Chronic systemic administration of salmeterol to rats promotes pulmonary β₂-adrenoceptor desensitization and down-regulation of G_sα

Abstract

1. The aim of the present study was to examine the effects of chronic infusion of the long-acting agonist salmeterol on pulmonary β₂-adrenoceptor function in Sprague-Dawley rats in vivo and to elucidate the molecular basis of any altered state.

2. Systemic administration of rats with salmeterol for 7 days compromised the ability of salmeterol and prostaglandin E₂ (PGE₂), given acutely by the intravenous route, to protect against ACh-induced bronchoconstriction when compared to rats treated identically with vehicle.

3. β₁- and β₂-adrenoceptor density was significantly reduced in lung membranes harvested from salmeterol-treated animals, which was associated with impaired salmeterol- and PGE₂-induced cyclic AMP accumulation ex vivo.

4. Three variants of G_sα that migrated as 42, 44 and 52 kDa peptides on SDS polyacrylamide gels were detected in lung membranes prepared from both groups of rats but the intensity of each isoform was markedly reduced in rats that received salmeterol.

5. The activity of cytosolic, but not membrane-associated, G-protein receptor-coupled kinase was elevated in the lung of salmeterol-treated rats when compared to vehicle-treated animals.

6. The ability of salmeterol, administered systemically, to protect the airways of untreated rats against ACh-induced bronchoconstriction was short-acting (t_off ∼45 min), which contrasts with its long-acting nature when given to asthmatic subjects by inhalation.

7. These results indicate that chronic treatment of rats with salmeterol results in heterologous desensitization of pulmonary G_s-coupled receptors. In light of previous data obtained in rats treated chronically with salbutamol, we propose that a primary mechanism responsible for this effect is a reduction in membrane-associated G_sα. The short-acting nature of salmeterol, when administered systemically, and the reduction in β-adrenoceptor number may be due to metabolism to a biologically-active, short-acting and non-selective β-adrenoceptor agonist.

Introduction

Repeated administration of high-dose β₂-adrenoceptor agonists renders susceptible individuals tolerant to their beneficial effects in asthma. With regard to long-acting agonists, there is evidence that tolerance develops both to their dilator and protective effects in the airways. Lipworth and colleagues (Newnham et al., 1994; 1995) have reported that 4 weeks treatment of asthmatic subjects with inhaled eformoterol and salmeterol produced a rightwards shift in the bronchodilator dose-response curve for both agonists. This state of tolerance was reflected in terms of a decrease in forced expiratory volume in 1 s (FEV₁) and forced expiratory flow at 25% and 75% of vital capacity (FEF_{25 – 75}), indicating that subsensitivity of β₂-adrenoceptors to salmeterol and eformoterol had occurred in the large and small airways. Equally, tolerance to the protective effect of salmeterol against methacholine- and exercise-induced bronchoconstriction has been documented in mild to moderate asthmatic subjects after chronic administration (Boulet et al., 1998; Cheung et al., 1992; Drotar et al., 1998; Simons et al., 1997).

The molecular aetiology of β₂-adrenoceptor sub-sensitivity in asthma is unclear. Airways resected from asthmatic patients fail to relax normally to isoprenaline, supporting a possible defect in β-adrenoceptor function (Bai et al., 1992; Cerrina et al., 1986; Goldie et al., 1986), although it is equivocal whether this is the result of treatment or a consequence of the disease process itself. Nevertheless, down-regulation of β-adrenoceptor number in lung and airways smooth muscle has been reported in animals given isoprenaline and noradrenaline chronically by infusion (Nerme et al., 1990; Nishikawa et al., 1993; 1994) and this is associated with a reduced functional responsiveness towards β-adrenoceptor agonists ex vivo (Nerme et al., 1990; Nishikawa et al., 1994). Thus, it is possible that compromised bronchodilatation and loss of protection against various bronchoconstrictor challenges in humans is due, at least in part, to pulmonary β₂-adrenoceptor desensitization.

