INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Asthma is a chronic, inflammatory disease of the airways in which a number of proinflammatory and immune cells have been implicated, including mast cells, monocytes/macrophages, T-lymphocytes, and eosinophils (1). The inflammatory basis of allergic asthma has led to the view that inhaled glucocorticosteroids should be considered a first-line treatment since they are effective in controlling many facets of the inflammatory response (2, 3). Paradoxically, the increased use of steroids in asthma therapy has not seen a reduction in morbidity or mortality. In fact in many countries, the prevalence of asthma and the number of asthma deaths have increased dramatically (4), a phenomenon that has been attributed to increased use of inhaled ₂-adrenoceptor agonists (reviewed in reference 7). One theory that could explain this paradox is that chronic use of sympathomimetic bronchodilators compromises the anti-inflammatory action of glucocorticosteroids (8). Indeed, a negative interaction between ₂-adrenoceptor agonists and glucocorticosteroids has been reported at the molecular level in epithelial cells and in human and rat lung (8). In a study by Peters and coworkers (8), albuterol and fenoterol added together with dexamethasone reduced the binding of the activated glucocorticoid receptor (GR) to DNA without altering GR number or the affinity of dexamethasone. The activation of the transcription factor cyclic AMP-response element-binding protein (CREB) and its associated coactivator CREB-binding protein (CBP) by cyclic AMP is believed to underlie this effect, as forskolin also reduced the binding of activated GR to DNA (8). This paradigm is supported further by the demonstration of a direct protein-protein interaction between GR and CREB within the nucleus of rat hepatoma cells that is apparently required for steroids to enhance maximally transcription of the phosphoenol pyruvate carboxykinase gene (11).

Taken together, these data suggest that ₂-adrenoceptor agonists might reduce the ability of inhaled steroids to negatively influence the transcription of proinflammatory genes, which would have the effect of worsening asthma. According to this hypothesis, the GR-CREB interaction would only have pathophysiologic relevance with high doses of ₂-adrenoceptor agonists, which have been associated with increased asthma morbidity and mortality (12). This is an attractive idea, as activated CREB can persist within the nucleus, implying that a brief exposure to ₂-adrenoceptor agonists could lead to a prolonged effect on the ability of steroid-GR complex to influence gene transcription (13).

Despite molecular data demonstrating an interaction of GR with other transcription factors, including CREB (8, 13, 14), evidence that this phenomenon has functional or clinical relevance is scant. However, in one study Wong and colleagues (15) found that regular treatment of asthmatic subjects with the inhaled glucocorticosteroid, budesonide, improved lung function and protected against the bronchoconstriction experienced after allergen provocation. However, when the ₂-adrenoceptor agonist, terbutaline, was administered concurrently with the steroid, the protection against allergen-induced bronchoconstriction was lost.

In this report, we present the results of experiments designed to address formally the potential functional consequences of the GR-CREB interaction that has been described in molecular studies. Specifically, we have determined if albuterol can reduce the ability of dexamethasone to inhibit the generation of tumor necrosis factor- (TNF) and granulocyte/macrophage colony-stimulating factor (GM-CSF). We selected human peripheral blood monocytes as a test system since they produce an array of proinflammatory cytokines, which are induced by lipopolysaccharide (LPS) in a steroid-sensitive manner. In addition, it has been reported that the number of immature macrophages with phenotypic characteristics of monocytes is significantly increased in asthmatic airways, suggesting a role in the inflammatory response (see 16). GM-CSF and TNF were chosen since they are believed to play an important role in initiating and/or perpetuating some of the chronic manifestations of allergic asthma (17). Indeed, protein and mRNA transcripts for TNF are elevated in a number of proinflammatory cells resident in the lung after activation of the IgE receptor in vitro (17), and there is evidence for enhanced expression of TNF mRNA in asthmatic airways (20). In vivo, it has been documented that the amount of TNF in bronchoalveolar lavage fluid recovered from patients with symptomatic asthma is significantly increased with respect to asthmatic subjects who are essentially asymptomatic (17). Moreover, exposure of asthmatic subjects to allergen evokes the release of TNF at a time that coincides with the late-phase response (21). Similarly, GM-CSF is produced by a range of proinflammatory and immune cells implicated in the pathogenesis of asthma; there is evidence for increased expression in epithelial cells from asthmatic subjects and in T-lymphocytes and eosinophils after endobronchial allergen provocation (22). In the bloodstream of patients with severe asthma the concentration of GM-CSF is elevated (23), which might relate in part to the finding that peripheral blood monocytes harvested from patients with asthma secrete more GM-CSF than do those from nonasthmatic subjects (24). GM-CSF also acts on other cells within the lung, including eosinophils, where it enhances their survival (25) and ability to release leukotrienes and superoxide anions (26).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolation and Purification of Human Mononuclear Cells

