IκB Kinase Is a Critical Regulator of Chemokine Expression and Lung Inflammation in Respiratory Syncytial Virus Infection

Abstract

Respiratory syncytial virus (RSV) is the major etiologic agent of severe epidemic lower respiratory tract infections in infancy. Airway mucosal inflammation plays a critical role in the pathogenesis of RSV disease in both natural and experimental infections. RSV is among the most potent biological stimuli that induce the expression of inflammatory genes, including those encoding chemokines, but the mechanism(s) that controls virus-mediated airway inflammation in vivo has not been fully elucidated. Herein we show that the inoculation of BALB/c mice with RSV results in rapid activation of the multisubunit IκB kinase (IKK) in lung tissue. IKK transduces upstream activating signals into the rate-limiting phosphorylation (and proteolytic degradation) of IκBα, the inhibitory subunit that under normal conditions binds to the nuclear factor (NF)-κB complex and keeps it in an inactive cytoplasmic form. Mice treated intranasally with interleukin-10 or with a specific cell-permeable peptide that blocks the association of the catalytic subunit IKKβ with the regulatory protein NEMO showed a striking reduction of lung NF-κB DNA binding activity, chemokine gene expression, and airway inflammation in response to RSV infection. These findings suggest that IKKβ may be a potential target for the treatment of acute or chronic inflammatory diseases of the lung.

Respiratory syncytial virus (RSV), a single-stranded negative-sense RNA virus of the Paramyxoviridae family, is well recognized as the major cause of serious lower respiratory disease in infancy and early childhood as well as in the elderly. In bronchiolitis, the most severe clinical manifestation of RSV infection, an intense peribronchial infiltration of mononuclear cells (lymphocytes and monocytes) occurs concomitantly with considerable edema, sloughing of the respiratory epithelium, and plugging of the small bronchioles with fibrin and mucus (1, 8).

Chemokines, a superfamily of small, structurally related molecules, induce the migration and activation of leukocytes and have emerged as central regulatory molecules in inflammatory, immune, and infectious processes of the lung (29). Indirect evidence suggests that these inflammatory molecules may play a critical role in the pathogenesis of RSV disease in infants. We recently reported that in the BALB/c mouse model, which shows close similarity to the pathogenesis of RSV-induced lower airway disease in humans, intranasal (i.n.) infection with RSV results in the rapid inducible expression of lung chemokines belonging to the CXC, CC, and C families. The levels of chemokines were dependent on the dose of RSV inoculum and paralleled the intensity of lung cellular inflammation. Furthermore, genetically altered mice with a selective deletion of the chemokine MIP-1α gene (−/−) had a significant reduction in lung inflammation following RSV infection compared to control (+/+) littermates (14).

Studies in vitro and in vivo have also demonstrated that RSV is among the most potent biological stimuli that are able to induce chemokine and cytokine production, a process that is largely controlled by virus-mediated activation of the transcription factor NF-κB (4, 5, 10). NF-κB constitutes a family of inducible transcription factors that include the potent RelA (p65) transactivator, RelB, c-Rel, and the NF-κB1 (p50) and NF-κB2 (p52) subunits. The inducible NF-κB subunits interact with cytoplasmic inhibitors, collectively known as IκBs (IκBα, IκBβ, IκBɛ, p105, and p100), through motifs contained within a conserved NH₂-terminal Rel homology domain. Extracellular stimuli initiate a signaling cascade that leads to the rapid phosphorylation of IκBα by two IκB kinases, IKKα (IKK-1) and IKKβ (IKK-2), at specific N-terminal serine residues. This event is coupled to the rapid ubiquitination and proteolytic degradation of phosphorylated IκBα through the 26S proteasome, thereby exposing the nuclear localization signal and freeing NF-κB to translocate to the nucleus, where it binds and activates target genes (20).

The pleiotropic cytokine interleukin-10 (IL-10) has been implicated as an important regulator of the functions of lymphoid and myeloid cells (26). In addition to its effects on T helper 1 (Th1) and Th2 cytokines, IL-10 also acts to inhibit proinflammatory cytokines, growth factors, and chemokine production by mononuclear phagocytes, neutrophils, and eosinophils (26). In vitro studies indicate that IL-10 inhibits proinflammatory cytokine production by specifically suppressing NF-κB activation. However, the in vivo mechanism(s) and signaling pathways responsible for NF-κB inhibition by IL-10, especially during viral infection, are largely unknown. For this study, by using a BALB/c mouse model of experimental RSV infection, we provide novel evidence that i.n. inoculation of IL-10 or a specific peptide that blocks IKKβ catalytic activity inhibits RSV-induced IκBα phosphorylation and degradation and NF-κB nuclear translocation. In addition, we show that lung chemokine expression and histopathology are significantly reduced in RSV-infected mice treated with IL-10 or the IKKβ inhibitory peptide compared to untreated controls. We suggest that mucosally delivered pharmacologic interventions targeting IKKβ-mediated NF-κB activation may have a significant impact on the development of acute inflammatory responses that are associated with RSV-induced lung pathology.

MATERIALS AND METHODS

RSV preparation.

Human RSV (A2) was grown in HEp-2 cells (American Type Culture Collection, Rockville, Md.). RSV was purified by polyethylene glycol precipitation, followed by centrifugation in 35 to 65% discontinuous sucrose gradients as described elsewhere (28). The virus was divided into aliquots, quick-frozen, and stored at −70°C until use. Virus titers were determined by a methylcellulose plaque assay (21). As a control, noninfected HEp-2 cells were accordingly prepared and diluted before application. RSV was UV inactivated as described previously (28) and was analyzed for infectious virus. No contaminating endotoxin or cytokines, including IL-1, IL-6, IL-8, tumor necrosis factor alpha, granulocyte-macrophage colony-stimulating factor, and interferons, were found in the sucrose-purified viral preparations (18).

