Bitransgenesis with β₂‐Adrenergic Receptors or Adenylyl Cyclase Fails to Improve β₁‐Adrenergic Receptor Cardiomyopathy

Abstract

Cardiomyopathic effects of β‐adrenergic receptor (βAR) signaling are primarily due to the β₁AR subtype. β₁/β₂AR and β₁/adenylyl cyclase type 5 (AC5) bitransgenic mice were created to test the hypothesis that β₂AR or AC5 co‐overexpression has beneficial effects in β₁AR‐mediated cardiomyopathy. In young mice, β₁/β₂ hearts had a greater increase in basal and isoproterenol‐stimulated contractility compared to β₁/AC5 and β₁AR hearts. By 6 months, β₁AR and β₁/β₂ hearts retained elevated basal contractility but were unresponsive to agonist. In contrast, β₁/AC5 hearts maintained a small degree of agonist responsiveness, which may be due to a lack of β₁AR downregulation that was noted in β₁‐ and β₁/β₂ hearts. However, by 9 ‐months, β₁, β₁/β₂, and β₁/AC5 mice had all developed severely depressed fractional shortening in vivo and little response to agonist. p38 mitogen activated protein kinase (MAPK) was minimally activated by β₁AR, but was markedly enhanced in the bitransgenics. Akt activation was only found with the bitransgenics. The small increase in cystosolic second mitochondria‐derived activator of caspase (Smac), indicative of apoptosis in 9‐month β₁AR hearts, was suppressed in β₁/AC5, but not in β₁/β₂, hearts. Taken together, the unique signaling effects of enhanced β₂AR and AC5, which have the potential to afford benefit in heart failure, failed to salvage ventricular function in β₁AR‐mediated cardiomyopathy.

Introduction

Heart failure from virtually every etiology is accompanied by enhanced sympathetic activity, an adaptation in response to decreased cardiac output. While this response is effective in increasing contractility during acute decompensation, prolonged activation is deleterious, leading to worsening failure. ¹ , ² Both β₁‐adrenergic receptors (β₁AR) and β₂AR are expressed on cardiomyocytes and participate in catecholamine‐mediated enhancement of cardiac inotropy or chronotropy. The deleterious effects of catecholamine signaling at the cardiomyocytes have generally been attributed to their activation of the β₁AR. Indeed, we and others have shown that moderate overexpression of β₁AR in cardiomyocytes of transgenic mice results in a time‐dependent heart failure, ³ , ⁴ , ⁵ while β₂AR expression at similar levels is well tolerated. ⁶ This diference in the propensity to evoke failure is not readily reconciled with the enhanced contractility observed in young transgenic overexpressing mice, as the degree of increased contractility is similar in β₁‐ and β₂AR‐overexpressing mice. ³ , ⁶ Nor is it altogether apparent that the pathogenic effects of β₁AR activation are entirely due to cAMP/protein kinase A (PKA) activation; cardiac adenylyl cyclase type 5 (AC5)‐overexpressing mice do not develop failure, yet have levels of (elevated) AC activities similar to those of young β₁AR‐overexpressing transgenic mice. ³ , ⁷ It has been postulated that intrinsic differences between β₁AR and β₂AR signaling accounts for the more pathogenic nature of β₁AR. ¹ And furthermore, certain properties of the β₂AR subtype may be “protective” in heart failure. ¹ These properties include coupling to the inhibitory G‐protein G_αi, signaling to antiapoptotic pathways, and receptor/cAMP microdomain localization. In addition to such potential distinct signaling events evoked by the two subtypes in cardiomyocytes, the heart failure milieu also includes stimuli (elevated catecholamines) for desensitization and downregulation of βAR. And indeed, the β₁‐ and β₂AR vary in a number of ways in regard to agonist‐promoted desensitization and trafficking. ¹ Taken together, these differences have suggested that β₂AR activation might mitigate against β₁AR‐mediated heart failure, and that stabilizing, activating, or mimicking the signaling of this subtype might have therapeutic potential. ⁸ Similarly, AC5/6 levels are reduced in β₁AR‐mediated cardiomyopathy, ⁹ and methods to replace, or overexpress, AC5 or AC6 have been considered as therapeutic interventions. ¹⁰ While AC6 overexpression has “rescued” certain forms of left ventricular dysfunction from genetic manipulation, ¹¹ such an approach has not been taken with a model of transgenic overexpression of β₁AR, which leads to a time‐dependent heart failure. To investigate these two potential avenues for altering β₁AR‐mediated cardiomyopathy, we utilized overexpressing transgenic mice that we have previously developed to create β₁/β₂AR and β₁/AC5 bitransgenic mice, which were compared to β₁AR‐overexpressing mice over a 9‐month time period for physiologic or biochemical modification of the β₁AR phenotype.

