IRAK-M Associates with Susceptibility to Adult-Onset Asthma and Promotes Chronic Airway Inflammation

Asthma is a heterogeneous disease characterized by chronic airway inflammation and airway remodeling (1, 2), affecting patients of all ages (3, 4). Genome-wide association studies and positional cloning have identified multiple single nucleotide polymorphisms (SNPs) of several asthma-associated genes expressing in airway epithelium, illustrating the critical role of the respiratory epithelial barrier in the pathogenesis of asthma (5, 6). Epithelial cells, in conjunction with type 2 innate lymphoid cells, dendritic cells (DCs), macrophages, and other cell types, constitute an innate immune network in the airway. Activation of this immune network regulates the Th2/Th17–dominant adaptive immune response (7). A predisposition to asthma was shown to be related to the impaired host innate immunity (8). Moreover, genome-wide association studies reveal that the locus of IL-1R–associated kinase (IRAK) gene on human chromosome 12q14.2 is one of the most consistently replicated regions linked to asthma or asthma-related traits in diverse populations (9, 10). A positive association between IRAK-M SNPs and early-onset asthma was reported in an Italian population (11). However, genetic studies on IRAK-M SNPs in asthma or atopy have yielded inconsistent results in children and adult patients in different populations (12–14). In addition to asthma, several IRAK-M SNPs have been implicated in other acute inflammatory diseases, such as acute lung injury (15).

IRAK-M, primarily expressed by DCs, macrophages, and airway epithelium, is an intracellular negative regulator of TLRs that regulates the innate immune homeostasis (11, 16). Evolving evidence has shown IRAK-M also participates in asthmatic immune responses. For instance, the expression of IRAK-M was upregulated in peripheral blood leukocytes in asthmatic children (8), and high mRNA expression of IRAK-M in sputum was a marker for frequent exacerbations of asthma (17). A unique feature of IRAK-M–mediated regulation is that it varies in a disease stage–dependent manner (18). Accumulating evidence from us and others indicates that IRAK-M may exert contradictory roles, inhibitory or activating, in inflammatory and immune responses, depending on the duration of stimulation and the kind of stimuli. In our previous study using an animal model, we demonstrated that IRAK-M knockout (KO) enhanced the acute airway inflammation in response to 3 days of cigarette smoking but attenuated the chronic inflammatory responses to 7 weeks of cigarette smoking via differentially modulating the function of regulatory T (Treg) cells, Th17 cells, DCs, and macrophages (19). Similarly, others have demonstrated IRAK-M KO reduced the acute inflammatory responses to a noninfectious injury and subsequent fibrosis (20, 21). By contrast, overexpression of airway epithelial IRAK-M by IL-13 stimulation impaired the innate immunity and increased susceptibility to respiratory infection (22, 23). In particular, we have studied the regulatory effect of IRAK-M on the asthmatic airway inflammation using the acute OVA mouse model and found that IRAK-M ablation exacerbated airway inflammation by enhancing Th2 and Th17 differentiation (24).These results support the multifactorial roles of IRAK-M signaling in modulating inflammatory responses (25). However, the impact of IRAK-M on chronic asthmatic inflammation remains unclear.

To fill in the knowledge gaps, in the current study, we assessed the role of IRAK-M in adult-onset asthma and the development of chronic asthmatic inflammation. First, using samples from a hospital-based asthma cohort with Chinese Han ethnicity, we explored the genetic association between IRAK-M SNPs and human adult-onset asthma and the functionalities of susceptible IRAK-M SNPs. Next, using a chronic OVA mouse model, we investigated the regulation of IRAK-M expression following repeated allergen exposure and the influence of IRAK-M loss on the chronic allergic inflammation and airway remodeling.

There were a total of 1348 subjects with Han ethnicity from the northern region of China enrolled in this study. Six hundred and seventy-nine patients with adult-onset asthma were recruited from two medical centers. Adult-onset asthma was defined as previously described (26). The asthma patients, who had their first symptoms at the age of 18 y or older, were fully characterized by clinical presentation, functional status, and inflammatory markers. Asthma patients with forced expiratory volume in the first second (FEV1) ≥70% of predicted underwent the methacholine (Mch) challenge test. Patients whose FEV1 was <70% of predicted underwent an airway reversibility test by β2 agonist. Six hundred and sixty-nine age- and ethnicity-matched controls were selected from nonasthmatic, nonatopic, healthy individuals with normal lung function. All participants were unrelated. Detailed characteristics of these individuals were listed in Table I.

