Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Airway inflammation is a central feature of the pathophysiology of asthma, and cytokine networks play a fundamental role in the chronic inflammatory process (1). Anti- inflammatory treatment with inhaled steroids provides the mainstay of asthma management in conjunction with bronchodilator therapy (2). Short-acting ₂-adrenoceptor (₂AR) agonists have been used as bronchodilator treatment in asthma for decades, and although these agents produce useful bronchodilatation through airway smooth-muscle (ASM) relaxation, concerns have been raised that they may have deleterious proinflammatory effects. When given on their own, short-acting -agonists can cause a rebound increase in bronchial responsiveness after cessation of therapy (3), and regular salbutamol (Salbu) use can increase the early response to allergen and neutrophilic airway inflammation (4). Recently, however, a number of studies have shown, in contrast, that when ₂-agonists (particularly long-acting ₂-agonists) are given in conjunction with inhaled steroids they produce beneficial effects on symptoms, airflow, and asthma exacerbations (5). A possible explanation for this effect is that there is a beneficial interaction between ₂AR agonists and steroids on part of the inflammatory process.

Although most anti-inflammatory research in asthma has concentrated on eosinophilic inflammation, asthma exacerbations are characterized by neutrophilic airway infiltration (9), and neutrophil influx occurs during allergen challenge (14). Neutrophil influx contributes to bronchial hyperresponsiveness. Neutrophils are recruited into the airways in asthma by a number of mediators and chemokines, the most important of which is interleukin (IL)-8 (15). Recent studies from ourselves and others have shown that human ASM is a rich source of biologically active chemokines and mediators including IL-8, which is released in large quantities in response to bradykinin (20), tumor necrosis factor (TNF)-, or IL-1 (21). Inasmuch as ASM mass is increased in chronic asthma, it may serve as an important source of chemokines, which may amplify the inflammatory response (22).

We hypothesized that if steroids and ₂AR have a beneficial interaction in asthma, it would seem likely that this interaction would be maximal in cells such as human ASM cells, which express large numbers of ₂ARs (23). Because we have previously shown that IL-8 production by human ASM cells is regulated by prostaglandin (PG) E₂ (20), which is coupled to adenylyl cyclase and elevations in cyclic monophosphate (cAMP), we postulated that alterations in cAMP in response to ₂-agonists might also regulate IL-8 release. We therefore studied the effect of ₂-agonists on IL-8 release by human ASM cells. We also tested for any interaction between ₂-agonists and the corticosteroids dexamethasone (Dex) and fluticasone (Flut) on IL-8 production by the cytokine TNF- and studied the cAMP dependence of the effect of ₂-agonists. TNF- is increased during asthma exacerbations (24) and causes neutrophilia and airway hyperresponsiveness (25) and therefore seemed a relevant proinflammatory cytokine to study with respect to IL-8 release.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture

Human tracheas were obtained from four postmortem individuals within 12 h of death. The donors had no history of respiratory diseases and no evidence of airway abnormalities. Primary cultures of human ASM cells were prepared from explants of ASM according to methods previously reported (26, 27). Cells at passage 3-4 were used for all experiments. We have previously shown that the cells grown in this manner depict the immunohistochemical and light-microscopic characteristics of typical ASM cells (26).

