INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Airway wall inflammation is a central feature to the pathophysiology of asthma and other airway diseases such as chronic obstructive pulmonary disease. The induction and perpetuation of inflammation involves a complex and coordinated response of multiple inflammatory cells, mediators, and cytokines. Evidence from our own studies (1) and from others (4), suggests that airway smooth muscle (ASM) cells, which have long been regarded as having predominantly contractile properties in response to inflammatory mediators, can potentially contribute to the pathogenesis of asthma by expressing and secreting multiple proinflammatory cytokines and mediators (9, 10). In particular, the production by airway smooth muscle of RANTES (regulated upon activation, normal T cell expressed and secreted) (1) and eotaxin (4, 5) implies a role for these structural cells to participate directly in the inflammatory process through recruitment and activation of eosinophils. In addition to recruitment, enhanced survival of infiltrating eosinophils is also thought to contribute to airway inflammation in asthma (11), and we have recently demonstrated that airway smooth muscle cells stimulated with interleukin (IL)-1 enhance the survival of peripheral blood eosinophils in culture by the release of granulocyte-macrophage colony-stimulating factor (GM-CSF) (2). Determining the role of airway smooth muscle in inflammation may be useful for understanding the pathogenesis of airway disease. As a consequence, there is now considerable interest in characterizing the stimuli and their associated intracellular signaling mechanisms that regulate the production of such cytokines by airway smooth muscle cells.

Many inflammatory cytokines and polypeptide growth factors elicit specific cellular responses through activation of mitogen-activated protein (MAP) kinase cascades (12, 13). Among the best characterized of the mammalian MAP kinase superfamily of serine/threonine kinases are the p42- and p44-kD extracellular signal-regulated kinases (ERKs) ERK2 and ERK1 (collectively defined as p42/p44 ERK), the p38 MAP kinase, and the p46 to p54-kD c-Jun amino-terminal kinase (JNK) or stress-activated protein kinase (SAPK). Each of these highly homologous MAP kinases is phosphorylated and catalytically activated by a signaling cascade comprising a series of sequentially activated conserved protein kinases, which include MAP kinase kinase kinases (MKKK) and dual-specificity MAP kinases (MKK), which include the MAP kinase, ERK-activating kinase (MEK). Activation of p42/p44 ERK MAP kinases occurs in response to many mitogenic stimuli and is important in cellular growth and differentiation as well as cell survival and death (12, 14). p38 MAP kinases and SAPK/JNK are activated by environmental stresses such as hyperosmotic shock, heat shock, endotoxins (lipopolysaccharide), and ultraviolet (UV) irradiation, and are thought to be important in apoptosis and cytokine expression (14). The p38 MAP kinases are also activated by proinflammatory cytokines such as IL-1 and tumor necrosis factor (TNF)-, implicating MAP kinase pathways as important intracellular signaling mechanisms in the inflammatory response. Although IL-1 and TNF-, through activation of distinct cell surface receptors, are known to induce proinflammatory cytokine production from airway smooth muscle cells (for reviews see [9] and [10]), the intracellular signaling mechanisms that regulate cytokine expression and release from these cells have not been determined.

In the present study, using specific chemical inhibitors of p38 MAP kinase (SB 203580 and SB 202190) and p42/p44 ERK (PD 98059 and U 0126) activation, experiments were performed to characterize the dependency of MAP kinase activation in the release of eosinophil-activating cytokines from human airway smooth muscle cells that were stimulated by IL-1. Our results demonstrate that in human airway smooth muscle cells, both the p42/p44 ERK and p38 MAP kinase pathways differentially regulate the release of GM-CSF, RANTES, and eotaxin in response to proinflammatory stimuli. A preliminary account of this study has been communicated to the American Thoracic Society (15).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reagents

All chemicals were of analytical grade or higher. Recombinant human IL-1 and matched antibody pairs for GM-CSF, RANTES, and eotaxin enzyme-linked immunosorbent assays (ELISAs) were purchased from R&D Systems (Abingdon, UK). The p38 MAP kinase inhibitors, SB 203580 (4-[4-fluorophenyl]-2-[4-methylsulfinylphenyl]- 5-[4-pyridyl]1H-imidazole) and SB 202190 (4-[4-fluorophenyl]-2-[4-hydroxyphenyl]-5-[4-pyridyl]1H-imidazole), and the inactive negative control compound, SB 202474 (4-[ethyl]-2-[4-methoxyphenyl]-5-[4-pyridyl]1H-imidazole), as well as the MEK inhibitors PD 98059 (2'-amino-3'-methoxy flavone), U 2016 (1,4-diamino-2,3-dicyano-1,4-bis [2-aminophenylthio]butadiene), and the Protease Inhibitor Cocktail Set III (Cat. 539134) were purchased from Calbiochem (Nottingham, UK). All cell culture medium (DMEM and RPMI 1640), fetal bovine serum (FBS), and cell culture reagents were obtained from Gibco Life Technologies (Paisley, UK). Collagenase (type CLS 1) was obtained from Worthington Biochemical Corporation (Freehold, NJ). All cell culture plasticware was purchased from Falcon Labware (Becton Dickinson, Oxford, UK). All other chemical reagents including those for bicinchoninic acid protein assay were obtained from Sigma (Poole, UK).

Isolation and Culture of Human Airway Smooth Muscle Cells

In accordance with procedures approved by the Guy's and St. Thomas' Hospitals Research Ethics Committee, human bronchial smooth muscle was obtained from the lobar or main bronchus of 26 patients of either sex (mean age 63 ± 2 yr; range 34-84 yr; 18 male, eight female) undergoing lung resection for carcinoma of the bronchus. After removal of the epithelium, portions of the smooth muscle not invaded by the carcinoma were dissected free of adherent connective and parenchymal tissue under aseptic conditions in Hanks' balanced salt solution and placed in culture as previously described (2). Briefly, finely chopped (1 mm³ approx.) pieces of smooth muscle were digested overnight in 1 ml Dulbecco's modified Eagle medium (DMEM) (supplemented with 1 mM sodium pyruvate, 2 mM L-glutamine, 1× nonessential amino acid mixture, 50 µg/ml gentamicin, and 1.5 µg/ml amphotericin B) containing 1 µM insulin, 5 µg/ml transferrin, 100 µM ascorbate, 1 mg/ml bovine serum albumin (BSA), and 1 mg/ml collagenase. The resulting cell suspension was centrifuged (200 × g for 5 min) and the pellet was washed in supplemented DMEM containing 10% FBS. Cells were seeded at 5 × 10⁵ viable cells in 25-cm² culture flasks and maintained in a humidified atmosphere at 37° C in 5% CO₂, with the medium replaced every 72 h. Using fluorescent immunocytochemistry techniques, growth-arrested cultured human airway smooth muscle cells (passage 1 and 2) stained (> 95%) for both smooth muscle -actin and smooth muscle myosin heavy chain. When examined by light and electron microscopy, these cells displayed all the reported characteristics of viable smooth muscle cells in culture (16).

