Partial Agonist Activity of Bucindolol Is Dependent on the Activation State of the Human β1-Adrenergic Receptor
Background— In contrast to other β-blockers, bucindolol has failed to reduce mortality in patients with chronic heart failure. It is currently debated whether this is due to partial agonist activity of this agent. We investigated whether conflicting results previously reported concerning the intrinsic activity of bucindolol can be explained by species differences or by different activation states of β-adrenergic receptors (β-ARs) in the respective tissues.
Methods and Results— On isolated right atria from transgenic mice with cardiac overexpression of human β1-ARs, bucindolol led to a greater increase in beating frequency (P<0.05) compared with wild-type mice. The increase amounted to 47% of the effect of xamoterol and was blocked by propranolol. On isolated, electrically stimulated, left ventricular muscle-strip preparations from failing human myocardium, bucindolol did not change the force of contraction under control conditions. In myocardial preparations pretreated with metoprolol (30 μmol/L, 90 minutes, subsequent washout), bucindolol significantly increased the force of contraction (P<0.001 vs control). In nonfailing atrial myocardium, isoproterenol pretreatment (1 μmol/L, 60 minutes) abolished the positive inotropic effect of xamoterol that was present under control conditions (P<0.05 vs control). The inotropic effects of bucindolol or xamoterol were inversely correlated to the inotropic response to forskolin in the respective specimens (r=−0.75 and −0.74, respectively; P<0.005).
Conclusions— We conclude that bucindolol is a partial agonist at the human β1-AR. In human failing myocardium, its partial agonist activity is masked by increased activation states of β-ARs and is unmasked after in vitro pretreatment with metoprolol. Thus, the partial agonist activity of bucindolol is dependent on the activation state of the human β1-AR.
Received February 18, 2003; revision received April 1, 2003; accepted April 4, 2003.
Beta-adrenergic receptor (β-AR)-blocking agents are currently regarded as standard therapy for chronic heart failure.1 Clinical studies, however, have demonstrated differences among β-blockers. Although carvedilol, metoprolol, and bisoprolol have significantly reduced mortality (reviewed in Bristow1), xamoterol2 and, most recently, bucindolol3 have shown adverse or neutral effects on mortality.
The reason why β-blockers differ in their efficacy to reduce mortality in heart failure patients might be related to different intrinsic activities of β-blockers. β-Blockers can be categorized as inverse agonists, neutral antagonists, or partial agonists.4 Whereas inverse agonists reduce the basal activation state of β-ARs, neutral antagonists bind to the receptor without changing its activation state. In contrast, partial agonists increase the activation state of β-ARs, leading to coupling of the receptor to the stimulatory G protein (Gs) with subsequent stimulation of adenylyl cyclase. Recent data from both in vitro and in vivo experiments5–7 suggest that continuous stimulation of cardiac β1-ARs is detrimental to the heart. Xamoterol causes β1-adrenergic stimulation in animal8,9 and human10 tissue, and this partial agonist activity is believed to be the culprit for the disappointing performance of xamoterol in heart failure.2
Whereas it has been clearly demonstrated that xamoterol exerts partial agonist activity at the human β1-AR,9,10 the situation for bucindolol is less clear. Although partial agonist activity has been described in several animal species,8,11 studies investigating the effect of bucindolol on human tissue led to conflicting results. Some groups did not observe partial agonist activity,12,13 whereas we and others observed partial agonist activity of bucindolol in human myocardium.14,15 Partial agonist activity of bucindolol, however, was not constant in human failing myocardium and varied, depending on the respective tissue investigated.14
To determine whether the putative partial agonist activity of bucindolol is species dependent, we compared the effect of bucindolol on atria from wild-type mice (murine β1-AR) and transgenic mice with cardiac-specific overexpression of the human β1-AR. Furthermore, we sought to investigate reasons underlying the great variability of results reported for the intrinsic activity of bucindolol in human myocardium. Determination of partial agonist activity in human myocardium is complicated because of the different degrees of β-adrenergic desensitization in chronic heart failure.1 Desensitization of β-ARs has been shown to convert partial agonist activity of β-AR ligands to inverse agonist activity in Sf9 insect cells.16 Thus, we hypothesized that in human failing myocardium, partial agonist activity of bucindolol in vitro might be masked by desensitization of the β-adrenergic signaling cascade in vivo. To elucidate this issue, we pretreated human failing myocardium with either the agonist isoproterenol or the inverse agonist metoprolol to achieve further desensitization or resensitization of β-ARs, respectively.
