INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

OVER THE YEARS, investigators have reported abnormalities in the myocardial -adrenergic receptor (-AR) signaling pathway in heart failure (3-5, 21, 24, 27, 35). It is generally believed that in the myocardium, downregulation of the -AR system results in desensitization due to elevated circulating and tissue levels of catecholamines. In addition to the abnormalities in the myocardium, decreases in -AR density and adenylyl cyclase activity have been reported in skeletal muscle in a dog model of heart failure (25). The -AR system plays a major role in the pathophysiology of heart failure not only in the heart but also in the vasculature. Understanding the changes that occur in both the heart and the vasculature is now even more important because of the data demonstrating the beneficial effects of -AR blockade in the treatment of patients with heart failure (1, 14, 28). Although -ARs are located in both myocardial and vascular tissues, most of the work on changes in -AR signaling in heart failure has focused on the myocardium, and alterations of -AR signaling in the vasculature have received less attention. The data on changes in the vascular -AR system show that in a dog model of heart failure -AR density was decreased in the mesenteric arteries (24), and in a rat model of heart failure (26) both cAMP- and cGMP-mediated vasorelaxation were decreased.

Abnormalities in the vascular -AR system are important in heart failure because this system is primarily responsible for vascular smooth muscle-dependent vasodilatation. In heart failure, decreased vasodilatation at baseline and a decreased vasodilatory response to -AR stimulation leads to impaired peripheral vasodilatation and inappropriate vasoconstriction. The end result is increased left ventricular (LV) afterload and worsening LV function. To help understand the mechanisms that control vascular -AR function in heart failure, we measured in vivo and in vitro -AR-stimulated responses and vascular -AR molecular signaling components in the rat coronary artery ligation model of heart failure. The purpose of the current study was to define the abnormalities in the vascular -AR signaling cascade in an attempt to determine how these changes affect the pathophysiology of heart failure.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Two groups of Sprague-Dawley rats weighing 175-275 g were used in this study: sham noninfarcted rats and myocardial infarction (MI) rats. In both sham and MI rats, in vivo -AR-stimulated responses were examined by measuring hemodynamic parameters at baseline and after administration of escalating doses of isoproterenol (0-2 µg/kg). In vitro -AR-stimulated vascular responses were studied by measuring isoproterenol (10⁸-10⁴ M)-induced vasorelaxation in isolated aortic rings. Molecular signaling components of the vascular -AR system measured in the current study included -AR density, cAMP and cGMP nucleotide levels, inhibitory G protein subunit (G_i), stimulatory G protein subunit (G_s), and -AR-coupled kinase 1 (-ARK1) protein levels.

Experimental MI. Heart failure was created in rats using standard techniques developed in our laboratory (18); these techniques result in 50% mortality within the first 3 wk. Animals without confirmation of infarction at time of death were employed as sham-operated controls. Animals were studied 6 wk after surgery.

LV hemodynamics (-AR-mediated response in vivo). Studies were performed 3 wk after coronary artery ligation. The hemodynamic measurements were obtained using methods that have been reported previously by our laboratory (19, 29, 35). In brief, after anesthesia, a 2-F catheter with two pressure sensors was inserted through the left carotid artery such that one sensor was located in the LV and the second sensor was located in the ascending aorta. We measured heart rate, aortic pressure, LV end-diastolic pressure, and rate of LV pressure change (LV dP/dt) using a two-sensor pressure transducer (Millar). After the baseline measurements were obtained, the data were recorded after -AR stimulation with increasing doses of isoproterenol (0-2 µg/kg).

Relaxation of arterial rings (-AR-mediated response in vitro). In separate groups of animals, after the baseline in vivo measurements were completed, the rat was euthanized and a 3- to 4-mm segment was cut from the ascending aorta for physiological studies before and after removal of the endothelial layer by gentle rubbing. The rest of the aorta and both carotid arteries were frozen for biochemical studies. The relaxation response of the ascending aorta was examined using a commercially available mounting apparatus attached to a force transducer. The arterial segment was attached to stainless steel wire stirrups. One wire was fixed in place and the other was attached to a force transducer. This preparation was suspended in an aerated organ bath chamber filled with 7 ml of modified Krebs-Henseleit solution maintained at 37°C by an outer water jacket. Studies were carried out by passively stretching the segments to 1 g, a predetermined baseline tension that results in maximum developed tension, for at least 45 min. Rings were then precontracted with 30 mM KCl. The presence of intact endothelium was verified physiologically by the ring response to 10⁶ M ACh and histologically by staining of the endothelial cells with factor VIII.

