Effects of Cardiac Denervation on Development of Heart Failure and Catecholamine Desensitization
Background Two signatures of heart failure are activation of the sympathetic nervous system and catecholamine desensitization. However, whether or not the elimination of cardiac nerves affects either the progression of heart failure or catecholamine desensitization is not clear.
Methods and Results We studied 8 dogs with selective ventricular denervation (VD) (surgical technique) and 10 intact dogs, chronically instrumented for measurement of left ventricular (LV) and arterial pressures, LV dP/dt, LV internal diameter, and wall thickness before and after heart failure was induced by rapid pacing (240 bpm) for 3 to 4 weeks. VD was confirmed by the absence of reflex effects induced by intracardiac veratrine and depletion of tissue norepinephrine and by supersensitive responses to norepinephrine. During the development of heart failure, LV end-systolic and end-diastolic stresses and heart rate increased, while myocardial contractility, as reflected by LV dP/dt and mean velocity of circumferential fiber shortening corrected for heart rate (Vcfc), decreased in both intact and VD dogs. However, the increases in LV end-diastolic stress and decreases in LV dP/dt as well as the relationship between LV systolic stress and Vcfc in heart failure were less (P<.05) in VD dogs. The responses of LV dP/dt and heart rate to both isoproterenol and norepinephrine in intact dogs were reduced in heart failure. The physiological desensitization to the inotropic effects of isoproterenol and norepinephrine was less in dogs with VD (P<.05), but chronotropic responses were similar because atrial innervation remained intact. Plasma norepinephrine levels were not different in VD dogs (592±79 pg/mL) compared with intact dogs (576±81 pg/mL) in heart failure.
Conclusions Dogs with selective VD tolerated the development of heart failure better than intact dogs and demonstrated significantly less catecholamine desensitization. The latter indicates that intact ventricular innervation is required for physiological expression of catecholamine desensitization despite comparable elevation of plasma catecholamines during the development of heart failure.
It is well known that the sympathetic nervous system is activated and circulating catecholamine levels are increased in heart failure.1 2 3 4 5 However, whether or not these compensatory mechanisms affect the progression of heart failure is controversial. Whereas some studies have suggested that compensatory sympathetic nerve activity could play a supportive role in heart failure,6 7 other studies have suggested the opposite, that β-adrenergic receptor blockade is beneficial in heart failure.8 9 10 11 12 13
Catecholamine desensitization is another hallmark of heart failure.14 15 16 17 18 19 20 21 22 The trigger for catecholamine desensitization could be either increased circulating catecholamine levels,15 17 23 increased cardiac neural activity, or both. Both of these questions, ie, the role of cardiac nerves in the progression of heart failure and their role in catecholamine desensitization in heart failure, could be approached with the use of models of either surgical cardiac denervation or pharmacological therapy. The surgical technique does not have the disadvantage of complicating influences of chronic β-adrenergic receptor blockade therapy.24 Moreover, when a model of selective VD is used, heart rate control remains intact.
Accordingly, to achieve these goals, we compared the effects of pacing-induced heart failure in intact, conscious dogs and dogs with selective surgical VD. This study had four components: (1) documentation of completeness of VD, (2) documentation of denervation supersensitivity to the sympathetic neurotransmitter NE, (3) determination of whether dogs with VD fare better hemodynamically during the development of heart failure, and (4) determination of the mechanism of catecholamine desensitization in heart failure, ie, whether it involved cardiac nerves or circulating catecholamine levels.