The generation of transgenic mice overexpressing the human β1-AR under the control of the α-myosin heavy-chain promoter has been described previously.5 Male wild-type and transgenic littermates derived from crosses of a heterozygous, transgenic (line β1TG4) and wild-type mice were studied at an age of 2 to 4 months.
Human ventricular myocardium was obtained from 4 male patients with heart failure due to ischemic or dilated cardiomyopathy (n=2/2) during heart transplantation (mean±SEM age, 53±5 years). Hearts were used within 1 hour after explantation. Human atrial myocardium was obtained from 19 patients undergoing cardiac arterial bypass grafting or valve replacement (n=15/4; 17 male, 2 female; mean±SEM age, 65±4 years; mean±SEM left ventricular ejection fraction, 64±4%).
Mouse heart experiments were performed as described previously.9 In brief, mouse atria were placed in a carbonated, 35°C tissue bath with modified Tyrode’s solution. Before addition of the pharmacological substances (1 μmol/L, unless indicated otherwise), the bath solution was changed 5 times with 5-minute intervals between the individual washing steps. The atria were allowed to contract spontaneously. The basal frequencies were comparable and not significantly different in the groups studied in Figure 1. The respective values were 402±14 beats per minute (bpm) in the xamoterol group (n=6), 397±15 bpm in the bucindolol group (n=9), and 363±24 bpm in the isoproterenol group (n=4). Propranolol (1 μmol/L) was added 10 minutes before the addition of bucindolol. The median effective concentrations (EC50s) for bucindolol, xamoterol, and isoproterenol were 59±30, 84±4, and 5.4±1.2 nmol/L, respectively.
Experiments on human left ventricular and right atrial myocardium were performed as described previously.14,17 The bath solution was maintained at 37°C, pH 7.4, and aerated with 95% O2 and 5% CO2. Muscles were stretched to the length at which the force of contraction was maximal.
Failing left ventricular myocardium (baseline) was preincubated with isoproterenol (1 μmol/L), metoprolol (30 μmol/L), or vehicle for 90 minutes. After a 4-fold washout of pretreatment agents (pretreatment), muscles were incubated with forskolin (0.3 μmol/L). Bucindolol was then applied in cumulative doses (0.001 to 1 μmol/L). Each concentration was allowed to equilibrate for 30 minutes.
Human atrial myocardium was exposed to isoproterenol (1 μmol/L), metoprolol (30 μmol/L), or vehicle for 60 minutes. After a 4-fold washout of pretreatment agents, forskolin (0.1 μmol/L) was added. Hereafter, xamoterol was added in cumulative concentrations (0.001 to 1 μmol/L). After removal of xamoterol and forskolin, isoproterenol was added in cumulative concentrations (0.001 to 1 μmol/L).
Human atrial myocardium was pretreated as in protocol 2. After removal of pretreatment agents, forskolin was added in cumulative concentrations (0.01 to 10 μmol/L). After maximum force was reached, forskolin was removed by replacing the bath solutions at least 4 times. Finally, isoproterenol was added in cumulative concentrations (0.001 to 10 μmol/L).
