INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Beta-adrenoceptor agonists are clearly efficacious for the relief of acute bronchospasm in asthma. Newer agents provide increased potency and longer duration of action. However, a possible association of beta-agonist therapy with increased morbidity or mortality has been reported (1). High doses and frequent, prolonged treatment may be most associated with possible adverse beta-agonist actions. Although beta- agonists are well established to relax bronchial smooth muscle constriction, any effects on the underlying inflammatory pathogenesis of asthma are not well understood. Because corticosteroids are frequently used to suppress inflammation in patients being treated with beta-adrenoceptor agonists (4, 5), any interaction between corticosteroids and beta-agonists could be of importance.

Although inflammation in asthma is heterogeneous in character and involves multiple cell types and mediators, the eosinophil may be of particular significance. Pulmonary infiltration with eosinophils is a common characteristic of asthma (6, 7) and eosinophil mediators can cause pulmonary injury as well as bronchial smooth muscle spasm. Eosinophils have a potent respiratory burst and generate high levels of superoxide anion, which may contribute to epithelial damage in asthma. Superoxide anion and other toxic oxygen metabolites generated during the eosinophil respiratory burst can be very sensitively detected with lucigenin-dependent luminescence. Beta-adrenoceptor agonists introduced immediately before eosinophil stimulation have been previously shown to inhibit the respiratory burst (8, 9), an effect that could reduce mediator release and be of clinical value. However, beta-agonists are frequently used chronically, and potent, long-acting beta-agonists are of increasing clinical importance (4, 10). The effects of prolonged beta-adrenoceptor agonist exposure on eosinophil function have not been well studied.

Corticosteroids, which are commonly used in asthma with the objective of reducing inflammation, are well recognized to reduce peripheral eosinophil counts. Corticosteroids may inhibit eosinophil function in vivo (11). Corticosteroids have also been reported to reduce eosinophil mediator generation (12) although studies are not extensive.

Because beta-adrenoceptor agonists and corticosteroids are very commonly used in asthma, the actions and interactions of these agents on eosinophil function and survival were studied. Prolonged beta-agonist exposure was found to increase mediator generation. The proinflammatory beta-agonist effect was not significantly reduced by corticosteroid. Rather, in the presence of corticosteroid, beta-agonists appeared to remain proinflammatory and anti-inflammatory corticosteroid effects were blocked. Beta-agonist proinflammatory actions may not be related to increased eosinophil survival since no reduction in apoptosis could be demonstrated with independent beta- agonist. However, corticosteroid-induced apoptosis was reduced by beta-agonist and blockade of anti-inflammatory corticosteroid actions may be mediated at least partially from a reduction in apoptosis and increased eosinophil survival.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hansel stain was obtained from Lide Labs, Inc. (Florissant, MO). Salmeterol and RU 40555 were provided by D. M. Bain (Glaxo, Greenford, Middlesex, UK) and D. Martini (Roussel Uclaf, Romainville, France), respectively. Nucleic acid stains (Syto 16 and propidium iodide) were obtained from Molecular Probes (Eugene, OR). All other chemicals were obtained from Sigma Chemical Company (St. Louis, MO).

Eosinophils were isolated from health subjects by phlebotomy. Subjects gave informed consent to a study protocol approved by the University of Washington, Seattle, Office of Research, Human Subjects Division. Percoll density gradient centrifugation was used to separate eosinophils from other leukocytes. Briefly, 67.5 ml of blood was obtained by venipuncture and anticoagulated with 11.25 ml of ethylenediaminetetraacetic acid (10 mg/ml) in saline solution. After the addition of 13.5 ml of 4.5% dextran in saline solution, blood was sedimented (1 g) for 45 min at room temperature (23° C). Leukocyte-rich plasma was collected, diluted 1:1 (vol/vol) with phosphate-buffered saline-sodium citrate-bovine serum albumin (PBS-citrate-BSA) buffer. The PBS-citrate-BSA buffer consisted of NaCl (136 mmol/L), KCl (2.7 mmol/L), Na₂PO₄ (8.1 mmol/L), KH₂PO₄ (1.47 mmol/L), sodium citrate (13 mmol/L), BSA (0.5% wt/vol); adjusted to pH 7.4, 290 mOsm/kg, 1.012 g/ml.