Materials and Methods

Transgenic mice

Transgenic mice overexpressing the human β₁AR (the most common variant, Arg389), β₂AR, and AC5 were generated using the α‐myosin heavy chain (α‐MHC) promoter to target expression to cardiomyocytes, and have each been previously described. ⁷ , ⁹ , ¹² All mice were of the FVB/N background. Heterozygous β₁AR transgenics were mated with heterozygous β₂AR or AC5 transgenics to create the bitransgenic mice, which are denoted as β₁/β₂AR and β₁/AC5 mice. Genomic DNA from tail‐cuts was screened for the presence of transgenes by targeted PCRs, which included one primer in the α‐MHC promoter and one in the cDNA of transgene. Mice were fed a normal diet and maintained under identical conditions, and either sex was studied.

Physiologic studies

The studies were approved by the University of Maryland School of Medicine Institutional Animal Care and Use Committee. A hemodynamic evaluation was performed, as described previously, using the work‐performing mouse heart preparation. ³ Mice were anesthetized via intraperitoneal injection with 100 mg/kg sodium nembutal and 1.5 units of heparin to prevent microthrombi. The hearts were removed and attached by the aorta to a 20‐gauge cannula and temporarily retrogradely perfused with oxygenated Krebs–Henseleit solution (in mM: 118 NaCl, 4.7 KCl, 2.5 CaCl₂, 0.5 Na‐EDTA, 25 NaHCO₃, 1.2 KH₂PO₄, and 11 glucose) saturated with 95% O₂, 5% CO₂. A polyethelene‐50 catheter was inserted into the apex of the left ventricle for the measurement of intraventricular pressure. The pulmonary vein was connected to a second cannula, and antegrade perfusion with oxygenated Krebs–Henseleit solution was initiated with a basal workload of 300 mmHg × mL/min (6 mL venous return and 50 mmHg mean aortic pressure). Hearts were allowed to equilibrate for 20 minutes. Atrial pressure was monitored through a sidearm in the left atrial cannula. The left ventricular pressure signals were digitized at 1 kHz and analyzed ofine by the computer software Biobench (National Instruments, Austin, TX, USA). The first positive and negative derivatives of the left intraventricular pressure curve (+dP/dt and –dP/dt) and duration of contraction and relaxation (time to peak pressure: TPP) and time to half relaxation (TR_1/2) were calculated. Afer establishment of baselines, infusions of the nonselective βAR agonist isoproterenol were undertaken using doses from 0.1 nM to 0.1 μM. The maximal response during the 5‐minute infusion was utilized to construct the dose–response curves.

Echocardiography

Mice were sedated with isofurane delivered by a nasal cone and secured in the supine position to a warming pad maintained at 37°C. Transthoracic echocardiography was performed with a Vev0770 echocardiograph with the 707B probe (Visualsonics, Toronto, CA, USA), as previously described. ³ The heart was imaged in the two‐dimensional mode (M‐mode) in the parasternal long‐axis views. The measurements of intraventricular septal (IVS) thickness, left ventricular posterior wall (LVPW) thickness, and left ventricular internal diameter were made from the left ventricle in systole and diastole. The diastolic measurements were made at the time of maximal left ventricular end‐diastolic dimension (LVEDD). Left ventricular end‐systolic dimensions (LVESD) were performed at the time of the most anterior systolic excursion of the LVPW. Left ventricular percent fractional shortening, chamber volume, and mass were calculated using methods as previously described. ¹³ In some mice, afer these baseline measurements were obtained, echocardiography was repeated 3 minutes after a 2‐μg/g body weight intraperitoneal injection of isoproterenol was administered.