8

4

9

6

6

4

-AR density assay. Arterial membranes were prepared as previously described (17, 25) for heart tissue. Arterial tissue had to be pooled to yield enough protein for the binding assay. In brief, tissue was homogenized in ice-cold lysis buffer [5 mM Tris · HCl at pH 7.4, 5 mM EDTA, 10 µg/ml leupeptin, 20 µg/ml aprotinin, and 1 mm phenylmethylsulfonyl fluoride (PMSF)]. Nuclei were obtained by centrifugation at 500 g for 15 min and the supernatant was filtered through two layers of cheesecloth. Final membranes were sedimented at 40,000 g for 15 min and washed in binding buffer (in mM: 75 Tris · HCl at pH 7.4, 12.5 MgCl₂, and 2 EDTA) before resuspension. Ligand-binding assays were performed in triplicate on membranes in 500 µl volume of binding buffer with saturating concentrations of the ¹²⁵I-labeled-AR ligand cyanopindolol (~500 pM). Nonspecific binding was measured in the presence of 1 µM alprenolol. Assays were performed at 37°C for 60 min and samples were filtered over glass-fiber filters then washed and counted using a gamma counter. Specific binding (B_max) was normalized to membrane protein.

Protein immunoblotting for G_i and G_s. Western analysis was performed for G_i and G_s as previously described (31). In brief, thoracic aortas were ground using a mortar and pestle and added to ice-cold buffer (25 mM Tris · HCl at pH 7.5, 5 mM EDTA, 5 mM EGTA, 10 µg/ml leupeptin, 20 µg/ml aprotinin, and 1 mM PMSF). After homogenization, nuclei and tissue were separated by centrifugation at 800 g for 15 min. Protein concentrations were determined on the supernatant (cytosolic fraction). Sedimented proteins (membrane fraction) were resuspended in 50 mM HEPES (at pH 7.3 with 5 mM MgCl₂), then 10 µg of membrane proteins were run on a 12% Tris-glycine gel and transferred to nitrocellulose membrane. G_i and G_s were both identified using polyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) and chemiluminescent detection of anti-rabbit IgG conjugated with horseradish peroxidase (Blaze, Pierce).

G protein receptor kinase activity. The activity assay was performed as described previously (31). In brief, supernatant prepared from pooled arterial tissues (3-5 mg) were adjusted to 50 mM NaCl concentration and partially purified using DEAE Sephacel chromatography. Between 50 and 60 µg protein of purified samples were incubated in a volume of 100 µl of lysis buffer (20 mM Tris · HCl at pH 7.4, 250 mM sucrose, 5 mM EDTA, 10 µg/ml leupeptin, 10 µg/ml aprotinin, and 0.1 mM PMSF) supplemented with 0.1 mM ATP (containing [-³²P]ATP), 10 mM MgCl₂, and rhodopsin-enriched rod outer segments. After 15 min incubation in white light, the reactions were quenched with 300 µl of ice-cold lysis buffer and centrifuged for 15 min at 13,000 g. The sedimented proteins were resuspended in 25 µl of protein gel-loading dye and electrophoresed on 12% polyacrylamide/Tris-glycine gels. Gels were then dried and phosphorylated rhodopsin was quantitated with a PhosphorImager (Molecular Dynamics).

Statistical analysis. Student-Newman-Keuls tests were used to determine alterations in all measured variables in heart failure compared with sham rats. The effects of heart failure on hemodynamics and in vitro isoproterenol, forskolin, and IBMX dose responses were determined by two-way ANOVA (disease × intervention) with repeated measures.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline hemodynamics. Changes in body weight, systolic pressure, diastolic pressure, LV dP/dt, and LV end-diastolic pressure after MI are listed in Table 1. The important changes with heart failure after MI were decreases in systolic and diastolic pressures, LV dP/dt, and increases in LV end-diastolic pressure. Right ventricular (RV) weight was increased, although LV weight-to-body weight was unchanged. These data are consistent with previous reports from our laboratory in the same model (17).

