INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CURRENT THERAPEUTIC strategies for chronic heart failure (HF) are centered on revascularization, afterload reduction, inotropic therapy, sympathetic blockade, and cardiac transplantation. Each of these approaches is limited in their overall efficacy, and as the number of people afflicted with HF continues to increase (27), more effective therapies are indicated. A sound approach for the elucidation of improved and novel therapeutic regimens for the treatment of chronic HF is to gain a broader understanding of the molecular mechanisms involved in the pathogenesis of HF. Recently, knowledge of the pathophysiology of HF has greatly increased with the study of cardiovascular changes associated with the neurohormonal activation observed in the disease (23). The findings of increased activity of the renin-angiotensin and sympathetic nervous system have led to the use of drugs such as angiotensin-converting enzyme inhibitors and -adrenergic antagonists, both of which have recently been demonstrated in large trials to impart dramatic and additive survival benefits (16, 24).

The successful use of -blockers in the treatment of HF underscores the importance of the myocardial -adrenergic receptor (-AR) system, which has long been at the forefront of conventional HF therapies. Signaling through -ARs located on the sarcolemmal membrane of cardiomyocytes plays a critical role in heart function, especially in disease states such as HF. In human HF, regardless of the cause, there is a constellation of molecular alterations that occur rendering myocardial -AR signaling defective, which undoubtedly exacerbates myocardial dysfunction (2, 6). Reduction of cardiac -AR density in the failing human heart was first observed in 1982 by Bristow et al. (3), who demonstrated the loss of specific radioligand-binding sites in failing human hearts removed at the time of transplantation. In addition, the remaining receptors appeared desensitized (3). Both ₁- and ₂-ARs are present in mammalian myocardium with the ₁-AR being the most abundant (2, 6). The understanding of the decreased density of -ARs has been enhanced by molecular studies, which show the mRNA for the ₁-AR to be significantly reduced (5, 25). ₂-AR mRNA is not altered in human HF (5, 25), but these receptors also appear desensitized (4). Interestingly, the levels of the -AR kinase (-ARK1) have been shown to be significantly elevated in human HF (25). -ARK1 is a member of the G protein-coupled receptor kinase (GRK) family that can phosphorylate agonist-occupied -ARs, triggering the process of desensitization (12). The fact that -ARK1 is elevated in human HF represents a potential mechanism for the loss of -AR responsiveness seen in HF. Another potential contributing factor to decreased -AR signaling in HF is increased levels of the adenylyl cyclase inhibitory G protein -subunit G_i (9).

Advanced understanding of the molecular changes that are associated with HF (i.e., what occurs in the -AR system) is critical to the development of more efficacious treatments for the failing human heart. Of particular importance is the development of reliable animal models of HF that recapitulate the molecular changes that accompany the disease in humans. Moreover, understanding the molecular and physiological milieu of the failing myocyte in the laboratory will help in the elucidation of novel approaches to HF treatment. The purpose of this study is to define the molecular changes of the -AR system associated with a rabbit infarct model of HF. Our hypothesis is that the molecular abnormalities observed in the failing human heart can be mirrored in an established animal model of HF.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals and induction of myocardial infarction. Animals used in this study were adult male New Zealand White rabbits (wt 3-5 kg). Animals were housed under standard conditions and were fed ad libitum. The Animal Care and Use Committee of Duke University Medical Center approved all procedures performed in accordance with the regulations adopted by the National Institutes of Health. To surgically produce myocardial infarction (MI), rabbits underwent left circumflex coronary artery (LCX) ligation or a thoracotomy only (sham operation). The procedure of LCX ligation and subsequent MI was adapted from a previously published model, which produces reliable ventricular remodeling in rabbits (15). Rabbits were pretreated with an intramuscular injection of 500,000 U penicillin, anesthetized with a mixture of ketamine (30 mg/kg) and xylazine (2 mg/kg), intubated, and then mechanically ventilated. We performed a left thoracotomy through the third or fourth intercostal space, and the large marginal branch of the LCX was identified and ligated with a 5-0 Prolene suture. We encircled the LCX in sham animals and withdrew the suture without ligating the artery. Anatomic closure was performed, the chest was evacuated of residual air using a 14-gauge angiocatheter attached to a syringe, and the rabbit was extubated when it was able to breathe spontaneously. Animals were allowed to recover and then returned to their cages when they were awake and responsive. MI size was determined as a percentage of the left ventricular (LV) free wall surface area. Areas of the LV free wall that were grossly pale, fibrotic, and thinned were considered to be infarcted. The hearts were removed and rinsed; and the atria, great vessels, and valves were trimmed away. The LV was cut away, and opened flat, and a paper tracing of the LV was made with the infarcted area marked. The paper tracing was then cut out and weighed to the nearest centigram. The tracing of the infarcted area was cut out and weighed separately. The ratio of the weights was used to estimate the percentage of LV infarcted.

