) in a low adherence round-bottom 96-well plate for 4 h, at 37°C, 5% CO2 and then thoroughly washed prior to flow cytometry staining. In the dark, cells were incubated for 20 min with eFluor 780 viability dye (eBioscience, San Diego, CA). Cells were then washed and incubated at 4°C for 10 min with anti-CD32/16 to block Fc receptors. ILC2s were stained using eF450 anti-Thy1.2 (Invitrogen, clone 53-2.1), PeCy7 anti-CD127 (BioLegend, clone A7R34), PerCP-eF710 anti-ST2 (eBioscience, clone RMST2-2), BV605 anti-KLRG1 (BioLegend, clone 2F1/KLRG1), and a combination of PE-conjugated Abs that recognize CD3e (BioLegend, clone 145-2611), CD11c (BioLegend, clone N418), CD11b (eBioscience, clone M1/70), CD49b (BD Biosciences, clone DX5), CD45R (Invitrogen, clone RA3-6B2), TCRγδ (BD Biosciences, clone GL3), and Ly6G (BioLegend, clone 1A8). Mϕs and Eos were stained together with BUV395 anti-CD45.2 (BioLegend, clone 104), AF700 anti-Ly6G (eBioscience, clone 1A8), BV510 anti-F4/80 (eBioscience, clone BM8), eF450 anti–MHC class II (eBioscience, clone M5/114.15.2), FITC anti-CD38 (eBioscience, clone HIT2), PeCy7 anti-CD11c (eBioscience, clone N418), PE-Vio770 CD101 (Miltenyi Biotec, Bergisch Gladbach, Germany), allophycocyanin anti-EGR2 (eBioscience, clone erongr2), PE anti–Siglec F (BD Biosciences, clone E50-2440), and PerCP-eF710 anti-ST2 (Invitrogen, clone RMST2-2). Cells were then fixed with IC (intracellular) fixation buffer (eBioscience, San Diego, CA) overnight. For ILC2s, following overnight fixation, cells were permeabilized with BD Perm/Wash buffer and subsequently intracellularly stained with AF488 anti–IL-13 (eBioscience, clone eBio13A) and allophycocyanin anti–IL-5 (BD Biosciences, clone TRFK5). Acquisition of all samples was performed with a BD LSRFortessa (Immununophenotyping Core Facility, Research Institute of the McGill University Health Centre) flow cytometer. Analysis was performed with FlowJo v10 (FlowJo, Ashland, OR). Positive populations were defined by fluorescence minus one control.
Randomization was respected by assigning mice from the same cage/litter to different treatment groups. Blinding to sample group allocation was conserved from sacrifice to data analysis.
All analyses and graphs were generated using Prism (GraphPad Software, San Diego, CA). Data were analyzed by one-way ANOVA followed by multiple comparisons using Tukey’s post hoc test. A p value of <0.05 was considered significant (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001). Outcomes are presented as mean ± SEM. Grubb’s test with α = 0.05 was employed to remove outliers.
Animal studies were approved by the McGill University Animal Care Committee and performed in accordance with the guidelines of the Canadian Council on Animal Care. Upon conducting research using animals, the investigators adhered to the laws of the United States, and the regulations of the Department of Agriculture.
STAT6-IP reduces IL-33–induced BALF eosinophilia
To better understand how STAT6-IP inhibits responses in cells of the innate immune system prior to the onset of Th2 adaptive immunity, male BALB/c mice were treated with IL-33 either alone or in the presence of STAT6-IP or negative control, STAT6-CP, according to the timeline shown in (Fig. 1. The profile of BALF inflammatory cells was assessed 72 h after the last IL-33 administration. As expected, BALF cell counts were low in SAL-treated mice and most cells were Mϕs (Fig. 2A–C). IL-33 induced prominent BALF inflammation: total BALF cell counts were dramatically increased (∼13-fold) (Fig. 2A–C), and large numbers of Eos were recruited into the airways of these mice (Fig. 2). In mice treated with STAT6-IP, but not STAT6-CP, total BALF cell counts and Eos influx were reduced to levels comparable to those in SAL-treated mice (Fig. 2).