Two major molecular mechanisms have been delineated in isolated cells that result in short-term β₂-adrenoceptor desensitization. One of these promotes homologous refractoriness and involves the uncoupling of the agonist-occupied form of the receptor from the stimulatory guanine nucleotide binding protein, G_s, by mechanisms that require phosphorylation of serine and threonine residues at the carboxy-terminus of the agonist-occupied receptor (Fredericks et al., 1996; Premont et al., 1995). This reaction can be catalysed by at least three members of the G-protein receptor-coupled kinase (GRK) superfamily family, including GRK2 and GRK3 (Krupnick & Benovic, 1998; Premont et al., 1995). The subsequent binding of β-arrestin, a soluble protein that prevents further coupling to G_s (Lohse et al., 1990), then halts signalling through the receptor. Short-term desensitization is also effected by cyclic AMP-dependent protein kinase (PKA) following phosphorylation of serine and threonine residues present within the third intracellular loop of the protein in response to an increase in intracellular cyclic AMP (Lohse et al., 1990). In contrast, prolonged periods of desensitization can involve physical internalization and subsequent degradation of receptors (Lohse, 1993) due to an inhibition of transcription and/or increased post-transcriptional processing of β₂-adrenoceptor mRNA (Lohse, 1993; Mayor et al., 1998). In addition, another less-well characterized process is a reduction of membrane-associated G_sα (Milligan, 1993), although the functional relevance of this process has not been rigorously explored in vitro or in vivo.

Although desensitization of G-protein-coupled receptors (GPCR) has been studied extensively, most of the information to date has been gathered from cultured cell systems and the extent to which this applies to the in vivo situation is little investigated. Here we describe pulmonary β₂-adrenoceptor desensitization in rats treated chronically with the long-acting β₂-adrenoceptor agonist, salmeterol and have investigated the molecular aetiology of this altered state. Although little is known of the regulation of the β₂-adrenoceptor in the rat, it was selected for two reasons. First, the rat is relatively more steroid-sensitive than other small laboratory animals such as the guinea-pig (Hirshman & Downes, 1985). Second, we have shown previously that glucocorticoids, such as dexamethasone, protect against pulmonary β₂-adrenoceptor desensitization in the rat in vivo (Mak et al., 1995), which mimics the effect of steroids in humans (Barnes, 1995).

Methods

Animals and surgery

Male Sprague-Dawley rats (Charles River Ltd, Kent) of 275 – 300 g body weight were housed in a temperature-controlled (21°C) environment with food and water available ad libitum. Animals were sedated with 2% v v⁻¹ Hypnorm (0.3 ml kg⁻¹) and implanted sub-cutaneously with osmotic mini pumps (Alzet model 2001) delivering salmeterol (10 μg kg⁻¹ h⁻¹) or vehicle (0.1% glacial acetic acid in phosphate-buffered saline) for 7 days as described previously (Finney et al., 2000). The dose of salmeterol is approximately 25% of the dose of salbutamol used in patients hospitalized with acute severe asthma (Bohn et al., 1984; Cheong et al., 1988; Cluzel et al., 1990; O'Driscoll et al., 1988) and is based on equivalence studies in human asthmatic subjects, where a weight-for-weight dose ratio for salmeterol:salbutamol of 1 : 4 was found (Higham et al., 1997). Salmeterol was given as an infusion to allow a comparison with a previous study where desensitization of pulmonary β₂-adrenoceptors was induced in rats treated systemically with salbutamol (Finney et al., 2000).

Instrumentation of rats for measurement of airway mechanics

Each rat was anaesthetized with urethane and the left carotid artery and left jugular vein were cannulated for measuring changes in blood pressure and for the injection of drugs respectively. The trachea was cannulated and the animal ventilated at constant volume with a pump operating at 75 strokes per minute. Changes in respiratory insufflation pressure were measured using a modification of the method described by Konzett & Rössler (1940) as described in Finney et al. (2000).

Assessment of pulmonary β₂-adrenoceptor desensitization in vivo

Vehicle- and salmeterol-treated rats were given ACh (500 μg kg⁻¹ i.v, bolus) every 5 min to establish a constant degree of bronchoconstriction. When airway function had normalized (15 min after last dose of ACh), a submaximal concentration of salmeterol (100 μg kg⁻¹ i.v. bolus) was administered and the magnitude of ACh-induced bronchoconstriction was reassessed 5 min later. An identical protocol was used to assess the potential bronchoprotective effect of PGE₂ (300 μg kg⁻¹ i.v. bolus).

