Ligand-independent Activation of the Glucocorticoid Receptor by β2-Adrenergic Receptor Agonists in Primary Human Lung Fibroblasts and Vascular Smooth Muscle Cells*

The glucocorticoid receptor (GR) is a ubiquitously expressed transcription factor present in most cell types. Upon ligand binding, the GR is activated and translocates into the nucleus where it transmits the anti-inflammatory actions of glucocorticoids. Here, we describe the ligand-independent activation of GR by the β2-adrenergic receptor (β2-AR) agonists, salbutamol and salmeterol, in primary human lung fibroblasts and vascular smooth muscle cells. Immunohistochemistry demonstrated expression of GR and the β2-AR by fibroblasts and vascular smooth muscle cells. Treatment of the cells with the β2-AR agonists, salbutamol or salmeterol, resulted in translocation of GR into the nucleus beginning at 30 min, as shown by immunohistochemistry and Western blotting of cytosolic and nuclear cell extracts. In comparison, activation of GR induced by the corticosteroids dexamethasone and fluticasone occurred at the same time after treatment (30 min) but resulted in a more complete depletion of GR from the cytosolic compartment. Electrophoretic mobility shift assays confirmed that nuclear GR, activated by both β2-AR agonists and glucocorticoids, actively bound to the GR consensus sequence (GR element). Functional activation of the GR was confirmed by a Luciferase reporter gene assay, using a GR driven promoter fragment from the p21(WAF1/CIP1) gene. The effects of the β2-AR agonists, salbutamol and salmeterol, were dependent upon binding to the β2-AR, because blocking of β2-AR with propranolol abrogated GR activation. GR activation appeared to involve cAMP. In summary, these data show that β2-AR agonists are potent activators of GR. Ligand-independent activation of GR by β2-AR agonists may substantially mediate the anti-inflammatory actions of these drugs observed in vitroand in vivo.

The glucocorticoid receptor (GR) 1 belongs to the family of intracellular ligand-inducible transcription factors termed the steroid/vitamin D/retinoic acid superfamily (1). All members of this superfamily share essential structural and functional features, which are an amino-terminal transactivation domain, a central zinc-finger DNA binding domain, and a carboxyl-terminal ligand binding domain. Unliganded GR is present in the cytosol and is associated with a large multiprotein complex of chaperones, including heat shock proteins Hsp90, Hsp70, and Hsp56 (1)(2)(3)(4)(5). This conformation is essential for allowing GR to bind steroid ligands with high affinity. Ligand binding induces a conformational change of the complex leading to dissociation of the chaperones. Activated GR then translocates to the nucleus and binds to its specific target DNA sequences termed the glucocorticoid response elements (GRE) (5,6). Depending on the structure of the affected gene promoter, glucocorticoids thereby lead to either increased or decreased gene transcription (6 -8). Biological effects of glucocorticoids on most human cell types are generally anti-inflammatory, characterized by decreased expression of inflammatory mediators or increased expression of protective mediators (1)(2)(3)(4).
A frequent inflammatory disease that requires the administration of glucocorticoids is asthma (9,10). In asthma, treatment regimen combining glucocorticoids with ␤ 2 -agonists results in better symptom control and lesser airway inflammation than simply increasing the dose of glucocorticoids (11)(12)(13). These clinical observations strongly suggest an interaction of both classes of drugs at a molecular level. In contrast to GR, a soluble intracellular receptor that resides in the cytosolic compartment when unliganded (1)(2)(3)(4), the ␤ 2 -adrenergic receptor (␤ 2 -AR) constitutes a plasma membrane-anchored, G-protein-coupled receptor with seven transmembrane spanning domains (14,15). The ␤ 2 -AR is highly expressed in fibroblasts, vascular smooth muscle cells (USMCs), and epithelial cells of the human lung (14,16). Signal transduction by ␤ 2 -AR occurs upon ligand binding via activation of adenylate cyclase, which increases the concentration of intracellular cAMP (14). In addition to the short term effects of ␤ 2 -AR agonists, such as VSMC relaxation (17), considerable evidence suggests that ␤ 2 -AR agonists have potent anti-inflammatory effects in vitro and in vivo (11, 18 -20). Because activation of GR is essentially responsible for the majority of anti-inflammatory effects, we thus addressed the question of whether ␤ 2 -agonists could lead to activation of GR in a ligand-independent manner.
