Augmentation of Cardiac Contractility Mediated by the Human β3-Adrenergic Receptor Overexpressed in the Hearts of Transgenic Mice
Background Stimulation of β1- and β2-adrenergic receptors (ARs) in the heart results in positive inotropy. In contrast, it has been reported that the β3AR is also expressed in the human heart and that its stimulation leads to negative inotropic effects.
Methods and Results To better understand the role of β3ARs in cardiac function, we generated transgenic mice with cardiac-specific overexpression of 330 fmol/mg protein of the human β3AR (TGβ3 mice). Hemodynamic characterization was performed by cardiac catheterization in closed-chest anesthetized mice, by pressure-volume-loop analysis, and by echocardiography in conscious mice. After propranolol blockade of endogenous β1- and β2ARs, isoproterenol resulted in an increase in contractility in the TGβ3 mice (30%), with no effect in wild-type mice. Similarly, stimulation with the selective human β3AR agonist L-755,507 significantly increased contractility in the TGβ3 mice (160%), with no effect in wild-type mice, as determined by hemodynamic measurements and by end-systolic pressure-volume relations. The underlying mechanism of the positive inotropy incurred with L-755,507 in the TGβ3 mice was investigated in terms of β3AR–G-protein coupling and adenylyl cyclase activation. Stimulation of cardiac membranes from TGβ3 mice with L-755,507 resulted in a pertussis toxin–insensitive 1.33-fold increase in [35S]GTPγS loading and a 1.6-fold increase in adenylyl cyclase activity.
Conclusions Cardiac overexpression of human β3ARs results in positive inotropy only on stimulation with a β3AR agonist. Overexpressed β3ARs couple to Gs and activate adenylyl cyclase on agonist stimulation.
Received June 25, 2001; Prevision received August 29, 2001; accepted August 30, 2001.
β-Adrenergic receptors (βARs) are members of a family of G protein–coupled receptors that are stimulated by naturally occurring catecholamines. In the heart, both the β1- and β2AR subtypes are known to modulate cardiac function by producing positive inotropic and chronotropic effects. A third βAR subtype, the β3AR,1,2 has been found primarily in adipose tissue of rodents, in human omental tissue, and in the brown adipose tissue of newborns.2 In human heart3,4 and mouse heart,5 β3AR transcripts have been detected by sensitive methods, such as RNase protection assays and reverse transcription–polymerase chain reaction assays. Like β1- and β2ARs, the β3AR couples to Gs to activate adenylyl cyclase, which in adipose tissue leads to lipolysis or thermogenesis.2 It has also been shown that the β3AR can couple to Gi, resulting in the attenuation of adenylyl cyclase stimulation and in the activation of the mitogen-activated protein kinase (MAPK) pathway.6,7 Gauthier et al8 proposed that the β3AR is present and functional in the human heart. They showed that stimulation of human ventricular endomyocardial biopsies with BRL 37344, a β3AR agonist, leads to a pertussis toxin (PTX)–sensitive negative inotropic effect, suggesting that in this system the β3AR is coupled to Gi.8
To further explore the physiological consequences of activation of the β3AR in cardiac contractility, we generated transgenic mice with cardiac-specific overexpression of the human β3AR (TGβ3 mice). β3AR signal transduction was assessed both in vitro in cardiac membranes and in vivo by catheterization in intact mice.
The αMHC-HAβ3AR transgene was constructed from a 5.5-kb SalI-SalI fragment containing the murine αMHC promoter9 and the EcoRI-XbaI fragment containing the human β3AR coding sequence (1 to 402 amino acids) with an NH2-terminal hemagglutinin (HA) tag.10 The αMHC-HAβ3AR transgene was digested with SpeI-SacI, purified with CsCl, and used for nuclear injection of oocytes by the Duke Comprehensive Cancer Center Transgenic Mouse Facility. One line of C57B6SJL/J mice that expressed the HAβ3AR transgene was established. Studies were performed on mice 2 to 8 months of age.
mRNAs from heart tissue of both TGβ3 and wild-type (WT) mice were separated by electrophoresis and transferred onto a nylon filter (Schleicher & Schuell) by standard techniques.11 The filter was hybridized to a random primer radiolabeled probe corresponding to the entire coding region of the human β3AR clone.