Quantification of β-adrenoceptor number in lung membranes

Lung membranes were prepared according to Mak et al. (1995) and Nishikawa et al. (1993), and total, β₁- and β₂-adrenoceptor density (B_max) and the affinity (K_D) of [¹²⁵I]-iodocyanopindolol ([¹²⁵I]-ICYP) was determined as described previously (Finney et al., 2000).

Measurement of steady-state β₁- and β₂-adrenoceptor mRNA levels

Total RNA was isolated from lung as described by Chomczynski & Sacchi (1987) and the abundance of β₁- and β₂-adrenoceptor mRNA transcripts estimated by Northern blot analysis. Samples of denatured mRNA (20 μg lane⁻¹) were size fractionated on 1% agarose/formaldehyde gels containing 20 mM MOPS (pH 7), 5 mM sodium acetate and 1 mM EDTA, and transferred to Hybond-N⁺ membranes (Amersham) by capillary action using 20× standard sodium citrate (SSC; 1× SCC, 150 mM NaCl, 15 mM sodium citrate, pH 7). The RNA was permanently fixed to the membranes using a UV Stratalinker 2400 (Stratagene, Cambridge).

Random primer labelling was carried out with the 851 bp SmaI/PvuII and 439 bp I fragments from the human β₁- and β₂-adrenoceptor cDNAs respectively using [α-³²P]-dCTP. A 1272 bp PstI fragment from the rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA was used as a ‘house-keeping' gene. Pre-hybridization and hybridization were carried out at 42°C with the probes labelled to approximately 1.5×10⁶ c.p.m. ml⁻¹ in buffer A (50 mM Tris-HCl (pH 7.5), 50% formamide, 2× SSC, 5× Denhardt's solution, 0.1% SDS, 5 mM EDTA, 250 μg ml⁻¹ denatured salmon sperm DNA). After hybridization, the blots were washed to a stringency of 0.1× SCC/0.1% SDS at 65°C and exposed to Kodak X-OMAT film at −80°C for 1 to 4 days. The autoradiograms were quantified by laser-scanning densitometry and expressed as a ratio to GAPDH.

Ex vivo cyclic AMP accumulation studies

Lung was chopped (∼5 mm³) and equilibrated for 30 min at 37°C in oxygenating (95% O₂/5% CO₂) Krebs-Henseleit (KS) solution (in mM: 118 NaCl, 5.9 KCl, 1.2 MgSO₄.7H₂O, 1.2 NaH₂PO₄.2H₂O, 2.5 CaCl₂.6H₂O, 10 glucose, 25 NaHCO₃) containing 10 μM indomethacin. Tissue was incubated for 30 min in the presence of IBMX (100 μM), and salmeterol (1 μM) or PGE₂ (1 μM) was added for 5 min. Tissue was removed from the KH solution, blotted on absorbent paper, immersed in liquid N₂ and stored at −80°C. When required the cyclic AMP content was extracted from the lung and measured by radioimmunoassay as described previously (Finney et al., 2000; Seybold et al., 1998).

Measurement of GRK activity

Cytosolic and particulate GRK was prepared from rat lung according to Benovic et al. (1987) and GRK activity was determined immediately by the method of Mayor et al. (1987).

Western immunoblot analyses

Frozen lung was homogenized, denatured and subjected to electrophoresis on 10% SDS/Tris polyacrylamide gels as described previously (Finney et al., 2000). Proteins were transferred onto Hybond nitrocellulose paper and blocked overnight in 10 mM Tris-base/0.05% Tween 20 containing 5% skimmed milk. Membranes were then incubated at 25°C for 1 h with a rabbit polyclonal antibody specific to either G_sα, Gβ or the β₂-adrenoceptor diluted 1 : 1000, 1 : 1000 and 1 : 750 respectively. Membranes were washed, incubated with a donkey, anti-rabbit horseradish peroxidase-conjugated antibody and treated with ECL reagent according to the manufacturer's instructions. Proteins were visualized by exposure of the membranes to Kodak X-OMAT film and quantified by laser-scanning densitometry.

Protein estimation

Protein was measured using a BioRad kit according to the manufacturer's instructions.

