The human near-term myometrial β₃-adrenoceptor but not the β₂-adrenoceptor is resistant to desensitisation after sustained agonist stimulation

Abstract

1. In order to compare the β₂- and β₃-adrenoceptor (β-AR) desensitisation process in human near-term myometrium, we examined the influence of a pretreatment of myometrial strips with either a β₂- or a β₃-AR agonist (salbutamol or SR 59119A, respectively, both at 10 μM, for 5 and 15 h) on the relaxation and the cyclic adenosine monophosphate (cAMP) production induced by these agonists.

2. To assess some of the mechanisms potentially implicated in the β-AR desensitisation process, we studied the influence of such treatment on the number of β₂- and β₃-AR binding sites, the β₂- and β₃-AR transcripts expression and the phosphodiesterase 4 (PDE4) activity.

3. Salbutamol, but not SR 59119A, concentration–response curve (CRC) was shifted by a 15 h salbutamol preincubation, with a significant difference in −log EC₂₀ values (6.31±0.13 vs 5.58±0.24, for control and 15 h salbutamol pretreatment, respectively, P<0.05). Neither salbutamol nor SR 59119A CRCs were modified after a 15 h preincubation with SR 59119A.

4. A 15 h exposure of myometrial strips to salbutamol significantly reduced the salbutamol-induced (0.60±0.26 vs 1.54±0.24 pmol mg⁻¹ protein, P<0.05), but not the SR 59119A-induced, cAMP production. No decrease in cAMP production was observed after a 15 h SR 59119A exposure.

5. A 15 h salbutamol exposure of myometrial strips significantly reduced the β₂- but not the β₃-AR binding site density, whereas no decrease in the number of β₂- and β₃-AR binding sites was observed after a 15 h SR 59119A treatment.

6. Neither PDE4 activity nor the β₂- and β₃-AR mRNA expression levels were affected by salbutamol or SR 59119A treatments.

7. Our results indicate that β₃-AR, but not β₂-AR, are resistant to the agonist-induced desensitisation. In our model, β₂-AR desensitisation is mediated by a decreased number of β₂-AR that was not explained by transcriptional regulation of the receptor.

Introduction

Despite preventative measures, the incidence of premature birth (a gestational age at birth of less than 37 completed gestational weeks) remains relatively high, at around 6%, in developed countries (Goldenberg & Rouse, 1998; Ananth et al., 2001). The reduction in the number of premature deliveries depends on the early and specialised management of pregnancy and parturition, including the development of new tocolytic agents (Tsatsaris et al., 2001). β₂-adrenoceptor (β₂-AR) agonists are the most common tocolytic agents used worldwide, but a loss of responsiveness to β-mimetics occurs physiologically in late-pregnant myometrium and after chronic exposure to these drugs (Berg et al., 1983; Litime et al., 1989). In addition, β-mimetics are known to be responsible for serious foetal and maternal side effects. For these different reasons, the use of β-mimetics as tocolytic agents has been repeatedly questioned (The Canadian Preterm Labor Investigators Group, 1992; Berkman et al., 2003).

Originally subclassified into β₁- and β₂-AR (Lands et al., 1967a, 1967b), another subtype, the β₃-AR subtype has since been reported (Emorine et al., 1989). β₃-AR have been shown to mediate lipolysis in white adipose tissue and thermogenesis in brown adipose tissue (Arch et al., 1984; Zaagsma & Nahorski, 1990; Lonnqvist et al., 1995), and also to inhibit the contractile activity of ileum and colon (Bond & Clarke, 1988; Manara & Bianchetti, 1990; Bardou et al., 1998). The presence of a functional β₃-AR was recently demonstrated in the human near-term myometrium (Bardou et al., 2000). In this work, the stimulation of myometrial tissue by a specific β₃-AR agonist led to a higher degree of relaxation than specific stimulation by β₂-AR agonist, and was associated with a cAMP production. Several arguments suggest that β₃-AR might be less prone than β₂-AR to desensitisation (Carpene et al., 1993; Nantel et al., 1995). This particularity may be due to a lack of recognition sites in β₃-AR for the cAMP-dependent protein kinase and for the β-AR kinase (Strosberg, 1993) implicated in the desensitisation process of β₂-AR.