Thirty adult mongrel dogs of either sex weighing 20 to 30 kg each were anesthetized with halothane (0.5 to 1.5 vol%) and ventilated with a Harvard respirator after induction with thiamylal sodium (10 to 15 mg/kg IV). A left thoracotomy was performed through the fourth intercostal space by use of a sterile technique. In 15 dogs, surgical VD was performed.25 A solid-state miniature pressure gauge (model 22, Königsberg Instruments, Inc) was implanted in the LV through the apex. Tygon catheters (Norton Elastic and Synthetic Division) were implanted in the descending thoracic aorta and left atrial appendage. Piezoelectric ultrasonic dimension crystals were implanted on opposing anterior and posterior endocardial surfaces of the LV to measure the LV internal diameter and on opposing endocardial and epicardial surfaces to measure wall thickness. A screw-in type of pacing lead was attached to the RV free wall, and left atrial pacing electrodes were implanted. Catheters and leads were externalized, and the thoracotomy was closed. The chest was evacuated. Each dog was treated with 1 g of cephalothin for 10 days after surgery. In 15 dogs, the same instrumentation was implanted without VD (intact group). All dogs tolerated the operation. Two intact dogs and 3 VD dogs were euthanatized due to infection or bleeding within the first postoperative month before pacing was initiated. Animals used in this study were maintained in accordance with guidelines of the Committee on Animals of the Harvard Medical School and the Guide for the Care and Use of Laboratory Animals (Department of Health and Human Services publication No. NIH 83-23, revised 1985).
Experiments were initiated 2 to 3 weeks after recovery from surgical instrumentation, when the dogs were healthy, ie, body temperature, blood cell count, and chemistry were within normal limits. Hemodynamic measurements were recorded with the dogs fully awake, lying quietly on their right side. The fluid-filled catheters in the aorta and left atrium were connected to strain-gauge manometers (Statham Instruments) for the measurement of arterial and atrial pressures. LV pressure was measured with a solid-state miniature pressure gauge and calibrated in vivo against the measurement of systolic arterial and left atrial pressures. LV wall thickness and internal diameter were measured with an ultrasonic transit-time dimension gauge. A cardiotachometer triggered by the pressure pulse provided instantaneous and continuous records of heart rate. The position of all catheters and crystals was confirmed after euthanasia.
Eighteen dogs (intact dogs, n=10; VD dogs, n=8) were studied in the control state and 1 day, 1 week, and 3 to 4 weeks after pacing, ie, as heart failure developed. An additional 7 dogs (intact dogs, n=3; VD dogs, n=4) were studied as controls over the same time period without inducing heart failure. There were four components involved in this study: (1) documentation of completeness of VD, (2) documentation of denervation supersensitivity, (3) determination of the effects of VD on the progression of heart failure, and (4) determination of the effects of VD on expression of catecholamine desensitization during the development of heart failure.
Tests of Adequacy of Denervation Procedure
1. Electrical stimulation of cardiac nerves at surgery: VD was confirmed by the elimination of both the ECG and positive inotropic responses of the left ansa subclavia and thoracic vagi and positive inotropic responses of the right ansa subclavia by electrical stimulation (10 to 20 Hz, 5.0 ms, 5 to 10 V).
2. Reflex stimulation of cardiac nerves in conscious dogs: In 10 intact and 8 VD dogs, the responses to nitroglycerin (5 μg/kg IV), phenylephrine (5 μg/kg IV), and intracardiac (left atrial) veratrine alkaloid (5 μg/kg), which stimulates ventricular nerves, were tested 2 to 3 weeks after surgery.
3. Measurement of cardiac tissue NE: After the physiological studies were completed, the animals were anesthetized with sodium pentobarbital (30 to 50 mg/kg IV), and at the time of euthanasia, LV tissue samples were removed immediately and placed in liquid nitrogen. Tissue NE levels were measured by a radioenzymatic assay.20
4. To determine whether the VD procedure was successful over the full time course of these experiments, both heart rate and mean arterial pressure responses to nitroglycerin, phenylephrine, and veratrine as well as inotropic responses to NE (0.05, 0.1, and 0.2 μg·kg−1·min−1 IV) were tested in three intact dogs and four VD dogs at 2, 3, and 8 to 9 weeks after surgical denervation.
Documentation of Denervation Supersensitivity
The 5-minute infusion of each dose of the sympathetic neurotransmitter NE (0.05, 0.1, and 0.2 μg·kg−1·min−1 IV) was examined in 10 intact and 8 VD dogs. On a separate day, 5-minute infusions of NE (0.05, 0.1, and 0.2 μg·kg−1·min−1 IV) were studied in the presence of ganglionic blockade with hexamethonium bromide (30 mg/kg IV) and methylatropine bromide (0.1 mg/kg IV) in 8 intact and 6 VD dogs. Absence of reflex heart rate changes in response to changes in arterial pressure induced by phenylephrine (5 μg/kg IV) and nitroglycerin (5 μg/kg IV) confirmed the adequacy of ganglionic blockade.
