Naunyn-Schmiedeberg's Archives of Pharmacology

, Volume 369, Issue 2, pp 151–159

Comparative pharmacology of human β-adrenergic receptor subtypes—characterization of stably transfected receptors in CHO cells

  • C. Hoffmann
  • M. R. Leitz
  • S. Oberdorf-Maass
  • M. J. Lohse
  • K.-N. Klotz
Original Article

DOI: 10.1007/s00210-003-0860-y

Cite this article as:
Hoffmann, C., Leitz, M.R., Oberdorf-Maass, S. et al. Naunyn-Schmiedeberg's Arch Pharmacol (2004) 369: 151. doi:10.1007/s00210-003-0860-y

Abstract

Although many β1-receptor antagonists and β2-receptor agonists have been used in pharmacotherapy for many years their pharmacological properties at all three known subtypes of β-adrenergic receptors are not always well characterized. The aim of this study was, therefore, to provide comparative binding characteristics of agonists (epinephrine, norepinephrine, isoproterenol, fenoterol, salbutamol, salmeterol, terbutalin, formoterol, broxaterol) and antagonists (propranolol, alprenolol, atenolol, metoprolol, bisoprolol, carvedilol, pindolol, BRL 37344, CGP 20712, SR 59230A, CGP 12177, ICI 118551) at all three subtypes of human β-adrenergic receptors in an identical cellular background. We generated Chinese hamster ovary (CHO) cells stably expressing the three β-adrenergic receptor subtypes at comparable levels. We characterized these receptor subtypes and analyzed the affinity of routinely used drugs as well as experimental compounds in competition binding studies, using the non-selective antagonist 125I-cyanopindolol as a radioligand. Furthermore, we analyzed the β-receptor-mediated adenylyl cyclase activity in isolated membranes from these cell lines. The results from our experiments show that all compounds exhibit distinct patterns of selectivity and activity at the three β-receptor subtypes. In particular, a number of β2- or β3-receptor agonists that are inverse agonists at the other subtypes were identified. In addition, β1-receptor antagonists with agonistic activity at β2- and β3-receptors were found. These specific mixtures of agonism, antagonism, and inverse agonism at different subtypes may have important implications for the therapeutic use of the respective compounds.

Keywords

β-Adrenergic receptor Antagonist Inverse agonist G protein-coupled receptors Stable transfection β1-receptor β2-receptor β3-receptor 

Introduction

Adrenergic receptors belong to the superfamily of seven transmembrane spanning proteins that activate heterotrimeric guanine nucleotide binding proteins. The three known β-adrenergic receptor subtypes are designated β1, β2, and β3 and are all positively coupled to adenylyl cyclase. The β1-receptor is the predominant subtype in mammalian heart (Brodde 1993) whereas the β2-receptor represents the predominant subtype in most vascular and bronchial smooth muscle cells (Guimaraes and Moura 2001; Kotlikoff and Kamm 1996). The β3-subtype is primarily found in white and brown adipose tissues (Krief et al. 1993). The β-adrenergic receptors signaling pathways play a key role in regulating cardiac function (reviewed by Brodde and Michel 1999) and β-adrenergic receptor mediated relaxation of smooth muscles plays an important physiological role in the regulation of vascular and bronchial tone (reviewed by Guimaraes and Moura 2001; Kotlikoff and Kamm 1996). These findings have led to the development of a large number of compounds modulating the activity of β-adrenergic receptors for the clinical treatment of diseases including hypertension, heart failure (reviewed by Berkin and Ball 2001; Krum 2003) and asthma (Waldeck 2002). The pharmacological characteristics of β1-receptors and β2-receptors have been studied exhaustively, and a large number of clinically relevant agonists and antagonists have been characterized in competition binding and functional studies. However, this information is spread over a vast number of studies and derived from different organ preparations and different species (Michel 1991). Therefore, a direct comparison of such data is often difficult and possibly not relevant for the human receptor subtypes. After the cloning of the human β-receptor subtypes their stable expression in eukaryotic expression systems like Chinese hamster ovary (CHO) cells became possible (Tate et al. 1991). However, detailed studies with such human receptor models typically focused on the characterization of single receptor subtypes (Liggett 1992; Lattion et al. 1999; Candelore et al. 1999; Del Carmine et al. 2002).