To analyze agonist-mediated phosphorylation of the β1-AR, a hemagglutinin tag was added to the human β1-AR at its N-terminus, and the construct was transfected into HEK293 cells with the calcium phosphate method. Forty hours after transfection, the cells were loaded with 140 μCi/mL [32P]orthophosphate in phosphate-free Dulbecco’s modified Eagle’s medium for 2 hours. Labeled cells were stimulated with isoproterenol or bucindolol (10 μmol/L) for 5 minutes and solubilized in lysis buffer containing (in mmol/L, unless indicated otherwise) Tris-HCl 50 (pH 7.4), NaCl 300, EDTA 5, NaN3 0.01%, NaF 50, Na4P2O7 5, Na3VO4 1, Triton X-100 1%, phenylmethylsulfonyl fluoride 1, and iodoacetamide 10; β1-ARs were immunoprecipitated with 12CA5 antibodies (Roche Biosciences) directed against the hemagglutinin tag. Immunoprecipitated receptors were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and phosphoimaging (PhosphorImager) analysis.
Average data are presented as mean±SEM. Statistical analyses (t tests for pairwise comparisons or ANOVA) were used with the InStat software package (GraphPad). Differences were considered significant when P<0.05. Linear and nonlinear regression analyses and calculation of EC50 values were performed with GraphPadPrism (GraphPad).
To determine whether the intrinsic activity of bucindolol on β1-ARs was species dependent, experiments were performed in murine and human myocardium. First, we compared the effect of bucindolol and 2 well-characterized β-adrenergic agonists on the spontaneous beating frequency of isolated right atria from wild-type and transgenic mice overexpressing the human β1-AR. Addition of bucindolol at a concentration of 1 μmol/L resulted in a marked and rapid increase of spontaneous right atrial beating frequency, which was significantly greater in transgenic compared with wild-type mice (Figure 1A and 1B). Bucindolol induced a mean rise of right atrial beating frequency by 78±11 bpm in transgenic animals (P<0.001 vs control; Figure 1B). This increase was completely abolished by the β-AR antagonist propranolol (1 μmol/L). The partial agonist activity of bucindolol amounted to 47% or 38% of the increase in beating frequency exerted by xamoterol or isoproterenol, respectively (Figure 1B). The relatively high partial agonist activity of xamoterol is consistent with the high expression level of β1-ARs, resulting in a high receptor reserve. In addition, as a read-out for the intracellular effects of β1-AR activation, we determined ligand-induced receptor phosphorylation. Hemagglutinin-tagged β1-ARs were expressed in 32P-labeled HEK293 cells and stimulated with bucindolol or isoproterenol, respectively. β1-AR phosphorylation was assessed after immunoprecipitation with an anti-hemagglutinin antibody (Figure 1C). Bucindolol induced rapid and prominent phosphorylation of the human β1-AR. Together with recent data demonstrating cAMP stimulation in response to bucindolol,15 this clearly classifies bucindolol as a partial agonist at the human β1-AR.
We then studied the effects of bucindolol on contractility in human ventricular myocardium. To elucidate whether the variability in the inotropic effects of bucindolol in previous studies on human myocardium12–15 was due to different sensitization states of β-ARs,16 failing myocardium was pretreated with isoproterenol, metoprolol, or vehicle for 90 minutes. The baseline force of contraction was not different among the 3 groups (Table 1). Isoproterenol increased force by 122±26% (P<0.05), and metoprolol reduced force by 68±3% (P<0.001). After washout of the pretreatment drugs, forskolin (0.3 μmol/L) was added to the organ bath to facilitate the coupling of Gs to adenylyl cyclase.18 Under these conditions, bucindolol changed the force of contraction in neither untreated nor isoproterenol-pretreated muscles. In contrast, in muscles pretreated with metoprolol, bucindolol concentration-dependently increased the force of contraction by a maximal 43±9% at 1 μmol/L (P<0.005), indicating partial agonist activity (Figure 2A).