Effects of VD on the Progression of Heart Failure
Hemodynamic measurements were made in the control state in the absence and presence of ganglionic blockade, after which rapid ventricular pacing was initiated at a rate of 240 bpm and controlled by use of a programmable pacemaker (model EV4543, Pace Medical Inc), which was worn externally in a vest. Additionally, in controls, the relationship between LV end-systolic stress as an index of afterload and Vcfc was examined by use of infusions of phenylephrine (1, 2, and 5 μg·kg−1·min−1 IV) at a constant heart rate (150 bpm). Baseline hemodynamics were recorded subsequently 1 day, 1 week, and 3 to 4 weeks after pacing, when heart failure was manifest. Baseline hemodynamics were also recorded in the absence and presence of ganglionic blockade. All data were collected during normal sinus rhythm after a 30-minute stabilization period after cessation of pacing. Plasma catecholamine samples were taken in the control state and before the animals were killed and were measured by a radioenzymatic assay.20
Effects of VD on Expression of Catecholamine Desensitization During the Development of Heart Failure
Infusions of NE (0.05, 0.1, and 0.2 μg·kg−1·min−1 IV, each dose for 5 minutes) were repeated at 1 day, 1 week, and 3 to 4 weeks after induction of rapid pacing. The infusions of NE (0.05, 0.1, and 0.2 μg·kg−1·min−1 IV) after ganglionic blockade plus atropine were also repeated at 3 to 4 weeks after pacing. All data were collected in sinus rhythm with the pacemaker turned off except for selective experiments at a constant heart rate (180 bpm) before and after heart failure. Infusions of ISO (0.05, 0.1, 0.2, and 0.4 μg·kg−1·min−1 IV, each dose for 5 minutes) were also examined in both intact and VD dogs before and after heart failure in the presence and absence of ganglionic blockade, with and without heart rate held constant with atrial pacing (240 bpm). During ISO infusions in the presence of ganglionic blockade, mean arterial pressure was maintained at baseline levels by use of phenylephrine infusions.
Hemodynamic measurements were recorded on a multichannel tape recorder (Honeywell) and played back on a direct-writing oscillograph (Gould-Brush). LV, arterial, and left atrial pressures, LV internal diameter, and wall-thickness analog signals were digitized (500 Hz), and LV wall thickening, LV diameter, LV fractional shortening, Vcfc, and LV wall stress were calculated with the use of a computer-based (Data Acquisition) system (Notocord System).
Data were expressed as mean±SE and were collected from the same dogs before and after heart failure. Baseline hemodynamics before and after heart failure were tested by Student’s t test. Baseline hemodynamics before and 1 day, 1 week, and 3 to 4 weeks after the development of heart failure in the two groups and responses to NE and ISO before and after heart failure between both groups were tested statistically with a repeated measures ANOVA procedure of Super ANOVA (Abacus Concepts). Regression lines were compared by differences in both slopes and elevations of the lines by use of the F test. A value of P<.05 was considered indicative of a significant difference.
Confirmation of VD
The completeness of VD was confirmed in three different ways. At surgery, electrical stimulation of cardiac nerves (left ansa subclavia) increased LV dP/dt by 86±14% before VD and did not increase LV dP/dt after VD. After recovery from surgery, in the intact group, phenylephrine (5 μg/kg IV) increased mean arterial pressure by 34±4 mm Hg and decreased heart rate reflexly by 28±5 bpm, whereas nitroglycerin (5 μg/kg IV) decreased mean arterial pressure by 14±1 mm Hg and increased heart rate by 47±8 bpm, and veratrine decreased mean arterial pressure by 25±8 mm Hg and heart rate by 38±9 bpm. In contrast, in the VD dogs, the responses to phenylephrine and nitroglycerin were similar, confirming the integrity of sinoaortic reflexes and sinoatrial nodal innervation. However, the responses of heart rate and mean arterial pressure to veratrine were abolished (Fig 1⇓), confirming selective VD.