STAT6-IP reduces IL-33–induced BALF cytokine and chemokine levels
To further understand the effects of STAT6-IP on IL-33–induced inflammation, we quantified a broad spectrum of cytokines and chemokines in the BALF. Our data demonstrate that the type 2 cytokines, IL-5 and IL-13, and chemokines recruiting Eos (CCL11/Eotaxin-1), neutrophils (CXCL1/KC), and Th2 cells (CCL22/MDC and CCL17/TARC) were significantly greater in IL-33–treated mice compared with their SAL-treated counterparts (Fig. 3). Each of these was significantly reduced in mice treated with STAT6-IP (Fig. 3) and unaffected by delivery of STAT6-CP. Levels of IL-4, IFN-γ, and IL-17 fell largely below the limit of detection (data not shown).
STAT6-IP reduces IL-33-induced lung recruitment and activation of Eos
Having shown that STAT6-IP reduced the number of BALF Eos induced by IL-33, we examined effects of STAT6-IP on levels of Eos in the lung as well as their state of activation. We (17) and others (39) have shown that Eos upregulate CD11c upon activation and that only these Eos transit into the airways where they can be recovered in the BALF. As expected, IL-33 significantly increased both CD11c− “mature” Eos as well as CD11clow “activated” Eos (39) (Fig. 4A, 4C). In addition, the proportion of activated Eos within the total Eos population increased from 31 ± 2% in SAL-treated mice to 47 ± 4% in IL-33–treated mice (Fig. 4B). Administration of STAT6-IP but not STAT6-CP effectively reduced both Eos populations in the lung (Fig. 4A, 4C). The proportion of activated Eos in STAT6-IP–treated mice was reduced to levels comparable to those in SAL-treated control mice (32 ± 4%) (Fig. 4B). Similar to data from Mesnil et al. (40), who showed that activated lung Eos in mice exposed to house dust mite express increased levels of CD101, our data demonstrate that Eos from IL-33–treated mice also expressed CD101. In fact, most Eos from IL-33–treated mice coexpressed CD11c and CD101 (Supplemental Fig. 1A), and these cells (Supplemental Fig. 1B) as well as the mean fluorescence intensity (MFI) of CD101 (Supplemental Fig. 1C) were reduced by STAT6-IP but not STAT6-CP.
STAT6-IP reduces IL-33–induced polarization of AAM and inhibits IL-33–induced increases in Ym1, Fizz1, and IL-13 mRNA levels
AAMs are a potential source of the chemokines and cytokines (Fig. 3) present in IL-33–treated mice. To confirm STAT6-IP targeting of AAMs, we evaluated responses in alveolar Mϕs in IL-33–treated mice gating on CD45+Ly6G−/intCD11chiSiglec F+ cells (Fig. 4A). IL-33 increased the number of alveolar Mϕs by ∼3-fold, an increase that was not significantly affected by STAT6-IP or STAT6-CP, although there was a trend for STAT6-IP to reduce the number of these cells (Fig. 5A, upper left panel). We then assessed AAM polarization by analyzing EGR2 (41) (Fig. 5A, upper right panel and lower left panel) and Arg1 (Supplemental Fig. 2) AAM markers in these cells. IL-33 increased the number of both EGR2+ AAMs and Arg1+ AAMs (Fig. 5A, Supplemental Fig. 2), and STAT6-IP, but not STAT6-CP, reduced the differentiation of these cells (Fig. 5A, Supplemental Fig. 2). There were no changes in the number of cells expressing markers of classically activated Mϕs (CD38, inducible NO synthase) in response to IL-33 or upon delivery of STAT6-IP or STAT6-CP (Fig. 5A, Supplemental Fig. 2, and data not shown). We further quantified lung mRNA levels of AAM markers Ym1 and Fizz1 as well as IL-13. IL-33 induced large increases in mRNA levels of each of these mediators: Ym1 increased 44.5 ± 3.5-fold, Fizz1 increased 370.8 ± 74.0-fold, and IL-13 increased 24.3 ± 3.5-fold (Fig. 5B). Treatment with STAT6-IP but not STAT6-CP significantly reduced each of these.
