Granulocyte/Macrophage-Colony-stimulating Factor (GM-CSF ...

3 downloads 2 Views 213KB Size Report
Aug 2, 2002 - mice and placed in a 1.5-ml Eppendorf tube containing 250 μl of RPMI. For vehicle treatments, 0.1% Me2SO (final v/v) was added. Wortmannin.

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2002 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 277, No. 45, Issue of November 8, pp. 42808 –42814, 2002 Printed in U.S.A.

Granulocyte/Macrophage-Colony-stimulating Factor (GM-CSF) Regulates Lung Innate Immunity to Lipopolysaccharide through Akt/Erk Activation of NF␬B and AP-1 in Vivo* Received for publication, August 2, 2002, and in revised form, August 26, 2002 Published, JBC Papers in Press, September 2, 2002, DOI 10.1074/jbc.M207840200

Steven Bozinovski, Jessica E. Jones, Ross Vlahos‡, John A. Hamilton‡, and Gary P. Anderson§ From the ‡Arthritis and Inflammation Research Center, Department of Medicine, Cooperative Research Center (CRC) for Chronic Inflammatory Diseases, Royal Melbourne Hospital, The University of Melbourne, Parkville, VIC 3010, Australia and the Lung Disease Research Laboratories, CRC for Chronic Inflammatory Diseases, Department of Pharmacology, The University of Melbourne, Parkville, VIC 3010, Australia

The lung innate immune response to lipopolysaccharide (LPS) coordinates cellular inflammation, mediator, and protease release essential for host defense but deleterious in asthma, chronic obstructive pulmonary disease, and cystic fibrosis. In vitro, LPS signals to the transcription factors NF␬B via TLR4, MyD88, and IL-1Rassociated kinase (IRAK), to AP-1 by mitogen-activated protein (MAP) kinases, and via an alternate route in IRAK-deficient mice, but the in vivo lung signaling pathway(s) are not understood. We investigated the role of Akt and Erk1/2 as LPS intensely stimulates granulocyte/ macrophage-colony-stimulating factor (GM-CSF) release, and neutralizing GM-CSF profoundly suppressed LPS-induced inflammation, suppressed expression and activity of lung proteases, significantly reduced GMCSF and tumor necrosis factor ␣ (TNF␣) mRNA expression, and dampened nuclear localization of both NF␬B (p50/65) and AP-1. LPS markedly activated Akt and Erk1/2, but not p38, in a GM-CSF-dependent manner in direct temporal association with NF␬B and AP-1 activation. Pharmacological inhibition of Akt or Erk activation in LPS-treated tracheal explants ex vivo inhibited the release of GM-CSF. These data implicate GM-CSFdependent activation of Akt in the amplification of this response and demonstrate the role of Erks rather than p38 in lung LPS inflammatory responses. Inhibition of GM-CSF may be of therapeutic benefit in inflammatory diseases in which LPS contributes to lung damage.

The immediate recognition of bacteria and their products in the lung is mediated by an ancient immune response that utilizes conserved pattern recognition receptors to distinguish pathogen-associated molecular pattern signatures of microbes (1). LPS1 is a component of the Gram-negative bacterial cell * This work was supported by the National Health and Medical Research Council (NHMRC) of Australia and the Cooperative Research Center (CRC) for Chronic Inflammatory Diseases. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § To whom correspondence should be addressed. Tel.: 61-3-8-3448602; Fax: 61-3-8-344-0241; E-mail: [email protected] 1 The abbreviations used are: LPS, lipopolysaccharide; GM-CSF, granulocyte/macrophage-colony-stimulating factor; Erk, extracellular signal-regulated kinase; MAP, mitogen-activated protein; MAPK, MAP kinase; IRAK, IL-1R-associated kinase; TNF, tumor necrosis factor; TLR4, Toll-like receptor 4; mAb, monoclonal antibody; JNK, c-Jun NH2-terminal kinase; PBS, phosphate-buffered saline; BAL, Bronchoalveolar lavage; BALF, BAL fluid; ELISA, enzyme-linked immunosorbent assay; ANOVA, analysis of variance; MMP, matrix metalloprotease; MEK, MAP/ERK kinase; MEKK, MEK kinase; r, recombinant.

wall and a potent endotoxin capable of activating innate immunity (2). The subsequent inflammatory defense reaction coordinates neutrophil recruitment and macrophage activation through release of TNF␣, chemokines, oxygen radical products, and proteases. Although this LPS response is essential for host defense, LPS is also implicated in several human lung diseases, notably cystic fibrosis, chronic obstructive pulmonary disease, and asthma (3– 6). LPS binds to soluble LPS-binding protein present in the alveolar fluid and is then transferred to CD14 expressed on alveolar macrophages and bronchial epithelial cells (2) and, in conjunction with MD-2 signals via, at least, Toll-like receptor 4 (TLR4). TLR4 initiates an IL-1␤ receptor-like signaling (7) promoting NF␬B activation. Currently, this is thought to involve the myeloid differentiation protein (MyD88) and MyD88adapter-like (Mal) adapter molecules, recruitment and activation of IL-1R-associated kinase (IRAK), and complexing with tumor necrosis factor receptor-associated factor 6 (TRAF6), which activates NF␬B through activation of NF␬B-inducing kinase (NIK) as reviewed in Refs. 8 and 9. NF␬B subsequently promotes the release of inflammatory cytokines such as TNF␣, IL-1␤, neutrophil-recruiting chemokines, and matrix metalloproteases. However, mice lacking the IRAK gene remain partially responsive to IL-1 and LPS (10), suggesting that TLR4 signals via IRAK-dependent and -independent pathways to promote NF␬B-mediated cytokine expression. GM-CSF, which was originally purified and cloned from lung tissue of endotoxin-treated mice, is a major survival and proliferative factor for hematopoietic cells and primes mature macrophages, eosinophils, neutrophils, and respiratory epithelium for inflammatory effector functions (11). In contrast to almost all other tissues, low level basal GM-CSF is required for normal pulmonary physiology as mice deficient in GM-CSF develop fatal alveolar proteinosis that can be rescued by overexpression of the PU.1 transcription factor (12). Since endogenous GM-CSF regulates the intensity of inflammation and strongly activates Akt and Erk kinases, we reasoned that these GM-CSF-regulated kinases might contribute to the lung LPS response (13, 14). In particular, Akt promotes NF␬B activity by phosphorylating I␬B kinase-␣ (IKK-␣) in response to TNF␣ and platelet-derived growth factor (15, 16), which liberates the transcription factor from I␬B to initiate nuclear translocation and DNA binding. In addition, Akt induces trans-activation of the p65/RelA subunit of NF␬B in response to oncogenic Ras and IL-1␤ (17), a process required for full activation that is distinct from nuclear translocation and DNA binding. GM-CSF also promotes activation of Erk1/2 MAPK homolog, which belongs to a family of universal signal transduction molecules involved