Experiment Protocol

The cells were cultured to confluence in 10% fetal calf serum (ICN Pharmaceuticals, Basingstoke, Hampshire, UK)-Dulbecco's modified Eagle's medium (Sigma, Poole, Dorset, UK) in humidified 5% CO₂/95% air at 37°C in 24-well culture plates and growth- arrested in serum-deprived medium for 24 h before experiments. Immediately before each experiment, fresh serum-free medium containing TNF- (Sigma) was added. In the time-course experiments the cells were incubated with TNF- (10 ng/ml) for 1 to 24 h, whereas in the concentration-response experiments the cells were incubated for 16 h with 0.1 to 100 ng/ml TNF-. In most experiments thereafter the cells were incubated with 10 ng/ml TNF- for 16 h. At the indicated times, the culture media were harvested and stored at 20°C until the enzyme-linked immunosorbent assay (ELISA) for IL-8. To test the inhibition by various drugs on the effect of TNF-, the ₂-agonists Salbu and salmeterol (Salme), the ₂-antagonist ICI-118,551 (ICI), the direct adenylyl cyclase activator forskolin (FSK), the membrane-permeable cAMP analogue 8-bromoadenosine 3',5'-cAMP (8-Br-cAMP), the specific type IV cAMP-dependent phosphodiesterase inhibitor rolipram (Roli), the corticosteroids Dex (all from Sigma) and Flut propionate (kindly provided by Dr. Malcolm Johnson, GlaxoWellcome Research and Development, Uxbridge, Middlesex, UK), and the specific cAMP-dependent protein kinase (PK) A inhibitor KT-5720 (Calbiochem-Novabiochem, Nottingham, Notts, UK) were added 1 or 2 h before the addition of TNF- as specified in figure captions. The time-course and concentration response of IL-8 release by the ₂-agonists Salbu and Salme and the direct adenylyl cyclase activator FSK were conducted in the same way as TNF-.

IL-8 Assay

The concentration of IL-8 in the culture medium was determined using an ELISA kit (CLB, Amsterdam, the Netherlands) according to the manufacturer's instructions. Briefly, 96-well ELISA plates were coated overnight at room temperature with 200 µL antihuman IL-8 coating antibody diluted in 0.1 M carbonate/bicarbonate buffer (pH 9.6). Plates were then washed five times with phosphate-buffered saline (PBS) (pH 7.2-7.4) and blocked for 1 h at room temperature with 200 µL blocking buffer. Plates were washed again with washing buffer (PBS with 0.05% Tween 20) and 100 µL of recombinant human IL-8 standards (1 to 240 pg/ml) as well as study samples (diluted 1/10-50 with dilution buffer) were added in duplicate to individual wells and incubated at room temperature for 1 h. After five washes, 100 µL of biotinylated IL-8 antibody diluted in dilution buffer was added for 1 h. After another five washes, 100 µL of streptavidin-horseradish peroxidase (HRP) conjugate, diluted 1/10,000 in dilution buffer, was added for 30 min. After the final washes, 100 µL of the substrate buffer containing HRP substrate tetramethylbenzidine dihydrochloride and hydrogen peroxide in 0.05 M phosphate-citrate buffer (pH 5.0) was added for 30 min in the dark and color-developed in proportion to the amount of IL-8 present. The reaction was stopped by adding 100 µL of stop solution (1.8 M sulfuric acid) and the degree of color generated was determined by measuring the optical density at 450 nm in a Dynatech MR5000 microplate reader (Dynatech, Billinghurst, Sussex, UK). The standard curve was linearized and subjected to regression analysis. The IL-8 concentration of unknown samples was extracted by using the standard curve. The results were expressed as picograms per milliliter of culture medium. The sensitivity of the ELISA kit at our hands was at least 5 pg/ml, which was consistent with the manufacturer's specifications. According to the kit insert, the anti-IL-8 antibody does not cross-react with IL-1 through -7 and IL-9 through -11, TNF, interferon-, granulocyte macrophage colony-stimulating factor, and regulated on activation, normal T cells expressed and secreted. All reagents used in the assay were supplied by the ELISA manufacturer except the HRP substrate tetramethylbenzidine dihydrochloride, which was obtained from Sigma.

Cell Viability

The toxicity of all the chemicals used in this study and their vehicles dimethyl sulfoxide (DMSO) and ethanol (Sigma; final concentration  0.6% vol/vol) to human ASM cells was determined by thiazolyl blue, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltertrazolium bromide (MTT) assay (26). After 16 or 24 h incubation with the chemicals, 20 µl 5 mg/ml MTT (Sigma) was added to the culture medium in 96-well plates and incubated for 1 h at 37°C. After removing the medium, 200 µl DMSO was added to solubilize the blue-colored tetrazolium, the plates were shaken for 5 min, and the optical density₅₅₀ values were read in a Dynatech MR5000 microplate reader. Viability was set as 100% in control cells.