STAT6-IP targets ILC2 to reduce type 2 inflammation
ILC2s are considered one of the main targets of IL-33 in the lung, and in their absence, IL-33–induced type 2 inflammation is compromised (12, 14). We hypothesized that STAT6-IP targeted ILC2s, reducing production of IL-5 and IL-13 to limit Eos influx and activation and AAM differentiation induced by IL-33. Our gating strategy to identify lung ILC2s is shown in (Fig. 6A. Because our experiments were performed only in male mice, we focused on KLRG1+ ILC2s (42), although we also examined responses in ILC2s lacking KLRG1 expression (data not shown). IL-33 induced robust increases in both total and KLRG1+ ILC2s, and these were both largely prevented by STAT6-IP, but not STAT6-CP (Fig. 6B). Similarly, STAT6-IP reduced the number of KLRG1+ ILC2s producing IL-13 alone or with IL-5 (Fig. 6C, second and third panels). MFI of each cytokine was not significantly affected by either IL-33 or STAT6-IP (or STAT6-CP) (Supplemental Fig. 3A). Very few ILC2s produced only IL-5, and these were similarly inhibited by STAT6-IP (Fig. 6C, first panel). Similar results were found in ILC2s lacking KLRG1 expression (data not shown), although the number of these cells was much lower, as expected in male mice (42). Thus, STAT6-IP, but not STAT6-CP, potently inhibited expansion of cytokine-producing ILC2 populations induced by IL-33.
STAT6-IP reduces IL-33–induced airway inflammation and mucus production
Consistent with the enhanced inflammatory responses to IL-33 noted above, mice treated with IL-33 also had a modest increase in inflammatory cell infiltration and mucus production in the lung and these responses were reduced in mice treated with STAT6-IP, but not STAT6-CP (Fig. 7, Supplemental Fig. 3B).
STAT6-IP reduces the ability of IL-33 to promote increases in lung cells that produce IL-13 upon ex vivo IL-33 re-exposure
To better understand how STAT6-IP regulates IL-13 levels in this model, we examined IL-13 protein production from minced lung explants harvested from mice exposed to SAL, IL-33, or IL-33 plus either STAT6-IP or STAT6-CP (Fig. 1). Explants were restimulated ex vivo with either SAL (Fig. 8A) or IL-33 (Fig. 8B). We reasoned that IL-13 production from lung explants cultured with SAL would reflect cytokine production from cells in response to in vivo treatment whereas that from lung explants stimulated with IL-33 would reflect levels produced upon reactivation of those cells. Compared to lung explants from mice treated with SAL, those from mice treated with IL-33 produced ∼15-fold more IL-13. Treatment of mice with STAT6-IP (but not STAT6-CP) reduced IL-13 production from lung explants, whether cultured with SAL or IL-33 (Fig. 8). These data likely reflect the reduced numbers of ILC2, and possibly other innate cells, such as Eos, present in STAT6-IP–treated mice, with the ability to produce IL-13 in response to IL-33.
In the current study, we demonstrate that delivery of STAT6-IP to the lungs diminishes IL-33–induced type 2 inflammation to levels comparable to those of SAL-treated controls. In this model, compatible with our data (17) and those of others (15, 20, 28, 29), IL-33 potently increased BALF cytokine and chemokine levels, AAM polarization, lung Eos recruitment and activation, and ILC2 expansion with associated IL-13 and IL-5 cytokine production. Each of these responses was significantly reduced upon treatment with STAT6-IP, but not with STAT6-CP. These data are in agreement with the well-defined role of STAT6 as a crucial promoter of type 2 inflammation in the lung (4) as well as our published data examining inhibitory activity of STAT6-IP (32, 34–36). Importantly, our data in this study provide important insight into how STAT6-IP targets cells of the innate immune system to inhibit type 2 inflammation prior to the onset of Th2 adaptive immunity.
A number of studies have shown that STAT6 in nonstructural, innate cells promotes production of chemokines that recruit Eos into the airways (11, 43–45). Data from Medoff et al. (43) support a role for STAT6 specifically in myeloid cells for Eos recruitment. Moreover, IL-7 receptor–deficient, Rag2−/−IL2rg−/−, and RORα−/− mice, all of which lack ILC2s, fail to recruit Eos into the lung and airways and have diminished levels of BALF IL-5 (11, 14). Hence, the reduction in BALF Eos in STAT6-IP–treated mice can be attributed to a number of factors: lower levels of upstream STAT6-dependent chemoattractants (Fig. 3) from AAMs and other cells, diminished IL-5 production from ILC2s (Fig. 7), and/or reduced activation of Eos in the lung (Fig. 4), likely through inhibition of the number of IL-13–producing ILC2s (17). These findings are in agreement with our previously published data showing that STAT6 is required for movement of Eos from the lung into the airways in response to IL-33 (17).