This paper is available on line at

Akt/Erk Pathways Promote LPS Responsiveness in Mouse Lung in a wide variety of biological responses including regulation of the AP-1 transcription factor complex. The MAPK homologs (Erk1/2, p38, and JNK) promote AP-1 expression and activity by direct phosphorylation. Since MAPK homologs display overlapping substrate specificities that respond in a cell type- and stimuli-specific manner (18), we investigated the growth factorregulated Erk1/2 and stress-activated p38 homologs. In the present study, we performed detailed kinetic analyses of the LPS response and report the in vivo activation of Akt and Erk1/2 kinases but not p38 in LPS-inflamed lungs, which is GM-CSF-dependent. Furthermore, the activation status of Akt and Erk1/2 temporally correlated with NF␬B and AP-1 DNA binding activity, implicating the kinases as central signal transduction modules in LPS-induced lung inflammation. Our data indicate the molecular mechanisms by which GM-CSF augments innate immunity to Gram-negative pathogens but also suggest that anti-GM-CSF-mediated suppression of Akt and Erk may have clinical implications in inflammatory diseases characterized by exuberant cytokine and protease expression, such as chronic obstructive pulmonary disease, cystic fibrosis, and asthma, in which LPS has been implicated. MATERIALS AND METHODS

Animals—Specific pathogen-free male BALB/c mice aged 6 –7 weeks old and weighing ⬃20 g were obtained from the Animal Resource Center Pty. Ltd. (Perth, Australia), housed at 20 °C on a 12-h day/night cycle in sterile micro-isolators, and fed a standard sterile diet of Purina mouse chow with water allowed ad libitum. All animal handling and experimental procedures, which were performed aseptically, were approved by the Animal Experimental Ethics Committee of the University of Melbourne and conformed to International standards of animal welfare as specified in the National Health and Medical Research Committee (NHMRC) of Australia guidelines. GM-CSF-deficient mice (GMCSF⫺/⫺) and wild-type litter mate controls (GM-CSF⫹/⫹) were the generous gift of Professor Ashley Dunn, Ludwig Institute for Cancer Research, Melbourne, Australia (19). Transnasal Instillation of LPS and Anti-GM-CSF mAb—Inflammation was induced by transnasally instilling a maximally tolerated dose of LPS (10 ␮g of Escherichia Coli Serotype 026:B6, Sigma, in 35 ␮l of in PBS vehicle) into the lungs of groups of 8 –10 mice anesthetized with 2% enflurane (Abbott) in air, which uniformly distributes LPS throughout the lungs (20). Solutions administered to the mice, alone or in combinations at specified time points, were (a) PBS; (b) PBS containing isotype control (rat anti-mouse IgG2a mAb of irrelevant specificity, 100 ␮g/mouse); (c) PBS containing anti-GM-CSF mAb (22E9, rat antimouse IgG2a ultrapurified in-house, 100 ␮g/mouse); (d) PBS containing LPS. Anti-GM-CSF, PBS, or isotype control was administered 3 h prior to LPS challenge. All reagents were LPS-free and negative in the limulus lysate assay. For necroscopy, mice were anesthetized with ketamine/xylazine (15 mg/kg and 30 mg/kg, intraperitoneally, respectively). Bronchoalveolar lavage (BAL) (four pooled aliquots of 0.3 ml of PBS) recoveries (85 ⫾ 5%) did not differ significantly between groups. In the GM-CSF⫺/⫺ mice only, which suffer from alveolar proteinosis (19), BAL samples were centrifuged over a 15% BSA in PBS (with 0.5 M EDTA) density gradient. Total cell counts and viabilities were determined by ethidium bromide/acridine orange (Molecular Probes) fluorescent viability stains using Neubauer hemocytometer. Cytocentrifuge preparations (Shandon Cytospin 3) using 100 ␮l of BAL were differentiated according to standard morphological criteria counting at least 500 cells/(DiffQuik, Ziess, Axiolab, ⫻1000). ELISAs—Murine GM-CSF and TNF␣ levels were analyzed according to the manufacturer’s instructions (Pharmingen, limit of detection 10 ng/ml). Preparation of Whole Cell Protein and Nuclear Lung Extracts—Following lavage, the lungs were perfused via the right ventricle with 5 ml of PBS to remove intravascular leukocytes, removed, snap-frozen in liquid nitrogen, ground, and stored at ⫺80 °C. For preparation of whole cell extracts, 10 mg of ground lung tissue was resuspended in 500 ␮l of lysis buffer (50 mM Tris-HCl (pH 7.5), 120 mM NaCl, 1% (v/v) Nonident P-40, 1 mM EDTA, 50 mM NaF, 40 mM ␤-glycerophosphate, 1 mM benzamadine, and 0.5 mM phenylmethylsulfonyl fluoride). Following a 15-min incubation on ice, homogenates were cleared by centrifugation for 10 min (16000 ⫻ g at 4 °C). For nuclear extraction, 10 mg of lung tissue was resuspended in 500 ␮l of nuclear lysis buffer 1 (10 mM


