Mechanism of β-Adrenergic Receptor Desensitization in Cardiac Hypertrophy Is Increased β-Adrenergic Receptor Kinase*

Pressure overload cardiac hypertrophy in the mouse was achieved following 7 days of transverse aortic constriction. This was associated with marked β-adrenergic receptor (β-AR) desensitization in vivo, as determined by a blunted inotropic response to dobutamine. Extracts from hypertrophied hearts had ≈3-fold increase in cytosolic and membrane G protein-coupled receptor kinase (GRK) activity. Incubation with specific monoclonal antibodies to inhibit different GRK subtypes showed that the increase in activity could be attributed predominately to the β-adrenergic receptor kinase (βARK). Although overexpression of a βARK inhibitor in hearts of transgenic mice did not alter the development of cardiac hypertrophy, the β-AR desensitization associated with pressure overload hypertrophy was prevented. To determine whether the induction of βARK occurred because of a generalized response to cellular hypertrophy, βARK activity was measured in transgenic mice homozygous for oncogenic ras overexpression in the heart. Despite marked cardiac hypertrophy, no difference in βARK activity was found in these mice overexpressing oncogenic ras compared with controls. Taken together, these data suggest that βARK is a central molecule involved in alterations of β-AR signaling in pressure overload hypertrophy. The mechanism for the increase in βARK activity appears not to be related to the induction of cellular hypertrophy but to possibly be related to neurohumoral activation.

The regulation of myocardial ␤-adrenergic receptors (␤-ARs) 1 involves a process characterized by a rapid loss of receptor responsiveness despite continued presence of agonist. Two classes of kinases regulate receptors through rapid receptor phosphorylation; the second messenger activated protein kinases, such as cAMP-dependent kinase A and protein kinase C (1), and the G protein-coupled receptor kinases (GRK), which phosphorylate only activated receptors leading to a process termed homologous desensitization (2,3). Of the six known members of the emerging GRK family, GRK2 (commonly known as ␤ARK1) and GRK5 appear to be dominantly expressed in the heart (4,5). Desensitization of agonist-occupied receptors by the primarily cytosolic ␤ARK1 requires a membrane-targeting event prior to receptor phosphorylation by a direct physical interaction between residues within the carboxyl terminus of ␤ARK and the dissociated, membrane-anchored ␤␥ subunits of G proteins (6,7). Unlike ␤ARK1, GRK5 does not undergo agonist-dependent translocation from cytosol to membrane but rather is constitutively membrane-bound (5).
Decreased responsiveness to ␤-AR agonists is a characteristic of chronic heart failure. In heart failure, ␤-AR desensitization is due to both diminished receptor number (receptor down-regulation) and impaired receptor function (receptor uncoupling) (8), in part related to enhanced ␤ARK activity (9,10). Recent data suggest a step wise increase in plasma norepinephrine levels in individuals from normal cardiac function to asymptomatic left ventricular (LV) dysfunction and symptomatic heart failure (11). Thus, high levels of circulating catecholamines early in the transition from stable cardiac hypertrophy (12) to LV dysfunction may account for the observed ␤-AR desensitization in the disease process.
Although cardiac hypertrophy may be regarded as an adaptive response to increased work load (13), ventricular hypertrophy (especially when accompanied by prolonged periods of hypertension) is associated with an increased incidence of heart failure (14). In models of cardiac hypertrophy, heterologous desensitization of adenylyl cyclase associated with down-regulation of ␤ 1 -ARs and increased G i␣ protein are found, which are associated with reduced positive inotropic response to isoproterenol (15,16).
We have recently shown that ␤ARK is a critical modulator of in vivo contractile function (17). Both the ␤ 1 -adrenergic and angiotensin II receptors are targets for ␤ARK1-mediated desensitization, whereas selected desensitization of ␤ 1 -ARs occurs with GRK5 (17,18). Transgenic mice, which overexpress a peptide inhibitor of ␤ARK1, show decreased desensitization of ␤-ARs supporting the in vivo role for ␤ARK1 as a modulator of cardiac function (17). Since abnormalities in ␤-adrenergic signal transduction may be one of the earliest changes in the transition from compensated hypertrophy to decompensated heart failure, we wanted to determine whether cardiac hypertrophy is associated with ␤-AR desensitization, and we tested the hypothesis that in LV pressure overload, hypertrophy desensitization occurs as a result of enhanced ␤ARK1 activity.