Several processes have been implicated in β₂-AR desensitisation; for example, the downregulation of the receptor and its uncoupling from G-stimulating (Gs) protein after phosphorylation by kinases (Berg et al., 1985; Litime et al., 1989). Another mechanism accounting for the reduced responsiveness of myometrium to β₂-AR agonists might be the enhancement of the phosphodiesterase (PDE) capacity to degrade cyclic AMP (cAMP). PDE families represent the sole mechanism for inactivating cAMP or cGMP (Conti et al., 2003; Houslay & Adams, 2003). In human myometrium, among the five PDE families represented, it is now well established that the PDE4 family, specific for cAMP hydrolysis and inducible by its own substrate cAMP, is predominantly expressed (Leroy et al., 1989; Bardou et al., 1999). There is now defined evidence that cAMP induces an upregulation of PDE4, accelerating the degradation of this cyclic nucleotide (Perry et al., 2002). Interestingly, previous studies on cellular models such as U937 monocytes (Torphy et al., 1995) and human neutrophils (Ortiz et al., 2000) indicated that cAMP-PDE4 activity was increased after a treatment by β₂-agonists. In human myometrial cells, we demonstrated that compounds elevating intracellular cAMP level, such as 8-Br-cAMP or forskolin, were able to increase PDE4 activity (Mehats et al., 1999). This mechanism could explain the limited relaxant effect of cAMP on myometrial tissue.

The purpose of the present study was to assess the desensitisation process of β₂- and β₃-AR in human near-term myometrium. We investigated the effects of long-term exposure of myometrial strips to a β₂-AR agonist, salbutamol, or a β₃-AR agonist, SR 59119A, on the inhibition of spontaneous contractile activity, on the cAMP production and on the number of β₂- and β₃-AR binding sites, induced by these agonists. Furthermore, we evaluated the possible implication of an upregulation of PDE4 activity or of a modulation in the mRNA expression levels for both β₂- and β₃-AR as some of the possible contributing process to desensitisation.

Methods

Biological samples

Myometrial biopsies were obtained from pregnant women who presented normal uncomplicated pregnancies, but who were delivered by elective caesarean section prior to the onset of labour (38–40th weeks of pregnancy) because of previously diagnosed cephalopelvic disproportion. Myometrial strips, excised from the longitudinal layer at the antiplacental site, were immediately placed in preoxygenated Krebs solution at 4°C (composition, mM: NaCl, 118; KCl, 5.4; CaCl₂, 2.5; KH₂PO₄, 0.6; MgSO₄, 1.2; NaHCO₃, 25; glucose, 11.7; ascorbic acid, 0.1) and transported immediately to the laboratory. Tissues were dissected free from serosa and used fresh (functional studies) or quickly frozen in liquid nitrogen and stored at −80°C after pretreatment (biochemical, binding and molecular studies). The use of human myometrial tissue for the experiments was approved by the ‘Comité Consultatif de Protection des Personnes pour la Recherche Biomédicale' (CCPPRB de Bourgogne, France), and by the Ethics Committee of the La Fe Hospital (Valencia, Spain). Written informed consent was obtained from all donors.

Tissue incubation process

Myometrial tissues were cut into strips (8–10 mm long by 2–3 mm in cross section) and incubated for 5 or 15 h in preoxygenated Krebs solution (composition as above) containing either a β₃-AR agonist, SR 59119A (10 μM), or a β₂-AR agonist, salbutamol (10 μM), or their respective solvents, at room temperature. We previously determined that maximal effects of salbutamol and SR 59119A were obtained at the concentration of 10 μM (Bardou et al., 2000).

Functional study

After the incubations described above, each muscle strip was suspended isometrically under a resting tension of 2g in a 10 ml organ bath containing Krebs solution (composition as above) at 37°C, and continuously gassed with a mixture of 95% oxygen and 5% carbon dioxide (pH 7.4). The 2 g initial tension was chosen according to previous reports including ours (Bardou et al., 2000; Dennedy et al., 2001). One end of each strip was connected to a force–displacement transducer and tension changes were measured with Pioden strain gauges (UF1), amplified (EMKA, Paris, France), and recorded on a pen-writing oscillograph (Linseis, L65514, Munich, Germany). After 1 h, during which the myometrial strips were washed every 15 min and the resting tension readjusted to 2g, the strips were allowed to equilibrate for a further 1 h until they showed regular spontaneous rhythmic contractile activity. Once contractions became regular in amplitude and frequency, cumulative concentration–response curves (CRCs) (from 0.1 to 10 μM) were established for each compound studied: SR 59119A (β₃-AR agonist) and salbutamol (β₂-AR agonist). The effect of each agonist was expressed as a percentage of the initial amplitude of spontaneous contractions. For all concentrations, the amplitude of contraction was recorded at steady state. Since 3–4 h were necessary for the construction of a cumulative CRC (the contact time for each concentration was around 40 min), only one complete curve was obtained for each strip. E_max indicates the maximal response obtained at the maximal concentration tested (10 μM), for each agent. EC₂₀ values indicate the concentration producing a 20% inhibition of tone, and were calculated using the GraphPad Prism 4.01 computer program (GraphPad Software, San Diego, CA, U.S.A.). For analysis, EC₂₀ values were log transformed and expressed as −log EC₂₀.