HEPES (pH 7.6), 15 mM KCl, 2 mM MgCl2, 0.1 mM EDTA, 5 mM ␤-mercaptoethanol, 0.2% Nonident P-40, and 0.5 mM phenylmethylsulfonyl fluoride) for 10 min on ice. Nuclei were pelleted by centrifugation at 800 ⫻ g for 30 s and lysed in 500 ␮l of nuclear lysis buffer 2 (50 mM HEPES (pH 7.6), 400 mM KCl, 0.1 mM EDTA, 10% glycerol, 5 mM ␤-mercaptoethanol, and 0.5 mM phenylmethylsulfonyl fluoride). Following a 30-min incubation on ice, the nuclei extract (supernatant) was retained following centrifugation for 10 min (800 ⫻ g at 4 °C). Protein concentrations from whole and nuclear extracts were determined using the Dc protein assay (Bio-Rad), and all extracts were stored at ⫺80 °C. RNA Extraction and Real Time Polymerase Chain Reaction—Total RNA was isolated from 10 mg of whole lung tissue according to the manufacturers’ instructions using the Rneasy kit (Qiagen). The purified total RNA prep was used as a template to generate cDNA. The reaction mix containing 1 ␮g of RNA, 0.5 ␮g of random hexamers (Promega), 15 units of avian myeloblastosis virus reverse transcriptase enzyme (Promega), and 20 units of Rnasin (Promega) in reverse transcriptase buffer (50 mM Tris-HCl (pH 8.3), 50 mM KCl, 10 mM MgCl2, 10 mM dithiothreitol, and 0.5 mM spermadine) was incubated at 37 °C for 60 min. The reaction was terminated at 90 °C for 2 min and cooled to 4 °C. The quantitative PCR was performed by ABI PRISM 7900HT sequence detection system (Applied Biosystems) using predeveloped Taqman probe/primer combinations for GM-CSF, 18 S rRNA, and TNF␣ optimized by the manufacturer. Taqman PCR was performed in 10-␮l volumes using AmpliTaq Gold polymerase and universal reaction buffer (Applied Biosystems). Threshold cycle numbers were transformed using the ⌬⌬Ct and relative value method as described by the manufacturer and were expressed relative to 18 S rRNA, which was used as a housekeeping gene by multiplexing single reactions. The data were then compared with levels in the PBS control group and are presented as fold increase over PBS alone. Electromobility Shift Assay— The following pairs of complimentary oligonucleotides with 4 nucleotide overhangs were used to generate double-stranded DNA for NF␬B and AP1 binding sites: NF␬B sense, 5⬘-CATGCAACAGAGGGGACTTTCCGAGAGG and NF␬B antisense, 5⬘-CATGCCTCTCGGAAAGTCCCCTCTGTTG-3⬘; AP1 sense, 5⬘-CATGCGCTTGATGAGTCAGCCGGAA-3⬘ and AP1 antisense, 5⬘-CTAGTTCCGGCTGACTCATCAAGCG-3⬘. Complimentary pairs of oligonucleotides (50 ng) were annealed and radiolabelled using 5 units of Klenow fragment (Promega) in the presence of 50 mM Tris-HCl (pH 7.2), 10 mM MgSO4, 0.1 mM DTT, 0.2 mM dGTP, 0.2 mM dCTP, 0.2 mM dTTP, and 100 ␮Ci of [␣-32P]dATP (Geneworks). Labeled probes were then separated from unincorporated isotope by size exclusion chromatography using Microspin G-25 columns (Amersham Biosciences). Nuclear extracts (5 ␮g) were used in binding assays in 20-␮l reactions and incubated at room temperature with labeled probe (50,000 cpm) in the presence of 10 mM Tris-HCl (pH 7.5), 0.05 mg/ml poly[dI-dC]䡠poly[dIdC], 4% glycerol, 1 mM MgCl2, 0.5 mM dithiothreitol, and 50 mM NaCl for 20 min. Following incubation, 15 ␮l of binding reaction was immediately resolved on a 7.5% gel slab (37.5 acrylamide: 1 bis-acrylamide) in 0.5⫻ TBE at 200 V for 20 –30 min. Gels were prepared 1 day prior to experiment and prerun for 30 min before binding reactions were loaded. Gels were then dried and exposed to autoradiography using an Intensifier screen at ⫺80 °C. Densitometry was performed using Kodak EDAS 1D image analysis software. The specificity of the assay was confirmed in separate shift assays on the same nuclear extracts, demonstrating that antibody to the p50 subunit of NF␬B supershifted both bands (not shown). Western Analysis—Whole cell extracts (30 ␮g) were diluted in Laemmli sample buffer, denatured at 90 °C for 5 min, and subjected to SDS-PAGE on 10% gel slabs with 5% stacking gels at 200 V for 45 min. Resolved proteins were transferred onto Hybond polyvinylidene difluoride membrane (Amersham Biosciences) using Trans-blot S.D. transfer cell (Bio-Rad) at 10 V for 30 min. Following transfer, polyvinylidene difluoride membranes were incubated in blocking solution (Tris-buffered saline containing 5% (w/v) skim milk powder and 0.5% (v/v) Tween 20) for 1 h at room temperature. The activation status of Akt, Erk1/2, and p38 was then assessed using phospho-specific antibodies (Cell Signaling, Beverly, MA), which recognize activated forms of these kinases. Actin levels were also assessed as a loading control using an antibody (Santa Cruz Biotechnology, Santa Cruz, CA) that reacts with a broad range of actin isoforms. Membranes were incubated with primary antibodies diluted in blocking solution overnight at 4 °C. Following primary incubation, membranes were washed three times in blocking solution and incubated for 1 h with horseradish peroxidaseconjugated anti-IgG (Bio-Rad). A final wash over 30 min with six changes in wash buffer (Tris-buffered saline containing 0.5% Tween 20) was performed. Immunoreactive bands were finally visualized by auto-