Culturing of epithelial cells and infection.

A549 cells (American Type Culture Collection) were grown as a monolayer in minimal essential medium (MEM) supplemented with 10% fetal bovine serum, 100 U of penicillin per ml, and 0.1 mg of streptomycin per ml (all from Gibco BRL, Grand Island, N.Y.). The cells were maintained at 37°C in 5% CO₂. A549 cells grown at 90% confluence were infected with sucrose-gradient-purified RSV at a multiplicity of infection of 1 in MEM supplemented with 2% fetal bovine serum for 1 h at 37°C in 5% CO₂. Thereafter, additional medium was added and the infection was continued for five more hours (for a total of 6 h of infection) (28).

Mouse and infection protocol.

Female 3- to 4-week-old BALB/c mice were purchased from Harlan (Houston, Tex.) and were housed in pathogen-free conditions in the animal research facility of the University Texas Medical Branch (UTMB), Galveston, Tex., in accordance with the National Institutes of Health and UTMB institutional guidelines for animal care. Under light anesthesia, mice were inoculated i.n. with 50 μl of sucrose-gradient-purified RSV diluted in phosphate-buffered saline (PBS) (final administered dose, 10⁷ PFU). The 50-μl volume was selected to allow infection of the mice with a high titer of purified RSV and distribution of the inoculum mainly in the lung tissue (12). Control mice were inoculated with 50 μl of either the supernatant obtained from uninfected HEp-2 cells that were processed in the same way as the purified RSV pools (also adjusted for total protein content) or PBS. As we previously reported, the use of these two control preparations gave similar results and therefore subsequent studies were performed with PBS-inoculated control animals (referred to hereafter as sham infection) (13, 14). In some experiments, 10 μg of either murine or human recombinant IL-10 (kindly provided by S. Narula, Schering-Plough) or control vehicle (PBS) was inoculated together with RSV in a final volume of 50 μl. This dose was chosen based on published information on the effect of IL-10 in a mouse model of lung injury (22) and on preliminary experiments conducted in our laboratory. In separate experiments, BALB/c mice were infected with RSV and concomitantly instilled i.n. with NBD peptide (20 μg/20 μl) or a mutated control peptide. On day 5, 8 h before mice were sacrificed, a second dose of peptide (20 μg/20 μl) or vehicle was instilled i.n. The optimal dose and schedule of administration were chosen based on published data (24) and preliminary experiments conducted with our mouse model. At the indicated time points after infection, mice were anesthetized with an intraperitoneal injection of ketamine and xylazine before the thoracic cavity was opened. The mice were exsanguinated via heart puncture, and the trachea was opened by incision of the cricothyroid membrane. For collection of bronchoalveolar lavage (BAL), the lungs were flushed twice with ice-cold sterile PBS (0.8 ml). Lungs were then removed for RSV titration, RNA isolation, and protein extraction. For virus titration, the lungs were homogenized in Dulbecco's modified Eagle medium supplemented with 2% fetal calf serum in a 10% ratio (wt/vol). Homogenized samples were centrifuged at 2,000 × g for 10 min. Serial dilutions of the supernatants were tested in a methylcellulose plaque assay (14). Northern blot analysis of RSV N mRNA in lung tissue was performed exactly as described previously (39).

Extraction of lung nuclear proteins and EMSA.

Lung tissue was quick-frozen in liquid nitrogen immediately after removal from the thoracic cavity. Nuclear and cytoplasmic proteins were isolated from the lung tissue exactly as described previously (13) and were stored at −80°C until use. For electrophoretic mobility shift assays (EMSAs), nuclear proteins were normalized by a protein assay (Protein Reagent; Bio-Rad, Hercules, Calif.) and used to bind to duplex oligonucleotides corresponding to the RANTES NF-κB binding site (wild type) or mutated (Mut) under previously described conditions (10, 13). Briefly, DNA binding reactions contained 10 μg of nuclear proteins, 6% glycerol, 5 mM HEPES-KOH, 100 mM NaCl, 0.4 mM MgCl₂, 0.05 mM EDTA, 1 μg of poly(dA-dT), and 40,000 cpm of ³²P-labeled double-stranded oligonucleotides in a total volume of 40 μl. The nuclear proteins were incubated with the probe for 20 min at room temperature before they were fractionated in a 6% nondenaturing polyacrylamide gel in TBE buffer (22 mM Tris-HCl, 22 mM boric acid, 0.25 mM EDTA, pH 8.0). Gels were dried and exposed for autoradiography with Kodak XAR films at −70°C.

Western immunoblots.

Western immunoblotting for phosphorylated IκBα (IκBα-P) or total IκBα was performed with whole-cell extracts (WCE). Lung tissue was homogenized in lysis buffer (50 mM HEPES, 150 mM NaCl, 1 mM EDTA, 1.5 mM MgCl₂, 10% glycerol, 1% Triton X-100, 1 mM Na₃VO₄, 10 mM NaF, 10 mM Na₄P₂O₇, 1 μg of leupeptin/ml, 1 μg of aprotinin/ml, 400 μM phenylmethylsulfonyl fluoride). After incubation on ice for 20 min, cell debris was removed by centrifugation at 4°C (16,000 × g, 15 min). The protein concentration was determined as described above, and samples were frozen in aliquots at −70°C. Briefly, 50-μg protein samples were separated by sodium dodecyl sulfate (SDS)-4 to 20% polyacrylamide gel electrophoresis (PAGE) (Bio-Rad) and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore, Bedford, Mass.) by electroblotting in 1× 3-[cyclohexylamino]-1-propanesulfonic acid buffer and 10% (vol/vol) methanol. Membranes were then washed, blocked, and probed overnight at 4°C with a 1:100 dilution of rabbit anti-IκBα-P (Calbiochem, La Jolla, Calif.) or 200 μg of anti-IκBα antibodies (Santa Cruz Biotech)/ml, followed by secondary detection with horseradish peroxidase-conjugated anti-rabbit immunoglobulin G by enhanced chemiluminescence (Amersham Corp., Arlington Heights, Ill.).

