|Home | About | Journals | Submit | Contact Us | Français|
Influenza virus is a common cause of respiratory infection and morbidity, which is often due to deleterious host immune responses directed against the pathogen. We investigated the role of IL-1 receptor-associated kinase-M (IRAK-M), an inhibitor of MyD88-dependent TLR signaling, in modulating the innate inflammatory response during influenza pneumonia using a murine model. The intranasal administration of influenza resulted in the upregulation of IRAK-M mRNA and protein levels in the lungs within 2 d after infectious challenge. Pulmonary influenza infection in mice deficient in IRAK-M (IRAK-M−/−) resulted in substantially increased mortality compared with similarly treated wild-type animals. Increased mortality in IRAK-M−/− mice was associated with enhanced early influx of neutrophils, high permeability edema, apoptosis of lung epithelial cells, markedly increased expression of inflammatory cytokines/chemokines, and release of neutrophil-derived enzymes, including myeloperoxidase and neutrophil elastase. Early viral clearance was not different in mutant mice, whereas viral titers in lungs and blood were significantly higher in IRAK-M−/− mice compared with wild-type animals. Increased lethality observed in IRAK-M−/− mice after influenza challenge was abrogated by Ab-mediated blockade of CXCR2. Collectively, our findings indicate that IRAK-M is critical to preventing deleterious neutrophil-dependent lung injury during influenza infection of the respiratory tract.
Influenza virus infection is a major cause of respiratory disease worldwide. Pneumonia is one of the most important complications associated with influenza (1–5) and influenza-associated pneumonia can be due either to primary infection or predisposition to the development of secondary bacterial pneumonia (3, 5–7). Excessive mortality in the elderly is often due to such complications during influenza seasons.
In murine pneumonia models, influenza infection results in the release of cytokines, chemokines, platelet-activating factor-related molecules, and enhanced expression of TLRs required for protective innate and acquired immunity against influenza (5–12). However, tissue damage in primary influenza pneumonia is believed to be partially caused by deleterious inflammation mediated by host-derived cytokines and chemokines (5, 6, 13). For example, we have found enhanced severity of influenza pneumonia in patients with increased serum levels of cytokines and high-mobility group box chromosomal protein-1 (7, 14). These results suggest that specific immunological reactions, especially innate immune responses, must be appropriately curtailed to limit collateral damage to lung structures in influenza pneumonia. Molecular mechanisms negatively regulating innate responses in influenza pneumonia are incompletely described.
TLRs are a critical family of pathogen recognition receptors that allow host cells to respond to a broad array of microbial pathogens. IL-1 receptor-associated kinase-M (IRAK-M or IRAK-3) is a protein in the IRAK family that has been shown to be an important negative regulator of TLR-mediated cell signaling (15–18). In contrast to IRAK-1 and IRAK-4, IRAK-M lacks kinase activity and negatively regulates signaling through MyD88-dependent TLRs, including TLR2, TLR4, TLR7/8, and TLR9. This protein has been shown to regulate critical aspects of innate immunity, including the development of endotoxin tolerance and sepsis-induced reprogramming of macrophage phenotype (15–17). For example, abdominal sepsis results in an impairment in innate lung antibacterial responses to Pseudomonas aeruginosa and sepsis-induced alterations in innate immunity are largely reversable in mice that are genetically deficient in IRAK-M (18). IRAK-M was initially believed to be expressed only by cells of monocytic lineage. However, emerging data suggest that cells of myeloid or nonmyeloid origin express biologically active IRAK-M. The role of this inhibitory protein in regulating key aspects of immunity against viral pathogens is unknown. In this study, we investigated the contribution of IRAK-M to host immunity during experimental murine influenza pneumonia.
A colony of IRAK-M–deficient mice bred on a B6 background for more than eight backcrosses was established at the University of Michigan (18). Age- and sex-matched specific pathogen-free 6- to 8-wk-old C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The animals were housed in specific pathogen-free conditions within the University of Michigan Animal Care Facility (Ann Arbor, MI) until the day of death. All animal experiments proceeded in accordance with National Institutes of Health policies on the human care and use of laboratory animals and were approved by the University Committee on Use and Care of Animals (UCUCA) at the University of Michigan.