Blood was collected from normal healthy subjects by antecubital venepuncture into acid citrate dextrose (in mM: disodium citrate, 160; glucose, 110; pH, 7.4) and mixed with 6% wt/vol Hespan (hydroxymethyl starch) to sediment erythrocytes. After standing at room temperature for 90 min, the leukocyte-rich plasma was removed and centrifuged at 312 g for 7 min. The resulting cell pellet was resuspended gently in approximately 7 ml of buffer A (in mM: KH₂PO₄, 5; K₂HPO₄, 5; NaCl, 110; pH, 7.4) made 50% vol/vol with Percoll and layered over a discontinuous Percoll density gradient (63% and 73% vol/ vol) in buffer A. Mononuclear cells were subsequently separated from polymorphonuclear cells by centrifugation at 1,200 g for 25 min at 18° C. Using this procedure, mononuclear cells were recovered from the 50%/63% vol/vol Percoll interface.

Mononuclear cells were washed three times in Ca²⁺/Mg²⁺-free HBSS to remove Percoll and finally suspended in Ca²⁺/Mg²⁺-free HBSS at a concentration of 10⁶ cells/ml. Cells (5 × 10⁵) were added to 24-well culture plates (Greiner Labortecnik Ltd, Dursley, Gloucestershire, UK) containing 500 µl Dutch modified RPMI 1640 (RPMI 1640 supplemented with 10% FCS, 2 mM L-glutamine, 100 units/ml penicillin, and 100 µg/ml streptomycin) and allowed to adhere to the plastic for 90 min at 37° C in a humidified incubator under an atmosphere of 5% CO₂. The purity of the adherent cell population was routinely > 90%. Plates were agitated, nonadherent cells were decanted, and the resulting monocytes were cultured for various times (see text and figure legends for details) in 1 ml supplemented Dutch-modified RPMI 1640 in the absence and presence of the drugs under investigation. The TNF and GM-CSF released into the culture supernatant were subsequently measured by specific ELISAs (see below).

Measurement of TNF

TNF was measured using an amplified sandwich ELISA. Ninety-six-well, round-bottom plates (Greiner Labortecnik) were coated with 100 µl of a mouse antihuman TNF monoclonal antibody diluted 1:400 in buffer B (in mM: Na₂CO₃, 15; NaHCO₃, 36; NaN₃,15: pH, 9.6) and left for 2 h at 37° C. Plates were washed subsequently in buffer C (in mM: NaCl, 145; KCl, 4; NaH₂PO₄, 10; Na₂HPO₄, 3; and 0.05% vol/ vol Tween-20 at pH 7.4) and treated immediately with 100 µl BSA (5% wt/vol) for 30 min at 37° C. After a further wash with buffer C, 100 µl TNF standards, quality controls, and unknown samples (diluted 1 in 4 in buffer C) were added to the plates and left for 18 h at 4° C. Plates were washed in buffer C, incubated for 2 h with 100 µl of a rabbit antihuman polyclonal TNF antibody (diluted 1:500 in buffer C supplemented with 10% vol/vol FCS), washed again, and then incubated for a further 2 h at room temperature with 100 µl of an alkaline-phosphatase-labeled sheep antirabbit polyclonal IgG antibody (diluted 1:2,000 in buffer C supplemented with 10% FCS). Plates were washed once more and developed with a p-nitrophenyl phosphate assay kit according to the manufacturer's instructions. TNF was measured colorimetrically at 405 nm and quantified by interpolation from a standard curve constructed to known concentrations of hrTNF. The detection limit of this assay is 8 pg /ml.