In vivo hemodynamic measurements. To measure in vivo cardiac hemodynamic data, rabbits were sedated with ketamine (30 mg/kg) and acepromazine (0.5-1.0 mg/kg), and then a small incision was made in the neck to expose the right carotid artery. A 2.5-Fr micropressure transducer (Millar Instruments) was zeroed to atmospheric pressure and passed into the right carotid artery and down into the LV cavity to record pressures and heart rate. Confirmation of the position of the catheter was determined by fluoroscopy as well as by the shape of the pressure waveform. Data acquisition was recorded on a PC-based system, sampled at 200 Hz, and analyzed using custom software as previously described (1). Hemodynamic pressures were determined for all rabbits before their initial surgery and again 3 wk after LCX-ligation or sham operation to determine pre- and post-MI values.

Echocardiographic measurements. All echocardiographic measurements were performed after animals were sedated as described earlier in METHODS. A commercially available cardiovascular ultrasound system (Hewlett-Packard SONOS 1500) with a 7.5-MHz pediatric transducer was used for all studies. Data were recorded on 0.5-in. S-VHS videotape. Off-line measurements were performed using the system software. Two-dimensional echocardiographic views of the midventricular short axis were obtained at the level of the papillary muscle tips below the mitral valve. A minimum of 200 beats were recorded during each study and analyzed by two independent observers who were blinded to the experimental protocol. M-mode measurements of LV internal dimensions (LVID) were determined at the plane bisecting the papillary muscles according to the American Society of Echocardiography leading edge method on five heartbeats chosen at random by each observer. Values for LVID were obtained by averaging the 10 measurements taken by both observers. We calculated the percentage of fractional shortening as (LVID_d  LVID_s)/LVID_d × 100%, where subscripts d and s represent diastole and systole, respectively.

Myocyte isolation and culture. Study animals were anesthetized as mentioned earlier, heparinized (2,000 U), and intubated. After opening the chests, we rapidly excised the rabbit hearts, which were then rinsed in normal saline and perfused by the Langendorff technique as previously described (8) with Joklik's modified minimum essential medium containing hyaluronidase, collagenase, bacterial protease, and 12.5 µM CaCl₂. When the ventricles were noticeably soft to the touch, they were dissected free, and gentle agitation and filtration was used to obtain cells. Yield from this procedure typically approached 1-2 × 10⁷ myocytes per rabbit heart with 50-80% in rod-shaped morphology (8). Myocytes were plated at a density of 1 × 10⁵/35-mm well on tissue culture plates that were precoated with 20 µg/ml of mouse laminin.

Intracellular cAMP assay. Cells were labeled overnight in 3.0 µCi/ml (1 Ci = 37 GBq) [³H]adenine (DuPont/NEN) in medium 199 and then preincubated in minimal essential medium with 10 mM HEPES and 1 mM 3-isobutyl-1-methylxanthine for 30 min as described previously (8). Cells were then incubated in varying concentrations of isoproterenol (10⁴-10⁸ M) for 15 min. After incubation, the medium was aspirated, 1 ml of ice-cold stop solution (2.5% perchloric acid · 100 µM cAMP¹ · 10,000 cpm ¹⁴C¹) was added to each well, cAMP production was determined by anion-exchange chromatography, and the percentage of incorporation of the total ³H uptake was calculated as previously described (8).