The study protocols were approved by the Ethics Committee for Human Research of Peking Union Medical College Hospital (S-767) and Beijing Aviation General Hospital (MHZYY 2014-05-01), and all subjects gave informed written consent.

Genomic DNA from peripheral blood leukocytes was extracted by standard protocols. A total of nine markers were genotyped. SNPs were genotyped using SNaPshot. The primers for PCR amplification and extension of IRAK-M gene are listed in Supplemental Tables 1 and 2. The PCR products were sequenced on an 3730xl DNA Analyzer (Applied Biosystems) and analyzed using the GeneMapper 4 software.

The genotyping efficiency for IRAK-M SNPs was >95%. Minor allele frequency was >5%. All SNPs tested were in Hardy–Weinberg equilibrium (p cutoff-value = 0.05) in asthma cases and controls.

Total RNA was extracted from PBMCs in asthmatics using TRIzol reagents (Invitrogen) and then reverse transcribed into cDNA using M-MLV Reverse Transcription System (Promega). Quantitative real-time PCR (qRT-PCR) was employed to analyze mRNA expression of IRAK-M by a Bio-Rad CFX96 Touch Real-Time PCR system using PowerUp SYBR Green Master Mix according to the manufacturer’s instructions (Applied Biosystems). The housekeeping gene GAPDH was used as internal control to normalize IRAK-M mRNA expression. All reactions were carried out in triplicate. The qRT-PCR primers were as follows: IRAK-M transcript variant 1 and 2, 5′-TCAGAGAAGAGTTATCAGGAAGGTG-3′ and 3′-AAGCTGATGGAAGATTTAAGCAGTA-5′; IRAK-M transcript variant 2, 5′-TGGCTGGATGTTCGTCATATTG-3′ and 3′-CTCCTGGAGGACCTGTAAAAGGT-5′; and GAPDH, 5′-GGAGCGAGATCCCTCCAAAAT-3′ and 3′-GGCTGTTGTCATACTTCTCATGG-5′.

Allele-specific alteration in rs1624395 binding to transcription factor c-Jun was assayed by using EMSA. Nuclear protein c-Jun was prepared from HeLa cells transfected with an expression pET-28α-His-MBP vector for c-Jun. EMSA were performed using dsDNA probes. Sequences of the oligonucleotides used were as follows: for IRAK-M rs1624395 allele A, forward 5′-ACAGAGCACAGAGGAGGAAATGATAACTTGGG-3′ and reverse 5′-CCCAAGTTATCATTTCCTCCTCTGTGCTCTGT-3′; for IRAK-M rs1624395 allele G, forward 5′-ACAGAGCACAGAGGAGGGAATGATAACTTGGG-3′ and reverse 5′-CCCAAGTTATCATTCCCTCCTCTGTGCTCTGT-3′. The 5′-biotin–labeled probes were synthesized by Invitrogen.

To ensure specificity of DNA–protein binding, an unlabeled probe was used in a 30-fold molar excess as a binding competitor before incubation of nuclear protein with biotin-labeled probes. A LightShift Chemiluminescent EMSA Kit was used according to manufacturer’s instructions (Thermo Fisher Scientific).

IRAK-M KO mice have been described previously elsewhere (16). IRAK-M KO mice (B6.129S1-Irak3^tmlFlv/J) were a kind gift by Dr. N.G. Frangogiannis (Albert Einstein College of Medicine), and wild type (WT) mice (purchased from Experimental Animal Research Center, Beijing, China) with C57BL/6 background were maintained in a pathogen-free mouse facility at Peking Union Medical College Hospital Animal Care Center. Ten- to twelve-week-old mice were used for experiments.

All experiments were performed according to international and institutional guidelines for animal care and approved by Peking Union Medical College Hospital Ethics Committee for Animal Experimentation.

The chronic asthma model was established according to that previously reported, with minor modifications (24, 27). Briefly, mice were given 50 μg of OVA (Grade V, Sigma-Aldrich) absorbed to 2.25 mg aluminum hydroxide in 200 μl of sterile saline on days 0 and 13 by i.p. injection, then challenged with 1% aerosolized OVA via a PARI nebulizer for 30 min daily from day 19 to day 24, then three times per week from day 26 to day 60. Sham mice received aluminum hydroxide and were exposed to 0.9% NaCl solution alone. Mice were sacrificed 24 h after the last challenge for further analysis.

Airway responsiveness to inhaled Mch was determined in mice 24 h after the final aerosol challenge, as we have previously described (24, 28).