Measurement of GM-CSF

GM-CSF released from cultured monocytes was measured with an amplified sandwich ELISA. Ninety-six-well, round-bottom plates were coated with 50 µl of a rat, antihuman GM-CSF monoclonal antibody diluted 1:250 in buffer D (0.1 M NaHCO₃ and 15 mM NaN₃ at pH 8.2) and left overnight at 4° C. Plates were subsequently washed in buffer C (145 mM NaCl, 4 mM KCl, 10 mM NaH₂PO₄, 3 mM Na₂HPO₄, and 0.05% vol/vol Tween-20 at pH 7.4) and immediately blocked with 200 µl FCS (10% vol/vol in buffer C) for 2 h at room temperature. After an additional wash with buffer C, 100 µl GM-CSF standards, quality controls, and unknown samples, in supplemented Dutch-modified RPMI 1640, were added to the plates and left for 18 h at 4° C. Plates were washed in buffer C, incubated for 45 min at room temperature with 100 µl of a biotinylated rat, antihuman monoclonal GM-CSF antibody diluted 1:500 in buffer C supplemented with 10% FCS, washed again, and then incubated for an additional 30 min at room temperature with 100 µl of avidin-peroxidase diluted 1:400 in buffer C supplemented with 10% FCS. Plates were washed again and developed with 100 µl ABTS substrate solution (0.55 mM 2,2'-azino- bis(3-ethylbenzthiazoline-6-sulphonic acid) and 0.1 M citric acid at pH 4.35 plus 0.1% vol/vol (30%) H₂O₂. GM-CSF was measured colorimetrically at 405 nm and quantified by interpolation from a standard curve constructed to known concentrations of GM-CSF. The detection limit of this assay is 16 pg /ml.

Protocol

To establish optimal conditions for interaction studies, concentration-response curves were constructed to dexamethasone and racemic albuterol (Figure 1). Monocytes were pretreated with dexamethasone and racemic albuterol for 20 and 5 min, respectively, prior to the addition of a submaximal concentration (EC₉₀) of LPS (3 ng/ml). The amount of immunoreactive TNF and GM-CSF released into the culture medium was subsequently measured at 18 h by sandwich ELISAs as described above.

View larger version (27K): [in this window] [in a new window]   Figure 1.   Inhibitory effect of dexamethasone and racemic albuterol on LPS-induced TNF (top panel ) and GM-CSF (bottom panel ) biosynthesis from human peripheral blood monocytes. Adherent cells were pretreated for 20 min with varying concentrations of the glucocorticosteroid, dexamethasone, or for 5 min with the 2-adrenoceptor agonist, albuterol, before being exposed to LPS (3 ng/ml). Cells were maintained at 37° C in a thermostatically controlled incubator under a 5% CO2 atmosphere, and the amount of TNF and GM-CSF released into the culture supernatant was measured at 18 h by sandwich ELISAs. The elaboration of cytokine under basal and LPS-stimulated conditions was 0.19 ± 0.07 and 2.91 ± 0.63 ng/ml, and 0.024 ± 0.14 and 0.52 ± 0.17 ng/ml for TNF and GM-CSF, respectively. Each data point represents the mean ± SEM of four to 11 determinations made from different cell preparations. See METHODS and Table 1 for further details.

For interaction studies, concentration-response curves to dexamethasone were constructed in the absence and presence of a low (1 µM) and relatively high concentration (100 µM) of albuterol (determined from the ₂-adrenoceptor agonist response curve). In these experiments, monocytes were incubated with albuterol for 5 and 180 min. Dexamethasone (0.03 to 300 nM) was then added for a further 20 mm followed by LPS (3 ng/ml). The amount of immunoreactive TNF and GM-CSF released into the culture medium was subsequently measured at 18 h by sandwich ELISAs as described above.

Drugs and Analytical Reagents

Percoll was obtained from Pharmacia/LKB (Milton Keynes, Buckinghamshire, UK), and LPS (from Salmonella enteritidis), albuterol, dexamethasone, avidin-peroxidase (code A3151) FCS, RPMI 1640, HBSS, and ABTS substrate solution were obtained from the Sigma Chemical Company (Poole, Dorset, UK). TNF and GM-CSF standards were obtained from British Biotechnology (code BDP 28) and Genzyme (code RH-CSF-C), respectively. TNF and GM-CSF quality control were from the National Institute for Biological Standards and Controls (codes 87/650 and 88/646, respectively). Rat antihuman GM-CSF monoclonal and biotinylated rat antihuman GM-CSF monoclonal antibodies (codes 18581D and 18592D, respectively) were purchased from Pharmingen/Cambridge Bioscience (Cambridge, UK), and mouse antihuman TNF (code C4 014), rabbit antihuman TNF (code IP 300), and alkaline-phosphatase-labeled sheep antirabbit IgG antibodies were purchased from Autogen-Bioclear (Calne, Wiltshire, UK), Genzyme Corporation (West Malling, Kent, UK) and Stratech Scientific Ltd (Luton, Bedfordshire, UK), respectively. The p-nitrophenyl phosphate assay kit (code 50-80-00) was from KPL/Dynatech Laboratories Ltd, (Billingshurst, Sussex, UK). All other reagents were from BDH (Poole, Dorset, UK).