A mouse-adapted influenza A virus strain (strain A/PR8/34: H1N1 isotype, kindly provided by Drs. K. Kuroda and K. Shimizu, Nihon University, Tokyo, Japan) was inoculated into mice as described. Briefly, each mouse was inoculated intranasally with 50 μl influenza virus (105 PFU/μl) per 20 g mouse. Inoculated viral titers were corrected for mouse weight. Saline was inoculated intranasally into mock-infected mice. Whole lungs were harvested at 2, 4, 6, or 7 d after influenza virus infection. Paraffin embedding and tissue staining with H&E were performed using standard methodologies (9).
Mice were euthanized by CO2 inhalation at various times. The pulmonary vasculature was perfused with 1 ml PBS containing 5 mM EDTA via the right ventricle, or blood was collected in a heparinized syringe from the right ventricle before the lungs were removed. Lymph nodes were carefully dissected away and then the lungs were homogenized in 1 ml PBS containing protease inhibitor (Roche Diagnostics, Indianapolis, IN). Homogenates and blood were then serially diluted in medium and plated on Madin-Darby canine kidney cell cultures to assay viral titers as PFU (9, 10, 13)
Bronchoalveolar lavage (BAL) was performed for assessment of leukocyte recruitment as previously reported (9, 18). The trachea was exposed and intubated using a 1.7-mm outer diameter polyethylene catheter. BAL was performed by instilling PBS containing 5 mM EDTA in 1-ml aliquots. In each mouse, ~4 ml PBS was instilled and 3 ml lavage fluid was retrieved. Lavaged cells from each group of animals were pooled and counted after lysis with hypotonic solution. Cytospins (Thermo Electron Corp. Waltham, MA) were prepared for determination of BAL differentials using a modified Wright stain.
Lungs were removed from euthanized animals and leukocytes were prepared as described (18). Briefly, lungs were minced with scissors to a fine slurry in 15 ml digestion buffer (RPMI 1640, 10% FCS, 1 mg/ml collagenase [Roche Diagnostics], 30 μg/ml DNase [Sigma-Aldrich, St. Louis, MO]) per lung and enzymatically digested for 30 min at 37°C. Any undigested fragments were further sheared by repeatedly withdrawing and ejecting the suspension through a 10-ml syringe. The total lung cell suspension was pelleted, re-suspended, and separated by centrifugation through a 40% Percoll gradient to enrich leukocytes.
Cells were isolated from lung digests as described previously, and cells were recovered at various times after influenza virus administration. Isolated leukocytes from lung digests were stained with the following FITC- or PE-labeled Abs to analyze T cell subsets: anti-CD4, anti-CD8, and anti-CD69 (BD Biosciences, San Jose, CA). Cells were collected on a FACScan or FACSCalibur flow cytometer (BD Biosciences) using CellQuest software version 3.2 (BD Biosciences). Data were analyzed using the CellQuest software package.
Primary murine pulmonary macrophages (PMs) and type II alveolar epithelial cells (AECs) were isolated using the method described previously (18, 19). Briefly, PMs were isolated from dispersed lung homogenates by adherence purification and plated at a density of 5 × 105 cells/ml. For the isolation of murine AECs, the pulmonary vasculature was perfused via the right ventricle with 1 ml PBS. The lungs were then filled via the trachea with 1–2 ml dispase (Worthington, Lakewood, NJ). Subsequently, 0.45 ml low-melting point agarose was infused via the trachea and the lungs were placed on ice for 2 min to harden the agarose. The lungs were then submerged in dispase for 45 min at 24°C before the lung tissue was teased from the airways and minced in DMEM with 0.01% DNase. After swirling for 10 min, followed by passage through a series of nylon filters, the cell suspension was collected by centrifugation and incubated with biotinylated Abs (anti-CD32 and anti-CD45; BD Pharmingen). After incubation with streptavidin-coated magnetic particles, myeloid cells were removed using a magnetic tube separator. Mesenchymal cells were removed by overnight adherence in a Petri dish. The nonadherent cells after this initial plating were plated at a density of 2 × 105 to 2 × 106/cm2 on plastic dishes coated with fibronectin and maintained in DMEM with penicillin/streptomycin and 10% FCS. Cells were washed with PBS for 1 h after plating. This technique routinely generates 5–6 × 106 cells/mouse, with a final adherent population that includes ≤4% nonepithelial cells (by intermediate filament staining). Cells (PMs or AECs) in culture were incubated with influenza A virus at a multiplicity of infection of 1.0, PBS (mock), or LPS at 100 ng/ml (positive control) and culture supernatants were collected 18 h later for measurement of selected cytokines by ELISA.