Finally, at the time of euthanasia, LV and RV NE levels were 510±97 and 566±130 pg/mg wet wt, respectively, in intact sham dogs, whereas in VD control shams, tissue NE levels were 1.7±0.1 and 1.8±0.2 pg/mg wet wt, respectively. After heart failure, LV and RV NE levels were significantly decreased in intact dogs (P<.05) compared with sham intact dogs, ie, 137±28 and 152±33 pg/mg wet wt, respectively. In VD dogs with heart failure, LV and RV NE levels were similar to those in sham VD dogs, ie, 2.4±0.3 and 3.1±0.9 pg/mg wet wt, respectively (Fig 2⇓).
Denervation Supersensitivity Before Heart Failure
In intact dogs, NE (0.2 μg·kg−1·min−1) increased LV dP/dt by 39±6% from 3098±237 mm Hg/s. In contrast, in VD dogs, the same dose of NE increased LV dP/dt significantly more, by 143±19% from 2404±56 mm Hg/s. Dose-response relationships confirmed these individual effects (Fig 3⇓, left panel).
In the presence of ganglionic blockade (Fig 3⇑, right panel), the response of LV dP/dt to NE was still enhanced in VD dogs compared with intact dogs, eg, during NE (0.2 μg·kg−1·min−1), LV dP/dt was increased more (by 255±35%, from 2080±133 mm Hg/s) than in intact dogs (169±21%, from 2140±146 mm Hg/s) (P<.05). Dose-response relationships confirmed these individual effects.
Confirming the persistence of VD, both heart rate and mean arterial pressure responses to phenylephrine, nitroglycerin, and veratrine were not changed 2, 3, and 8 to 9 weeks after surgery in three intact dogs and four VD dogs. Furthermore, responses to NE (0.2 μg·kg−1·min−1) also were not changed, ie, supersensitive responses for LV dP/dt persisted (148±26% at 2 weeks, 147±27% at 3 weeks, and 155±29% at 8 to 9 weeks after surgery) in the four VD dogs.
Effects of VD on Baseline Hemodynamics
Before heart failure, heart rate, LV systolic pressure, mean arterial pressure, LV end-diastolic pressure, fractional shortening, Vcfc, and end-systolic and end-diastolic wall stresses were not different between intact dogs (n=10) and VD dogs (n=8). However, LV dP/dt was slightly (P<.05) lower in VD dogs (2576±142 mm Hg/s) than in intact dogs (2991±163 mm Hg/s). (See Table 1⇓.)
In the presence of ganglionic blockade, there were no significant hemodynamic differences between intact and VD dogs.
Effects of VD on Responses to Heart Failure
In intact conscious dogs, heart failure increased heart rate (+29±4 bpm), LV end-diastolic pressure (+18.3±1.8 mm Hg), and end-systolic (+25.5±4.3 g/cm2) and end-diastolic (+41.1±4.5 g/cm2) wall stresses, whereas mean arterial pressure (−18±3 mm Hg), LV dP/dt (−1486±178 mm Hg/s), fractional shortening (−11.9±2.4%), and Vcfc (−0.59±0.12 s−1) decreased significantly. In VD dogs, the decreases in LV dP/dt (−950±170 mm Hg/s) and increases in LV end-diastolic pressure (+9.4±1.6 mm Hg/s) and wall stress (+23.8±4.0 g/cm2) were significantly attenuated (P<.05) as heart failure developed compared with responses in intact dogs (Fig 4⇓). Furthermore, the inverse relation between Vcfc and LV end-systolic wall stress was steeper in intact dogs as heart failure developed compared with responses to an acute increase in afterload induced with phenylephrine. Importantly, this relationship was better maintained in VD dogs than in intact dogs (Fig 5⇓). Heart rate changes were similar between intact and VD dogs.
In the presence of ganglionic blockade, the differences between intact and VD dogs appeared to be more pronounced, ie, in VD dogs, end-diastolic pressure (+8.7±1.4 mm Hg) and wall stress (+24.1±3.8 g/cm2) increased less during the development of heart failure than in intact dogs (+19.5±1.7 mm Hg and +49.0±4.2 g/cm2, respectively; P<.05). Similarly, LV dP/dt (−509±135 mm Hg/s), Vcfc (−0.13±0.08 s−1/2), and fractional shortening (−2.0±1.8%) tended to decrease to a lesser extent in VD dogs than they did in intact dogs as heart failure developed (−1106±163 mm Hg/s, −0.33±0.08 s−1/2, and −6.9±1.6%, respectively).