-AR-mediated response in vivo. The -AR-stimulated changes in heart rate, LV dP/dt, and peripheral vascular resistance are shown in Fig. 1. Heart failure resulted in no change in heart rate, a decrease in LV dP/dt, and an increase in systemic vascular resistance. There was a predictable isoproterenol dose response for heart rate, LV dP/dt, and systemic vascular resistance. Although LV dP/dt increased at high-dose isoproterenol, the maximal response in heart failure was flatter, which suggests a limit in contractile reserves in heart failure. For systemic vascular resistance, the two curves converge with higher doses of isoproterenol; this suggests that there is a physiological limit to the decrease in systemic vascular resistance. The data reported here for isoproterenol dose response are similar to those previously reported by our laboratory in an investigation of the effects of inducible versus endothelial-mediated nitric oxide (NO) synthase in heart failure (17).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Isoproterenol-stimulated hemodynamic response in sham control and heart failure rats. Data are presented as the changes with each data point (sham control and congestive heart failure, CHF) normalized for the respective baselines. A: changes in heart rate. B: changes in left ventricular rate of pressure change (LV dP/dt). C: changes in systemic ventricular resistance (SVR). Each data point presented as mean ± SD; n = 8-11 in each group; *P < 0.05.

-AR control of arterial relaxation in vitro. There was a dose-dependent increase in arterial vasorelaxation in response to isoproterenol in both the sham and heart failure rats. Arteries from heart failure rats demonstrated markedly decreased arterial vasodilatation compared with sham control rats (Fig. 2). Removal of the endothelium decreased isoproterenol-induced vasorelaxation by 30% in sham rats but induced no change in heart failure rats.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Isoproterenol dose response in rat aorta from sham (n = 6) and CHF (n = 6) rats with endothelium intact (E) and endothelium removed (NoE). With the endothelium removed, there is decreased arterial vasorelaxation, which suggests that there is a -adrenergic receptor (-AR)-mediated component of endothelial-dependent vasorelaxation. The vasorelaxation response is decreased in heart failure and removal of the endothelium does not change the response. No differences were found in CHF with or without endothelium. Only one curve is shown for clarity. Each data point presented as mean ± SD; *P < 0.05.

Mechanisms of decreased -AR responsiveness. To elucidate the mechanisms of diminished -AR-stimulated vasorelaxation, vascular responses to forskolin (a direct adenylyl cyclase activator) and IBMX (a phosphodieasterase inhibitor) were measured in vitro. Heart failure resulted in a decrease in IBMX dose response with no change in the vasorelaxation to forskolin (Figs. 3 and 4).

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Forskolin dose response in aorta from sham (n = 6) and CHF (n = 6) rats. No differences were noted between sham and CHF responses. Each data point presented as mean ± SD.

View larger version (11K): [in this window] [in a new window]   Fig. 4.   IBMX dose response in aorta from sham (n = 6) and CHF (n = 6) rats. CHF rats exhibited decreased response to IBMX. Each data point presented as mean ± SD; *P < 0.05.

Alterations in -AR signaling components. In heart failure rats, the following changes were observed: decreases (P < 0.05; n = 5) in -AR density as shown in Fig. 5 (58.7 ± 6.0 vs. 35.7 ± 1.9 fmol/mg membrane protein for the aorta, and 29.6 ± 5.6 vs. 18.0 ± 3.9 fmol/mg membrane protein for the carotid artery) and increases (P < 0.05; n = 5) in G protein-coupled receptor kinase (GRK) activity levels as shown in Fig. 6 (relative phosphorimage counts of 191 ± 39 vs. 259 ± 26 in the aorta and 115 ± 30 vs. 202 ± 7 in the carotid artery). It should be noted that these activities were measured in cytosolic extracts which are predominantly -ARK1 (9).

View larger version (11K): [in this window] [in a new window]   Fig. 5.   -AR density in aorta and carotid artery from sham (n = 5) and CHF (n = 5) rats; *P < 0.05. Note there are decreases in receptor density in both aorta and carotid artery in heart failure.