In spite of the fact that a large number of β-receptor agonists and antagonists have been in clinical use for decades there is a lack of comparative pharmacological data about their affinity, subtype selectivity and efficacy at the human receptor subtypes. Such data appear to be important to better understand the potential clinical benefits and risks of individual compounds. The β2-selectivity of β-agonists in the treatment of asthma, for instance, is a critical determinant of the risk for cardiac side effects. In an in vitro study of the so-called β2-selective agonist broxaterol, which is not in clinical use, we discovered that this compound binds to β1-and β2-receptors with similar affinity (Conti et al. 1998). However, it turned out that it exhibits agonistic activity at the β2-subtype only whereas it was an antagonist at β1-receptors (Conti et al. 1998), which makes it a functionally selective compound. In order to provide similar information for the characteristics of relevant compounds acting at β-receptors we have now compared the detailed pharmacological characteristics of a number of clinically and experimentally relevant compounds on all three human β-adrenergic receptor subtypes under carefully standardized conditions in an identical cellular background.

Materials and methods

Materials

(-)-Epinephrine, (-)-norepinephrine, CGP-20712 ((±)-2-hydroxy-5-[2-[[2-hydroxy3-[4-[1-methyl-4-(trifluoromethyl)-1H-imidazol-2-yl]phenoxy]propyl]amino]ethoxy]-benzamide), (-)-isoproterenol, metoprolol, propranolol, and terbutaline were purchased from Sigma (Taufkirchen, Germany). Atenolol, bisoprolol, BRL-37344 ((R*,R*)-(±)-4-[2-[(2-(3-chlorophenyl)-2-hydroxyethyl)amino]propyl]phenoxyacetic acid), CGP-12177 (4-[3-[(1,1-dimethylethyl)amino]2-hydroxypropoxy]-1,3-dihydro-2H-benzimidazol-2-one)), SR 59230A (1-(2-ethylphenoxy)-3-[[(1S)-1,2,3,4-tetrahydro-1-naphthalenyl]amino]-(2S)-2-propanol), pindolol, formoterol and salbutamol were purchased from Tocris/Biotrend (Cologne, Germany). The sources of further compounds were: carvedilol (Boehringer Mannheim, Mannheim, Germany); fenoterol, ICN (Eschwege, Germany); alprenolol, Astra (Wedel, Germany); ICI-118551 ((±)-1-[2,3-(dihydro-7-methyl-1H-inden-4-yl)oxy]-3-[(1-methylethyl)amino]-2-butanol), RBI/Sigma (Taufkirchen, Germany). Salmeterol was kindly provided by Dr. H. Krohn (GlaxoSmithKline, Bad Oldeslohe, Germany). Broxaterol was kindly synthesized by Prof. M. De Amici (Istituto di Chimica Farmaceutica e Tossicologica, Università degli Studi di Milano, Italy). (-)-3-125I-Iodocyanopindolol (125I-CYP) was from Amersham Biosciences (Freiburg, Germany; specific activity, 2,200 Ci/mmol). [α-32P]ATP was from PerkinElmer LifeScience (Rodgau-Jügesheim, Germany). Cell culture media and fetal calf serum were from PanSystems (Aidenbach, Germany), penicillin (100 U/ml), streptomycin (100 μg/ml), L-glutamine and G-418 were purchased from Gibco-Life Technologies (Eggenstein, Germany). All other materials were from sources as described earlier (Klotz et al. 1998).

cDNA of human β-adrenergic receptors

cDNAs coding for human β-adrenergic receptors in pcDNA3 expression vectors were verified by sequencing and comparison with the respective GeneBank entries. The translated amino acid sequences correspond to the published sequences for the β1-adrenergic receptor (Frielle at al. 1987; GenBank entry J03019), β2-adrenergic receptor (Schofield et al. 1987; GenBank entry Y00106), and β3-adrenergic receptor (Emorine et al. 1989; GenBank entry X72861). With respect to polymorphisms, the β adrenergic receptors used in this study correspond to the following variants: β1-receptor 49-Ser, 389-Gly; β2-receptor 16-Arg, 27-Gln, 164-Thr; β3-receptor 64-Trp. All of the variants correspond to the sequences originally termed wild-type.

Stable transfection of cells

Chinese hamster ovary cells (CHO-K1 cells; CCL61, American Type Culture Collection, Rockville, MD, USA) were transfected with plasmid DNA for stable expression using the calcium phosphate precipitation method (Chen and Okayama 1987) as described for the rat A1 adenosine receptor (Freund et al. 1994). Positive clones were selected with 600 μg/ml of the neomycin analog G-418, and single clonal lines were isolated by limiting dilution. Expression of the receptor was verified by radioligand binding.

