We have previously reported that interleukin (IL)-1β causes β-adrenergic hyporesponsiveness in cultured human airway smooth muscle (HASM) cells by increasing cyclooxygenase (COX)-2 expression. The purpose of this study was to determine whether p38 mitogen-activated protein (MAP) kinase is involved in these events. IL-1β (2 ng/ml for 15 min) increased p38 phosphorylation fourfold. The p38 inhibitor SB-203580 (3 μM) decreased IL-1β-induced COX-2 by 70 ± 7% (P < 0.01). SB-203580 had no effect on PGE2 release in control cells but caused a significant (70–80%) reduction in PGE2 release in IL-1β-treated cells. IL-1β increased the binding of nuclear proteins to the oligonucleotides encoding the consensus sequences for activator protein (AP)-1 and nuclear factor (NF)-κB, but SB-203580 did not affect this binding, suggesting that the mechanism of action of p38 was not through AP-1 or NF-κB activation. The NF-κB inhibitor MG-132 did not alter IL-1β-induced COX-2 expression, indicating that NF-κB activation is not required for IL-1β-induced COX-2 expression in HASM cells. IL-1β attenuated isoproterenol-induced decreases in HASM stiffness as measured by magnetic twisting cytometry, and SB-203580 abolished this effect. These results are consistent with the hypothesis that p38 is involved in the signal transduction pathway through which IL-1β induces COX-2 expression, PGE2 release, and β-adrenergic hyporesponsiveness.
- mitogen-activated protein
- human airway smooth muscle
- nuclear factor-κB
- activator protein-1
- prostaglandin E2
- β-adrenergic responsiveness
- cytoskeletal mechanics
- magnetic twisting cytometry
the cyclooxygenase(COX) enzymes COX-1 and COX-2 convert arachidonic acid (AA) to prostaglandins (PGs) and thromboxane. COX-1 is constitutively expressed in most tissues, whereas COX-2 is induced by mitogens (16), bacterial lipopolysaccharide (15), and cytokines (4, 31). Several recent reports (21, 22,31) indicated that interleukin (IL)-1β induces COX-2 expression and markedly increases PGE2 release in cultured human airway smooth muscle (HASM) cells. The induction of COX-2 by IL-1β in these cells has important functional consequences because IL-1β decreases the responsiveness of HASM cells to β-agonists, and this effect is ablated by COX-2 inhibitors and mimicked by exogenous PGE2 (21, 22, 37). The mechanistic basis for this effect of COX-2 has not been established, but our results are consistent with the hypothesis that the marked increase in PGE2 that results from the induction of COX-2 by IL-1β leads to increased cAMP formation, phosphorylation of the β-adrenergic receptor by protein kinase A (PKA), and consequent desensitization of the receptor (22, 37). The signal transduction pathway leading from IL-1β to the induction of COX-2 in HASM cells has not been established.
In mammalian cells, at least three different subfamilies of mitogen-activated protein (MAP) kinases have been described. They include the extracellular signal-regulated kinases (ERKs), c-Jun amino-terminal kinase (JNK), and p38. These kinases are activated by distinct upstream MAP kinase kinases (MEKs) that recognize and phosphorylate threonine and tyrosine residues within a tripeptide motif (Thr-X-Tyr) on MAP kinases that is required for their activation. Once activated, MAP kinases, in turn, phosphorylate a variety of intracellular substrates including certain transcription factors (18). IL-1β is known to activate all three MAP kinase subfamilies in HASM cells (27), and Laporte et al. (21) have previously reported that activation of the ERK MAP kinases is required for the IL-1β induction of COX-2 and consequent β-adrenergic hyporesponsiveness in HASM cells.
The purpose of this study was to determine whether p38 activation is also required for IL-1β-induced COX-2 expression, increased PGE2 release, and β-adrenergic hyporesponsiveness in HASM cells. To do so, we examined the effect of the highly specific inhibitors of p38 SB-203580 and SB-202190 (6) on COX-2 expression and PGE2 release induced by IL-1β. We also examined the effect of SB-203580 on the IL-1β-induced changes in HASM cell responses to the β-agonist isoproterenol (Iso). Because our results provided evidence of an important role for p38 in these events, we sought to determine the role of p38 in the signal transduction pathway leading from IL-1β to the induction of COX-2. Moore et al. (26) have previously reported that IL-1β increases activator protein (AP)-1 and nuclear factor (NF)-κB DNA-binding activity in HASM cells, consistent with its effects in other cell types (32). Because the promoter region of the humanCOX-2 gene contains NF-κB- and AP-1-like consensus sequences (1) and because p38 has been reported to be capable of activating both these transcription factors (35,44), we examined the effect of IL-1β and SB-203580 on the binding of nuclear proteins to the oligonucleotides encoding the consensus sequences for AP-1 and NF-κB using electrophoretic mobility shift assay (EMSA). We also examined the effect of NF-κB inhibitors on COX-2 expression induced by IL-1β.
