Dual Inhibition of β-Adrenergic and Angiotensin II Receptors by a Single Antagonist
A Functional Role for Receptor–Receptor Interaction In Vivo
Background— Although the renin–angiotensin and the β-adrenergic systems are interrelated, a direct interaction between β-adrenergic receptors (βARs) and angiotensin II type 1 receptors (AT1Rs) has not been identified.
Methods and Results— Here, we provide evidence for a functional and physiological interaction between 2 G protein–coupled receptors: the βAR and the AT1R. Selective blockade of βARs in mouse cardiomyocytes inhibits angiotensin-induced contractility with an IC50 that is similar to its inhibition of isoproterenol-mediated contractility. Furthermore, administration of the angiotensin receptor blocker valsartan to intact mice results in a significant reduction in the maximal response to catecholamine-induced elevation of heart rate. The mechanism for this transinhibitory effect of β-blockers and angiotensin receptor blockers is through receptor–G protein uncoupling; ie, β-blockers interfere with AT1R-Gq coupling, and valsartan interferes with βAR-Gs coupling. Finally, we demonstrate that AT1Rs and βARs form constitutive complexes that are not affected by ligand stimulation. As a result of these interactions, a single receptor antagonist effectively blocks downstream signaling and trafficking of both receptors simultaneously.
Conclusions— We show that direct interactions between βARs and AT1Rs may have profound consequences on the overall response to drugs that antagonize these receptors.
Received July 16, 2003; revision received August 8, 2003; accepted August 8, 2003.
Heart failure is a progressive disorder that involves dysfunction of the renin–angiotensin system and the adrenergic nervous system.1,2 Sustained adrenergic drive after injury to the heart results in abnormalities of β-adrenergic receptor (βAR) signaling (downregulation of β1ARs and uncoupling of both β1 and β2ARs from G proteins3,4), as well as a diminished number of angiotensin II (Ang II) type 1 receptors (AT1Rs).1 Despite evidence for positive feedback regulation between the renin–angiotensin system and the βAR systems,5 to date there is no proof of direct cross talk at the receptor level between β-adrenergic and angiotensin receptors.
Both βARs and AT1Rs belong to the superfamily of G protein–coupled receptors (GPCRs), which, once activated, interact with effector molecules through coupling to G proteins. AT1Rs translate the actions of the agonist Ang II by coupling to the heterotrimeric Gq/11 family of G proteins and activation of phospholipase Cβ (PLCβ), whereas βARs mediate the actions of norepinephrine through Gs-dependent activation of adenylyl cyclase. Although receptors in the GPCR superfamily were initially believed to function as monomeric entities, a growing body of evidence suggests that they exist as homodimers or heterodimers.6
In the present study, we demonstrate the existence of a direct interaction between AT1Rs and βARs and show that interfering specifically with the signaling of one receptor (ie, either the βAR or the AT1R) results in the uncoupling and inhibition of signaling by the reciprocal, interacting receptor.
Preparation of Adult Cardiac Myocytes and Contractility Studies
Mice used in these studies were female adult wild-type C57BL/6 mice, 3 to 6 months of age and weighing 30 to 40 g (Jackson Laboratories, Bar Harbor, Maine). Animals were handled according to approved protocols and animal welfare regulations of the Institutional Review Board at Duke University Medical Center. Myocytes were isolated as described previously7 and visualized with an inverted microscope (Nikon Eclipse TE 300). Single-cell contraction was measured by video edge detection, and recordings were made under basal conditions and after administration of the different agents, as described in the figure legends.
Cell Culture and Transfection
COS-7 and human embryonic kidney (HEK) 293 cells were grown in DMEM or modified Eagle’s medium, respectively. Human umbilical vein endothelial cells (HUVECs) were cultured in EGM-2 complete medium. The rat Ang II type 1A–green fluorescent protein (AT1R-GFP) and HA-tagged AT1R cDNAs were a generous gift from Dr Larry Barak (Duke University, Durham, NC), and the human FLAG–β2AR was a generous gift from Dr Robert Lefkowitz (Duke University, Durham, NC). Transient transfections were performed on 60% to 80% confluent monolayers in 100-mm dishes by use of FUGENE6 (Roche Molecular Biochemicals). Agonist stimulations were performed at 37°C in serum-free media after preincubation with the indicated inhibitors.