Cell culture and membrane preparation

Chinese hamster ovary cells stably transfected with human β-adrenergic receptor subtypes were grown adherently and maintained in Dulbecco’s Modified Eagle’s Medium with nutrient mixture F12 (DMEM/F12), containing 10% fetal calf serum, penicillin (100 U/ml), streptomycin (100 μg/ml), L-glutamine (2 mM) and geneticin (G-418, 0.2 mg/ml) at 37°C in 5% CO2/95% air. Cells were split 2 or 3 times weekly at a ratio between 1:5 and 1:15. In order to harvest cells the culture medium was removed, cells were washed twice with PBS and membranes were prepared or cells were frozen on the dishes for later preparation of membranes. Crude membrane fractions were prepared from fresh (measurement of adenylyl cyclase) or frozen cells (radioligand binding) according to two different protocols, which have been described recently (Klotz et al. 1998). The resulting membrane pellets were resuspended in 50 mM Tris/HCl buffer pH 7.4 to give a final protein concentration of 1–2 mg/ml. Protein content was determined by the method of Bradford (1976) with bovine serum albumin (Sigma) as a standard.

Radioligand binding studies and adenylyl cyclase activity

The radioligand binding experiments were performed with membranes prepared as described above. Assays were done in a volume of 200 μl in 50 mM Tris/HCl, pH 7.4 (assay buffer) in the presence of 100 μM GTP to ensure monophasic binding curves for agonists. For saturation binding experiments at human β1- and β2-receptors we used up to 400 pM 125I-CYP and for β3-receptors up to 1,500 pM 125I-CYP. Non-specific binding was determined in the presence of 10 μM alprenolol. For competition binding we used 50 pM 125I-CYP in the case of β1- and β2-receptors, or 80 pM 125I-CYP for β3-receptors. For most of the competition binding experiments membranes with intermediate receptor expression (β1: 367±75 fmol/mg protein, β2: 282±19 fmol/mg protein, β3: 377±82 fmol/mg protein) were used. For selected compounds it was demonstrated that higher receptor expression did not affect Ki-values (data not shown). Membranes were incubated for 90 min at 30°C, filtered through Whatman GF/C filters, and washed three times with ice-cold assay buffer. Samples were counted in a γ-counter (Wallac 1480 wizard 3”). KD-values for 125I-CYP were calculated by nonlinear curve fitting with the program SCTFIT (De Lean et al. 1982). Ligand IC50-values were calculated using Origin 6.1 (OriginLab Corporation, Northampton, MA, USA) and were transformed to Ki-values according to Cheng and Prusoff (1973).

Adenylyl cyclase activity in cell membranes was determined according to Jakobs et al. (1976). Fifty micrograms of membrane protein were added to an incubation mixture with final concentrations of 50 mM Tris/HCl pH 7.4, 100 μM cAMP, 0.2% BSA, 10 μM GTP, 100 μM ATP, 1 mM MgCl2, 100 μM IBMX, 15 mM phosphocreatine, and 300 U/ml of creatine kinase. Membranes were incubated with about 200,000 cpm of [α-32P]-ATP for 20 min in the incubation mixture as described (Klotz et al. 1985). Accumulation of [α-32P]-cAMP was linear over at least 20 min under all conditions. The reaction was stopped by addition of 400 μl 125 mM ZnAc-solution and 500 μl 144 mM Na2CO3. Samples were centrifuged for 5 min at 14,000 rpm in a laboratory microcentrifuge. Eight hundred microliters of the resulting supernatant were finally applied to aluminia WN-6 (Sigma) columns that were eluted twice with 2 ml 100 mM Tris/HCl pH 7.4. The eluates were counted in a β-counter (Beckmann LS 1801). Data were calculated using Origin 6.1. The basal and isoproterenol-stimulated adenylyl cyclase activity, respectively, in the different cell clones characterized in Table 3 was (all values in pmol/mg membrane protein/min) 30.9±4.5 and 65.4±5.8 (β1, clone 2); 4.3±0.5 and 11.1±1 (β2, clone 1); 27.0±2.8 and 36.0±1.5 (β2, clone 2) 104±4.5 and 138±6.8 (β2, clone 3); 26±2.8 and 93.7±6.0 (β3, clone 2).

Results

The CHO cells utilized for stable transfection with β-adrenergic receptor subtypes appear to have negligible amounts of endogenous β-adrenergic receptors according to two criteria: first, in untransfected CHO cells 125I-CYP binding was only slightly different from non-specific binding, corresponding to an endogenous β-adrenergic receptor level of <15 fmol/mg protein (less than 5% of the binding observed in transfected cells), and second, no isoproterenol-induced adenylyl cyclase stimulation was observed with membranes prepared from untransfected cells (data not shown). After transfection with either of the human β-adrenergic receptor subtypes substantial specific binding of the antagonist 125I-CYP was observed and receptor-mediated stimulation of adenylyl cyclase demonstrated functional coupling of the β-adrenergic receptors to adenylyl cyclase.