Radioligand binding and Western blots

For radioligand binding, the hearts were homogenized in 5 mM Tris, 2 mM EDTA, pH 7.4 buffer at 4°C with a Polytron for 15 seconds, diluted, and centrifuged at 400 ×g for 10 minutes. The supernatant was recovered and centrifuged at 30,000 ×g for 15 minutes, and the pellet was resuspended in 75 mM Tris/12 mM MgCl₂/2 mM EDTA pH 7.4 at 25°C. Quantitative radioligand binding was performed, as previously described, ¹⁴ using the βAR radioligand¹²⁵ I‐cyanopindolol (¹²⁵I‐CYP) with 1.0‐μM propranolol used to define nonspecific binding. In the β₁/β₂AR mouse hearts, differentiation of the densities of the two subtypes was determined using competition with the β₁AR‐specific antagonist CGP20712 and β₂AR‐specific antagonist ICI118551, as previously described. ¹⁴ The results are provided as fmol/mg protein and are from six hearts from each group. Protein was determined by the copper bicinchoninic method. ¹⁵ Western blots were carried out as previously described. ³ Briefly, homogenized hearts were solubilized in 10 mM Tris and 1 mM EDTA pH 7.6 with 1% SDS. Protease inhibitor cocktail (Roche, Nutley, NJ, USA), and phosphatase inhibitor cocktails 1 and 11 (Calbiochem, San Diego, CA, USA) were included in all steps. The samples were clarified by centrifugation at 10,000 ×g, the proteins fractionated on 10% SDS‐polyacrylamide gels and transferred to nitrocellulose membranes. Immunoblotting was carried out with antibodies (from Cell Signaling, Danvers, MA, USA) using the following titers: ERK1/2 MAPK (1:2,000), phospho‐ERK1/2 MAPK(1:1,000), p38 MAPK(1:2,000), phospho‐p38 MAPK (1:1,000), Akt (1:1,500), and phospho‐Akt (1:1,000). Detection was by enhanced chemiluminescence (PerkinElmer, Waltham, MA, USA), and the signals were acquired directly from the membranes using a Fuji LAS‐3000 charged coupled device camera and quantitated with the included software. For each blot, the ratios of phosphorylated signal to the total signal was calculated and then normalized to the mean nontransgenics (NTG) ratio.

Statistical analysis

Unpaired t‐tests were used to compare data from the indicated groups, typically comparing results from β₁AR transgenic hearts with β₁/β₂AR and β₁/AC5 transgenics, or between time periods (2, 6, or 9 months). When dose–response studies were performed, the minimal response (R _min) and maximal response (R _max) were determined by fitting the data to a sigmoid curve using Prism (GraphPad, San Diego, CA, USA). Paired t‐tests were utilized for radioligand binding and Western blot data as indicated. p values < 0.05 were considered significant.

Results

Physiologic function at 2 months of age

βAR expression in 2‐month‐old mice as determined by ¹²⁵I‐CYP radioligand binding was: NTG mice 31 ± 3.5 fmol/mg (∼70%β₁AR), β₁AR mice 3,446 ± 352 fmol/mg (essentially all β₁AR), and β₁/AC5 mice 1,666 ± 285 fmol/mg (essentially all β₁AR; p < 0.01 vs. β₁AR mice). For the β₁/β₂AR mice, β₁AR expression was 2,624 ± 256 fmol/mg (p < 0.05 vs. β₁AR mice) and β₂AR expression was 1,350 ± 60 fmol/mg. None of the transgenic mice had heart/body weight ratios that differed from NTG mice (Table 1). The hearts from age‐and sex‐matched mice were studied using the work‐performing model at baseline (the absence of agonist) and in response to the nonselective βAR agonist isoproterenol. In the hearts from these young mice, overexpression of β₁AR resulted in an increased baseline +dP/dt and –dP/dt ( Table 2 ). Co‐overexpression in the β₁/β₂AR bitransgenic mice did not significantly enhance +dP/dt over β₁AR overexpression alone, nor did co‐overexpression of AC5, as observed in the β₁/AC5 bitransgenic mice, enhance +dP/dt over β₁AR overexpression. However, a modest increase in baseline –dP/dt was observed in the β₁/β₂AR bitransgenic hearts compared to β₁AR hearts. This increase in –dP/dt was only found when β₁AR was co‐expressed with β₂AR, and not with the β₁/AC5 bitransgenics ( Table 1 ).

Table 2.

Cardiac contractile parameters at baseline stratifed by age and transgene.