In intact dogs, plasma NE increased (to 576±81 from 223±18 pg/mL) and plasma EPI increased significantly (to 332±86 from 142±21 pg/mL; P<.05) during the development of heart failure. In VD dogs, baseline plasma NE and EPI levels were not different (NE, 219±43 pg/mL; EPI, 169±24 pg/mL), nor were the observed increases in plasma catecholamines (NE, 592±79 pg/mL; EPI, 344±60 pg/mL) different during the development of heart failure compared with intact dogs.
Despite the attenuated changes in LV geometry in VD dogs during the development of heart failure, there were no significant differences in LV plus septum weight/body weight among all groups, ie, LV plus septum weight/body weight was 5.1±0.1 g/kg in intact dogs without heart failure, 4.9±0.3 g/kg in intact dogs with heart failure, 5.1±0.2 g/kg in VD dogs without heart failure, and 4.8±0.3 g/kg in VD dogs with heart failure.
Effects of VD on Responses to ISO in Heart Failure
Chronotropic Effects of ISO
In controls, the heart rate responses to ISO (0.4 μg·kg−1·min−1) increased significantly (P<.05) and similarly in both intact dogs (+103±15% from 104±3 bpm) and VD dogs (+100±10% from 104±6 bpm). After the development of heart failure, the heart rate response to ISO was attenuated similarly and significantly in both groups (intact, +33±6%; VD, +32±7%) (Table 2⇓). Similar chronotropic desensitization to ISO was observed in the presence of ganglionic blockade for both groups (Table 2⇓). Taken together, these heart rate data indicate that desensitization to the β-adrenergic agonist ISO occurred to a similar extent in the atria of both intact and VD dogs, in which cardiac innervation was intact. (See Figs 6⇓ and 7⇓.)
Inotropic Effects of ISO
In the control state, the LV dP/dt response to ISO (0.4 μg·kg−1·min−1) increased significantly (P<.05) in both intact (+165±15% from 3128±212 mm Hg/s) and VD (+205±19% from 2469±92 mm Hg/s) dogs. After the development of heart failure, the LV dP/dt response to ISO was attenuated significantly (P<.05) in the intact dogs (+72±10% from 1529±74 mm Hg/s) compared with VD dogs in which the ISO response was preserved (+198±23% from 1555±60 mm Hg/s) (Table 2⇑; Fig 8⇓).
To be certain that differences in heart rate responses did not contribute to the differential inotropic responses between intact and VD dogs, the response to ISO (0.4 μg·kg−1·min−1) was compared with heart rate held constant by atrial pacing at 240 bpm. In intact dogs, ISO (0.4 μg·kg−1·min−1, n=10) increased LV dP/dt by 151±13% from 3081±340 mm Hg/s before the development of heart failure, whereas the LV dP/dt response to ISO was significantly less (+68±10% from 1756±139 mm Hg/s; P<.05) after the development of heart failure. In contrast, in VD dogs, ISO (0.4 μg·kg−1·min−1, n=7) increased LV dP/dt similarly before and after heart failure (before, 184±24%; after, 180±25%) (Fig 9⇓). The presence of ganglionic blockade did not further alter the pattern or the magnitude of desensitization in either intact or VD dogs (Table 2⇑). Thus, selective VD prevented the development of desensitization to the inotropic effects of the β-adrenergic receptor agonist ISO.
Effects of VD on Responses to NE in Heart Failure
Chronotropic Effects of NE
Heart rate decreased significantly (P<.05) by 15±2% from 99±4 bpm in response to NE (0.2 μg·kg−1·min−1) in intact dogs. In VD dogs, the chronotropic effects of NE were not different compared with those in intact dogs, ie, −12±3% from 101±6 bpm. After the development of heart failure, NE decreased heart rate significantly less in both intact and VD dogs.
In the presence of ganglionic blockade, NE (0.2 μg·kg−1·min−1) increased heart rate by 27±8% in intact dogs before heart failure, and after heart failure, the increase in heart rate was less (4±2%; P<.05). In VD dogs, NE (0.2 μg·kg−1·min−1) increased heart rate by 30±9% before heart failure. After heart failure, the increase in heart rate in response to NE was less (12±4%; P<.05). These values were not different from those in intact dogs, suggesting that desensitization during the development of heart failure occurred similarly in the atria of both intact and VD dogs, in which cardiac innervation was intact.
