Inverse Agonism and Neutral Antagonism at a1a- and a1b-Adrenergic Receptor Subtypes

We have characterized the pharmacological antagonism, i.e., neutral antagonism or inverse agonism, displayed by a number of a-blockers at two a1-adrenergic receptor (AR) subtypes, a1aand a1b-AR. Constitutively activating mutations were introduced into the a1a-AR at the position homologous to A293 of the a1b-AR where activating mutations were previously described. Twenty-four a-blockers differing in their chemical structures were initially tested for their effect on the agonistindependent inositol phosphate response mediated by the constitutively active A271E and A293E mutants expressed in COS-7 cells. A selected number of drugs also were tested for their effect on the small, but measurable spontaneous activity of the wild-type a1aand a1b-AR expressed in COS-7 cells. The results of our study demonstrate that a large number of structurally different a-blockers display profound negative efficacy at both the a1aand a1b-AR subtypes. For other drugs, the negative efficacy varied at the different constitutively active mutants. The most striking difference concerns a group of N-arylpiperazines, including 8-[2-[4-(5-chloro-2-methoxyphenyl)-1-piperazinyl]ethyl]-8-azaspiro[4,5] decane-7,9-dione (REC 15/3039), REC 15/2739, and REC 15/3011, which are inverse agonists with profound negative efficacy at the wild-type a1b-AR, but not at the a1a-AR. Adrenergic receptors (ARs) mediate the functional effects of epinephrine and norepinephrine by coupling to several of the major signaling pathways modulated by guanine nucleotide regulatory proteins (G proteins). The AR family includes nine different gene products: three b (b1, b2, b3), three a2 (a2A, a2B, a2C), and three a1 (a1a, a1b, a1d) receptor subtypes. Like all G protein-coupled receptors (GPCR), the ARs share seven hydrophobic regions that form a transmembrane a-helical bundle and are connected by alternating intracellular and extracellular hydrophilic loops. Mutational analysis of the ARs has revealed that the a-helical bundle contributes to form the ligand binding site of the receptor, whereas amino acid sequences of the intracellular regions appear to mediate the interaction of the receptor with G proteins as well as with different signaling and regulatory proteins (Wess, 1997). Both selective and nonselective antagonists for different AR subtypes are widely used in a variety of pathological conditions, including hypertension, heart failure, and prostate hypertrophy as well as in mental diseases such as depression. Several studies have demonstrated that b-blockers can behave either as neutral antagonists or inverse agonists at the wild-type b2-AR or at a constitutively active b2-AR mutant (Samama et al., 1993b; Chidiac et al., 1994). However, inverse agonism at other AR subtypes has been less extensively investigated. It has been previously reported that a small range of a-blockers could inhibit the agonist-independent phospholipase C as well as phospholipase D responses mediated by constitutively active mutants of the a1b-AR (Lee et al., 1997). A recent study demonstrated that some a-blockers can inhibit the spontaneous activity of the a1d-AR subtype (Garcı́a-Sáinz and Torres-Padilla, 1999). The main aim of this study was to characterize the pharmacological antagonism, i.e., neutral antagonism or inverse agonism, displayed by a number of a-blockers at two a1-AR subtypes, a1aand a1b-AR. To achieve this goal, constitutively activating mutations were first introduced into the a1a-AR at the position homologous to A293 of the a1b-AR where activating mutations were previously described (Kjelsberg et al., 1992). Several ligands were then screened for their effect on the agonist-independent activity of both the wild-type a1aand a1b-ARs and their constitutively active mutants. Our study provides a number of findings that might represent a solid basis to further elucidate the activation process of the a1aand a1b-AR subtypes, and the mechanism This work was supported by the Fonds National Suisse de la Recherche Scientifique (Grant 31-51043.97) and by the European Community (Grant BMH4-CT97-2152). ABBREVIATIONS: AR, adrenergic receptor; GPCR, G protein-coupled receptor; DMEM, Dulbecco’s modified Eagle’s medium; [I]HEAT, [I]iodo-2-[b-(4-hydroxyphenyl)-ethyl-aminomethyl]tetralone; IP, inositol phosphate; CAM, constitutively active mutant; R, inactive; R*, active. 0026-895X/99/050858-09$3.00/0 Copyright © The American Society for Pharmacology and Experimental Therapeutics All rights of reproduction in any form reserved. MOLECULAR PHARMACOLOGY, 56:858–866 (1999). 