Dissolution and Storage of Drugs

Stock solutions of albuterol and dexamethasone were dissolved in distilled water and subsequently diluted to the desired working concentration in enriched Dutch-modified RPMI 1640. LPS was dissolved in distilled water and stored at 70° C. Human recombinant TNF and GM-CSF for both quality controls and standards were obtained as lyophilized powders and reconstituted at 1 µg/ml in distilled water and stored at 70° C.

Data and Statistical Analyses

Data points, and values in the text and table, represent the mean ± SEM of "n" independent determinations. Concentration-response curves were analyzed by least-squares nonlinear iterative regression with the PRISM curve-fitting program (GraphPad Software, San Diego, CA) and log EC₅₀ values interpolated from computer-generated lines of best-fit. When appropriate, data were analyzed by Student's t test for paired data. The null hypothesis was rejected when p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have reported previously that exposure of human peripheral blood monocytes to LPS evokes a time- and concentration-dependent elaboration of TNF and GM-CSF (27). In the experiments described herein, LPS was used at a submaximal concentration of 3 ng/ml (~ EC₉₀) and TNF and GM-CSF were measured in the culture supernatant 18 h after the addition of the stimulus, at a time when cytokine release was maximal. Under the conditions used in this study, the generation of TNF and GM-CSF in response to LPS is absolutely dependent upon new protein synthesis (see 27 for further details).

As shown in Figure 1, treatment of human monocytes with dexamethasone (100 pM to 300 nM) inhibited LPS-induced TNF and GM-CSF generation in a concentration-dependent manner with equal potency (EC₅₀ values ~ 2 nM) (Table ). In all cases, a maximally effective concentration of dexamethasone (300 nM) was more effective at suppressing the release of GM-CSF (~ 90%) than of TNF (~ 75 to 80%) (see Table ). Albuterol (1 nM to 100 µM) also prevented LPS- induced TNF generation in monocytes, but it was approximately 30-fold less potent (EC₅₀ ~ 60 nM) (see Table ) than dexamethasone and, at a concentration of 100 µM, produced only 38% inhibition (Figure 1). In contrast, albuterol failed to prevent the ability of LPS to generate GM-CSF over the same concentration range (Figure 1).

Pretreatment of human monocytes with albuterol (1 and 100 µM) for 5 and 180 min inhibited LPS-induced TNF generation by approximately 25% (Figures 2 and 3), although in some experiments a greater effect was noted (Figure 2, lower panel ). However, regardless of the concentration of albuterol, or the time of pre-incubation, the inhibitory effect of dexamethasone was not significantly affected with respect to the EC₅₀ or the maximal effect produced (Figures 2 and 3 and Table ). Qualitatively identical data were obtained when GM-CSF release was used as an index of monocyte activation (Figures 4 and 5 and Table ).

View larger version (26K): [in this window] [in a new window]   Figure 2.   Lack of effect of albuterol on the inhibition of LPS-induced TNF biosynthesis from human peripheral blood monocytes by dexamethasone. Adherent cells were pretreated for 5 min with 1 µM (top panel ) and 100 µM (bottom panel ) albuterol followed by varying concentrations of the glucocorticosteroid, dexamethasone (0.03 to 300 nM for 20 min), before being exposed to LPS (3 ng/ml). Cells were maintained at 37° C in a thermostatically controlled incubator under a 5% CO2 atmosphere, and the amount of TNF released into the culture supernatant was measured at 18 h by a sandwich ELISA. Each data point represents the mean ± SEM of three to four determinations made from different cell preparations. See METHODS and Table for further details.

View larger version (26K): [in this window] [in a new window]   Figure 3.   Lack of effect of albuterol on the inhibition of LPS-induced TNF biosynthesis from human peripheral blood monocytes by dexamethasone. Cells were pretreated for 180 min with 1 µM (top panel ) and 100 µM (bottom panel ) albuterol followed by dexamethasone before being exposed to LPS (3 ng/ml). TNF was measured at 18 h by ELISA. Each data point represents the mean ± SEM of four determinations made from different cell preparations. See METHODS, legend to Figure 2, and Table for further details.

View larger version (27K): [in this window] [in a new window]   Figure 4.   Lack of effect of albuterol on the inhibition of LPS-induced GM-CSF biosynthesis from human peripheral blood monocytes by dexamethasone. Cells were pretreated for 5 min with 1 µM (top panel ) and 100 µM (bottom panel ) albuterol followed by dexamethasone before being exposed to LPS (3 ng/ml). GM-CSF was measured at 18 h by ELISA. Each data point represents the mean ± SEM of four determinations made from different cell preparations. See METHODS, legend to Figure 2, and Table for further details.