Cell Stimulation and Collection of Cell-conditioned Medium

Cells at passages 2 to 6 were used for all experiments and were harvested from 75-cm² flasks by treatment with trypsin/EDTA (ethylenediamine tetraacetic acid) (0.2 mg/ml of each in phosphate-buffered saline) and washed in supplemented DMEM containing 10% FBS as previously described. Cells were then seeded into 24-well plastic tissue culture plates at an initial density of 2 × 10⁴ cells/well. When the cells approached confluence, growth was arrested by washing (1 × 0.5 ml/well) the monolayers with RPMI 1640 for 2 min and then replacing the washing medium with RPMI 1640 supplemented with 25 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid (HEPES), 2 mM L-glutamine, 100 U/ml:100 µg/ml penicillin/streptomycin (supplemented RPMI) with the addition of 1 µM insulin, 5 µg/mL transferrin, 100 µM ascorbate and 1 mg/ml BSA. After 72 h airway smooth muscle cell monolayers were washed (1 × 0.5 ml/well) with supplemented RPMI 1640 containing 1 mg/ml BSA and then cultured in the same medium (0.5 ml/well) for a further period up to 24 h in the absence or presence of varying concentrations of recombinant human IL-1 at 37° C in a CO₂ incubator. Where indicated, cells were also pretreated for 30 min with pharmacological inhibitors. Inhibitors were dissolved in dimethyl sulfoxide (DMSO) at 50 mM (100 mM for PD 98059) and further dilutions were made in cell culture medium. The highest concentration of DMSO to which cells were exposed was 0.02% (corresponding to 10 µM of inhibitor) or 0.2% (corresponding to 100 µM of inhibitor). DMSO vehicle controls were included in each experiment. When present, inhibitors remained in the media throughout the cytokine stimulation, except in experiments where the addition of inhibitors was delayed. Cell-conditioned medium (0.5 ml/well) was collected, and cell-free supernatants were stored at 70° C until measurement of cytokine levels by ELISA.

Measurement of Cytokine Levels by ELISA

Cytokine levels in airway smooth muscle cell-conditioned culture medium were determined in duplicate by specific sandwich enzyme-linked immunosorbent assays (ELISA) using matched monoclonal (anti-human) capture and biotinylated anti-human monoclonal (GM-CSF) or polyclonal (RANTES, eotaxin) detection antibody pairs. The manufacturer's instructions were followed throughout. Samples were diluted until the level of cytokine was within the limits of the standard curve of the assay. Concentrations of cytokines were detected spectrophotometrically (Anthos HTII; Salzburg, Austria) in cell-conditioned medium and initially expressed in ng/ml before normalization to ng/ml/million cells to correct for small differences in cell densities between patients. The lower limits of sensitivity for each of the assays were GM-CSF < 2.8 pg/ml; RANTES < 5 pg/ml; eotaxin < 5 pg/ml. None of the assays showed any cross-reactivity with each other or interference with IL-1.

Western Immunoblot Analysis of p38 MAP Kinase, ERK, and SAPK/JNK Phosphorylation

Human airway smooth muscle cells were grown to near confluence in 25-cm² flasks. After growth arrest as above for 96 h, cells were stimulated with 1 ng/mL IL-1 at 37° C in a CO₂ incubator. Where indicated, cells were also pretreated for 30 min with pharmacological inhibitors. When present, inhibitors remained in the media throughout the cytokine stimulation. At the indicated time points, cells were washed twice with ice-cold PBS containing protease inhibitors (200 µM Na₃VO₄, 2 mM phenylmethylsulfonyl fluoride) and lysed by scraping in RIPA buffer (PBS containing 1% Igepal, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 20 µl/10⁶ cells/400 µl Protease Inhibitor Cocktail Set III and 200 µM Na₃VO₄). Cellular extracts were heated to 90° C for 10 min and clarified by centrifugation (10,000 × g for 5 min). Supernatants were analyzed for protein content using the bicinchoninic acid assay (17). Samples of total protein extracts (10 µg/ lane) were separated by SDS-PAGE on 10% acrylamide precast gels (Novex, Frankfurt, Germany). The separated proteins were transferred electrophoretically to nitrocellulose membranes (25 V for 90 min) and the blots were blocked in PBS containing Tween 20 (0.05%) (PBS-T) and dried nonfat milk (5%) for at least 1 h. To detect MAP kinase activation, blots were probed with antibodies that recognize the phosphorylated forms of p42/p44 ERK, SAPK/JNK (New England Biolabs, Beverly, MA), and p38 (Calbiochem). Antibodies were diluted in PBS-T containing dried nonfat milk (1%) and BSA (1%) and blots were incubated as follows: p42/p44 at 1:2,000 dilution for 90 min at room temperature; p38 at 1:750 dilution overnight at 4° C; SAPK/JNK at 1:1,000 dilution overnight at 4° C. These antibodies specifically recognize the phosphorylated amino acids Thr²⁰² of p42/ p44 ERK, Thr¹⁸⁰/Tyr¹⁸² of p38, or Thr¹⁸³/Tyr¹⁸⁵ of SAPK/JNK. The blots were then washed with PBS-T and incubated at room temperature for 1 h with goat anti-rabbit immunoglobulin (Ig) G horseradish peroxidase (HRP)-conjugated secondary antibody (1:4,000) to detect p38 activation or goat anti-mouse IgG HRP-conjugated secondary antibody (1:5,000), to detect ERK or SAP/JNK activation, thoroughly washed in PBS-T and visualized by enhanced chemiluminescence (Amersham-Pharmacia, UK). Signals were quantified using ImageQuant software (Molecular Dynamics, Sunnyvale, CA) on autoradiographs that depicted bands within a linear range of exposure. Densitometric data were normalized to unstimulated control cells, which were set to 1.0.