Biochemical study

cAMP assay

After the 5 or 15 h pretreatment, myometrial strips were equilibrated in Krebs solution (composition as above) continuously gassed with a mixture of 95% oxygen and 5% carbon dioxide at 37°C (pH 7.4) for 50 min, and then exposed to the agonists (SR 59119A or salbutamol, each at 10 μM) for 5 min (Andersson et al., 1980). During this 5-min time period, 0.5 mM 3-isobutyl-1-methylxanthine (IBMX), a nonselective inhibitor of cAMP-PDE activity, was added in order to prevent cAMP degradation (Fonteh et al., 1993; Bardou et al., 2000; Boulven et al., 2001). cAMP levels (expressed as pmol mg⁻¹ protein) were determined after agonist pretreatments, and were compared with those obtained in time-matched paired strips incubated either in Krebs solution alone (basal value) or in Krebs solution containing the respective solvents of each agonist (solvent value). Tissues were then processed as previously reported by Bardou et al. (1999). The samples were immediately transferred into liquid nitrogen and homogenised in ice-cold 10% trichloroacetic acid. The homogenate was centrifuged at 10,000 × g for 15 min at 4°C. The pH of the supernatant was neutralised by the addition of excess of calcium carbonate, followed by low-speed centrifugation. Aliquots of the supernatant were assessed for cAMP, using an enzyme immunoassay kit (RPN 225, Amersham Pharmacia Biotech, Little Chalfont, U.K.), following the instructions of the manufacturer.

cAMP-PDE assay

Myometrial tissues, treated as described above, were homogenised in ice-cold homogenisation buffer (100 mg tissue ml⁻¹) with an Ultra-Turrax. The composition of the homogenisation buffer was 100 mM Tris-HCl (pH 7.4), 2 mM MgSO₄, 2 mM EDTA, 10% glycerol, 1 mM β-mercaptoethanol, 1 μM leupeptin, 10 μg ml⁻¹ aprotinin, 25 μg ml⁻¹ pefabloc, 130 μg ml⁻¹ benzamidine and 50 μg ml⁻¹ soybean trypsin inhibitor. cAMP PDE activity was determined using the Kincaid & Manganiello (1988) method. Activity was measured in high-affinity conditions with 1 μM cAMP as substrate, and was expressed as specific activity (pmol min⁻¹ mg⁻¹ protein). All assays were carried out in linearity conditions with respect to time and protein concentration. PDE4 activity was gauged as the fraction of total cAMP PDE activity that was inhibited by 10 μM rolipram, a selective PDE4 inhibitor. Protein concentrations were determined using the Bradford (1976) method with bovine serum albumin (BSA) as a standard. IBMX was not used in this part of the study.

Binding studies

After a 15 h incubation of myometrial strips with salbutamol or SR 59119A (as described above), crude membranes were obtained as reported previously (Breuiller et al., 1987). Membranes were prepared at a protein concentration of 1–3 mg ml⁻¹, in a cold 50 mM Tris-HCl (pH 7.4) buffer. Protein concentration was determined by the Bradford method. Binding studies were performed at 30°C for 60 min in a 0.25 ml crude membrane preparation (12 μg of protein per sample) that contained 50 mM Tris-HCl (pH 7.4), 154 mM NaCl, 0.1% BSA and (−)-[¹²⁵I]-iodocyanopindolol (ICYP) from 2 to 1000 pM (saturation analysis). The ICYP (specific activity, 2200 Ci mmol⁻¹) was obtained from Perkin-Elmer-Life Sciences (Boston, MA, U.S.A.). Reactions were terminated by dilution with 5 ml of ice-cold 50 mM Tris-HCl (pH 7.4), 0.01% Triton X-100 and filtration over glass-fiber filters (Whatman GF/C) presoaked in 0.1% BSA in 50 mM Tris-HCl (pH 7.4). The filters were washed with an additional 15 ml of 50 mM Tris-HCl (pH 7.4), 0.01% X-100 Triton, dried and counted on a Packard Gamma counter with 85% efficiency. In previous studies examining β₂-AR, specific binding was defined as the difference between the amount of ICYP bound in the absence (total binding) and the presence (nonspecific binding) of 500 μM unlabeled (−)-isoproterenol (Sigma). For β₃-AR binding studies, nonspecific binding was defined by the amount of ICYP bound in the presence of 500 μM unlabeled (±)-alprenolol (Sigma). Competition experiments were carried out at a fixed concentration of ICYP (400 pM) with increasing concentrations of competing agents, SR 59119A or salbutamol, and in the presence of 0.1 μM (−)-propranolol in order to block β₁- and β₂-AR (Roberts et al., 1993). At ligand concentrations near the ICYP dissociation constant (K_D), nonspecific binding was <15 and 30% of total binding for β₂- and β₃-AR, respectively. Results are expressed as means±standard error mean (s.e.m.). The density of receptors (B_max) and dissociation constants (K_D) were calculated by Scatchard analysis of saturation curves and linear regression analysis using the GraphPad Prism 4.01 computer program (GraphPad Software, San Diego, CA, U.S.A.). The inhibition constants were calculated by the Cheng & Prusoff (1973) equation: K_i=IC₅₀/(1+[L]/K_D), where IC₅₀ is the concentration of competing agent producing 50% of a maximal response, [L] is the concentration of ICYP used in the assay and K_D is the dissociation constant of ICYP, as determined by Scatchard analysis.