Radioligand binding. Myocardial membranes were prepared by homogenization of excised hearts in ice-cold lysis buffer [5 mM Tris · HCl (pH 7.4) and 5 mM EDTA] as described previously (14). Care was taken to avoid use of infarcted heart tissue in MI animals. Final purified cardiac membranes were suspended at a concentration of 1-2 mg/ml in ice-cold -AR binding buffer [75 mM Tris · HCl (pH 7.4), 12.5 mM MgCl₂, and 2 mM EDTA], and receptor binding was performed as previously described using the nonselective ¹²⁵I-labeled -AR ligand cyanopindolol (¹²⁵I-CYP) (14). Nonspecific binding was determined in the presence of 20 µM of alprenolol. Reactions were conducted in 500 µl of binding buffer at 37°C for 1 h and then terminated by suction through glass-fiber filters. All assays were performed in triplicate, and -AR density (in fmol) was normalized to milligrams of membrane protein.

Membrane adenylyl cyclase activity. Myocardial membranes were prepared as described above. Membranes (20-30 µg of protein) were incubated for 15 min at 37°C with [-³²P]ATP under basal conditions or in the presence of either 100 µM isoproterenol or 10 mM NaF, and cAMP was quantitated by standard methods as we have described (14).

Protein immunoblotting. Immunodetection of myocardial levels of -ARK1 was performed on cardiac cytosolic or membrane-protein extracts after immunoprecipitation as previously described (7, 11). Excised hearts were homogenized in ice-cold lysis buffer [25 mM Tris · HCl (pH 7.5), 5 mM EDTA, 5 mM EGTA, 10 µg/ml leupeptin, 20 µg/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride (PMSF)], and the soluble and particulate fractions were separated by centrifugation. Separated fractions were then solubolized in ice-cold radioimmunoprecipitation buffer [50 mM Tris · HCl (pH 8.0), 5 mM EDTA, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 10 mM NaF, 5 mM EGTA, 10 mM sodium pyrophosphate, and 1 mM PMSF]. -ARK1 was immunoprecipitated from 1 ml of clarified extract (equal protein amounts) using a monoclonal anti--ARK1 antibody (1:2,000; Refs. 7, 11) and 35 µl of a 50% slurry of protein A-agarose conjugate agitated for 1 h at 4°C. After extensive washing, immune complexes were electrophoresed through 12% polyacrylamide Tris-glycine gels and transferred to nitrocellulose. The 80-kDa -ARK1 protein was visualized using standard chemiluminescence (ECL, Amersham). Immunodetection of G protein-coupled receptor kinase 5 (GRK5) was performed by Western blotting of myocardial membranes using a polyclonal anti-GRK5 (7, 11). For protein immunoblotting of G_i, membrane fractions were prepared as described above, and protein immunoblots were carried out using commercially available antibodies (Santa Cruz Biotechnology) as previously described (1). The G_i antibodies used were either specific for isoforms G_i-1 (Santa Cruz, I-20) or G_i-2 (Santa Cruz, T-19); additionally, a nonspecific antibody to G_i isoforms 1-3 (G_i-1-3) was also used (Santa Cruz, C-10). Quantitation of immunoreactive products was done by scanning the final autoradiography films and using ImageQuant software (Molecular Dynamics).

GRK activity assays. Myocardial extracts were prepared by homogenization of excised hearts or cultured myocytes in 2 ml of ice-cold lysis buffer [25 mM Tris · HCl (pH 7.5), 5 mM EDTA, 5 mM EGTA, 10 µg/ml leupeptin, 20 µg/ml aprotinin, and 1 mM PMSF] as previously described (7, 11, 14). Soluble cytosolic fractions and membrane fractions were separated by centrifugation, and GRK activity was assessed in 50-60 µg of membrane and 100-150 µg of cytosolic protein by light-dependent phosphorylation of rhodopsin-enriched rod outer segment membranes in lysis buffer with 10 mM MgCl₂ and 0.1 mM ATP (containing [-³²P]ATP) as we have described (7, 11, 14). After incubating over white light for 15 min at room temperature, reactions were quenched with ice-cold lysis buffer and centrifuged for 15 min. Inhibition of GRK activity in cytosolic extracts was performed by adding 2 µl of monoclonal -ARK1 antibody before light exposure as described previously (1, 7). Pelleted material was resuspended in 35 µl of protein gel loading dye and electrophoresed through 12% Tris-glycine gels. Phosphorylated rhodopsin was visualized by autoradiography of dried gels and quantified using a PhosphorImager (Molecular Dynamics).