IKK activity in lung tissue.

For immunoprecipitation of the IKK signalosome, lung WCE (500 μg) were gently rotated for 3 h at 4°C in the presence of 1 μg of anti-IKKβ antibody (Santa Cruz Biotechnology). After the addition of 100 μl of Sepharose-protein A beads (Sigma), the mixture was further incubated at 4°C for 3 h with gentle rotation. The immunoprecipitate was washed three times with RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.25% deoxycholate, 1 mM Na₃VO₄, 1 mM NaF, 1 μg of leupeptin/ml, 1 μg of aprotinin/ml, 1 mM PMSF), followed by three washes with 5× assay buffer (100 mM MOPS [morpholinepropanesulfonic acid], 125 mM β-glycerophosphate [pH 7.2], 5 mM Na₃VO₄, 5 mM EGTA, 5 mM dithiothreitol, 1 μg of aprotinin/ml, 100 μM ATP, 75 mM MgCl₂). The kinase reaction was performed in 40 μl of kinase buffer in the presence of 10 μCi of [γ-³²P]ATP (Du Pont NEN Research Products, Boston, Mass.) and 0.6 μg of the fusion protein glutathione S-transferase (GST)-IκBα(1-54) or a mutated form of IκBα in which the serine phosphorylation sites were mutated to alanine at positions 32 and 36 (S32A and S36A). The reaction was incubated at 30°C for 30 min and was terminated by the addition of 20 μl of 5× SDS-PAGE sample buffer. For visualization of phosphorylated IκBα, the reaction was subjected to SDS-10% PAGE, transferred to a PVDF membrane, and exposed to XAR film (Eastman Kodak). Western blotting for IKKβ was performed as a loading control by using anti-IKKβ antibody (a gift of Frank Mercurio, Signal Pharmaceuticals).

Synthesis of the NEMO binding domain fusion peptide.

The NEMO binding domain (NBD) fusion peptide is an 11-amino-acid peptide spanning the region from T735 to E745 of IKKβ fused with a 17-amino-acid sequence of the Antennapedia homeodomain that has been previously described as a specific inhibitor of IKKβ catalytic activity (24). Based on the described sequence, the NBD peptide and a mutated control peptide were chemically synthesized in the Protein Chemistry Core at UTMB by 9-fluorenylmethoxy carbonyl chemistry and were purified by high-performance liquid chromatography with a C₁₈ reverse-phase column to >95% purity. The peptide sequence was confirmed by matrix-assisted laser desorption ionization-time of flight mass spectrometry. Before use, the peptides were resuspended in 1 mM HCl to a stock solution of 10 mg/ml. Final concentrations were prepared in Dulbecco's PBS, pH 7.2. The NBD and mutated peptides were injected i.n. into mice (20 μg/20 μl) as described below.

Chemokine mRNA expression by RNase protection assay.

After excision, lungs were quick-frozen in liquid nitrogen and stored at −80°C until total RNA was isolated by the thiocyanate-phenol chloroform method. Chemokine mRNA expression was determined by a multiprobe RNase protection assay (RPA) with a RiboQuant kit (Pharmingen, San Diego, Calif.), as previously described (14). The probe was labeled with [α-³²P]UTP (Du Pont NEN Research Products) by use of T7 polymerase. After overnight hybridization with 3 μg of total RNA and RNA digestion, the samples were treated with a proteinase K-SDS mixture, extracted with phenol-chloroform, and precipitated in the presence of ammonium acetate. The samples were finally loaded into a QuickPoint sequence gel (Novex, San Diego, Calif.), exposed to XAR film (Eastman Kodak), and developed at −70°C. The identity of each protected fragment was established by analyzing its migration distance against a standard curve (migration distance versus the log nucleotide length for each undigested probe). The quantity of each mRNA species in the original RNA sample was then determined based on the signal intensity (measured in an AlphaImager 2200 optical densitometer [Alpha Innotech Corp., San Leandro, Calif.]) given by the appropriately sized, protected probe fragment bands. Sample loading was normalized with the housekeeping gene L32, included in each template set.

Determination of chemokine proteins in BAL samples.

BAL samples were collected as described above. Cells were removed by centrifugation at 400 × g for 15 min at 4°C, and supernatants were tested by use of a commercial enzyme-linked immunosorbent assay for RANTES, MIP-1α (R&D Systems, Minneapolis, Minn.), and MCP-1 (Biosource International, Camarillo, Calif.). The sensitivity of the assay for RANTES, MIP-1α, and MCP-1 was 2 pg/ml, 1.5 pg/ml, and 9 pg/ml, respectively.

Pulmonary histopathology.