To test whether these findings were specific for bucindolol or might also apply to other partial agonists, we studied the effect of different pretreatments on the inotropic effects of xamoterol in human myocardium from patients without heart failure. As indicated in Table 2, the baseline force of contraction was equal in the 3 groups. As expected from the strong partial agonist xamoterol, this agent exerted a positive inotropic effect in control as well as metoprolol-pretreated myocardium (Figure 2B). In contrast, in isoproterenol-pretreated myocardium, the positive inotropic effect of xamoterol was completely abolished (Figure 2B). Interestingly, the inotropic responses to xamoterol and bucindolol showed a close inverse correlation with the inotropic effects induced by forskolin in the respective tissue specimens (xamoterol, r=−0.74, P<0.0001, Figure 2C; bucindolol, r=−0.75, P<0.005, not shown). Because the sensitivity of adenylyl cyclase to forskolin stimulation is dependent on the amount of precoupled Gsα,18,19 this might indicate that the response to a partial agonist is dependent on the activation state of the receptor.
To further elucidate this issue, we sought to determine the activation and sensitization states of β-ARs after the respective pretreatment protocols in human myocardium. Figure 3 illustrates that after isoproterenol pretreatment, forskolin concentration-response curves were shifted to the left (increased sensitivity, Figure 3A and 3B), whereas isoproterenol concentration-response curves were slightly shifted to the right (decreased sensitivity, Figure 3C and 3D). In contrast, after metoprolol pretreatment, forskolin sensitivity was significantly decreased and isoproterenol sensitivity increased (Figure 3A through 3D). This indicates that agonist-pretreated β-ARs are slightly desensitized but also more activated, whereas metoprolol-pretreated β-ARs are sensitized but less activated compared with control conditions.
The main results of the present study are that bucindolol displays partial agonist activity at the human β1-AR and that in human failing myocardium, this effect is masked by an increased activation state and desensitization of β-ARs and can be unmasked by in vitro resensitization of β-ARs by inverse agonist pretreatment.
Although stimulation of cardiac β-ARs represents the strongest mechanism to increase contractility of the heart,20 in heart failure, chronic stimulation of the β-adrenergic signaling pathway due to increased sympathetic activation contributes to the progression of ventricular dysfunction.1,5 Treatment of heart failure patients with the β-blockers carvedilol, metoprolol, and bisoprolol improves ventricular function and survival.1 In contrast, bucindolol failed to increase survival in heart failure patients in the recent Beta-blocker Evaluation of Survival Trial (BEST).3 Several lines of evidence suggest that partial agonist activity of β-AR antagonists has detrimental effects on the prognosis of heart failure patients. Xamoterol is a partial agonist at the human β1-AR9,10 and leads to excess mortality in heart failure patients.2 In this context, it is of great clinical importance to determine whether bucindolol has partial agonist activity or not.
Different investigations of the intrinsic activity of bucindolol in human myocardium have led to conflicting results. Whereas some groups did not observe partial agonist activity,12,13 we and others observed partial agonist activity of bucindolol in human myocardium.14,15 In samples from failing human hearts, however, bucindolol behaved as a partial agonist in one part and as an inverse agonist in another part of the experiments.14 It appears plausible that these discrepant results might be due to highly variable sample conditions. Chidiac et al16 observed that pretreatment of β2-ARs with an agonist can alter the properties of subsequently applied ligands. In particular, pretreatment with the agonist isoproterenol converted the partial agonist activity of dichloroisoproterenol to inverse agonist activity.16 We reasoned that endogenous catecholamines might exert similar effects in heart failure patients. Therefore, we sought to establish better defined models by (1) investigating under standardized conditions the human β1-AR overexpressed in mouse hearts and (2) standardizing the experiments on human myocardium as far as possible by defined pretreatment protocols, extending the observations of Chidiac et al16 by applying agonist as well as inverse agonist pretreatment.