Mice were sacrificed, and left lungs were taken out and inflated to 25 cm H₂O with 10% formalin, fixed overnight, then embedded in paraffin and sectioned to 3-μm thickness. H&E, periodic acid–Schiff (PAS), and Masson trichrome (MT) stains were performed to evaluate the severity of airway inflammation, mucus secretion, and collagen deposit.

Semiquantitative scoring for airway inflammation in H&E slides was evaluated in a blind manner by counting the inflammatory cells that infiltrated around airways and vessels according to the method of a previously reported study (29). The scores were as follows: 0 = normal; 1 = <3 cells thick in diameter; 2 = 4–10 cells thick; 3 = >10 cells thick. A score of 0–3 on each section was used to reflect overall extent of airway inflammation (0: normal, 1: <25% of each section, 2: 25–75%, 3: >75%). The index was calculated by multiplying severity by overall extent with a maximal score of 9.

Mucus scoring was evaluated based on the previously published method (30). Mucous scores were as follows: 0, no mucous; 1, a few cells secreting mucous; 2, many cells secreting mucous; and 3, extensive mucous production.

Mouse trachea were cannulated with a 20-gauge catheter, through which 0.8 ml of ice-cold PBS (pH 7.4) was used to lavage the lung three times. The recovered fluid was centrifuged at 1500 rpm for 5 min at 4°C. Cytospins were prepared from cell pellets.

qRT-PCR was used to reflect mRNA expression of cytokines, including IL-4, IL-13, IFN-γ, cytotoxic lymphocyte Ag 4 (CTLA-4), and TGF-β1. Total RNA was prepared from the right lung of the animal using TRIzol reagent and reverse transcribed into cDNA using a commercial kit (Takara Bio). qRT-PCR was performed on the 7500 Fast Real-Time PCR System (Applied Biosystems) in a 20-μl reaction that contained 2 μl of cDNA and SYBR Premix Ex Taq (Takara Bio). The expression level of GAPDH was used as an internal control. The relative expression was calculated with the comparative cycle threshold method using Applied Biosystems 7500 Real-Time PCR Software v2.0.6 (Applied Biosystems) and was expressed as the fold change compared with the control. The murine primers used to amplify genes were as follows: IRAK-M, 5′-GGGGCATCAACGAGCTATCC-3′ and 3′-GGTTCCAGAGGTCCAGGGTC-5′; IL-4, 5′-TCGAATGTACCAGGAGCCATAT-3′ and 3′-GAAGCACCTTGGAAGCCCTA-5′; IL-13, 5′-AGGCCAGCCCACAGTTCTA-3′ and 3′-CCACCAAGGCAAGCAAGAG-5′; IFN-γ, 5′-GCTGATCCTTTGGACCCTCT-3′ and 3′-AGGCTTTCAATGACTGTGCC-5′, CTLA-4, 5′-TCCGGAGGTACAAAGCTCAACTG-3′ and 3′-ACCATGGCTGCTAGCCAACAC-5′; TGF-β1, 5′-AATGGTGGACCGCAACAAC-3′ and 3′-GCACTGCTTCCCGAATGTC-5′; and GAPDH, 5′-TTGTCTCCTGCGACTTCAACA-3′ and 3′-TGGTCCAGGGTTTCTTACTCC-5′.

Commercial ELISA kits were employed to detect concentrations of IL-4, IL-13, and IFN-γ in bronchoalveolar lavage (BAL) fluid (R&D Systems) and levels of IL-17A, IL-33, and thymic stromal lymphopoietin (TSLP) in the supernatant of lung homogenates (eBioscience). Measurements of IL-10, IL-12p40, IL-23, and IL-27 were analyzed by LEGENDplex software (BioLegend).

Total collagen contents in the lung were determined by analysis of hydroxyproline. Hydroxyproline contents in the lung were measured using a commercial kit (Jiancheng Bioengineering Company, Nanjing, China).

The whole protein from the lung was prepared as previously described, with modifications (31). The lungs were harvested and homogenized in lysis buffer containing a protease inhibitor mixture (Sigma-Aldrich). Cell debris was removed by centrifugation at 12,000 × g for 10 min at 4°C. The supernatants were eluted by adding an equal volume of sample loading buffer, then run on 12% SDS–PAGE gels before transferring to PVDF membranes. Membranes were incubated in the presence of the indicated Abs. The Abs were as follows: anti-mouse IRAK-M (Abcam) and anti-mouse α-smooth muscle actin (α-SMA) (Boster Biological Technology, Wuhan, China). Primary Ab application was followed by incubation with HRP-conjugated secondary Abs (rabbit anti-mouse IgG) (Bioleaf Biotech, Shanghai, China). Immunopositive blots were visualized using an ECL with AlphaEaseFC imaging system (Alpha Innotech, San Leandro, CA) and semiquantitatively analyzed with AlphaEase FC software (Alpha Innotech).