View larger version (27K): [in this window] [in a new window]   Figure 5.   Lack of effect of albuterol on the inhibition of LPS-induced GM-CSF biosynthesis from human peripheral blood monocytes by dexamethasone. Cells were pretreated for 180 min with 1 µM (top panel ) and 100 µM (bottom panel ) albuterol followed by dexamethasone before being exposed to LPS (3 ng/ml). GM-CSF was measured at 18 h by ELISA. Each data point represents the mean ± SEM of three to four determinations made from different cell preparations. See METHODS and Table for further details.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study no evidence was obtained to support the concept (8) that high concentrations of albuterol compromise the anti-inflammatory effect of glucocorticosteroids. Thus, dexamethasone failed to suppress the generation of TNF and GM-CSF from LPS-stimulated human monocytes; neither the EC₅₀ of dexamethasone nor the maximal inhibition of cytokine production achieved was altered significantly. This result was obtained regardless of the two concentrations of albuterol selected (which were maximal at 1 µM and supramaximal at 100 µM in this system, respectively, and probably in excess of the concentration achieved in vivo after inhalation) or the period of exposure prior to dexamethasone (5 and 180 min).

One pharmacologic explanation for this discrepancy is functional antagonism. Because albuterol reduced TNF generation per se, increased potency and effectiveness of dexamethasone might be expected. However, according to the hypothesis proposed by Peters and coworkers (8), this would be countered by the ability of the ₂-adrenoceptor agonist to reduce the interaction of the steroid-bound GR with DNA and so limit repression of the TNF gene. To address this possibility, identical experiments were conducted measuring LPS- induced GM-CSF release. The advantage of using this marker of activation is its insensitivity to albuterol, which eliminates complications that can arise from functional antagonism. As shown in Figures 4 and 5, albuterol was unable to antagonize the inhibitory effect of dexamethasone on GM-CSF release at either concentration or pretreatment period studied, which excludes unequivocally a role for functional antagonism.

The results of this study are enigmatic and difficult to reconcile with data obtained in rat and human lung where, at the molecular level, albuterol and fenoterol decreased the binding of the GR to DNA by a mechanism that appears to be due to the activation of the transcription factor, CREB (8), and would be expected to limit the degree of gene repression effected by the steroid. The reason for this discrepancy is unclear, but several explanations are worthy of consideration, including the possibilities that it is a cell-specific phenomenon or dependent upon the nature of the activating stimulus (see 28). In this respect, recent studies have suggested that the ability of transcription factors to interact with DNA might depend upon the structure of the associated chromatin (29). Modulation of gene transcription by CREB occurs via an interaction with the so-called "coactivators" CBP or p300. These are large proteins that integrate signals from diverse stimuli via other transcription factors, including activator protein-1, nuclear factor B, and signal transducer and transactivator proteins, activators of cyclic AMP-dependent protein kinase and GR (30). CBP contains a functional histone acetylase activity that permits modulation of chromatin structure and subsequent induction of proinflammatory genes (31). Once bound to CBP, GR receptors cause a decrease in its intrinsic acetylase activity, resulting in compression of the chromatin structure and repression of gene transcription (32). Thus, competition for binding sites on CBP may result in varying acetylase activities, which would have a subsequent "knock-on" effect on gene transcription. Indeed, there is evidence that synergistic and antagonistic interactions can occur in different cell types. For example, Pennie and coworkers (29) described a synergistic interaction between the cell permeant 8-bromo analog of cyclic AMP and dexamethasone when transcription was measured in a transiently transfected reporter gene assay in which chromatin did not associate with the DNA. In contrast, when cells were stably transfected with the same reporter gene that is incorporated into the genome and thus influenced by chromatin, 8-bromo cyclic AMP antagonized the effect of the steroid (29). These data clearly imply that the way any of two transcription factors interact could depend on cell type, and that the nature of the association is probably dictated by the complement of different endogenous transcription factors expressed and/or their relative amounts, thereby providing multiple possible interactions. Another possibility is that monocytes taken from the blood of normal subjects and those harvested from asthmatic subjects do not respond to ₂-adrenoceptor agonists and steroids in the same way. This is a reasonable proposal given the significant differences in other parameters that have been reported, including the release of interleukin-1, acid phosphatase, and nonspecific esterases (32) and the expression of certain cell-surface receptors (33). Finally, an alternative consideration is that the ability of ₂-agonists to prevent the binding of the activated GR complex to DNA demonstrated at the molecular level may have no physiologic/pathophysiologic relevance and may not account for or contribute to the adverse effect of ₂-adrenoceptor agonists when administered chronically to asthmatic subjects (7).