Receptor expression

CHO cells were transfected with one of the three human β-adrenergic receptors and several clones stably expressing one of the individual subtypes were obtained. The saturation binding data shown in Table 1 reveal that for all three subtypes cell lines with comparable expression level (about 300 fmol/mg protein) were obtained. For both β1- and β2-receptors we obtained also cell lines with a 30-fold difference in receptor expression levels. Receptor expression was routinely controlled and was constant when cells were maintained in culture for 6–8 weeks (15–25 passages). Fig. 1 shows a representative saturation binding experiment for each human β-receptor subtype stably expressed in CHO cells. The KD-values in Table 1 show that 125I-CYP binds to β1- and β2-receptors with similar affinity which is in agreement with previous binding data for the rat subtypes (Neve et al. 1986). For different clones of β1- and β2-receptors KD-values of 62–95 pM and 45–51 pM respectively were determined. At the β3-receptor subtype a 3.5–6-fold lower affinity was observed (KD-values 210–360 pM).
Table 1

Characteristics of different CHO cell lines stably expressing the human β-adrenergic receptor subtypes. Receptor expression was determined by binding of the antagonist radioligand 125I-CYP. Each saturation experiment was repeated three times in triplicate

Receptor subtype

Clone

KD-value (pM)

95% Confidence limit (pM)

Bmax (fmol/mg protein) ± SD

β1

1

61.9

41.8–90.7

9,130±3,060

2

95.3

78.4–112.2

367±75

β2

1

47.9

43.2–53.2

40±3

2

44.6

42.1–47.1

282±19

3

50.7

49.0–52.4

1,280±135

β3

1

209

126–346

341±27

2

360

321–399

377±82

Fig. 1

125I-CYP saturation binding to human β-adrenergic receptor subtypes. The results of representative saturation binding experiments using membranes from CHO-cells stably expressing the respective receptor subtype and the antagonist radioligand 125I-CYP are shown as specific (black circles) and non-specific binding (white circles). Saturation binding at human β1-, β2-, and β3-receptors is shown. The results of the particular experiment shown were for β1-receptor, KD-value 68 pM and Bmax 380 fmol/mg protein; β2-receptor 59 pM and 1,350 fmol/mg protein and, β3-receptor 310 pM and 250 fmol/mg protein. The results of the all saturation binding experiments are summarized in Table 1

Competition binding

We continued to investigate the pharmacological profile of human β-adrenergic receptors in CHO cells in competition binding experiments. In addition to the endogenous agonists epinephrine and norepinephrine, clinically relevant drugs as well as experimental compounds were investigated for their binding profile at the three subtypes. Since binding experiments were performed in the presence of 100 μM GTP all Ki-values for agonists reflect low affinity binding.

For the endogenous agonists epinephrine and norepinephrine we observed no selectivity with respect to binding at the β1-adrenergic receptors. At the β2-adrenergic receptor epinephrine was more potent than norepinephrine (see Table 2). The β2-receptor exhibits a 35-fold higher affinity for epinephrine than for norepinephrine. At the β3-adrenergic receptor we observed the opposite, i.e. a very low binding affinity for epinephrine. According to our data, the β3-subtype exhibited a 30-fold higher affinity for norepinephrine compared to epinephrine.
Table 2

Binding affinities from competition experiments for agonists and antagonists at human β-adrenergic receptor subtypes. 50 to 80 pM 125I-CYP were used as radioligand. Experiments were done in the presence of 100 μM GTP. Ki-values were calculated from IC50-values using the Cheng-Prussow equation and represent geometric mean values of at least three different experiments done in triplicate

β1-adrenergic receptor

β2-adrenergic receptor

β3-adrenergic receptor

Compound

Ki (nM)

95% Confidence limit

Ki (nM)

95% Confidence limit

Ki (nM)