HUVECs or COS-7 cells were serum-starved for 2 hours before stimulation. Reactions were rapidly terminated, and samples were separated by SDS-PAGE, transferred to nitrocellulose membranes, and probed for phosphorylated extracellular signal–regulated kinase (ERK) using a polyclonal rabbit phospho-ERK1/2–specific antibody (Cell Signaling Technology). Blots were visualized with electrochemiluminescence (Amersham Pharmacia), and the autoradiographs were quantified by use of a Fluor-S MultiImager (Bio-Rad). Blots were then stripped and reprobed with an anti-ERK1/2 antibody (Upstate Biotechnology, Inc) to normalize the level of phosphorylated ERK to total ERK.
Ligand Binding Studies
COS-7 cells were plated into 24-well dishes and allowed to reach confluence. Binding reactions were performed in 0.25 mL binding buffer as described previously.8 125I-labeled angiotensin 0.5 nmol/L or 125I-labeled cyanopindolol (CYP) 125 pmol/L and various concentrations of unlabeled competing ligands were added, then plates were incubated on ice for 4 hours. Nonspecific binding was determined by the addition of 100 μmol/L unlabeled Ang II or propranolol, respectively. For internalization studies, COS-7 cells were serum-starved for 2 hours and stimulated with the different agents, as indicated in the figure legend. Protein concentrations were determined with the Bradford assay (Bio-Rad). Data were analyzed by use of Prism 3.0 (GraphPad).
Inositol Phosphate Measurements
COS-7 cells were grown in 12-well dishes and incubated with 2 μCi/well myo-[3H]-inositol (76 Ci/mmol, Amersham) in serum-free medium for 24 hours. Samples were preincubated with propranolol for 30 minutes before stimulation with Ang II for 60 minutes more. Inositol phosphate levels were measured exactly as described previously.8
Preparation of Cardiac Membranes
Crude cardiac membranes were prepared from excised hearts as described previously.9
[35S]-labeled GTPγS binding to isolated cardiac membranes was assayed in a total reaction volume of 50 μL as described previously.10
HEK 293 cells expressing AT1R-GFP or HA-AT1R were plated on 35-mm glass-bottomed culture dishes and serum-starved for 3 hours before stimulation. Live cells were treated with the different agents as described in the figure legends. Fixing and dual staining of cells were performed as described previously.11 Images were collected by use of an Olympus 1×70 laser scanning confocal microscope.
Heart Rate and Blood Pressure Measurements in Intact Mice
Mice were anesthetized with ketamine (100 mg/kg) and xylazine (2.5 mg/kg) and underwent bilateral vagotomy before catheterization to prevent reflex inhibition in heart rate as described previously.9 The left axillary artery was cannulated for continual recording of blood pressure and heart rate. Drugs were administered through the right external jugular vein. Control animals were injected with vehicle, followed by graded dose of isoproterenol (ISO; 50 to 10 000 pg). Two other groups of animals were injected with a single dose of valsartan (250 μg) or propranolol (1 μg) before administration of ISO.
Data are expressed as mean±SEM. Statistical significance was determined by a 1-way ANOVA or a 2-way repeated-measures ANOVA when appropriate. Post hoc analysis was performed with a Tukey-Kramer multicomparison test or a Newman-Keuls test when appropriate. A value of P<0.05 was considered significant.
To test the existence of possible cross-regulation between AT1Rs and βARs in the heart, we measured the effect of β-blockers on Ang II–mediated contractility of freshly isolated mouse cardiomyocytes. Under basal conditions, stimulation of the cells with Ang II resulted in a significant increase in cell shortening and rate of cell shortening (−dL/dt) (Figure 1, A and B), similar to the response observed by stimulation with the βAR agonist ISO.7 Remarkably, coadministration of the nonselective β-blocker propranolol together with Ang II abolished the Ang II effect, demonstrating a direct action of β-blockers on the Ang II–mediated contractile response (Figure 1, A and B). Because β1ARs are the predominant subtype of βARs in the heart, we tested the effect of the selective β1AR inhibitor metoprolol on the Ang II–mediated increase in contractility. As depicted in Figure 1B, metoprolol also abolished the Ang II–mediated response. Neither of the β-blockers had any effect on the basal levels of contractility. To confirm that the actions of β-blockers on the AT1R were receptor-mediated, we tested the effect of propranolol on a non–receptor-mediated augmentation in myocyte contractility induced by the Na+/K+-ATPase inhibitor ouabain. As shown in Figure 1B, ouabain induced a substantial increase in contractility that was not significantly affected by the presence of the β-blocker.