In the present study, cultures of primary human lung fibroblasts and vascular smooth muscle cells were used to demonstrate that both glucocorticoids and ␤ 2 -AR agonists activate the human GR. We show that the kinetics of GR activation are similar for both substances, leading to rapid nuclear translocation of functional activated GR after 30 min. Compared with ␤ 2 -AR agonists, the endogenous ligands for GR, glucocorticoids, led to complete depletion of GR from the cytosolic compartment. Activation of GR by ␤ 2 -AR agonists was mediated by the interaction of the drugs with the ␤ 2 -AR because a ␤ 2 -AR antagonist, propranolol, abolished the observed activation of GR in a concentration-dependent manner. Similarly to the effect of the drugs on GR activation, addition of dibutyryl-cAMP or 8-bromo-cAMP activated the GR, suggesting that the action of ␤ 2 -AR on GR activation involves the adenylate cyclase/cAMP pathway. Thus, our results suggest that ligand-independent activation of GR by ␤ 2 -AR agonists may be an essential mechanism contributing to the anti-inflammatory effects evoked by ␤ 2 -agonists.
Cell Culture-Primary cell lines of fibroblasts (n ϭ 3) or VSMC (n ϭ 3) were established from human lung tissue biopsies obtained from patients undergoing lobectomy or pneumectomy, as described previously (21). Fibroblasts were cultivated in RPMI 1640 supplemented with 10% FCS, 8 mM L-glutamine, and 20 mM HEPES. VSMC were cultivated in minimal essential medium, supplemented with 5% FCS, 8 mM L-glutamine, 1% minimal essential medium vitamin mix, and 20 mM HEPES. For experiments, cells were grown on 150-mm cell culture dishes until 80% confluent. Cells were then growth-arrested by serum starvation in low serum medium (0.1% FCS) for 48 h prior to stimulation with the indicated drugs (dexamethasone, formeterol, salmeterol, salbutamol, propranolol, 8-bromo-cAMP, or dibutyryl-cAMP). Low serum medium was replaced every 12 h. No antibiotics or antimycotics were added to the culture conditions at any time.
Immunohistochemistry-Expression of GR and ␤ 2 -AR by fibroblasts or VSMC was assessed by immunohistochemistry, as described previously (22). Cells were seeded onto 8-well chamber slides at 80% confluency and growth-arrested for 24 h. After washing three times with ice-cold phosphate-buffered saline (PBS), cells were fixed in PBS containing 4% formaldehyde. Cells were washed three times in PBS, and endogenous peroxidase activity was blocked by incubating the cells for 1 h in 80% methanol with 0.6% hydrogen peroxide. After washing three times in PBS, cells were incubated with whole goat serum to prevent unspecific binding of primary antibodies. Primary antibodies (1:50 each, in PBS) were added to the cells for 1 h. Cells were washed three times in PBS and incubated with the secondary, biotinylated antibody for 30 min. After washing the cells three times in PBS, avidin and biotinylated horseradish peroxidase was added and specific binding was visualized by staining with 3-amino-9-ethyl carbazole. Unspecific binding of the secondary antibody was examined by following the outlined protocol, without the addition of primary antibodies. Cells were counterstained with Mayer's hemalum (22).
Preparation of Cytosolic and Nuclear Extracts-Nuclear and cytosolic extracts from stimulated or unstimulated control cells were prepared at the indicated time points as originally described by Dignam et al. (23). Cells were washed twice in ice-cold PBS and harvested in 1 ml of PBS. The samples were centrifuged for 30 s at 10,000 ϫ g, and cell pellets were resuspended in 50 l of low salt buffer (20 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM NaVO 4 , 1 mM EDTA, 1 mM EGTA, 0.2% Nonidet P-40, 10% glycerol, supplemented with a set of proteinase inhibitors, Complete TM ). After 10 min of incubation on ice, the samples were centrifuged at 13,000 ϫ g for 2 min (4°C), and the supernatants were taken as cytosolic extracts. Nuclei were resuspended in high salt buffer (20 mM HEPES, pH 7.9, 420 mM NaCl, 10 mM KCl, 0.1 mM NaVO 4 , 1 mM EDTA, 1 mM EGTA, 20% glycerol, supplemented with Complete TM ), and nuclear proteins were extracted by shaking on ice for 30 min. Samples were then centrifuged at 13,000 ϫ g for 10 min (4°C), and the super-natants were taken as nuclear extracts.