p38 MAP Kinase, ERK Kinase, and SAPK/JNK Assay

The activity of p38 MAP kinase was analyzed using a commercially available kit (p38 MAP kinase assay kit, cat. 9820; New England BioLabs, Beverly, MA). The kit employs two different phospho-specific antibodies. The first is a monoclonal anti-phospho-specific p38 MAP kinase antibody (Thr¹⁸⁰/Tyr¹⁸²) that is used to selectively immunoprecipitate active p38 MAP kinase from cell lysates, and does not cross-react with p42/44 ERK or SAPK/JNK. Following an in vitro kinase assay of its substrate, activating transcription factor (ATF)-2, p38 MAP kinase activity was determined by Western blotting using a second anti-phospho-specific antibody that detected p38 MAP kinase-induced phosphorylation of ATF-2 at Thr⁷¹. The kinase activity of p42/p44 ERK was analyzed using a similar kit (p42/p44 MAP kinase assay kit, cat. 9800; New England BioLabs) in which a monoclonal anti-phospho-specific p42/p44 ERK (Thr²⁰²/Tyr²⁰⁴) antibody was used to selectively immunoprecipitate active p42/p44 ERK from cell lysates. This did not cross-react with p38 MAP kinase or SAPK/JNK. Following an in vitro kinase assay of its substrate protein, erythroleukemia virus 26-like (Elk-1), p42/p44 MAP kinase activity was determined by Western blotting using a second anti-phospho-specific antibody that detected p42/p44 ERK-induced phosphorylation of EIk-1 at Ser³⁸³. SAPK/JNK activity was also analyzed using a kit (SAPK/JNK kinase assay kit, cat. 9810; New England BioLabs). The kit employed an amino-terminal c-Jun (codons 1-89) fusion protein bound to glutathione sepharose beads to precipitate SAPK/JNK from cell lysates. c-Jun (1-89) contains a high affinity binding site for SAPK/JNK. Following the in vitro kinase reaction step, phosphorylation of c-Jun at Ser⁶³/Ser⁷³ was determined by Western blotting using an anti-phospho-specific c-Jun antibody that detects SAPK/JNK-induced phosphorylation of c-Jun at Ser⁶³.

After stimulation by IL-1 (1 ng/mL) in the absence or presence of inhibitors, near-confluent cells in 75-cm² flasks were washed once with ice-cold PBS containing protease inhibitors (200 µM Na₃VO₄, 2 mM phenylmethylsulfonyl fluoride) and lysed by scraping in ice-cold lysis buffer supplied in the kits (20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM -glycerolphosphate, 2 mM Na₃VO₄, 1 µg/ml leupeptin). Cellular extracts were sonicated (4 × 10 s) and clarified by centrifugation (10,000 × g for 5 min at 4° C). Cell lysates containing 200 µg (250 µg for assay of SAPK/JNK activity) of protein were incubated overnight at 4° C with anti-phospho-specific p38 or p42/p44 MAP kinase antibody to selectively immunoprecipitate active p38 or p42/p44 MAP kinase, respectively. Immunoprecipitates were incubated for 30 min at 30° C with the appropriate fusion proteins, ATF-2 (for assay of p38 MAP kinase activity) or Elk-1 (for assay of p42/p44 ERK activity) in the presence of ATP (200 µM). For determination of SAPK/JNK activity cell lysates were incubated directly with an amino-terminal c-Jun fusion protein bound to glutathione sepharose beads. Kinase reactions were stopped by addition of 4× SDS-PAGE sample buffer and heating to 90° C (5 min), and samples were separated by SDS-PAGE on 10% acrylamide precast gels (Novex). The separated proteins were transferred electrophoretically (25 V for 90 min) to nitrocellulose membranes and blotted with anti-phospho-specific antibodies for either ATF-2, Elk-1, or c-Jun. Membranes were incubated with an HRP-conjugated anti-rabbit antibody (New England BioLabs; 1:4,000) and then visualized by enhanced chemiluminescence (LumiGLO; KPL Europe, Guildford, UK). Signals were quantified using ImageQuant software (Molecular Dynamics) on autoradiographs that depicted bands within a linear range of exposure. Densitometric data were normalized to unstimulated control cells, which were set to 1.0.

Data and Statistical Analysis

Data in the text and figure legends are expressed as mean ± SEM of observations obtained from airway smooth muscle cells cultured from n patient donors. EC₅₀/IC₅₀ values and extrapolated maximum responses were estimated for individual concentration-response curves using nonlinear least-squares regression (SigmaPlot; Jandel Scientific, Erkrath, Germany) where appropriate. EC₅₀/IC₅₀ values were converted to negative logarithmic values (pD₂) for all statistical analysis, although for ease of comprehension EC₅₀/IC₅₀ values are given in the text. Data were compared using one-way analysis of variance (ANOVA) followed by Bonferroni's post hoc t test to determine statistical differences after multiple comparisons (SigmaStat; Jandel Scientific, Erkrath, Germany). A probability value of less than 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Release of Eosinophil-activating Cytokines and MAP Kinase Activation in Human Airway Smooth Muscle Cells Stimulated by IL-1

Cytokine release. Levels of GM-CSF, RANTES, and eotaxin were determined in the same supernatants by ELISA. In all cultures examined, the release of GM-CSF and eotaxin by unstimulated cells was less than could be detected by ELISA (< 2.8-10 pg/ml). Constitutive release of RANTES was observed at 24 h, but did not exceed 5 ng/ml/million cells (Figure 1a).

View larger version (17K): [in this window] [in a new window]   Figure 1.   IL-1 induces release of eosinophil-activating cytokine from human airway smooth muscle. Release of GM-CSF (circles), RANTES (squares), and eotaxin (triangles) into the culture medium was determined by ELISA in unstimulated cells (Unstim, closed symbols) or cells treated with IL-1 (open symbols). (a) Time course showing the relative kinetics of GM-CSF, RANTES, and eotaxin release from cells stimulated with a maximally effective concentration of IL-1 (1 ng/ml). (b) Relative efficacy and potency of IL-1 on induction of GM-CSF, RANTES, and eotaxin release at 24 h. Points represent mean ± SEM of duplicate values from independent experiments using cells cultured from four or five different donors, cell passages 2-6. Note that due to large differences in absolute levels of cytokines, GM-CSF levels are presented on a separate expanded scale.