Ligand Binding, GTPγS Loading, and Adenylyl Cyclase Assays
Crude membranes were prepared from excised hearts, and ligand binding assays were performed as previously described.9 [35S]GTPγS loading and adenylyl cyclase assays were performed as previously described.12,13
2D guided M-mode echocardiography was performed in conscious mice with an HDI 5000 echocardiograph (ATL) as previously described.14
Adult myocytes were isolated from WT and TGβ3 mice as previously described.15 After isolation, myocytes were fixed in 3% paraformaldehyde, and length and width were measured with a video edge-detection system (Crescent Electronics).
Hemodynamic Evaluation in Intact Anesthetized Mice
Cardiac catheterization was performed as described previously.14 Mice were anesthetized with a mixture of ketamine (100 mg/kg IP) and xylazine (2.5 mg/kg IP), and after bilateral vagotomy, a 1.4F high-fidelity micromanometer catheter (Millar Instruments) was inserted into the right carotid artery and advanced retrogradely across the aortic valve.
Hemodynamic measurements were recorded at baseline and 45 to 60 seconds after the injection of isoproterenol (1000 pg IV). After hemodynamics returned to baseline, propranolol (0.05 μg/g body weight [BW] IV) was administered to block β1- and β2ARs. After return to baseline, hemodynamic measurements were again recorded before and after administration of isoproterenol.
Hemodynamic measurements were recorded at baseline and 90 to 120 seconds after the injection of an incremental dose of L755,507 (0.25 to 4.0 μg IV).
In separate experiments, in vivo pressure-volume (P-V) relations were determined as previously described.14 Mice were anesthetized as described above and maintained by the administration of 0.5% to 1.0% isoflurane. The space and time resolutions of the sonomicrometry system are 0.015 mm and 0.001 seconds, respectively.
Data and Statistical Analyses
The digitized data were analyzed with a computer algorithm as previously described.14 Data are expressed as mean±SEM. Unpaired Student’s t tests and repeated-measures ANOVA were performed for statistical comparisons of the WT and TGβ3 mice after agonist stimulation. Post hoc analysis was performed with a Scheffé test. For all tests, a value of P<0.05 was considered significant.
Generation of β3AR Overexpressing Mice
To investigate the biochemical and physiological consequences of overexpression of β3ARs in heart, we generated transgenic mice with cardiac-restricted overexpression of the human β3AR (TGβ3). Transgene expression was documented by Northern analysis of mRNA from the heart (Figure 1A). As shown, expression was detected only in TGβ3 mice and was absent in WT mice. Several transcripts were detected, corresponding to 1.4 kb and 3.0 to 4.0 kb in size. The range in sizes is probably the result of the utilization of a variety of transcriptional termination signals downstream of the integration site.
Characterization of the β3AR Expression in TGβ3 Mice
The level of β3AR expression in cardiac membranes of the TGβ3 mice was quantified by use of competition ligand-binding assays (Figure 1B) with the radioligand [125I]iodocyanopindolol ([125I]ICYP) and increasing concentrations of unlabeled pindolol. Binding data from membranes prepared from WT and TGβ3 heart extracts were fit by a biphasic curve with a very small high-affinity component corresponding to β1- and β2ARs and a low-affinity phase corresponding to displacement of pindolol from the β3ARs. From these data, the number of β3ARs expressed in hearts of TGβ3 mice (Bmax) was calculated with GraphPAD software and was determined to be 330±36 fmol/mg membrane protein.
To characterize the endogenous expression levels of β1- and β2ARs in the transgenic heart, saturation binding experiments were performed with 5 to 200 pmol/L [125I]ICYP. At these relatively low concentrations and because of the low affinity of the β3ARs for the radioligand, only β1- and β2ARs will be bound, with minimal contribution from the β3ARs. Results in Figure 1C show that the Bmax for β1- and β2ARs is reduced from 53.6±4.1 fmol/mg in the WT mice to 34.5±2.0 fmol/mg in TGβ3 mice (P<0.002 between WT and TGβ3 mice). Competition binding experiments with the β2AR-selective antagonist ICI 118,551 were performed to assess the proportion of β1- and β2ARs expressed in the heart (Figure 1D). In cardiac membranes from WT mice, the data were fit by a biphasic curve with 33.8±3.4% high-affinity binding sites (β2AR) and 66.2±3.4% low-affinity sites (β1AR). In the TGβ3 mice, however, the proportion was 51.8±0.7% β2ARs and 48.5±0.8% β1ARs (P<0.02 between WT and TGβ3 mice), which suggests that the expression of the endogenous β1ARs was downregulated by ≈50%, from 35.5 fmol/mg in the WT mice to 16.7 fmol/mg in the TGβ3 mice. These data also demonstrate that there was no compensatory change in the β2AR expression in TGβ3 hearts.