To discriminate between βAR-mediated effects of propranolol versus βAR-independent actions (eg, direct antagonism of angiotensin receptors or nonspecific effects on membrane fluidity), we measured the IC50 for inhibition of the ISO- and Ang II–enhanced contractility by propranolol. As shown in Figure 1C, increasing concentrations of propranolol resulted in comparable inhibition of both ISO- and Ang II–mediated contractility. Furthermore, the IC50 values for propranolol inhibition of the effects of ISO and Ang II were not significantly different (3.0×10−7 and 5.3×10−7 mol/L, respectively, P=NS). The same result was obtained for inhibition in the rate of contractility in the presence of the β-blocker (Figure 1D). The ability of β-blockers to inhibit both Ang II– and ISO-stimulated myocyte contractility within the same range of concentrations supports our hypothesis that propranolol exerts its effect on angiotensin signaling via the βAR and not through binding to the angiotensin receptor.
To directly exclude the possibility that β-blockers inhibit Ang-mediated signaling by interfering with Ang II binding to its receptor, we performed competitive radioligand binding analyses in intact COS-7 cells. COS cells express β2ARs and AT1Rs at comparable levels, with little expression of β1ARs. As shown in Figure 2A, increasing concentrations of unlabeled Ang II effectively displaced 125I-Ang II binding. Conversely, both propranolol and the β2AR antagonist ICI-118,551 (ICI) were unable to compete with the labeled ligand, suggesting that β-blockers do not interfere with binding of Ang II to the AT1R. In a similar manner, we found that the Ang II receptor blocker valsartan, which binds specifically to the AT1R,12 did not displace the β2AR antagonist 125I-cyanopindolol from the endogenously expressed β2ARs (Figure 2B). Together, these data demonstrate that propranolol and valsartan bind specifically to the βAR and AT1R, respectively, and that the transinhibitory effect on their reciprocal pathways is exerted through specific blockade at a point that is distal to ligand binding.
To test whether selective antagonists for one receptor are able to uncouple the reciprocal receptor from binding to its cognate G protein, we assessed receptor–G protein coupling in mouse cardiac membranes. As depicted in Figure 3A, stimulation with Ang II enhanced AT1R–Gq coupling, as determined by [35S]GTPγS loading of Gq. However, preincubation of the membranes with propranolol completely prevented AT1R–Gq coupling. Consistent with receptor uncoupling, inositol phosphate formation after Ang II stimulation in COS-7 cells was abrogated in the presence of either propranolol (Figure 3B) or ICI (not shown). To test whether blockade of the AT1R induced a similar effect on coupling of βARs to Gs, cardiac membranes were stimulated with ISO in the absence or presence of valsartan. ISO-induced coupling of βAR to Gs, which resulted in a significant increase in [35S]GTPγS loading, was completely blocked in the presence of valsartan (Figure 3C). Consistent with βAR–Gs uncoupling, ISO-mediated generation of the second messenger cAMP was impaired in the presence of valsartan in COS-7 cells endogenously expressing both receptors (Figure 3D).
To test whether receptor transinhibition, either by Ang II receptor blockers or β-blockers, affects downstream signaling by β2ARs and AT1Rs, we measured ERK mitogen-activated protein kinase activation in intact COS-7 cells endogenously expressing both receptors. Stimulation of cells with either ISO or Ang II induced a significant increase in ERK phosphorylation (Figure 4, A and B) in a dose- and time-dependent manner (not shown), confirming the existence of both receptors in these cells. Similar to the observations made in the contractility studies, pretreatment of cells with propranolol abolished the Ang II–mediated activation of ERK (Figure 4B, left). In a reciprocal manner, stimulation of the cells with ISO in the presence of valsartan abrogated the ISO-mediated activation of the ERK pathway (Figure 4A). As expected, the β1AR-selective antagonist metoprolol had no effect on ERK activation by either ISO or Ang II (Figure 4A, right), consistent with the fact that COS-7 cells do not endogenously express detectable levels of β1ARs. The presence of primarily β2ARs in COS-7 cells was further confirmed by the observation that the β2AR-selective inverse agonist ICI blocked the Ang II response (Figure 4B). These data confirm that the effect of β-blockers on Ang II–mediated signaling is specific to the subtype of receptors expressed in the tissue studied.