Lungs were perfused and fixed in 10% buffered formalin and embedded in paraffin. Multiple 4-μm-thick lung sagittal medial sections were stained with hematoxylin and eosin (H&E). Slides were analyzed and scored for cellular inflammation by light microscopy by two independent investigators, as previously described (14, 36). Briefly, inflammatory infiltrates were scored by enumerating the layers of inflammatory cells surrounding the vessels and bronchioles throughout the whole lung section. The presence of zero to three layers of inflammatory cells was considered normal. A moderate to abundant amount of infiltrate (more than three layers of inflammatory cells surrounding 50% or more of the circumference of the vessel or bronchioles) was considered abnormal (14, 36). The number of abnormal perivascular and peribronchial spaces divided by the total number of perivascular and peribronchial spaces was the percentage reported as the pathology score. A total of ∼15 perivascular and peribronchial spaces were counted for each slide.

Statistical analysis.

Statistical analysis was performed with the Sigma Stat 3.0 software package (Jandel Corp., San Rafael, Calif.). The data were analyzed by the Wilcoxon test, and for inflammatory scores, separate sample t tests were performed for each independent experiment.

RESULTS

Time course of NF-κB binding activity in mouse lungs.

A time course of NF-κB activation in the lung was established by i.n. infection of BALB/c mice with RSV or a sham inoculum for 1.5, 3, 6, 12, and 24 h. The activation of NF-κB was determined by EMSA of nuclear extracts from whole lung tissue. Although there were slight differences for each experiment, NF-κB binding activity was first detectable between 0.5 and 1.5 h and was strongly present 24 h after RSV inoculation. Moreover, activation of lung NF-κB was still detectable 5 days after infection, as shown below (see Fig. 8). In contrast, for sham-inoculated animals, a weak complex first appeared between 1.5 h and 3 h and was not further inducible over the experimental period (Fig. 1). We have previously reported this weak binding activity of NF-κB at the earliest time points in mice treated with a sham inoculum (13).

FIG. 8.

NBD peptide blocks NF-κB activation and reduces lung chemokine expression and inflammation in RSV-infected mice. BALB/c mice were infected with RSV and concomitantly injected i.n. with NBD peptide (20 μg/20 μl) or a mutated control peptide. On day 5, 8 h before mice were sacrificed, a second dose of peptide or vehicle was inoculated i.n. (A) Nuclear proteins were extracted from the lungs and used for EMSAs as described for Fig. 1. Results for one sample of sham infection (lane 1) and two separate samples of RSV plus mutated NBD (mNBD; lanes 2 and 3) or RSV plus NBD (NBD; lanes 4 and 5) are shown. (B) Densitometry of lung chemokine mRNA expression by RPA (means ± SEM of three mice per treatment group). (C) Multiple H&E-stained lung sections were scored for cellular inflammation as described in Materials and Methods (means ± SEM of three mice per treatment group). *, P < 0.05 for RSV-infected mice treated with NBD compared to RSV-infected mice.

FIG. 1.

Time course of lung NF-κB activation in RSV-infected mice. Nuclear proteins were extracted from the lungs of sham- or RSV-infected mice (at 1.5, 3, 6, 12, and 24 h postinoculation) and used for EMSA (with the RANTES NF-κB-1 oligonucleotide probe) as described in Materials and Methods. Each data point is for one animal that is representative of three animals tested for each time point.

Inhibition of RSV-induced lung NF-κB binding activity by IL-10.

Previous studies, mostly in isolated cells, have established that the cytokine IL-10 inhibits NF-κB binding activity (32). However, to our knowledge, IL-10 has not yet been evaluated as a potential therapy to modulate NF-κB-mediated lung inflammation induced by respiratory viral pathogens. Experiments were therefore conducted to establish the effect of a single i.n. dose of murine recombinant IL-10 (rIL-10) (10 μg), given at the time of RSV inoculation, on lung NF-κB binding activity. An EMSA of lung nuclear proteins showed that treatment of BALB/c mice with IL-10 inhibited RSV-induced NF-κB binding activity in lung extracts obtained at 0.5, 1.5, 6, and 24 h postinfection (Fig. 2C, compare lane 2 with lane 3, lane 4 with lane 5, lane 6 with lane 7, and lane 8 with lane 9).

FIG. 2.

Exogenous IL-10 blocks RSV-induced NF-κB activation by partially preventing phosphorylation and degradation of IκBα. The effect of IL-10 on IκBα phosphorylation (A) and total IκBα expression (B) was assessed by Western blot analysis and NF-κB activation by EMSA (C). BALB/c mice were infected with RSV alone (lanes 2, 4, 6, and 8) or together with i.n. administration of rIL-10 (10 μg) (lanes 3, 5, 7, and 9) or were sham infected (lane 1). At 0.5, 1.5, 6, and 24 h postinfection, lungs were removed and nuclear proteins or WCE was isolated as described in Materials and Methods. For Western blots, 50 μg of WCE was fractionated by SDS-PAGE (4 to 20%), transferred onto a PVDF membrane, and probed with an antibody against phosphorylated IκBα (A) or total IκBα (B). For determination of NF-κB activation, EMSAs were performed in parallel, using nuclear proteins from mouse lungs as described for Fig. 1 (C). Each data point and/or treatment condition is from a single animal and is representative of data for three animals. Results are representative of three independent experiments.