Parameters	NTG 2m n= 7	β1 2m n= 5	β1/β2 2m n= 5	β1/AC5 2m n= 5	NTG 6m n= 5	β1 6m n= 6	β1/β2 6m n= 5	β1/AC5 6m n= 5
+dP/dt, mmHg/s	4104 ± 118*	5588 ± 217	5928 ± 408	5988 ± 356	3863 ± 85*	5718 ± 594	3021 ± 311*	4420 ± 205
‐dP/dt, mmHg/s	3128 ± 169*	4639 ± 360	5608 ± 344*	4567 ± 321	2852 ± 272*	5085 ± 603	3393 ±*280	4264 ± 225*
TPP, ms/mmHg	0.37 ± 0.04*	0.29 ± 0.03	0.22 ± 0.01*	0.24 ± 0.01	0.41 ± 0.04*	0.26 ± 0.02	0.34*± 0.03	0.28 ± 0.02*
TR1/2, ms/mmHg	0.60 ± 0.03*	0.47 ± 0.02	0.38 ± 0.02*	0.37 ± 0.03*	0.57 ± 0.03*	0.46 ± 0.03	0.49 ± 0.04	0.44 ± 0.03

*Different from β₁A transgenic hearts at the same age, p < 0.01

Table 1.

Heart‐to‐body weight ratios stratified by age and transgene.

	2‐month n= 5	6‐month n= 5	9‐month n= 5
NTG	3.5 ± 0.04	3.6 ± 0.1	3.8 ± 0.2*
β1	3.5 ± 0.1	3.9 ± 0.2	5.3 ± 0.2
β1/β2	3.6 ± 0.1	5.1 ± 0.2*	6.7 ± 0.2*
β1/AC5	3.6 ± 0.2	4.6 ± 0.1*	5.2 ± 0.2
β2	3.5 ± 0.2	3.9 ± 0.1	4.2 ± 0.1*
AC5	3.6 ± 0.1	3.7 ± 0.2	3.9 ± 0.1*

*Ratio different from β₁AR transgenic hearts of the same age, p < 0.01.

The contractile response to isoproterenol in these young mice is shown in Figure 1 . While baseline +dP/dt was equivalent in β₁AR, β₁/β₂AR, and β₁/AC5 hearts at 2 months of age (as discussed above), the response to agonist was significantly greater for β₁/β₂AR hearts compared to the other two transgenics, which were not different between each other ( Figure 1A ). Thus, the co‐expression of β₂AR further enhanced agonist‐promoted contractility in β₁AR‐overexpressing hearts, but co‐overexpression of AC5 with β₁AR had no effect over β₁AR alone. For relaxation in the 2‐month‐old hearts, essentially the same pattern was found, except that, as reported above, the baseline –dP/dt was somewhat increased for β₁AR and β₁/AC5 hearts compared to β₁/β₂AR hearts ( Figure 1B ).

Figure 1.

Agonist‐promoted contractile responses in hearts from IβAR,β/IβAR, andβ,/ACS transgenic mice at 2 months of age. Results are from 5–6 experiments performed with each of the indicated lines using the ex vivo work‐performing model. (A) +dP/dt, (B) −dP/dt. *R _min′†R _max′ different from β₁AR transgenic hearts at p < 0.01.

Physiologic function at 6 months of age

In these older mice, the β₁AR transgenic mice had heart/body weight ratios (T a b l e 1) that trended toward being greater than those of NTG mice (3.9 ± 0.2 mg/g vs. 3.6 ± 0.1 mg/g, p= 0.20). However, both β₁/β₂ and β₁/AC5 hearts had increased heart‐to‐body weight ratios (5.1 ± 0.2 and 4.6 ± 0.1, respectively) compared to NTG hearts (p < 0.005) as well as β₁AR transgenics (p < 0.005). In these 6‐month‐old mice, the baseline +dP/dt remained increased over NTG in the β₁AR hearts; however, in β₁/β₂AR hearts, this effect was not observed, and indeed they did not differ from NTG (T a b l e 2). β₁/AC5 mice at 6 months of age also displayed a decrease in +dP/dt compared to 2‐month‐old hearts (4,420 ± 205 mmHg/s vs. 5,986 ± 356 mmHg/s, p < 0.01), but this parameter was still slightly greater than NTG (3803 ± 85 mmHg/s, p < 0.05). This pattern was mimicked in regard to baseline –dP/dt.

While at 6 months of age β₁AR‐overexpressing hearts maintained a somewhat increased baseline +dP/dt, as we have previously noted, ³ they were unresponsive to isoproterenol ( Figure 2A ). Similarly, β₁/β₂AR mice were unresponsive to agonist. In contrast, β₁/AC5 mice had a positive inotropic response that was similar to that of NTG, but the maximal response was depressed compared to that of the hearts from 2‐month‐old β₁/AC5 bitransgenic mice (6,240 ± 526 mmHg/s vs. 9,065 ± 240 mmHg/s, p < 0.005). For relaxation, neither β₁AR nor β₁/β₂AR hearts responded to agonists. The maximal β₁/AC5 relaxation response was greater than NTG, but essentially parallel in nature, with the difference in maximal increase attributable to the increased baseline –dP/dt. Nevertheless, the maximal isoproterenol‐promoted –dP/dt in 6‐month‐old β₁/AC5 bitransgenics was not different from that in 2‐month‐old mice of the same genotype, nor was the change in –dP/dt from the baseline afected by age in these mice ( Figure 1B and 2B ).