MATERIALS AND METHODS
Experimental Animals-Adult wild-type (strain C57/B6) and transgenic mice of either sex and 3-6 months of age were used for this study. Two types of transgenic mice were used for this study, 1) mice with cardiac-specific overexpression of a ␤ARK1 inhibitor (␤ARKct) shown previously to have diminished desensitization to ␤-AR stimulation (17) and 2) transgenic mice homozygous for cardiac-targeted oncogenic ras known to develop severe cardiac hypertrophy in the absence of hemodynamic overload (19). The animals in this study were handled according to the animal welfare regulations of the University of California, San Diego, and the protocol was approved by the Animal Subjects Committee of this institution.
Microsurgical Techniques-Mice were anesthetized with a mixture of ketamine and xylazine (20). After endotracheal intubation, mice were connected to a rodent ventilator. Using microsurgical procedures as described previously (21), the chest cavity was entered in the second intercostal space, and the transverse aorta between the right (proximal) and left (distal) carotid arteries was isolated. Transverse aortic constriction (TAC) was performed by tying a 7-0 nylon suture ligature against a 27-gauge needle, the latter being promptly removed to induce pressure overload cardiac hypertrophy (20,22). After aortic constriction, the chest was closed, the pneumothorax was evacuated, and the mice were extubated and allowed to recover from the anesthesia. Shamoperated animals underwent the same operation except for aortic constriction.
Hemodynamic Evaluation in Intact Anesthetized Mice-After 7 days of aortic constriction, mice were anesthetized and reweighed, and the left carotid artery (distal to the stenosis) was cannulated with a flamestretched PE-50 catheter connected to a modified P-50 Statham transducer. Either a 1.4 French (0.46 mm diameter) or 1.8 French (0.61 mm diameter) high fidelity micromanometer catheter (Millar Instruments) was inserted into the right carotid (proximal to the stenosis), and simultaneous aortic pressures were measured. Following bilateral vagotomy, the micromanometer was advanced retrograde into the LV. Hemodynamic measurements were recorded at base line and following 2-min infusions of dobutamine at 0.5, 1.0, 1.5, and 2.0 g/kg/min. Continuous high fidelity LV and fluid-filled aortic pressure was recorded simultaneously on an 8 channel chart recorder and in digitized form at 2000 Hz for later analysis. Experiments were then terminated, hearts were rapidly excised, and individual chambers were separated, weighed, and frozen in liquid N 2 for later biochemical analysis.
GRK Activity by Rhodopsin Phosphorylation-Myocardial extracts were prepared by homogenization of excised hearts in ice-cold lysis buffer (2 ml) (25 mM Tris-Cl (pH 7.5), 5 mM EDTA, 5 mM EGTA, 10 g/ml leupeptin, 20 g/ml aprotinin, and 1 mM PMSF) and centrifuged at 48,000 ϫ g for 30 min. The supernatants, which contain soluble kinases (i.e. ␤ARK1), were kept. The pelleted membranes were rehomogenized in lysis buffer containing 250 mM NaCl, put on ice for 30 min to dissociate membrane-bound kinase (i.e. GRK5), and centrifuged at 48,000 ϫ g for 30 min. This supernatant (membrane fraction) was kept. Both the cytosolic and membrane fractions were further purified by adding a slurry of 50% (v/v) diethylaminoethyl Sephacel (pH 7.0) in the presence of 50 mM NaCl for the cytosolic fraction (␤ARK1) and 250 mM NaCl for the membrane fraction (250 mM NaCl was used to dissociate membrane associated GRKs) and incubating at 4°C for 30 min. To remove NaCl from the suspension, fractions were run over an ion exchange column (18,23). Final supernatants were eluted with no salt lysis buffer and concentrated using a Centricon (Amicon-30) microconcentrator. Protein concentration was determined by modified Lowry method (Bio-Rad DC Protein Assay kit). Concentrated cytosolic (300 -400 g of protein) and membrane extracts (12-15 g of protein) from sham and TAC ventricles were incubated with rhodopsin-enriched rod outer segments (ROS) (17) in reaction buffer (25 l) containing 10 mM MgCl 2 , 20 mM Tris-Cl, 2 mM EDTA, 5 mM EGTA, and 0.1 mM ATP (containing [␥-32 P]ATP). After incubating in white light for 15 min at room temperature, reactions were quenched with ice-cold lysis buffer (300 l) and centrifuged for 15 min at 13,000 ϫ g. Sedimented proteins were resuspended in 20 l of protein-gel loading dye and electrophoresed through 12% SDS-polyacrylamide gels (24). Phosphorylated rhodopsin was visualized by autoradiography of dried polyacylamide gels and quantified using a phoshorimaging system, Molecular Imager GS-250 (Bio-Rad Laboratories).