The major regulatory pathway that controls NF-κB activation occurs at the level of its interaction with the inhibitory protein IκBα (2). Thus, to determine whether RSV-induced activation of NF-κB was associated with the degradation of IκB inhibitors and whether the inhibitory effect of IL-10 on NF-κB binding was due to a preservation of IκB, lung whole-cell protein extracts were tested in parallel for expression levels of IκBα by Western blotting. IκBα was chosen as a potential candidate since it was previously shown that it is by far the most abundant IκB subunit present in lung cells (17). As shown in Fig. 2B, RSV infection induced a rapid, partial degradation of IκBα in the lung (compare lane 1, sham-infected mouse, with lane 2, RSV-infected mouse). Furthermore, total IκBα protein levels in the lung were consistently higher in mice that had been treated with IL-10 at the time of RSV inoculation than in those that had been infected but not treated with IL-10 (Fig. 2B, compare lane 2 with lane 3, lane 4 with lane 5, and lane 6 with lane 7). In parallel experiments, levels of phosphorylated IκBα in lung extracts were determined. At 6 h postinfection, mice treated with IL-10 had consistently lower levels of phosphorylated IκBα (IκBα-p) than those that had been infected but not treated with IL-10 (Fig. 2A, compare lanes 6 and 7). However, despite a clear inhibitory effect of IL-10 on RSV-induced NF-κB binding activity and total IκBα degradation, differences in IκBα-P at the earlier time points were difficult to detect in Western blots.

IL-10 inhibits IKKβ activity in vivo.

The activation of NF-κB following in vitro stimulation of isolated cells with cytokines (25) or infection with RSV (Fig. 3A and our unpublished observations) is dependent upon IKKβ-mediated phosphorylation of IκBα. Since it is currently unknown if this event is also operational in vivo, we investigated the effect of RSV infection, with or without IL-10 treatment, on IKKβ activity in lung tissue. IKKβ was immunoprecipitated from whole lung extracts and tested for the ability to phosphorylate a GST-IκBα substrate. Western blotting for IKKβ was performed to control for sample loading in the kinase assay. The specificity of the IKKβ assay was initially determined in cultured RSV-infected epithelial cells by the inability of the immunoprecipitate to phosphorylate a GST-IκBα mutant in which the serine phosphorylation sites were mutated to alanine at positions 32 and 36 (S32A and S36A) (Fig. 3A, compare lanes 2 and 3). As shown in Fig. 3B, RSV strongly induced IKKβ activity in the lungs of mice that were infected for 1.5, 6, and 24 h (note that GST-IκBα appears as a double band when transferred to a PVDF membrane) (Fig. 3A and B). Western blotting for IKKβ was performed as a loading control and showed that similar amounts of IKKβ were immunoprecipitated from the lung extracts (Fig. 3B, lower panel). As was observed with the NF-κB EMSA (Fig. 1) (13), we noticed some increase in IKKβ activity in sham-inoculated mice, probably reflecting a stimulatory activity of the inoculum used as a control (supernatant of HEp-2 cells). However, the IKKβ activity in sham-inoculated mice appeared to peak at 6 h and at each time point tested was significantly lower than that induced by the virus (Fig. 3C). Furthermore, IL-10 inhibited RSV-induced IKKβ activity in the mouse lung to levels that were comparable to those detected in sham-inoculated mice (Fig. 3B and C).

FIG. 3.

RSV induces IKK activity: inhibitory effect of IL-10. WCE from A549 cells (250 μg) or mouse lung tissue (500 μg) were immunoprecipitated with an antibody against IKKβ, and the immunoprecipitates were subjected to a kinase assay using GST-IκBα wild type (WT) or a mutated form (Mut) as substrates. Samples were subjected to SDS-PAGE and transferred to PVDF membranes. Membranes were dried and exposed to an XAR film. (A) RSV infection (6 h) induces IKKβ activity in A549 cells, as shown by specific phosphorylation of wild-type GST-IκBα but not a mutated substrate. (B) Time course of IKKβ activity in lung tissues of sham-infected mice, RSV-infected mice, and RSV-infected mice treated i.n. with IL-10. RSV infection induces the activation of IKKβ activity (lanes 2, 5, and 8), which is decreased by IL-10 treatment (lanes 3, 6, and 9). Equal sample loading was determined by Western blotting for IKKβ. (C) Densitometry analysis of phosphorylated GST-IκBα complexes expressed as percentages of activation relative to those of RSV-infected mice.

IL-10 inhibits RSV-induced lung chemokine expression.

Chemokines are critically involved in RSV-mediated lung inflammation and pathology. We have recently shown that experimental RSV infection of mice results in a rapid, dose-dependent induction of several inflammatory chemokines in lung tissue (14). Since the majority of these chemokine genes are transcriptionally regulated by NF-κB, we reasoned that the observed inhibitory effect of IL-10 treatment on RSV-induced NF-κB activation in the lung would be reflected upon inducible expression of these inflammatory molecules. For this reason, we characterized the profile and abundance of lung chemokines in groups of BALB/c mice infected with RSV, with or without i.n. IL-10 treatment. RNA was isolated from mouse lung tissues 24 h after infection and were analyzed by RPA using a multiprobe containing DNA templates for nine murine chemokines (Fig. 4). For sham-infected mice, only the mRNA for RANTES was weakly visible. On the other hand, for RSV-infected mice, we consistently observed the appearance of mRNA-protected bands specific for RANTES, MIP-1β, MIP-1α, MIP-2, and MCP-1 (Fig. 4A). Densitometric analysis showed that the expression of mRNAs in the lung for RANTES (P < 0.05), MIP-1α (P = 0.02), MIP-2 (P = 0.001), and MCP-1 (P < 0.05) was significantly decreased in RSV-infected mice that were treated with IL-10 compared to infected, untreated animals (Fig. 4B). The inhibitory effect of IL-10 on RSV-induced lung chemokines was confirmed by the determination of RANTES, MIP-1α, and MCP-1 protein levels in BAL samples. As shown in Fig. 5, treatment with IL-10 at the time of infection significantly reduced the concentrations of all three chemokines in BAL compared to those in RSV-infected, untreated littermates.