Flow cytometry analysis was performed using single cell suspension. Single-cell suspension was prepared according to the method in a previously described study (24). Briefly, instilled lungs and spleens were minced and digested with collagenase type 1A and type IV bovine pancreatic DNase, then passed through a cell strainer to obtain cell suspension. The mononuclear cells from lung and spleen were isolated by centrifugation with Ficoll solution.

Cells were incubated with fluorochrome (FITC, PE, PerCP-Cyanine5.5, and allophycocyanin)–conjugated Abs (anti-Mouse CD4, CD8, ST2, MHC class II, CD80, CD86, OX40L, Siglec-F, Ly-6G, CD3, OX40L, programmed death–ligand 1 [PD-L1], and programmed death–ligand 2 [PD-L2]) at room temperature for 30 to 45 min. Cytokine intracellular staining was performed on cells that had been stimulated with 50 ng/ml PMA (Sigma-Aldrich) and 1 μM of ionomycin (Sigma-Aldrich) in the presence of GolgiStop (BD Biosciences) for 6 h. The activated cells were fixed, permeabilized, and stained with mAbs (anti-mouse IFN-γ, IL-4, IL-17A, T-bet, Foxp3, and ROR-γ). All samples were analyzed with Accuri C6 (BD Biosciences). The fluorochrome-conjugated Abs were purchased from eBioscience, BD Biosciences, or BioLegend, and the corresponding isotype control Abs were added to “isotype samples” at the same concentrations as the Abs of interest.

Bone marrow was obtained from C57BL/6 mice. The harvested bone marrow cells, depleted of RBCs, were incubated for 7 d with RPMI 1640 containing 10% FBS, 100 U/ml penicillin, 100 U/ml streptomycin, 20 ng/ml GM-CSF (for bone marrow–derived DCs [BMDCs]) or 50 ng/ml GM-CSF (for bone marrow–derived macrophages [BMDMs]). Purification of BMDCs was used a CD11c magnetic cell-sorting kit (Miltenyi Biotec, Gladbach, Germany). BMDMs were recovered from the adherent cells by digestion with 5% trypsin. Stimulation of BMDCs/BMDMs was performed with LPS (1 μg/ml) on day 8 for 24 h.

For association between IRAK-M SNPs and asthma susceptibility, all genotype and phenotype data were formatted using PLINK. Genotype and allele frequencies in cases and controls were compared by contingency table analysis. Genotype relative risk was calculated according to the statistical method described by Lathrop and colleagues (17). Quantile normalization was applied to all quantitative traits. Linkage disequilibrium and haplotypes were assessed using the Haploview version 4.0 (32). All markers were included in these analyses on a gene-by-gene basis.

Comparisons of the animal data were performed using one-way ANOVA with a Tukey post hoc test for multiple groups and Student t test for two groups with GraphPad Prism Version 5.01 (GraphPad Software, La Jolla, CA). Data are expressed as means ± SEM, and a p value <0.05 was considered significant.

Given the positive association of IRAK-M polymorphism with the susceptibility to early-onset asthma in children (11), we examined the impact of IRAK-M SNPs on the risk of developing adult-onset asthma (Table I). We selected nine IRAK-M SNPs that were studied in early-onset asthma (11). After correction for multiple tests, we found two IRAK-M SNPs, rs1624395 and rs1370128, were significantly associated with adult-onset asthma (odds ratio = 1.227, 95% confidence interval = 1.054 –1.428, p = 0.008 for rs1624395; odds ratio = 1.187, 95% confidence interval = 1.020 –1.381, p = 0.027 for rs1370128). Homozygote G/G for rs1624395 and homozygote C/C for rs1370128 were more often seen in adult-onset asthma patients than in controls (p values were 0.025 and 0.009, respectively) (Table II). We also demonstrated the associations between polymorphic markers in IRAK-M locus and numerous quantitative asthma-related traits, for instance the marker rs7970350 with high blood eosinophil counts (p < 0.05), rs2141709 with airflow obstruction (FEV1% and FEV1/forced vital capacity%, all p < 0.05), rs1821777 with FEV1% (p < 0.01) and serum total IgE level (p < 0.05), and rs1624395 with serum total IgE level (p < 0.05).