Statistical Analysis

Data are expressed as means ± standard error of the mean (SEM) from n determinations (wells). Statistical analysis was performed by using statistical software from SPSS, Inc. (28). One-way analysis of variance and/or unpaired two-tailed t test were used to determine the significant differences between the means. The results were adjusted for multiple testing by using Bonferroni's correction. P values of less than 0.05 were accepted as statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of TNF- on IL-8 Production

To investigate the time course of IL-8 production, human ASM cells were cultured in the presence or absence of TNF- (10 ng/ml). Cell-culture supernatants from control cells were collected at 0.5, 2, 8, and 16 h, and those from TNF--treated cells were collected at 0.5, 1, 2, 4, 8, 16, and 24 h. IL-8 release from control cells was low even though there was a slight increase over the incubation period up to 16 h (5.13 pg/ml at 0.5 h, 8.68 pg/ml at 2 h, 12.06 pg/ml at 8 h, and 21.16 pg/ml at 16 h). There was a marked and time-dependent increase in IL-8 release after stimulation with TNF-, significant difference was observed from after 0.5 h of stimulation as compared with IL-8 production from control cells (P < 0.001), and the highest IL-8 concentration was achieved after 16 h stimulation (P < 0.001) (Figure 1A). When the cells were cultured with TNF- at concentrations of 0.1, 1.0, 10, and 100 ng/ml for 16 h, a concentration-dependent increase in IL-8 production was also observed, which was significant from 0.1 ng/ml (P < 0.001) and peaked at 100 ng/ml (Figure 1B). With regard to the results, an incubation time of 16 h and a concentration of 10 ng/ml were chosen for the following experiments.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Time course (A) and concentration response (B) of TNF- on IL-8 production. Human ASM cells were incubated with 10 ng/ml TNF- for the times indicated or with increasing concentrations of TNF- for 16 h for the concentration response. IL-8 accumulation in the medium was measured by ELISA as described in MATERIALS AND METHODS. Each point represents the mean ± SEM of six determinations from two independent experiments.

Effects of Various cAMP Stimulants on IL-8 Release

The effect of the ₂-agonists Salbu and Salme and the direct adenylyl cyclase activator FSK on IL-8 production from human ASM cells was assessed. As shown in Figure 2A, Salbu (10 µM), Salme (1 µM), and FSK (10 µM) each caused a time-dependent increase in IL-8 accumulation that was significant after 4 h incubation with Salbu (P < 0.01 compared with control 10.13 pg/ml), 8 h incubation with Salme (P < 0.001 compared with control 12.66 pg/ml), and 2 h incubation with FSK (P < 0.001 compared with control 7.32 pg/ml). In the concentration response, all three cAMP stimulants enhanced IL-8 release in a concentration-dependent manner, with significant increase observed at 0.1 µM for Salbu (P < 0.05 compared with control 9.02 pg/ml), at 0.1 µM for Salme (P < 0.001), and at 1 µM for FSK (P < 0.001) (Figure 2B). Maximum effect was achieved with 10 µM Salbu and FSK and 1 µM Salme (Figure 2B). The magnitude of IL-8 release by these three cAMP stimulants was much smaller than that of TNF-, suggesting that the increase of cAMP is a weak inducer of IL-8 generation from human ASM cells.

View larger version (19K): [in this window] [in a new window]   Figure 2.   Time course (A) and concentration response (B) of cAMP stimulants on IL-8 production. Human ASM cells were incubated with 10 µM Salbu, 1 µM Salme, or 10 µM FSK for the times indicated or with increasing concentrations of Salbu, Salme, or FSK for 16 h for the concentration response. IL-8 accumulation in the medium was measured by ELISA as described in MATERIALS AND METHODS. Each point represents the mean ± SEM of six determinations from two independent experiments.

Effects of Steroids on TNF--Induced IL-8 Release

Pretreatment of the cells with steroids Dex (0.1 to 10 µM) and Flut (0.01 to 1 µM) before TNF- stimulation resulted in a concentration-dependent inhibition (P < 0.001 for all concentrations) but did not abolish TNF--induced IL-8 release (Figure 3). The amounts of 1 µM of Dex and 0.1 µM of Flut were chosen for subsequent experiments.