Whole lungs were harvested, immediately snap-frozen in liquid nitrogen, and stored at −70°C. RNA was isolated and real-time quantitative RT-PCR was performed as previously described (18). Measurement of gene expression was performed using the ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA). Briefly, primers and probes for β-actin and targeted molecules were designed using Primer Express software (Applied Biosystems). The primers, placed in different exons, were tested to ensure that they do not amplify genomic DNA. The following primers and probe nucleotide sequences were used: for murine IRAK-M: forward 5′-TGAGCAACGGGACGCTTT-3′, reverse 5′-GATTCGAACGTGCCAGGAA-3′, probe 5′-TTACAGTGCACAAATGGCACAACCCC-3′; for murine TNF-α: forward 5′-CAGCCGATGGGTTGTACCTT-3′, reverse 5′-TGTGGGTGAGGAGCACGTAGT-3′, probe 5′-TCCCAGGTTCTCTTCAAGGGACAAGGC-3′; murine MIP-2: forward 5′-GAACATCCAGAGCTTGAGTGTGA-3′, reverse 5′-CCTTGAGAGTGGCTATGACTTCTGT-3′, probe 5′-CCCCCAGGACCCCACTGCG-3′; for murine keratinocyte-derived chemokine (KC): forward 5′-CAAGAACATCCAGAGCTTGAAGGT-3′, reverse 5′-GTGGCTATGACTTCGGTTTGG-3′, probe 5′-TTGCCCTCAGGGCCCCACTG-3′; for murine IFN-α1: forward 5′-CCTGGCGGTGCTGAGCTA-3′, reverse 5′-TCCTGAGGTTATGAGTCTGAGGAA-3′, probe 5′-TGGCCAACCTGCTCTCTAGGATGTGAC-3′; for murine IFN-β1: forward 5′-CTGCGGCCTAGCTCTGAGA-3′, reverse 5′-CAGCCAGAAACAGCCATGAG-3′, p robe ′-CA- 5 CACTGCATCTTGGCTTTGCAGCTCT-3′; for murine caspase-3; forward 5′-CTGGACTGTGGCATTGAGACA-3′, reverse 5′-CAGCCTCCACCGG-TATCTTC-3′, probe 5′-AGTGGGACTGATGAGGAGATGGCTTGC -3′; and for murine β-actin: forward 5′-CCGTGAAAAGATGACCCAGATC-3′, reverse 5′-CACAGCCTGGATGGCTACGT-3′, probe 5′-CACACTGCATCTTGGCTTTGCAGCTC -3′. Specific thermal cycling parameters used with the TaqMan One-Step RT-PCR Master Mix Reagents kit (Applied Biosystems) included 30 min at 48°C, 10 min at 95°C, and 40 cycles involving denaturation at 95°C for 15 s and annealing/extension at 60°C for 1 min. Relative quantitation of cytokine mRNA levels was plotted as fold change relative to untreated control cells of the lungs. All experiments were performed in duplicate.
Myeloperoxidase (MPO) activity in the cells from lung homogenates was detected by a previously reported methodology (10, 11). Neutrophil elastase (NE) activity in BAL fluid (BALF) was determined using the synthetic substrate Suc-Ala-Ala-Pro-Val pNA, which is highly specific for NE, as described previously (11). Briefly, samples were incubated in 0.1 M NaCl and 1 mM substrate for 24 h at 37°C and the amount of pNA released was measured spectrophotometrically at 405 nm and was considered to represent the NE activity.
Proteins in the lung homogenates were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes (Millipore, Bedford, MA). The membranes were blocked with 5% skimmed milk in TBS (pH 7.2–7.4) containing 0.05% TBST and incubated with rabbit Abs directed against IRAK-M (Abcam, Cambridge, U.K.; diluted 1:1,000) or β-actin (Sigma-Aldrich; diluted 1: 5000) for 1 h at room temperature. The membranes were then washed with TBST and incubated with HRP-conjugated anti-rabbit IgG (Abcam; diluted 1:20,000) for 1 h at room temperature. After two additional washes, the signals were developed with an ECL Plus Western blot detection kit (Amersham, Arlington Heights, IL).