Haplotype analyses of all markers on a gene-by-gene basis (including haplotypes present in ≥5% of the genotyped population) revealed the significant association between one IRAK-M haplotype and asthma after 1000 permutations (GTAT, χ² = 0.787, permutation p = 0.0079). This haplotype was significantly under-represented in asthmatic cases in comparison with controls, with case and control frequencies being 0.098 and 0.138, respectively (Table III). This haplotype, therefore, appears to confer protection against the risk of having asthma.

The asthma-susceptible genes contain polymorphisms that modify function. We next studied the functionality of the risk for SNPs of IRAK-M influencing asthma susceptibility. We chose the top-associated polymorphism rs1624395, which is in linkage disequilibrium with rs1370128 (r² = 0.9) (Supplemental Fig. 1). We found that rs1624395 influenced the mRNA expression of the IRAK-M transcript variant 1 in peripheral monocytes from asthma patients (n = 60, AA = 30, AG + GG = 30). As shown in Fig. 1, asthmatics carrying AG/GG genotypes had significantly higher mRNA levels in blood monocytes than the AA genotype carriers.

To define the possible molecular mechanism underlying the difference in mRNA expression of IRAK-M by PBMCs in asthmatics carrying the rs1624395 A or G allele, we used EMSA analysis to detect the binding properties of rs1624395 G allele and A allele to NF C-Jun. c-Jun has been demonstrated to directly bind the human IRAK-M promoter to upregulate epithelial IRAK-M expression (22). Using the 5′-biotinylated probe against rs1624395 to coincubate with NF c-Jun protein from human HeLa cells, we demonstrated significantly greater affinity for allele G than allele A (Fig. 2A, 2B).

In an acute asthma mouse model, we have shown significantly higher mRNA expression of IRAK-M in the lungs exposed to OVA (around 50-fold) (24). We further examined the change of IRAK-M expression in a chronic asthma mouse model sensitized and challenged by OVA for 2 mo (Fig. 3A). The expression of IRAK-M mRNA was significantly higher (by ∼100-fold) in mouse lungs with chronic OVA treatment than those with PBS treatment (Fig. 3B, p < 0.01). Consistently, a significantly higher level of lung IRAK-M protein was observed in chronically OVA-treated mice than PBS-treated mice (Fig. 3C, p < 0.05).

In our previous study, we have demonstrated a clear inhibitory role of IRAK-M on the acute allergic airway inflammation with OVA asthmatic mouse model (24). However, this conclusion cannot be simply extrapolated to the regulation of chronic allergic inflammation, as the role of IRAK-M is known to vary between the acute and chronic inflammatory phases (25). In this study, we further assessed the impact of IRAK-M on chronic asthma, with a simulated model established by repeated OVA administration (Fig. 3A) to WT or IRAK-M KO mice. We first assessed the airway inflammation by examining the epithelial thickness and the inflammatory cell infiltration in the airway walls. We found no inflammation around the airways in PBS-treated WT or IRAK-M KO mice (Fig. 4A, left panel). Following chronic OVA challenge, WT mice showed worse pathological characteristics of airway inflammation than IRAK-M KO mice did, such as thicker airway epithelia, heavier mucus production, and more inflammatory cells around the bronchioles and vessels (Fig. 4A, right panel). We semiquantitatively scored the histopathological findings. The inflammation scores of WT mice were significantly higher than those of IRAK-M KO mice (3.28 ± 0.40 versus 2.47 ± 0.39, p < 0.05) (Fig. 4B). Consistently, the number of total inflammatory cells, eosinophils, neutrophils, lymphocytes, but not macrophages in BAL fluid was significantly more in WT mice than IRAK-M KO mice (Fig. 4C). Using flow cytometry analysis, we found a significantly higher percentage of Th2 cells in the BAL fluid from WT mice than that in IRAK-M KO mice (Fig. 4D), even though the percentages of total T cells or B cells were not different between the two groups.

Airway remodeling is one of characteristics of chronic asthma and is related to airway obstruction (2, 33). We subsequently examined the impact of IRAK-M KO on airway reactivity and airway remodeling in mice of the chronic OVA model. Following chronic OVA challenge, WT and IRAK-M KO mice developed increased airway resistance in response to the increasing concentrations of Mch. Of note, the airway resistance was significantly higher in WT mice as compared with the IRAK-M KO mice (p < 0.05) (Fig. 5A).