Human tracheae were obtained from lung transplant donors in accordance with procedures approved by the University of Pennsylvania (Philadelphia, PA) Committee on Studies Involving Human Beings. Tracheal smooth muscle cells were isolated and placed in culture as previously described (14, 21, 22, 30, 37). The cells were cultured on plastic in Ham's F-12 medium supplemented with 10% fetal bovine serum, 102 U/ml of penicillin, 100 μg/ml of streptomycin, 200 μg/ml of amphotericin B, 12 mM NaOH, 1.7 μM CaCl2, 2 mM l-glutamine, and 25 mM HEPES. The medium was replaced every 3–4 days. The cells were passaged with 0.25% trypsin and 1 mM EDTA every 10–14 days. Cells from the nine different donors studied were used in passages 4–7 in the studies described in Experimental protocol. In all cases, the cells were grown to confluence, and 24 h before use were serum deprived and hormone supplemented with 5.7 μg/ml of insulin and 5 μg/ml of transferrin because these conditions maximize expression of smooth muscle-specific contractile proteins (30).
To demonstrate p38 activation by IL-1β, we measured p38 phosphorylation in whole cell lysates using Western blotting. HASM cells from the same passage of the same donor cells were treated with IL-1β (2 ng/ml) for 0, 5, 10, 15, or 30 min. The medium was removed, and the cells were washed with PBS and then lysed in 400 μl of extraction buffer [10 mM Tris · HCl buffer with 50 mM NaCl, 50 mM NaF, 10 mM d-serine, 1 mM EDTA, 1 mM EGTA, 1% sodium dodecyl sulfate, 1% Triton X-100, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), 5 μg/ml of leupeptin, 1 μg/ml of pepstatin, and 10−2 U/ml of aprotinin]. The cells were scraped off the flasks, passed through a 25 -gauge needle, and solubilized by sonication. Western blot analysis with an antibody to phosphorylated p38 (New England Biolabs, Beverly, MA) was performed as described, with details from the manufacturer's protocol. To determine the specificity of the p38 inhibitor, we measured IL-1β-induced phosphorylation of p42/p44 and JNK proteins in cells treated with SB-203580 (3 μM) by Western blot analysis using antibodies to phosphorylated p42/p44 and JNK (New England Biolabs).
In other experiments, we used Western blotting to assess the effect of the p38 inhibitor SB-203580 on the expression of COX-2 induced by IL-1β. For these experiments, four flasks of HASM cells from the same passage of the same donor cells were used. Two were treated with SB-203580 (3 or 30 μM). Two hours later, IL-1β (2 ng/ml) was added to both flasks and to a third flask. The fourth flask served as a control. Approximately 22 h later, the cells were washed with PBS, and the proteins were isolated as described above. Western blotting for COX-2 was performed as previously described (21). Similar experiments were conducted with another p38 inhibitor, SB-202190. A similar protocol was used to examine the effect of the NF-κB inhibitors MG-132 and pyrrolidine dithiocarbamate (PDTC) on COX-2 expression induced by IL-1β (24, 29).
We also examined the effect of SB-203580 on IL-1β-induced PGE2 release. For these experiments, HASM cells in 24-well plates were treated with SB-203580 (3 μM) or were left untreated. Two hours later, IL-1β (2 ng/ml) was added to some wells of cells treated with the inhibitor and also to the untreated wells. Eighteen hours later, the medium was replaced with 0.5 ml of fresh serum-free hormone-supplemented medium, at which point the cells were left untreated or AA (10−5 M) or bradykinin (BK; 10−6 M) was added. After a 15-min incubation at 37°C, the supernatants were harvested and stored at −20oC until a subsequent assay with a PGE2 enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI). The antibody to PGE2 had <1% cross-reactivity to 6-keto-PGF1α and <0.01% to thromboxane B2 and other PGs according to the manufacturer's specifications.
To determine whether the role of p38 on the IL-1β effects on COX-2 expression might be through the effects on AP-1 or NF-κB activation, we examined the effect of IL-1β and SB-203580 on the binding of nuclear proteins to the oligonucleotides encoding the consensus sequences for AP-1 and NF-κB with EMSA. For these experiments, HASM cells were treated with IL-1β (20 ng/ml for 2 h) either alone or in combination with SB-203580 (30 or 3 μM for 4 h) or were left untreated (control). The nuclear extracts were harvested, and EMSA performed as described in Nuclear protein extracts and EMSA for NF-κB and AP-1. We also examined the effect of the ERK inhibitor U-0126 (10 μM) in these assays.