Akt/Erk Pathways Promote LPS Responsiveness in Mouse Lung

radiography using chemiluminescence (ECL, Amersham Biosciences) with exposure times of 30 s to 2 min. Densitometry was performed using Kodak EDAS 1D image analysis software. Protease Expression and Activity in BAL Fluid—Zymography was used to assess protease expression in response to LPS treatment. Briefly, SDS-PAGE mini-gels (10%) were prepared with the incorporation of gelatin (2 mg/ml, Labchem) before casting. BALF (10 ␮l) was run into gels at a constant voltage of 200 V under non-reducing conditions. When the dye front reached the bottom, gels were removed and washed twice for 15 min in 2.5% Triton X-100 and incubated at 37 °C overnight in zymography buffer (50 mM Tris-HCl (pH 7.5), 5 mM CaCl2, 1 mM ZnCl2 and 0.01% NaN3). The gels were then stained for 45 min with Coomassie Brilliant Blue R-250 (Sigma) and extensively destained. Following destaining, zones of enzyme activity appeared clear against the Coomassie Blue background. BALF was also tested for net gelatinase activity using fluorescence-conjugated gelatin (Molecular Probes). The gelatin substrate (10 ␮g) was diluted in 50 mM Tris pH 7.5, 150 mM NaCl, 5 mM CaCl2, 0.01% NaN3 and incubated at room temperature for 16 h with 100 ␮l of neat BALF. The digested substrate has absorption/ emission maxima at 495 nm/515 nm, and its fluorescence intensity was measured in a microplate reader (Victor II, Wallac) to detect quantitative differences in activity. Preparation of Trachea Explants—Trachea and main bronchi were isolated free of adherent connective tissue from untreated anesthetized mice and placed in a 1.5-ml Eppendorf tube containing 250 ␮l of RPMI. For vehicle treatments, 0.1% Me2SO (final v/v) was added. Wortmannin (Sigma) and U0126 (Cell Signaling) were added to a final concentration of 1 and 10 ␮M, respectively, for 30 min at 37 °C, rotating at 750 rpm in a thermo-mixer. Mouse serum (1%, final v/v) prepared in-house and LPS (1 ␮g/ml) were then added to individual tubes, and at 0, 15 min and 3 h, the trachea explants were briefly spun down, and the neat supernatant was retained for GM-CSF ELISA. Tracheas were immediately snap-frozen in liquid nitrogen, and protein extracts were prepared as detailed previously using 200 ␮l of lysis buffer. Data Analysis—As all data were normally distributed, responses were analyzed by ANOVA followed by Dunnett’s multiple comparison test. Unless otherwise specified, data were reported as mean ⫾ S.E. Group sizes in all experiments were at least 8 based on power calculations made to detect a 15% difference from control for p ⬍ 0.05, the level of statistical significance. Note that all experiments were replicated at least three times with identical statistical outcomes. RESULTS

Anti-GM-CSF mAb Blocks LPS-induced Lung Neutrophilia and Neutralizes All Free GM-CSF—Consistent with previous studies (20), 10 ␮g of LPS/mouse caused marked neutrophilic inflammation but not lung hemorrhage or acute respiratory distress. Administration of anti-GM-CSF 3 h prior to LPS challenge markedly suppressed LPS-induced neutrophilia (⬎85%), whereas an equivalent dose of isotype control exerted no significant effect on neutrophil recruitment into the lavage of BALB/c mice (Fig. 1A). The degree of inhibition was equivalent to the suppression of neutrophilia observed in confirmatory experiments in GM-CSF⫺/⫺ mice (Fig. 1B). Anti-GM-CSF recovered by BAL after instillation completely neutralized recombinant GM-CSF (rGM-CSF) but not rTNF␣ (Fig. 2A). Anti-GMCSF also blocked detection of secreted GM-CSF but not TNF␣ protein in LPS-challenged mice, which were markedly enhanced in saline- and LPS-treated mice (Fig. 2B). To establish whether anti-GM-CSF exerted its inhibitory effect at the transcript level, real time fluorescence (Taqman) PCR was used to assess GM-CSF and TNF␣ transcript expression normalized to 18 S rRNA. Consistent with in vitro models (21), LPS promoted expression of GM-CSF and TNF␣ mRNA in vivo (Fig. 2C). Neutralizing GM-CSF significantly inhibited expression of both GM-CSF and TNF␣ mRNA by ⬃26 and 44% respectively, suggesting that GM-CSF may play a role in priming or promoting activation of the inflammatory transcriptional machinery responsible for coordinating gene expression. The discrepancy between TNF␣ protein and transcript implies that preformed TNF␣ stores may be secreted in an anti-GM-CSFresistant manner and are consistent with previous studies demonstrating that protein synthesis inhibition dampens but

FIG. 1. Anti-GM-CSF suppresses LPS-induced lung inflammation. As shown in A, animals were pretreated for 3 h with vehicle (PBS), 100 ␮g/mouse of IgG2a isotype control (Iso), or 100 ␮g of anti-GM-CSF mAb (anti-GM) per mouse and then challenged with 10 ␮g of LPS/ mouse to elicit an intense neutrophilic lung inflammation. As shown in B, wild-type (WT) and GM-CSF-deficient mice were challenged with vehicle or 10 ␮g of LPS as above. BAL was performed at necroscopy at 24 h after challenge. Data are expressed as total macrophages (clear bars) and neutrophils (black bars) in BAL (mean ⫾ S.E.) for 8 –12 mice/treatment group. No significant changes were observed in either cell type. Results are significantly different from vehicle-treated LPSexposed animals; #, p ⬍ 0.005; ns, not significant, ANOVA/Dunnett’s test.

does not completely abolish LPS-induced TNF␣ release in the lung in vivo (22). Anti-GM-CSF Inhibits LPS-induced Matrix Metalloprotease (MMP) Release—Since MMPs contribute to the movement of neutrophils into the lung parenchyma and MMP9 expression in response to LPS is transcriptionally regulated by NF␬B and AP-1 (23, 24), we investigated the secretion of MMP9 in the BALF of LPS-challenged mice by gelatin zymography (Fig. 3A). LPS potently stimulated protease release in the BALF as assessed by clear regions corresponding to zones of degradation. Major bands of protease activity were identified at ⬃90 kDa, corresponding to the molecular size of latent MMP9 (92 kDa) and a major band activity band corresponding to the active form of MMP9 (86 kDa). Additionally, higher molecular size protease bands were observed, corresponding to previously described dimers or complexes of MMP9 (25). We did not observe bands corresponding to MMP2 (72 kDa) or matrilysin (MMP7, 28 kDa). Anti-GM-CSF reduced the expression of MMP9 by about 50%. Net protease activity in neat lavagates was also assessed using gelatin conjugated to a fluorescence probe. Consistent with the zymography, LPS-induced net gelatinase activity was partially suppressed by prior treatment with antiGM-CSF (Fig. 3B). The Activation of NF␬B and AP-1 Is Partially Inhibited by Anti-GM-CSF—As expression of GM-CSF, TNF␣, and MMP9