Western Blot Analysis-For Western blotting, cells were seeded onto 150-mm cell culture dishes and allowed to reach 80% confluence. After 48 h of serum starvation, cultures were stimulated with the indicated concentrations of dexamethasone, fluticasone, salbutamol, or salmeterol and harvested at the indicated time points. Expression and localization of glucocorticoid receptor was determined in cytosolic and nuclear extracts of the cells by Western blot analysis using gradient SDS-polyacrylamide gels (4 -15%) as described earlier (24). Aliquots of cytosolic and/or nuclear extracts along with prestained molecular weight markers were applied to the gels and run at 25 mA constant current for 4 h at room temperature. After electrophoresis, proteins were electroblotted on Immobilon-P transfer membranes for 90 min with 1 mA/cm 2 at room temperature. Membranes were blocked in 5% skimmed milk in TBS-Tween (10 mM Tris, 150 mM NaCl, 0.05% Tween 20, pH 8.0) for 1 h at room temperature. After blocking, membranes were incubated with an antibody specific to GR at 4°C overnight. The following day, membranes were washed three times with TBS-Tween and then incubated with the secondary, peroxidase-coupled antibody at a dilution of 1:5000 for 1 h at room temperature. Membranes were then washed three times in TBS-Tween, and specific bands were visualized using an ECL system according to the manufacturer's instructions.
Electrophoretic Mobility Shift Assays-DNA mobility shift assays were performed as originally described by Sen and Baltimore (25). Oligonucleotides comprising the consensus sequences for GR (5Ј-AAG ATT CAG GTC ATG ACC TGA GGA GA-3Ј) (GRE) were end-labeled with [␥-32 P]ATP using T4-polynucleotide-kinase. Aliquots of nuclear extracts (1 g) were incubated with the labeled consensus oligonucleotides under binding conditions (4% glycerol, 1 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM dithiotreitol, 50 mM NaCl, 10 mM Tris-HCl, pH 7.5, 50 g/ml poly(dI-dC)) in a total volume of 10 l. Reactions were carried out at room temperature for 30 min, and protein-DNA complexes were analyzed on a 4% polyacrylamide gel. The identity of GR was confirmed by addition of competitive unlabeled consensus sequence oligonucleotides or by addition of monoclonal antibody specific for GR.
Luciferase Reporter Gene Assay-Two days before transfection, cells were seeded into 12-well plates (1 ϫ 10 4 cells/well) and were serumdeprived for 24 h. Cells were then subjected to liposomal transfection using the cationic lipid Tfx-50 at a DNA to lipid ratio of 1:3 (0.5 g of plasmid/well). A plasmid containing a GR-driven promoter fragment of the human p21 (WAF1/CIP1) gene, WWP-Luc subcloned in front of a Luciferase gene was kindly provided by Prof. Dr. B. Vogelstein (Johns Hopkins University, Baltimore, MD) (26). Transfection was carried out in the absence of FCS for 2 h at 37°C in a 100% humidified atmosphere. Cells were then incubated with either salmeterol, salbutamol, or dexamethasone. After 36 h cells were washed twice with ice-cold PBS and lysed, and equal amounts of lysates were analyzed for Firefly luciferase expression. In brief, 10-l aliquots of cell lysates were mixed with 50 l of luciferase reagent buffer, and luminescence of the samples was integrated over a time period of 10 s in a LUMAC Biocounter M1500P (Landgraaf, Netherlands).
Statistical Analysis-For statistical analysis Student's t test and ANOVA analysis were performed. A p value Ͻ0.05 was estimated significant.
Ethic Committee Approval-The protocol for establishing primary human cell cultures from biopsies obtained during surgery was approved by the ethical committee of the Faculty of Medicine, University Hospital Basel (approval number M75/97).