Inotropic Effects of NE
In intact dogs, NE (0.2 μg·kg−1·min−1) increased LV dP/dt by 39±6% from 3098±237 mm Hg/s. After the development of heart failure, the LV dP/dt response to NE was desensitized significantly, ie, NE increased LV dP/dt significantly less (P<.05) by 24±5% from 1460±80 mm Hg/s. In contrast, the LV dP/dt responses to NE (0.2 μg·kg−1·min−1) were not desensitized after the development of heart failure in VD dogs. NE increased LV dP/dt by 143±19%, from 2404±56 mm Hg/s before heart failure. After heart failure, NE increased LV dP/dt by 136±20% from 1586±72 mm Hg/s (Fig 10⇓). These responses were not different before and after the development of heart failure (Table 3⇓).
To be certain that differences in heart rate responses did not contribute to the differential inotropic responses between intact and VD dogs, the response to NE (0.2 μg·kg−1·min−1) was compared with heart rate held constant with ventricular pacing at 180 bpm. In intact dogs, NE (0.2 μg·kg−1·min−1, n=10) increased LV dP/dt by 68±12% from 2642±211 mm Hg/s before the development of heart failure, whereas the LV dP/dt response to NE was significantly decreased (+38±4% from 1218±54 mm Hg/s; P<.05) after the development of heart failure. In contrast, in VD dogs, NE (0.2 μg·kg−1·min−1, n=7) increased LV dP/dt similarly before and after heart failure (before, 187±19%; after, 176±15%) (Fig 11⇓).
In the presence of ganglionic blockade, NE (0.2 μg·kg−1·min−1) increased LV dP/dt by 169±21% from 2140±146 mm Hg/s in intact dogs before the development of heart failure, whereas the LV dP/dt response to NE was attenuated significantly (+50±7% from 1141±101 mm Hg/s; P<.05) after the development of heart failure. In contrast, the LV dP/dt response to NE was not desensitized after the development of heart failure in VD dogs. NE (0.2 μg·kg−1·min−1) increased LV dP/dt by 255±35% from 2080±133 mm Hg/s before the development of heart failure and by 226±26% from 1648±17 mm Hg/s after the development of heart failure (Table 3⇑).
There are two unique features to the current investigation: (1) the use of an animal model with selective VD to study the role of cardiac nerves on the progression of experimental heart failure and (2) the use of this model to study mechanisms of catecholamine desensitization in heart failure. There were two major findings. First, the elimination of ventricular nerves attenuated the decline of cardiac function during the progression of heart failure. Second, we demonstrated that cardiac nerves, rather than circulating catecholamines, were responsible for physiological desensitization to sympathomimetic amines during the development of heart failure.
The approach used in the current investigation was different from prior studies in that selective VD was used to interrupt neural traffic, as opposed to pharmacological blockers. Selective VD did not exert major effects on baseline cardiac function. One reason is that heart rate and atrial function were not disturbed because atrial innervation remained intact. Second, in healthy, conscious dogs, baseline sympathetic tone is relatively low. Therefore, it was not surprising that selectively interrupting the ventricular cardiac nerves did not exert a major effect on baseline cardiac function.