Murine cytokines/chemokines (TNF-α, MIP-2, and KC; R&D Systems, Minneapolis, MN), and albumin (Albumin Quantification Kit; Bethyl Laboratories, Montgomery, TX) for lung permeability assessment were quantified using a modified double ligand method as described (18).
For detection of apoptotic cells in the lungs of mice infected with influenza virus, tissue sections were analyzed using the TUNEL assay. After digestion with 0.5% trypsin, sections were treated with TUNEL reaction mixture using the In Situ Cell Death detection kit (Boehringer, Mannheim, Germany) for 1 h at 37°C in the dark. Slides were then rinsed three times with 1× PBS and incubated with alkaline phosphatase-conjugated FITC-labeled Ab for 30 min at 37°C. Sections were then washed and developed with nitro-blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl phosphate p-toluidine.
Polyclonal goat CXCR2 Ab was generated by immunization of goats with a 17-aa peptide consisting of the ligand binding site and Ab purified by standard techniques for in vivo use. For neutralization experiments, anti-CXCR2 Ab or control goat polyclonal IgG (1 mg) was injected i.p. 2 h before inoculation of the virus and then 500 μg was injected 2 d after inoculation of the virus as previously described (20–22).
All data are expressed as means ± SD and analyzed using StatView software (Abacus Concepts, Cary, NC). The significance of differences between or among groups was examined using Student t test or an ANOVA, followed by Tukey or Dunnett tests. Kaplan-Meier analysis was used for analysis of survival rates. Differences in viral titers were tested by geometric analysis. A p value < 0.05 was considered to indicate a statistically significant difference.
We first examined the expression of IRAK-M in the lungs of wild-type (WT) C57B/6 mice infected with influenza virus. We found that influenza virus infection resulted in a maximal 2-foldupregulationof IRAK-M mRNA in lung homogenates at day 2 after inoculation, with decline by day 7 (Fig. 1A). Similarly, the expression of IRAK-M protein in lungs was most abundant 2 d after viral administration, with a return to baseline levels by day 7 (Fig. 1B).
To determine whether IRAK-M contributed to the regulation of innate immune responses in lungs during influenza infection, we administered 1 × 105 PFU of influenza virus to IRAK-M−/− and WT mice, then assessed survival out to 14 d. Compared with WT mice, IRAK-M−/− animals appeared more visibly ill, exhibiting lethargy, ruffled fur, and reduced oral intake during the several days after influenza virus infection. Long-term survival in WT mice was ~40% and no WT mice died before day 8. In contrast, IRAK-M−/− mice died much more rapidly and no survival was observed in infected IRAK-M−/− mice past day 7 after viral challenge (Fig. 2A, p < 0.001).
Experiments were next performed to determine whether the decreased survival observed in IRAK-M−/− animals was due to impairment of viral clearance after influenza administration. Lung and blood viral titers were quantitated in WT and IRAK-M−/− mice at 2 and 6 d after viral inoculation. There was no difference in lung or blood viral titers noted between IRAK-M−/− and WT mice at day 2 after intranasal influenza administration (Fig. 2B, left panel). However, by day 6, IRAK-M−/− mice had ~10- and 100-fold greater influenza PFU in blood and lungs, respectively, than observed in infected WT mice (Fig. 2B, right panel).
Histological examination of the lungs of WT mice infected with influenza virus revealed mild bronchopneumonia on day 2 post-infection (Fig. 3Ab, 3Af, 3Aj), whereas IRAK-M−/− mice showed evidence of severe pneumonia (Fig. 3Aa, 3Ae, 3Ai). At day 6, bronchopneumonia in WT mice was more readily apparent(Fig. 3Ad, 3Ah, 3Al). However, IRAK-M−/− mice showed not only severe pneumonia, but also evidence of extensive diffuse alveolar damage with alveolar hemorrhage (Fig. 3Ac, 3Ag, 3Ak), including detachment of respiratory epithelium, airspaces filled with inflammatory cells and RBCs, and destruction of alveolar architecture.
To quantitate the magnitude of alveolar injury, the permeability of the alveolar-capillary membrane was assessed by measurement of albumin concentrations in BALF. At day 2, BALF albumin levels were significantly increased in IRAK-M−/− mice compared with that observed in WT mice (p < 0.01, Fig. 3B). Alveolar permeability was greatest at day 6 after viral infection, with BALF albumin levels 3-fold greater in IRAK-M−/− mice than in similarly treated WT animals (Fig. 3B).