Fibroblasts and VSMC Express the Glucocorticoid and ␤ 2 -
Adrenergic Receptors-To analyze whether primary human lung fibroblasts or VSMC, in culture, express ␤ 2 -AR or GR, we performed immunohistochemistry of the cells with antibodies specific for ␤ 2 -AR or GR We found that GR was expressed by both cell types. Immunohistochemical staining for GR in serum-deprived cells was predominantly cytoplasmic with no staining of nuclei (Fig. 1A) and was most prominent in a perinuclear compartment of the cells (arrows, Fig. 1A), indicating that unliganded GR resides in close proximity to the nucleus. No signal was obtained when the secondary antibody alone was applied (Fig. 1C).
As demonstrated in Fig. 1B, primary human fibroblasts expressed the ␤ 2 -AR. Clear expression of the ␤ 2 -AR is shown with the staining being localized at the plasma membrane (arrows, Fig. 1B), in contrast to the lack of signal when the secondary antibody alone was applied (Fig. 1C). Staining for ␤ 2 -AR was not exclusively found at the plasma membrane, an observation that may indicate the cytoplasmic recircularization of ␤ 2 -AR as recently described (14). Similar observations were made using VSMC.
␤ 2 -Agonists Activate the Glucocorticoid Receptor-Activated GR rapidly disappears from the cytoplasmic compartment and translocates to the nucleus (1)(2)(3)(4)(5). To analyze activation of GR, we assessed the cytosolic depletion and nuclear translocation of GR by immunohistochemistry and Western blotting of cytosolic and nuclear extracts. We determined the effects of two ␤ 2 -AR agonists, salbutamol and salmeterol, and compared their effects to two glucocorticoids, fluticasone and dexamethasone, serving as positive controls.
Immunohistochemical analysis revealed that incubation of cells with dexamethasone ( Fig. 2C) or fluticasone (Fig. 2D) for 4 h, with both glucocorticoids at a concentration of 10 Ϫ8 M, induced complete translocation of the GR from the cytosol into the nucleus. Interestingly, incubation of cells with the ␤ 2 -AR agonists, salmeterol (Fig. 2E) or salbutamol (Fig. 2F), both at a concentration of 10 Ϫ8 M, also resulted in translocation of the GR to the nucleus, 4 h after addition of the drugs. However, the effect of ␤ 2 -AR agonists seemed to be not as complete as observed in the presence of glucocorticoids These results suggest a ligand-independent activation of GR by ␤ 2 -AR agonists mediated via ␤ 2 -AR. Fig. 3A gives additional evidence that dexamethasone and fluticasone rapidly induced depletion of GR from the cytosolic, as shown by Western blotting. Depletion was rapid occurring 30 min after treatment of the cells and completed with no GR left in the cytosolic compartment between 30 min and 4 h after addition of the glucocorticoids.
Interestingly, the two ␤ 2 -AR agonists, salmeterol and salbutamol, induced depletion of GR from the cytosolic compartment when added to the cell cultures (Fig. 3B) in a time-dependent manner. Activation of GR by the ␤ 2 -AR agonists, salmeterol and salbutamol, began at 30 min and continued over a time period of 4 h. Compared with glucocorticoid-induced translocation of GR, the ␤ 2 -AR agonist-mediated depletion of GR from the cytosolic fraction was not complete. Low amounts of GR were still detected in the cytosol of the cells after 4 h. Thus, ␤ 2 -AR agonists seemed to provoke a more sustained signal that leads to prolonged activation of GR, as compared with the rapid onset and complete signal provided by glucocorticoids.
As shown in Fig. 3C, salmeterol (Fig. 3C, lane 2) and fluticasone (Fig. 3C, lane 3), at 1 h, led to significant depletion of GR from the cytosolic compartment (Fig. 3C, lane 1). The glucocor-FIG. 1. Expression of ␤ 2 -AR and GR in primary human lung fibroblasts. Human lung fibroblasts were cultivated from human tissue biopsies and seeded onto 8-well chamber slides. Cells were fixed and subsequently incubated with rabbit polyclonal antibody specific for human GR (Fig. 1A) or human ␤ 2 -AR (Fig. 1B). Background staining was assessed by incubation with secondary antibody alone (Fig. 1C). Immunoreactivity was visualized using an avidin-biotin peroxidase stain. Pictures are representative for cell lines of primary human fibroblasts (n ϭ 3) and vascular smooth muscle cells (n ϭ 3).