Statistical analysis. Values are means ± SE. Hemodynamic values comparing pre- and post-MI measurements and signaling data from different groups of rabbit hearts and myocytes were evaluated using a Student's t-test. An analysis of variance (ANOVA) with repeated measurements with a grouping factor was performed on the isoproterenol dose-response data in myocytes. P < 0.05 was considered significant for all analyses.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Assessment of MI. MI rabbits, 21 days after LCX ligation, compared with sham-operated animals show clinical signs of congestive HF, including pulmonary congestion, hepatomegaly, pleural effusion, weight loss, and ascites. Infarct size was measured by LV dissection and determination of the percentage of the LV free wall affected by the infarct (see METHODS). Our method of LCX ligation procedure reproducibly results in LV infarctions of 30-40%. Overall mortality for infarcted rabbits was 27% at 3 wk after surgery, whereas there was no mortality in the sham group.

In vivo post-MI physiological function. MI after LCX ligation resulted in changes in hemodynamic measures of cardiac contractility after 3 wk as summarized in Table 1. Significant decrements of both LV +dP/dt_max (15%) and LV dP/dt_min (23%) as measurements of contractility and relaxation, respectively, were apparent in MI rabbits compared with sham-operated rabbits (Table 1). End-diastolic pressure in the LV was significantly elevated in rabbits after LCX ligation 3 wk after surgery compared with corresponding presurgical values (9 ± 3 vs. 0 ± 1 mmHg, P < 0.05) indicative of a failing heart. Systolic blood pressure and heart rate were not altered by MI (Table 1). No significant variations between preoperative and postoperative states were noted for any of the parameters in the sham-operated control group (Table 1).

-AR signaling characteristics after MI. Similar to human HF, post-MI rabbits showed a global biventricular downregulation of total -AR number after 3 wk (Table 3). Compared with sham animals, -AR density in post-MI rabbits was reduced by 44% in membranes purified from both the right ventricle (RV) and LV, demonstrating that the infarct present on the LV free wall causes global cardiac -AR changes. Competition binding isotherms with the ₂-AR-selective antagonist ICI-188,551 revealed that the decrease in -AR density is apparently due to the specific loss of ₁-ARs, as in human HF (5). The percentage of ₂-ARs found in control (sham) rabbits (23.4 ± 1.9%, n = 4) was significantly increased in MI rabbits (38.1 ± 2.5%, n = 4, P < 0.01, Student's t-test). This near doubling of ₂-AR percentage in failing rabbit hearts is consistent with ₁-ARs being solely responsible for the total -AR density loss (44%). Myocardial -AR functional coupling was assessed by measuring membrane adenylyl cyclase activity and, as seen in Table 3, significant uncoupling is present 3 wk after LCX ligation. Basal adenylyl cyclase activity is significantly depressed in membranes from both the RV and LV (32 and 43% decrease, respectively) compared with the activity in membranes purified from sham-operated control hearts. Furthermore, stimulation with the -agonist isoproterenol produced significantly less adenylyl cyclase activation in ventricular membranes from MI hearts compared with control hearts (Table 3). Maximal stimulation of adenylyl cyclase by NaF, which activates adenylyl cyclase in a postreceptor manner, was similar between treatment groups (Table 3), indicating no profound abnormalities in the myocardial adenylyl cyclase system.