View larger version (17K): [in this window] [in a new window]   Figure 3.   Effects of Dex (A) and Flut (B) on TNF--induced IL-8 production. Human ASM cells were pretreated with or without increasing concentrations of Dex (0.1 to 10 µM) or Flut (0.01 to 1 µM) for 1 h before incubation with TNF- (10 ng/ml) for 16 h. IL-8 production was measured by ELISA as described in MATERIALS AND METHODS. Each point represents the mean ± SEM of six determinations from two independent experiments. ***P < 0.001 compared with the effect of TNF-.

Effects of cAMP Stimulants on TNF--Induced IL-8 Release

To investigate whether the increase of cAMP had an impact on cytokine-induced IL-8 release, we studied the effects of Salbu, Salme, and FSK on TNF--induced IL-8 production from human ASM cells. As shown in Figure 4, TNF- strongly stimulated IL-8 release, but pretreatment of the cells with the cAMP stimulants (all at 0.1 to 10 µM) had no significant effect on the increase. The results suggest that even though cAMP increase on its own weakly but significantly stimulates IL-8 release, it has no effect on the strong accumulation of IL-8 induced by TNF-.

View larger version (38K): [in this window] [in a new window]   Figure 4.   Effect of cAMP stimulants on TNF--induced IL-8 production. Human ASM cells were pretreated with or without increasing concentrations (0.1 to 10 µM) of Salbu, Salme, or FSK for 1 h before incubation with TNF- (10 ng/ml) for 16 h. IL-8 production was measured by ELISA as described in MATERIALS AND METHODS. Each point represents the mean ± SEM of six determinations from two independent experiments.

Effect of the Combination of Steroids and cAMP Stimulants on TNF--Induced IL-8 Release

We then went on to study whether the combined use of steroids and cAMP stimulants could further affect TNF-- induced IL-8 release. As shown before, pretreatment of the cells with either 1 µM Dex or 0.1 µM Flut strongly inhibited but did not abolish TNF--induced IL-8 release (Figure 5; P < 0.001). However, when Dex and Flut were used in combination with Salbu (1 and 10 µM), Salme (0.1 and 1 µM), or FSK (1 and 10 µM), a significant further inhibition over the effects of Dex and Flut alone on IL-8 release was observed (Figure 5).

View larger version (21K): [in this window] [in a new window]   Figure 5.   Effects of the combination of steroid Dex (A) and Flut (B) with cAMP stimulants on TNF--induced IL-8 release. Human ASM cells were pretreated with or without Dex (1 µM), or with Dex + Salbu, Dex + Salme, Dex + FSK, Flut (0.1 µM), Flut + Salbu, Flut + Salme, or Flut + FSK for 1 h before incubation with TNF- (10 ng/ml) for 16 h. IL-8 production was measured by ELISA as described in MATERIALS AND METHODS. Each point represents the mean ± SEM of six determinations from two independent experiments. ***P < 0.001 compared with the effect of TNF-; ++P < 0.01 and +++P < 0.001 compared with the effect of Dex or Flut alone.

cAMP Dependence of the Synergistic Inhibition

To clarify the role of cAMP in the synergistic inhibition of IL-8 production described earlier, we examined whether KT5720, a potent and selective inhibitor of the cAMP-dependent PKA, could reverse the synergistic inhibition and whether Roli, the specific inhibitor of the cAMP-dependent phosphodiesterase, could further enhance the synergistic inhibition. As shown in Figure 6A, combined pretreatment with Flut (0.1 µM) + Salme (1 µM) and Flut + FSK (10 µM) resulted in marked further inhibition of TNF-- induced IL-8 release as compared with the effect of Flut alone. KT5720 on its own did not affect TNF--induced IL-8 release; however, when it was used together with Flut + Salme and Flut + FSK it significantly reduced the synergistic inhibitory effects of Flut + Salme (P < 0.05) and Flut + FSK (P < 0.01). Roli on its own did not have any effect on TNF--induced IL-8 release; however, when it was used together with Flut + Salme and Flut + FSK, an enhanced inhibition was observed with Flut + Salme (P < 0.05) but not with Flut + FSK, which already appeared to have a much stronger inhibition than Flut + Salme (Figure 6B). The results indicate that the synergistic inhibition of TNF--induced IL-8 release by the combination of steroids and cAMP stimulants is mediated by cAMP increase.