To confirm ␤ARK-dependent phosphorylation of ROS, the protein kinase A inhibitor, PKI (1 M, Sigma), and heparin (10 g/ml) were incubated with purified ␤ARK, and the capacity to phosphorylate lightactivated rhodopsin was determined. Heparin inhibited phosphorylation while PKI did not (data not shown).
Total ␤-AR density was determined by incubating 25 g of the above membranes with a saturating concentration of 125 I-cyanopindolol in 500 l of binding buffer (17). Nonspecific binding was determined in the presence of 20 M alprenolol. Binding assays were conducted at 37°C for 60 min and terminated by rapid vacuum filtration over glass fiber filters, which were subsequently washed and counted in a gamma counter. Specific binding was normalized to membrane protein and reported as picomoles of receptor per milligram of membrane protein.
Statistical Analysis-Data are expressed as mean value Ϯ S.E. To examine the effect of the dobutamine on changes in hemodynamic parameters between the control and TAC groups, a two-way repeated measures analysis of variance (ANOVA) was used. Post hoc analysis with regard to differences in mean values between the groups at a specific dose was conducted with a Newman-Keuls test. To test for statistical difference in adenylyl cyclase activity, a one-factor ANOVA was used. A Student's t test was used to test for statistical difference in the parameters of LV hypertrophy, ␤-AR density, and GRK activity. For all analyses, p Ͻ 0.05 was considered significant.

RESULTS
Pressure overload cardiac hypertrophy was achieved following 7 days of TAC, which resulted in a significant increase in LV weight to body weight ratio (34%) and LV to tibia length ratio (39%) compared with sham-operated mice (Fig. 1A). The hemodynamic response to chronic aortic constriction was followed by monitoring the pressure gradient between the two carotid arteries (proximal and distal) in anesthetized animals prior to the measurement of ventricular hemodynamics. A significant gradient across the surgically induced stenosis was present at 7 days after TAC (Fig. 1B). Systolic aortic pressure proximal to the stenosis in the TAC group was slightly higher than sham-operated animals but did not reach statistical significance (p ϭ 0.08). This moderate degree of constriction provides an adequate mechanical stimulus for the development of cardiac hypertrophy (Fig. 1A) without the development of cardiac failure (20,26).
To determine whether the development of LV hypertrophy is associated with ␤-AR uncoupling in vivo, cardiac catheterization was used to measure catecholamine responsiveness in intact anesthetized mice (Fig. 2). Marked blunting of the dobutamine (a ␤ 1 -agonist) induced inotropic response ( Fig. 2A) as well as an attenuated fall in LV dP/dt min (an index of myocardial relaxation) (Fig. 2B) was observed in mice with cardiac hypertrophy. The LV systolic pressure and heart rate response to dobutamine between the groups were not different (Fig. 2, C  and D). These data demonstrate that ␤-AR desensitization is associated with the development of cardiac hypertrophy.