Drugs and analytical reagents

Salmeterol was provided by GlaxoWellcome (Stevenage, Hertfordshire, U.K.). ECL reagent, horseradish peroxidase-conjugated secondary antibody and [α³²P]-dCTP and [¹²⁵I]-ICYP (specific activities >3000 Ci mmol⁻¹ and 2000 Ci mmol⁻¹ respectively) were supplied by Amersham International (Amersham, Buckinghamshire, U.K.). Antibodies against G_sα and Gβ were purchased from NEN/Dupont (code NEI 805 and NEI 808 respectively) and the β₂-adrenoceptor antibody was from Santa Cruz (sc # 569). All other reagents were from Sigma (Poole, Dorset, U.K.).

Statistical analysis

Data points and bars represent the mean±s.e.mean of ‘n' independent observations. When appropriate, data were analysed non-parametrically using Mann – Whitney U-test. The null hypothesis was rejected when P<0.05.

Results

Repeated administration of ACh (500 μg kg⁻¹ i.v. ∼ED₅₀) to saline- and salmeterol-treated rats provoked bronchoconstriction that was highly reproducible over 60 min (data not shown). The resting mean arterial blood pressure was not significantly different (P>0.05) between vehicle (76.1±3.7 mmHg, n=28)- and salmeterol (77.9±8.2 mmHg, n=31)-treated animals. Similarly, intravenous administration of ACh (500 μg kg⁻¹) evoked a depressor response in both groups of rats of equivalent magnitude (vehicle: 34.4±1.5 mmHg, n=28; salmeterol: 29.1±3.1 mmHg, n=28, P>0.05).

Effect of chronic treatment of rats with salmeterol on the ability of salmeterol and PGE₂ to protect against ACh-induced bronchoconstriction

Intravenous administration of salmeterol (100 μg kg⁻¹ bolus) or PGE₂ (300 μg kg⁻¹ bolus) to vehicle-treated rats significantly reduced (by 44 and 32% respectively) the magnitude of ACh-induced bronchoconstriction whereas in the salmeterol-treated group of animals no significant protection was observed with either agonist indicating that the desensitization was heterologous (Figure 1a,b; Table 1). In none of the experiments did salmeterol or PGE₂ affect resting airways tone.

Figure 1.

Effect of chronic systemic treatment of rats with salmeterol on the ability of salmeterol and PGE₂, administered acutely, to protect against ACh-induced bronchoconstriction. Rats were treated with vehicle or salmeterol (10 μg kg⁻¹ h⁻¹). After 7 days, each animal was instrumented for the measurement of lung function. ACh (500 μg kg⁻¹ i.v.) was administered and the maximum increase in overflow pressure was measured. When baseline lung function was re-established, salmeterol (100 μg kg⁻¹ i.v. bolus) or PGE₂ (300 μg kg⁻¹ i.v. bolus) was given and 5 min later ACh was administered again and any change in overflow pressure was noted. Each bar represents the mean±s.e.mean of four determinations made in different rats. ^*P<0.05, significant protection of ACh-induced bronchoconstriction.

Table 1.

Effect of chronic systemic treatment of rats with salmeterol on the ability of salmeterol and PGE₂, given acutely, to protect against ACh-induced bronchoconstriction

Effect of chronic treatment of rats with salmeterol on β₁- and β₂-adrenoceptor density and mRNA expression in lung

Chronic treatment of rats with salmeterol (10 μg kg⁻¹ h⁻¹) produced a significant (∼70%) reduction in β-adrenoceptor density when compared to naïve animals, without affecting the affinity of the non-selective ligand, [¹²⁵I]-ICYP (Table 2). In the presence of selective antagonists it was established that salmeterol reduced β₁- and β₂-adrenoceptor number by 50 and 70% respectively but did not alter the K_D of [¹²⁵I]-ICYP for either receptor subtype (Table 2). To determine whether the in vivo binding of salmeterol to an ‘exosite' within the β₂-adrenoceptor reduced, artificially, the total number of binding sites available to [¹²⁵I]-ICYP ex vivo, β₂-adrenoceptor protein in the lung from both groups of animals was measured by Western blotting. Consistent with the radioligand binding data, chronic treatment of rats with salmeterol reduced β₂-adrenoceptor protein in lung membranes by ∼50% (Figure 2). Despite the reduction in receptor expression at day 7, the steady state level of β₁- and β₂-adrenoceptor mRNA transcripts was unchanged in lung harvested from salmeterol-treated rats when compared to animals that received vehicle (Figure 3).