The model of transgenic mice overexpressing the human β1-AR allowed the generation of robust and reproducible data in a simple physiologic assay.9 In this model, bucindolol displayed pronounced partial agonist activity at the human β1-AR. Moreover, stimulation of HEK293 cells with bucindolol increased β1-AR phosphorylation. Because only β-ARs in the activated conformation are a substrate for receptor phosphorylation by β-AR kinase,21,22 this clearly indicates that the partial agonist activity of bucindolol translates into postreceptor events involving the activation of adenylyl cyclase and cAMP-dependent activation of kinases, ie, β-AR kinase and protein kinase A.22 Phosphorylation of the β-AR is a prerequisite for receptor desensitization by uncoupling from Gs and (in the long term) downregulation by sequestration and internalization.22 This is in concert with results from Asano et al,23 who observed downregulation of β-ARs in chick heart myocytes after bucindolol treatment for 24 hours.
The studies by Chidiac et al16 suggest that partial agonist activity is best discovered in cells with sensitized β-ARs. Accordingly, when human myocardium was pretreated with the inverse agonist metoprolol followed by washout, an unambiguous partial agonist activity of bucindolol was observed. In contrast, in isoproterenol-treated or untreated myocardium, this partial agonist effect was absent. Similarly, agonistic effects of the (stronger) partial agonist xamoterol were observed in metoprolol-treated or untreated human atrial myocardium but not in isoproterenol-pretreated tissue.
The effects exerted by pretreatment are most likely induced by slowly reversible alterations in the activation and desensitization states of the β-adrenergic signal transduction system. The desensitization state of β-ARs—which is thought to be due to phosphorylation and β-arrestin binding—is best assessed by measuring agonist concentration-response curves. These experiments revealed that pretreatment with metoprolol sensitized and pretreatment with isoproterenol desensitized the β-AR-Gs-adenylyl cyclase system. The fact that control human myocardium had an intermediate position in these experiments indicates that the β-adrenergic system is partially desensitized in these samples.
The pretreatment also affected the responsiveness to forskolin. Forskolin directly activates adenylyl cyclase; however, it has a higher efficacy and potency when Gsα is precoupled to the adenylyl cyclase complex.18,19 Accordingly, high-affinity [3H]forskolin binding can be applied to assay Gsα-adenylyl cyclase complexes.19 In cell systems overexpressing the β2-AR, increased basal [3H]forskolin binding indicates that Gsα-adenylyl cyclase complexes reflect constitutive (nonliganded) receptor activity.24 In our functional experiments, the potency of forskolin (EC50 values) was dependent on the type of pretreatment. This indicates that pretreatment affects the amount of Gsα coupled to adenylyl cyclase and thus, the activation state of the system.
Pretreatment affected the responsiveness to isoproterenol and to forskolin in opposite directions. Isoproterenol pretreatment desensitized β-ARs (decreased potency of isoproterenol) but activated the Gsα-adenylyl cyclase system (increased potency of forskolin). In contrast, metoprolol pretreatment increased the potency of isoproterenol but decreased the potency of forskolin. These effects explain the inverse correlation between partial agonist effects of both bucindolol and xamoterol and the potency of forskolin in individual human myocardial samples (Figure 2C). It suggests that partial agonist effects are more readily detected in samples with a low level of β-AR preactivation (as indicated by low sensitivity to forskolin).
These observations might explain the highly variable results that were obtained in earlier studies with partial β-AR agonists in general and with bucindolol in particular on human myocardium. Especially in failing myocardium, detection of partial agonist activity might be critically hampered by pronounced activation and desensitization of β-ARs due to chronic stimulation with endogenous catecholamines.1 This is supported by the fact that in failing myocardium, maximal stimulation of β-ARs with isoproterenol did not change the inotropic effect of bucindolol compared with control conditions. In stark contrast, under resensitized conditions (ie, metoprolol pretreatment), the partial agonist activity of bucindolol could be unmasked in this tissue. Bucindolol is structurally similar to carvedilol. However, despite these similarities, carvedilol never displayed partial agonist activity, even in systems extremely sensitive to test this property, including studies from our own laboratory9 and experiments in the pithed rat.8 This is in agreement with studies on human myocardium, where carvedilol is equivocally classified as an inverse agonist.1,13,14,17
Regarding the ongoing debate on why bucindolol is less effective than other β-blockers in reducing mortality in heart failure patients,1,3 it should be noted that in human failing myocardium, the partial agonist activity of bucindolol was observed only after metoprolol pretreatment. However, the extensive sympathetic activation in heart failure1 might be mimicked best by isoproterenol pretreatment—a condition wherein partial agonism was observed for neither bucindolol nor xamoterol. In contrast, the therapeutically beneficial compounds metoprolol9,13,14 and bisoprolol9,17 behave as inverse agonists in a broad range of experimental tissues and conditions.