To determine whether alterations in GRK activity could account for the ␤-AR uncoupling seen in these hypertrophied hearts, extracts were prepared and assayed for phosphoryla-tion of the G protein-coupled receptor rhodopsin. Cytosolic extracts from TAC hearts had Ϸ3-fold increase in GRK activity compared with extracts from hearts of sham-operated animals (Fig. 3, A and B). Protein immunoblotting was used to detect the level of ␤ARK1 in the cytosol of hypertrophied hearts (Fig.   FIG. 3. Effect of pressure overload hypertrophy on ␤ARK activity and protein. A, cytosolic extracts from sham-operated and TAC hearts were measured for their capacity to phosphorylate rhodopsin. 300 g of cytosolic protein was incubated with 350 pmol of rhodopsinenriched ROSs in lysis buffer (total volume 25 l). Phosphorylated rhodopsin was visualized by autoradiography following electrophoresis through 12% SDS-polyacrylamide gels. Lane 1, C, ROS in the absence of heart extract; lanes 2-5, cytosolic extracts from individual shamoperated mice;, lanes 6 -9, individual TAC-operated hearts. Each lane (2-9) represents extracts from a separate heart. Lane 10 -12, incubation of ROS with 12.5, 25, and 50 g of purified ␤ARK1, respectively. Shown is a representative autoradiograph of a dried gel where phosphorylated rhodopsin (Rho) is visualized. B, the level of ␤ARK activity in 12 sham-operated and 11 TAC hearts. Activities were calculated as 32 P incorporation (fmol/min/mg of cytosolic protein), *, p Ͻ 0.001. C, immunodetection of myocardial level of ␤ARK1 in cytosolic extracts from individual sham-operated (lanes 1-4) and individual TAC operated (lanes 5-8) mice. An Ϸ80-kDa protein was visualized by Western blotting and chemiluminescence following solubilization of cytosolic extracts and immunoprecipitation. Lane 9, 50 g of purified ␤ARK1. and left ventricular weight to tibia length (LV/T) was measured in sham-operated (n ϭ 24) and 7 days following transverse aortic constriction (TAC) (n ϭ 23). The largest body weight (either preoperative or postoperative) was used to calculate the ratios of (LV) weight to body weight to eliminate potential bias from postoperative weight loss in the aortic-constricted animals. Significant cardiac hypertrophy resulted following TAC. *, p Ͻ 0.001 TAC versus sham. B, systolic aortic pressure was measured in the carotid arteries proximal and distal to the region of stenosis created using a suture ligature chronically tied around the transverse aorta. Hemodynamic measurements were performed in 12 sham-and 11 TAC-operated animals. TAC resulted in a significant difference between proximal and distal pressures confirming the presence of a chronic hemodynamic load in vivo in the mice. †, p Ͻ 0.001, t test with Bonferroni correction for four comparisons.
FIG. 2. In vivo assessment of LV contractile function in response to ␤-agonist stimulation. Cardiac catheterization was performed in the intact anesthetized mice using a 1.8 French high fidelity micromanometer. Parameters measured were heart rate, aortic pressure, LV systolic, and end diastolic pressure, and the maximal and minimal first derivative of LV pressure (LV dP/dt max and dP/dt min , respectively). Ten sequential beats were averaged for each measurement. Four measured parameters are shown at base line and after progressive infusion of dobutamine in sham-operated (E, n ϭ 12), and TAC (q, n ϭ 11) mice. A, LV dP/dt max ; B, LV dP/dt min ; C, LV, systolic pressure; D, heart rate. Data were analyzed with a two-way repeated ANOVA and post hoc analysis with regard to differences in mean values between the groups at a specific dose was conducted with a Newman-Keuls test. *p Ͻ 0.001; †, p Ͻ 0.05; control versus transgenic. The pattern of change between groups was statistically significantly for LV dP/dt max , p Ͻ 0.005 (A) and LV dP/dt min , p Ͻ 0.001 (B). 3C). Consistent with the marked increase in GRK activity, a clear increase in ␤ARK1 protein levels was observed by Western blotting in hearts following TAC (Fig. 3C).
Since GRK5 is associated with the membrane, the level of GRK activity was measured in cardiac membranes of hypertrophied and non-hypertrophied hearts by assessing their capacity to phosphorylate light-activated rhodopsin (Fig. 4). A 2.5fold increase in membrane GRK activity was observed. These data suggest that in cardiac hypertrophy, increased GRK levels in both cytosol and membrane fractions may account for the observed desensitization of ␤-ARs in vivo.
To identify the specific GRK that mediates ␤-AR desensitization in the hypertrophied heart, monoclonal antibodies that neutralize ␤ARK1 and ␤ARK2 (C5/1) or GRK4, GRK5, and GRK6 (A16/17) were used to inhibit endogenous GRK activity (25). In a rhodopsin phosphorylation assay using purified ␤ARK1 as a control, pre-incubation with the monoclonal antibody C5/1, significantly inhibited phosphorylation. (Fig. 5A). When the monoclonal antibody C5/1 was added to cytosolic extracts from hypertrophied hearts, phosphorylation of rhodopsin was significantly inhibited as shown by the negligible GRK activity (Fig. 5A). Incubation of either purified ␤ARK1 or cytosolic extracts with the monoclonal antibody A16/17 did not inhibit rhodopsin phosphorylation (data not shown). This indicates that ␤ARK1 accounts for the increase in cytosolic kinase activity with cardiac hypertrophy.