Physiological and Basal Hemodynamic Parameters in WT and TGβ3 Mice
To determine the functional consequences of β3AR overexpression in the heart, cardiac catheterization was performed and hemodynamic measurements were recorded. As shown in Table 1, TGβ3 mice showed a significantly lower left ventricular (LV) systolic pressure and reduced LV dP/dtmax and LV dP/dtmin compared with WT mice. There was no difference in heart rate or LV end-diastolic pressure. Interestingly, in the TGβ3 mice, LV weight was lower, resulting in a lower LV/BW ratio than in WT mice. Morphometric analysis of the hearts, however, revealed no differences in myocyte size between the WT and TGβ3 hearts (2365±97 μm2, n=100, versus 2555±88 μm2, n=100, respectively, P<0.147), suggesting that overexpression of β3AR results in a decreased number of cells or a reduced amount of nonmyocyte tissue.
Echocardiography in Conscious WT and TGβ3 Mice
Because anesthesia can affect the contractile state of the ventricle, we sought to measure echocardiographic parameters in conscious mice. Chamber dimensions, wall thickness, % fractional shortening, and heart rate did not show any difference between WT and TGβ3 mice (Table 2).
Effect of Isoproterenol on Hemodynamics in WT and TGβ3 Mice
To determine whether the β1/β2/β3AR agonist isoproterenol could augment contractile function in the TGβ3 mice, hemodynamic measurements were made before and after isoproterenol administration. LV contractility, as assessed by LV dP/dtmax at baseline conditions, was lower in TGβ3 mice than in WT mice (6664±388 mm Hg/s, n=10, versus 9470±921 mm Hg/s, n=9, P<0.02, Figure 2A, Table 1), whereas the effect of isoproterenol on LV dP/dtmax was comparable between TGβ3 and WT mice (Figure 2A). To further characterize the increase in ΔLV dP/dtmax with isoproterenol, the TGβ3 mice were pretreated with the nonselective β1/β2AR antagonist propranolol. As shown, the positive inotropic effect of isoproterenol was completely abolished by pretreatment with propranolol in WT mice; a small but significant increase in contractility was still observed, however, in the TGβ3 mice (Figure 2B), indicating that a small fraction of the positive inotropic action of isoproterenol in TGβ3 mice may be attributed to the stimulation of overexpressed β3ARs.
Effect of the Selective β3AR Agonist L755,507 on Hemodynamics in WT and TGβ3 Mice
To test directly whether stimulation of β3ARs could augment contractility, hemodynamic parameters in WT and TGβ3 mice were measured in response to the selective human β3AR agonist L-755,507. This compound is >1000-fold more selective for the activation of the β3AR than for the β1AR and has no measurable β2AR agonist activity.16 As shown in Figure 3A and 3B, L-755,507 (0.25 to 4.0 μg IV) led to a marked increase in cardiac contractility in TGβ3 mice (n=10) that was completely absent in WT mice (n=5). Similarly, a large dose-dependent increase in heart rate in response to L-755,507 was also observed in TGβ3 mice and was again absent in WT mice (Figure 3C). There was no significant difference in the response of LV pressure to L-755,507 between WT and TGβ3 mice (Figure 3D). These results indicate that L-755,507 acts selectively on the human β3AR and exerts positive inotropic and chronotropic actions in TGβ3 mice.