Release of GM-CSF, RANTES, and eotaxin into culture supernatants from IL-1-stimulated (1 ng/ml) human airway smooth muscle cells was time dependent with significant increases readily detected at 6, 12, and 24 h (p < 0.05-0.001, n = 4) (Figure 1a). Culture supernatants from cells stimulated with varying concentrations of IL-1 were therefore collected 24 h after stimulation. Significant release of all three cytokines was observed with 0.01 ng/ml IL-1 (p < 0.001, n = 4) (Figure 1b). Near maximum release of each of the cytokines occurred at 1 ng/ml IL-1, and this concentration was used in all subsequent experiments. Concentrations of IL-1 inducing 50% of the maximum release (EC₅₀) of each cytokine were GM-CSF, 47 ± 5 pg/ml, RANTES, 40 ± 3 pg/ml, and eotaxin 22 ± 5 pg/ml. The rank order for maximal cytokine release stimulated by IL-1 from human airway smooth muscle cells was RANTES > eotaxin > > GM-CSF. Because large differences were present in the absolute levels of each cytokine measured, the data were normalized against the response to IL-1 alone, and this allowed direct comparison of the effects of chemical inhibitors on each of the cytokines.

p42/p44 ERK, p38, and SAPK/JNK MAP kinase activation. IL-1 was examined for its capacity to activate the p42/p44 ERK, p38, and SAPK/JNK MAP kinases. Immunoblot analysis showed that amounts of phosphorylated tyrosine of p42/p44 ERK in cells stimulated with 1 ng/ml IL-1 were increased at 5 min (p < 0.05, n = 3). Peak phosphotyrosine levels occurred at 15 min and were 18.9 ± 2.5-fold above unstimulated levels (p < 0.001, n = 3), remaining significantly elevated up to 4 h (p < 0.001, n = 3), after which levels returned to near baseline (Figure 2a). Activation of p42/p44 ERK was dependent on the concentration of IL-1 (EC₅₀ 4.9 ± 1 pg/ml) (Figure 2b). Significant tyrosine phosphorylation of p42/p44 ERK was observed at 0.001 ng/ml (p < 0.05, n = 3). Maximum p42/p44 ERK activation occurred with 1-10 ng/ml IL-1 and was 14.8 ± 0.6-fold (p < 0.001, n = 3), compared with unstimulated cells.

View larger version (51K): [in this window] [in a new window]   Figure 2.   IL-1 induces the threonine and tyrosine phosphorylation of p42/p44 ERK, p38 MAP kinase, and SPK/JNK. (a) Growth- arrested cells were incubated with medium alone (Unstim) or treated with IL-1 (1 ng/ml) for the times indicated. Proteins in cell lysates were separated by SDS- PAGE and immunoblotted using phospho-specific antibodies to p42/p44 ERK (upper pair of panels), p38 MAP kinase (middle pair of panels), and SAPK/ JNK (lower pair of panels). (b) Cells were stimulated with IL- (0.0001-100 ng/ml) for 15 min, lysed, and proteins were separated by SDS-PAGE prior to immunoblot detection of phosphorylated p42/p44 ERK (upper panel ), p38 MAP kinase (middle panel ), and SAPK/JNK (p46 and p54 isoforms, lower panel ). Lysates from unstimulated cells prior to (t0), and after, the 15 min period of stimulation (Unstim) were also included. Data in graphs are mean ± SEM of fold increases in phosphorylated p38 MAP kinase (open circles), p42/p44 ERK (open squares), or SAPK/JNK (open triangles) derived from densitometric quantification of autoradiographs using ImageQuant software, and are relative to unstimulated levels of MAP kinase activation in the absence of IL-1, which was set to 1.0. Because of cross-reactivity of phosphorylated p42/ p44 ERK with the phospho-specific JNK antibody, only the p54 isoform was quantitated. Autoradiographs and data in graphs are representative of independent experiments using cells cultured from three different donors, cell passages 3-4.

Similarly, amounts of phosphorylated threonine and tyrosine of p38 MAP kinase in IL-1-stimulated cells were significantly increased at 5 min (p < 0.05, n = 3). Peak phosphotyrosine levels (16.2 ± 1.7-fold; p < 0.001, n = 3) occurred at 15 min and remained significantly elevated up to 4 h (p < 0.05, n = 3), after which they returned to near baseline levels (Figure 2a). Phosphorylated threonine and tyrosine of p38 by IL-1 was also concentration dependent (EC₅₀ 9.7 ± 2.8 pg/ml) (Figure 2b). Significant p38 MAP kinase phosphorylation was observed at 0.001 ng/ml (p < 0.05, n = 3). Maximum p38 activation occurred with 1-10 ng/ml IL-1 and was 28.6 ± 6.5-fold (p < 0.001, n = 3), compared with unstimulated cells. Blots were stripped and reprobed with phosphorylation state-independent specific antibodies, and confirmed that equal amounts of p42/p44 ERK or p38 MAP kinase protein were blotted (data not shown, n = 2).

IL-1 also induced phosphorylation of threonine and tyrosine of SAPK/JNK, which was time (peak activation at 15 min and was 13.8 ± 3.2-fold; p < 0.01, n = 3) and concentration dependent (EC₅₀ 75 ± 8 pg/ml) (Figure 2). The duration of SAPK/JNK activation by IL-1 was shorter than for p42/p44 or p38 activation and had returned to baseline levels by 60 min. Maximum SAPK/JNK activation occurred with 1-10 ng/ml IL-1 and was 15.6 ± 0.6-fold (p < 0.001, n = 3), compared with unstimulated cells. Extended time courses up to 24 h of stimulation by IL-1 did not reveal significant (p > 0.05, n = 3) increases in the activation of p42/p44 ERK, p38, or SAPK/ JNK beyond 8 h.

Effect of p38 MAP Kinase Inhibitors on Cytokine Release and on p42/p44 ERK, p38 MAP Kinase, and SAPK/JNK Activity in Cells Stimulated by IL-1

To determine whether IL-1-stimulated release of GM-CSF, RANTES, or eotaxin was dependent upon the activation of p38 MAP kinase, SB 203580 and SB 202190, potent and specific inhibitors of p38 MAP kinase activity (18, 19), were examined. Preincubation for 30 min with SB 203580 (0.1 nM-100 µM) inhibited (p < 0.001, n = 6) IL-1-induced release of eotaxin to baseline levels (Figure 3a). The maximum inhibitory effect of SB 203580 occurred at >100 µM. The IC₅₀ value for this effect was 7.5 ± 2.7 µM. In the same culture supernatants, induction of RANTES was unaffected by SB 203580 (p > 0.05, n = 6). In two independent experiments, pretreatment of the cells with SB 203580 (0.1 nM-100 µM) selectively attenuated IL-1-induced increases in p38MAP kinase activity (IC₅₀ 2.7 µM; range 4.1 µM and 1.3 µM), as demonstrated by the reduced phosphorylation of its substrate, ATF-2 (Figure 3a, upper panels). No effect on IL-1-stimulated increases in the activity of p42/p44 ERK (i.e., Elk-1) or SAPK/JNK (i.e., c-Jun) was observed except at 100 µM, the highest concentration of SB 203580 investigated. SB 202474, an inactive structural analogue of SB 203580 (18), was examined against IL-1-stimulated cytokine production and p38 MAP kinase activity. This compound (0.1 nM-10 µM) had no effect (p > 0.05, n = 6) on IL-1-stimulated eotaxin levels. An apparent increase (< 20%) in IL-1-stimulated RANTES levels was detected (Figure 3b), but this did not achieve significance (p > 0.05, n = 6). Pretreatment of cells for 30 min with SB 202474 (0.1 nM-10 µM) also had no effect on IL-1-induced increases in the activity of p38 MAP kinase, ERK or SAPK/JNK (Figure 3b, upper panels).