Akt/Erk Pathways Promote LPS Responsiveness in Mouse Lung


FIG. 3. Anti-GM-CSF suppresses MMP9 levels and activity. As shown in A, animals pretreated with PBS or 100 ␮g of anti-GM-CSF mAb/mouse (Anti-GM) were challenged with saline (clear bars) or 10 ␮g of LPS/mouse for 2 h (black bars) were lavaged, and BALF was incubated with fluorescence-conjugated gelatin. The fluorographic measurement of net free gelatinase activity in the BALF is presented. As shown in B, BALF was also subjected to gelatin zymography. #, p ⬍ 0.05; ANOVA/Dunnett’s test, significantly different from saline and LPS group (n ⫽ 4). FIG. 2. Anti-GM-CSF neutralizes recombinant and endogenous GM-CSF present in BALF and inhibits expression of GM-CSF and TNF␣ transcript. As shown in A, ELISA was used to measure recombinant GM-CSF and TNF␣ prepared in the BALF of PBS (f) or anti-GM-CSF-treated and LPS-exposed animals (Œ) over the specified concentration range. B, endogenous secreted GM-CSF and TNF␣ secreted in BALF of PBS (clear bars) or LPS (black bars, 10 ␮g for 2 h)-challenged mice pretreated with and without 100 ␮g of anti-GMCSF/mouse. #, p ⬍ 0.05; ANOVA/Dunnett’s test, significantly different from saline and LPS group (n ⫽ 4). As shown in C, total RNA was isolated from PBS (clear bar) and LPS-challenged lungs (black bars, 10 ␮g for 2 h) pretreated in the absence or presence of anti-GM-CSF. RNA was reverse-transcribed using random hexamers and subjected to Taqman PCR to detect GM-CSF and TNF␣ transcript relative to 18 S rRNA by multiplexing single reactions. The data were then compared with levels in the PBS control group and are presented as fold increase over PBS alone. #, p ⬍ 0.05; ANOVA/Dunnett’s test, significantly different from saline and LPS group (n ⫽ 4).

are under the regulation of multiple transcription factors including NF␬B and AP-1 (26, 27), we examined their DNA binding activity by electromobility shift assay using nuclear extracts from LPS-challenged lungs. Their respective activities were assessed in a kinetic study in the absence or presence of neutralizing anti-GM-CSF (Fig. 4A) and presented graphically following densitometry analysis (Fig. 4B). LPS induced the DNA binding activity of nuclear NF␬B and AP-1, and maximal activity was maintained between 1 and 6 h. Importantly, antiGM-CSF reduced the activity of NF␬B by 30 – 40% during the maximal activation period, consistent with the decline in TNF␣ transcript and MMP9 expression. The baseline activity of AP-1 was slightly elevated by anti-GM-CSF mAb, possibly due to anti-GM-CSF antibody binding to Fc receptors on lung mast cells or macrophages (28). Fc receptor signaling is known to feed into the Ras pathway for AP-1 activation (28), but in the absence of concurrent basal NF␬B activation (Fig. 1), this AP-1 response alone is clearly not sufficient to cause neutrophilic inflammation in vivo. This LPS-induced activation of AP-1 was

suppressed by anti-GM-CSF, as maximal activation between 1 and 6 h was reduced by 50%. Anti-GM-CSF Inhibition of Akt and Erk1/2 Signaling Correlates with Dampened NF␬B and AP-1 Nuclear Localization— 1L-1␤ receptor activates NF␬B by promoting the dissociation of I␬B from NF␬B, which is followed by the Akt-mediated transactivation of p65/RelA for complete activation (7). Similarly, TLR4 promotes I␬B ubiqitination; however, the role of Akt in LPS-induced lung inflammation remains unclear. Furthermore, the activation of two MAPK homologs, Erk1/2 and p38, is known to be important in AP-1 regulation (18); however MAPK substrate specificities in our mouse model have not been established. Therefore, we examined Akt, Erk1/2, and p38 activity in whole lung extracts derived from LPS-challenged mice by Western analysis using phospho-specific antibodies that only recognize activated forms of their respective kinases. Kinase activity was assessed in a kinetic study over 24 h, and actin was also immunoblotted as an internal loading control (Fig. 5A). Densitometry analysis was performed, and activities were presented as a fold increase above baseline (Fig. 5B). LPS rapidly induced activation of Akt and Erk1/2 within 15 min, suggesting that Akt and Erk are functional modules of the TLR4 signaling cascade. A minor secondary phase of Akt/Erk activity was prevalent between 2 and 6 h, falling to baseline by 24 h. This secondary wave of kinase activity correlates with expression of GM-CSF and TNF␣ levels that peak at 2 h following LPS challenge. In contrast, p38 was not significantly activated in LPS-inflamed lungs, indicating an important role for Erks in AP-1 activation. Prophylactic administration of anti-GM-CSF mAb suppressed the primary phase of LPS-induced Akt activation by about 50%. This decline in activity correlates with the 30 – 40% decline in NF␬B DNA binding activity and is consistent with Akt contributing to promote cytokine production by regulating NF␬B. Anti-GM-CSF elevated basal Erk1/2 activity, which is consistent with the profile


Akt/Erk Pathways Promote LPS Responsiveness in Mouse Lung

FIG. 4. Anti-GM-CSF therapy suppressed the LPS-induced activation of NF␬B and AP-1. As shown in A, whole lungs isolated from both LPS-exposed (⫺) and anti-GM-CSF-treated and LPS-exposed (⫹) mice were harvested at the indicated time points over 24 h. Nuclear lysates were prepared and incubated with radiolabelled probe corresponding to the consensus-binding site for NF␬B and AP-1, and their respective DNA binding activity was determined by electromobility shift assay. As shown in B, autoradiograph densitometry data are expressed as fold induction relative to baseline (value at 0 h) and are representative of at least three replicates per time point.; LPS-challenged (f) and anti-GM-CSF-treated and LPS-challenged animals (Œ).