Membrane GRK activity was also studied with monoclonal antibodies since the increase in membrane GRK activity could be due to either enhanced expression of GRK5 or increased membrane translocation of ␤ARK1 (a process required for activation) (7). Membrane extracts from sham-operated and TAC mice were incubated in the presence of the monoclonal C5/1 or the monoclonal antibody A16/17 (Fig. 5B). As shown (Fig. 5B,  lanes 5-8), a 2.6-fold increase in GRK activity was present in membrane extracts of hypertrophied hearts compared with sham-operated hearts. Incubation with A16/17 (Fig. 5B, lanes  9 -12) partially reduced membrane GRK activity (1.6-fold higher than non-hypertrophied controls). In contrast, incubation with the monoclonal antibody C5/1 (Fig. 5B, lanes 13-16) markedly inhibited membrane GRK activity to levels near that of non-hypertrophied control hearts. These data suggest that the predominant GRK responsible for the increase in membrane kinase activity in cardiac hypertrophy is ␤ARK1.
␤-AR density was determined in cardiac membranes prepared from sham (n ϭ 12) and TAC (n ϭ 12) operated hearts. No difference was found in the number of receptors between the two groups (21.8 Ϯ 1.3 versus 25.1 Ϯ 2.2, fmol/mg of membrane protein, p ϭ not significant). In contrast to the lack of change in ␤-AR density, hypertrophied hearts had significantly lower isoproterenol-stimulated adenylyl cyclase activity (Table I). Thus, functional uncoupling of ␤-ARs as assessed by the physiologic response to ␤-agonist stimulation and adenylyl cyclase activation was hindered with the development of myocardial hypertrophy. To confirm that the observed uncoupling of ␤-ARs was not due to the increase in inhibitory G protein (G i␣ ), we measured the level of G i␣ immunoreactivity with antibodies directed against G i␣ 1-3 (Fig. 6). No difference in the level of G i␣ was observed by immunoblotting in sham and hypertrophied hearts. These data suggest that impaired catecholamine responsiveness with myocardial hypertrophy is not due to alterations in either ␤-AR density or to levels of inhibitory G protein.
To determine whether the impaired catecholamine responsiveness and ␤-AR desensitization that accompany the development of cardiac hypertrophy could be accounted for by the induction of ␤ARK, we used a strategy of in vivo ␤ARK inhibition achieved with cardiac-specific overexpression of a ␤ARK1 inhibitor in transgenic mice (17). The inhibitor utilized was the carboxyl terminus of ␤ARK1 (␤ARK1ct), which contains the G␤␥-binding domain and competes with endogenous ␤ARK for binding and subsequent translocation/activation (6,7,17). Transgenic mice overexpressing the ␤ARKct underwent TAC, and the in vivo hemodynamic response to dobutamine was determined 7 days later. Following TAC, hearts from transgenic mice with overexpression of the ␤ARKct developed LV hypertrophy to the same degree as wild-type mice (a 38% increase in LV to body weight ratio compared with shamoperated ␤ARKct inhibitor mice, Fig. 7A). To determine whether this increase in LV mass was associated with increased ␤ARK activity, cytosolic extracts from sham-and TACoperated ␤ARKct mouse hearts were used to test for their capacity to phosphorylate rhodopsin. A 3-fold increase in GRK activity was found in cytosolic extracts from hypertrophied  1-4), and TAC hearts (lanes [5][6][7][8] were assessed for the capacity to phosphorylate rhodopsin (Rho). Lanes 9 -11 represent 1:5000, 1:2000, and 1:1000 dilution of purified GRK5, respectively. See Fig. 3 for details.  compared with sham-operated ␤ARKct mice (Fig. 7B). To determine whether forced overexpression of the ␤ARKct would prevent functional ␤-AR desensitization with the development of cardiac hypertrophy, we performed cardiac catheterization in sham and TAC mice overexpressing the ␤ARKct inhibitor. As shown in Fig. 8, the response in LV dP/dt max and heart rate to ␤-AR stimulation was similar in the hypertrophied transgenic mice compared with sham-operated transgenic liter mates (Fig. 8). Similarly, no statistical difference in the LV dP/dt min response to dobutamine was found (sham, Ϫ9992 Ϯ 786 to Ϫ12356 Ϯ 874 mmHg/s; TAC, Ϫ9654 Ϯ 952 to Ϫ10750 Ϯ 1237, mmHg/s, p ϭ not significant ANOVA).