We next quantitated differences in lung inflammatory cell accumulation and activation in WT and IRAK-M−/− mice during the evolution of severe influenza pneumonia. No differences in total BAL leukocytes, polymorphonuclear cells (PMN), or mononuclear cells were observed in uninfected WT and IRAK-M−/− mice at baseline (data not shown). However, as shown in Fig. 4A, afer administration of the influenza virus, we observed a time-dependent increase in total leukocytes in BALF at days 2, 4, and 6, which was significantly greater in infected IRAK-M−/− mice than in WT mice. Of the total BAL leukocyte populations, neutrophils were substantially increased in IRAK-M−/− mice compared with WT mice on all days. In contrast, the accumulation of BAL mononuclear cells (monocytes/macrophages and lymphocytes) after influenza infection was not different between IRAK-M−/− and WT mice.
Given the importance of T cells, especially CD8+ T cells, in antiviral immunity, we examined the accumulation and activation of specific lymphocyte populations using four-color flow cytometry of leukocyte suspensions obtained from collagenase-digested whole lungs. As shown in Fig. 4B, we noted an increase in the number of both CD4+ and CD8+ T cells in the lungs of mice after intranasal influenza administration, which was maximal at day 6 after viral infection. The numbers of CD4+ and CD8+ T cells were significantly greater in infected IRAK-M−/− mice, especially at day 6 (p <0.01 as compared with WT mice). Moreover, greater numbers of activated CD4+ and CD8+ T cells (as indicated by coexpression of the activational marker CD69) were noted in IRAK-M–deficient mice.
Our previous studies have identified the importance of inflammatory cytokines and chemokines in severe influenza pneumonia (6, 9, 10, 12, 14) and the expression of these mediators has been causally linked to influenza-induced lung damage (5, 6, 13). We therefore examined whether IRAK-M was required to modulate the expression of inflammatory cytokines/chemokines during influenza pneumonia. WT and IRAK-M−/− mice were challenged with influenza virus, then their lungs were harvested at days 2 and 6 to quantitatively assess the mRNA and protein expression of TNF-α, the neutrophil-active ELR+ CXC chemokine MIP-2, and KC. Administration of influenza to WT mice resulted in an induction of TNF-α, MIP-2, and KC mRNA (Fig. 5A) and protein (Fig. 5B), peaking at 2 d and returning toward baseline levels by 6 d. By comparison, IRAK-M−/− mice expressed considerably greater quantities of TNF-α and CXC chemokine mRNA and protein, particularly at early time points (day 2).
We next analyzed expression of type 1 IFNs, IFN-α1 and -β1, to further explore their possible role in antiviral immunity and excessive host immune responses known to be mediated through the RIG-1/TLR3 pathway (23–25). At day 2 after influenza administration, both IFN-α1 and -β1 mRNA expressions were significantly enhanced in IRAK-M−/− mice compared with WT mice (Fig. 5C). Type 1 IFN mRNA levels decreased to near baseline expression by day 6 in both infected WT and IRAK-M mutant mice.
In addition to enhanced PMN influx, we also found more abundant release of MPO and NE in IRAK-M−/− mice; PMN products that are injurious to host tissues (11) (Fig. 5D). The activity of MPO and NE was minimal in homogenates from un-infected lungs, but increased in influenza-infected lungs recovered from WT mice. The activity of MPO and NE was significantly enhanced in influenza-infected lungs of IRAK-M−/− mice, as compared with WT mice, particularly at day 2 after infectious challenge.
To determine possible cellular sources of robust cytokine and chemokine production observed in IRAK-M−/− mice, we measured cytokine and chemokine production from primary PMs and AECs isolated from WT and IRAK-M−/− mice ex vivo. As shown in Fig. 6, influenza infection induced the secretion of TNF-α from PMs and the production of TNF-α was significantly greater in IRAK-M–null PMs incubated with virus than in WT PMs. Compared with incubation with influenza, LPS stimulation induced substantially more TNF-α production by PMs. On the other hand, influenza induced considerable quantities of MIP-2 and KC from both PMs (Fig. 6A) and AECs (Fig. 6B), and cells isolated from IRAK-M−/− mice secreted significantly more chemokines than WT cells. Moreover, the secretion of KC and MIP-2 was higher in influenza virus-infected cells than that observed with LPS stimulation.