FIG. 2. Immunostaining analysis of the cellular distribution of GR in human primary fibroblasts.
Serum-deprived human primary fibroblasts expressed GR in a nucleus-associated compartment but not in the nucleus (Fig. 2A). Unspecific staining of the second antibody was excluded when cells were only incubated with the second antibody (Fig.  2B). When cells were cultivated in the presence of either dexamethasone (Fig. 2C) or fluticasone (Fig. 2D), the GR translocated into the nucleus. GR was also translocated to the nucleus when cells were treated with salmeterol ( Fig. 2E) or salbutamol (Fig. 2F). All drugs were used at a concentration of 10 Ϫ8 M, and experiments were repeated in three different fibroblast and two VSMC lines. ticoid fluticasone was clearly more potent than the ␤ 2 -AR agonist salmeterol. The depletion of GR from the cytosol coincided with an increase in nuclear GR of the respective samples (Fig.  3C, lanes 4 -6). Compared with untreated cells (Fig. 3C, lane 4), salmeterol-treated (Fig. 3C, lane 5) and fluticasone-treated (Fig. 3C, lane 6) cells exhibit a significantly higher content of GR in nuclear extracts.
Translocated Glucocorticoid Receptor Binds to Its Consensus Sequence-To assess whether GR that is translocated into the nucleus in response to both classes of drugs was capable of DNA binding, we performed electrophoretic mobility shift assays (EMSA) of nuclear cell extracts using an oligonucleotide comprising the GR consensus sequence, the GRE. Fig. 4 depicts characteristic EMSA demonstrating GRE binding activity within nuclear extracts from glucocorticoid-treated cells (Fig.  4A) as well as from ␤ 2 -AR agonist-treated cells (Fig. 4B). Untreated cells (0 h) served as a control and exhibited weak binding activity for the 32 P-labeled GRE oligonucleotide, suggesting small amounts of active GR being present in the nuclei of serum-starved cells. This observation was in accordance with immunohistochemistry and Western blots, both demonstrating that low amounts of GR were present within the nucleus of fibroblasts and VSMC.
When cells were stimulated with fluticasone or dexamethasone, GR was rapidly activated as observed by its increased binding to the GRE oligonucleotides. This observation supports the nuclear translocation of GR as seen in immunohistochemistry and Western blot analyses. GR activation was observed as early as 30 min after stimulation with glucocorticoids (Fig. 4A).
A similar effect was noted with salmeterol-or salbutamoltreated cells (Fig. 4B). Both ␤ 2 -AR agonists induced rapid activation of GR after 30 min with a subsequent decline of GRE binding activity thereafter.
The identity of GRE binding activity in nuclear extracts of the cells was confirmed by adding unlabeled competitor GRE oligonucleotides or antibodies specific for GR to the samples (Fig. 5). Addition of unlabeled competitor GRE oligonucleotides led to complete disappearance of the GR⅐GRE complex (Fig. 5,  A, lane 3, and C, lane 3). When the samples were incubated with monoclonal antibodies to GR, specific bands diminished in a concentration-dependent manner (Fig. 5A, lanes 3-6). Thus, these observations together identified GRE binding activity within nuclear extracts of the cells as GR.
Similar to the results obtained with ␤ 2 -agonists both 8-bromo-cAMP or dibutyryl cAMP induced a DNA mobility shift for GRE (data not shown), indicating that the mechanism for GR activation by ␤ 2 -agonists may be due to an increase of intracellular cAMP levels (Fig. 5C, lane 6). The stimulatory effect of cAMP on GR activation was dose-dependent and occurred at physiologically relevant concentrations (data not shown). As shown in Fig. 5C (lane 7) a protein kinase A (PKA) inhibiting peptide partly reduced the salmeterol-induced activation of GR. This may further suggests the involvement of the cAMP/ Cytosolic and nuclear extracts were prepared, and equal aliquots were separated on gradient 4 -15% SDS-polyacrylamide gel electrophoresis gels and transferred on nitrocellulose membranes. GR (92 kDa) was detected after incubation with rabbit polyclonal anti-GR antibody (Santa Cruz) following standard protocol for ECL Western detection (Amersham). In C, activation of GR is shown by parallel detection of GR in cytosolic (lanes 1-3) and nuclear extracts (lanes 4 -6). Extracts were prepared at 30 min as described for A and B. Extracts from untreated cells (lanes 1 and 4) were applied together with extracts from salmeterol-treated (10 nM) (lanes 2 and 5) and fluticasone-treated cells (lanes 3 and 6). Nuclear extracts were prepared at the indicated time points as described under "Experimental Procedures." Equal aliquots of nuclear extracts were incubated with [␥-32 P]ATP end-labeled GRE oligonucleotides, and the formation of specific GR⅐GRE complexes was analyzed on 4% polyacrylamide gels. Free probe without nuclear extracts was applied as negative control. FR, fragment.