RT–PCR analysis

Total RNA was extracted from the pretreated myometrial strips using the Trizol reagent method (Gibco-Life Technologies, Cergy-Pontoise, France). Samples were homogenised in Trizol buffer (100 mg tissue ml⁻¹) with an Ultra-Turrax. RNA preparations were recovered by phenol/chloroform extraction, isopropanol precipitation and ethanol washing, according to the manufacturer's instructions. The first strand of complementary DNA (cDNA) was generated from 8 μg total RNA using random hexamers to prime the reverse transcription (RT) in a total reaction volume of 25 μl. Total RNA was denatured by heating at 72°C for 10 min and cooling immediately on ice. This preparation was then incubated with 800 U Moloney Murine Leukemia Virus reverse transcriptase M-MLV RT (Gibco-Life Technologies) in the presence of 10 mM dithiothreitol, 20 μM random hexamers (pd(N)6), 600 μM dNTP and 20 U ribosomal RNAsin ribonuclease inhibitor (Gibco-Life Technologies), for 60 min at 39°C. The reaction was stopped by heating at 95°C for 5 min, followed by cooling. RT products were stored at −20°C. Preparations achieved without reverse transcriptase were routinely used as a control of each RNA sample. No PCR product was detected in the absence of reverse transcriptase during the RT step, indicating that the RNA preparations were free from intact genomic DNA. Amplification was performed in 1 × PCR buffer (50 mM KCl, 20 mM Tris-HCl, pH 9.3) in a 25 μl total reaction volume. This contained 200 μM of each deoxy-NTP and 0.5 mM MgCl₂, together with 1 μM of each specific primer, sense and antisense, 1.25 U Taq DNA polymerase (Gibco-Life Technologies), and 2–3 μl of the RT product. The primers were designed according to Roberts et al. (1997). The PCR conditions for amplification were 28 cycles for β₂-AR and 36 cycles for β₃-AR. The amplification profile for β₂- and β₃-AR consisted in denaturation at 94°C for 1 min, annealing at 64°C for 1 min and extension at 72°C for 1 min with a final extension at 72°C for 10 min.

After the indicated subsaturing cycles of denaturation and extension, a 15 μl aliquot from each reaction mixture was resolved by electrophoresis on 2% agarose gel (Agarose 1000, Gibco-Life Technologies), and visualised by ethidium bromide staining under UV light. The DNA molecular mass standard ladder consisted of fragments between 123 and 861 bp in multiples of 123 bp (λ123 DNA ladder, Gibco-Life Technologies). Additional validity control was achieved by a Southern blot analysis of the PCR product using an internal oligonucleotide probe as previously described (Roberts et al., 1997) (data not shown). Amplification of an endogenous marker, the human β₂-microglobulin cDNA, was used as an internal control, because its related protein is found on the surface of many nucleated cells (Suggs et al., 1981).

Drugs and solutions

The drugs and chemicals used and their sources were: N-((7-methoxy-1,2,3,4-tetrahydronaphtalen-(2R)-2-yl)methyl)-(2R)-2-hydroxy-2-(3-chlorophenyl)ethanamine hydro-chloride (SR 59119A) that was a gift from Sanofi-Synthelabo, Sanofi-Midy Research Centre (Milan, Italy), and salbutamol sulfate (Sigma, France). SR 59119A was dissolved in a mixture of absolute ethanol 30%, DMSO 2%, distilled water for the 10 μM solution and thereafter diluted in distilled water, whereas salbutamol was dissolved in distilled water. Drug concentrations are given as final bath concentrations.

Statistical analysis

In the functional experiments, differences among groups were determined by analysis of variance (ANOVA), followed by the Bonferroni-corrected t-test. In biochemical and binding experiments, differences among groups were determined by Student's t-test for paired data. All differences were considered significant when P<0.05.

Results

Viability after 15 h pretreatment

In order to assess the viability of the myometrial smooth muscle tissue, we compared the amplitude of spontaneous contractions after the resting and equilibration period of either freshly harvested and set-up tissues or tissues incubated for a 15 h period in Krebs solution at room temperature. The amplitudes for spontaneous contractions were 4.81±0.65 and 4.19±0.45 g, respectively (n=21 for each time, difference not statistically significant).