Figure 2.

Agonist‐promoted contractile responses in hearts from |i,AR, p/p₂AR, andβ/ACS transgenic mice at 6 months of age. Results are from 5–6 experiments performed with each of the indicated lines using the ex vivo work‐performing model. (A) +dP/dt (B) ‐dP/dt. *R _min′†R _max′ different from β₁AR transgenic hearts at p < 0.01

Physiologic function at 9 months of age

At 9 months of age, an increase in the heart‐to‐body weight ratio was apparent for the β₁AR‐overexpressing hearts compared to NTG. And, the β₁/β₂ and β₁/AC5 hearts continued to have increased ratios, as was observed at 6 months. At this age, a number of the transgenic mouse hearts were unstable once they were removed and thus could not be studied by the ex vivo method. So, noninvasive M‐mode echocardiography at rest and in response to a single subcutaneous dose of isoproterenol was carried out in the 9‐month‐old mice ( Table 3 and Figure 3 ). Left ventricular chamber dilatation was observed for β₁AR, β₁/β₂AR, and β₁/AC5 mice compared to NTG, readily observed in the LVEDD and LVESD measurements. Substantial increases in calculated left ventricular systolic (2‐fold) and diastolic (2–5‐fold) volumes were noted. As previously reported, ³ β₁AR‐mediated cardiomyopathy results in markedly reduced fractional shortening compared to NTG at 9 months of age (T a b l e 2). Neither co‐expression of β₂AR or AC5 had any notable effect on this phenotype, and indeed β₁/β₂AR mice had the lowest baseline fractional shortening of all transgenics (11.2 ± 1.85%). Consistent with the 6‐month ex vivo contractile studies, β₁AR and β₁/β₂AR mice at 9 months had minimal increases in fractional shortening in response to isoproterenol ( Figure 3 ). However, while some contractile responsiveness was observed at 6 months with the β₁/AC5 mice, by 9 months of age, isoproterenol stimulation of fractional shortening in these mice was virtually absent (from 19 ± 1.4% at baseline to 26 ± 1.6% after isoproterenol).

Table 3.

Echocardiography results in 9‐month‐old mice.

Parameter	NTG*n= 5	β1 n= 9	β1/β2 n= 7	β1/AC5 n= 5
IVsd, mm	1.1 ± 0.07	0.78 ± 0.06	0.65 ± 0.04	0.83 ± 0.03
IVSs, mm	1.56 ± 0.11	1.05 ± 0.09	0.76 ± 0.15	1.11 ± 0.10
LVPdd, mm	1.3 ± 0.05	1.0 ± 0.03	0.79 ± 0.1†	0.97 ± 0.04
LVPds, mm	1.7 ± 0.09	1.37 ± 0.15	1.17 ± 0.11	1.42 ± 0.06
LVEsd, mm	2.37 ± 0.2	3.68 ± 0.09	4.91 ± 0.07	3.80 ± 0.15
Lvedd, mm	3.76 ± 0.2	4.71 ± 0.15	5.43 ± 0.02	4.63 ± 0.11
Lv% fractional shortening	35 ± 3.9	19 ± 1.6	11 ± 1.85	19 ± 1.4
LVVD, μL	67 ± 11.0	128 ± 11.5	156 ± 14.8	101 ± 10
LVVS, μL	20.1 ± 5.9	84 ± 9.7	125.1 ± 14.9	78.8 ± 6.9
LVM, mg	135 ± 9.3	215 ± 11	238 ± 31	177 ± 18

*Parameters different from those of β₁AR mice at p < 0.01.

†Different from β₁AR mice at p < 0.01. LVVS, left ventricular volume (systole); LVM, left ventricular

mass; LVPD, left ventricular — dimension.

Figure 3.

Baseline and agonist‐stimulated fractional shortening in hearts fromβ, AR, p/p₂AR andβ,/ACS transgenic mice at 9 months of age. Results are from 5–9 experiments performed with each line using echocardiography in the anesthetized mouse. *Basal, †isoproterenol stimulated, LVEFs different from β₁AR transgenics at p < 0.01.