P-V Loops in WT and TGβ3 Mice
To rigorously investigate whether basal function was different in TGβ3 mice and WT mice, as suggested by Table 1 and Figures 2A and 3⇑A, we obtained end-systolic P-V relations for both groups (Figure 4B and 4C). Under basal conditions, the end-systolic P-V relation was curvilinear and the slope, Emax′, of TGβ3 mice at baseline was comparable to that of WT mice (Table 3). The administration of L-755,507 resulted in a steeper and more curvilinear end-systolic P-V relation only in the TGβ3 mice (Figure 4B and 4C, Table 3). Furthermore, no significant difference in the volume intercept of the end-systolic P-V relation, LV end-systolic volume, or LV end-diastolic volume was observed between WT and TGβ3 mice. These data show that overexpression of the β3AR does not affect the basal contractile state of the ventricle but can result in a significant enhancement of contractility with administration of L-755,507. Interestingly, whereas baseline LV dP/dtmax suggested depressed contractility in TGβ3 mice, a more rigorous analysis using P-V data showed basal contractility similar to that of WT mice. This is in agreement with known limitations of using LV dP/dtmax as an index of contractile function.14
Characterization of β3AR Signaling in TGβ3 Hearts
To investigate the mechanism by which the selective stimulation of the β3AR in the TGβ3 mice enhances contractility, biochemical analysis of the physical coupling of the receptor to its cognate G protein and measurement of adenylyl cyclase activity were performed. The agonist-mediated activation of the G protein α-subunit was measured by the quantitative binding of the radiolabeled nonhydrolyzable GTP analog [35S]GTPγS. Figure 5A shows that stimulation with L-755,507 results in GTPγS loading only in TGβ3 membranes (1.33±0.04-fold over basal) and not in WT. To determine whether L-755,507–mediated GTPγS loading in TGβ3 membranes is due to incorporation into Gs or Gi, TGβ3 mice were treated with PTX (0.1 μg/g BW) overnight. GTPγS loading stimulated by the Gi-coupled A1 adenosine receptor agonist N6-cyclopentyladenosine (CPA) was completely abrogated in the PTX-treated sample, confirming Gi blockade by PTX (Figure 5A). Stimulation of the PTX-treated TGβ3 membranes with L-755,507 resulted in GTPγS loading that was not significantly different from that observed in the untreated TGβ3 membranes. These data show that the overexpressed human β3ARs in the TGβ3 mice are coupled mostly to non-Gi proteins.
To assess the functional coupling of the overexpressed human β3AR to Gs in the TGβ3 mice, L-755,507–stimulated adenylyl cyclase activity was measured in cardiac membranes prepared from WT and TGβ3 hearts. The L-755,507–stimulated adenylyl cyclase activity in TGβ3 membranes was 1.6-fold over basal, whereas there was no stimulation in WT membranes (Figure 5B). In addition, there was no measurable increase in basal cyclase activity in the TGβ3 membranes compared with WT controls (Figure 5B, legend). This observation suggests that despite the marked overexpression of β3ARs, they are functionally inactive until specifically stimulated with a β3 agonist.
Numerous studies have shown that stimulation of β1- and β2ARs in intact animals or cardiac preparations can lead to positive chronotropic and inotropic effects. Conversely, from initial work,5,8,17 it appears that stimulation of β3ARs can lead to a negative inotropic effect. To study the potential for the β3AR to affect cardiac function, a transgenic mouse model was constructed that overexpresses the human β3AR. The most salient feature of this model is enhanced cardiac contractility after intravenous injection of a selective β3AR agonist, L-755,507. The increase in LV dP/dtmax is attributable to the overexpression of β3AR, because WT animals showed no enhancement of contractility with administration of the β3AR agonist. Furthermore, in TGβ3 mice, administration of isoproterenol could overcome the blockade produced by pretreatment with the selective β1- and β2AR antagonist propranolol. Unfortunately, the unavailability of a specific β3AR antagonist without partial agonist activity hampers the ability to further dissect the inotropic effect of isoproterenol in the TGβ3 mice. Another characteristic of the TGβ3 mouse is the downregulation of its endogenous β1ARs by 50%. This result was not totally unexpected, because in the β3AR knockout mouse,18 the mRNA levels for β1AR are upregulated in white and brown adipose tissues, whereas those for β2AR are unchanged, suggesting a compensatory regulation of β1- and β3AR gene expression.
Gauthier et al8 initially described functional coupling of the β3AR to Gi when stimulated with the β3AR agonist BRL37344, resulting in negative inotropic effects in the human heart. Subsequently, negative inotropic effects of BRL37344 have been demonstrated in the isolated guinea pig heart.17 A similar conclusion was inferred from studies in β3AR-knockout mice, in which isoproterenol produced an augmented contractile response in comparison to WT mice.5 In our study, however, β3AR-Gs coupling is clearly evident in the TGβ3 mouse model from the PTX-insensitive GTPγS loading, the activation of adenylyl cyclase in cardiac membranes, and the in vivo positive inotropic effect caused by selective β3AR stimulation. The reason for the apparent disparity between the results presented in this work and those of Gauthier et al8 is unknown. They may be attributable, however, to the use of different β3AR agonists or to different experimental models, ie, overexpression of human β3AR in a mouse versus endogenous β3AR in human endomyocardial biopsies or in guinea pig heart. Nonetheless, the β3AR-Gs coupling described in the TGβ3 mouse is consistent with reports of β3AR dually coupling to Gs and Gi in a variety of cell types.6,7 Stimulation of β3ARs in these models led to increased adenylyl cyclase activity, despite potential inhibitory effects from its coupling to Gi.2
In addition to the studies describing β3AR activation in heart, there is evidence of a vasodilatory effect after β3AR agonist treatment.19 In conscious dogs, β3AR stimulation with selective agonists induced marked peripheral vasodilation and positive inotropic and chronotropic effects.20,21 It is notable that in WT mice, the selective rodent β3AR agonist CL316243 leads to a hypotensive response, and in β1-/β2AR knockout mice, this effect is augmented.22 These effects do not appear to be directly related to β3AR stimulation of cardiac myocytes, however, because CL316243 has no chronotropic or inotropic effects in atrial or ventricular preparations from these knockout animals.22 The variability of the cardiac and vascular responses to β3AR agonists in different species highlights the fact that the function of the β3AR in these tissues is still poorly understood.
