INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

ALTHOUGH A RECENT preliminary study has suggested that growth hormone (GH) may improve hemodynamics in patients with idiopathic dilated cardiomyopathy (1), it remains unclear whether GH plays an important role in the experimental setting of heart failure. Several studies have reported that GH or insulin-like growth factor 1 (IGF-1) treatment enhances left ventricular (LV) performance in rats with myocardial dysfunction induced by myocardial infarction (4, 5, 22). However, the data reported in those previous studies, particularly regarding the changes in myocardial contractility, LV end-diastolic pressure, ejection fraction, and cardiac index, were inconsistent among the studies (4, 5, 22). Also, our recent study using hypophysectomized and intact rats with moderate-to-large myocardial infarcts demonstrated that neither GH replacement nor excess GH treatment significantly affected myocardial function (20). Because of the limitations of the rat infarction model, including the lack of direct measurement of cardiac and systemic hemodynamics during the progression of myocardial dysfunction and the significant influence of anesthesia or recent surgery on hemodynamic measurements, it is difficult to reconcile the contradictory findings.

Accordingly, the primary goal of the present investigation was to use a rapid ventricular pacing-induced heart failure model in chronically instrumented, conscious dogs to determine whether chronic GH treatment affected resting cardiac and systemic vascular function during the development of heart failure and when severe heart failure was manifested. A second goal was to determine whether GH treatment could prevent the diminished LV contractile reserve, as assessed by inotropic response to -adrenergic receptor stimulation, that occurs during heart failure (10, 14, 15). The final goal of the study was to determine whether GH treatment affected myocardial morphological changes or renal function during the late stages of heart failure.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animal preparation. Fourteen adult mongrel dogs, weighing 18.4 ± 0.4 kg, were anesthetized with thiopental sodium (12-15 mg/kg iv). After tracheal intubation and ventilation with a ventilator (North American Dräger, Telford, PA), isoflurane anesthesia (1.0-2.0 vol% in oxygen) was maintained during surgery. With sterile technique, a left thoracotomy was performed at the fifth intercostal space. Tygon catheters (Norton Plastics, Akron, OH) were implanted in the descending aorta and left atrium for measurement of their respective pressures. A solid-state miniature pressure gauge (Konigsberg Instruments, Pasadena, CA) was implanted in the LV cavity through the apex for high-fidelity measurements of LV pressure and rate of change of LV pressure (LV dP/dt). A flow probe (Transonic Systems, Ithaca, NY) was placed on the ascending aorta to measure aortic blood flow. One pair of piezoelectric ultrasonic dimension crystals was implanted on opposing anterior and posterior endocardial surfaces of the LV to measure LV internal diameter. Proper alignment of the crystals was achieved during surgical implantation by positioning the crystals so as to obtain a signal with the greatest amplitude and shortest transit time. A pacing lead (Medtronic, Minneapolis, MN) was attached to the right ventricular free wall. Additionally, stainless steel pacing leads were attached to the left atrial appendage. The pericardium was left open. Catheters and leads were externalized between the scapulae, and the thoracotomy was closed in layers. An additional two dogs not surgically instrumented or subjected to rapid pacing-induced heart failure served as controls for histological studies. The dogs used in this study were maintained in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals [DHHS Publication No. (NIH) 85-23, Revised 1985], and the studies were approved by the Merck Research Laboratories (West Point, PA) Institutional Animal Care and Use Committee.