After centrifugation (10 min at 200 g) of the diluted leukocyte-rich plasma, the supernatant was decanted, and the cell pellet was resuspended in 9 ml of PBS-citrate-BSA (25 to 30 × 10⁶ cells/ml). Cell suspension (2.2 ml) was gently layered atop a multiple, discontinuous, Percoll density gradient consisting of the following volumes and densities (top to bottom): 2 ml at 1.077 g/ml, 2 ml at 1.0825 g/ml, 2 ml at 1.085 g/ml, 2 ml at 1.0875 g/ml, 2 ml at 1.090 g/ml, 1 ml at 1.0925 g/ml, and 1 ml at 1.095 g/ml. Percoll was mixed with 10-fold concentrated PBS to obtain a stock Percoll solution of 1.12 g/ml and 290 mOsm/kg, pH 7.4. This stock Percoll solution was then mixed with appropriate volumes of PBS-citrate-BSA to obtain solutions of the desired density. All solutions were adjusted to pH 7.4, 290 mOsm/kg. Gradient fractions were collected after centrifugation for 30 min at 1,000 g by piercing the bottom of the polypropylene centrifuge tube with an 18-gauge needle and collecting fractions. After washing and hypotonic lysis of erythrocytes, gradient fractions were suspended in ice-cold Dulbecco's phosphate-buffered saline (DPBS) containing glucose (4.44 mmol/L) and 0.1% BSA. Cells were stored on ice until they were used. Cell counts in each fraction were determined with a hemacytometer, and percentage of eosinophils in each fraction was evaluated by means of differential counts of Hansel-stained cytospin slides. The percentage of eosinophils in the peripheral blood from normal subjects used for these studies averaged to 3.3%. The yield of eosinophils (total from all fractions) averaged to 0.128 × 10⁶ cells per milliliter of blood obtained by phlebotomy. Fractions of density  1.085 g/ml containing purified eosinophils (> 98% pure eosinophils) were used for luminometry studies.

Purified eosinophils were washed once with sterile RPMI 1640 containing 10% FBS and penicillin-streptomycin (100 units/ml and 0.1 mg/ml, respectively) and then cultured in this media (approximately 1-2 × 10⁵ cells per ml). After addition of treatment or vehicle, eosinophil cultures were incubated for 24 h at 37° C under 5% CO₂ atmosphere. Viability of cultured (24 h) eosinophils was assessed by exclusion of trypan blue and ranged from 85-95% (200 cells counted). Significant differences in eosinophil viability as a result of treatments were not detected at 24 h cell culture. Eosinophils (2 × 10⁴ cells) were stimulated with 0.1 ml FMLP (final concentration 1 µM), and lucigenin-dependent luminescence (10 µM lucigenin) from each tube was monitored in a Berthold LB953 luminometer for 1 s at 2-min intervals for 26 min at 37° C. Lucigenin-dependent luminescence was linearly related to superoxide anion generated from the degradation of xanthine by xanthine oxidase (1 to 10 × 10³ units/ml xanthine oxidase and 100 fmol/L xanthine for 26 min at 37° C with 10 µmol/L lucigenin; linear regression R² = 0.97). This range of superoxide anion generation was similar to that generated by stimulated eosinophils (2.5 × 10⁴ cells/ml).

Apoptosis in cultured eosinophils was evaluated by two methods: examination of nuclear morphology on cytospin slides by fluorescence microscopy subsequent to staining with Syto 16 and analysis of cellular DNA content by flow cytometry subsequent to extraction of fragments (13). Cultured eosinophils were fixed with 100 µl of 4% paraformaldehyde for 30 min at room temperature, stained with 1 µM Syto 16, and cytospun 7 min at 44 g with low acceleration. Slides were examined with a Nikon Diaphot fluorescence microscope at ×400 magnification, and 200 cells on each slide were evaluated for nuclear morphology as described by Stern (14).