In addition to the decline in BALF Eos infiltration, recruitment of Eos into the lungs in response to IL-33 was also reduced in STAT6-IP–treated mice. IL-33 exerts its effects on Eos-committed progenitors in the bone marrow, both directly (46) and via IL-5 production (47), the result of which is augmented Eos precursor development, differentiation and maturation, enhanced expression of adhesion molecules, and chemotaxis, ultimately amplifying recruitment of Eos precursors and mature Eos, both of which express Siglec F, into the lung (46, 47). In fact, IL-33–deficient mice display reduced bone marrow and peripheral blood Eos, and exogenous IL-33 expands Eos precursors as well as their IL-5Rα expression, culminating in heightened IL-5 responsiveness (46). As expected, ILC2-deficient mice have diminished levels of IL-5 in the lung as well as significantly reduced lung Eos numbers (14), and IL-5–neutralizing Ab inhibits IL-33–induced expansion of Siglec F+ Eos precursors in the bone marrow and blood (46). In agreement with the role of the IL-33–induced ILC2/IL-5 axis in Eos development and recruitment into the lung, in the current study we demonstrate that the STAT6-IP–associated decrease in ILC2 numbers and, more specifically, IL-5–producing ILC2s, was accompanied by a significant reduction in the number of lung Eos. Reduced Eos recruitment into the lung could also be attributed to STAT6-IP–dependent inhibition of AAM differentiation and associated Eos chemokine production (20, 41). Nevertheless, we speculate that the reduced Eos chemokine levels in STAT6-IP–treated mice may be less important for Eos recruitment into the lung compared with reduced ILC2 production of IL-5 based on our previous findings that Eos recruitment into the lung in STAT6 knockout mice is not compromised, even though AAM polarization is dramatically reduced (17). Our data from this previous study indicate that in STAT6 knockout mice (as opposed to STAT6-IP–treated mice from the current study) ILC2 expansion and cytokine production, as well as Eos recruitment into the lung, are intact. Taken together, these data highlight an important difference between STAT6-IP–treated mice and STAT6 knockout mice (17), and they suggest that STAT6-IP, although originally designed as a cell-penetrating STAT6 inhibitor, has additional immunomodulatory activities.
We (17) and others (37) have identified distinct Eos subsets that differ in their CD11c expression. To our knowledge, we are the first to provide evidence that IL-33-induced Eos activation is almost completely dependent on STAT6 (17). Specifically, we showed (using STAT6 knockout mice) that IL-33–induced Eos recruitment into the lung was independent of STAT6, whereas activation and transit into the airways were almost completely STAT6-dependent. These data are in line with data from our current study showing that IL-33–induced Eos activation was significantly reduced upon treatment with STAT6-IP. Little is known about the function of activated Eos expressing CD11c, aside from their peribronchial localization and their ability to transit into the airways (39, 40). We speculate, however, that these activated Eos promote type 2 inflammation based on data from Jacobsen et al. (48), who showed in an Eos adoptive transfer model that Eos exposed to IL-33 (along with GM-CSF and IL-4) promoted allergic inflammation, differentiation of AAMs, and mucus production, all in a manner that was dependent on their ability to produce IL-13.
Our data also demonstrate that STAT6-IP inhibited AAM differentiation induced by IL-33: YM1, Fizz1, and IL-13 lung mRNA levels were reduced as were the numbers of Mϕs expressing EGR2 or Arg1, markers of AAMs. Diminished AAMs in STAT6-IP–treated mice is in agreement with the well-characterized role of cell-intrinsic STAT6 activity on Mϕ differentiation into AAMs, both in vivo and in vitro (17, 20, 45, 49). Although it is not clear whether STAT6-IP targets Mϕs directly, or whether the reduction of AAMs is due to inhibition of ILC2s and/or Eos, the decline in AAMs and IL-13 may be responsible for the reduced BALF levels of CCL17 and CCL22 in STAT6-IP–treated mice (20).