To address the question of whether the induction of ␤ARK1 in pressure overload hypertrophy is simply a generalized response to cellular hypertrophy, we measured the level of ␤ARK activity in transgenic mice homozygous for oncogenic ras overexpression in the heart, which develop cardiac hypertrophy in the absence of hemodynamic overload (19). Compared with wild-type, age-matched controls, overexpression of oncogenic ras resulted in a 62% increase in LV to body weight ratio (5.8 Ϯ 0.35 versus 3.59 Ϯ 0.08 mg/g, p Ͻ 0.001). Cytosolic extracts from hearts of these transgenic mice were used to measure ␤ARK activity in a rhodopsin phosphorylation assay. Despite the development of severe LV hypertrophy, no difference in the level of ␤ARK activity was found in mice overexpressing oncogenic ras compared with wild-type matched controls (Fig. 9). These data suggest that the increase in ␤ARK protein and activity in response to pressure overload is not a general phenomenon that accompanies the development of cellular hypertrophy but is specific for pressure overload-induced hypertrophy.

DISCUSSION
The present study demonstrates that ␤-AR desensitization occurs with the development of pressure overload cardiac hypertrophy, and that the uncoupling of ␤-ARs with cardiac hypertrophy can be accounted for by an increase in ␤ARK1. The blunted inotropic response to ␤-AR stimulation that accompanies LV hypertrophy can be reversed by ␤ARK inhibition. Furthermore, the mechanism for increased ␤ARK levels and activity appear not to be related to the induction of cellular hypertrophy but rather is specific for that which occurs in response to hemodynamic overload. These data demonstrate that in the mouse, ␤ARK1 is responsible for altered ␤-AR signaling associated with pathologic states such as pressure overload cardiac hypertrophy.  ␤ARK inhibitor (␤ARKct). A, left ventricular weight to body weight (LV/BW) and left ventricular weight to tibia length (LV/T) was measured in ␤ARKct animals 7 days following either sham-operation (n ϭ 10) or transverse aortic constriction (TAC) (n ϭ 13). *, p Ͻ 0.05 TAC versus SHAM. B, cytosolic extracts from sham-operated and TAC hearts from ␤ARKct transgenic mice were measured for their capacity to phosphorylate rhodopsin. Lane 1, C, ROS in the absence of heart extract; lanes 2-5, cytosolic heart extracts from individual sham-operated ␤ARKct mice; lanes 6 -9, individual TACoperated ␤ARKct mice. Lanes 2-9 each represents extracts from separate hearts; lane 10, incubation of ROS with 25 g of purified ␤ARK1.

FIG. 8.
In vivo contractile function of transgenic mice overexpressing a ␤ARK inhibitor following the development of cardiac hypertrophy. Cardiac catheterization was performed in intact anesthetized transgenic animals. The methods used were as in Fig. 2. Shown is LV dP/dt max (A), heart rate (B), and the difference from base line for LV dP/dt max (C, ⌬LV dP/dt max ) and heart rate (D, ⌬heart rate) with dobutamine infusion in sham-operated (E), n ϭ 10, and TACoperated (q) mice, n ϭ 13. A two-way repeated ANOVA showed no significant difference between sham and TAC ␤ARK inhibitor mice. To determine if the values for basal dP/dt max were different and whether it influenced the response ␤-AR stimulation, an analysis of covariance was performed. No significant effect of basal LV dP/dt max (covariate) was found. Furthermore, just treating basal dP/dt max as 2 isolated groups using a Student's t test (without correction for multiple comparisons) no significance was reached. No significant difference was found with respect to difference from base line for LV dP/dt max , and heart rate (C and D).