To better define the mechanisms of increased alveolar permeability in IRAK-M–mutant mice during influenza infection, we investigated apoptotic cell death in the lungs of IRAK-M−/− and WT mice after intranasal influenza challenge. As shown in Fig. 7A, TUNEL staining revealed early evidence of apoptosis in infected IRAK-M−/− mice by day 2, observed primarily in bronchial epithelial cells, infiltrating PMN, and to a lesser extent AECs (Fig. 7Aa, 7Ae). By day 6, fewer apoptotic cells were found in IRAK-M−/− mice, which might be due to nearly complete denuding of the respiratory epithelium. In WT mice, little apoptosis was observed at day 2 postinfection (Fig. 7Ab, 7Af). However, frequent apoptotic leukocytes and epithelial cells were observed in WT mice at Day 6 (Fig. 7Ad, 7Ah).
To confirm the induction of apoptosis pathways during influenza pneumonia, we assessed the expression of caspase-3 mRNA in whole-lung homogenates. Caspase-3 is an essential effector of apoptosis in influenza infection and the effects of this kinase are mediated through cytokines and the TLR/RIG-1/mitochondrial antiviral signaling protein pathway (26, 27). Caspase-3 mRNA expression was significantly upregulated in IRAK-M−/− mice at day 2 after influenza infection compared with WT mice, and returned to baseline levels by day 7 (Fig. 7B).
To determine whethe the enhanced expression of inflammatory cytokines and chemokines observed in IRAK-M−/− mice contributed to increased inflammation, lung injury, and mortality in influenza pneumonia, we targeted these cytokines and chemokines in neutralization experiments. In initial experiments, we treated WT and IRAK-M−/− mice with 1 mg polyclonal rabbit anti-murine TNF-α Ab i.p. 2 h before intranasal influenza administration, then 0.5 mg i.p. every 48 h twice. Treatment with anti–TNF-α Ab did not alter short- or long-term survival mortality in influenza-infected IRAK-M−/− mice and, in fact, accelerated mortality in infected WT mice (data not shown). To define the contribution of ELR+ CXC chemokines, WT and IRAK-M−/− mice were treated with a polyclonal goat anti-murine neutralizing Ab against CXCR2, the sole receptor for ELR+ CXC chemokines in mice, which has been shown to block CXC chemokine-mediated effects in vitro and in vivo (20–22, 28, 29). As shown in Fig. 8A, treatment with anti-CXCR2 Ab did not alter survival in WT mice infected with influenza. In contrast, anti-CXCR2 Ab administration in IRAK-M−/− mice significantly improved short- and long-term survival, approximating survival observed in control IgG-treated WT mice (p < 0.02, Fig. 8A). Furthermore, administration of anti-CXCR2 Ab resulted in a trend toward reduced alveolar permeability in IRAK-M−/− mice at 6 d, reflected by decreased BALF albumin concentrations relative to control IgG-treated IRAK-M−/− mice (Fig. 8B). Improved survival in influenza-infected IRAK-M−/− mice after neutralization of CXCR2 was associated with a striking reduction in BALF neutrophils compared with the control IgG-treated IRAK-M−/− mice (Fig. 8C). There was no difference in lung viral titers between anti-CXCR2 Ab and control IgG treatment at day 2 in IRAK-M−/− mice (Fig. 8D, left panel). However, by day 6, IRAK-M−/− mice administered anti-CXCR2 Ab had an ~10-fold reduction in viral PFU in their lungs as compared with that observed in IRAK-M−/− mice treated with control IgG (Fig. 8D, right panel).
Influenza A virus is a negative-stranded RNA virus that infects epithelial cells of the upper respiratory tract and bronchi. Infection usually is limited to the trachea and bronchi but may extend to bronchioles and alveoli, resulting in interstitial pneumonia. AECs represent a primary target of infection in influenza pneumonia and if the host innate response is not adequately controlled, influenza-induced pneumonitis can progress to acute lung injury and acute respiratory distress syndrome (6, 30). Influenza-induced stimulation of innate host responses has been suggested as a cause of severe influenza-related pneumonia (9, 10, 31, 32).