In the current study we also demonstrate that STAT6-IP reduced both IL-33–induced ILC2 expansion as well as the number of cytokine-producing ILC2s. Because the IL-13/IL-5 MFI was unaffected, these data demonstrate that the STAT6-IP inhibitory impact on ILC2s is attributed to compromised ILC2 expansion in the lung, effectively reducing the number of ILC2s with the potential to produce cytokines as opposed to altering the ability of individual ILC2s to produce IL-5 and/or IL-13. Although a role for cell-intrinsic STAT6 activity in ILC2 proliferation has now been established (50), and STAT6-IP inhibition of ILC2 expansion is in agreement with these findings, they are in contrast to published data from our laboratory demonstrating that in STAT6-knockout mice ILC2 expansion in the lung in response to IL-33 is unaffected (17). Thus, these data provide further evidence that STAT6-IP has immunomodulatory activity distinct from its ability to enter cells and directly inhibit STAT6.
Finally, we have demonstrated in this study that in vivo IL-33 administration enhanced ex vivo IL-13 production from minced lung explants, the production of which was further increased upon restimulation with IL-33 and that in vivo delivery of STAT6-IP effectively reduced ex vivo IL-13 production. Although we have not yet identified which cells produced IL-13 in these lung explant cultures, likely sources include ILC2s, AAMs, and activated Eos (20, 48, 49). Thus, the reduced ex vivo IL-13 production from lungs harvested from STAT6-IP–treated mice can likely be attributed to reductions in each of these cell types, whether through diminished proliferation, recruitment, and/or activation. We cannot, however, rule out the possibility that STAT6-IP has also modified the innate lung environment so that even upon secondary exposure to IL-33, the ability of these innate cells to respond is affected. Our published data from both RSV infection (34, 35) and allergic airways disease models (36) indicate that delivery of STAT6-IP at the time of Ag priming disrupts Th2-biased inflammation upon secondary infection with RSV or challenge with OVA (34–36). Combined with data from the current study, we propose that STAT6-IP targets cells of the innate immune system, effectively inhibiting Th2 differentiation and adaptive immunity and disrupting the ability of these innate cells (e.g., AAMs, ILC2s) to respond even upon restimulation.
We propose the following working model based on our current findings (Fig. 9). Delivery of the first dose of IL-33 to the lung induces ILC2 cytokine production, the outcome of which is recruitment of Eos into the lung via IL-5 and their subsequent transition into the activated phenotype via IL-13. IL-13 (from ILC2 and activated Eos) synergizes with additional doses of IL-33 to induce AAM polarization, which, through chemokine and cytokine production, further promotes Eos recruitment and activation. STAT6-IP effectively constrains each of these innate cell types, interrupting a series of positive feedback loops such that upon restimulation with IL-33 responses are diminished. Our ongoing experiments aim to identify the specific innate cell types targeted by STAT6-IP and to better define its activity as an immunomodulatory peptide.
We thank the immunophenotyping platform of the Research Institute of McGill University Health Centre for excellent support for flow cytometry services. We also thank Fatima Hubaishi for help with quantification of the histology.
This work was supported by Canadian Institutes of Health Research Operating Grant PJT-162254 as well as U.S. Department of Defense Award W81XWH-15-1-0697 through the Peer-Reviewed Medical Research Program to E.D.F. Opinions, interpretations, conclusions, and recommendations are those of the author and are not necessarily endorsed by the U.S. Department of Defense. V.M. and L.L. were supported by studentships from the Research Institute of the McGill University Health Centre (to V.M.) and Les Fonds de Recherche du Québec–Santé (to V.M. and L.L.). The Meakins-Christie Laboratories, Research Institute of the McGill University Health Centre are supported in part by a Centre Grant from Les Fonds de Recherche du Québec–Santé.
V.M. and E.D.F. were responsible for the design of the experiments and writing the manuscript. V.M. performed all experiments (with help from V.G., L.L., H.Z., and J.S.) and analyzed the data, except for data presented in Supplemental Fig. 1, which was performed and analyzed by R.K., and Fig. 7, which was performed and analyzed by V.G. and R.K.
The online version of this article contains supplemental material.
Abbreviations used in this article:
alternatively activated Mϕ
bronchoalveolar lavage fluid
group 2 innate lymphoid cell
mean fluorescence intensity
respiratory syncytial virus
The authors have no financial conflicts of interest.