For the experiments in which we examined the effect of SB-203580 on IL-1β-induced changes in the cell stiffness responses to Iso or dibutyryl cAMP (DBcAMP), four flasks of cells from the same passage of the same donor were used. Two flasks were treated with SB-203580 (3 μM). Two hours later, IL-1β (2 ng/ml) was added to one of the flasks treated with SB-203580 and also to an untreated flask. Eighteen hours later, the cells were harvested by a brief exposure to trypsin and EDTA, resuspended in serum-free medium with and without IL-1β and SB-203580, and plated at 20,000 cells/well on collagen I (500 ng/cm2)-coated bacteriological plastic dishes (6.4-mm 96-well Removawells, Immunlon II). Two to six hours later, measurements of cell stiffness were made with magnetic twisting cytometry. Cumulative concentration-response curves to Iso or DBcAMP were performed as follows. First, three to five measurements of cell stiffness were made under baseline conditions. After these measurements, 2 μl of a solution containing Iso or DBcAMP were added to the well that contained 200 μl of medium. After a 1-min incubation with the agent, two to four measurements of cell stiffness were again obtained. This procedure was repeated with increasing concentrations of the agent. The concentration ranges used were 10−8 to 10−5 M Iso and 10−4 to 3 × 10−3 M DBcAMP. Only one agonist was studied per well. Details of the methodology for magnetic twisting cytometry are found inMagnetic twisting cytometry.
Nuclear protein extracts and EMSA for NF-κB and AP-1.
Nuclear extraction was performed with standard methods (20,25). Briefly, confluent HASM cells were harvested by scraping and centrifugation (3,000 rpm for 5 min) at 4°C in PBS containing a protease inhibitor cocktail (1 μg/ml of aprotinin, 1 μg/ml of leupeptin, 10 μg/ml of soybean trypsin inhibitor, and 1 μg/ml of pepstatin). The pellet was washed twice with 1 ml ice-cold buffer A [10 mM HEPES, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol (DTT), and 0.5 mM PMSF] and recentrifuged as described above. The supernatant was removed, and the nuclei were isolated by treating the pellet for 5 min on ice with 60 μl ofbuffer A that also contained 0.1% Nonidet P-40 followed by centrifugation at 14,000 rpm for 10 min at 4°C. The crude pellet was resuspended in 10 μl of buffer containing HEPES (20 mM, pH 7.9), MgCl2 (1.5 mM), NaCl (0.42 M), EDTA (0.2 mM, pH 8.0), glycerol (25% vol/vol), DTT (0.5 mM), and PMSF (0.5 mM) for 15 min on ice and clarified by centrifugation at 14,000 rpm for 10 min at 4°C. This supernatant containing the nuclear protein extract was subsequently diluted with 10 μl of buffer containing HEPES (20 mM, pH 7.9), KCl (50 mM), glycerol (20% vol/vol), DTT (0.5 mM), and PMSF (0.5 mM). Protein concentrations were determined by the bicinchoninic acid system. The nuclear extracts were stored at −70°C.
A double-stranded oligonucleotide probe containing the NF-κB or AP-1 consensus sequence (Gel Shift Assay System, Promega, Madison, WI) was end labeled with [γ-32P]ATP (3,000 Ci/mmol at 10 mCi/ml; New England Nuclear, Boston, MA) with T4 kinase. A total of 4 μg of nuclear extract were incubated with binding Tris · HCl buffer (50 mM) containing NaCl (250 mM), EDTA (2.5 mM), glycerol (25%), and poly(dI-dC) (25 mg) for 10 min at room temperature and then incubated with radiolabeled probe for another 20 min at room temperature. The specificity of the DNA-binding protein for the putative binding site was established by competitive experiments with an oligonucleotide containing a binding site for the protein of interest. DNA-protein complexes were resolved by electrophoresis on a 6% nondenaturing polyacrylamide gel (Novex, San Diego, CA) at 65 V. The gels were then dried and exposed to X-ray films by autoradiography, which were quantified by densitometer.
Transfection of HASM cells.
HASM cells were grown in complete medium for 72 h (60–80% confluence) in six-well tissue culture plates. Before transfection, the medium was changed to 1% fetal bovine serum. HASM cells were transfected with 0.5 μg of pNF-κB-Luc, designed for monitoring the NF-κB signal transduction pathway (Clontech, Palo Alto, CA), and 0.5 μg of a β-galactosidase control vector to normalize for differences in transfection efficiency from well to well. The NF-κB and β-galactosidase vectors were cotransfected with Fugene 6 (Roche, Indianapolis, IN) according to the manufacturer's protocol. Twenty-four hours posttransfection, the medium was replaced with serum-free medium containing insulin and transferrin as described inCell culture. The cells were then incubated with 25 μM MG-132 or 10 μM PDTC for 17 h and either 0 or 2 ng/ml of IL-1β for 15 h. The cells were lysed with reporter lysis buffer (Promega, Madison, WI) and harvested. The samples were assayed for luciferase activity by scintillation counting and for β-galactosidase activity by spectrophotometry with the β-galactosidase enzyme assay system (Promega). The results of the experiments are reported as mean luciferase activity normalized to β-galactosidase activity. With this system, the transfection efficiency averaged 16.4 ± 2.8% (n = 5 experiments) as assessed by flow cytometry of cells transfected with a green fluorescence protein-expressing vector.
Magnetic twisting cytometry.