Table 2.

Effect of chronic systemic treatment of rats with salmeterol on the density of cell-surface β₂-adrenoceptors and on the affinity of [¹²⁵I]-ICYP

Figure 2.

Effect of chronic systemic treatment of rats with salmeterol on β₂-adrenoceptor protein expression in lung membranes. Rats were treated with salmeterol (10 μg kg⁻¹ h⁻¹) or vehicle for 7 days and the expression of β₂-adrenoceptor protein was measured. (a,b) show a representative Western blot and bar chart of the mean±s.e.mean of six determinations respectively. Equal loading was confirmed by staining of the gels with Coomassie Blue (not shown). ^*P<0.05, significant reduction in β₂-adrenoceptor protein in salmeterol-treated rats when compared to vehicle-treated animals.

Figure 3.

Effect of chronic systemic treatment of rats with salmeterol on the steady-state level of β₁- and β₂-adrenoceptor mRNA transcripts. Rats were treated with salmeterol (10 μg kg⁻¹ h⁻¹) or vehicle for 7 days. Lungs were excised and β₁- and β₂-adrenoceptor mRNA was extracted and estimated by Northern analysis. (a) shows the mean±s.e.mean of six independent determinations that are expressed relative to the ‘house-keeping' gene GAPDH. A representative autoradiogram is shown in (b).

Effect of chronic treatment of rats with salmeterol on the ability of salmeterol and PGE₂ to increase cyclic AMP mass in lung ex vivo

To determine if down-regulation of β-adrenoceptor number was accompanied by compromised signal transduction, the ability of salmeterol to increase cyclic AMP mass in lung parenchyma was assessed ex vivo. No significant difference in the basal cyclic AMP content was detected between lung excised from either group of animals (Figure 4). However, salmeterol (1 μM)-induced cyclic AMP accumulation was significantly attenuated in lung taken from salmeterol-treated rats (1.56 fold increase over basal level) when compared to animals that received vehicle (2.74 fold increase over basal level; Figure 4a). The desensitization of β-adrenoceptor-mediated cyclic AMP accumulation by salmeterol was heterologous in that the ability of PGE₂ to increase the cyclic AMP content in lung was profoundly impaired (Figure 4b).

Figure 4.

Effect of chronic systemic treatment of rats with salmeterol on salmeterol- and PGE₂-induced cyclic AMP accumulation in lung ex vivo. Rats were treated with salmeterol (10 μg kg⁻¹ h⁻¹; filled bars) or vehicle (hatched bars) for 7 days and the ability of salmeterol (a) and PGE₂ (b) to increase cyclic AMP mass in lung was determined ex vivo in the presence of IBMX (100 μM). Each bar represents the mean±s.e.mean of six independent determinations using tissue from different animals. ^*P<0.05, significant reduction in salmeterol- and PGE₂-induced cyclic AMP formation.

Effect of chronic treatment of rats with salmeterol on the expression of G_sα and G_β subunits in lung

Western blotting was performed to determine if the compromised ability of salmeterol to elevate the cyclic AMP content in lung taken from salmeterol-treated rats was associated with a reduction in the expression of G_sα (Figure 5). A primary antibody raised against an epitope at the carboxy-terminus of the α-subunit of G_s, identified three bands in lung membranes prepared from vehicle-treated rats that migrated as 42, 44 and 52 kDa peptides on SDS polyacrylamide gels. Treatment of rats for 7 days with salmeterol significantly reduced the intensity of each of these bands by between 45 and 70% (Figure 5a,b). A primary antibody raised against an epitope at the carboxy-terminus of the common β-subunit of heterotrimeric GTP-binding proteins, detected a single 35 kDa band in lung membranes prepared from vehicle-treated rats. However, unlike G_sα variants, G_β was not altered in lung taken from salmeterol-treated animals (data not shown). Western blotting failed to detect G_sα or G_β subunits in the cytosolic fraction of lung tissue taken from either group of animals at day 7.