Eosinophils were prepared for flow cytometric evaluation of apoptosis following the methods of Gong and coworkers (13). Eosinophils (0.5 × 10⁶ cells) were resuspended in 1 ml of PBS, 10 ml of ice cold 70% ethanol was slowly added with mixing, and cells were stored at least 1 h at 20° C. Cells were then pelleted by centrifugation at 800 g for 5 min and resuspended in 750 µl PBS for 5-10 min at room temperature, transferred to microfuge tubes with a 750 µl PBS rinse, centrifuged 10 min and pellets were resuspended in 40 µl of hypertonic phosphate-citrate buffer consisting of 192 parts 0.2 M Na₂HPO₄ and 8 parts of 0.1 M citric acid (pH 7.8). Tubes were incubated 30-45 min at room temperature, centrifuged 5 min at 1,000 g, pellets were resuspended in 200 µl PBS with 50 µl of RNase (10 mg/ml, 50 units/mg) incubated 15 min at room temperature, 250 µl of 200 µM propidium iodide was added (final concentration 100 µM) for 15 min at room temperature. Samples were stored on ice until analyzed on a Becton Dickinson FACSCAN flow cytometer equipped with an argon laser. Fluorescence in channel 2 (propidium iodide-DNA fluoresence) from 10,000 gated events was analyzed. The gate was establised on a forward scatter/side scatter display from a population of freshly isolated eosinophils that had been processed and stained similarly to cultured eosinophils and was set to exclude debris. Apoptotic eiosinophils exhibited a decrease in DNA content and were slightly smaller in size than nonapoptotic eosinophils.

Each experiment was conducted in triplicate, and the means from three to five experiments were used in the analyses. Statistical analysis was conducted with Minitab Statistical Software, Version 9 (State College, PA). Treatments were compared with controls by a paired t test or, for multiple comparisons, by one-way analysis of variance and Dunnett's test (family error rate = 5%).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Eosinophil responses following 24-h exposure to the long-acting beta-2 agonist salmeterol and the nonspecific beta agonist isoproterenol were evaluated with lucigenin-dependent luminescence detection of the respiratory burst induced by chemotactic peptide (1 µM FMLP). Lucigenin detection of superoxide anion was linear and very sensitive (9). As previously reported, the respiratory burst was moderately inhibited when eosinophils were acutely exposed to beta-agonist at the time of stimulation by FMLP (9). In contrast, eosinophils exposed to beta-agonist for 24 h had a significant increase in the respiratory burst (Figure 1). The increase was not apparent after 2-h exposure to isoproterenol (99% control) and progressively increased at 4 h (112%) with further increase at 6 h (120% control) and plateau at 24 h. A maximal increase in respiratory burst was induced by pharmacologically relevant concentrations of beta-agonist (10 nM) using either salmeterol or isoproterenol (Figure 1). The proinflammmatory effect of beta agonist was dose-dependent with a significant increase evident at isoproterenol concentrations as low as 0.1 nM (124% of control). Although cells from different subjects vary in magnitude of the FMLP induced respiratory burst, prolonged exposure to the beta-agonists consistently had a proinflammatory effect. To determine whether beta-agonist was inducing a persistent change in signal transduction or cell function, eosinophils were evaluated 6 h after removal of beta-agonist. The enhancement of the respiratory burst persisted and remained detectable even if eosinophils had been removed from culture media containing beta-agonist for 6 h prior to analysis (data not shown). Because corticosteroids are commonly used to suppress inflammation in asthma, effects of dexamethasone on the eosinophil respiratory burst were evaluated.

View larger version (47K): [in this window] [in a new window]   Figure 1.   Increased FMLP (1 µM) stimulated respiratory burst following 24 h culture in beta-agonists as indicated by lucigenin-dependent luminescence. Eosinophils were stimulated with FMLP (1 µM), and luminescence was monitored for 1 s at 2-min intervals over 30 min. Means of the absolute FMLP-induced luminescence were 54K, 115K, 124K, 124K, and 130K for control, 10 nM iso, 100 nM iso, 10 nM salm, and 100 salm, respectively. Data are shown as percent of internal control responses (same sample eosinophils cultured in the absence of beta-agonists). Mean ± SEM (n = 4) are shown (*p < 0.05 compared with control).