The ability of the antagonists used in the present study to inhibit reciprocal receptor signaling was found to depend primarily on the expression levels of both receptors. For example, stimulation of HUVECs with ISO resulted in a marked elevation of ERK phosphorylation (Figure 4C). However, these cells did not respond to stimulation by Ang II, and the ISO-induced ERK activation was not sensitive to treatment with valsartan. Complementary binding studies showed that as opposed to the heart, HUVECs express βARs and AT1Rs at a ratio of ≈46:1 (79 versus 1.7 fmol/mg protein, respectively).
The finding that pharmacological blockade of either the βAR or the AT1R results in functional uncoupling of the reciprocal receptor to its cognate G protein led us to postulate that this effect is mediated through a direct interaction between AT1R and βARs. To determine whether AT1Rs and βARs form stable complexes, we used the strategy of differential epitope tagging and selective coimmunoprecipitation.6 FLAG-epitope–tagged β2ARs and HA-tagged AT1Rs were expressed alone or together, at similar low expression levels, and the cells were exposed to Ang II, ISO, or propranolol. As shown in Figure 5A, immunoprecipitation of FLAG-β2ARs resulted in coimmunoprecipitation of HA-AT1Rs in the absence of ligand (lane 3). Furthermore, the amount of complex precipitated was not affected by the presence of either receptor agonist or antagonist (lanes 4 to 6), suggesting that these receptors are assembled as oligomers before their localization on the plasma membrane.13
To further determine whether there is a direct interaction between βARs and AT1Rs, we examined whether stimulation or blockade of 1 receptor affected the ability of the other receptor to undergo a process of internalization after agonist stimulation.14,15 HEK 293 cells expressing similar levels of endogenous βARs and transfected GFP-tagged AT1Rs were stimulated with Ang II in the absence or presence of propranolol. Internalization of the GFP-AT1R was visualized in the form of aggregates of receptors under the laser scanning confocal microscope. As expected, GFP-AT1Rs underwent marked internalization after Ang II stimulation (Figure 5B, left). However, pretreatment of endogenous βARs with propranolol prevented Ang II–induced GFP-AT1R internalization, and the receptors remained on the surface of the plasma membrane (Figure 5B, right).
Analysis of AT1R internalization showed that AT1Rs internalize by either Ang II or ISO stimulation (Fig 5, C and D). Importantly, only 30% to 40% of AT1Rs internalized in response to ISO, compared with ≈80% in response to Ang II. Because βARs and AT1Rs display different internalization patterns on the basis of their ability to bind β-arrestin,16 the presence of AT1R–βAR dimers may alter the recycling properties of the AT1R to that of a βAR.
Because endocytosis of GPCRs is preceded by phosphorylation of the receptor, it is possible that βAR-mediated activation of PKA may promote phosphorylation, desensitization, and internalization by means of heterologous desensitization.17 To test this possibility, we examined the effect of the PKA-selective inhibitor H-89 on ISO-induced internalization of the AT1R. As seen by both confocal microscopy (Figure 5C) and ligand binding studies (Figure 5D), inhibition of PKA did not prevent internalization of the AT1R in response to ISO, supporting the idea that a direct interaction between the 2 receptors, rather than the stimulation of protein kinase activity, is responsible for the effect on receptor trafficking.
The effect of valsartan on internalization of βARs was assessed in HEK 293 cells, which did not show detectable levels of AT1Rs by ligand binding. Cells transfected with either FLAG-β2AR alone or with both FLAG-β2AR and GFP-AT1R were stimulated with ISO in the presence of valsartan. Cells that expressed only β2ARs (red) showed marked internalization after ISO stimulation (Figure 5E, white arrows). However, internalization in cells expressing both βARs (red) and GFP-AT1Rs (green) in response to ISO was significantly impaired in the presence of valsartan.