View larger version (42K): [in this window] [in a new window]   Figure 3.   SB 203580 selectively inhibits IL-1-stimulated increases in p38 MAP kinase activity, and prevents release of eotaxin, but not RANTES. Growth-arrested cells were incubated with medium alone (Unstim) or treated with IL-1 (1 ng/ml for 15 min) in the presence or absence of increasing (log10) concentrations of (a) SB 203580, an inhibitor of p38 MAP kinase activity, or (b) SB 202474, a structurally related compound that does not inhibit p38 MAP kinase activity. MAP kinase activity was determined in cell lysates by selective immunoprecipitation (p42/p44 and p38) or chemical precipitation (SAPK/ JNK) followed by an in vitro kinase assay of an appropriate substrate prior to SDS-PAGE and immunoblot analysis. p42/p44 ERK activity (upper panels) was analyzed using a phospho-specific antibody to Elk-1 (p-Elk-1), p38 MAP kinase activity (middle panels) was analyzed using a phospho-specific antibody to ATF-2 (p-ATF-2), and SAPK/JNK activity (lower panels) was analyzed using a phospho-specific antibody to c-Jun (p-c-Jun). Autoradiographs are representative of two independent experiments, cell passage 3. Data in the graphs show the effects of (a) SB 203580, (b) SB 202474, and (c) SB 202190, a more potent p38 MAP kinase inhibitor, on RANTES (open squares) and eotaxin (open triangles) release from cells stimulated with IL-1 (1 ng/ml) for 24 h. Data are expressed as a percentage of the control response to IL-1 alone for either RANTES (filled bar) or eotaxin (hatched bar) and represent mean ± SEM of duplicate values from independent experiments using cells cultured from six different donors, cell passages 4-6. *p < 0.05, ***p < 0.001 compared with stimulated levels in the absence of inhibitor by Bonferroni's t test. NS p > 0.05 by ANOVA.

SB 202190 (0.1 nM-100 µM), a more potent inhibitor of p38 MAP kinase activity, was also examined against IL-1-stimulated cytokine release. In keeping with the effects of SB 203580, SB 202190 also inhibited (p < 0.001, n = 5) the release of eotaxin (Figure 3c). Inhibition of eotaxin release to baseline levels occurred at 10 µM and the IC₅₀ for this effect was 732 ± 94 nM. In the same culture supernatants, SB 202190 had no effect on the release of RANTES (p > 0.05, n = 5).

In contrast, in the same supernatants, levels of GM-CSF were found to be markedly increased by inhibitors of p38 MAP kinase activity (Figure 4). Maximum augmentation of GM-CSF levels occurred with SB 202190 at 1 µM and was more than 6-fold above the levels that were stimulated by IL-1 alone (p < 0.001, n = 5). A similar increase in IL-1-stimulated GM-CSF release occurred with SB 203580. A control compound, SB 202474, which does not inhibit p38 MAP kinase activity, also increased production of GM-CSF (p < 0.001, n = 6) by IL-1, but the extent of this increase was less than half of that seen with the active p38 MAP kinase inhibitors. Approximate EC₅₀ values for the enhancement of IL-1-stimulated GM-CSF release were 184 ± 32 nM (SB 202190), 1.8 ± 0.4 µM (SB 203580), and > 10 µM (SB 202474). None of the compounds had any effect on cytokine levels that were collected from unstimulated cells (data not shown).

View larger version (21K): [in this window] [in a new window]   Figure 4.   Inhibitors of p38 MAP kinase activity augment IL-1-stimulated GM-CSF release from human airway smooth muscle cells. Release of GM-CSF into the culture medium was determined by ELISA in cells incubated with medium alone (Unstim) or cells treated with IL-1 in the absence (open bar) or presence of increasing (log10) concentrations of SB 203580 (open circles), SB 202190 (open squares), or SB 202474 (open triangles). Data are expressed as a percentage of the control response to IL-1 alone and represent mean ± SEM of duplicate values from independent experiments using cells cultured from six different donors, cell passages 4-6. *p < 0.05, **p < 0.01, ***p < 0.001 compared with stimulated levels in the absence of inhibitor by Bonferroni's t test.

Effect of MEK Inhibitors on Cytokine Release and on p42/p44 ERK, p38 MAP Kinase, and SAPK/JNK Activity in Cells Stimulated by IL-1

PD 98059 or U 0126, specific noncompetitive inhibitors of the p42/p44 ERK activators, MEK1 and MEK2 (20, 21), were examined to determine if the release of GM-CSF, RANTES, or eotaxin induced by IL-1 was dependent upon ERK activation. Preincubation for 30 min with PD 98059 (0.1 nM-100 µM) had no effect (p > 0.05, n = 6) on the levels of RANTES or eotaxin present in the culture supernatants of cells that were stimulated with IL-1 (1 ng/ml) for 24 h (Figure 5a). In contrast, levels of GM-CSF in the same supernatants were found to be increased following incubation with either 0.01, 0.1, or 1 µM (p < 0.05-0.01, n = 6), but not with 10 or 100 µM PD 98059 (Figure 5a). Pretreatment of the cells with PD 98059 (0.1 nM-100 µM) selectively attenuated IL-1-induced increases in p42/p44 ERK activity (IC₅₀ > 15 µM) as demonstrated by the reduced phosphorylation of its substrate, Elk-1, and partially prevented (< 60%) phosphorylation of p42/p44 ERK (data not shown). However, although clear inhibition of Elk-1 phosphorylation was observed at 10 µM and 100 µM, complete inhibition to baseline levels did not occur even at these concentrations (Figure 5a, upper panel). In the same experiment, no effect of PD98059 was seen against IL-1-stimulated p38 MAP kinase or SAPK/JNK activity.