Dexamethasone caused a marked, dose-dependent decrease in the eosinophil respiratory burst in cells incubated with the corticosteroid for 24 h (Figure 2). In contrast, inclusion of dexamethasone during beta-agonist exposure caused only a statistically insignificant trend to decreased superoxide anion generation (Figure 3). The 42% inhibition (p < 0.05) of superoxide anion (Figure 2) induced by 100 nM dexamethasone (respiratory burst was 58% of that in control eosinophils evaluated in parallel but without corticosteroid exposure) was reduced to 22% inhibition (N.S.) when 100 nM isoproterenol was included during culture (Figure 3). Dexamethasone had no detectable effect in the presence of salmeterol (Figure 3). Thus, dexamethasone did not effectively block the proinflammatory respiratory burst increase induced by beta-agonist. Magnitude of the respiratory burst in the presence of combined beta-agonist and corticosteroid was significantly greater than in eosinophils cultured without the drugs.

View larger version (10K): [in this window] [in a new window]   Figure 2.   Dexamethasone inhibition of FMLP (1 µM) stimulated respiratory burst as indicated by lucigenin-dependent luminescence. Absolute FMLP-induced luminescence was 36K (mean control). Data are shown as percent of internal control response (same sample eosinophils cultured 24 h in the absence of dexamethasone). Mean ± SEM (n = 3) are shown.

View larger version (49K): [in this window] [in a new window]   Figure 3.   Beta-adrenoceptor agonists (100 nM) block inhibition of the respiratory burst induced by dexamethasone (100 nM). FMLP (1 µM) induced respiratory burst was indicated by lucigenin-dependent luminescence. Data represent integrated (30 min) luminescence (cps) from a representative experiment repeated three times. Mean ± SEM (n = 5) are shown (*p < 0.05 compared with eosinophils exposed to dexamethasone without beta-agonist).

To determine whether pharmacologically relevant concentrations of beta-agonist would affect the eosinophil in the presence of corticosteroid, low concentrations of the clinically important beta-agonist albuterol were included during a 24-h incubation of eosinophils with dexamethasone (100 nM). Very low concentrations of beta-agonist induced an increase in the respiratory burst of eosinophils incubated with 100 nM dexamethasone (Figure 4). A dose-dependent proinflammatory increase in the respiratory burst was induced by concentrations of albuterol as low as 3 nM (p < 0.05, n = 3).

View larger version (23K): [in this window] [in a new window]   Figure 4.   Dexamethasone (100 nM) inhibition of the eosinophil respiratory burst is antagonized by low concentrations of albuterol. The FMLP (1 µM) induced respiratory burst was indicated by lucigenin-dependent luminescence. Means of the absolute FMLP-induced luminescence were 66K, 55K, 61K, 80K, and 111K for control, 100 nM dex, 100 nM dex + 0.1 nM alb, 100 nM dex + 1 nM alb, 100 nM dex + 3 nM alb, respectively. Data are shown as percent of internal control response (same sample eosinophils incubated for 24 h in the absence of dexamethasone). Mean ± SEM (n = 3) are shown (*p < 0.05).

Because corticosteroids may require several hours before actions are manifest, beta-agonists were introduced after initial corticosteroid exposure to determine whether prior corticosteroid could block proinflammatory beta-agonist actions. Although 24-h exposure to dexamethasone (100 nM) inhibited the respiratory burst to 44 ± 5.6% of control in samples from the three subjects studied, isoproterenol (100 nM) included during the last 4 h increased the respiratory burst to 74 ± 19% (p < 0.05, Figure 5). Thus, pre-exposure to 100 nM dexamethasone for 20 h was unable to block beta-agonist effects and as little as 4 h beta-agonist exposure was sufficient to induce a significant increase in oxygen metabolite generation.