Effect of a long-term exposure to a β₂-AR agonist, salbutamol, or to a β₃-AR agonist, SR 59119A, on the further inhibition of spontaneous contractions of near-term myometrial strips induced by the same agonists

Without pretreatment, salbutamol and SR 59119A both inhibited the spontaneous contractions of myometrial strips in a concentration-dependent manner (Figures 1a and b). A 15 h pretreatment of the strips with salbutamol (10 μM) induced a significant shift (Figure 1a) of the CRC of the same agonist, whereas the CRC of SR 59119A was not modified (Figure 1b). The effect of salbutamol obtained at the maximal concentration tested (10 μM) was not significantly affected by long-term treatment with this agonist (41±9 vs 27±8% for control and 15 h salbutamol pretreatment experiments, respectively), whereas a significant difference was reached for the −log EC₂₀ values (6.31±0.13 vs 5.58±0.24, for control and 15 h salbutamol pretreatments, respectively, P<0.05). The pretreatment of myometrial strips with SR 59119A (10 μM) did not affect significantly either the CRC of the same agonist (Figure 1b) or the CRC of salbutamol (Figure 1a). A 5 h pretreatment of the strips with either salbutamol or SR 59119A did not significantly affect the two CRC (data not shown).

Figure 1.

Effect of a 15 h pretreatment with salbutamol (10 μM) or with SR 59119A (10 μM) on the (a) salbutamol- or (b) SR 59119A-induced inhibition of spontaneous contractions of human near-term myometrial strips. Results are expressed as mean±s.e.m. of seven experiments performed on myometrial strips obtained from seven women. A statistically significant difference for the concentration–response curves for salbutamol between salbutamol pretreatment and control experiments was determined (ANOVA, P<0.05).

Effect of salbutamol or SR 59119A pretreatment on cAMP production induced by the same agonists

The cAMP production was measured in myometrial strips pretreated with salbutamol or with SR 59119A for 5 or 15 h. Results were expressed as mean±s.e.m. of cAMP increase (pmol min⁻¹ mg⁻¹ protein) over basal value (3.19±0.39 and 2.96±0.41 pmol min⁻¹ mg⁻¹ protein at 5 and 15 h, respectively). In time-matched control experiments (5 and 15 h), salbutamol and SR 59119A (both at 10 μM) were both able to induce cAMP production. The increase in cAMP production was not statistically different between salbutamol and SR 59119A in 5 and 15 h time-matched controls (Figures 2a, b, c and d). The 15 h pretreatment with salbutamol significantly decreased the ability of this agonist to stimulate cAMP production (0.60±0.26 pmol min⁻¹ mg⁻¹ protein after 15 h pretreatment with salbutamol vs 1.54±0.24 pmol min⁻¹ mg⁻¹ protein without pretreatment, P<0.05) (Figure 2c). The decrease in salbutamol-induced cAMP production obtained after a 5 h pretreatment with salbutamol did not reach statistical significance (Figure 2a).

Figure 2.

Effect of a 5 h (a, b) or a 15 h (c, d) pretreatment with salbutamol (10 μM) or SR 59119A (10 μM) on the further salbutamol- and SR 59119A-induced cAMP production of human near-term myometrial strips. Results are expressed as mean±s.e.m. of cAMP increase (in pmol min⁻¹ mg⁻¹ protein) over basal value (3.19±0.39 and 2.96±0.41 pmol min⁻¹ mg⁻¹ protein at 5 and 15 h, respectively). Six experiments were performed on myometrial strips obtained from six women. The asterisk indicates a cAMP level that was statistically different from control value (^*P<0.05).

Neither a 5 nor a 15 h pretreatment with salbutamol affected the SR 59119A-induced cAMP production (Figures 2b and d). Indeed, the 5 or 15 h pretreatments with SR 59119A were without effect either on the salbutamol-induced or the SR 59119A-induced cAMP production (Figures 2a, b, c, d).

Effect of salbutamol or SR 59119A pretreatment on the number of β₂- and β₃-AR binding sites

Pharmacological characterisation of β-AR in human pregnant myometrium was accomplished by performing saturation and competition-binding studies using ICYP as a labelled ligand. Saturation radioligand-binding studies were performed to quantify the density of β-AR subtypes in membrane preparations of control myometrial tissue, and after salbutamol or SR 59119A pretreatment. In control myometrial tissue, Scatchard analysis of the saturation curve demonstrated the appearance of a two-site plot (data not shown), indicating both a low-affinity site (pK_D=8.98±0.11; B_max=116.2±19.0 fmol mg⁻¹ of protein) and a higher affinity site (pK_D=10.55±0.17; B_max=7.6±1.3 fmol mg⁻¹ of protein) (Table 1). Consistent with our previous report where β-AR binding properties were described in human pregnant myometrium (Breuiller et al., 1987), the high-affinity binding site found here resembled that of a β₂-AR profile. The identity of the low-affinity binding site was confirmed by competition-binding experiments. As shown in Figure 3a, ICYP binds to a class of receptors that demonstrate pharmacologic specificity consistent with the β₃-AR subtype: the selective β₃-AR agonist SR 59119A was more potent than the selective β₂-AR agonist salbutamol in inhibiting ICYP binding.