Details of the magnetic twisting cytometry technique have been previously described (14, 21, 22, 37, 40, 41). Briefly, ferromagnetic beads (∼4 μm in diameter) were coated Arg-Gly-Asp-containing peptides (Peptite 2000, Telios Pharmaceuticals, San Diego, CA) and allowed to bind to adherent HASM cells through integrins that recognize the Arg-Gly-Asp sequence. Individual wells containing bead-coated cells were placed inside the magnetic twisting cytometer and maintained at 37°C with a circulating water bath that is built into the system. The beads were first magnetized with a brief 1,000-G pulse oriented parallel to the surface on which the cells were plated. The magnetic field vector generated by the beads in this direction was measured by an in-line magnetometer. Subsequently, a much smaller magnetic field was applied orthogonally to the first, causing the beads to rotate to align their magnetic moments with this field. This applied torque, or twisting stress (80 dyn/cm2 in this case), is opposed by reaction forces developed within the cytoskeleton to which the beads are bound through the integrin molecules. Magnetic twisting cytometry uses the applied twisting stress and resulting angular rotation of the magnetic beads and expresses the ratio as cell stiffness. Bead rotation increases with the strength of the applied twisting field and is inversely proportional to the resistance of the cell to shape distortion.
The tissue culture reagents and drugs used in this study were from Sigma (St. Louis, MO) with the exception of amphotericin B and the trypsin-EDTA solution that were from GIBCO BRL (Life Technologies, Grand Island, NY), IL-1β that was from Genzyme (Cambridge, MA), SB-203580 that was from Calbiochem-Novabiochem (La Jolla, CA), and U-0126 that was a gift from Dupont Pharmaceuticals (Wilmington, DE). SB-203580, SB-202190, U-0126, and MG-132 were dissolved in DMSO at concentrations such that the concentration of DMSO in the cell wells never exceeded 0.1%. PDTC was dissolved in water; and DBcAMP was dissolved at 10−1 M in distilled water, frozen in aliquots, and diluted appropriately in medium on the day of use. Iso (10−1 M in distilled water) was made fresh each day. Because Iso is rapidly oxidized, dilutions of Iso in medium were made immediately before addition to the cells.
Changes in basal and BK- and AA-stimulated PGE2 release induced by IL-1β and SB-203580 were examined by ANOVA, with treatment and experimental day as the main effects. The effect of SB-203580 on the IL-1β-induced changes in cell stiffness responses to Iso was examined by repeated-measures ANOVA, with treatment (control, IL-1β, SB-203580, and IL-1β plus SB-203580) and experimental day as the main effects. Follow-up t-tests were used to determine where the treatment effect lay. Effects of p38 inhibitor (SB-203580 or SB-202190) and NF-κB inhibitor (MG-132 or PDTC) treatment on increased COX-2 expression induced by IL-1β and effects of SB-203580 and U-0126 on the percent changes in IL-1β-induced AP-1 and NF-κB DNA-binding activity were examined by paired t-tests of optical densitometry measurements. A P value <0.05 was considered significant.
Activity of p38 in whole HASM cell lysates was estimated by determining the level of phosphorylated p38 with Western blotting (Fig. 1 A). There was some p38 phosphorylation even in untreated cells, but p38 phosphorylation began to increase within 5 min after the addition of IL-1β (2 ng/ml), peaking after 15 min. On average, phosphorylation of p38 increased fourfold 15 min after treatment with IL-1β (2 ng/ml;n = 4 donors). In contrast, IL-1β did not alter the expression of p38 per se (Fig. 1 B). To ensure that the effects of SB-203580 described below were not the result of nonspecific effects on other MAP kinase pathways that have also been shown to be activated by IL-1β, we examined the effects of SB-203580 (3 μM) on MEK1/2 and stress-activated protein kinase (SAPK) activity by measuring its effects on IL-1β-induced phosphorylation of ERK and JNK. Both ERK and JNK were phosphorylated in the presence of IL-1β, but phosphorylation was not altered in the presence of SB-203580 (Fig.1 C). Similar results were obtained in cells from three different donors.
IL-1β (2 ng/ml for 22 h) induced COX-2 expression in HASM cells (Fig. 2) as previously reported (23), and SB-203580 caused a concentration-dependent inhibition of this response. In cells treated with 3 and 30 μM SB-203580 2 h before IL-1β, the density of the COX-2 band was reduced by 70 ± 7 and 90 ± 7%, respectively (P < 0.01; n = 6 donors/group) compared with cells treated with IL-1β alone (Fig. 2 A). Similar results were obtained with SB-202190 (3 μM), which caused a 40 ± 20% reduction (n = 3 donors) in the density of the COX-2 band (Fig. 2 B). In contrast, SB-203580 did not alter the expression of p38 (Fig. 1 B), indicating that the effect on COX-2 was not a nonspecific effect on protein synthesis.
We also examined the effect of SB-203580 on the changes in basal PGE2 release induced by IL-1β (2 ng/ml for 22 h). Compared with release in the control cells, IL-1β caused a significant increase (10- to 20-fold) in basal PGE2release. Laporte et al. (22) have previously reported that this increased PGE2 release can be completely ascribed to the induction of COX-2 evoked by IL-1β. SB-203580 (3 μM) had no effect on the basal PGE2 release in the control cells but decreased IL-1β-evoked PGE2 release by 80% (P < 0.001; Fig.3). The inhibitory effect of SB-203580 was not due to cytotoxicity because release of lactate dehydrogenase into the culture medium was not altered by 24 h of incubation with 30 μM SB-203580 (data not shown).