95% Confidence limit

Norepinephrine

3,570

2,440–5,210

26,400

23,400–29,900

4,300

4,240–4,360

Epinephrine

3,970

2,840–5,530

735

510–1,050

126,000

116,000–136,000

Isoproterenol

224

145–343

458

377–556

1,570

1,370–1,810

Fenoterol

13,600

11,100–16,700

719

565–915

55,700

45,100–68,800

Salbutamol

2,440

1,770–3,380

2,170

1,600–2,950

53,700

37,100–77,700

Salmeterol

1,600

1,110–2,290

24.6

15.9–37.8

7,180

4,680–11,000

Formoterol

1,710

1,430–2,060

2,570

1,590–4,160

8,090

4,380–14,900

Terbutaline

31,300

19,000–51,600

15,400

11,900–20,000

79,800

40,400–157,000

BRL-37344

37,900

34,100–41,800

9,170

7,520–11,200

430

389–475

Alprenolol

5.8

4.5–7.5

1.2

0.8–1.8

35.0

15.2–80.8

Pindolol

2.6

1.3–5.2

4.8

3.8–6.1

44.1

30.9–63.0

Carvedilol

1.7

1.3–2.2

1.1

0.8–1.6

247

234–261

Atenolol

388

298–504

8,140

5,970–11,100

65,100

47,100–89,900

Bisoprolol

22.4

18.1–27.7

1,150

1,100–1,190

9,070

7,370–11,200

Metoprolol

47.0

27.3–81.0

2,960

2,090–4,190

10,100

8,130–12,400

S-Propranolol

1.8

1.2–2.8

0.8

0.6–1.0

186

134–259

CGP-20712

4.7

4.0–5.5

4,040

2,790–5,860

2,360

1,770–3,150

SR 59230A

16.4

14.0–19.2

61.9

30.2–127

122

39.2–383

CGP-12177

4.5

3.3–6.1

4.3

2.3–8.1

77.1

70.8–84.0

ICI-118551

49.5

40.0–61.4

0.7

0.4–1.1

611

531–703

Broxaterol

1,310

930–1,860

1,290

916–1,810

3,990

3,470–4,590

Among the clinically used β2-adrenergic receptor agonists only salmeterol and to a lesser extent fenoterol exhibited some selectivity for the β2-adrenergic receptor. From this group salmeterol was also the most selective compound for β2- vs. β3-adrenergic receptors. Both salbutamol and terbutaline showed a similar affinity profile, i.e., almost the same affinity for β1- and β2-receptors and low affinity for the β3 subtype.

Within the group of β-blockers pindolol, carvedilol, and S-propranolol were found to be non-selective for β1- vs. β2-adrenergic receptors, while atenolol, bisoprolol, and metoprolol were up to 63-fold selective for β1- vs. β2-adrenergic receptors. Of this group, bisoprolol was the most selective compound with respect to selectivity for β1- vs. β3-adrenergic receptors. All three compounds were 170–400-fold selective for β1 compared to the β3 receptor. It is interesting to note that Gille et al. (1985) found very similar values for β1- and β2-affinities for S-propranolol in functional experiments with preparations from human myocardium.

Among the experimental substances CGP-20712 was the most selective compound for β1- vs. β2- and β3-receptors, while ICI-118551 was the most selective compound for β2- vs. β1- and β3-receptors.

According to our data, BRL-37344 exhibited the highest selectivity for the human β3-adrenergic receptor (90-fold vs. β1, 20-fold vs. β2). However, the highest affinity for β3-adrenergic receptors was observed for alprenolol and pindolol (Table 2).

Adenylyl cyclase stimulation

Next, we tested adenylyl cyclase stimulation by isoproterenol, norepinephrine, and epinephrine in isolated membranes from stably transfected CHO cells with comparable receptor expression for each receptor subtype (Fig. 2). All concentration-response curves for these agonists were shifted to the left compared to the respective binding curves. However, the rank order of potency in the binding and cyclase experiments was identical (Fig. 2, Table 2) and was found to be Iso >NA = A at the β1-receptor, Iso = A >>NA at the β2-receptor, and Iso >NA >A at the β3-receptor. Isoproterenol, norepinephrine and epinephrine were full agonists at each receptor.
Fig. 2

Agonist-stimulated adenylyl cyclase activity in membranes isolated from CHO-cells stably expressing human β-adrenergic receptor subtypes. Membranes were prepared as described in Materials and methods. Adenylyl cyclase stimulation was calculated as % of the maximal stimulation achieved by each compound. Maximal adenylyl cyclase was achieved with concentrations corresponding to 100-fold Ki-values as given in Table 2. Adenylyl cyclase stimulation is shown for isoproterenol (squares), norepinephrine (triangles), and epinephrine (circles) at human β1-, β2- and β3-receptor. Data are representative of three to four experiments done in duplicate

We then characterized all compounds investigated in Table 2 and determined their efficacy and classification as agonists, antagonists or inverse agonists. Therefore, we investigated single maximally effective concentrations of the ligands in adenylyl cyclase assays in cell lines expressing the different β-adrenergic receptors at comparable expression levels. Compound concentrations used in this assay were generally 100×Ki-value (as given in Table 2), or maximally 1 mM. As judged by the data for epinephrine, norepinephrine and isoproterenol (Fig. 2), at these concentrations a full effect on adenylyl cyclase should be achieved for all compounds. The maximal stimulation with isoproterenol and unstimulated basal values were determined in each individual experiment and were used as a reference for the efficacy of the other compounds. The data were calculated as percent stimulation by isoproterenol. For compounds that reduced the adenylyl cyclase activity below the unstimulated basal activity the degree of inverse agonist activity was determined as percent reduction of basal adenylyl cyclase activity.