Table 1.

Effect of a 15 h exposure of myometrial strips to salbutamol (10 μM) or to SR 59119A (10 μM) on the number of β₂- and β₃-AR binding sites and on the pK_D values

		Control	Salbutamol pretreatment	SR 59119A pretreatment
β2-AR	Bmax	7.6±1.3	4.7±1.0*	10.7±2.7
	pKD	10.55±0.17	10.70±0.10	10.23±0.14
β3-AR	Bmax	116.2±19.0	107.5±14.4	127.5±26.0
	pKD	8.98±0.11	8.88±0.07	8.95±0.11

Results are expressed as mean±s.e.m. of pK_D and of maximal binding (B_max in fmol mg⁻¹ protein). Experiments were performed on five myometrial membrane preparations obtained from five women. ^*The asterisk indicates the number of binding sites that were statistically different from control (P<0.05).

Figure 3.

Competition between ICYP and SR 59119A or salbutamol for β₃-AR binding sites in membrane preparations of (a) control, (b) 15 h salbutamol- and (c) 15 h SR 59119A-treated pregnant myometrial strips. Results are given for three to four different pregnant women.

Scatchard analysis revealed that the number of β₂-AR binding sites was significantly reduced in the membrane preparations of the 15 h salbutamol-treated strips compared to the control strips (B_max=4.7±1.0 vs 7.6±1.3 fmol mg⁻¹ of protein, P<0.05) (Table 1). However, a 15 h exposure to SR 59119A did not significantly modify the number of β₂-AR binding sites. pK_D values for 15 h salbutamol and 15 h SR 59119A exposed membrane preparations were similar to that of control (Table 1). By contrast, Scatchard analysis of saturation curves demonstrated that the number of β₃-AR binding sites was not modified either by a 15 h salbutamol or by a 15 h SR 59119A treatment (B_max=107.5±14.4 and 127.5±26.0 fmol mg⁻¹ of protein in membrane preparations of 15 h salbutamol- and 15 h SR 59119A-treated strips, respectively) (Table 1). pK_D values for β₃-AR in the membrane preparations of 15 h salbutamol- and of 15 h SR 59119A-treated strips were unchanged in comparison to strips without pretreatment.

The effect of salbutamol or SR 59119A pretreatment on the agonist-binding properties of β₂- and β₃-AR was further analysed (Figure 3). In control membrane preparations, SR 59119A competed at the propranolol-resistant ICYP-binding site with a pK_i value of 6.58±0.68 (Figure 3a). Competition-binding data were similar in membrane preparations of 15 h salbutamol- (pK_i=6.19±0.33) and 15 h SR 59119A- (pK_i=6.76±0.49) treated myometrial strips (Figures 3b and c). It should be noted that the shallow slope for these curves indicates that two sites may be present but cannot be significantly differentiated by computer analysis. The β₂-AR agonist salbutamol did not compete at the β₃-AR-binding sites (Figures 3a, b and c).

Analysis of β₂- and β₃-AR mRNA expression in near-term myometrial strips treated with salbutamol or SR 59119A

Using an RT–PCR method, we analysed the expression of β₂- and β₃-AR transcripts in myometrial strips treated either with 10 μM salbutamol or SR 59119A. Amplification of β₂-microglobulin cDNA, the internal control, allowed a semiquantitative analysis of the β₂- and β₃-AR transcripts expression, as we verified the successful normalisation of RNA amounts by obtaining equivalent intensity for the β₂-microglobulin band in control vs treated strips. None of the treatments (5 and 15 h) with salbutamol or with SR 59119A induced a change in β₂- and β₃-AR mRNA expression (Figures 4 and 5).

Figure 4.

Effect of a 5 or 15 h treatment with salbutamol (10 μM) or SR 59119A (10 μM) on the β₂-AR transcript expression of human near-term myometrium. The ethidium bromide-stained gel pictures are representative of five separate experiments with the myometrium of different women. DNA molecular mass standards appear in the lane at the far right.

Figure 5.