Physiological studies. Hemodynamic recordings were made using a data tape recorder (TEAC, Montebello, CA) and a multiple-channel oscillograph (Gould, Cleveland, OH). Aortic and left atrial pressures were measured using strain-gauge manometers (Argon, Athens, TX), which were previously calibrated using a mercury manometer, connected to the fluid-filled catheters. The solid-state LV pressure gauge was cross-calibrated with aortic and left atrial pressure measurements. LV dP/dt was obtained by electronically differentiating the LV pressure signal with a frequency response of 700 Hz. A triangular wave signal was substituted for the pressure signals to directly calibrate the differentiator (Triton Technology, San Diego, CA). Aortic blood flow was measured using a volume flowmeter (Transonic Systems). Mean arterial pressure, left atrial pressure, and ascending aortic blood flow (cardiac output) were measured using an amplifier filter. Stroke volume was calculated as the quotient of cardiac output and heart rate. Cardiac output was normalized by body weight to yield cardiac index. LV dimension was measured with an ultrasonic transit-time dimension gauge (Triton Technology). Total peripheral resistance was calculated as the quotient of mean arterial pressure and cardiac output. LV end-diastolic dimension (EDD) was measured at the beginning of the upstroke of the LV dP/dt signal. LV end-systolic dimension (ESD) was defined as the point of maximum negative dP/dt. The percent shortening of LV internal diameter was calculated as [(EDD  ESD)/EDD] × 100. Mean velocity of LV circumferential fiber shortening corrected for heart rate (V_cfc) was calculated as [(EDD  ESD)/EDD]/(ET/), where ET and R-R denote ejection time and R-R interval (in s), respectively. Ejection time was measured as the interval between maximum and minimum LV dP/dt. The LV isovolumetric relaxation time constant () was calculated beat by beat, on-line, from the minimum value of the time derivative of the LV pressure signal (LV dP/dt_max) to 36% of LV dP/dt_max (Modular Instruments, Malvern, PA) (12). A cardiotachometer triggered by the LV pressure pulse provided instantaneous and continuous records of heart rate.

1

1

1

1

Biochemical and morphological analysis. Plasma and urine electrolyte concentrations were measured by ion-selective electrode methodology (Beckman Synchro Elise). Urine protein concentration was measured by the Coomassie blue technique. Blood urea nitrogen (BUN) levels were measured with an automated blood analyzer (Gem Profiler: Schiapparelli Biosystems, Fairfield, NJ). Plasma ANP levels were measured by standard radioimmunoassay (American Laboratory Products, Windham, NH). Plasma IGF-1 levels were assessed by a previously described method (3). A total of eight hearts (3 vehicle-treated failure dogs, 3 GH-treated failure dogs, and 2 normal dogs) were perfused with saline followed by a 4% formaldehyde-2% glutaraldehyde solution. The fixed tissues were processed for light and transmission electron microscopic evaluation.

Data analysis. Data before and after development of heart failure, and responses to inotropic challenge were compared using the Student's t-test for paired data with a Bonferroni correction. Data between the GH-treated and vehicle-treated groups were compared using unpaired Student's t-test. Analysis by two variable linear regression and by multiple linear regression was used to compare plasma ANP with left atrial pressure and LV V_cfc. All values are expressed as means ± SE. Statistical significance was accepted at the P < 0.05 level.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Effects of GH on body weight and plasma level of IGF-1. The baseline values of body weight, i.e., before the initiation of treatment and rapid ventricular pacing, were 18.7 ± 0.6 and 18.0 ± 0.6 kg in the vehicle-treated and GH-treated groups, respectively. The changes in body weight after 2, 3, and 4 wk of pacing were significantly (P < 0.05) greater in the GH-treated group than in the vehicle-treated group (Fig. 1). After 4 wk of pacing, the body weight significantly (P < 0.05) increased to 21.1 ± 0.7 kg in the GH-treated group, whereas in the vehicle-treated group, body weight was 19.2 ± 1.1 kg. Because body weight can be affected by ascites at the late stage of heart failure, body weight also was measured after removing the abdominal fluid at the final experiment, i.e., after 4 wk of pacing. Under this condition, body weight still was significantly (P < 0.01) increased by 2.0 ± 0.5 from 20.0 ± 0.6 kg in the GH-treated group and was significantly (P < 0.05) greater than the vehicle-treated group, in which the body weight was decreased by 0.3 ± 1.1 from 18.4 ± 1.0 kg.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   A: plasma levels of insulin-like growth factor 1 (IGF-1) at baseline (C) and after 1-4 wk of rapid ventricular pacing in growth hormone (GH)- and vehicle-treated dogs. B: effects of GH on body weight in conscious dogs during the development of heart failure. Data are changes from baseline (i.e., before initiation of treatment and rapid ventricular pacing).

Effects of GH on basal hemodynamics before and after heart failure development. Figure 2 shows representative waveforms from vehicle-treated and GH-treated conscious dogs before and after the development of heart failure. Note that in the vehicle-treated dog, LV dP/dt was decreased, whereas mean left atrial pressure, LV dimension, and heart rate were increased after heart failure (Fig. 2, left). However, similar changes in these parameters also were observed in the GH-treated dog (Fig. 2, right).