858 at A PE T Jornals on O cber 3, 2017 m oharm .aspeurnals.org D ow nladed from of action of drugs acting at these receptors as well as their structure-activity relationships. Experimental Procedures Mutagenesis and Transfections. The cDNA encoding human a1a-AR (Schwinn et al. 1995; cDNA was a kind gift from Dr. J.P. Hieble, SmithKline Beecham, Van Nuys, CA) or hamster a1b-AR (Cotecchia et al., 1992) were mutated by polymerase chain reactionmediated mutagenesis technique with Taq DNA polymerase. The mutated DNA fragments obtained were digested with the appropriate enzymes and cloned into the expression vector pRK-5 containing the wild-type a1aor a1b-AR cDNA. Recombinant clones were isolated and sequenced. COS-7 cells grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and gentamicin (100 mg/ml) were transfected with the diethylaminoethyl-dextran method. The transfected DNA ranged between 0.5 and 3 mg/10 cells. Ligand Binding. Membranes derived from cells expressing the a1a-AR subtypes and their mutants were prepared as previously described (Cotecchia et al., 1992). The binding was performed at 25°C in 50 mM Tris-HCl (pH 7.4), 150 mM NCl, and 5 mM EDTA. For saturation binding experiments of [I]iodo-2-[b-(4-hydroxyphenyl)-ethyl-aminomethyl]tetralone ([I]HEAT), the radioligand concentration ranged from 12 to 400 pM (150-ml assay volume) and prazosin (10 M) was used to determine nonspecific binding. For saturation binding experiments of [H]prazosin, the radioligand concentration ranged from 25 to 4400 pM (300-ml assay volume) and phentolamine (10 M) was used to determine nonspecific binding. In competition-binding experiments, the final concentrations of [I]HEAT and [H]prazosin were 80 and 400 pM, respectively. In some competition-binding experiments, the concentration of [I]HEAT was 10 pM (500-ml assay volume). Results of ligand binding experiments were analyzed with Prism 2.0 (GraphPAD Software, San Diego, CA). Inositol Phosphate (IP) Measurement. Transfected COS-7 cells (0.15 3 10) seeded in 12-well plates were labeled for 15 to 18 h with myo-[H]inositol (New England Nuclear, Boston, MA) at 5 mCi/ml in inositol-free DMEM supplemented with 1% fetal bovine serum. Cells were preincubated for 10 min in PBS containing 20 mM LiCl and then treated for 45 to 100 min with different ligands. Total IPs were extracted and separated as previously described (Cotecchia et al., 1992). Molecular Modeling of Ligands. The protonated structures of the ligands considered in this study were fully optimized by means of semiempirical molecular orbital calculations (AM1) (Dewar et al., 1985) with the MOPAC 6.0 (QCPE 455) program. QUANTA molecular modeling package (release 96; Molecular Simulation Inc., Waltham, MA) was used for building and analyzing the molecular structures. Statistical Analysis. Statistical analysis was perfomed as indicated in the figure legends with Prism 2.0 (GraphPAD Software). Materials. COS-7 cells were obtained from American Type Culture Collection (Rockville, MD). DMEM, gentamicin, fetal bovine serum, and restriction enzymes were purchased from Life Technologies, Inc. (Grand Island, NY). Taq polymerase was obtained from Roche Laboratories (RotKruez, Switzerland). [I]HEAT, [H]prazosin, and [H]inositol were obtained from New England Nuclear. (2)-Epinephrine and corynanthine were purchased from Sigma Chemical Co. (St. Louis, MO). 5-Methylurapidil, prazosin, WB 4101, phentolamine, spiperone, S-(1)-niguldipine were obtained from Research Biochemicals Inc. (Natick, MA). (1)-Cyclazosin and (2)-cyclazosin were a gift from Dr. D. Giardinà (University of Camerino, Camerino, Italy). Indoramin and AH11110A were a gift from Dr. J.P. Hieble, (SmithKline Beecham), and BE 2254 was a gift from Dr. D. Hoyer (Novartis, Basel, Switzerland). BMY 7378, WAY 100635, SNAP 5089, RS-17053, alfuzosin, terazosin, tamsulosin, REC 15/ 2739, REC 15/3039 (8-[2-[4-(5-chloro-2-methoxyphenyl)-1-piperazinyl-ethyl]-8-azaspiro[4,5]decane-7,9-dione), REC 15/2869, REC 15/ 3011, and REC 15/2615 were obtained from Recordati (Milano, Italy). Results and Discussion Activating Mutations of a1aand a1b-AR Subtypes. One strategy to identify inverse agonists is to enhance the basal activity of GPCR by introducing activating mutations and to screen drugs for their ability to inhibit the agonistindependent activity of the constitutively active receptor mutants (CAM). We have previously reported that in the a1b-AR mutations of A293 at the C-terminal end of its 3i loop with any amino acid enhanced the constitutive activity of the receptor, and was highest when alanine was substituted with lysine or glutamic acid (Kjelsberg et al., 1992). To identify inverse agonists at both the a1aand a1b-AR subtypes, we constructed CAMs of the a1a-AR by mutating A271 (homologous to A293 of the a1b-AR) to lysine or glutamic acid. As shown in Fig. 1, mutations of either A271 or A293 markedly enhanced the basal activty of a1aand a1b-ARs, respectively, resulting in increased agonist-independent accumulation of IPs. Saturation binding analysis of [I]HEAT or [H]prazosin indicated that the expression levels of the wild-type and CAM receptors were good, ranging between 1.7 and 4.3 pmol/mg protein (Table 1). Our findings support the notion previously suggested for other GPCRs (Wess, 1997) that the C-terminal end of the 3i loop plays a crucial role in the conformational switch underlying the transition between the inactive (R) and active states (R*) of the a1a-AR subtype. Two main differences can be highlighted between a1aand a1b-ARs. First, the agonist-independent activity of both the wild-type a1b-AR and its CAMs was significantly higher