Figure 5.

Effect of chronic systemic treatment of rats with salmeterol on G_sα expression in lung membranes. Rats were treated with salmeterol (10 μg kg⁻¹ h⁻¹) or vehicle for 7 days and the expression of membrane-associated G_sα variants was measured by Western blotting. (a,b) show representative blot and a bar charts of the mean data (±s.e.mean of six determinations) respectively. Equal loading was confirmed by staining of the gels with Coomassie Blue (not shown). ^*P<0.05, significant reduction in G_sα variant expression.

Effect of chronic treatment of rats with salmeterol on GRK activity in lung

Chronic administration of rats with salmeterol produced a significant (∼2 fold over basal level) increase in cytosolic GRK activity in lung parenchyma when compared to vehicle-treated animals (from 164±25 to 341±87 fmol min⁻¹ mg protein⁻¹; n=12, P<0.05). No change in GRK activity was detected in the particulate fraction of salmeterol-treated rats at day 7.

Duration of action of intravenous salmeterol

The duration of action of salmeterol was assessed by monitoring the magnitude of ACh (500 μg kg⁻¹ i.v.)-induced bronchoconstriction repeatedly in naïve rats over a period of 60 min. The anti-spasmogenic activity of salmeterol (100 μg kg⁻¹; bolus) was short-lasting with an offset halftime (defined as the time required for the peak of ACh-bronchoconstriction to recover by 50%) of ∼45 min (Figure 6). Protection against ACh-induced bronchoconstriction was re-established when salmeterol was re-administered 65 min after the first dose, indicating that the short duration of action was not due to β₂-adrenoceptor desensitization.

Figure 6.

Duration of action of intravenous salmeterol in rats. Naïve rats were given ACh (500 μg kg⁻¹, i.v.) and changes in overflow pressure were measured. When basal lung function was re-established salmeterol (100 μg kg⁻¹ i.v.; bolus) was given and 5, 10, 15, 30, 45 and 60 min later the magnitude of ACh-induced bronchoconstriction was measured. At 65 min salmeterol (100 μg kg⁻¹ i.v.; bolus) was given again and the protection against the original bronchoconstrictor response was measured. Data points represent the mean±s.e.mean of six independent determinations.

Discussion

Systemic infusion of salmeterol to rats for 7 days promoted pulmonary β₂-adrenoceptor desensitization that was seen both at the functional and molecular level. This finding is consistent with a number of previous studies where desensitization of β₂-adrenoceptor-mediated relaxation has been demonstrated ex vivo in tracheae and lung taken from rats treated chronically with isoprenaline and noradrenaline (Avner & Noland, 1978; Nishikawa et al., 1994), and in vivo in rats given salbutamol (Finney et al., 2000). Significantly, the anti-spasmogenic activity of PGE₂, which relaxes rat tracheae through EP₄-like prostanoid receptors (Lydford & McKechnie, 1994), was also impaired indicating that the desensitization was heterologous. Although this mode of desensitization is poorly documented in vivo, it has been observed in the lung and heart of rats given salbutamol (Finney et al., 2000) and isoprenaline (Zeiders et al., 1997) respectively.

Many processes can effect desensitization of β₂-adrenoceptors and, as the molecular aetiology of long-term subsensitivity in vivo is largely unexplored, several mechanisms were investigated.

Down-regulation of β₂-adrenoceptor number

Systemic treatment of rats with salmeterol produced a marked (70%) reduction in pulmonary β₂-adrenoceptor density relative to animals that received vehicle. These results agree with a previous study where rats were treated chronically with salbutamol (Finney et al., 2000), but the degree of receptor down-regulation effected by salmeterol was much greater. The reason for this disparity is currently unclear. One explanation is that the decrease in B_max may have been over-estimated due to pseudo-irreversible in vivo binding of salmeterol to the ‘exosite' within the β₂-adrenoceptor, which would reduce the total number of binding sites available to [¹²⁵I]-ICYP ex vivo. However, this possibility seems unlikely as a comparable (52%) reduction in β₂-adrenoceptor protein was found in lung membranes by Western analysis.