The results of these experiments are summarized in Table 3. Functionally, the compounds fall into several different categories. A group of five compounds, namely isoproterenol, epinephrine, norepinephrine, formoterol and fenoterol are agonists at all three receptor subtypes. The so-called β2-agonists salbutamol, terbutaline, and broxaterol behaved as partial agonists at the β2-receptor and as opposed to fenoterol were antagonists at the β1-receptor. Their efficacy at the β3-receptor subtype ranged from 42–87%. The long-acting asthma drug salmeterol was found to be the only β2-receptor agonist with no agonistic activity at the other receptor subtypes, where it acted as an inverse agonist. All other compounds tested were antagonists or inverse agonists at β1- and β2-receptors and exhibited characteristic differences at the β3-receptor, with activities ranging from partial agonistic activity (e.g., CGP-12177) to inverse agonistic activity (ICI-118 551).
Table 3

Adenylyl cyclase responses of human β-adrenergic receptor subtypes to selected ligands. Membranes were prepared from cells with comparable receptor expression level (β1: 367±75 fmol/mg, β2: 282±19 fmol/mg, β3: 377±82 fmol/mg protein) as described in Materials and methods. Adenylyl cyclase stimulation represents percentage maximal stimulation achieved by isoproterenol (25 μM for β1, 50 μM for β2, 130 μM for β3). Inverse agonist activities were calculated as % reduction of the basal adenylyl cyclase activity. Compound concentrations used correspond to 100-fold Ki-values as given in Table 2, but did not exceed 1 mM. Data represent mean values ± SEM of 3–10 different experiments done in triplicate. ND not determined

Compound

β1-adrenergic receptor (mean ± SEM)

β2-adrenergic receptor (mean ± SEM)

β3-adrenergic receptor (mean ± SEM)

Isoproterenol

100

100

100

Basal

0

0

0

Epinephrine

133±8

110±8

106±7

Norepinephrine

123±7

103±9

122±8

Fenoterol

66±6

76±5

110±6

Formoterol

58±6

ND

139±12

Salbutamol

2±5

33±3

87±6

Terbutaline

6±16

41±5

45±4

Broxaterol

−11±9

20±11

42±4

Salmeterol

−33±8

44±8

−13±16

BRL−37344

−5±4

−7±2

28±3

CGP−12177

−19±13

−32±10

36±7

Alprenolol

−24±1

ND

11±3

Pindolol

−25±10

−34±8

13±4

SR 59230A

−19±4

ND

5±1

Atenolol

−23±11

−26±8

7±7

Carvedilol

−28±9

−30±6

2±6

Metoprolol

−24±11

−34±7

2±4

Bisoprolol

−33±9

−30±5

−1±7

Propranolol

−35±6

−35±10

−4±3

CGP−20712

−25±5

−30±7

−20±6

ICI−118551

−22±6

−32±11

−30±3

Since differential expression of β2-adrenergic receptors had been shown to modify agonist stimulation of adenylyl cyclase (Bouvier et al. 1988; Lohse 1992; Whaley et al. 1994) we wanted to investigate whether receptor expression also had an influence on inverse agonist activity. Therefore, we measured modulation of adenylyl cyclase activity in membranes from CHO cells expressing β2-adrenergic receptors at different levels. The results are summarized in Table 4. The agonist epinephrine was found to be a full agonist and its efficacy was independent of receptor expression within the tested range. The partial agonist fenoterol, however, varied in its efficacy with respect to receptor expression, exhibiting partial agonist activity at low receptor expression levels and an increasing efficacy with higher receptor expression levels. A similar variation in efficacy was observed for compounds that exhibited inverse agonist activity. They were found to behave as neutral antagonists at low receptor expression and inverse agonism became more prominent with increasing receptor expression levels.
Table 4

Adenylyl cyclase responses of human β-adrenergic receptor subtypes at different expression level to selected agonists and inverse agonists. Membranes were prepared as described in Materials and methods. Adenylyl cyclase stimulation was calculated as % maximal stimulation achieved by 50 μM isoproterenol. Inverse agonist activities were calculated as % reduction of the basal adenylyl cyclase activity. Compound concentrations used correspond to 100-fold Ki-values as given in Table 2, but did not exceed 1 mM. Data represent mean values ± SEM of 3–10 different experiments done in triplicate

β2-adrenergic receptor

Compound

40±3 fmol/mg protein (mean)

282±19 fmol/mg protein (mean)

1,260 ±135 fmol/mg protein (mean)