To further evaluate the role of p38 in IL-1β-induced prostanoid release, we examined the effect of prior administration of SB-203580 on BK (1 μM)- and AA (10 μM)-stimulated PGE2 release in control and IL-1β-stimulated cells. AA-stimulated PGE2release requires both the COX and PGE2 synthase enzymes but not phospholipase A2 (PLA2), whereas in the case of BK-stimulated PGE2 release, PLA2 must also be activated. In control cells, SB-203580 (3 μM for 24 h) had no significant effect on either the BK- or AA-evoked PGE2 release (Fig. 3 A). The results with AA indicate that at this concentration, SB-203580 does not have any nonspecific effect on PGE2 synthase or COX activity. The results with BK suggest that p38 does not play a role in BK-induced PLA2 activation, whereas Laporte et al. (21) have previously reported that ERK is involved in these events. In contrast to its lack of effect in control cells, SB-203580 (3 μM) caused a marked and significant decrease in both BK- and AA-induced PGE2 release in IL-1β-treated cells (Fig. 3 B), consistent with the effect of the p38 inhibitor on COX-2 expression. We also examined the effect of a very short preincubation with SB-203580 (3 μM) on AA-stimulated PGE2 release in IL-1β-treated cells. In this case, SB-203580 was added to the cells for only 30 min before the addition of AA and not throughout the 22-h period of IL-1β pretreatment, thus being unable to influence IL-1β-induced COX-2 expression. Although long-term (24-h) SB-203580 treatment caused an ∼75% reduction in AA-stimulated PGE2 release in IL-1β-treated cells (P < 0.001; Fig. 3), short-term (30-min) SB-203580 treatment had no significant effect on AA-stimulated PGE2 release in IL-1β-treated cells (24.22 ± 2.6 and 18.48 ± 3.1 ng/ml in IL-1β- and SB-203580 plus IL-1β-treated cells, respectively). These results suggest that at this concentration (3 μM), SB-203580 does not have any nonspecific effects on COX-2 activity and support the idea that at this concentration, the effect of long term SB-203580 treatment on PGE2 release in IL-1β-treated cells (Fig. 3) is through effects of the inhibitor on COX-2 expression.
We also examined the effect of IL-1β and SB-203580 on the binding of nuclear proteins to the oligonucleotides encoding the consensus sequences for AP-1 and NF-κB with EMSA. Representative blots are shown in Fig. 4. Compared with binding in the control cells, IL-1β increased the binding of nuclear proteins to the DNA consensus sequences for both AP-1 (Fig. 4 A) and NF-κB (Fig. 4 B), and SB-203580 (30 μM) significantly reduced this effect. When quantified by laser densitometry for 6 experimental days, IL-1β caused a significant increase in nuclear protein binding to the oligonucleotides encoding the consensus sequences for both AP-1 and NF-κB (P < 0.01; Fig. 5). For AP-1, treatment with 30 μM SB-203580 caused a 57% reduction in IL-1β-induced binding of nuclear proteins (P < 0.05; Fig. 5 A), but a lower concentration of SB-203580 (3 μM) had no effect even though this concentration did have significant effects on COX-2 expression and PGE2 release (Fig. 2). For NF-κB, SB-203580 at 30 μM again caused a significant reduction in both the higher and lower molecular weight protein-DNA complexes (32 and 34% reduction, respectively; P < 0.05 for both), but again, a lower SB-203580 concentration (3 μM) had no effect (Fig.5 B). Because Laporte et al. (21) have previously reported that ERK also participates in IL-1β-induced COX-2 expression in these cells and because ERK has been reported to be capable of evoking both AP-1 and NF-κB activation (2,43), we also examined the effect of U-0126 (10 μM) in these experiments. U-0126 is an inhibitor of the MEK1/2 enzyme that activates ERK (8). Although there was a trend for U-0126 to inhibit IL-1β-induced binding of nuclear proteins to the oligonucleotides encoding the consensus sequences for AP-1, the effect was not significant. However, U-0126 did cause a small but significant inhibition of nuclear protein binding to the NF-κB consensus sequences (P < 0.05 for both bands). To ensure that the effect of U-0126 was not the result of nonspecific effects on the p38 MAP kinase pathway, we examined the effects of U-0126 (10 μM) on p38 phosphorylation induced by IL-1β. Phosphorylation of p38 was not altered in the presence of U-0126 (data not shown).