The results of the present and a previous study (Finney et al., 2000) are consistent with the concept that chronic exposure of animals to β₂-adrenoceptor agonists effects tolerance by stimulating the internalization and degradation of the cognate receptors (Lohse, 1993). Although destabilization of β₂-adrenoceptor mRNA can account for this effect, there was no evidence for a reduction in mRNA transcripts in the lung of salmeterol-treated rats at day 7. This was unexpected and contrary to previous in vivo experiments performed in rats and guinea-pigs given isoprenaline (Mak et al., 1995; Nishikawa et al., 1993) and noradrenaline (Nishikawa et al., 1994) respectively. At least three explanations could account for this discrepancy. First, the efficacy and selectivity of salmeterol, noradrenaline and isoprenaline differ between β-adrenoceptor subtypes, which may influence the magnitude and kinetics of mRNA degradation. Second, β-adrenoceptor expression and function vary between strains of rat (Van Liefde et al., 1994). Thus, mRNA in the lung of Sprague-Dawley rats treated with salmeterol (this study) may be degraded relatively rapidly and recover to steady-state levels by day 7 when compared to Wistar rats treated identically with isoprenaline (Mak et al., 1995; Nishikawa et al., 1993). Third, salmeterol may down-regulate pulmonary β₂-adrenoceptors in Sprague-Dawley rats by a process unrelated to mRNA destabilization.

While tolerance to salmeterol could be due to the reduction in pulmonary β₂-adrenoceptor density, other processes must be involved, as PGE₂-induced responses were also impaired. Indeed, the desensitization of salmeterol-induced cyclic AMP accumulation in lung ex vivo was very similar in magnitude to that produced in the lung of salbutamol-treated rats (Finney et al., 2000), which would not be expected if β₂-adrenoceptor down-regulation was the only determinant of tolerance. In addition, heterologous desensitization may have resulted, at least in part, from down-regulation/uncoupling of the EP₄-like receptor. Although the mechanism by which this could occur is unclear, PKA is an unlikely effector as activation of this enzyme by salbutamol was significantly impaired in lung taken from salbutamol-treated rats (Finney et al., 2000).

Activation of GRKs

Phosphorylation of GPCRs by GRKs can promote desensitization. However, GRK activity was not elevated in lung membranes prepared from salmeterol- (this study) or salbutamol-treated rats (Finney et al., 2000), which is consistent with GRK-induced receptor phosphorylation being a short-term process. Moreover, as GRKs effect homologous desensitization it is difficult to envisage how this process could also inhibit PGE₂-induced responses. In contrast, cytosolic GRK activity was significantly increased in the lung of salmeterol-treated rats, in agreement with a previous study where salbutamol was used (Finney et al., 2000), which may be due to increased GRK2 gene transcription and/or mRNA stability (Iaccarino et al., 1998).

Down-regulation of G_sα

Heterologous desensitization GPCRs can be accounted for by a reduction in the abundance of membrane-associated G_sα (Milligan, 1993). Indeed, three molecular weight species of G_sα were detected in rat lung membranes as reported previously (Finney et al., 2000) but the expression was significantly reduced in tissue taken from those animals that received salmeterol. Further support for this mechanism derives from a previous study where chronic treatment of rats with salbutamol reduced the ability of cholera toxin to increase the cyclic AMP content in lung ex vivo (Finney et al., 2000). Moreover, the bronchoprotective effect of forskolin and the phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine, was preserved in rats treated chronically with salbutamol providing further evidence that the site of desensitization is upstream of adenylyl cyclase (Finney et al., 2000). Mechanistically, down-regulation of G_sα would blunt the ability of salmeterol, and other agonists (e.g. PGE₂) that utilize the same pool of G_sα, to activate adenylyl cyclase and protect the airways against ACh-induced bronchoconstriction. A similar explanation is suggested for the impaired activation of adenylyl cyclase by isoprenaline and glucagon ex vivo in cardiac membranes purified from rats treated chronically with isoprenaline (Zeiders et al., 1997).