To determine whether our in vitro findings could be translated into the in vivo context, we tested the effect of a single dose of valsartan on the ISO-stimulated elevation of heart rate in intact, vagotomized, wild-type mice. As shown in Figure 6A, increasing doses of ISO yielded a marked elevation in heart rate, with half of the maximal response occurring at a dose of 1 ng per mouse. Pretreatment with a single dose of propranolol resulted in a marked shift of the ISO response curve to the right, as expected from a classic competitive β-antagonist. In contrast, a single dose of valsartan resulted in a significant 25% reduction in heart rate, with half the maximal response occurring at a dose of 1.4 ng per mouse. The ability of valsartan to attenuate ISO-stimulated heart rate without altering the half-maximal response indicates a decrease in agonist efficacy without an effect on agonist potency. Administration of valsartan or propranolol had no effect on peripheral blood pressure compared with the control group, suggesting a direct effect of valsartan on the heart (Figure 6B).
The major finding of this study is that AT1Rs can directly interact with both βAR subtypes and that this interaction elicits a phenomenon by which selective βAR blockade inhibits signaling of AT1Rs, whereas selective AT1R blockade inhibits downstream signaling of βARs. Moreover, the mechanism for dual transinhibition of 2 independent receptors by a single antagonist is via functional uncoupling of the signaling receptor from its cognate G protein.
The accepted paradigm for signaling of GPCRs is that these receptors function as single units (monomers), independently capable of coupling to a G protein and activating/inhibiting effector molecules. On the basis of our results and the results obtained for other GPCRs,8 we propose a shift in this paradigm by suggesting that more than 1 receptor may be involved in efficient receptor–G protein coupling, thus providing a new role for in vivo oligomerization of GPCRs. Because one of the receptors in the complex (the “nonsignaling” receptor) has to be free of antagonist to permit coupling and signaling of the ligand-activated receptor, it may have a role in stabilizing the interaction of the activated receptor with its cognate G protein.18 It is therefore possible that blockade of either βARs or AT1Rs induces a conformational change that is no longer favorable to support the interaction of the other receptor with its G protein (ie, AT1R–Gq or βARs–Gs). The nature of conformational changes in the presence of either agonists or antagonists and their effect on the reciprocal receptor is currently under investigation.
Accumulating evidence indicates that oligomer formation by GPCRs adds a level of complexity to their signaling. Here, we demonstrate an interaction between AT1Rs and βARs and present evidence that as a result of this interaction, a single receptor antagonist simultaneously blocks signaling and trafficking of both receptors. These findings are supported by recent observations made in prostate cancer cells, in which blockade of either bradykinin type 1 receptor or bradykinin type 2 receptor resulted in impaired coupling and signaling by the reciprocal receptor without interfering with ligand binding.8 Because our data, as well as other studies,19 indicate that dimerization is a constitutive process, the overall response of an organ to an agonist or an antagonist will most likely depend on the levels and combinations of GPCR complexes it expresses. In this respect, it was recently demonstrated that AT1R–bradykinin receptor dimers increase the efficacy and potency of Ang II20 and that an increase in the number of these dimers mediates Ang II–induced hypersensitivity in preeclampsia.21 Here, we also demonstrate that whereas βARs are sensitive to valsartan in the heart, they are unaffected by this drug in endothelial cells, which express a far greater ratio of βARs to AT1Rs. Together, these findings indicate that expression levels of different receptors have a critical effect on the overall response of a specific tissue to selective receptor antagonists.
Selective blockade of angiotensin and adrenergic signaling pathways in patients with moderate and severe heart failure has been shown to improve survival and quality of life.22–25 Because some of the hallmarks of the failing heart are distinct abnormalities in the βAR system, it is currently believed that many of the beneficial effects of β-blockers are a result of antagonism of the cytotoxic effects of elevated circulating catecholamines and normalization of βAR signaling.2 In view of our in vitro and in vivo data, we propose that β-blockers may have an additional novel role in directly blocking Ang II–mediated pathways, thus gaining control over 2 signaling pathways that play a pivotal role in cardiac function and are strongly implicated in the pathophysiology of the failing heart.