View larger version (10K): [in this window] [in a new window]   Fig. 6.   -AR-coupled kinase 1 (-ARK1) activity levels in aorta and carotid artery from sham (n = 6) and CHF rats (n = 5). Note the increased -ARK1 activity levels in CHF. Each data point presented as mean ± SD; *P < 0.05.

G protein, cGMP, and cAMP. As shown in Fig. 7, in heart failure rats we found a trend toward decrease (P = 0.054) in G_s in the aorta (9,502 ± 965 vs. 7,173 ± 544 arbitrary densitometry units; n = 9) but no change in G_i in the aorta (11,521 ± 1,636 vs. 14,015 ± 1,840 arbitrary densitometry units; n = 9). In heart failure rats there were also decreases (P < 0.05; n = 5) in cGMP and cAMP in the carotid artery (0.85 ± 0.10 vs. 0.31 ± 0.06 and 2.3 ± 0.3 vs. 1.2 ± 0.1 pmol/mg protein, respectively); see Fig. 8.

View larger version (11K): [in this window] [in a new window]   Fig. 7.   Content of inhibitory G protein subunit (Gi) and stimulatory G protein subunit (Gs) in both sham (solid bars, n = 9) and CHF (open bars, n = 9) rats. CHF rats showed a trend to decrease (P = 0.047) in the Gs level; however, there was no difference in Gi between the two groups. Each data point presented as mean ± SD.

View larger version (8K): [in this window] [in a new window]   Fig. 8.   cAMP and cGMP levels in carotid artery from sham and CHF (n = 6 in each group) rats. Note both cAMP and cGMP levels are decreased in carotid arteries from CHF rats. Each data point presented as mean ± SD; *P < 0.05.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The most important finding of this study is that heart failure after MI results in functional and molecular alterations in the vascular -AR signaling pathway. Interestingly the changes in the vasculature are similar to those that are known to occur in the myocardium: a depressed functional response to isoproterenol accompanied by decreases in -AR density, cGMP and cAMP levels, and upregulation of -ARK1. The fact that these alterations seen in the vasculature are similar to those previously reported in the myocardium suggests a common systemic neurohormonal mechanism.

Although previous workers have shown decreases in vascular -AR density and cAMP and cGMP-mediated vasorelaxation in heart failure (24, 26), our data examine the alterations in the vascular -AR signaling pathway and its physiological consequences. It is clear from our data that the primary defect results in a decrease in receptor density accompanied by decreases in the second messenger systems. The increase in -ARK1 implies increased phosphorylation of the agonist-occupied receptor, triggering desensitization in the vasculature similar to what has been postulated for the heart. Based on our work these alterations in vascular -AR signaling have specific physiological (hemodynamic) effects, i.e., blunted in vivo hemodynamic responses to isoproterenol and decreased in vitro vasorelaxation.

Chronic stimulation of -ARs causes "desensitization" and downregulation of cardiovascular -receptors, which may be the basis of their long-term harmful effects in patients with heart failure. Mechanisms underlying -AR desensitization (homologous and heterologous) have been elucidated in the heart. Homologous desensitization occurs in response to -AR stimulation and results in decreased responsiveness to -AR agonists such as isoproterenol but not to other agonists linked to adenylyl cyclase. Homologous desensitization is associated with phosphorylation of the agonist-occupied receptor by a cAMP-independent protein kinase, -ARK1, or other members of the GRK family. -ARK1 can phosphorylate agonist-occupied -ARs, triggering desensitization when responsiveness of adenylyl cyclase to other stimulatory responses is diminished (23). Previous work with -ARK1 has focused on the heart, where -ARK1 has been shown to be elevated in LV tissue from explanted hearts in patients with heart failure (33) as well as from the hearts of rats and rabbits after coronary artery ligation (27, 34). Because our data define alterations in -AR signaling in the vasculature in heart failure, the possibility is raised that action of -ARK1 is a possible mechanism for the desensitization of vascular -ARs and the decrease in vasorelaxation response to isoproterenol in heart failure.