To further evaluate the role of NF-κB in COX-2 expression induced by IL-1β, we examined the effect of two inhibitors of NF-κB activation, MG-132 (29) and PDTC (24), on IL-1β-induced COX-2 expression with Western blotting. There was no effect of MG-132 (10 or 25 μM) treatment on IL-1β-induced COX-2 expression (Fig. 6 A). Similar results were obtained in cells from three different donors (Fig.6 B). In contrast, PDTC caused a decrease in COX-2 expression. At 10 μM, PDTC caused a 32% reduction in COX-2 expression as assessed by densitometric analysis of experiments in three different donors (Fig. 6 B). To verify the efficacy of the NF-κB inhibitors used, we transfected HASM cells with a construct consisting of κB enhancer elements and a luciferase reporter. Transfected cells were stimulated with IL-1β in the presence and absence of PDTC or MG-132. Compared with basal luciferase activity measured in unstimulated cells, IL-1β increased luciferase activity (normalized by β-galactosidase activity to control for transfection efficiency) approximately fivefold (Fig. 6 C) and pretreatment with either MG-132 or PDTC before the addition of IL-1β caused a marked reduction in luciferase activity.
Laporte et al. (22) have previously reported that COX-2-generated prostanoids are implicated in IL-1β-induced β-adrenergic hyporesponsiveness. Because our results indicated that p38 was implicated in IL-1β-induced COX-2 expression and prostanoid release, we sought to confirm that p38 is also involved in the IL-1β-induced β-adrenergic hyporesponsiveness. To do so, we examined the effect of SB-203580 (3 μM) on the IL-1β-induced changes in HASM cell stiffness responses to Iso (Fig.7). Neither SB-203580, IL-1β, nor their combination had any effect on baseline cell stiffness [126.15 ± 11.6 dyn/cm2 in control cells; 112.73 ± 5.99 dyn/cm2 in IL-1β-treated cells; 117.13 ± 10.54 dyn/cm2 in SB-203580 (3 μM)-treated cells; 127.08 ± 9.75 dyn/cm2 in SB-203580 plus IL-1β-treated cells]. In control cells, Iso caused a dose-related decrease in cell stiffness (Fig. 7). Repeated-measures ANOVA indicated a significant effect of drug treatment (P < 0.0001) on Iso-induced changes in cell stiffness. Follow-up analysis indicated that the treatment effect lay in the response to IL-1β (2 ng/ml), which reduced the capacity of Iso to decrease cell stiffness as previously described (21, 22,37). SB-203580 on its own had no effect on the responses to Iso but abolished the effect of IL-1β.
Shore et al. (37) have previously shown that IL-1β decreases HASM cell stiffness responses to Iso but has no effect on the cell stiffness responses to DBcAMP. To ensure that the effect of SB-203580 (Fig. 7) was not the result of nonspecific effects of the drug on the ability of HASM cells to decrease stiffness, we studied its effect on the responses to DBcAMP. DBcAMP caused a concentration dose-related decrease in cell stiffness, consistent with previous reports (21, 22, 37), but neither IL-1β, SB-203580 (30 μM), nor their combination had any effect on these responses (Fig.8).
In this study, we demonstrated that p38 MAP kinase activation is involved in the IL-1β signaling pathway leading to COX-2 expression and PGE2 release in HASM cells. IL-1β caused phosphorylation of p38 (Fig. 1), and selective p38 MAP kinase inhibitors (SB-203580 and SB-202190) decreased IL-1β-induced COX-2 expression (Fig. 2) and PGE2 release (Fig. 3). The role of p38 is unlikely to involve activation of AP-1 and NF-κB because SB-203580 at 3 μM, a concentration that substantially decreased IL-1β-induced COX-2 activation (Fig. 2), had no effect on AP-1 or NF-κB activation after IL-1β stimulation (Fig. 5). PDTC reduced COX-2 expression induced by IL-1β, but another class of NF-κB inhibitors, MG-132, did not block IL-1β-induced COX-2 expression, suggesting that pathways other than NF-κB activation are involved in the induction of COX-2 by IL-1β (Fig. 6). SB-203580 also blocked the effects of IL-1β on the HASM responses to the β-agonist Iso (Fig.7) without affecting the responses to DBcAMP (Fig. 8), consistent with the role of COX-2 in these events (22).
Our results indicate that IL-1β causes a marked increase in the level of phosphorylated p38 MAP kinase (Fig. 1), with a time course consistent with reports in other cell types (11, 12). In HASM cells and other airway smooth muscle preparations, p38 MAP kinases have been also shown to be activated by seven-transmembrane-domain receptor ligands such as endothelin, angiotensin, and BK and by growth factors such as platelet-derived growth factor (9, 23,28). In each of the studies cited above, as in this study, there was a small amount of p38 phosphorylation even under basal conditions.
The observation that SB-203580 blocked COX-2 expression and PGE2 release induced by IL-1β is consistent with reports in other cell types (13, 15, 34) and suggests that p38 is involved in these events. However, IL-1β induction of COX-2 likely requires other signaling pathways as well. For example, Laporte et al. (21) have previously reported that IL-1β increases ERK (p42/p44) phosphorylation and that inhibitors of ERK phosphorylation substantially reduce IL-1β-increased PGE2 release and COX-2 expression in HASM cells. In other cell types, both JNK/SAPK and ERK signaling pathways have been shown to play a role in COX-2 expression and/or PGE2 production (13, 45). For example, Guan et al. (12) have shown that overexpression of the dominant negative form of JNK1 or p54 JNK2/SAPKβ reduces COX-2 expression and PGE2 production by mesangial cells.