Selected protein expression or activity by genotype and age

Although the primary goals of this study relate to physiologic function, we also examined expression or activity of several proteins previously identifed as playing important roles in adrenergic signaling and heart failure progression. β₁AR expression decreased over time in some mice, as summarized in Figure 4 . β₁AR expression in 6‐month‐old β₁AR transgenic hearts decreased compared to 2‐month‐old hearts (1,983 ± 203 fmol/mg vs. 3,446 ± 352 fmol/mg, p < 0.01). In contrast, there was no change in β₁AR expression in the β₁/AC5 bitransgenic mice over this time period (1,666 ± 248 fM/mg vs. 2,005 ± 720 fM/mg). While overall βAR expression did not change over time in the β₁/β₂ bitransgenic hearts (3,694 ± 290 fM/mg vs. 3,727 ± 235 fM/mg), the absolute levels of the two subtypes, and their ratios, clearly changed. By 6 months of age, β₁AR expression in these mice decreased to 989 ± 204 fM/mg (from 2,624 ± 256 fM/mg, p < 0.01), while β₂AR expression actually increased (2,746 ± 207 fM/mg at 6 months, from 1,350 ± 60 fM/mg at 2 months, p < 0.01). Thus, β₁AR expression decreased 2 to 6 months. At 2 months, ERK1/2 MAPKactivity was not elevated in any heart over NTG except for the β₂AR transgenic, which was utilized as a positive control ( Figure 5A ). At 6 months of age, these β₂AR‐overexpressing hearts showed no enhancement of ERK1/2 MAPK activity compared to NTG ( Figure 5D ), but the β₁/β₂ hearts revealed a small increase. In young mice, β₁AR overexpression resulted in a 2‐fold increase in p38 MAPK activation; co‐expression of β₂AR, and AC5, with β₁AR resulted in an even more marked increase in the activity of this kinase at 2 months, and this pattern was maintained in the 6‐month‐old mice ( Figure 5B and 5E ). Akt was not activated by β₁AR overexpression; however, both of the bitransgenics at 2 and 6 months of age revealed activation of Akt by 3–4‐fold ( Figure 5C and 5F ).

Figure 4.

Age‐dependent changes inβ,‐ and β₂AR expression in hearts fromβ,AII, pypjAR, and β,/ACS transgenic mice. Results are from fove experiments. *β₁AR expression decreased compared to 2 months of age, p < 0.01.

Figure 5.

Alterations in cardiac ERK1/2 MAPK, p38 MAPK, and Akt presented by age and transgene. Results are plotted as the ratio of phosphorylated to total kinase expression, normalized to the mean NTG values. Results are from fve experiments. *, p values of < 0.05 to < 0.01 versus β₁AR.

Finally, we measured an index of apoptosis to assess the potential for modification by co‐expression. β₁AR transgenic hearts overexpressing the receptor at the levels utilized here do not exhibit overt apoptosis (such as would be detected by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining) at any age, including at 9 months. ³ We thus utilized a highly sensitive immunoblot assay for cystosolic second mitochondria‐derived activator of caspase (Smac) (also termed direct IAp binding protein, DIABLO). In the intrinsic cell death pathway, activation of caspases is by release from the mitochondria of proapoptotic proteins such as Smac. ¹⁶ As shown in F i g u r e 6,β₁AR‐overexpressing hearts display a small increase in cystosolic Smac only at 9 months of age. This was also observed with β₂AR hearts, and the extent of increase in cystosolic Smac from 2 months to 9 months was the same in the bitransgenic β₁/β₂AR hearts as the single βAR transgenics. However, β₁/AC5 bitransgenics showed no evidence of enhanced Smac release, indicating a potential protective role for overexpressed AC5.