PKA pathway.
GR Activation by ␤ 2 Agonists Depends on Their Interaction with the ␤ 2 -Adrenergic Receptor-We analyzed whether the described effects of ␤ 2 -AR agonists were dependent upon binding of the drugs to the ␤ 2 -AR by preincubating the cells with the ␤ 2 -AR antagonist, propranolol. When propranolol was administered to the cells 30 min prior to stimulation with salmeterol (10 Ϫ8 M), GR activation was dose-dependently inhibited as shown by EMSA (Fig. 5A). Propranolol at a concentration of 10 Ϫ7 M (Fig. 5A, lanes 4 -7) clearly inhibited GR activation by salmeterol (Fig. 5A, lane 2), whereas propranolol at 10 Ϫ9 M (Fig. 5A, lane 4) had no effect on salmeterol-induced GR activation. The data thus demonstrate that functional interaction of ␤ 2 -agonists with their respective receptor, the ␤ 2 -AR, is responsible for ␤ 2 -AR agonist-induced activation of GR.
Translocated Nuclear GR Is Functional-To assess whether GR activated by glucocorticoids or ␤ 2 -AR antagonist was functional, we assessed whether the drugs affected a glucocorticoid driven p21 (WAF1/CIP1) promoter/Luciferase construct (26). At concentrations higher than 10 Ϫ9 M dexamethasone activated the reporter gene p21 (WAF1/CIP1) . Although p21 (WAF1/CIP1) activation was inconsistent at a dexamethasone concentration of 10 Ϫ8 M, it was constantly expressed at a concentration of either 10 Ϫ7 M or 10 Ϫ6 M (data not shown). The achieved expression of p21 (WAF1/CIP1) with dexamethasone at 10 Ϫ7 M was about 197 Ϯ 29% of control (Fig. 6). Similar to the glucocorticosteroid the two ␤ 2 -AR antagonists, salmeterol (136 Ϯ 6%) and salbutamol (149 Ϯ 21%), induced the expression of Firefly luciferase in a concentration range of 10 Ϫ8 -10 Ϫ6 M. In accordance to the above described results on GR activation determined by Western blotting and EMSA compared with the effect of glucocorticoid treatment the activation of p21 (WAF1/CIP1) was less prominent in the presence of both salmeterol or salbutamol (Fig. 6). Thus, although not as potent as the glucocorticoids, the ␤ 2 -AR agonists clearly activated the p21 (WAF1/CIP1) promoter, suggesting that GR activation by ␤ 2 -AR agonists also leads to altered gene transcription.

DISCUSSION
The mechanism of GR activation in a ligand-independent manner has been the subject of several recent investigations (27,28). In this study, we demonstrate ligand-independent activation of GR by two ␤ 2 -agonists, salmeterol and salbutamol. Both drugs were potent activators of GR in the absence of agonistic ligands as demonstrated by immunohistochemistry, Western blotting, EMSA, and reporter gene assays. GR was rapidly depleted from the cytosol of primary human lung fibroblasts and VSMC and translocated into the nucleus. Translocated GR was functional as it bound to its specific DNA recognition sequence, GRE, and also activated a GR-inducible luciferase reporter gene assay, p21 (WAF1/CIP1) . Finally, GR activation by ␤ 2 -agonists was dependent upon their binding to ␤ 2 -AR, reasoning that downstream signaling initiated by ␤ 2 -AR agonist/␤ 2 -AR interaction is responsible for the observed GR activation.