To reflect the effect of IRAK-M deficiency on airway pathology (33), we performed PAS scorings, MT staining, and α-SMA expression. We found the following after chronic OVA challenge: 1) more severe mucus hypersecretion in the airways from the WT mice than that from IRAK-M KO mice (Fig. 5B), 2) significant higher PAS scores in WT mice than IRAK-M KO mice (2.78 ± 0.44 versus 1.84 ± 0.38, p < 0.05) (Fig. 5B), 3) more collagen deposition in the peribronchiolar area by MT staining and higher hydroxyproline content in the WT lungs than IRAK-M KO lungs (0.60 ± 0.11 μg/mg versus 0.41 ± 0.09 μg/mg, p < 0.05) (Fig. 5C), and 4) higher expression of α-SMA in WT lungs than IRAK-M KO lungs by Western blotting assay (Fig. 5D).

Asthma is typically characterized by Th2-type inflammation in the airways (7). However, the immune regulation of asthmatic inflammation involves a complex entity of T cell subtypes. Therefore, we further characterized the influence of IRAK-M KO on the composition of T subsets in mouse lung homogenates and spleens of the chronic OVA model. As shown in Fig. 6A, chronic OVA challenge resulted in an increased Th2 cell percentage and a reduced Th1 cell percentage in both WT and IRAK-M KO mouse lungs. However, the changes in Th1 cell and Th2 cell percentage were significantly less in IRAK-M KO mouse lungs than WT mouse lungs. Additionally, we observed significantly higher percentage of Th1 cells in the spleens from OVA-treated IRAK-M KO mice than WT mice (Fig. 6B). Moreover, chronic OVA treatment significantly increased the Th17 and Treg cell percentage in both IRAK-M KO and WT mouse lungs, with no significant intergroup difference (Fig. 6A, 6B).

Asthmatic airway inflammation is mediated by cytokines produced from structural and immune cells (1). To understand the mechanisms by which IRAK-M supports the Th2 immune responses following chronic OVA stimulation, we examined the levels of both proinflammatory and anti-inflammatory cytokines in BAL fluid and lungs of chronically immunized and challenged WT and IRAK-M KO mice.

The airway epithelium constitutes the first line of defense against the environmental stimuli. Upon exposure to an inhaled allergen, airway epithelial cells release the cytokines to induce resident DC maturation and further skew the naive T cells differentiation to Th2 type (1, 34). In the acute mouse model of asthma, we have demonstrated OVA-induced elevation of epithelium-derived Th2 instructive cytokines, IL-17A, IL-33, and TSLP, with a higher level in the lungs of IRAK-M KO mice than WT mice (24). However, we didn’t find a significant difference in the elevated concentrations of these cytokines in the lung homogenates between WT and IRAK-M KO mice following chronic OVA administration (Fig. 7A).

At protein level, the concentrations of Th2 cytokines IL-4 and IL-10 but not IL-13 were significantly higher in BAL fluid from chronically OVA-treated WT mice, whereas the concentration of Th1 cytokine IFN-γ was significantly higher in BAL fluid from chronically OVA-treated IRAK-M KO mice (Fig. 7B). Moreover, we found that chronic OVA-stimulation caused significantly higher levels of IL-12p40, IL-23, and IL-27 in BAL fluid from IRAK-M KO mice than those from WT mice. At mRNA level, lung expression of IL-4 and IL-13 was significantly increased in WT mouse lungs but not in IRAK-M KO mouse lungs by chronic OVA stimulation. By contrast, mRNA expression of IFN-γ and CLTA-4 in the lungs was significantly elevated in IRAK-M KO mice (Fig. 7C). TGF-β1, increased in the airways of patients with asthma (35, 36), mediates the association between airway remodeling and airway hyperresponsiveness during asthmatic process (37). In supporting of this finding, we found that lung expression of TGF-β1 mRNA was significantly enhanced by chronic OVA challenge in WT mice but not in IRAK-M KO mice (Fig. 7C). Together, these data suggested IRAK-M KO suppressed the Th2 inflammation and promoted Th1 phenotype exampled by a greater INF-γ production.

To further understand the mechanism underlying the suppression of chronic allergic airway inflammation by IRAK-M KO, we further evaluated the function of IRAK-M KO on APCs in the chronic asthmatic inflammation.