Effect of a 5 or 15 h treatment with salbutamol (10 μM) or SR 59119A (10 μM) on the β₃-AR transcripts expression of human near-term myometrium. The ethidium bromide-stained gel pictures are representative of five separate experiments with myometrium of different women. DNA molecular mass standards appear in the lane at the far right.

Influence of salbutamol or SR 59119A pretreatment on cAMP-PDE4 activity

In order to assess whether cAMP-PDE4 activity could be affected by β₂- or β₃-AR agonist treatment, we measured the PDE4 activity in homogenates of myometrial strips treated either with salbutamol or with SR 59119A. No increase in PDE4 activity was observed after 5 or 15 h of treatment with salbutamol or SR 59119A (Figure 6). We also controlled so that the total cAMP-PDE activity was not modified by these treatments (data not shown).

Figure 6.

Effect of a 5 or 15 h treatment with salbutamol (10 μM) or SR 59119A (10 μM) on the PDE4 activity of human near-term myometrial strips. Results are expressed as mean±s.e.m. of six experiments performed on myometrial strips obtained from six women.

Discussion

The purpose of the present study was to examine whether desensitisation occurred in human near-term myometrium after a sustained stimulation of β₂- and β₃-AR with their respective agonists. In our experimental conditions, we confirmed that, without pretreatment of the myometrial strips, both salbutamol and SR 59119A were able to inhibit spontaneous contractions and to induce cAMP production. These findings provide confirmation of β₂- and β₃-AR functional integrity, and positive coupling with the adenylyl cyclase pathway (Bardou et al., 2000). Our functional and biochemical results demonstrate that β₂-AR undergoes functional desensitisation after long-term exposure to salbutamol, which is associated with a significant decrease in cAMP production. As it was shown in binding experiments, this desensitisation occurs at the receptor level. Indeed, the density of β₂-AR binding sites was decreased after a 15 h exposure of myometrial strips to salbutamol. In contrast, a sustained stimulation by the β₃-AR agonist SR 59119A did not modify its subsequent functional effect and the number of β₃-AR binding sites remained unchanged after such treatment. Our results provide strong evidences for resistance of β₃-AR to the agonist-induced desensitisation in human near-term myometrium. Nevertheless, our binding experiments show that β₃-AR are more numerous than β₂-AR in near-term myometrium, while it was admitted that the predominant subtype was the β₂-AR subtype in this tissue (Hayashida et al., 1982; Breuiller et al., 1987). β₃-AR have not been previously detected, probably because of insensitive experimental conditions such as unadapted (i.e. too low) ICYP concentration needed to recognise the β₃-AR subtype. Indeed, the pharmacological properties of the human β₃-AR differ from those of conventional β₁- and β₂-AR. The main differences include atypically low affinity for conventional β-AR antagonist radioligands such as ICYP, as well as low potencies for β-AR reference agonists (Emorine et al., 1994).

Our results are in general agreement with the numerous literature on myometrial β₂-AR desensitisation. For example, it was shown that treatment of pregnant women (Berg et al., 1985; Engelhardt et al., 1997), or in vitro treatment of human myometrial tissue (Andersson et al., 1980) with a β₂-AR agonist was associated with a loss in efficacy of the β₂-AR agonist, due to a reduction of the number of β₂-AR binding sites.

We hypothesised that part of the β₂-AR desensitisation process consists in a transcriptional regulation of their receptors and that long-term treatment with salbutamol would be associated with a decrease in β₂-AR mRNA levels, as it was previously described in pregnant rats with isoprenaline, a nonselective β-AR agonist (Lecrivain et al., 1998). Contrary to the report of Lecrivain et al. (1998), in our study, the decrease in the number of β₂-AR binding sites was not associated with a significant change in the β₂-AR mRNA levels. But our results are in accordance with the study by Engelhardt et al. (1997) performed in human myometrial biopsies treated with fenoterol, a selective β₂-AR agonist.