View larger version (34K): [in this window] [in a new window]   Figure 5.   Pre-exposure to dexamethasone does not prevent proinflammatory effects of isoproterenol. Eosinophils were cultured in the presence of dexamethasone (100 nM) for 24 h and isoproterenol (100 nM) was added at 0, 18, 20, 22 h after dexamethasone. FMLP (1 µM) induced respiratory burst at 24 h was measured using lucigenin-dependent luminescence. Means of the absolute FMLP-induced luminescence were 67K, 40K, 128K, 74K, 83K, and 54K for control, no iso, iso at 0 h, 18 h, 20 h, 22 h, respectively. Data are shown as percent of internal control response (same sample eosinophils incubated for 24 h in the absence of dexamethasone. Mean ± SEM (n = 3) are shown (*p < 0.05).

The proinflammatory effect of beta-agonists did not require continued presence of beta-agonist. Eosinophils incubated with 100 nM isoproterenol and 100 nM dexamethasone were removed from beta-agonist containing medium after 18 h and incubated in drug-free media for an additional 6 h before stimulation with FMLP. An increase in the respiratory burst induced by 18 h of beta-agonist/glucocorticoid incubation remained apparent at 24 h (143 ± 20% of control, n = 3), after 6 h of incubation in drug-free medium.

Both corticosteroid and beta-agonist effects could be pharmacologically blocked by antagonists. Inhibitory effects of dexamethasone were blocked by the antagonist RU-40555 (data not shown). Proinflammatory effects of beta-agonist were blocked by nadolol (propranolol was not used because it could interfere with phospholipase D signal transduction with potential inhibition of the respiratory burst). (+)Isoproterenol also had no significant proinflammatory effect in contrast to ()isoproterenol (Figure 6). Thus, the beta-agonist induced increase in superoxide anion generation was blocked by beta-antagonist and was stereospecific, suggesting a beta-adrenoceptor mediated effect.

View larger version (38K): [in this window] [in a new window]   Figure 6.   Proinflammatory action of isoproterenol (3 nM) on the respiratory burst from cultured eosinophils is stereospecific and antagonized by nadolol (30 nM) in the presence of dexamethasone (100 nM). FMLP (1 µM) induced respiratory burst at 24 h was detected with lucigenin-dependent luminescence. Means of the absolute FMLP-induced luminescence were 69K, 59K, 199K, 93K, and 93K for control, dex, ()iso, (+)iso, and ()iso with nadolol. Data are shown as percent of internal control response (same sample eosinophils incubated for 24 h in the absence of dexamethasone). Mean ± SEM (n = 5) are shown (*p < 0.05 compared with eosinophils exposed to dexamethasone without beta-agonist).

Consistent with reports that corticosteroids can reduce eosinophil survival (15, 16), dexamethasone increased eosinophil apoptosis as determined by fluorescence microscopy and flow cytometry.

Morphologic examination with fluorescence microscopy revealed that apoptosis was nearly doubled in cells exposed to dexamethasone (100 nM) for 24 h (Figure 7). Although morphologic studies may be most specific for apoptosis, flow cytometric analysis showed a similar increase in cells with less than G0 DNA content, a result consistent with loss of small DNA fragments as expected in cells undergoing apoptosis (Figure 8). A significant increase in cell necrosis, as judged by trypan blue exclusion, was induced by dexamethasone at 48 h (data not shown).

View larger version (19K): [in this window] [in a new window]   Figure 7.   Dexamethasone (100 nM) increases eosinophil apoptosis after 24 h cell culture, and beta-agonists (isoproterenol: 100 or 3 nM, albuterol: 3 nM) inhibit this response. Data represent percent of eosinophils with condensed nuclear morphology as evaluated with fluorescence microscopy of nucleic acid-stained eosinophils on cytospin slides. Mean ± SEM (n = 5) are shown (*p < 0.05 compared with relevant control).