The present data add further evidence to the hypothesis that the initial activation state of human β-ARs rather than the species investigated determines the intrinsic activity of a β-AR ligand.16 Furthermore, these observations provide an explanation for different results obtained in different studies and experimental systems concerning the intrinsic activity of bucindolol. We conclude that bucindolol is a partial agonist at the human β1-AR and that detection of partial agonist activity in human failing myocardium is dependent on the activation state of the human β1-AR. ⇓
These studies were supported by grants from the Deutsche Forschungsgemeinschaft (DFG; SFB 355 and Leibniz Award) and the Fonds der Chemischen Industrie. Dr Maack is a recipient of the Emmy Noether Program of the DFG.
Bristow MR. Beta-adrenergic receptor blockade in chronic heart failure. Circulation. 2000; 101: 558–569.
Engelhardt S, Hein L, Wiesmann F, et al. Progressive hypertrophy and heart failure in beta1-adrenergic receptor transgenic mice. Proc Natl Acad Sci U S A. 1999; 96: 7059–7064.
Communal C, Singh K, Sawyer DB, et al. Opposing effects of beta1- and beta2-adrenergic receptors on cardiac myocyte apoptosis : role of a pertussis toxin-sensitive G protein. Circulation. 1999; 100: 2210–2212.
Willette RN, Aiyar N, Yue TL, et al. In vitro and in vivo characterization of intrinsic sympathomimetic activity in normal and heart failure rats. J Pharmacol Exp Ther. 1999; 289: 48–53.
Engelhardt S, Grimmer Y, Fan GH, et al. Constitutive activity of the human beta1-adrenergic receptor in beta1-receptor transgenic mice. Mol Pharmacol. 2001; 60: 712–7.
Andreka P, Aiyar N, Olson LC, et al. Bucindolol displays intrinsic sympathomimetic activity in human myocardium. Circulation. 2002; 105: 2429–2434.
Chidiac P, Nouet S, Bouvier M. Agonist-induced modulation of inverse agonist efficacy at the beta2-adrenergic receptor. Mol Pharmacol. 1996; 50: 662–669.
Darfler FJ, Mahan LC, Koachman AM, et al. Stimulation of forskolin of intact S49 lymphoma cells involves the nucleotide regulatory protein of adenylate cyclase. J Biol Chem. 1982; 257: 11901–11907.
Alousi AA, Jasper JR, Insel PA, et al. Stoichiometry of receptor-Gs-adenylate cyclase interactions. FASEB J. 1991; 5: 2300–2303.
Brodde OE, Michel MC. Adrenergic and muscarinic receptors in the human heart. Pharmacol Rev. 1999; 51: 651–690.
Benovic JL, Staniszewski C, Mayor F, Jr, et al. Beta-adrenergic receptor kinase: activity of partial agonists for stimulation of adenylate cyclase correlates with ability to promote receptor phosphorylation. J Biol Chem. 1988; 263: 3893–3897.
Hausdorff WP, Caron MG, Lefkowitz RJ. Turning off the signal: desensitization of beta-adrenergic receptor function. FASEB J. 1990; 4: 2881–2889.
Loisel TP, Ansanay H, Adam L, et al. Activation of the beta(2)-adrenergic receptor-Galpha(s) complex leads to rapid depalmitoylation and inhibition of repalmitoylation of both the receptor and Galpha(s). J Biol Chem. 1999; 274: 31014–31019.