Airway APCs, mainly DCs and macrophages, play crucial roles not only to initiate the specific Th2 cell immune responses but also to restimulate local residential T (memory) lymphocytes and recruit effector cells following the repeated allergen exposure (1, 34). We have previously demonstrated that IRAK-M KO led to a downregulation of costimulatory molecules, such as OX40L, CD80, CD86, and CD124, on the airway APCs in acute asthma model of mice (24). However, in the current study of chronic OVA model of asthma, we found that IRAK-M KO did not make a difference on the expression of costimulatory molecules MHC class II, CD80, and CD86 on lung DCs despite a higher expression of CD80 by airway macrophages of IRAK-M KO mice (Supplemental Fig. 2). Notably, the expression of PD-L1 (CD274) and PD-L2 (CD273) was significantly higher on the lung DCs from IRAK-M KO mice than WT mice after chronic OVA exposure (Fig. 8A, p < 0.05). Similarly, a higher surface expression of PD-L1 (CD274) was revealed on the BAL fluid macrophages from IRAK-M KO mice as compared with WT mice (Fig. 8B).

To confirm that the absence of IRAK-M modifies the instructive role of APCs on T cell responses, we examined the cytokines released by activated WT and IRAK-M KO DCs and macrophages. We used BMDCs and BMDMs and stimulated them with LPS, a specific TLR4 agonist. We found that LPS induced a significant elevation of IRAK-M expression on WT BMDCs and BMDM (Fig. 9A).

A lack of IRAK-M did not make a difference in the CD124, CD206, and OX40L expression of BMDC or BMDM to LPS stimulation (data not shown). Instead, IRAK-M KO led to a significantly higher level of IFN-γ, TNF-α, and IL-6 production by BMDCs and BMDMs to LPS stimulation (Fig. 9B, 9C). We did not find a significant difference in the LPS-induced IL-13 production between WT and IRAK-M KO BMDCs or BMDMs (data not shown). These data support that loss function of IRAK-M in DCs or macrophages favors Th1 promotion.

In the present study, we reported an association of two IRAK-M SNPs, rs1642395 and rs1370128, with susceptibility to adult-onset asthma and one haplotype (GTAT) of IRAK-M potentially protecting against the risk of asthma. These findings provide evidence for the hypothesis that IRAK-M plays a role in the pathogenesis of adult-onset asthma. It is by far, to our knowledge, the first genetic study of IRAK-M as an adult-onset asthma susceptible gene in Han Chinese ethnicity and provides invaluable supplementary information to the comprehensive role of IRAK-M in the pathogenesis of asthma.

The susceptible IRAK-M SNPs reported in this study are different from the original findings from early-onset persistent asthmatic patients (11). Several factors could attribute to the discrepancy. First, the patient age, ethnicity, and disease stage are different among the study groups. Asthma susceptibility is the result of interaction of genetic profiles and environmental factors at various times throughout the entire life. Some susceptible genes were reported to cause early-onset mild or intermittent asthma under the risk of environmental exposure (6). Genetic evidence has shown different gene expression profiles in the airways between adult- and childhood-onset severe asthma (38). Moreover, case-control association and next-generation sequencing studies have revealed significant discrepancies in the frequencies of SNPs and haplotype blocks for asthma genes between Chinese and other ethnicities (39). Second, the inconsistencies among studies could reflect real differences in the genetic effects of IRAK-M variants or linkage disequilibrium patterns between typed and true causative variants (11, 14, 40). Also, we found in the current study that genetic markers in IRAK-M were closely associated with asthma susceptibility but not linked to asthma-related phenotypes. A similar phenomenon has been reported in the previous genetic studies of the asthmatic populations (6).

Many asthma-associated SNPs are located in noncoding intergenic and intronic regulatory regions. These SNPs may modify mRNA expression and transcription factor binding (41, 42). In the current study, we analyzed the top-associated SNP rs164395 and identified the risk allele G that linked to significantly higher mRNA expression of IRAK-M in PBMCs from asthmatics. Using EMSA, we reveal the potential underlying mechanism is that G allele, as compared with A allele, has a higher binding capacity to C-Jun, a transcription factor directly binding to the human IRAK-M promoter (22).