View larger version (16K): [in this window] [in a new window]   Figure 8.   Flow cytometric analysis of eosinophil populations subsequent to 24 h cell culture in the presence of dexamethasone (100 nM). Figure depicts representative bivariate log-density contour plots of control cell population (left) and dexamethasone-treated population (right). Abscissa represents forward scatter signal (linear amplification) and ordinate represents channel 2 fluorescence (DNA-propidium iodide fluorescence, logarithmic amplification). Contour intervals are 50% log density. Please note the increase in a subpopulation with decreased DNA content (aneuploid, or less than Go) and diminished size in the contour plot on the right.

Beta-agonist tended to decrease spontaneous apoptosis, although results were not statistically significant. Effects of beta-agonist were much more profound in the presence of glucocorticoid. Inclusion of beta-adrenoceptor agonist (3-100 nM) during glucocorticoid exposure markedly reduced apoptosis such that eosinophil apoptosis approached that observed in control cultures without dexamethasone (Figure 7).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of this study demonstrate that beta-adrenoceptor agonists can increase at least one component of eosinophil-mediated inflammation, oxygen metabolite generation, and block potentially important anti-inflammatory actions of corticosteroids. Eosinophils are clearly relevant to the pathogenesis of asthma (7, 17) and toxic oxygen metabolites may contribute to epithelial injury as well as bronchial edema (18). Clearly, increases in eosinophil activity or survival have the potential to prolong or exacerbate asthma. Proinflammatory beta-agonist effects were apparent in vitro at concentrations of 1-3 nM. Since inhaled beta-adrenoceptor agonists can deliver much higher drug concentrations (bronchial beta-agonist concentrations can be calculated to potentially exceed 100 µM or 1,000-fold above the "high" beta-agonist concentration studied) and plasma beta-agonist concentrations have been measured over 200 nM in studies of asthmatic patients (3, 21), the observed increase in oxygen metabolite generation occurred at beta-agonist concentrations which are clinically relevant. An exacerbation of inflammation and inhibition of corticosteroid action could be consistent with the controversial (10, 22) suggestion that prolonged use of higher doses is associated with increased morbidity and possibly mortality in asthma (2, 3).

Three different beta-adrenoceptor agonists were studied and, although direct comparisons were not performed in all studies, proinflammatory effects were similar. The proinflammatory actions appear to be beta-adrenoceptor-mediated as shown by beta-antagonist and stereoisomer studies. The time required before onset of a proinflammatory effect was apparent and the persistence after removal of beta-agonist suggest transcription and translation may be required. Changes in the NADPH oxidase (25), enhancement of signal transduction or degradation of a respiratory burst inhibitor (28, 29) are potential effects of chronic beta-agonist exposure that may warrant further study.

Beta-agonist increased the respiratory burst and inhibited glucocorticoid-induced apoptosis while glucocorticoid inhibited the respiratory burst and increased apoptosis. The association suggests that changes in cell viability may contribute to changes in the measured respiratory burst. Apoptosis-associated depletion of NAD and ATP (30) could diminish NADPH required for the respiratory burst. However, the relative magnitude of changes in apoptosis and oxygen metabolite generation were sufficiently distinct that additional mechanisms are likely to be of importance.

Because it has been proposed that the combination of corticosteroid with beta-adrenoceptor agonist may be of therapeutic value, our results, which suggest corticosteroid may not effectively block beta-agonist-induced increases in oxygen metabolite generation, are of interest. In addition, a potentially desirable effect of corticosteroid to increase apoptosis was blocked by beta-agonist, results which may be consistent with past studies, which suggest beta-agonist can block corticosteroid receptor binding to DNA (33). Thus, in vitro the actions of glucocorticoids and beta-agonists in combination did not appear to meet what may be clinically desirable objectives.

In conclusion, beta-adrenoceptor agonists were found to increase the eosinophil respiratory burst, mildly reduce the rate of spontaneous apoptosis and at least partially block corticosteroid induced apoptosis. The clinical ramifications of such in vitro observations remain unclear. Further studies may be of value since beta-adrenoceptor agonists and corticosteroids are primary agents in the treatment of asthma and are commonly used simultaneously. Drug interactions and any actions to exacerbate inflammation could clearly be of clinical importance.