Discussion

Enhancement of cardiac contractility can be accomplished by increasing βAR signaling with transgenic overexpression of the β₁AR or β₂AR subtype, AC5, or AC6. ³ , ⁶ , ⁷ , ¹⁷ Early studies with these mice revealed several intriguing findings that have provided new insights into how the heart responds over time to these different interventions and the potential for new therapeutics. Two early studies, one in the C57BL/6 background ¹⁸ and one in the FVB/N background, ⁶ showed that overexpression of the β₂AR increased resting contractility and the response to the agonist isoproterenol, with no deleterious effects throughout the life of the animal. In addition, we showed that transgenic mice overexpressing mutated β₂AR, known to have depressed coupling in transfected cells, also displayed depressed contractility and agonist responsiveness compared to wild‐type β₂AR. ⁶ This indicated that despite background levels of β₁‐ and β₂AR expression together (<100 fmol/mg), transgenic overexpression provides a model that, within limits, is useful for linking receptor signaling to physiologic function. Subsequent studies in the FVB/N background revealed that as β₂AR expression levels were increased (from ∼60‐ to ∼350‐fold over background) a progressive, time‐dependent cardiomyopathy developed in those with overexpression of approximately 100‐fold or more. ¹⁹ In contrast, several reports have indicated that relatively low levels of β₁ AR overexpression (5–20‐fold), while initially increasing contractility and the response to agonist, ultimately result in cardiomyopathy and heart failure. ³ , ⁴ While the age of onset of ventricular failure and the upper limits of nonpathogenic expression levels differ between investigators and strains, the paradigm that modest cardiac β₂AR overexpression is well tolerated in mice, while β₁ AR overexpression is not, has been generally accepted. With the mice generated in our laboratory, ³ β₁AR (the most common human allele, Arg389) overexpression at levels of 1,000–3,000 fmol/mg results in three time‐dependent physiologic phases: an early (< 4‐month‐old period) enhancement of contractility and agonist response, an approximately 5–7‐month‐old period where agonist responsiveness is absent but fractional shortening is maintained, and a > 9‐month‐old period where chamber dilatation, depressed fractional shortening, and heart failure are observed. Cardiac overexpression of AC5 and AC6 has also been reported. ⁷ , ¹⁷ Enhanced baseline and agonist‐stimulated contractility were observed with AC5 overexpression, ⁷ while the AC6 overexpressors had enhanced agonist‐stimulated contractility without an increase in the baseline contractility ¹⁷ None of these AC transgenic mouse lines displayed loss of agonist responsiveness or overt cardiomyopathy with age.

These studies of single‐gene transgenics prompted the development of potential therapeutic strategies for heart failure. For example, AC5 and AC6 overexpression has been reported to “rescue” ventricular function evoked by G_β overexpression, ¹¹ , ²⁰ as has overexpression of the β₂AR. ²¹ It has been hypothesized that unique properties of AC5/6 or the β₂ARs, aside from their cAMP/PKA‐dependent effects, may be responsible for these salutary effects. With AC6 overexpression, phospholamban expression is reduced due to enhancement of the transcriptional repressor ATF3, which binds to the phospholamban promoter. ²² This effect is not observed with isoproterenol or forskolin treatment, indicating cAMP‐independent effect. The cardiac effects of altering AC5/6 expression is nevertheless still somewhat unclear, as AC5 (‐/‐) mice show protective effects in heart failure models. AC5 (‐/‐mice have been reported to be resistant to apoptosis during chronic isoproterenol infusion, with an accompanying increase in phosphorylated Akt and Bcl2. ²³ Furthermore, AC5 (‐/‐) mice have been shown to be largely protected from pressure overload‐induced ventricular dysfunction and apoptosis. ²⁴ The β₂AR is now recognized to have distinct properties compared to ₁AR. The β₂AR subtype has been shown to couple to G. (after PKA‐mediated receptor phosphorylation), which leads to activation of ERK1/2 MAPK, Akt and c‐Src family members. β₂AR also activate proapoptotic pathways including p38 MAPK, but collectively, β₂AR signaling has been considered to be antiapoptotic. A recent study has shown that inhibition of G_i via transgenesis with a G_i‐inhibitory peptide resulted in mice with greater infarct size and apoptosis during ischemia/reperfusion, ²⁵ supporting the notion that β₂AR‐G_i signaling is protective under such conditions. In addition, β₂AR when co‐overexpressed with β₁AR using adenovirus vectors with isolated cardiomyocytes enhances isoproterenol‐stimulated myocyte contractile responses. ²⁶ This has been suggested to be due to heterodimerization of the two subtypes that form a distinct signaling unit. However, the physiologic effects of such co‐expression in the intact heart have not been demonstrated.