We might also hypothesise that an increased degradation of cAMP by PDE4 could play a role in β₂-AR desensitisation. This mechanism has been well documented in several systems including S49 lymphoma cells, platelets, hepatocytes, adipocytes and Sertoli cells (for review, Giembycz, 1996; Perry et al., 2002). It has been demonstrated, in different human blood cells, that treatment with either β₂-AR agonists (Manning et al., 1996; Seybold et al., 1998; Ortiz et al., 2000) or with a cAMP structural analogue, 8-Bromo-cAMP, leads to an upregulation of PDE4 enzymes. More recently, new evidences for a role of cAMP on PDE4 promoters have been defined (Vicini & Conti, 1997; D'Sa et al., 2002; Le Jeune et al., 2002). Furthermore, it has been shown, in an in vivo model of rat pulmonary β₂-AR desensitisation induced by salbutamol, that the β₂-AR binding site density was significantly reduced, and associated with an increase in the PDE3 and PDE4 activities in lung membranes of these animals (Finney et al., 2000). Interestingly, the implication of an upregulation of PDE4 in the adaptation of myometrium to long-term treatment with β₂-AR has been emphasised by Mehats et al. (1999). Furthermore, Bardou et al. (1999) and Mehats et al. (2001) have both shown that pretreatment of myometrial strips of women at term with a PDE4-selective inhibitor potentiates the relaxant effect of salbutamol. In our model, the long-term exposure of myometrial strips to salbutamol did not modify the PDE4 activity. Such an absence of PDE4 increase can partly be explained by the degree of intensity of the intracellular cAMP rise (1.5–2-fold) observed following β-AR agonist treatment. This increase might be too weak, with regard to the 10-fold cAMP increase observed in forskolin-treated myometrial cells, to induce PDE4 de novo synthesis (Mehats et al., 1999). Another reason for the lack of PDE4 upregulation observed in our study might be linked to specific cAMP intracellular compartmentalisation, as already described (Jurevicius & Fischmeister, 1996; Zaccolo & Pozzan, 2002).

The myometrial β₂-AR desensitisation process may also occur through other different pathways that have not been assessed in the present study. Simon et al. (2001) demonstrated that G-protein-coupled receptor kinase (GRK) activation, particularly GRK2, could be an important mediator of myometrial β₂-AR desensitisation. Another possible mechanism for β₂-AR desensitisation is uncoupling to adenylyl cyclase, as it was described either in pregnant rat myometrium (Lecrivain et al., 1998) or in the human failing heart (Bohm et al., 1997). This finding could be attributed either to an increased activity and gene expression of β-AR kinase leading to phosphorylation and uncoupling of receptors (Bohm et al., 1997), or to an increase in the functional activity of Gi-proteins, negatively coupled with the cAMP-signalling pathway (Lecrivain et al., 1998). We can also hypothesise that modifications in the coupling mechanisms (Gs proteins) between receptors and the catalytic component of adenylyl cyclase are implicated in the loss of β₂-AR adenylyl cyclase stimulation in the myometrium after long-term exposure to a selective β₂-AR agonist, as it has been shown to occur at the end of pregnancy (Litime et al., 1989).

The originality of our work is to demonstrate for the first time that, at the end of pregnancy, the myometrial β₃-AR seems to be less prone to sustained agonist-induced desensitisation. Such refractoriness of β₃-AR to desensitisation has already been suggested in other models. For example, it was reported that both the rat β₃-AR expressed in white adipocytes (Granneman, 1992) and the human β₃-AR transfected in Chinese Hamster Ovary (β₃-CHO) cells (Nantel et al., 1993) were completely resistant to short-term agonist-promoted desensitisation, whereas β₂-CHO cells were not (Chambers et al., 1994). Our functional studies showed that human myometrial β₃-AR is resistant to desensitisation. The biochemical approach demonstrated the lack of increased PDE4 activity, consequent to a long-term treatment of myometrial strips with SR 59119A, and the binding experiments showed the lack of downregulation of β₃-AR binding sites. Furthermore, we did not notice any modification in the β₃-AR mRNA levels. These results are conflicting with those of Thomas et al. (1992), who demonstrated that a long-term treatment of 3T3-F442A adipocytes with a β₃-AR agonist led to an upregulation of β₃-AR mRNA and receptor expression. This might suggest that β₃-AR regulation is tissue specific. Indeed Chaudhry & Granneman (1994) demonstrated that β₃-AR desensitisation is dependent upon the cellular background in which this receptor is expressed. In this study, the mechanism responsible for β₃-AR desensitisation does not appear to involve cAMP-dependent phosphorylation or downregulation of the receptor. Oriowo (1998) also described that in the rat lower oesophageal activation of β₃-AR was not linked to either ATPase-, Ca²⁺-dependent K⁺ channels or to the NO/cGMP signalling pathway. We found that, in the human near-term myometrium, β₃-AR stimulation was not associated with an increase of cGMP production (Bardou et al., 2000).

In conclusion, we have demonstrated that, in contrary to β₂-AR, β₃-AR are resistant to agonist-induced desensitisation in near-term human myometrium. Interestingly, long-term treatment with salbutamol did not affect the inhibitory properties and the ability to induce cAMP production of SR 59119A, and reciprocally. This study confirms the potential therapeutic interest of β₃-AR agonists, particularly for women requiring rescue therapy after failure of tocolysis with a β₂-AR agonist. Although their clinical properties remain to be demonstrated, the use of these new molecules could be of interest in pharmacological treatment of preterm labour.

Abbreviations

AR

adrenoceptor

cAMP

cyclic adenosine monophosphate

CRC

concentration–response curve

ICYP

(−)-[¹²⁵I]-iodocyanopindolol

PDE

phosphodiesterase