View larger version (47K): [in this window] [in a new window]   Figure 5.   U0126 selectively inhibits IL-1-stimulated increases in the kinase activity of p42/ p44 ERK, and prevents release of GM-CSF, RANTES, and eotaxin. Growth-arrested cells were incubated with medium alone (Unstim) or treated with IL-1 (1 ng/ml for 15 min) in the presence or absence of increasing (log10) concentrations of the MEK inhibitors (a) PD 98059 or (b) U 0126. MAP kinase activity was determined in cell lysates by selective immunoprecipitation (p42/p44 and p38) or chemical precipitation (SAP/ JNK) followed by an in vitro kinase assay of an appropriate substrate prior to SDS-PAGE and immunoblot analysis. p42/ p44 ERK activity (upper panels) was analyzed using a phospho-specific antibody to Elk-1 (p-Elk-1), p38 MAP kinase activity (middle panels) was analyzed using a phospho-specific antibody to ATF-2 (p-ATF-2), and SAPK/JNK activity (lower panels) was analyzed using a phospho-specific antibody to c-Jun (p-c-Jun). Autoradiographs are representative of two independent experiments, cell passage 3. Data in the graphs show the effects of (a) PD 98059 and (b) U 0126 on GM-CSF (open circles), RANTES (open squares), and eotaxin (open triangles) release from cells stimulated with IL-1 (1 ng/ml) for 24 h. Data are expressed as a percentage of the control response to IL-1 alone for either GM-CSF (open bar), RANTES (filled bar), or eotaxin (hatched bar) and represent mean ± SEM of duplicate values from independent experiments using cells cultured from six to eight different donors, cell passages 3-6. *p < 0.05, ***p < 0.001 compared with stimulated levels in the absence of inhibitor by Bonferroni's t test.

Preincubation for 30 min with U 0126 (0.1 nM-10 µM) inhibited IL-1-induced release of GM-CSF, RANTES, and eotaxin by 60-70% (p < 0.001, n = 6) (Figure 5b). IC₅₀ values were 1.9 ± 0.7 µM for GM-CSF, 5.0 ± 1.1 µM for RANTES, and 1.8 ± 0.5 µM for eotaxin. In two separate experiments, pretreatment of the cells with U 0126 (0.1 nM-10 µM) inhibited IL-1-induced increases in p42/p44 ERK activity (IC₅₀ 186 nM, range 247 nM and 125 nM) to baseline levels, as indicated by the reduced Elk-1 phosphorylation (Figure 5b, upper panel). Partial inhibition (approximately 20%) of p38 MAP kinase activity, indicated by reduced ATF-2 phosphorylation, was detected at 10 µM U 0126. No inhibition against IL-1-stimulated SAPK/JNK activity was observed. Neither PD 98059 nor U 0126 had any effect on cytokine levels in supernatants collected from unstimulated cells (data not shown).

Effect of Delayed Addition of p38 MAP Kinase and MEK Inhibitors on IL-1-stimulated Cytokine Release

To investigate the duration of IL-1-stimulated p38 MAP kinase and p42/p44 ERK activation that was necessary for induction of eosinophil-activating cytokine release, pharmacological inhibitors were added to cell cultures at varying times before or following stimulation with IL-1 (1 ng/ml, 24 h). Pretreatment with either SB 203580 or U 0126 at 10 µM was chosen as this concentration of the inhibitors was found to attenuate cytokine levels in the conditioned media by approximately 50%-60% (Figures 3a and 5b). Consistent with data in Figure 3a, the p38 MAP kinase inhibitor, SB 203580, inhibited the release of eotaxin, and had no effect on RANTES levels (Figure 6a). SB 203580 (10 µM) prevented eotaxin release even when added up to 2 h after stimulation (p < 0.001, n = 3). Beyond 2 h the extent of the attenuation diminished, but remained significant up to 8 h (p < 0.001, n = 3). In parallel experiments, delayed addition of the inactive structural analogue SB 202474 (10 µM) had no effect (p > 0.05, n = 3) on IL-1-stimulated eotaxin production (data not shown). In the same culture supernatants, significant enhancement (p < 0.001, n = 3) of GM-CSF levels was observed with a 30-min pretreatment of SB 203580 and to a lesser extent (p < 0.05, n = 3) with the control compound SB 202474 (Figure 6b). The magnitude of this enhancement was greater with SB 203580 compared with SB 202474 (p < 0.01, n = 3), but this difference in the ability of both compounds to enhance GM-CSF release was present only when they were added up to 1 h after stimulation. Thereafter, differences in magnitude were not maintained (Figure 6b). GM-CSF levels were significantly elevated (p < 0.01 compared with IL-1 alone, n = 3) even when SB 203580 was added up to 2 h after stimulation.

View larger version (29K): [in this window] [in a new window]   Figure 6.   Effect of delayed addition of SB 203580 on IL-1-stimulated cytokine release from human airway smooth muscle cells. Release of (a) RANTES (filled bars) and eotaxin (hatched bars) into the culture medium was determined by ELISA in cells stimulated with IL-1 (1 ng/ml) for 24 h. SB 203580 (10 µM) was added 30 min prior to (30 min), simultaneously with (t0), or at stated times after cell stimulation. Attenuation of eotaxin release by SB 203580 was significant (p < 0.05- 0.001, compared with stimulated levels in the absence of inhibitor, n = 3 by Bonferroni's t test) up to 8 h after stimulation. In (b) the effect of SB 203580 (filled bars) was compared with SB 202474 (open bars) on IL-1-stimulated GM-CSF release. GM-CSF levels were significantly increased by addition of SB 203580 (p < 0.05-0.001, compared with stimulated levels in the absence of inhibitor) up to 2 h after stimulation, but not at 4 and 8 h. In each case, data are expressed as a percentage of the control response to IL-1 alone and represent mean ± SEM of duplicate values from independent experiments using cells cultured from three different donors, cell passages 3-5. Statistical annotations are omitted for clarity.

Similar experiments were performed with the MEK inhibitor, U 0126. Attenuation of IL-1-stimulated GM-CSF, RANTES, and eotaxin production by U 0126 remained significant (p < 0.01-0.001 compared with levels in the absence of inhibitor for each cytokine, n = 3) despite addition of the inhibitor up to 4 h after stimulation (Figure 7). When added 8 h after stimulation, its capacity to prevent cytokine release was much reduced (p > 0.05 compared with levels in the absence of inhibitor for each cytokine, n = 3).