There is a report (42) that p38 MAP kinase is involved in the activation of cytosolic PLA2 in other cell types. However, this does not appear to be the case in HASM cells because in control cells, SB-203580 had no effect on BK-stimulated PGE2 release (Fig. 3 A). SB-203580 did inhibit BK-stimulated PGE2 release in IL-1β-treated cells, but this effect is likely to have been the result of inhibition of COX-2 expression rather than of PLA2 activation because the magnitude of the effect on BK-induced PGE2 release, which requires PLA2 activation, and AA induced PGE2release, which does not, was similar. In contrast to p38, Laporte et al. (21) have previously reported that ERK appears to increase PGE2 release both by effects on the induction of COX-2 and by effects on PLA2 activation.
The inhibitory effects of SB-203580 on COX-2 expression and PGE2 release were not the result of cytotoxicity because lactate dehydrogenase release into the culture medium was not altered by 24 h of incubation even with 30 μM SB-203580 and because we observed no effect of SB-203580 on the expression of another protein, p38. We cannot exclude the possibility that the effect of SB-203580 might be the result of nonspecific effects on enzymes other than p38 MAP kinase. However, another p38 inhibitor, SB-202190, with a somewhat different chemical structure and hence likely to have different nonspecific effects, had similar effects to SB-203580 on COX-2 expression (Fig. 2 B). In addition, other investigators (6) have shown that SB-203580 does not block activation of other MAP kinases (p42/p44, JNK/SAPK), MAP kinase kinase, protein phosphatase, p90 S6 kinase, PKA, or c-Raf. Our results indicate that SB-203580 also does not block ERK or JNK phosphorylation in HASM cells (Fig. 1 C). It is theoretically possible that the nonspecific effects of SB-203580 might have contributed to its effect on IL-1β-induced PGE2 release because SB-203580 has been reported to inhibit COX activity in platelets (5). However, we do not think that this is likely. First, SB-203580 (3 μM) had no significant effect on AA-stimulated PGE2 release in control cells (Fig. 3 A). SB-203580 did reduce AA-stimulated PGE2 release in IL-1β-treated cells (Fig. 3 B), but this effect is likely to have been the result of the effects on COX-2 expression rather than on activity because when SB-203580 (3 μM) was administered to IL-1β-treated cells too late to influence COX-2 expression, it had no effect on AA-stimulated PGE2release. Higher concentrations of SB-203580 may have some inhibitory effects on COX activity because with 30 μM SB-203580, we did observe a 48% and significant reduction in AA-induced PGE2 release even in control HASM cells whether SB-203580 was given for 30 min or 24 h (data not shown). Kalmes et al. (17) have shown that concentrations of SB-203580 > 10 μM can have other nonspecific effects, and it is possible that such effects may be responsible for the effects of SB-203580 at 30 μM but not at 3 μM that we observed on AP-1 and NF-κB activation (Figs. 4 and5).
To begin to address the signal transduction pathway by which p38 activation leads to COX-2 expression, we examined the effect of SB-203580 on AP-1 and NF-κB activation using EMSA. The promoter region of the COX-2 gene contains putative binding sequences for both these transcription factors, and both have been implicated in the induction of COX-2 in other cell systems (19). In response to IL-1β, we observed an increase in the binding of nuclear proteins to the oligonucleotides containing the consensus sequences for either AP-1 or NF-κB as previously reported in HASM and other cell types (26, 32, 35, 36). SB-203580 partially inhibited both AP-1 and NF-κB activation but only at a high concentration (30 μM) and not at a concentration (3 μM) at which substantive effects of SB-203580 on IL-1β-induced COX-2 expression were observed (Figs. 4and 5), suggesting that p38 is unlikely to contribute to the induction of COX-2 by inducing activation of AP-1 or NF-κB. In contrast, results with the MEK inhibitor U-0126 (10 μM) at concentrations that Laporte et al. (21) have previously shown to strongly inhibit COX-2 expression suggest that ERK appears to contribute only to the activation of NF-κB, although there was a nonsignificant trend for U-0126 to inhibit AP-1 DNA-binding activity as well.
NF-κB is normally inactive and kept sequestered in the cytoplasm by its interaction with the inhibitory subunit IκB (3). On cell activation, IκB is rapidly phosphorylated, ubiquitinated, and then degraded, resulting in the release and subsequent nuclear translocation of active NF-κB (3). Although the mechanism by which MAP kinases participate in the regulation of NF-κB activation is still not firmly established, it has been reported than MAP kinases or kinases downstream from them may be involved in the phosphorylation of IκB (2). Consistent with these results, we observed that ERK inhibition caused a small but significant reduction in NF-κB activation in HASM cells. Similarly, Reddy et al. (32) reported than ERK participates in IL-1β-stimulated NF-κB activation in a human epidermal cell line. In contrast, and consistent with our results, inhibition of p38 does not appear to influence NF-κB translocation in a kidney cell line stimulated with tumor necrosis factor-α and hydrogen peroxide (43). It is possible that in HASM cells, the effects of p38 on COX-2 expression are mediated not at the transcriptional but at the posttranscriptional level. Several studies (7, 33) have reported that SB-203580 caused a rapid degradation of COX-2 mRNA in cells treated with IL-1β or lipopolysaccharide, suggesting a role for p38 MAP kinase in the stability of COX-2 mRNA. The 3′-untranslated region of COX-2 contains 22 AUUUA motifs that are recognized to be important determinants of mRNA instability (7).