than that of the wild-type a1a-AR or its CAMs. Second, for both the a1a-AR and its CAMs the epinephrine-induced IP accumulation above basal was significantly higher than that of the a1b-AR or its CAMs (Fig. 1). This suggests that the agonistoccupied a1a-AR has greater efficacy in activating phospholipase C than the a1b-AR, whereas its spontaneous or mutation-induced isomerization toward the R* is lower. Our findings are in agreement with those from a previous study (Theroux et al., 1996) describing the coupling efficiencies of Fig. 1. Constitutively active mutants of the a1a-AR and a1b-AR subtypes. A271 of a1a-AR and A293 of a1b-AR were mutated into lysine and glutamic acid. Receptors expression in COS-7 cells ranged from 1.5 to 2.5 pmol/mg protein. Control (Con) indicates cells not expressing the receptors. IP ([H]IP) accumulation was measured in cells expressing the wild-type or mutated receptors after incubation in the absence (Bas) or presence of 10 M epinephrine (Epi) for 45 min. Values represent means 6 S.E. of three independent experiments. Statistical significance was analyzed by unpaired Student’s t test. a, P , .05 Bas of the mutants was compared with that of their respective wild-type receptor. b, P , .05 Bas or Epi of the a1b-AR, A293K, and A293E were compared with those of a1a-AR, A271K, and A271E, respectively. c, P , .05 Epi was compared with Bas of each respective receptor. Inverse Agonism at a1-Adrenergic Receptor Subtypes 859 at A PE T Jornals on O cber 3, 2017 m oharm .aspeurnals.org D ow nladed from different a1-AR subtypes expressed in HEK 293 or SK-N-MC cells. In that study, the agonist-induced IP response mediated by the a1a-AR was higher, whereas its agonist-independent activity was lower compared with a1b-AR expressed at a similar level. Inhibition of Receptor-Mediated Basal IP Accumulation. Twenty-four a-blockers differing in their chemical structures were tested for their effect on the basal activity of the constitutively active A271E and A293E mutants expressed in COS-7 cells (Fig. 2). All the ligands used in this study, except REC 15/3039, were previously described for their structure, binding affinity at recombinant as well as native a1-AR subtypes, and some of their pharmacological effects in different tissues (Michel et al., 1995; Giardinà et al., 1996; Leonardi et al., 1997; Testa et al., 1997). Our results show that the majority of a-blockers displayed inverse agonism as demonstrated by their ability to decrease the basal activity of both CAMs. However, the various a-blockers differed in their negative efficacy and some of these differences depended on the a1-AR subtype. Drugs with the highest negative efficacy (defined as $70% inhibition of the basal activity) at both CAMs included WAY 100635, WB 4101, all the tested quinazolines (prazosin, terazosin, both (1)and (2)-cyclazosin, REC 15/2615, and alfuzosin), indoramin, corynanthine, spiperone, and AH11110A (Fig. 2). For the other drugs, their negative efficacy differed at the two CAMs. The most striking difference concerned some N-arylpiperazines that displayed modest negative efficacy (e.g., 5-methylurapidil, BMY 7378, and REC 15/2869) or neutral antagonism (e.g., REC 15/3039, REC 15/2739, and REC 15/3011) at the A271E mutant. However, the negative efficacy of these compounds was more pronounced at the A293E, resulting in at least 45% inhibition of the receptor-mediated basal activity. For phentolamine, BE 2254, and tamsulosin negative efficacy was also greater at the A293E than at the A271E. In contrast, for S-(1)-niguldipine negative efficacy was greater at the A293E than at the A271E mutant. The concentration-dependence of the inhibitory effect was determined for those ligands that displayed the most profound negative efficacy at both the A293E and A271E receptors. The EC50 values of the compounds in inhibiting the basal activity of the CAMs (Table 2) were in the same order of magnitude as their ligand-binding affinities (Table 3). To further investigate the mode of action of inverse agonists at the two CAMs we focused on prazosin and 5-methylurapidil whose properties were similar at the two mutants, the first being almost a full inverse agonist and the second having only modest negative efficacy. Treatment of cells with prazosin did not have any effect on aluminum fluoride-induced accumulation of IP (results not shown). The inhibition of basal IP accumulation by prazosin was competitively inhibited by increasing concentrations of 5-methylurapidil. The apparent Kb values of 5-methylurapidil calculated according the Schild equation were in good agreement with its binding affinity for the receptor mutants (Table 3) (Kb 5 4.4–5.4 nM and 267–505 nM for the A271E and A293E, respectively). These results confirm that the inhibitory effect of the inverse agonist prazosin is mediated by the receptor and not by other unknown mechanisms on the signaling cascade. To assess whether the effect of the a-blockers observed on the CAMs reflected their behavior at the wild-type receptors, a selected number of drugs were tested for their effects in COS-7 cells expressing the wild-type a1a and a1b-AR sub-