View larger version (20K): [in this window] [in a new window]   Figure 6.   Effect of KT-5720 (A) and Roli (B) on the synergistic inhibition of TNF--induced IL-8 release by Flut and cAMP stimulants. Human ASM cells were pretreated with or without KT-5720 (1 µM) or Roli (10 µM) for 1 h before the addition of Flut (0.1 µM), Flut + Salme (1 µM) or Flut + FSK (10 µM), respectively, for another 1 h before the final incubation with TNF- (10 ng/ml) for 16 h. IL-8 production was measured by ELISA as described in MATERIALS AND METHODS. Each point represents the mean ± SEM of six determinations from two independent experiments. **P < 0.01 and ***P < 0.001 compared with the effect of Flut; +P < 0.05 and ++P < 0.01 compared with the effect of Flut + Salme or Flut + FSK.

We also tested whether the effect of cAMP stimulants could be mimicked by the membrane-permeable cAMP analogue 8-Br-cAMP. The results showed that 8-Br-cAMP, like the cAMP stimulants tested, weakly but significantly stimulated IL-8 release from the cells in a concentration-dependent manner (Figure 7A). Pretreatment with 8-Br-cAMP (1 to 100 µM) did not alter TNF--induced IL-8 release (Figure 7B); however, when 8-Br-cAMP was used in combination with 0.1 µM Flut, a significant further inhibition over the effect of Flut alone on IL-8 release was achieved in a concentration-dependent manner (Figure 7C).

View larger version (18K): [in this window] [in a new window]   Figure 7.   Effect of 8-Br-cAMP on IL-8 release (A) and TNF-- induced IL-8 release (B), and effect of the combination of Flut with 8-Br-cAMP on TNF--induced IL-8 release (C). Human ASM cells were treated with increasing concentrations of 8-Br-cAMP (1 to 10 µM) for 16 h (A), or pretreated with 8-Br-cAMP (B), or Flut (0.1 µM) or Flut + 8-Br-cAMP (C) for 1 h before incubation with TNF- (10 ng/ml) for 16 h. IL-8 production was measured by ELISA as described in MATERIALS AND METHODS. Each point represents the mean ± SEM of six determinations from two independent experiments. *P < 0.05 and ***P < 0.001 compared with control (A); **P < 0.01 compared with the effect of TNF- (C); ++P < 0.01 and +++P < 0.001 compared with the effect of Flut alone (C).

₂-Receptor Involvement in the Synergistic Inhibition

Finally, we verified whether the synergistic inhibition by ₂-receptor agonists and steroids was mediated by ₂-receptors. We found that when the cells were pretreated with the ₂-receptor-specific antagonist ICI before Flut + Salme or Flut + Salbu, the synergistic inhibition was significantly reversed (P < 0.01; Figures 8A and 8B, respectively). However, the antagonist had no effect on the synergistic inhibition by Flut + FSK (Figure 8A).

View larger version (21K): [in this window] [in a new window]   Figure 8.   Effect of ICI on the synergistic inhibition of TNF--induced IL-8 release by Flut and cAMP stimulants. Human ASM cells were pretreated with or without ICI (10 µM) for 1 h before the addition of Flut (0.1 µM) (A and B), Flut + Salme (1 µM), Flut + FSK (10 µM) (A), or Flut + Salbu (10 µM) (B) for another 1 h before the final incubation with TNF- (10 ng/ml) for 16 h. IL-8 production was measured by ELISA as described in MATERIALS AND METHODS. Each point represents the mean ± SEM of six determinations from two independent experiments. **P < 0.01 and ***P < 0.001 compared with the effect of Flut alone; ++P < 0.01 compared with the effect of Flut + Salme (A) or Flut + Salbu (B).