Other investigators (10, 39) have reported that in other cells, the induction of COX-2 can be blocked by NF-κB blockers such as the proteasome inhibitor MG-132 and the oxidant scavenger PDTC. In our study, we found that MG-132 did not prevent the induction of COX-2 by IL-1β even though PDTC did. We do not know why we found different effects with the two inhibitors. The lack of effect of MG-132 is not the result of lack of efficacy because we found that both MG-132 and PDTC markedly reduced IL-1β-induced NF-κB activity as measured by a NF-κB reporter assay. The effect of PDTC is not the result of cell toxicity because we found no effect of PDTC on trypan blue dye exclusion. Instead it is likely that the NF-κB does not participate in IL-1β-induced COX-2 expression in these HASM cells because MG-132 did not inhibit IL-1β-induced COX-2 expression even though it markedly reduced NF-κB activation. The ability of PDTC to inhibit COX-2 expression may reflect a role for oxidants rather than a role for NF-κB in these events because the primary activity of the compound is as an oxidant scavenger (24). The reason for the reported efficacy of MG-132 against COX-2 expression in another study (10) but not in this one may be related to differences in the cell type or the stimulus used to induce COX-2.
Laporte et al. (22) have previously reported that the mechanism by which IL-1β causes decreased HASM cell responses to β-agonist involves COX-2-generated prostanoid release. In particular, they showed that exogenous PGE2 mimics the effects of IL-1β, whereas inhibitors of COX (NS-398 or indomethacin) block the effects of IL-1β. Our results were consistent with the hypothesis that the marked increases in basal PGE2 release that result from the induction of COX-2 by IL-1β lead to increased basal cAMP, consequent PKA activation, and subsequent phosphorylation and heterologous desensitization of the β-adrenergic receptor (22,37). Because the results of this study indicated that p38 activation was required for IL-1β-induced PGE2 release and COX-2 expression, we reasoned that p38 activation should also be involved in IL-1β-induced β-adrenergic desensitization. Our results support that hypothesis. We demonstrated that a selective p38 inhibitor (SB-203580) inhibited the effects of IL-1β on cell stiffness changes induced by Iso. These results are consistent with the effect of SB-203580 on basal PGE2 release: SB-203580 (3 μM) caused a 70% inhibition in COX-2 expression and an 80% reduction in PGE2 release.
Cytoskeleton stiffness as measured here is an index of the ability of cells to resist distortions of shape in response to shear stress applied through magnetic beads linked to the cytoskeletal network via integrin receptors. Theoretical modeling studies of such networks indicate that increasing the interconnectedness of the members, as would occur during actomyosin interactions, increases the stiffness of the network (38). Indeed, application of a variety of contractile agonists to smooth muscle cells results in increased stiffness, whereas a bronchodilating agonist reduces stiffness (14, 37). However, changes in cell adhesion can also influence cytoskeleton stiffness (14, 37), and we cannot rule out the possibility that the observed effects of SB-203580 might be the result of some role for p38 in cell adhesion. However, we believe that such an explanation is very unlikely. First, changes in cell adhesion influence basal cell stiffness in these experiments (14, 37), but neither IL-1β, SB-203580, nor the combination of these agents altered baseline stiffness. Second, such changes would have been expected to alter cell stiffness response to any dilating agonist, but the responses to DBcAMP were unaffected by IL-1β, SB-203580, or their combination (Fig. 8).
In summary, our results indicate that IL-1β activates p38 MAP kinase and that activation of p38 leads to COX-2 expression and an increase in PGE2 release but that the role of p38 is unlikely to involve activation of the transcription factors NF-κB or AP-1. Our results also indicate that p38 MAP kinase activation is required for IL-1β-induced β-adrenergic hyporesponsiveness. Understanding the role of MAP kinases in the mechanism by which cytokines lead to β-adrenergic receptor dysfunction may provide new information for pharmacological intervention for asthma.
We gratefully acknowledge the help of Drs. Geoff Maksym and Ben Fabry in maintaining the magnetometer used in the magnetic twisting cytometer experiments.
This study was supported by National Heart, Lung, and Blood Institute Grants HL-56383, HL-33009, HL-55301, and HL-64063 and National Institute of Allergy and Infectious Diseases Grant AI-40203.
J. D. Laporte was the recipient of an American Lung Association fellowship.
Address for reprint requests and other correspondence: S. A. Shore, Harvard School of Public Health, 665 Huntington Ave., Boston, MA 02115.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society