for AP-1. Notably, however, maximal activation of Erk1/2 in response to LPS challenge was suppressed. Inhibition of Akt and Erks by Wortmannin and U0126, Respectively, Blocked GM-CSF Release from Trachea Explants—As pharmacological inhibitors of Akt and Erk directly cause lung inflammation in our hands, to confirm the roles of Akt and Erks, tracheal segment explants were prepared and treated with pharmacological inhibitors of Akt (wortmannin) and Erks (U0126) prior to LPS challenge. The activities of both Akt and Erks were assessed by Western analysis using phospho-specific antibodies (Fig. 6A). LPS induced activation of Akt within 15 min, and this activation was abrogated by wortmannin. The degree of activation we observed in isolated trachea ex vivo was more intense than that we observed in whole lung but similar to levels reported previously for LPS-stimulated macrophages in vitro (29). The lower peak activation observed in the whole lung may be due to the presence of LPS-insensitive cell populations. It is noteworthy that Akt activation in vivo was biphasic and that the duration of activation was sustained in comparison with the ex vivo trachea and is consistent with the time course of induction of detectable GM-CSF and TNF protein. Consistent with our in vivo data, the decline in Akt activity in trachea explants correlated with a reduction in GM-CSF secretion, as assessed by ELISA (Fig. 6B). In contrast, Erks displayed a high level of phosphorylation at baseline, most likely due to tissue reaction to preparing the explants (30), which was not significantly enhanced by LPS. Importantly, prior treatment with U0126 completely abolished Erk activity and also potently suppressed GM-CSF release after LPS challenge, suggesting that Erk is essential, although alone not sufficient to promote GM-CSF production.

FIG. 5. GM-CSF-dependent activation of Akt and Erk in LPSexposed lungs. As shown in A, whole lungs isolated from LPS-challenged (⫺) and anti-GM-CSF-treated and LPS-challenged (⫹) mice were harvested at the indicated time points over 24 h. Whole cell extracts were prepared and analyzed by Western blotting with phospho-specific antibodies that recognize activated forms of Akt, Erk1/2, and p38, respectively. As a control, lysates were also immunoblotted with an antibody that reacts with a broad range of actin isoforms. As shown in B, autoradiograph densitometry data are expressed as fold induction and are representative of at least three replicates per time point; LPS-challenged (f) and anti-GM-CSF-treated and LPS-challenged animals (Œ). DISCUSSION

Despite the importance of LPS as a trigger of innate host defenses and lung inflammation, very little is known of the actual in vivo transduction pathways activated by this endotoxin in the lung. Consistent with previous reports in mice and humans (6, 31), LPS promoted an intense neutrophilic inflammation that was initiated by TNF␣. Our data indicate that GM-CSF is critical to this process and are consistent with the near complete suppression of LPS-induced neutrophilia resistance to systemic endotoxin shock (32) in GM-CSF-deficient mice and the inflamed lung phenotype of transgenic mice that selectively overexpress GM-CSF in lungs (33). Extensive in vitro studies implicate NF␬B and AP-1 as obligatory transcription factors in inflammation and innate immunity (34). Since we have demonstrated that anti-GM-CSF reduces their respective transcriptional activities, we reasoned that alternate pathways contribute to NF␬B activity and identify Akt as a likely IRAK-independent module. Furthermore, we identify Erk as the most likely regulator of AP-1 activity. Despite the recent finding using putative pharmacological p38 inhibitors that implicate a key role for p38 in LPS-triggered inflammation (35), we observed no significant changes in lung p38 activity at any time point, suggesting that the p38 inhibitors most likely dampen leukocyte diapedesis within the granulocyte.

Akt/Erk Pathways Promote LPS Responsiveness in Mouse Lung

FIG. 6. Wortmannin and U0126 suppress GM-CSF release from trachea explants. As shown in A, trachea explants isolated from healthy mice were treated with 0.1% Me2SO (DMSO) (vehicle), 1 ␮M wortmannin, or 10 ␮M U0126 for 30 min prior to LPS challenge (1 ␮g/ml). At 0, 15 min, and 3 h after LPS exposure, tracheas were retained, and whole cell extracts prepared for Western analysis with phospho-specific antibodies that recognize activated forms of Akt and Erk1/2. As a control, lysates were also immunoblotted with an antibody that reacts with a broad range of actin isoforms. As shown in B, the supernatant was also retained following LPS exposure, and ELISA was used to measure GM-CSF levels. #, p ⬍ 0.05; ANOVA/Dunnett’s test, significantly different from Me2SO and LPS group (n ⫽ 4).

In most tissues and cell types, GM-CSF is not basally produced; however, lung epithelial cells produce significant amounts of GMCSF at baseline (36). Consistently, we detected a 5-fold higher basal level of GM-CSF transcript as compared with TNF␣ transcript, as assessed by Taqman PCR (data not shown). GM-CSF is also rapidly consumed by receptor internalization to account for its lack of detection by ELISA in untreated mice (37). Anti-GM-CSF (22E9) is a high affinity neutralizing monoclonal antibody that sequesters free GM-CSF, thereby blocking its multipotential functions, which include augmenting neutrophil survival and activities. As anti-GM-CSF almost completely inhibited neutrophilic inflammation, it is likely that LPS-induced GM-CSF secretion augments innate immunity and retards apoptosis of recruited neutrophils in the lung. This suggestion is strengthened by results in GM-CSFdeficient mice since these animals have normal hematopoiesis and normal circulating neutrophil numbers (19). GM-CSF mediates survival of isolated neutrophils via an Akt-dependent mechanism, and suppression of Akt activity promotes the proapoptotic molecules BAD and caspase-3 to initiate programmed cell death (14, 38). Importantly, we established that this neutralizing antibody can also partially block release of GM-CSF and inhibit TNF␣ transcript as well as MMP9 activity in LPS-exposed lungs. Collectively, these observations suggest that steady state GM-CSF may also be required to prime LPS responses in the lung. One explanation supported by in vitro evidence in macrophages and in vivo evidence in GM-CSF-deficient mice is that GM-CSF regulates expression of CD14 and TLR4 (12). However, macrophages recovered from antiGM-CSF-treated lungs do not down-regulate TLR4 expression in our studies. Very recently, LPS has been shown to activate Akt in monocytes under in vitro conditions (39), although the signaling cascade promoting its activation has not yet been established. We observed rapid activation of Akt in LPS-inflamed lungs, suggesting that the kinase is a functional component of the LPS/TLR4 pathway. Since deletion of IRAK attenuates without eliminating cytokine responsiveness in these mutant mice (40),