Cell Viability

Cell viability after 16 h (some 24 h) treatment with the chemicals used in this study was consistently > 95% compared with cells treated with the vehicles.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

There are several novel findings in our studies. First, we showed that the short- and long-acting ₂-agonists Salbu and Salme, when given alone, increased release of IL-8. Second, we showed that Salbu or Salme, given on their own, did not alter TNF--stimulated IL-8 production but when given in the presence of Dex or Flut synergistically increased the inhibitory effects of these corticosteroids on TNF--stimulated IL-8 release. The effects of Salbu and Salme on steroid-mediated suppression of TNF--stimulated IL-8 release were mimicked by the direct activator of adenylyl cyclase FSK and the membrane-permeable cAMP analogue 8-Br-cAMP; inhibited by KT5720, a selective inhibitor of cAMP-dependent PKA (29); and potentiated by the type IV phosphodiesterase inhibitor Roli (30); suggesting that the effects of Salbu and Salme are cAMP-mediated. These studies may provide a mechanistic explanation for the paradoxical findings that ₂-agonists, given on their own, are proinflammatory in asthma, but that combining a long-acting ₂-agonist and an inhaled corticosteroid produces complementary benefits on symptoms and airflow and potentiates the ability of steroids to reduce asthma exacerbations.

The cell-culture methods and assays used in these studies have been previously validated in our laboratories (20, 26, 27). As we previously showed that PGE₂, which is coupled to adenylyl cyclase, elevated IL-8 production by human ASM cells (20), we hypothesize that other agents binding to receptors coupled to adenylyl cyclase might have similar effects. We found that Salbu and Salme caused a time- and concentration-dependent increase in IL-8 release when administered on their own. The fact that their effects were mimicked by FSK and 8-Br-cAMP suggests that the effects are cAMP-mediated. This is the first report of an increase in IL-8 release by human ASM cells in response to Salbu and Salme, although a similar effect of Salbu on IL-8 production by airway epithelial cells has previously been reported (31) and FSK has been shown to increase IL-8 release by colonic epithelial cells (32). The magnitude of IL-8 release by the three cAMP stimulants and 8-Br-cAMP given on their own was fairly small, suggesting that cAMP is a relatively weak inducer of IL-8 release from human ASM cells.

We then studied the effects of ₂-agonists and other cAMP stimulants, given either on their own or in combination with corticosteroids, on TNF--stimulated IL-8 release. TNF- has previously been shown to be a potent stimulant of IL-8 release by human ASM cells (21). We decided to use this as the cytokine of choice in our experiments because TNF- levels are increased in bronchoalveolar lavage fluid in asthma (24) and TNF- inhalation increases bronchial hyperresponsiveness associated with an increase in neutrophilia (25), suggesting that it may be an important mediator that contributes to neutrophil accumulation and bronchial hyperresponsiveness during asthma exacerbations. We chose to use TNF- in these studies rather than IL-1 because we have shown that IL-1, but not TNF-, induces cyclooxygenase (COX)-2, the inducible form of COX in human ASM cells causing substantial prostanoid release, and that this can impair adenylyl cyclase function (26, 33). In contrast, TNF- neither induces COX-2 nor impairs adenylyl cyclase function over the time course used in our experiments (26, 33). We found that the effect of TNF- on IL-8 release was much stronger than the weak positive effect of cAMP stimulants and was concentration-dependent but partially inhibited by both Dex and Flut. We also found that the effect of either Dex or Flut was considerably enhanced when cells were cotreated with either Salbu or Salme. In contrast, Salbu or Salme given in the absence of corticosteroids had no effect on TNF--induced IL-8 release. The subsequent studies to probe the mechanism of this synergistic inhibitory effect using FSK, 8-Br-cAMP, KT-5720, and Roli suggest that the effects are cAMP-mediated. The magnitude of the effects of Salbu, Salme, and FSK is consistent with the relative potency of these agents in increasing cAMP in these cells (data not shown). The fact that the effects of Salbu and Salme could be antagonized by the ₂-receptor-selective antagonist ICI (34) suggests that it is ₂-receptor-mediated. As would be expected, ICI had no effect on the synergistic inhibition produced by Flut and FSK because FSK activates adenylyl cyclase directly.