View larger version (108K): [in this window] [in a new window]   Fig. 2.   Representative waveforms of left ventricular (LV) pressure, rate of change of LV pressure (LV dP/dt), aortic pressure, left atrial pressure, LV short axis diameter, and heart rate from a conscious dog treated with vehicle (left) and from a conscious dog treated with GH (right) before and after heart failure. Note that after heart failure, LV dP/dt decreased, whereas left atrial pressure, LV diameter, and heart rate markedly increased in both dogs. bpm, Beats/min.

1

1

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Effects of GH on resting systemic hemodynamics in conscious dogs during the development of heart failure. Values are %changes from baseline (C) levels. After 4 wk of rapid ventricular pacing, left atrial pressure, total peripheral resistance, and heart rate increased, whereas cardiac index and stroke volume index decreased. There were no differences in any of these parameters between the GH- and vehicle-treated groups.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Effects of GH on resting LV function in conscious dogs during the development of heart failure. Values are %changes from baseline (C) levels. After 4 wk of rapid ventricular pacing, LV systolic pressure, LV dP/dt, LV fractional shortening, and LV velocity of circumferential fiber shortening corrected for heart rate (Vcfc) decreased, whereas LV end-diastolic and end-systolic diameters increased. There were no differences in any of these parameters between the GH- and vehicle-treated groups.

Effects of GH on inotropic response to -adrenergic receptor challenge. The basal hemodynamics and response to isoproterenol (0.2 µg · kg¹ · min¹) before and after 2 and 4 wk of pacing in the GH- and vehicle-treated groups are shown in Table 3. After 2 and 4 wk of pacing, LV dP/dt and heart rate responses to isoproterenol were markedly attenuated compared with control, i.e., before heart failure. However, the changes in LV systolic pressure, LV dP/dt, mean arterial pressure, mean atrial pressure, LV end-diastolic diameter, and heart rate induced by isoproterenol were similar in the two treatment groups. Figure 5 shows the dose-response effects of isoproterenol on LV dP/dt after 3 wk of pacing in the GH- and vehicle-treated groups. Clearly, isoproterenol elicited similar inotropic responses in these two groups at each dose.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Dose-dependent effects of isoproterenol on LV dP/dt in GH- and vehicle-treated conscious dogs at baseline (control) and after 3 wk of pacing. Values are %changes from baseline levels. After heart failure, LV dP/dt responses to isoproterenol were attenuated compared with control. However, there was no difference between these two groups.

1

1

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Dose-dependent effects of dobutamine on LV dP/dt in GH-treated (A) and vehicle-treated conscious dogs (B) at baseline (control) and after 2, 3, and 4 wk of pacing. Values are %changes from baseline levels. After heart failure, LV dP/dt responses to dobutamine were attenuated compared with control. However, there was no difference between these 2 groups.

Effects of GH on plasma ANP and renal function. The relationships between the plasma level of ANP and left atrial pressure or LV V_cfc during the development of heart failure are shown in Fig. 7. The release of ANP correlated with the level of left atrial pressure and V_cfc. The correlation coefficient was significant (P < 0.01) for all groups studied. GH treatment did not significantly modify these relationships.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Relationships between plasma atrial natriuretic peptide (ANP) levels and left atrial (LA) pressure (A) and LV Vcfc (B) in conscious dogs before and during heart failure. Data are absolute values. Both LA pressure and LV Vcfc were correlated with plasma ANP level. However, slopes for the GH-treated dogs (dotted line) were similar to those for the vehicle-treated dogs (solid line).