View this table: [in this window] [in a new window]   Table 3.   Ventricular -AR density and membrane adenylyl cyclase activity

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Concentration-response curve of isoproterenol (Iso) in cultured ventricular cardiomyocytes isolated 3 wk after left circumflex coronary artery (LCX) ligation (myocardial infarction, MI) or sham operation (sham). Intracellular cAMP production is expressed as percent conversion of total 3H uptake into [3H]cAMP. Values are means ± SE for 5 separate preparations/group performed in triplicate. Iso-response curve in MI was significantly attenuated compared with sham-response curve.  P < 0.01 ANOVA; ** P < 0.01 vs. sham values (Student's t-test).

Myocardial GRK levels and activity after MI. Because the decrements of contractility, -AR density, and adenylyl cyclase activity in post-MI animals closely parallel the abnormalities associated with human HF, we sought to further understand the mechanisms that might be responsible for these alterations. The uncoupling of the -AR system in human HF is associated with elevations of the expression and activity of -ARK1 (25). To examine this phenomenon, we first assessed cytosolic GRK activity present in soluble extracts from ventricular tissue or cultured myocytes using an in vitro phosphorylation assay using rhodopsin as our model substrate. Cytosolic GRK activity in soluble RV and LV tissue extracts and in extracts from cultured ventricular myocytes from 3 wk-post-MI rabbit hearts was found to be significantly greater than the GRK activity present in extracts from sham-operated rabbit hearts (as shown in Fig. 2A). Increased GRK activity in MI samples was 2.5-fold higher than sham GRK activity (Fig. 2A). Choi et al. (7) have previously shown that soluble myocardial GRK activity is primarily caused by the actions of -ARK1. Consistent with this fact, we found that the enhanced cytosolic GRK activity in extracts from cultured MI hearts was completely eliminated when a monoclonal -ARK1 antibody known specifically to inhibit -ARK1 (7) was included in the reaction (Fig. 2A).

View larger version (43K): [in this window] [in a new window]   Fig. 2.   A: cytosolic G protein-coupled receptor kinse (GRK) activity in extracts isolated from ventricular tissues or cultured myocytes from sham or MI rabbits. Equal soluble protein amounts were used in an in vitro phosphorylation assay with rhodopsin-enriched rod outer segment membranes as the substrate. 32P incorporation (fmol · mg1 · min1) was quantified using a PhosphorImager, and data were expressed as percentage of 32P incorporation in sham samples. Representative experiment (inset) using right ventricular (RV) and left ventricular (LV) extracts done in duplicate shows phosphorylated rhodopsin. GRK activity of myocyte cytosolic extracts was determined in presence and absence of a -adrenergic receptor kinase (-ARK1) monoclonal antibody (mAb). Values are means ± SE for 4 hearts performed in duplicate for each group. * P < 0.05 vs. sham RV; ** P < 0.01 vs. sham LV;  P < 0.01 vs. sham myocytes (Student's t-test). B: ventricular -ARK1 levels in soluble cytosolic extracts. -ARK1 protein was determined by protein immunoblotting of immunoprecipitated myocardial extracts from RV and LV specimens (from MI and sham rabbits; 3 wk postsurgery) using a -ARK1 mAb. Values are means ± SE for 4 hearts/group, expressed as percent control (sham) densitometry units of scanned chemiluminescent immunoblots. Representative result (inset) from 2 MI and 2 sham animals. Purified -ARK1 was a control for protein migration. * P < 0.05 vs. sham (Student's t-test).