The precise mechanism of the effects of cAMP on steroid-induced inhibition of IL-8 release has not yet been studied. However, because there is no cAMP response element on the IL-8 gene promoter, the effect is likely to be a secondary one on other transcriptional elements. In many cells, TNF--induced IL-8 release is mediated by the transcription factors nuclear factor-B and activator protein-1 (35, 36). Further studies are required to examine interactions between steroids and ₂-agonists on the transcriptional regulation of the IL-8 promoter to determine the mechanism of the synergistic effect in human ASM cells. It is of interest that in primary lung fibroblasts and vascular smooth-muscle cells ₂-agonists have been shown to increase the nuclear translocation of the glucocorticoid receptor, which might provide a mechanistic explanation for the effects seen in our studies (37) and the beneficial interaction between Salme and Dex on allergen-induced blood mononuclear cell activation reported by others (38).

We have considered the implications of our findings with respect to the interactions between ₂-agonists and corticosteroids in asthma. Some studies have shown that ₂-agonists, when given on their own, are proinflammatory (3, 4), and our findings with IL-8 production are consistent with a weak proinflammatory effect. The strong synergistic inhibitory interaction between ₂-agonists and steroids on TNF-stimulated IL-8 production seen in our study is, however, more interesting. A number of recent studies, particularly with long-acting ₂-agonists, have shown that these agents potentiate the effects of cortico-steroids on airflow and asthma symptoms and cause a reduction in asthma exacerbations (5). Our data may be particularly relevant to reduction in asthma exacerbations that occurred in these studies, inasmuch as asthma exacerbations are associated with significant neutrophilic inflammation of the lungs and the neutrophilia is thought to be IL-8-driven. The effects were seen in our studies only at high concentrations of ₂-agonists. It is difficult to know how the concentrations of -agonists used in our in vitro studies compare with the concentrations of -agonists found in vivo. It is also possible that the -receptor responsiveness of ASM cells in vitro may be different from that in vivo. Nevertheless, our studies suggest a possible mechanism whereby ₂-agonists might potentiate the reduction in cytokine-stimulated IL-8 release produced by corticosteroids. The studies that have shown the greatest effect on asthmatic subjects in vivo have used long-acting ₂-agonists in combination with steroids. In our experiments, we saw effects with both short- and long-acting ₂-agonists. This discrepancy may be explained by the fact that under the experimental conditions used in vitro, Salbu is likely to behave more as a long-acting ₂-agonist because it is not being washed out or metabolized, whereas in vivo it is metabolized, cleared, and excreted.

We have previously reported that bradykinin (BK) stimulates IL-8 release from human ASM cells largely via a prostanoid-dependent mechanism (20). Because PGE₂ is the major prostanoid from human ASM cells and it also causes cAMP accumulation, the effect of BK on IL-8 release, like that of ₂-agonists we showed in the present study, is also likely to be cAMP-dependent. It would be interesting to know why under certain circumstances cAMP exerts a stimulatory effect, whereas under others it exerts an inhibitory effect, on IL-8 release from human ASM cells. As we stated earlier, TNF- does not stimulate prostanoid generation from these cells, yet exerts a much stronger stimulatory effect on IL-8 release than both BK and ₂-agonists, thus providing a good model for studying cAMP manipulation of IL-8 release. It is reasonable to speculate from our data that when cAMP is the major mechanism for IL-8 release, it exerts a stimulatory effect; however, when mechanism(s) other than cAMP is the major mechanism, cAMP exerts an inhibitory effect. In view of this, BK and PGE₂ are both likely to have similar effects as ₂-agonists on TNF--induced IL-8 release. We have also found that IL-1 stimulates the release of PGE₂ and IL-8 simultaneously, and the COX inhibitor indomethacin, which stops prostanoid generation, enhances IL-8 release (unpublished observation), suggesting that IL-1- induced IL-8 release is not prostanoid-dependent and that endogenous prostanoids inhibit simultaneous IL-8 release. The results add weight to this speculation and to the suggestion that prostanoid generation as a result of COX-2 induction may exert a braking effect on inflammatory responses.

In conclusion, we found that ₂-agonists, given on their own, caused a small increase in IL-8 production from human ASM cells. In contrast, when these agents were given in combination with corticosteroids they potentiated corticosteroid-induced inhibition of TNF--mediated IL-8 release in a cAMP-dependent manner. Our findings may be relevant to the mechanisms of action of these drugs in asthma.