Myocardial morphological evaluation. Light microscopic evaluation of hematoxylin- and eosin-stained sections of LV showed no apparent changes among heart failure dogs treated with GH or vehicle, as well as control dogs without heart failure. However, light microscopic evaluation of Epon-embedded, toluidine blue-stained sections revealed clear cytoplasmic vacuolation in myocardial fibers of both the GH- and vehicle-treated dogs with heart failure. Transmission electron microscopic evaluation indicated the presence of myocardial fibers with vacuoles filled with cytoplasmic debris sometimes with a myelin-like structure. In addition, the Z bands in myocardial fibers appeared slightly distorted in both the GH- and vehicle-treated dogs with heart failure. These changes were not observed in control dogs without heart failure. Overall, no morphological differences between the GH- and vehicle-treated heart failure dogs were observed. Figure 8 shows representative transmission electron micrograph of LV myocardium from a control dog without pacing-induced heart failure (Fig. 8A) and a GH-treated dog with heart failure (Fig. 8B). Note that myocardial fibers contained vacuoles filled with cytoplasmic debris and that the Z bands were distorted in the GH-treated dog compared with nonfailure control. These findings were similar to those observed in the vehicle-treated heart failure dogs and also were consistent with previous findings in this canine heart failure model (11).

View larger version (173K): [in this window] [in a new window]   View larger version (168K): [in this window] [in a new window]   Fig. 8.   Representative transmission electron micrographs of LV myocardium from a control dog without pacing-induced heart failure (A) and a GH-treated dog with heart failure (B). Note that myocardial fibers contain vacuoles filled with cytoplasmic debris and that the Z bands were distorted in the GH-treated dog compared with nonfailure control. These findings were similar to those observed in the vehicle-treated heart failure dogs.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The present study demonstrates that GH treatment at a dose that increases body weight and plasma IGF-1 level does not significantly improve cardiac and systemic function in conscious dogs with congestive heart failure induced by rapid ventricular pacing. This conclusion is based on several findings: 1) the impairment of resting LV systolic (i.e., both isovolumic and ejection phases) and diastolic function (i.e., isovolumic relaxation time constant ) were similar for the GH- and vehicle-treated groups during the development of heart failure; 2) LV contractile reserve, as assessed with -adrenergic receptor stimulation, was attenuated similarly after heart failure for both GH- and vehicle-treated groups; 3) GH treatment did not modify the relationships between plasma ANP release and left atrial pressure or LV V_cfc during the development of heart failure; and 4) peripheral vascular resistance, renal function, and myocardial morphology were also similar for GH-treated and vehicle-treated dogs with heart failure.

A major concern with our conclusion was whether the dose of GH used was sufficient to produce any impact on the heart. In the present study, the plasma IGF-1 level was 3.5-fold higher in the GH-treated group than in the vehicle-treated group. Also, body weight in the treated animals was increased by 17% compared with control. These data are comparable to those from a previous study, in which administration of IGF-1 was shown to significantly enhance LV hypertrophy in rats (4). In addition, our previous study also demonstrated a dose-related increase in plasma IGF-1 levels associated with increased body weight in normal dogs (16). For example, administration of GH at a dose of 0.1 IU · kg¹ · day¹ increased body weight almost as much as a dose of 1 IU · kg¹ · day¹ (0.56 mg · kg¹ · day¹), which was the dose used in the present study, further indicating that 1 IU · kg¹ · day¹ can be considered as a maximal dose. Therefore, the negative results of the current study are unlikely to be related to an insufficient dose of GH.

Although the mechanism of rapid ventricular pacing-induced heart failure utilized in the present study is still unclear and does not ideally mimic the process of chronic congestive heart failure that occurs in humans, it meets many of the criteria for heart failure (19, 21). Indeed, in the present study, we observed progressive decreases in LV dP/dt, stroke volume, LV fractional shortening, and V_cfc, while the isovolumetric , total peripheral resistance, left atrial pressure, and LV end-diastolic and end-systolic diameters were increased after multiple weeks of rapid ventricular pacing. At the final stage of pacing, altered myocardial ultrastructure also was observed. In addition, several prior studies have shown that rapid pacing-induced heart failure in dogs is characterized by a blunted inotropic response to catecholamine stimulation (14, 15), which is thought to be related to an impairment of -adrenergic receptor signal transduction pathways, including decreased -adrenergic receptor density, uncoupling of -adrenergic receptors, and a defect in the adenyl cyclase catalytic units (10, 14). In the present study, a marked attenuation of the isoproterenol-induced increase in LV dP/dt was observed during the development of heart failure. Because isoproterenol stimulates both ₁- and ₂-adrenergic receptors, which could indirectly affect LV dP/dt via loading condition and heart rate, we also examined the effects of a selective ₁-adrenergic receptor agonist, dobutamine, during the development of heart failure. Without significantly changing the loading condition or heart rate, dobutamine induced dose-dependent inotropic responses that were similar to those produced by isoproterenol. Under these circumstances, we did not detect any significant differences between the GH-treated and vehicle-treated dogs, suggesting that GH does not play a role in preserving cardiac function either at rest or during inotropic stimulation in heart failure.