Isoproterenol

100

100

100

Basal

0

0

0

Epinephrine

102±8

110±8

103±4

Norepinephrine

88±14

103±9

103±6

Formoterol

91±9

ND

106±7

Fenoterol

53±8

76±5

88±4

CGP−12177

9±1

−32±10

−66±2

Alprenolol

−1±3

ND

−62±11

Pindolol

−9±7

−34±8

−66±2

Propranolol

7±3

−35±10

−80±1

Metoprolol

6±4

−34±7

−88±1

Bisoprolol

5±1

−30±5

−86±2

Atenolol

3±4

−26±8

−63±2

ICI−118551

1±5

−32±12

−84±2

CGP−20712

0±1

−30±7

−72±2

SR 59230A

0±3

ND

−68±9

Carvedilol

−7±3

−30±6

−75±2

Discussion

Stable transfection of receptors into CHO cells provides useful models for the investigation of the human receptor subtypes in an identical cellular background. In this study we establish such transfected cell systems for the three known subtypes of human β-adrenergic receptors along with a comparative pharmacological characterization using selected experimental compounds as well as clinically relevant drugs. In general, the stably transfected receptors exhibited the expected pharmacological profile with respect to ligand binding. However, with respect to functional assays several interesting discoveries were made.

In our study we used 125I-CYP as a radioligand at the different β-adrenergic receptor subtypes, which is widely used in many laboratories. Levin et al. (2002) characterized the 49-Ser variant of human β1-adrenergic receptor that was also used in the present study and published a 125I-CYP KD-value of 57 pM in stably transfected HEK-293 cells, which is in excellent agreement with our data (see Table 1). In another study a β1 KD-value of 17 pM was found for 125I-CYP (Tate et al. 1991) which is somewhat lower than our values of 62 and 95 pM for two different cell clones respectively. In the same study Tate et al. (1991) had observed KD-values of 31 pM at β2- and 230 pM at β3-receptors, which is again in close agreement with our data in Table 1.

The pharmacological profile of the three human β-adrenergic receptors in stably transfected CHO cells was established in competition binding experiments. Since binding experiments were performed in the presence of 100 μM GTP all Ki-values for agonists reflect low affinity binding. The results of the competition binding experiments are summarized in Table 2. In general, these data are in good agreement with individual results reported in different publications. For example, recently Levin et al. (2002) published Ki-values for low affinity binding of isoproterenol (268 nM), norepinephrine (3,600 nM), and metoprolol (54 nM) at the 49-Ser variant of the human β1-adrenergic receptor. Del Carmine et al. (2002) have recently published a binding study at the human β2-adrenergic receptor stably expressed in CHO cells. Since in that study binding affinities were determined in the absence of GTP, the data for agonist binding can not be directly compared to our data, however the rank order of isoproterenol >epinephrine = fenoterol >salbutamol >norepinephrine is identical to our findings. Data for antagonist binding (e.g., ICI-118,551 with a KD-value of 0.68 nM and S-propranolol with a KD-value of 0.43 nM), however, are similar to our findings.

Michel (1991) had summarized ligand binding data for β1- and β2-adrenergic receptor that were available from the literature. Since these data were from different species and tissues, they cannot be directly compared to our data. Again, however, the overall pharmacological profiles are comparable to our findings. Isoproterenol was about equally potent on β1- and β2-receptors, epinephrine bound slightly better to the β2-receptor and norepinephrine was 10-fold selective for β1- vs. β2-receptor. Among the clinically used β2-agonists, terbutaline and salbutamol were found to be non-selective whereas fenoterol was more potent at the β2-receptor. All these characteristics were reproduced for the transfected human receptors in CHO cells in this study. In addition, we found the expected β2-selectivity for sameterol whereas formoterol turned out to be nonselective (Table 2). The selectivity profile for β-AR antagonists reported by Michel (1991) was also found to be similar for the non-selective compounds pindolol and S-propranolol as well as atenolol, bisoprolol, and metoprolol which were β1 selective. Overall it appears that in the case of β1- and β2-adrenergic receptors species differences play only a minor role.

Less information is available about the pharmacological profile of the β3-adrenergic receptor. At the human β3-adrenergic receptor we observed a surprisingly low binding affinity for epinephrine. According to our binding data the β3-receptor exhibits a 30-fold higher affinity for norepinephrine over epinephrine. The low affinity of epinephrine at the β3–receptor was shown before (Liggett 1992) although the difference between epinephrine and norepinephrine was less pronounced than in our study.

Clinically used antagonists showed the expected selectivity patterns with β1-selectivity of atenolol, bisoprolol and metoprolol. Both carvedilol and S-propranolol do not discriminate β1- and β2-receptors but bind with ≥100-fold lower affinity to the β3-subtype. Of the experimental compounds CGP-20712 and ICI-118551 were confirmed as β1- and β2-selective antagonists, respectively. In contrast, SR 59230A did not show the expected β3-selectivity. In our hands this compound was about 4- and 7-fold more potent at the β1-subtype than at β2-and β3-receptors respectively.