FIG. 4.

Inhibitory effect of IL-10 on chemokine mRNA expression in lung tissue of RSV-infected mice. The expression of nine murine chemokine mRNAs was determined by RPA. (A) BALB/c mice were sham infected (lane 1), infected with RSV (lane 2), or infected with RSV together with i.n. administration of rIL-10 (lane 3). At 24 h postinfection, RNA was isolated from lung tissue and hybridized with a ³²P-labeled RiboQuant MultiProbe (Pharmingen) containing DNA templates for the mouse chemokines RANTES, eotaxin, MIP-1β, MIP-1α, MIP-2, MCP-1, and TCA-3 and the housekeeping genes L32 (rRNA) and GAPDH. After RNase treatment and purification, protected probes were run on a QuickPoint sequence gel, exposed to an XAR film, and developed. The identity of each protected fragment was established as described in Materials and Methods. (B) The quantity of each mRNA species in the original RNA sample was determined based on the signal intensity (by optical densitometry) given by the appropriately sized, protected probe fragment band. Sample loading was normalized with the housekeeping gene L32, which was included in each template set. The density of each chemokine mRNA is expressed relative to that of L32. The data shown are representative of four independent experiments. Data are expressed as means ± standard errors of the means (SEM) of four animals per group. *, P < 0.05, compared with RSV plus IL-10.

FIG. 5.

Inhibitory effect of IL-10 on RANTES, MCP-1, and MIP-1α protein production. The concentrations of RANTES, MCP-1, and MIP-1α were determined by enzyme-linked immunosorbent assay in BAL obtained from groups of mice that were sham infected, RSV infected, or RSV infected and treated with rIL-10 (at 24 h postinfection). Data are expressed as means ± SEM of five animals per group. #, P < 0.001 compared to sham-infected mice; *, P < 0.05 compared to mice infected with RSV plus IL-10.

IL-10 decreases lung inflammation and disease induced by RSV.

The mouse model shows close similarity to the pathogenesis of RSV-induced lower airway disease in humans. We have recently established that the experimental infection of BALB/c mice with highly purified preparations of RSV A, at a dose of 10⁷ PFU, induces a severe inflammatory response in lung tissue as early as 24 h after i.n. inoculation (14). Lung inflammation was characterized by an excess of monocytes/macrophages, lymphocytes, and to a lesser extent, neutrophils surrounding bronchioles and vessels, with evidence of the involvement of alveolar spaces. Thus, the following studies were designed to examine if i.n. IL-10 at a dose that effectively blocked RSV-induced NF-κB activation and reduced chemokine production is able to modify lung pathology. In four independent experiments, groups of BALB/c mice were infected i.n. with four different pools of RSV and concurrently inoculated with IL-10 or a control vehicle, or they were sham infected. Twenty-four hours after inoculation, mice were sacrificed, and multiple H&E-stained sagittal lung sections were analyzed and scored for inflammation by a pathologist who was blind to the treatment protocol, as previously described (14, 36). In sham-infected mice, no lung inflammation was detected. Mice infected with RSV had evident cellular infiltration in peribronchial and perivascular areas and intra-alveolar involvement, with slight differences in inflammatory scores depending on the pool of purified RSV used (Fig. 6). In all four experiments, treatment with IL-10 significantly reduced cellular perivascular and peribronchial infiltration and alveolar inflammation compared to RSV-infected, untreated animals. The inflammatory response in the lung mirrored the appearance of illness, since infected mice treated with IL-10 appeared to have less “ruffled hair” and were more actively moving than those that were untreated (not shown). Despite these differences in inflammation and illness, viral replication was comparable in IL-10-treated and untreated mice, as determined by a plaque assay (Fig. 7) or by Northern blot analysis of RSV nucleocapsid (N) transcripts (data not shown) (39).

FIG. 6.

Lung pathology scores in RSV-infected mice. Mice were sham infected, infected with RSV, or infected with RSV and treated with rIL-10. At 24 h postinfection, mice were sacrificed and lungs were removed, fixed in 10% buffered formalin, and embedded in paraffin. Multiple 4-μm-thick sections were stained with H&E and scored for cellular inflammation. The number of abnormal perivascular and peribronchial spaces divided by the total spaces counted is the percentage reported as the pathology score. Each data point represents the pathology score of an individual animal (group means are indicated by horizontal lines) in each independent experiment (n = 4 experiments). The pathology score for sham-infected mice is zero. P values were <0.01 for each experiment for RSV-infected mice compared to RSV-infected mice treated with IL-10.

FIG. 7.

Viral replication in lungs of BALB/c mice infected with pooled RSV and treated with IL-10. BALB/c mice were infected with RSV (final administered dose, 10⁷ PFU) together with 10 μg of IL-10 or control vehicle (PBS) in a final volume of 50 μl. One, 5 and 21 days after infection, the lung tissue was removed and homogenized and the concentration of infectious virus was determined by a plaque assay. The bar graph shows means ± SEM of three mice for each treatment group.

Inhibition of NF-κB activation by the NBD peptide reduces lung inflammation in RSV-infected mice.