we suggest that the IRAK-independent component of LPSinduced NF␬B activation corresponds to activation of Akt. Consistent with this proposal, we found that that LPS-induced activation of Akt in trachea explants was sensitive to prior incubation with wortmannin, a selective inhibitor of the upstream activator of Akt termed phosphatidylinositol-3 kinase. Furthermore, this reduction in activity correlated with a decline in GM-CSF expression to indicate the requirement of Akt for full activation of NF␬B. In addition to inducing the activation of NF␬B, LPS also induced activation of Erk1/2 in treated lungs. The Erk1/2 isoforms were also rapidly stimulated within 15 min of LPS challenge. Activation of Erk is relayed through a well characterized signal transduction pathway involving the small GTP-binding protein Ras. The GTP-bound form of Ras recruits and activates Raf at the cellular membrane, where it then activates the Erk kinases (MEKK1 and MEKK2). MEKK1/2 then phosphorylate and potently activate Erk1/2. More recently, receptor-bound TRAF6 has been shown to interact with Ras and activate Erk (41) in an analogous manner to the classical pathway. Our studies have shown that the LPS-induced primary activation of Erk1/2 was sensitive to prior administration of the neutralizing anti-GM-CSF mAb. The impairment of Erk signaling implicates the kinase as an important mediator of neutrophilic lung inflammation via its ability to regulate the AP-1 transcription factor. AP-1 exists either as a homodimeric c-Jun complex or as a c-Jun/c-Fos heterodimer that is regulated by transcription and direct phosphorylation. Phosphorylation of c-Jun stabilizes the transcription factor and enhances the trans-activation and DNA binding of AP-1 (42). Either Erk or the JNK homolog can mediate phosphorylation of c-Jun in a cell type-dependent manner (18). Here, we demonstrate that the activation of Erk1/2 by LPS challenge correlates temporally with the subsequent DNA binding activity of AP-1. Furthermore, suppression of the LPS-induced activation of Erk1/2 by anti-GM-CSF was consistent with its inhibition of the transcription factor. Pharmacological inhibition of Erk activity in trachea explants with U0126 also potently suppressed the release of GM-CSF ex vivo, suggesting a critical role in transcriptional regulation. Therefore, the Erk1/2 homologs appear to play a primary role in AP-1 activation in LPS-inflamed lungs. Normal cells and tissues, with the exception of neutrophils, tightly regulate MMP expression and activation by growth factors, cytokines, and inflammatory mediators to prevent tissue destruction. Recently, Akt has been shown to regulate the invasive properties of highly metastatic HT1080 cells by modulating the NF␬B-driven expression of MMP9 (43), and ectopic expression of GM-CSF in BALB/c mammary carcinoma F311 cells promotes protease-dependent invasiveness (44). In this study, we establish that the LPS-induced secretion of GM-CSF regulates the release of MMP9. Anti-GM-CSF-mediated suppression of NF␬B and AP-1 was associated with a significant reduction in the expression of MMP9 (active, latent, and complexed forms), consistent with a decline in net gelatinase activity in the BALF of LPS-challenged mice. NF␬B activity is known to regulate MMP9 (43, 44). The association between suppression of AP-1 and MMP9 levels, however, does not prove a causal relationship between these observations. This benefit of neutralizing GM-CSF may have clinical implications as repeated exposure to LPS causes emphysema in vivo (4). Our data provide a molecular basis for the beneficial role of GM-CSF in innate immune responses essential for host defense in the lung. Conversely, the ability of anti-GM-CSF to suppress activation of the upstream kinase Erks and Akt and to dampen AP-1 and NF␬B activation provides a rationale for the pronounced anti-inflammatory activity of anti-GM-CSF that may