A possible mechanism of decreased -AR density in the present study is the hypotension associated with heart failure. This mechanism is unlikely for the following reasons: 1) there are no data to support the contention that hypotension decreases -AR density in vascular smooth muscle; in contrast, there are data to show that hypotension produced by verapamil treatment of normotensive animals does not decrease but rather increases -AR density in the cardiac muscle (8); 2) hypertension in human and animal models is generally associated with decreased vascular -AR density (7, 10, 15); and 3) the question of whether downregulation of vascular -AR density in hypertension might be related nonspecifically to blood pressure level has been addressed (15). In that study, treatment with verapamil or hydrochlorothiazide lowered blood pressure without any change in -AR number. The study concluded that vascular -AR function is selectively impaired in hypertension independent of blood pressure (15).

Our data support the concept that there is a common "catecholamine trigger" in heart failure that is responsible for the accompanied alterations in -AR signaling. If the enhanced sympathetic tone in heart failure is responsible for the changes in -AR signaling, we would expect to see them present globally, i.e., in the vasculature as reported in the present study as well as in the RVs and LVs as has been reported in the rabbit coronary artery ligation model of heart failure (27). Other evidence supporting this is that chronic treatment of mice with isoproterenol results in enhanced expression of -ARK1 in the heart (22), and in transgenic mice overexpression of -ARK1 in the heart results in cardiac dysfunction (25).

This common neurohormonal pathway raises the interesting possibility that treatment in heart failure previously focused on the myocardium may also affect the vasculature. For example, to date investigators have not been able to clearly define why -AR blockade improves survival in patients with heart failure. One explanation of our inability to determine the mechanisms of action of -AR blockade may be that attention had been focused at the heart and not the vasculature. In view of our data, it would be appropriate to examine the vascular -AR signaling cascade after treatment with -blockers. Another experimental approach based on our data would be to use -ARK inhibition as a possible tool to rescue vascular -ARs in heart failure. This concept is based on previous work in the heart where -ARKct, a -ARK1 inhibitor, was able to rescue the disrupted muscle LIM protein gene (MLP) heart failure mouse (30). That data showed that by targeting myocardial -ARK1 in heart failure in mice, that phenotype could be rescued with -ARK inhibition. The parallel between the heart and the vasculature is obvious. Because we have shown that the -ARs in the vasculature are also desensitized and -ARK activity is increased, the vasculature could be an important target in heart failure because -ARK inhibition should lead to vasodilatation and potentially a decreased LV afterload.

Although -ARs are primarily located in vascular smooth muscle, there are -ARs on endothelial cells (32). It is not clear whether vascular endothelial -ARs participate in the -AR-mediated vasodilatation. Previous work indicates that activation of these receptors stimulates release of NO, a major component of endothelial-dependent vasodilatation (20). Both endothelium-independent (11) and endothelium-dependent (2) -AR-mediated vasorelaxation have been reported. Our data describing the responses with removed endothelium show that the vascular relaxation response to isoproterenol is reduced but not abolished, suggesting that both endothelial and vascular smooth muscle -ARs are involved in heart failure (Fig. 2). Support that endothelial release of NO contributes to -AR-mediated vasodilatation comes from human studies where forearm blood flow during isoproterenol and salbutamol infusions was attenuated with N^G-monomethyl-L-arginine (6, 12) and -AR stimulation increased NO release in endothelial cells from patients in heart failure (13).

The physiological coupling between NO and -AR pathways and how this alters in vivo hemodynamics in heart failure was recently reported by our laboratory (19). The interaction between these two pathways in human forearm vasculature was studied previously (12); however, the interaction between diminished NO and downregulation of vascular -AR-mediated relaxation in heart failure is still under investigation. At the cellular level the interaction between NO release and altered -AR signaling in the vasculature in heart failure is unclear.

In summary, our data clearly demonstrate that there are abnormalities in vascular -AR signaling in heart failure that directly alter hemodynamics. The fact that these changes in the vasculature are similar to those in the heart supports the concept that there is a common systemic neurohormonal pathway in heart failure that may be in part responsible for the progressive nature of the disease. Although -AR blockade has been found to improve survival in patients with heart failure, the mechanisms responsible for this benefit are not clear. It is possible that some of the beneficial effects of blocking -AR activation in heart failure are due to the effects of these agents on the vasculature.