View larger version (33K): [in this window] [in a new window]   Figure 7.   Effect of delayed addition of U 0126 on IL-1-stimulated cytokine release from human airway smooth muscle cells. Release of GM-CSF (open bars), RANTES (filled bars), and eotaxin (hatched bars) into the culture medium was determined by ELISA in cells stimulated with IL-1 (1 ng/ml) for 24 h. U 0126 (10 µM) was added 30 min prior to (30 min), simultaneously with (t0), or at stated times after cell stimulation. In each case, data are expressed as a percentage of the control response to IL-1 alone and represent mean ± SEM of duplicate values from independent experiments using cells cultured from three different donors, cell passages 3-5. Attenuation of cytokine release by U 0126 was significant (p < 0.05-0.001, compared with stimulated levels in the absence of U 0126 for each cytokine, n = 3 by Bonferroni's t test) when added up to 4 h after stimulation. Statistical annotations are omitted for clarity.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, we examined the possible role of two signal transduction pathways for the release of multiple cytokines following stimulation of human airway smooth muscle cells by IL-1. Our results demonstrate that in human airway smooth muscle cells, simultaneous release of biologically relevant quantities of several eosinophil-activating cytokines, including GM-CSF, RANTES, and eotaxin, induced by IL-1, was associated with threonine and tyrosine phosphorylation of p42/p44 ERK, p38 MAP kinase, and SAPK/JNK, as well as increased kinase activity of p42/p44 ERK, p38, and SAPK/JNK. Inhibition of IL-1-stimulated eotaxin release by either SB 203580 or U 0126 was observed, whereas inhibition of RANTES release occurred only in the presence of U 0126. These data indicate that eotaxin release is regulated by pathways that involve both p38 MAP kinase-dependent and p42/p44 ERK-dependent mechanisms; the release of RANTES is dependent on p42/p44 ERK, but not p38 MAP kinase activity. In addition, we provide evidence that release of GM-CSF by IL-1 is dependent on p42/p44 ERK activation, and is also tonically suppressed by a mechanism that is partially p38 MAP kinase dependent. These data are consistent with other recent reports that IL-1 stimulates MAP kinase activation in human airway smooth muscle (22). Here we explore the significance of these intracellular signaling events for the induction of cytokine release from human airway smooth muscle cells.

Cell-permeable chemical inhibitors of protein kinases provide one approach for investigating the biological functions of intracellular signaling events in cells and tissues. In this study, to elucidate the role of p38 MAP kinase-dependent pathways in IL-1-induced GM-CSF, RANTES, and eotaxin release from human cultured airway smooth muscle cells, SB 203580 and its more potent structural analogue, SB 202190, were used as specific inhibitors of p38 MAP kinase activity. Both compounds caused complete inhibition of IL1-stimulated eotaxin release, but had no effect on the release of RANTES in the same cell-conditioned medium samples. SB 203580 was also shown to prevent IL-1-stimulated increases in p38 MAP kinase activity, indicated by inhibition of ATF-2 threonine phosphorylation, without decreasing the kinase activity of IL-1-stimulated p42/ p44 ERK or SAPK/JNK, determined by Elk-1 or c-Jun serine phosphorylation, respectively. Furthermore, SB 202474, a control compound that does not inhibit p38 MAP kinase activity (18), had no inhibitory effect on IL-1-stimulated eotaxin or RANTES release, and was also without effect on the kinase activity of IL-1-stimulated p42/p44 ERK, p38, or SAPK/JNK. Taken together, these data suggest that induction of p38 MAP kinase activity by IL-1 in human airway smooth muscle cells is a necessary event in the release of eotaxin, but not RANTES. This is the first report implicating p38 MAP kinase activity in the release of eotaxin by airway smooth muscle. Our observation that induction of RANTES release by IL-1 was not sensitive to inhibition of p38 MAP kinase activity agrees with similar earlier findings in human airway smooth muscle cells by Maruoka and colleagues (25), who showed that TNF--stimulated RANTES release was not prevented by SB 203580.

When GM-CSF release was examined in the same cell-conditioned samples, the active inhibitors of p38 MAP kinase activity, SB 203580 and SB 202190, both enhanced its release. SB 202190 was approximately 7.5-fold more potent than SB 203580, consistent with its reported greater potency against p38 MAP kinase in other cell systems, compared with SB 203580 (18, 26). However, the inactive compound, SB 202474, was also found to enhance GM-CSF release, albeit this effect occurred at concentrations that were more than 5- to 50-fold greater than with the active pyrinylimidazole compounds. Nevertheless, enhancement of GM-CSF release was observed with the inactive compound, suggesting that a component of this response was unrelated to inhibition of p38 MAP kinase activity, and reflected another action of the compounds. The observed increase in GM-CSF release by IL-1 in the presence of the active p38 MAP kinase inhibitors was an unexpected finding, particularly as inhibition of eotaxin in the same samples had been observed. Similar enhancement of GM-CSF production by SB 203580 has been reported in human cultured mast cells, and was attributed to potentiation of JNK activation by this compound (27). In the present study, however, this seems an unlikely mechanism in accounting for enhanced GM-CSF release, as potentiation of IL-1-stimulated SAPK/ JNK activity was not observed with the inhibitor.

Another possibility is that SB 203580 and SB 202190 induced endogenous IL-1 expression, though a recent study, also in human airway smooth muscle, has reported that this does not occur with SB 203580 (24), and the concentration of IL-1 used in the present study was maximally effective. More likely, it is possible that IL-1-stimulated GM-CSF release is tonically suppressed by a mechanism that is dependent on p38 MAP kinase activity. Such a mechanism might also involve endogenous cyclooxygenase (COX) metabolites such as prostaglandlin E₂ (PGE₂) (28, 29), as IL-1, in addition to inducing cytokine release, also promotes COX-1 activity and expression of the COX-2 isoform in human airway smooth muscle cells (28, 29). PGE₂ promotes increases in intracellular cyclic AMP in these cells (29), and our own recent findings have shown that agents that increase intracellular cyclic AMP levels also inhibit the release of IL-1-stimulated GM-CSF (30). We also noted that indomethacin, a mixed COX-1/COX-2 inhibitor, selectively enhanced the production of GM-CSF, compared with RANTES or eotaxin (30). Lazzeri and colleagues have also reported enhanced GM-CSF production from human airway smooth muscle cells in the presence of indomethacin and other mixed COX-1 and COX-2 inhibitors (31). Thus, in addition to inducing release of GM-CSF, IL-1 may tonically suppress GM-CSF release by a p38 MAP kinase-dependent pathway that is also dependent on the activity of COX, consistent with reports that p38 MAP kinase activation is required for induction of COX-2 mRNA and protein expression (32).

Finally, SB 203580 and its related pyrinylimidazole derivatives may directly inhibit COX-1 and COX-2 activity in human airway smooth muscle cells. Indeed, SB 203580 was originally developed from drugs that are inhibitors of COX and reversibly inhibits PGE₂ formation by purified COX-1 and COX-2 with IC₅₀ values of 2 µM (35). This compares very closely with an IC₅₀ of 1.8 µM in the present study for enhancement of GM-CSF production by SB 203580. Thus, the observed increase in IL-1-stimulated GM-CSF release induced by these pyrinylimidazole derivatives should not be interpreted as the sole result of inhibition of p38 MAP kinase activity, preventing downstream activation or induction of COX activity, because at this stage a component of mixed inhibition that involves direct suppression of COX activity by these compounds cannot be excluded.