An additional feature of the present study involved comparing the relationships between the plasma levels of ANP and left atrial pressure or V_cfc, an index reflecting LV systolic performance, during the progression of heart failure. It is well accepted that neurohumoral activation is evident in congestive heart failure (2, 7). ANP specifically has been shown to have the strongest correlation with atrial stretch and LV ejection fraction (2), because ANP is synthesized in the myocardium and released mainly in response to increased atrial tension (9, 18). Notably, several previous studies either in patients or in animal models demonstrated that increases in plasma ANP levels correlate with the severity of congestive heart failure (1, 8). In the current study, we also found strong correlations between the plasma levels of ANP and both the increases in left atrial pressure and decreases in V_cfc during the development of heart failure. It is conceivable that if the GH treatment had affected cardiac dynamics, the relationship between ANP and LV function would have been modified. However, no differences in these relationships were observed between the GH- and vehicle-treated groups. Additionally, we did not find any significant differences between the GH-treated and vehicle-treated heart failure dogs. However, our findings do not exclude the possibility that GH affects myocardial remodeling during the development of congestive heart failure, because the pacing-induced heart failure model used in the current study does not exhibit significant myocardial remodeling.

Our findings in the current study appear to conflict with those reported by other investigators using the myocardial infarction-induced LV dysfunction rat model (4, 5, 22) but are consistent with our prior study demonstrating that neither GH replacement in hypophysectomized rats nor excess GH treatment in intact rats with myocardial infarction improved LV function (20). Although it is difficult to reconcile the controversies due to the major differences among the species, heart failure models, and methods used, it is noteworthy that significant inconsistencies in hemodynamic effects with GH were evident in previous studies utilizing the rat myocardial infarction model (4, 5, 22). Yang et al. (22) reported that treatment with GH resulted in a marked increase in myocardial contractility, as reflected by an increase in LV dP/dt and decrease in LV end-diastolic pressure in rats with myocardial infarction. However, Duerr et al. (5) found no difference in LV dP/dt or LV end-diastolic pressure between IGF-1 combined with GH-treated and vehicle-treated rats with myocardial infarction. In addition, an earlier study by Duerr et al. (4) reported significant relationships among LV ejection fraction, infarct size, and treatment, with a trend for ejection fraction to be higher in treated rats with larger infarcts, which led them to conclude that IGF-1 enhances LV function in heart failure. More recently, Duerr et al. (5) reported that there was only a significant interaction for cardiac index, but not for ejection fraction, between treatment and infarct size. A preliminary study in patients with dilated cardiomyopathy reported that GH treatment increased myocardial mass and improved cardiac hemodynamics (6). Because the treatment protocols in the preceding clinical study and in the present experimental study differ significantly, it is difficult to compare our data with those findings. In addition, the preceding clinical report was based on experience with seven treated patients without a control group (6). The ultimate therapeutic utility of GH treatment in specific subsets of heart failure patients should be addressed by more rigorous clinical trials (13).

To explore the potential effects of GH on other aspects of congestive heart failure, we examined renal function by measuring plasma and urine samples. The results indicate that there was a similar tendency for GFR and BUN to increase compared with baseline, i.e., before heart failure, for the GH-treated and vehicle-treated groups. However, the changes were not statistically significant, indicating that despite significant ventricular dysfunction, renal function was preserved in this model, a finding similar to data published previously (17).

In summary, GH treatment at a dose that increases body weight and plasma IGF-1 level does not significantly affect cardiac or systemic vascular dynamics during the progression of congestive heart failure in conscious dogs. However, it should be noted that because the pacing induced-heart failure model used in the current study yields severe failure within a relatively short period of time and the underlying mechanism of this model may also differ from human congestive heart failure, the potential beneficial effects of GH treatment, particularly on a chronic basis, in patients with heart failure cannot be excluded.