View larger version (34K): [in this window] [in a new window]   Fig. 3.   A: GRK activity of membrane extracts purified from cultured ventricular cardiomyocytes isolated 3 wk after MI or sham operation. Equal protein aliquots of these extracts were used in an in vitro phosphorylation assay using rhodopsin-enriched rod outer segment membranes as in Fig. 2A. Representative autoradiograph (inset) from a dried gel shows rhodopsin phosphorylation in myocyte membrane extracts from 2 MI and 2 sham animals. Values are means ± SE of 4 separate myocyte preparations. ** P < 0.01 vs. sham myocytes (Student's t-test). B: ventricular -ARK1 levels in membrane extracts from LV of MI and sham rabbit hearts. -ARK1 protein was visualized by protein immunoblotting of immunoprecipitated myocardial extracts as in Fig. 2B. Data are means ± SE of 6 hearts/group expressed as percent control (sham) densitometry units of scanned chemiluminescent immunoblots. * P < 0.05 vs. sham (Student's t-test).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   G protein-coupled receptor kinase 5 (GRK5) levels in myocardial membranes purified from ventricles of MI and sham rabbits and from extracts from cultured ventricular myocytes. GRK5 was identified by standard Western blot analysis using a polyclonal GRK5 antibody. Chemiluminescent blots were scanned, and densitometry of immunoreactive GRK5 was assessed. Representative Western blot (inset) of extracts from cultured myocytes isolated from 2 MI and 2 sham animals are shown. Purified GRK5 was included as a control for protein migration. Values are means ± SE of 3-4 hearts/group expressed as percent control (sham) densitometry units of scanned chemiluminescent immunoblots. P = not significant.

Myocardial G_i content in infarcted rabbit hearts. In the final biochemical characterization of the MI rabbit model, we examined the membrane content of G_i in LV tissue samples from MI and sham animals by protein immunoblotting (Fig. 5). Similar to findings in human HF (9), the level of G_i, determined with a nonspecific antibody to G_i isoforms 1-3, was found to be significantly increased 3.5-fold in post-MI rabbits compared with the levels found in sham-operated controls 3 wk after surgery. Protein immunoblotting with antibodies specific to G_i-1 and G_i-2 revealed that of these two isoforms only G_i-2 was increased (Fig. 5), which is consistent to what is seen in human HF. However, G_i-3 may also be increased (Fig. 5).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Gi protein levels in myocardial membranes purified from LV tissue samples from 8 MI and 8 sham rabbits (3 wk postsurgery). Gi was identified by standard Western blot analysis using a nonselective polyclonal antibody for Gi isoforms 1-3 (Gi-1-3, Santa Cruz Biotechnology). Values are means ± SE, expressed as percent control (sham) densitometry units of scanned chemiluminescent immunoblots using this antibody. Representative experiment (inset) shows Gi levels from 2 MI and 2 sham hearts using either the nonselective antibody or an antibody selective for Gi-2. Western blotting with a selective antibody for Gi-1 revealed no alterations of this isoform between sham and MI. ** P < 0.01 vs. sham (Student's t-test).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A novel finding of this study is that the ligation of the LCX of rabbits can produce both in vivo LV diastolic and systolic dysfunction. Furthermore, this animal model can recapitulate important biochemical alterations such as attenuated -AR signaling that are seen in human HF. The LCX ligation model of HF in rabbits as carried out in this study was also found to share some of the gross physiological characteristics of human HF at 3 wk post-MI, a relatively short time span. Previous work with this model has also demonstrated some of these physiological changes after 3 wk (15, 17). Pennock et al. (17) demonstrated echocardiographic and in vivo physiological parameters of cardiac dysfunction in MI rabbits similar to our own, including enlarged LVID_s and LVID_d, decrements in percentage of fractional shortening, and elevated LV end-diastolic pressure. Although several parameters such as enlargement of the atria and decreases in the pulmonary venous systolic-to-diastolic ratio were consistent with the severe diastolic abnormalities commonly seen in patients post-MI, Pennock et al. (17) were unable to demonstrate changes suggestive of classic LV systolic dysfunction, such as cardiac index or LV +dP/dt_max. Mahaffey et al. (15) also found no significant differences in LV +dP/dt_max or LV dP/dt_min between sham and MI rabbits (15). In contrast, in the present study, we demonstrated a 15 and 23% decrease in LV +dP/dt_max or LV dP/dt_min, respectively. This may be the result of an increase in infarct size as we typically see an LV infarct of ~35% compared with 24% in previous studies (15, 17). Our present data indicate that in combination with the detail of echocardiographic assessment available today, this rabbit HF model can provide an accurate recapitulation of the functional abnormalities found in human HF, and importantly, this model allows for the serial measurements of these parameters.