Receptor signaling was investigated in adenylyl cyclase assays using cell lines with similar expression levels (~300 fmol/mg membrane protein) for the three β-receptor subtypes. The affinity profile for the endogenous agonists and isoproterenol is shown in Fig. 2. From these data it can be concluded that the human β1-receptor does not distinguish between the endogenous ligands norepinephrine and epinephrine, while the β2-receptor would be the ‘adrenergic’ receptor and the β3-receptor would be a ‘noradrenergic’ receptor. Since the expression of the human β3-receptor has been reported to be mainly in adipose tissue it might escape adrenal stimulation and have its main role in responding to stimulated sympathetic neurons (Giacobino 1995). Steinle et al. (2003) reported that β3-receptors are also expressed on cultured human retinal endothelial cells and stimulation of these receptors by BRL-37344 stimulates cell migration, an early marker for angiogenesis. Therefore, they conclude that sympathetic nerves might also play a role in vascular disorders of the eye.

From Fig. 2 it can also be concluded that at compound concentrations of 100×Ki-value the adenylyl cyclase activation should be maximal at least for full agonists. The data in Table 3 show a measure for efficacy of each compound relative to isoproterenol as a reference. According to their efficacy at the different β-receptor subtypes the tested compounds could be divided into several groups. The clinically relevant β2-receptor agonists differed markedly in their efficacy at β1-receptors and β3-receptors. Fenoterol and formoterol were the only compounds to exhibit agonistic activity at all receptor subtypes, salbutamol and terbutaline were agonists at β2- and β3-receptors and antagonists at β1-receptors, whereas salmeterol was only agonistic at β2-receptors. Such mixed activity for β2-agonists was already described for the experimental compound broxaterol (Conti et al. 1998), but we did not find evidence in the literature for similar behavior of other β2-agonists. However, our findings may be of importance for clinical use of those compounds as the β1-activity of sympathomimetics determines the cardiac side effects. The ‘functional’ β2-selectivity of salbutamol and terbutaline may be superior to binding selectivity as it should not result in cardiac side effects even at higher doses.

When further analyzing Table 3 we found a large group of compounds showing inverse agonism at β1- and β2-receptors but differing in their efficacy at β3-receptors. BRL-37344 was the only compound that is close to being a neutral antagonist at β1- and β2- and was found to be functionally selective for β3-receptors. Previously, this compound has been described as an agonist at all β-receptor subtypes (Blin et al. 1993). Currently we can not explain the discrepancy to our findings. Although pindolol is generally claimed to be a partial agonist it showed inverse agonism at β1- and β2-receptors and only 13% agonistic activity at β3-receptors in comparison to isoproterenol. All other experimental and clinically used β-blockers behaved as antagonists or inverse agonists at all three receptor subtypes. Del Carmine et al. (2002) have recently published a thorough study on the behavior of compound efficacy at human β2-receptors expressed in CHO cells. Overall, we find an identical activity pattern in our study although the absolute degree of efficacy was different. These differences in efficacy may be related to different receptor expression levels. As seen in Table 4 the efficacy for fenoterol does indeed increase with increasing receptor expression. Del Carmine et al. measured adenylyl cyclase stimulation in cells expressing 8.4 pmol receptor/mg membrane protein while we used cells expressing about 0.3 pmol/mg protein, thus, this difference in receptor expression can easily explain the differences in efficacy. Table 4 shows that inverse agonistic activity can be best observed at high receptor density. Since inverse agonists are assumed to promote an inactive conformation of the receptor (Samama et al. 1994) their effect is most likely more striking due to a higher number of spontaneously active receptors with increasing receptor expression.

In summary, we provide cellular systems that allow the comparison of β-adrenergic receptor ligands directly at the human receptor subtypes in an identical cellular background. The stably transfected receptors exhibit the expected pharmacological profile with respect to ligand binding. In functional assays these model cells allow for a detailed investigation of ligand properties and with such analysis some surprising discoveries were made in this study.

Acknowledgements

We thank Ms. Martina Fischer and Mr. Nico Falgner for technical assistance. This study was supported by the BIOMED 2 program “Inverse agonism. Implications for drug design” and the Leibnitz award of the Deutsche Forschungsgemeinschaft.

Copyright information

© Springer-Verlag 2004

Authors and Affiliations

  • C. Hoffmann
    • 1
  • M. R. Leitz
    • 1
  • S. Oberdorf-Maass
    • 1
  • M. J. Lohse
    • 1
  • K.-N. Klotz
    • 1
  1. 1.Institut für Pharmakologie und ToxikologieUniversität WürzburgWürzburgGermany

Personalised recommendations