Several previous experimental studies have demonstrated ␤-AR desensitization with cardiac hypertrophy (16,(27)(28)(29). In two studies, desensitization was associated with a small decrease in ␤ 1 -AR subtype and the increase of the inhibitory G protein G i␣ , which was reversed with angiotensin converting enzyme inhibitor treatment (16,27). In this study, we show that the marked impairment of catecholamine responsiveness associated with cardiac hypertrophy is related to an increase in GRK activity that predominantly can be accounted for by the increase in ␤ARK1 subtype. There also may be a slight increase in GRK5 activity. The impairment in contractile function with ␤-agonist stimulation following the development of cardiac hypertrophy ( Fig. 2A) is remarkably similar to that observed with targeted overexpression of ␤ARK1 (17), providing further evidence that ␤ARK1 is a critical in vivo modulator of ␤-AR signal transduction.
Although ␤ARK1 is predominantly a cytosolic enzyme, translocation to the membrane by G␤␥ is required for desensitization of ␤ARs. In contrast, GRK5 is always membrane-associated. We used monoclonal antibodies to discriminate between ␤ARK1 and GRK5 subtypes in the hypertrophied heart. These experiments demonstrate that in both cytosolic and membrane fractions, ␤ARK1 was the subtype predominately responsible for the enhanced GRK activity and suggests that the majority of membrane activity can be accounted for by ␤ARK1, which is translocated to the membrane (Fig. 5). Furthermore, the fact that ␤ARK activity is enhanced in the ␤ARKct mice (in the presence of an inhibitor of G␤␥ translocation) without affecting cardiac function, suggests that just an increased level of ␤ARK will not enhance its association to the membrane to cause ␤-AR desensitization. This is similar to recent findings that suggest that brief cardiac ischemia in an isolated rat heart preparation is of sufficient stimulus to induce membrane GRK activity (30). Our studies here in cardiac hypertrophy go further in that we used monoclonal antibodies to show that a small degree of membrane kinase activity could be accounted for by GRK5.
Tissue distribution of the various members of the GRK family has been assessed by mRNA expression (4). GRK3 and GRK6, in addition to ␤ARK1 and GRK5, can be identified in cardiac tissue; however, the level of mRNA expression is significantly less than ␤ARK1 and GRK5. Therefore, it is possible that other GRK subtypes such as ␤ARK2 or GRK6 could account for some of the observed GRK activity. Quantitative immunoblotting of all subtypes would be required to demonstrate that ␤ARK1 and GRK5 are the sole GRKs present in heart tissue.
Induction of myocardial hypertrophy involves activation of several signaling pathways involving key effector molecules such as ras, mitogen-activated protein kinase, and other signaling cascades that are linked to the cell surface through stimulation of G protein-coupled receptors (22,31,32). An important issue is whether the increase in GRK activity in response to pressure overload involves specific interactions with the ␤-AR signaling pathway. We utilized a genetic model of cellular hypertrophy to demonstrate that the increase in ␤ARK associated with pressure overload hypertrophy was not simply a generalized cellular response of increase protein synthesis. ␤ARK activity in heart extracts from transgenic mice overexpressing oncogenic ras (which develop massive myocardial hypertrophy) had no increase in cytosolic ␤ARK activity (Fig. 9). Clearly, cardiac hypertrophy induced by pressure overload is a very different pathologic state than that induced through forced overexpression of oncogenic ras. Indeed, our previous report of the ras transgenic showed no abnormality in contractility at base line or with isoproterenol stimulation but altered relaxation presumably due to mechanical effects (19). Thus, enhanced GRK activity with LV pressure overload appears not to be directly related to myocyte hypertrophy but perhaps to activation of the sympathetic nervous system and/or the renin angiotensin axis.
A question that we attempted to address in this study is whether increased ␤ARK activity is required for the development of hypertrophy. Although pressure overload cardiac hypertrophy is associated with enhanced ␤ARK activity, transgenic mice that overexpress an inhibitor of endogenous ␤ARK activity develop cardiac hypertrophy to the same extent as wild-type mice (Fig. 7). Since it is possible that the increase in ␤ARK could have other cellular effects not inhibited by the ␤ARKct, we cannot definitively prove that ␤ARK is not required for the hypertrophic phenotype. Nonetheless, these data do suggest that inhibiting endogenous ␤ARK translocation and ␤-AR desensitization do not prevent the development of the hypertrophy in response to a mechanical/neurohumoral stimulus.