In addition to the functional indexes of HF, the rabbit LCX-ligation model also produced biochemical alterations that are present in human HF. Three weeks post-MI, rabbits exhibit a significant 44% decrease in global myocardial -AR density, which as in human HF (5, 25) was apparently due to a selective downregulation of ₁-ARs. Furthermore, animals also exhibit functional uncoupling of remaining receptors, consistent with what has been seen in the failing human heart (3, 4). An important aspect of this defective myocardial -AR system in the failing rabbit heart is our finding that -ARK1 levels and activity are selectively enhanced globally. The fact that -ARK1 is selectively increased in rabbit HF is in contrast to what has been observed in a porcine pacing-induced model of HF where GRK5 was shown to be increased (19). Specifically, GRK5 was not elevated in this model of HF, and importantly, the GRK-selective increases in the levels and activity of -ARK1 closely resemble what has been seen in human HF (25). Furthermore, a recent finding in a rat model (26), which showed that -ARK1 was enhanced after myocardial ischemia, increases the importance of -ARK1 in MI-induced HF. A final biochemical characteristic found in the MI rabbit model of HF and also found in human HF is elevated myocardial protein content of the adenylyl cyclase inhibitory G protein G_i, which can potentially contribute to -AR dysfunction (9). Our results with selective G_i antibodies revealed that G_i-2 and possibly G_i-3 were increased, whereas G_i-1 was not.

Importantly, -AR changes were seen in both the RV and LV in this rabbit model, indicating that biventricular failure is present after LCX ligation. Downregulation and functional uncoupling of myocardial -ARs in HF is hypothesized to result from enhanced sympathetic nervous system activity present in HF (23). If this is the trigger for such -AR changes in the failing heart, then one would expect global biochemical alterations as is the case in the MI rabbit model described here. Consistent with a "catecholamine trigger" hypothesis, Iaccarino et al. (11) recently discovered that chronic treatment of mice with the -agonist isoproterenol produces enhanced expression of -ARK1 in the heart (11). This increase in myocardial -ARK1 can lead to the desensitization of -ARs and uncoupling of the adenylyl cyclase system (11). Moreover, studies in transgenic mice with myocardial-targeted overexpression of -ARK1, at levels (three- to fivefold) seen in our rabbit model and human HF, have demonstrated that this alone can cause cardiac dysfunction and loss of inotropic reserve (14). Thus the increased -ARK1 seen in HF can be triggered by catecholamines (11), and this enhancement undoubtedly contributes to myocardial dysfunction and the pathology of HF.

Mounting evidence supports this critical role of -ARK1 in cardiac function. As in human and rabbit HF, -ARK1 was shown to be elevated in a genetic mouse model of cardiomyopathy and HF resulting from the disruption of the muscle LIM protein gene (21). Interestingly, Rockman et al. (21) recently found that targeted transgenic expression of a peptide inhibitor of -ARK1 to the hearts of these mice at birth prevents the development of cardiomyopathy and restores -AR hemodynamic responsiveness. Thus -ARK1 appears to be a novel target for the treatment of HF. Accordingly, it is important to have appropriate animal models of HF available in the laboratory to continue the pursuit of this as well as other potential therapeutic targets.

In summary, the rabbit LCX-ligation model of MI-induced HF appears to hold several key advantages for study. First, and foremost, is that it appears to closely resemble the clinical picture of human HF triggered by ischemic cardiomyopathy. Furthermore, unlike pacing-induced failure, this MI model produces stable and reproducible HF that is relatively inexpensive compared with larger chronic animal models (10). As many of the functional and biochemical characteristics of human HF are accurately reproduced by this small animal model at 3 wk post-MI, including -AR derangements and biventricular failure, additional studies in the rabbit could focus on the exact time course of events that lead to the eventual structural and functional changes within the heart. In addition, the straightforward physiological assessments that can be performed on this model invite the testing of potential molecular therapies (including gene therapy), such as targeting -AR density or -ARK1 inhibition via adenoviral-mediated in vivo myocardial delivery (18).