To investigate the requirement for p42/p44 ERK activation in IL-1-induced GM-CSF, RANTES, and eotaxin release, PD 98059 and U 0126, structurally unrelated, specific noncompetitive inhibitors of the p42/p44 ERK activators, MEK1 and MEK2, were examined. PD 98059 was ineffective against IL-1-stimulated RANTES and eotaxin release, although partially inhibiting IL-1-induced increases in p42/p44 ERK activity as demonstrated by its capacity to reduce phosphorylation of Elk-1 and to prevent phosphorylation of p42/p44. However, this was observed only at the higher concentrations (> 10 µM-100 µM) investigated, and even at these concentrations complete inhibition to baseline levels could not be demonstrated. Similar observations and concerns about the efficacy of this compound in human airway smooth muscle cells have been expressed by others (22, 23), where U 0126, another MEK inhibitor, was found to be more effective than PD 98059 in preventing mitogen-stimulated phosphorylation of p42/p44 ERK. This discrepancy has been attributed to differences in binding affinities of approximately 100-fold for MEK enzymes (21) between the two compounds, and to limitations of solubility of PD 98059, which serve to reduce the effective concentration of PD 98059 in intact cells. PD 98059 did not prevent release of RANTES or eotaxin induced by IL-1, but did enhance GM-CSF release. The underlying mechanism for this latter effect was not explored but is unlikely to be due to inhibition of p38 MAP kinase activity (discussed earlier), as PD 98059 did not prevent IL-1-induced p38 MAP kinase activity. However, like SB 203580, PD 98059 is reported to inhibit COX activity in platelets (35) at concentrations similar to those that enhanced IL-1 GM-CSF release (i.e., 10 nM-1 µM). Furthermore, this occurred under conditions that had little effect against increases in IL-1-stimulated p42/p44 ERK activity, as relatively high concentrations of PD 98059 (20-50 µM) are required to completely block the activation of p42/p44 ERK in intact cells (20). Thus, in human airway smooth muscle, the only effect of PD 98059 was to enhance GM-CSF release, probably as a result of the removal of a COX-dependent component that would otherwise tonically suppress IL-1-stimulated GM-CSF release (see earlier discussion and [30, 31]).

Thus, any initial interpretation of our PD 98059 data that ERK-dependent mechanisms are not important in IL-1-induced cytokine release should be considered in the context of our findings with the second MEK inhibitor, U 0126. This compound was found to completely and selectively inhibit IL-1-induced increases in p42/p44 ERK activity, and prevented IL-1-stimulated GM-CSF, RANTES, and eotaxin release by around 60-70%, perhaps indicating a partial inhibition. These data suggest that ERK-dependent mechanisms are required, but may not be sufficient for the regulation of cytokine production by IL-1 in human airway smooth muscle cells, and are consistent with other reports describing p42/p44 ERK-mediated cytokine expression in nonmuscle cells (36, 37). The data are also in general agreement with the report by Maruoka and colleagues (25) showing that platelet-activating factor, but not TNF--mediated RANTES production, was dependent on p42/ p44 ERK activation. Discrepancies between this report and our data may reflect a stimulus-dependent action of ERK or the interpretation of findings based on the use a single, poorly efficacious MEK inhibitor.

Examination of the time course of MAP kinase activation following exposure of human airway smooth muscle cells to IL-1 showed a rapid phosphorylation of p42/p44 ERK and p38 MAP kinase that peaked at 15 min, consistent with previous reports (22, 23), and remained elevated above basal levels at least to 4-8 h after stimulation. Whether these later increases in phosphorylation of MAP kinases were due directly to the persistence of IL-1 in the culture medium or occurred as a result of secondary events, including the autocrine induction of endogenous IL-1, was not investigated. Nevertheless, to determine any requirement for sustained activation of p42/ p44 ERK or p38 MAP kinase in the induction of eosinophil-activating cytokine release, inhibitors were added to the cell cultures at varying times after stimulation by IL-1. The effectiveness of U 0126 in attenuating cytokine release persisted even when the inhibitor was added up to 4 h after IL-1. Similarly, the capacity of SB 203580 to prevent eotaxin or to augment GM-CSF release remained up to 4 h after stimulation with IL-1. Together, these data suggest that late activation of both p42/p44 ERK and p38 MAP kinase is important for the magnitude of IL-1-stimulated cytokine release. Alternatively, a minimum level of sustained MAP kinase (p38 and p42/p44 ERK) activation, up to 4 h, is required throughout the period of cytokine gene activation, protein translation, and release. This is consistent with studies that suggest that cooperation between ERK and p38 MAP kinase pathways is necessary for optimal induction of cytokine gene expression (38, 39) and cytokine release (25). Although IL-1 induced a similar rapid and concentration-dependent phosphorylation of SAPK/JNK, approximately 10-fold higher concentrations of IL-1 were required and its activation was not sustained, returning to basal levels within 30-60 min. Further studies using dominant negative mutants or specific chemical inhibitors of JNK are required to determine the significance of its activation in this system.

In summary, our results indicate that IL-1 activates p42/p44 ERK-, p38 MAP kinase-, and SAPK/JNK-dependent processes in human airway smooth muscle. In addition, we provide evidence that implicates both p38 MAP kinase- and p42/p44 ERK-dependent pathways in the induction of eosinophil-activating cytokines by IL-1 from human airway smooth muscle cells. We conclude that the release of RANTES is dependent on activation of p42/p44 ERK but occurs independently of the activity of p38 MAP kinase. Eotaxin release, however, is dependent on both p38 MAP kinase- and p42/p44 ERK-dependent mechanisms. In addition, the release of GM-CSF was negatively coupled to p42/p44 ERK activity, and appeared to be tonically suppressed by a p38 MAP kinase-dependent pathway, though direct inhibition of COX activity due to poor inhibitor selectivity may also contribute to this component of tonic suppression. Thus, the release of eosinophil-activating cytokines from human airway smooth muscle cells is regulated by multiple MAP kinase-dependent signaling pathways. Clearly, the intracellular transduction pathways in inflammation are complex, and prediction of their effects cannot easily be based on the measurement of release of a single cytokine. Understanding the cellular mechanisms and intracellular pathways that modulate the release of cytokines by airway smooth muscle may be important for the development of future therapeutic interventions that are targeted at the airway smooth muscle remodeling process in the airway wall of the diseased lung.