It has been shown that during chronic heart failure, the cardiac β1-adrenergic receptors are downregulated, leading to a deficiency in contractility.23 We show that in this transgenic mouse model with cardiac-restricted overexpression, the human β3AR is quiescent until stimulated with a selective agonist, at which point there is a marked augmentation in LV contractility. In addition, because the β3AR is relatively insensitive to catecholamines, it would be minimally activated by endogenous catecholamines. Taken together, this approach could have important therapeutic potential in patients with heart failure, in which delivery of the human β3AR by gene therapy approaches to the heart could provide a functionally inactive signaling protein that becomes activated only when a highly selective agonist is exogenously administered.
This study was supported in part by National Institutes of Health grants HL-16037 (Dr Lefkowitz) and HL-61558 (Dr Rockman). Dr Rockman is a recipient of a Burroughs Wellcome Fund Clinical Scientist Award in Translational Research. Dr Lefkowitz is an Investigator of the Howard Hughes Medical Institute. We thank Dr Sheila Collins for her insightful comments, advice, and a gift of the L-755,507 compound (Merck Research Laboratories).
The first 2 authors contributed equally to this work.
Emorine LJ, Marullo S, Briend-Sutren MM, et al. Molecular characterization of the human β3-adrenergic receptor. Science. 1989; 245: 1118–1121.
Krief S, Lonnqvist F, Raimbault S, et al. Tissue distribution of β3-adrenergic receptor mRNA in man. J Clin Invest. 1993; 91: 344–349.
Gerhardt CC, Gros J, Strosberg AD, et al. Stimulation of the extracellular signal-regulated kinase 1/2 pathway by human β3 adrenergic receptor: new pharmacological profile and mechanism of activation. Mol Pharmacol. 1999; 55: 255–262.
Soeder KJ, Snedden SK, Cao WH, et al. The β3-adrenergic receptor activates mitogen-activated protein kinase in adipocytes through a Gi-dependent mechanism. J Biol Chem. 1999; 274: 12017–12022.
Milano CA, Allen LF, Rockman HA, et al. Enhanced myocardial function in transgenic mice overexpressing the β2-adrenergic receptor. Science. 1994; 264: 582–586.
Liggett SB, Freedman NJ, Schwinn DA, et al. Structural basis for receptor subtype-specific regulation revealed by a chimeric β3/β2-adrenergic receptor. Proc Natl Acad Sci U S A. 1993; 90: 3665–3669.
Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; 1989.
Esposito G, Santana LF, Dilly K, et al. Cellular and functional defects in a model of heart failure. Am J Physiol. 2000; 279: H3101–H3112.
Rockman HA, Choi DJ, Akhter SA, et al. Control of myocardial contractile function by the level of β-adrenergic receptor kinase 1 in gene-targeted mice. J Biol Chem. 1998; 273: 18180–18184.
Susulic VS, Frederich RC, Lawitts J, et al. Targeted disruption of the β3-adrenergic receptor gene. J Biol Chem. 1995; 270: 29483–29492.
Shen YT, Cervoni P, Claus T, et al. Differences in β3-adrenergic receptor cardiovascular regulation in conscious primates, rats and dogs. J Pharmacol Exp Ther. 1996; 278: 1435–1443.
Rohrer DK, Chruscinski A, Schauble EH, et al. Cardiovascular and metabolic alterations in mice lacking both β1- and β2-adrenergic receptors. J Biol Chem. 1999; 274: 16701–16708.
Bristow MR. Why does the myocardium fail? Insights from basic science. Lancet. 1998; 352 (suppl 1): SI8–SI14.