Akt/Erk Pathways Promote LPS Responsiveness in Mouse Lung

be of clinical utility in lung disease, such as asthma, chronic obstructive pulmonary disease, and cystic fibrosis, in which LPS is implicated in pathogenesis. REFERENCES 1. Medzhitov, R., and Janeway, C. A., Jr. (1997) Cell 91, 295–298 2. Martin, T. R. (2000) Am. J. Respir. Cell Mol. Biol. 23, 128 –132 3. Li, J. D., Dohrman, A. F., Gallup, M., Miyata, S., Gum, J. R., Kim, Y. S., Nadel, J. A., Prince, A., and Basbaum, C. B. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 967–972 4. Vernooy, J. H., Dentener, M. A., van Suylen, R. J., Buurman, W. A., and Wouters, E. F. (2002) Am. J. Respir. Cell Mol. Biol. 26, 152–159 5. Michel, O., Kips, J., Duchateau, J., Vertongen, F., Robert, L., Collet, H., Pauwels, R., and Sergysels, R. (1996) Am. J. Respir. Crit. Care Med. 154, 1641–1646 6. O’Grady, N. P., Preas, H. L., Pugin, J., Fiuza, C., Tropea, M., Reda, D., Banks, S. M., and Suffredini, A. F. (2001) Am. J. Respir. Crit. Care Med. 163, 1591–1598 7. Medzhitov, R., Preston-Hurlburt, P., and Janeway, C. A., Jr. (1997) Nature 388, 394 –397 8. Bowie, A., and O’Neill, L. A. (2000) J. Leukocyte Biol. 67, 508 –514 9. Beutler, B. (2000) Curr. Opin. Immunol. 12, 20 –26 10. Swantek, J. L., Tsen, M. F., Cobb, M. H., and Thomas, J. A. (2000) J. Immunol. 164, 4301– 4306 11. Klein, J. B., Buridi, A., Coxon, P. Y., Rane, M. J., Manning, T., Kettritz, R., and McLeish, K. R. (2001) Cell. Signal. 13, 335–343 12. Shibata, Y., Berclaz, P. Y., Chroneos, Z. C., Yoshida, M., Whitsett, J. A., and Trapnell, B. C. (2001) Immunity 15, 557–567 13. Dijkers, P. F., van Dijk, T. B., de Groot, R. P., Raaijmakers, J. A., Lammers, J. W., Koenderman, L., and Coffer, P. J. (1999) Oncogene 18, 3334 –3342 14. Klein, J. B., Rane, M. J., Scherzer, J. A., Coxon, P. Y., Kettritz, R., Mathiesen, J. M., Buridi, A., and McLeish, K. R. (2000) J. Immunol. 164, 4286 – 4291 15. Ozes, O. N., Mayo, L. D., Gustin, J. A., Pfeffer, S. R., Pfeffer, L. M., and Donner, D. B. (1999) Nature 401, 82– 85 16. Romashkova, J. A., and Makarov, S. S. (1999) Nature 401, 86 –90 17. Madrid, L. V., Wang, C. Y., Guttridge, D. C., Schottelius, A. J., Baldwin, A. S., Jr., and Mayo, M. W. (2000) Mol. Cell. Biol. 20(5), 1626 –1638 18. Leppa, S., Saffrich, R., Ansorge, W., and Bohmann, D. (1998) EMBO J. 17, 4404 – 4413 19. Stanley, E., Lieschke, G. J., Grail, D., Metcalf, D., Hodgson, G., Gall, J. A., Maher, D. W., Cebon, J., Sinickas, V., and Dunn, A. R. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 5592–5596 20. Blyth, D. I., Pedrick, M. S., Savage, T. J., Bright, H., Beesley, J. E., and Sanjar,

S. (1998) Am. J. Respir. Cell Mol. Biol. 19, 38 –54 21. Meja, K. K., Seldon, P. M., Nasuhara, Y., Ito, K., Barnes, P. J., Lindsay, M. A., and Giembycz, M. A. (2000) Br. J. Pharmacol. 131, 1143–1153 22. Goncalves de Moraes, V. L., Boris Vargaftig, B., Lefort, J., Meager, A., and Chignard, M. (1996) Br. J. Pharmacol. 117, 1792–1796 23. Kim, H., and Koh, G. (2000) Biochem. Biophys. Res. Commun. 269, 401– 405 24. Aljada, A., Ghanim, H., Mohanty, P., Hofmeyer, D., Tripathy, D., and Dandona, P. (2001) J. Clin. Endocrinol. Metab. 86, 5988 –5991 25. Zheng, T., Zhu, Z., Wang, Z., Homer, R. J., Ma, B., Riese, R. J., Jr., Chapman, H. A., Jr., Shapiro, S. D., and Elias, J. A. (2000) J. Clin. Invest. 106, 1081–1093 26. Coles, L. S., Diamond, P., Occhiodoro, F., Vadas, M. A., and Shannon, M. F. (2000) J. Biol. Chem. 275, 14482–14493 27. Jongeneel, C. V. (1995) Immunobiology 193(2– 4), 210 –216 28. Turner, H., and Cantrell, D. A. (1997) J. Exp. Med. 185, 43–53 29. Monick, M. M., Carter, A. B., Robeff, P. K., Flaherty, D. M., Peterson, M. W., and Hunninghake, G. W. (2001) J. Immunol. 166, 4713– 4720 30. Santiago, F. S., Lowe, H. C., Day, F. L., Chesterman, C. N., and Khachigian, L. M. (1999) Am. J. Pathol. 154, 937–944 31. Starcher, B., and Williams, I. (1989) Lab. Anim. 23, 234 –240 32. Basu, S., Dunn, A. R., Marino, M. W., Savoia, H., Hodgson, G., Lieschke, G. J., and Cebon, J. (1997) J. Immunol. 159, 1412–1417 33. Xing, Z., Ohkawara, Y., Jordana, M., Graham, F., and Gauldie, J. (1996) J. Clin. Invest. 97, 1102–1110 34. Guha, M., and Mackman, N. (2001) Cell. Signal. 13, 85–94 35. Underwood, D. C., Osborn, R. R., Bochnowicz, S., Webb, E. F., Rieman, D. J., Lee, J. C., Romanic, A. M., Adams, J. L., Hay, D. W., and Griswold, D. E. (2000) Am. J. Physiol. 279, L895–L902 36. Nakata, K., Akagawa, K. S., Fukayama, M., Hayashi, Y., Kadokura, M., and Tokunaga, T. (1991) J. Immunol. 147, 1266 –1272 37. Metcalf, D., Nicola, N. A., Mifsud, S., and Di Rago, L. (1999) Blood 93, 1579 –1585 38. Goyal, A., Wang, Y., Graham, M., Dossef, A., Bhatt, N., and Marsh, C. (2001) Am. J. Respir. Cell Mol. Biol. 26, 224 –230 39. Monick, M. M., Mallampalli, R. K., Carter, A. B., Flaherty, D. M., McCoy, D., Robeff, P. K., Peterson, M. W., and Hunninghake, G. W. (2001) J. Immunol. 167, 5977–5985 40. Thomas, J. A., Allen, J. L., Tsen, M., Dubnicoff, T., Danao, J., Liao, X. C., Cao, Z., and Wasserman, S. A. (1999) J. Immunol. 163, 978 –984 41. O’Neill, L. (2000) Biochem. Soc. Trans. 28, 557–563 42. Musti, A. M., Treier, M., and Bohmann, D. (1997) Science 275, 400 – 402 43. Kim, D., Kim, S., Koh, H., Yoon, S. O., Chung, A. S., Cho, K. S., and Chung, J. (2001) FASEB J. 15, 1953–1962 44. Gabri, M. R., Menna, P. L., Scursoni, A. M., Gomez, D. E., and Alonso, D. F. (1999) Pathobiology 67, 180 –185