Although there is little direct physical evidence for AT1R–βAR dimerization in the heart, we believe that our data strongly support the existence of direct receptor cross talk in vivo. Therefore, it is possible that transinhibition of receptor signaling has potentially broad implications when the use of AT1R and βAR antagonists is considered in treatment of disease states such as heart failure, particularly because the use of 1 antagonist may block more than 1 signaling pathway. Although caution is warranted with regard to the translation of in vitro data to the outcome of clinical trials, our data may provide a biological explanation for unexpected findings from the study of valsartan in heart failure patients (the Val-Heft clinical trial 6), in which treatment of patients with a combination of β-blockers, valsartan, and ACE inhibitors resulted in an increase in adverse events.26 If indeed AT1R–βAR interactions occur in vivo, as our data suggest, β-blockers in conjunction with valsartan would produce near-complete inhibition of both receptor signaling pathways, because each antagonist blocks not only its own receptor but also the signaling of the reciprocal receptor by a mechanism of transinhibition. The addition of ACE inhibitors would lower the levels of circulating Ang II and norepinephrine through inhibition of the renin–angiotensin system and the sympathetic nervous system.27 Therefore, a combination of all 3 drugs may act together to severely suppress the 2 signaling systems beyond a critical point necessary to maintain homeostasis.
In conclusion, the present study shows that direct interactions between βARs and AT1Rs in vivo have a profound role in determining the overall response to drugs designed to block these receptors. Better understanding of the complex interactions between different GPCRs in vivo may therefore have important implications for the development of highly specified pharmacological treatment for a wide range of cardiovascular disorders.
This work was supported by in part by National Institutes of Health grant HL-56687 and the Burroughs Wellcome Fund (both to Dr Rockman). Dr Rockman is a recipient of a Burroughs Wellcome Fund Clinical Scientist Award in Translational Research. The authors thank Kristine Hesser Porter, Dr Lan Mao, and Weili Zou for excellent technical assistance.
This article originally appeared Online on September 8, 2003 (Circulation. 2003;108:r79–r86).
Ungerer M, Bohm M, Elce JS, et al. Altered expression of β-adrenergic receptor kinase and β1-adrenergic receptors in the failing human heart. Circulation. 1993; 87: 454–463.
Rockman HA, Choi DJ, Akhter SA, et al. Control of myocardial contractile function by the level of beta-adrenergic receptor kinase 1 in gene-targeted mice. J Biol Chem. 1998; 273: 18180–18184.
Esposito G, Rapacciuolo A, Naga Prasad SV, et al. Genetic alterations that inhibit in vivo pressure-overload hypertrophy prevent cardiac dysfunction despite increased wall stress. Circulation. 2002; 105: 85–92.
Feldman DS, Zamah AM, Pierce KL, et al. Selective inhibition of heterotrimeric Gs signaling: targeting the receptor-G protein interface using a peptide minigene encoding the Galpha(s) carboxyl terminus. J Biol Chem. 2002; 277: 28631–28640.
Naga Prasad SV, Barak LS, Rapacciuolo A, et al. Agonist-dependent recruitment of phosphoinositide 3-kinase to the membrane by beta-adrenergic receptor kinase 1: a role in receptor sequestration. J Biol Chem. 2001; 276: 18953–18959.
Burnier M. Angiotensin II type 1 receptor blockers. Circulation. 2001; 103: 904–912.
Sullivan R, Chateauneuf A, Coulombe N, et al. Coexpression of full-length gamma-aminobutyric acid(B) (GABA(B)) receptors with truncated receptors and metabotropic glutamate receptor 4 supports the GABA(B) heterodimer as the functional receptor. J Pharmacol Exp Ther. 2002; 293: 460–467.
Jordan BA, Trapaidze N, Gomes I, et al. Oligomerization of opioid receptors with beta 2-adrenergic receptors: a role in trafficking and mitogen-activated protein kinase activation. Proc Natl Acad Sci U S A. 2001; 98: 343–348.
Oakley RH, Laporte SA, Holt JA, et al. Differential affinities of visual arrestin, beta arrestin1, and beta arrestin2 for G protein-coupled receptors delineate two major classes of receptors. J Biol Chem. 2000; 275: 17201–17210.
Hamm HE. How activated receptors couple to G proteins. Proc Natl Acad Sci U S A. 2001; 98: 4819–4821.