Our findings have identified IRAK-M as an essential negative regulator of host-derived inflammatory responses in murine influenza pneumonia. The expression of IRAK-M is induced during influenza infection coincident with the time of peak inflammation. Moreover, we observed a much more vigorous accumulation of inflammatory cells and expression of inflammatory cytokines/chemokines in IRAK-M−/− mice during viral infection of the respiratory tract. Our results are consistent with the finding of enhanced compartmentalized inflammatory responses in IRAK-M−/− mice during bacterial infection (18). The current study differs from previous reports, however, because microbial clearance was facilitated by the generation of a robust innate response during acute Gram-negative bacterial infection, whereas exuberant inflammation during influenza pneumonia is clearly detrimental to host outcome. Specifically, we observed more severe lung injury and substantially reduced survival in IRAK-M−/− mice during influenza infection.
Neutrophils are the dominant leukocyte population recruited to the lung early in influenza infection (33–37), and this process is markedly enhanced in IRAK-M−/− mice. Despite more vigorous recruitment of inflammatory leukocytes, we found no difference in early viral clearance in IRAK-M−/− mice compared with their WT counterparts at day 2. By 6 d, however, IRAK-M−/− mice showed a 10-fold increase in blood viral PFU and a 100-fold increase in viral burden in the lungs. Mechanisms accounting for impaired local clearance of virus in IRAK-M−/− mice are uncertain, but may be attributable to PMN-mediated damage to lung cells and structures, including the respiratory epithelium. Importantly, we found significantly higher levels of the neutrophil-derived products, MPO and NE, in infected IRAK-M-deficient mice. These enzymes are released during neutrophil activation and have been associated with deleterious tissue damage. Notably, NE complexes with α-antitrypsin have been found in the BALF from patients with acute lung injury and animal models of severe influenza pneumonia (11, 38). Alternatively, immunosuppression may occur as a result of uptake of apoptotic neutrophils or other cells. Breakdown of the alveolar capillary membrane (as indicated by increased lung permeability) is a rather plausible explanation for enhanced dissemination of virus to the bloodstream. The finding of reduced lung injury and improved survival in anti-CXCR2–treated IRAK-M−/− mice is consistent with the notion that heightened chemokine-dependent neutrophilic inflammation is deleterious during influenza virus infection. It has previously been reported that CXCR2 is required for neutrophil recruitment to the lungs, but is not essential for viral clearance in the case of influenza virus infection (35). Furthermore, anti-CXCR2 Ab-treated IRAK-M−/− mice had reduced neutrophil influx and prolonged survival during influenza infection. Neutrophils have been shown to promote epithelial damage and detachment induced by viral infection and contribute to the pathophysiology of viral disease (39, 40). Interestingly, anti-CXCR2 Ab administration reduced lung viral burden in IRAK-M−/− mice, which might also contribute to improved survival in these animals. Zhao and colleagues have recently reported that H5N1 influenza virus can replicate in neutrophils, suggesting that these cells might serve as a reservoir for virus (41). Although reduced PMN influx after anti-CXCR2 treatment could theoretically account for decreased viral load, we have not evaluated this in our model and further investigations are needed.
We found earlier and more impressive apoptosis of lung cells, including airway and AECs and infiltrating PMN, in IRAK-M−/− mice during the evolution of infection, as compared with WT mice. These results were consistent with the time course of caspase-3 induction in whole lungs. The molecules driving early apoptotic responses in the lungs of IRAK-M−/− mice have not been delineated. A likely candidate is TNF-α, which is a strong promoter of the apoptotic pathways, and its expression is enhanced in influenza-infected IRAK-M−/− mice (26, 27). Direct viral infection of AECs can lead to host-immune cell-mediated induction of apoptosis in epithelial cells and resultant lung damage. Alternative candidates include Fas-Fas ligand interactions and TRAIL, which are major mediators of AEC apoptosis in acute lung injury and severe influenza pneumonia (42, 43). We did not find increased levels of Fas ligand in the BALF of influenza-infected IRAK-M−/− mice (data not shown). It is tempting to speculate that lung structural cell apoptosis may contribute to the enhanced lung albumin leakage observed in mutant mice.