IRAK-M is known to regulate the cytokine production of monocytes/macrophages via intervening the intracellular the TLR/NF-κB signaling pathway (16). We have previously reported that IRAK-M deficiency exacerbated the OVA-induced acute airway inflammation through multiple pathways, such as promoting the production of proinflammatory cytokines by airway epithelial cells, enhancing the activities of DCs and macrophages, and deviating the CD4⁺ T cells differentiation to Th2 and Th17 cells (24). But, in the current study, we demonstrated an inhibitory effect of IRAK-M deficiency on the OVA-induced chronic mouse airway inflammation. Mechanistically, we found that IRAK-M KO could reduce the Th2/Th1 imbalance in the chronic allergic inflammation by increasing the Th1 cell percentage and Th1 cytokines (IFN-γ, IL-12p40, IL-23, and IL-27) production. The enhanced Th1 function can inhibit the Th2 cell proliferation and activity (1). Indeed, constant overexpression of airway epithelial IRAK-M could impair the innate immunity and compromise the capability of clearance of inhaled environmental Ags from the airways (22, 23, 43). Chronic stimulation of innate immunity is a major contributor to the persistent airway inflammation and irreversible airway remodeling (1, 34). At the cellular level, our in vitro data demonstrated IRAK-M KO BMDCs and BMDMs secreted more Th1 cytokines after TLR ligation by LPS. A similar finding, that IRAK-M inhibited the Th1 cytokine production by DCs, has been reported previously (44). Hence, the impact of IRAK-M KO on the chronic allergic airway inflammation could be attributed to the alternation of DC/macrophage–medicated T cell differentiation.

The mechanism underlying the distinct regulation of IRAK-M on the DC function at different inflammatory stages is not clear. Our studies suggested it could act via modulating the costimulatory signaling molecules on the DCs differentially. There is mounting evidence demonstrating that differentiation of Th cells relies on particular costimulatory molecules signaling from APCs (45). In the current study, we observed that IRAK-M KO led to a significantly higher expression of PD-L1 (CD274) and PD-L2 (CD273) on the lung DCs or macrophages following chronic OVA stimulation. The surface molecules PD-L1 (CD274) and PD-L2 (CD273) of DCs were known to present the inhibitory signal to T cell activation (46). Particularly, PD-L2 (CD273) was shown to attenuate Th2-predominant allergic response through an IFN-γ–dependent mechanism (47). Previous studies have demonstrated that CTLA-4 interacting with CD80 inhibited Th2 differentiation (48, 49). In supporting with these findings, our data showed that IRAK-M KO mice had a higher expression of CTLA-4 mRNA by lung and CD80 by airway macrophages after chronic OVA challenge.

In addition, IRAK-M also regulates the tissue remodeling process accompanying chronic inflammation. In an animal model of bleomycin-induced lung injury, IRAK-M was shown to promote inflammatory responses and collagen deposit in the lungs at the late stage (20). In another study of cardiac remodeling following myocardial infarction, IRAK-M was demonstrated to prompt the myofibroblast conversion via inducing the TGF-β1–mediated α-SMA expression in the fibroblasts (50). Similarly, in the current study, we found that IRAK-M loss attenuated the airway remodeling and reduced the α-SMA and TGF-β1 expression in the lung tissue with chronic allergic airway inflammation.

Our study has a few limitations. For instance, this study only included adult patients and lacked child asthmatics as replicating control. Therefore, we cannot conclude if the polymorphisms of IRAK-M identified in this study are an age-specific risk of asthma susceptibility or asthma-related traits. However, this current study demonstrated the functionalities of the associated IRAK-M SNPs, suggesting the importance of IRAK-M in asthma pathogenesis. In addition, the mice used in this study underwent whole-body IRAK-M KO, which might introduce some unrecognized confounding variables to impact the experimental results. An allelic-specific and temporally controllable KO model as well as an overexpression animal model is desired to precisely study the complexity of asthma.

In conclusion, we demonstrated in the current study that IRAK-M was genetically linked in adult-onset human asthmatics, and it has a proinflammatory effect in an animal model using mouse lungs following chronic allergen stimulation. This action of IRAK-M is opposite to its anti-inflammatory effect in the acute phase of asthma. The underlying mechanism of these findings is related to promoting the effect of IRAK-M on DC-mediated, Th2-deviated immune activation following a repetitive allergen stimulation. These findings illustrated the complexity of immune regulation at different stages of asthmatic inflammation. Moreover, it implicates that downregulating IRAK-M and the downstream signaling pathway may provide an alternative therapy to control chronic asthmatic airway inflammation.

We thank Prof. Nikolaos G. Frangogiannis for kindly gifting IRAK-M KO mice. We thank all subjects for agreeing to participate in this study. We thank Dr. Kewu Huang for the help in the FlexiVent measurement of airway reactivity. We thank Prof. Roy Pleasants for English editing.

The authors have no financial conflicts of interest.