In addition to IKKα and IKKβ, a third structural component of the IKK complex, known as IKKγ or NEMO, is required for proinflammatory activation of the complex (42). By extensive analysis of the interactive domains of NEMO with IKKβ, May et al. recently identified an amino-terminal α-helical region of NEMO that is associated with a carboxyl-terminal segment of IKKα and IKKβ and termed this the NEMO-binding domain (NBD) (24). An 11-amino-acid peptide spanning the region from T735 to E745 of IKKβ (NBD) fuses with a 17-amino-acid sequence of the Antennapedia homeodomain that mediates membrane translocation (24). The NBD was identified as a specific inhibitor of IKKβ catalytic activity and NF-κB activation (24). Therefore, we tested whether the NBD peptide could inhibit NF-κB activation in RSV-infected BALB/c mice. In addition, studies were designed to investigate whether blocking NF-κB activation could affect lung inflammation at the time of peak RSV replication and disease (day 5). In preliminary experiments, we established that the effect of the NBD peptide on RSV-induced NF-κB activation significantly decreased 8 h after injection (data not shown). Since frequent i.n. inoculation of the NBD in mice impractical, the peptide was given twice, once at the time of infection and once 8 h before the mice were sacrificed (day 5). RSV-infected mice treated with the NBD peptide had almost complete inhibition of viral-induced lung NF-κB nuclear translocation (by EMSA) (Fig. 8A) and a significant reduction in chemokine mRNA expression (by RPA) (Fig. 8B) compared to control animals injected with a mutated scrambled peptide. Peribronchial and perivascular inflammatory cell infiltrates were also unequivocally reduced in RSV-infected mice treated with NBD compared to controls (Fig. 8C). Overall, the results obtained with the NBD peptide are qualitatively comparable to those observed with exogenous rIL-10 and suggest that activation of IKKβ is a crucial step that regulates lung inflammation during RSV infection.

DISCUSSION

Lower respiratory tract infections caused by RSV are characterized by a profound cellular inflammation of the airway mucosa which contributes to disease manifestations, including airflow limitation, lung atelectasis or emphysema, and hypoxemia (31). The mouse model shows close similarity to the pathogenesis of RSV-induced lower airway disease in humans (11, 12). In BALB/c mice, RSV rapidly replicates in the lungs after i.n. inoculation and induces mononuclear cell infiltration around peribronchial and perivascular tissues (37) and objective plethysmographic signs of pulmonary dysfunction (30, 40). The molecular mechanisms that regulate airway inflammation in viral respiratory infections are not fully understood. We and others have previously established that in airway epithelial cell lines infected with RSV, a network of inflammatory cytokines and chemokines is inducibly regulated by the transcription factor NF-κB (4, 5, 10, 38). Herein and in other detailed studies, we showed that RSV induces a rapid and prolonged activation of NF-κB in lung tissue (Fig. 1 and 8), a process that is associated with nuclear translocation of the NF-κB subunits p50, c-Rel, and RelA (13). Moreover, we provided in vivo evidence that a viral respiratory infection activates NF-κB in lung tissue through a signaling pathway(s) that leads to the activation of IKKβ and the phosphorylation and degradation of IκBα. Functional blocking of IKKβ by IL-10 or the specific NBD peptide results in a marked attenuation of RSV-induced lung inflammation, due in part to the inhibition of inflammatory chemokine gene expression. These results indicate that IKKβ-mediated NF-κB activation is required for the development of airway inflammation in response to RSV infection. Interestingly, our data show that a reduction of lung inflammation does not affect RSV replication and clearance in the lung, as we previously showed in studies with MIP-1α-deficient mice (14). The mechanism of RSV lung clearance in the absence of inflammatory cells is not clear at the moment, but our data suggest that critical cellular components involved in antiviral immunity to RSV (such as CD8⁺ T lymphocytes) may not be affected by the overall reduction in cellular inflammation observed in our studies. Future investigations will focus on the identification of inflammatory cell types that are affected by blocking of the IKK-NF-κB pathway in the lung.

The major mechanism that regulates NF-κB activation involves phosphorylation followed by polyubiquitination and degradation of the inhibitory proteins called IκBs. In response to different stimuli, active NF-κB translocates to the nucleus as a result of proteolytic degradation of the IκB proteins. The large cytoplasmic multisubunit complex serine-protein kinase IKKβ has been identified as the major kinase responsible for the phosphorylation of IκBα (25). Previous work by our group has shown that proteolysis of IκB subunits, mainly IκBα, is temporally linked to RelA DNA-binding activation in RSV-infected lung epithelial cells (17). However, the upstream components of the NF-κB cascade activated by RSV infection, as well as its effect on airway inflammatory responses in vivo, had not been defined. To address these questions, we utilized an experimental model in which BALB/c mice were inoculated i.n. with RSV together with IL-10, a pleiotropic cytokine with profound anti-inflammatory and immunomodulatory activities. Indeed, IL-10 treatment has been shown to have beneficial effects in animal models of colitis (16), arthritis (41), and experimental endotoxemia (3). Interestingly, clinical trials in which IL-10 has been used for the treatment of inflammatory conditions have shown clinical improvement only in patients who had inhibition of NF-κB activation as a result of IL-10 therapy (33). Our study demonstrates that a single dose of IL-10 in RSV-infected mice has the following effects: (i) potently inhibits NF-κB DNA-binding activity in the lung, an effect which is sustained up to 24 h postinoculation; (ii) suppresses IKKβ activity and (partially) prevents IκBα phosphorylation and degradation; (iii) significantly reduces the concentrations of RSV-induced chemokines; and (iv) significantly reduces the pathological inflammatory response in the airways. All of these effects occurred in the absence of a direct antiviral effect, since viral transcription and the duration of viral shedding were unaffected by IL-10 treatment. These observations extend previous observations by Mulligan et al. on the protective role of IL-10 in a rat model of immune complex-induced lung injury (27). In subsequent studies by the same group, exogenous IL-10 was shown to inhibit NF-κB activation via a mechanism that was ascribed to sustained IκBα protein expression in the lung tissue of rats treated with the cytokine (22). Our present study provides experimental evidence that IKK is an upstream target of IL-10 inhibitory activity upon NF-κB activation and inflammation in virus-infected lungs. Consistent with previous results with highly controlled lung epithelial cell lines (17), as well as the results of Lentsch et al. with lung extracts (22), we observed an apparent temporal dissociation between the induction of NF-κB DNA-binding activity and IκBα phosphorylation in RSV-infected lung tissue extracts. In this regard, differences in the levels of IκBα-P expression between control and RSV-infected mice were difficult to detect at the 0.5- and 1.5-h time points (Fig. 2). We suspect that the sensitivity of the anti-IκBα-P antibody may have limited our ability to detect small changes in whole lung tissue extracts at the earlier time points. Alternatively, other IκB-independent mechanisms that are not yet fully understood may exist by which NF-κB nuclear translocation and DNA-binding activity are initially induced in response to a variety of different stimuli (19). Nonetheless, at the 6-h time point, we observed concordance between the EMSA for NF-κB and the Western blots for total IκBα or IκBα-P, as RSV-infected mice treated with IL-10 had a consistent reduction in lung NF-κB activation, lower levels of IκBα-P, and higher total IκBα levels than RSV-infected mice that were not treated with IL-10 (Fig. 2A and B). These studies, along with those in which the specific NBD peptide was used to probe for the requirement of IKKβ activity for NF-κB activation in lung tissue, strongly suggest that this signaling pathway is required for NF-κB nuclear translocation, chemokine expression, and the development of full airway inflammatory responses following a viral infection and perhaps other injurious stimuli.