In addition to enhanced PMN accumulation in IRAK-M−/− mice, we also found increased early and sustained influx of both CD4+ and CD8+ T cells, maximal at day 6 after influenza challenge. In particular, cytotoxic T cells can damage influenza-infected lung epithelial tissue in a manner that is partially dependent on TRAIL (44). Possible candidates that facilitate the recruitment/activation of CD4+ and/or CD8+ cells and whose mRNA expression was enhanced in influenza-infected IRAK-M−/− mice include MIG/CXCL9 and MIP-3α/CCL20 (data not shown). We also found enhanced IFN-α1 and -β1 mRNA expressions in influenza-infected IRAK-M−/− mice at day 2. These cytokines are important in antiviral immunity, but they have also been reported to promote deleterious immune responses mediated through the RIG-1/TLR3 pathway (23–25). Conversely, we found no differences in the mRNA expression of the T cell chemo-attractants MIP-1α/CCL3 or IP-10/CXCL10. Importantly, the enhanced accumulation of CD8+ T cells in IRAK-M−/− mice did not contribute meaningfully to the increased mortality observed in these mice, because Ab-mediated depletion of CD8+ did not alter mortality in IRAK-M−/− mice compared with that of mutant mice receiving the control Ab (data not shown).
TLRs play a key role in the innate immune recognition of multiple viral pathogens, including influenza. Among the TLR family, the MyD88-dependent toll receptors TLR7 and TLR8 are closely related endosomal TLRs that recognize single-stranded RNA, and they are thought to be the most important pathogen recognition receptors in influenza virus infection (30, 37, 45–47). In addition, TLR3, which requires the TIR domain-containing adaptor-inducing IFNβ, but not MyD88, has also been reported to participate in immunity against influenza virus (23). IRAK-M has been shown to negatively regulate MyD88-dependent, but not TIR domain-containing adaptor-inducing IFNβ-dependent TLR signaling, suggesting that TLR7 and TLR8, but not TLR3, assume dominant roles in the generation of innate response to influenza virus in the lungs. IRAK-M can also regulate MyD88-dependent TLR signaling activated by host-derived danger signals (alarmins or danger-associated molecular patterns ) and signaling through the IL-1 receptor type I.
IRAK-M was initially characterized and believed to be exclusively expressed by monocyte/macrophage populations. Since the initial description of the molecule, IRAK-M has been shown to be produced in other cells, including epithelial cells. For example, IRAK-M has been shown to be expressed by biliary epithelial cells and mediates LPS tolerance in these cells (48). Lung structural cells are major sources of chemokines during acute Gram-negative bacterial pneumonia (49). Primary human differentiated type II AECs secrete both neutrophil-active chemokines (IL-8 and MIP-2) and monocyte-active chemokines (MCP-1) in response to influenza infection ex vivo (50). Importantly, these studies indicate that IRAK-M regulates influenza-induced cytokine and chemokine expression in both myeloid cells (PMs) and structural cells (AECs). Moreover, IRAK-M also regulates cytokine production in macrophages and AECs stimulated with LPS. We found that AECs were a primary source of MCP-1 in response to influenza virus ex vivo and enhanced expression of MCP-1 was observed in IRAK-M–null AECs (data not shown). In addition, we found that primary AECs responded to LPS differently from influenza infection. Specifically, KC and MIP-2 were secreted more robustly from both PMs and AECs after incubation with influenza compared with LPS stimulation. However, TNF-α secretion from PMs was more dominant after LPS stimulation. These stimulus-specific differences may account for the disparity in the inflammatory response and microbial clearance observed in influenza infection as compared with lung bacterial infection. Given the myeloid and nonmyeloid cellular sources of IRAK-M, it is difficult to determine the cell populations regulated in an IRAK-M–specific manner during influenza infection in vivo. A bone marrow chimera model will be required to confirm the relative contribution of myeloid and nonmyeloid cells.
In conclusion, our findings indicate that IRAK-M serves as a vital regulator of innate responses during murine influenza pneumonia, allowing the host to fine tune the innate response to ensure adequate viral clearance mechanisms while limiting collateral damage to organ systems required for host survival. Furthermore, this protein may play a major role in modulating pathogen-induced inflammation at mucosal surfaces. Additional investigation is needed to define the context in which IRAK-M can be exploited therapeutically to improve outcomes in patients with serious infections of the respiratory tract.
This work was supported by Japanese Society for the Promotion of Science Grant-in-Aid for Scientific Research 21790774 (to M.S.) and National Institutes of Health/National Heart, Lung, and Blood Institute Grants HL57243 and P50 HL074024 (to T.J.S.).
The authors have no financial conflicts of interest.