The mechanism for the increase in ␤ARK is not certain. Recent evidence suggest that both phosphatidylinositol 4,5bisphosphate (33, 34) (although concentration-dependent (35)) and protein kinase C (36) can directly regulate ␤ARK activity to enhance ␤-AR phosphorylation and initiate desensitization. G protein-coupled signaling cascades that involve protein kinase C are generally considered to be importantly involved in the regulation of cell growth, which occurs in response to mechanical stimuli (37,38) and receptor stimulation (31). Whether this may account for the small but significant greater responsiveness in the hypertrophied transgenic mice overexpressing ␤ARKct is not certain.
GRK5 is expressed mostly in the heart compared with other tissues (5). It does not undergo agonist-dependent translocation from cytosol to membrane but rather is constitutively membrane bound (5) and will effectively phosphorylate and desensitize ␤-ARs both in vitro and in vivo (18,25). Our results in membrane fractions demonstrate that the predominant GRK activity in the membrane of hypertrophied hearts is ␤ARK1 (Fig. 5). Recent evidence suggest that GRK5, but not ␤ARK1, can be inhibited by Ca 2ϩ /calmodulin (39). Calmodulin appears to be an important molecule in cardiomyocyte growth as shown by the proliferative and hypertrophic response of atrial myocytes when overexpressed in transgenic mice (40). Although it is possible that in hypertrophied hearts GRK5 activity was partially inhibited by calmodulin, it is not known whether intracellular calmodulin levels were increased in these hearts with pressure overload cardiac hypertrophy. Nonetheless, our data suggest that the increased membrane GRK activity is due to enhanced ␤ARK translocation.
The high fidelity catheters used for these experiments were either a 1.8 French or a 1.4 French micromanometer, which have diameters of 0.6 and 0.46 mm, respectively. The length of the sensor is 2.5 mm. We have previously assessed LV enddiastolic and end-systolic diameter by echocardiography and show that, for normal mice of approximate weight and age used in this study, the average diameter of the ventricle is 3.73 mm (end-diastole) and 2.2 mm (end-systole), whereas for mice with FIG. 9. GRK activity in cytosolic extracts from transgenic mice overexpressing oncogenic ras. The capacity of cytosolic extracts from age-matched, wild-type mice (lanes 1-5) and transgenic mice overexpressing oncogenic ras (lanes 6 -10) is tested in a rhodopsin phosphorylation assay. See Fig. 3 for details. moderate hypertrophy, LV end-diastolic diameter is Ϸ3.4 mm and LV end-systolic diameter is Ϸ2 mm (26). Thus, the diameter of the catheter is considerably smaller than the diameter of the ventricle in either control or hypertrophied hearts and suggests that catheter size is unlikely to influence the measurement of hemodynamic variables.
Cardiac hypertrophy that develops following acute pressure overload is clearly different from that which occurs secondary to hypertension or valvular disease in the clinical setting, and thus, these data must be interpreted as such. In this regard, we have previously investigated the degree of myocardial injury in this overload model of mechanical-induced hypertrophy. In developing this mouse model of hypertrophy, we have chosen to constrict the transverse aorta that, in contrast to ascending aortic constriction, avoids excessive overload on the left ventricle because of the position of the innominate artery proximal to the constriction, allowing for a low resistance outlet (20,21). Furthermore, we have performed histological examinations on hypertrophied ventricles and find that only the occasional heart will have small localized foci of cardiac damage accounting for less than 5% of the ventricle (41).
␤-AR desensitization has been shown to occur in end-stage human heart failure and is considered to be an important alteration leading to contractile dysfunction and impaired exercise tolerance (42). As we observed in hypertrophy, this desensitization in part, may be secondary to elevated ␤ARK protein and activity (9,10). Chronic pressure overload resulting from hypertension is a leading cause of chronic heart failure (14), and the onset of hypertrophy is a well accepted prognostic indicator for subsequent cardiac dysfunction and morbid events (43). Whether ␤-AR desensitization, which occurs in hypertensive cardiac hypertrophy (44), is a factor in the pathological process for the transition from compensatory hypertrophy to overt heart failure is unknown. This study demonstrates that in an animal model of pressure overload hypertrophy, ␤ARK activity is increased early in the disease state and is a potential therapeutic target when considering reversal of the impaired catecholamine responsiveness.