The regulatory mechanisms of airway inflammation in RSV infection are not fully understood. However, recent studies with young infants suggested that inflammatory chemokines, including RANTES, eotaxin, and MIP-1α in particular, play a critical role in the pathogenesis of severe lower respiratory tract infections caused by RSV (9, 15, 35). We have previously described that the experimental infection of mice with RSV induces a dose-dependent expression of several inflammatory chemokines which parallels the intensity of lung cellular inflammation (14). As direct evidence for the role of chemokines in lung pathology, we have shown that genetically altered mice with a selective deletion of the MIP-1α gene have a striking reduction in lung inflammation following RSV infection compared to control littermates (14). The results of this study showing that the treatment of mice with IL-10 or the NBD peptide significantly reduced airway inflammation as well as the expression of several chemokine genes, including RANTES, MCP-1, and MIP-1α, further argue in favor of an essential role of these molecules in RSV-mediated lung pathology. Moreover, since NF-κB is required for RSV-induced transcription of chemokine genes in cultured cells (5, 10, 23, 38), we propose that the effect of IL-10 and the NBD peptide on RSV-induced lung chemokines is due to the inhibitory activity of IL-10 and the NBD peptide on NF-κB activation that we report in this study. While there was a similar inhibitory effect of IL-10 or the NBD peptide on RSV-induced RANTES, MIP-1α, MIP-1β, MIP-2, and MCP-1 mRNA expression, we observed an inhibitory effect of the NBD peptide but not IL-10 on TCA-3. The reason(s) for this discrepancy is unclear, and a better understanding of the (transcriptional) mechanisms that regulate the expression of this gene is necessary for formulating hypotheses. Interestingly, both subunits, p50 and c-Rel, that we demonstrate to be present in the RSV-inducible NF-κB heterodimer complex (13) have been shown to play a crucial role in the expression of lung chemokines. Mice deficient in the p50 subunit are incapable of mounting eosinophilic airway inflammation and are deficient in the production of eotaxin, MIP-1α, and MIP-1β (43). Mice deficient in the c-Rel subunit have decreased expression of MCP-1 in lung tissues following an allergen challenge compared to wild-type controls (7).

Both transcriptional and posttranscriptional mechanisms have been implicated in the inhibitory effect of IL-10 on cytokine and chemokine production (reviewed in reference 26). Most of these studies have been performed with isolated monocytes or macrophages. In addition, IL-10 has been recently shown to suppress IL-8 production in tumor necrosis factor-stimulated intestinal epithelial cells by a mechanism that involves the suppression of IKKβ activity and the inhibition of NF-κB binding (32). It has indeed been shown that intestinal epithelial cells express functional IL-10 receptors 1 and 2 (6), and we have preliminary data suggesting that these receptors are also expressed on certain lung epithelial cells (R. P. Garofalo, personal observation). Surprisingly, however, little is known about the regulation of chemokine expression by IL-10 in in vivo models of lung injury or inflammation. In the only study we are aware of, the role of endogenous IL-10 was investigated in a model of lipopolysaccharide (LPS)-induced lung inflammation. Mice lacking IL-10 (−/−) had increased levels of the chemokines MIP-1α and MIP-2 as well as neutrophilic lung inflammation in response to an intratracheal challenge with LPS compared to control IL-10^+/+ mice (34). However, we should consider the possibility that the effect of IL-10 and the NBD peptide on lung chemokine expression that we report in this study might represent a component of the more global down regulation of gene expression mediated by IL-10- and NBD-induced NF-κB inhibition. Therefore, although strongly suggested by other experimental models in animals or by observations of naturally acquired infections, the role of chemokines in mediating lung pathology should probably be seen in a larger context of proinflammatory gene products that RSV is able to induce through the activation of the “master switch,” NF-κB. For this reason, our data presented here highlight the importance of IKKβ activity as a novel potential target for anti-inflammatory and immunomodulatory therapies for virus-induced lung disease.