Activation of GR results in altered transcription of several cytokine genes involved in the inflammatory process leading to their repression or activation (6 -8). Our findings are especially important to understand the underlying molecular mechanism of drug action in the treatment of asthma. Here, glucocorticoids and ␤ 2 -AR agonists are the most effective drugs in the treatment of this disease (9,10). Clinical studies on asthmatic individuals suggested that the administration of salmeterol, in combination with glucocorticoids, resulted in an improved symptom control being more effective than increasing the dose of the glucocorticoid alone (11)(12)(13). We therefore assumed that an interaction between both substances might occur at the level of GR activation, intensifying the anti-inflammatory potency of glucocorticoids. It is known that glucocorticoids induce transcription of the ␤ 2 -AR gene (29,30), preventing agonistinduced desensitization of the ␤ 2 -AR itself and improving the therapeutic efficacy of ␤ 2 -AR agonists. It is unknown, however, whether ␤ 2 -agonists are capable of affecting GR activity.
An earlier report (31) analyzing the cellular effects of ␤ 2 -AR agonists in rat lung tissue cubes demonstrated an inhibitory effect of unphysiologically high concentrations (10 Ϫ6 M) of salbutamol on the binding of GR from crude tissue homogenates to their recognition sequence, GRE. This report failed to show direct activation and/or inhibition of GR itself. In our study, we compared GR activation by ␤ 2 -AR agonists with GR activation by specific ligands, two glucocorticoids, in a definite cell population, either primary human lung fibroblasts or VSMC. The comparison between ligand-dependent (induced by glucocorticoids) and ligand-independent (induced by ␤ 2 -AR agonists) GR activation demonstrated that both mechanisms occurred in a similar time frame. Ligand-dependent activation of GR was apparently more potent and resulted in total depletion of GR of the cytosolic compartment. In contrast, ligand-independent activation of GR was sustained over a longer period of time. The cellular function of ligand-independent activation of GR was illustrated by a luciferase reporter gene, WWP-Luc (21). The WWP-Luc construct contains the promoter region (Ϫ2.4 kilobases to ϩ1 base pair) of the p21 (WAF1/CIP1) gene. This construct was shown to be activated by glucocorticoids (32). p21 (WAF1/CIP1) is a cell cycle kinase inhibitor and accounts for some of the antiproliferative effects of glucocorticoids in fibroblasts (33).
The biochemical modulation of GR is suggested to be achieved by phosphorylation (activation) and dephosphorylation (inactivation) at seven different phosphorylation sites (34 -39). How can activation of GR by ␤ 2 -AR agonists be explained at the molecular level? In general, signal transduction upon interaction of ␤ 2 -AR agonists with the ␤ 2 -AR results in the activation of G-proteins that are coupled to the intracellular domain of the receptor (14,40). Although we could not demonstrate the immediate upstream event preceding GR activation in response to ␤ 2 -AR agonists, we assume an event in the ␤ 2 -AR agonist signal transduction pathway to be responsible for the observed GR activation. This is evidenced by the fact that the ␤ 2 -AR antagonist, propranolol, prevented ␤ 2 -AR agonist-induced GR activation. ␤ 2 -AR activation leads to an increase of intracellular cAMP (40), PKA (14,40), and calmodulin (CaM) (41). We assessed the effect of two cAMP mimetics, which also activated GR. These findings suggested a possible involvement of the known ␤ 2 -AR agonist-mediated cAMP pathway, which was further supported by the finding that a PKA inhibiting peptide abolished the salmeterol-induced activation of GR. To the contrary, CaM, a ubiquitous intracellular signaling molecule located to plasma membrane receptors and ion channels (42), directly activated GR in a ligand-independent manner (43). The mechanism responsible for GR activation by CaM was suggested to be a phosphorylation of specific tyrosine residues of GR (43). Analogously, CaM-dependent phosphorylation of the estrogen receptor, another member of the steroid/ vitamin D/retinoic acid superfamily, has been reported to correlate with the activation of the estrogen receptor itself (44). It is therefore possible that ␤ 2 -AR agonists activate GR involving the action of CaM in a similar way.
In conclusion, our study demonstrates, for the first time, ligand-independent activation of GR by two different ␤ 2 -agonists using primary human cell lines of lung fibroblasts and VSMC. GR activation by ␤ 2 -agonists explains, at least in part, the as yet unresolved anti-inflammatory potency of ␤ 2 -agonists seen in vivo and in vitro.