Down-regulation of G_sα could be effected by at least three mechanisms (discussed in Finney et al., 2000) including decreased G_sα gene transcription, destabilization of G_sα mRNA transcripts or redistribution/degradation of membrane-associated G_sα. In cell-based experiments there is strong evidence that β-adrenoceptor agonists promote the translocation of G_sα from the membrane to a cytosolic pool (Ransnas et al., 1989; Wedegaertner et al., 1996) by a process that involves depalmitoylation of the protein (Wedegaertner & Bourne, 1994; Loisel et al., 1999). This is followed, after chronic exposure, by degradation of G_sα (Milligan, 1993). The results presented herein are consistent with the degradation theory as the reduction in membrane-associated G_sα was not associated with a commensurate increase of G_sα in the cytosol at day 7.

A critical question that arises from the present study is whether a 45 to 70% reduction in G_sα is sufficient to effect tolerance to salmeterol. As discussed previously (Finney et al., 2000) evidence provided by Paulssen et al. (1992) suggests that changes in G_sα expression can, indeed, affect signalling through GPCRs in an agonist-dependent manner.

Duration of action of intravenous salmeterol

An unexpected observation of the present investigation was the short-acting behaviour of salmeterol when given to rats by the intravenous route that was not the result of acute β₂-adrenoceptor desensitization. If the duration of action of salmeterol is to be explained by binding to an exosite (Green et al., 1996; Rong et al., 1999) then this should be independent of the route of administration. Thus, other mechanisms must account for this difference. One possibility is that salmeterol was rapidly metabolized to a shorter-acting molecule that retained affinity and efficacy at the β₂-adrenoceptor. This situation would be in contrast to administration by inhalation where salmeterol's high lipophilicity would ensure that it is rapidly absorbed and retained in the airways; indeed, bronchoprotection effected by inhaled salmeterol in asthmatic subjects persists for more than 12 h (Rabe et al., 1993). Previous studies in rats have established that salmeterol is cleared predominantly by metabolism and at least two, albeit minor, products can be formed by O-dealkylation of the phenalkyloxalkyl ‘tail' that, theoretically could be short-acting (Manchee et al., 1993). In our own preliminary studies we have found that following bolus intravenous administration of salmeterol to Sprague-Dawley rats, the circulating level fell rapidly due to the combined effects of a very high volume of distribution (previously estimated at 42 ml min⁻¹ kg⁻¹) and metabolism with products resulting from O-dealkylation detected (Authors' unpublished observations). Metabolism of salmeterol could also account for the unanticipated reduction in pulmonary β₁-adrenoceptor density after chronic treatment if these O-dealkylated metabolites are agonists at both β₁- and β₂-adrenoceptor subtypes. Alternatively, the down-regulation of β₁-adrenoceptors might simply reflect the weak agonist activity of salmeterol at this subtype (Ball et al., 1991; Roux et al., 1996).

Conclusion

Chronic systemic treatment of rats with salmeterol produced pulmonary β₂-adrenoceptor desensitization. Although several processes could contribute to this effect, the heterologous nature of the desensitization and the results of a previous investigation with salbutamol (Finney et al., 2000) where the bronchoprotective effect of forskolin and IBMX were preserved, suggest that a primary molecular aetiology is a reduction in the abundance of membrane-associated G_sα. As salmeterol is a partial agonist on many tissues the marked reduction in β₂-adrenoceptor density may also have contributed to desensitization. It is important to appreciate that these experiments were performed with healthy rats and the relevance of the findings to patients with asthma is unknown. However, evidence that asthma, itself, does not predispose to β₂-adrenoceptor desensitization has been presented (Penn et al., 1996). Thus, down-regulation of G_sα may be pertinent to the treatment of asthma where susceptible individuals become tolerant to the beneficial effects of salmeterol or high doses of short-acting β₂-adrenoceptor agonists such as salbutamol.

Note added in proof

Since submission of this manuscript the short duration of action of salmeterol at pulmonary β₂-adrenoceptors has been confirmed in the guinea-pig when given by the intravenous route.

Abbreviations

FEF_{25 – 75}

forced expiratory flow at 25% and 75% of vital capacity

FEV₁

forced expiratory volume in 1 s

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

GRK

G-protein receptor-coupled kinase

ICYP

iodocyanopindolol

KH

Krebs-Henseleit

PGE₂

prostaglandin E₂

PKA

cyclic AMP-dependent protein kinase

ROS

rod outer segments

SSC

standard sodium citrate

IBMX

3-isobutyl-1-methylxanthine