In the current report, we examined the physiologic and signaling consequences of β₂AR and AC5 overexpression in the setting of β₁AR‐mediated cardiomyopathy by developing β₁/β₂ a n d β₁/AC5 bitransgenic mice. In young mice, co‐overexpression of β₂‐ with β₁AR resulted in enhanced agonist‐promoted contraction and relaxation compared to β₁AR overexpression alone. This is consistent with the reports in isolated myocytes, ²⁶ but as we show in the intact heart, over time, β₂AR co‐overexpression does not attenuate the cardiomyopathic effects of β₁AR. In contrast, no enhancement was observed when AC5 was overexpressed with β₁AR. These findings suggest that (a) enhancement of βAR responsiveness can occur over that of β₁AR by co‐expression of β₂AR, despite the fact that the latter can inhibit AC via G_i coupling, and (b) the level of AC is not the “limiting” component in βAR‐mediated cardiac contraction coupling, as has been claimed by some. ¹⁷ , ²⁷ In regard to the latter, additional signaling mechanisms other than via AC could also explain the enhanced agonist‐stimulated inotropy of the β₁/β₂ hearts compared to that of the β₁/AC5 hearts. ²⁸ By 6 months of age, β₁AR overexpressors maintained elevated contractility at baseline but failed to respond to agonist. β₁/β₂ hearts had a lower baseline contractility, potentially due to the increased G_i, which is known to occur with β₁AR overexpression ³ working in concert with β₂AR‐G_i coupling, and also failed to respond to agonist. β₁/AC5 mice displayed an approximately 2‐fold increase in agonist‐promoted contraction and relaxation. The former was similar to that seen with NTG mice, and was clearly depressed compared to the 2‐month response. Relaxation at 6 months in the β₁/AC5 mice was not significantly altered compared to young hearts. We have previously shown that AC5/6 expression is depressed at 6 months in β₁AR overexpressors, so concomitant transgenic expression of AC5 may have compensated to some extent. However, at least for contractility, while responsiveness is preserved, the maximal amplitude of the response is not maintained from 2 to 6 months, and thus frank “rescue” is not afforded.

The downstream signals measured relative to apoptosis failed to reveal a robust pattern, indicative of a beneficial effect of β₂AR or AC5 on β₁AR‐mediated cardiomyopathy. Of note, our results with the single β₁AR and β₂AR transgenics differ from those of a recent paper by Peter et al., ²⁹ but this may be attributable to the differences in strain or receptor expression levels. In that paper, β₂AR overexpression promoted a cardiomyopathy accompanied by significant apoptosis, p38 MAPK and Akt activation, and reduced left ventricular ejection fraction (LVEF). Although not speciffically stated, the cited source ¹⁸ of these mice indicate that β₂AR expression was 40 pmol/mg (∼30‐fold greater than in this report) and the strain was C57BL/6, as compared to FVB/N used here. Interestingly, we did find activation of p38 MAPK in our study with the β₁AR, but the activation was substantially amplified in the β₁/β₂ and β₁/AC5 bitransgenics. This suggests that a threshold effect (potentially cAMP –dependent, since it occurs with AC5 co‐overexpression) may be at play for activation of p38 MAPK. Peter et al., also found activation of p38 MAPK in a β₁AR overexpressor, as well as Akt in older mice. Based on the reference ⁴ for the source, these mice were in the FVB/N background and presumably expressed 600 fmol/mg (not unlike our β₁AR overexpressor). However, we do not find increases in these kinases in our β₁AR overexpressors. Indeed, the only hearts with elevated Akt were the bitransgenics.

Taken together, our findings fail to reveal a significant long‐term physiologic improvement in ventricular function, or changes in the development of heart failure, by β₂AR or AC5 overexpression in β₁AR‐mediated cardiomyopathy. Tere were differences in some phenotypes between β₁/β₂ and β₁/AC5 hearts. The latter had some preservation of ventricular function and agonist responsiveness at 6 months of age, did not show a progressive downregulation of β₁AR expression, had little or no effect on cardiac mass, and had no increase in cystosolic Smac. However, by 9 months of age, the relevance of these differences appeared to be minimal, as β₁AR, β₁/β₂, an d β₁/AC5 hearts all had markedly depressed fractional shortening, with little response to agonist. We cannot exclude the possibility that different levels of overexpression of β₂AR or AC5 than used here may have given different results. Or, that AC6 overexpression may have given different results compared to our AC5 bitransgenic. Nor does this study address other approaches to maintain βAR responsiveness in heart failure, such as reversal of desensitization by inhibition of G‐protein coupled receptor kinase (CRKs). ³⁰ Nevertheless, the current study places into question whether maintenance or enhancement of β₂AR signaling, or AC activity, by overexpression of this receptor or its effector has a long‐term beneficial effect in heart failure.