The first major finding of the current investigation is that dogs with selective VD tolerated the development of heart failure better than control animals with intact nerves that were subjected to the chronic rapid-pacing stimulus. In support of this concept, increases in LV end-diastolic pressure and wall stress and decreases in myocardial contractility in VD dogs were less as heart failure developed than they were in intact dogs. However, chronotropic changes were similar because, as noted above, atrial innervation remained intact. The protective effects of VD were even more apparent in the presence of ganglionic blockade (Table 1⇑). Under these conditions, atrial function, eg, heart rate, did not change with induction of heart failure, but decreases in LV dP/dt and Vcfc and increases in LV end-diastolic pressure and stress were clearly less in VD dogs. One other study26 examined the effects of total cardiac denervation in the response to heart failure but failed to demonstrate improvement. There are four major differences between that study and the present study: (1) the use of selective VD versus total cardiac denervation; (2) direct measurements of preload, afterload, and myocardial contractility in the current study, which may be more precise than those obtained by echocardiography; (3) the current study was conducted in the presence of ganglionic blockade, in which differences between VD and intact dogs were even more apparent; and (4) the current investigation was conducted in awake animals, whereas echocardiographic studies generally require sedation, which could affect both cardiac loading conditions as well as contractility.26
Of critical importance in studies involving cardiac denervation is documentation that the denervation is complete and persistent. To this end, completeness of denervation was confirmed initially at the time of surgery by observing the abolition of effects of electric stimulation of sympathetic nerves. At the time that experiments were conducted in the conscious animals, arterial baroreflex responses elicited by phenylephrine and nitroglycerin confirmed that baroreflex control of heart rate remained intact. However, stimulation of ventricular receptors with veratrine alkaloids demonstrated that reflex hypotension and bradycardia were not observed in dogs with VD. After the end of the experiments, demonstration of depletion of tissue norepinephrine further confirmed denervation. Finally, a separate group of animals with VD was studied for 2 months, ie, the time period for all protocols to be completed in the dogs with heart failure. In these animals, supersensitivity to NE, a signature of denervation, was observed without abatement during the entire 2-month observation period.
The extent to which different models of denervation demonstrate supersensitivity remains controversial. Because this topic has not been examined in a model of selective VD, it was important to investigate it as part of the present study. Denervation supersensitivity is mainly caused by lack of neural uptake mechanisms for the neurotransmitter NE.27 Second, loss of baroreflex buffering of cardiac function plays a role.27 Third, postsynaptic supersensitivity mechanisms are also involved.27 For these reasons, supersensitivity to NE was most significant because with NE, several mechanisms, ie, lack of neural uptake, baroreflex buffering, and postsynaptic mechanisms, all play a role. After ganglionic blockade, the supersensitivity was still present but less apparent.
The second major goal of the current investigation was the determination of the mechanism of catecholamine desensitization in heart failure. It has not been clear whether the chronically elevated circulating catecholamines are responsible for the catecholamine desensitization, which is another hallmark of heart failure. If circulating catecholamines were the predominant stimulus, then desensitization should not have been attenuated in the animals with VD. Heart failure induced classic catecholamine desensitization to inotropic responses induced by both ISO and NE in intact dogs, both in the presence and absence of ganglionic blockade. However, this was not observed in dogs with VD. In contrast, chronotropic responses induced by both ISO and NE were desensitized similarly in intact dogs and dogs with VD, again confirming the presence of intact atrial innervation. Because chronotropic changes were similar in both groups, it was not surprising that additional experiments conducted with heart rate held constant by electrical pacing confirmed the inotropic desensitization to ISO and NE in intact dogs with heart failure and the lack of desensitization in VD dogs with heart failure.
Therefore, because desensitization was significantly prevented with the development of heart failure in VD dogs, it suggests that NE released from neurons, rather than circulating catecholamines, is the dominant trigger. In support of this are studies showing a much higher concentration of NE in the synaptic cleft than occurs in the systemic circulation.28 Moreover, administration of catecholamines to mimic the rise in plasma NE and EPI observed in heart failure has surprisingly little effect on cardiac function in normal animals.29 These observations in combination suggest that high concentrations of catecholamines are required to initiate desensitization mechanisms, which can be achieved by neural stimulation but not by the levels of circulating catecholamines generally observed in heart failure.
It might be argued that the reason that VD dogs failed to demonstrate catecholamine desensitization was simply that the severity of heart failure was less. To address this point, it must be noted that catecholamine desensitization is apparent at 1 day of pacing30 and decreases further with 1 week of pacing and then with 3 to 4 weeks of pacing. Hemodynamic impairment in terms of decreases in LV dP/dt and increases in LV end-diastolic pressure were similar in the intact dogs at 1 week and the VD group at 3 to 4 weeks (Fig 4⇑; Table 3⇑), yet as noted above, catecholamine desensitization was clearly apparent after 1 day of pacing in intact dogs30 but was not evident after 3 to 4 weeks of pacing in VD dogs (Figs 8⇑ and 10⇑).
In summary, cardiac nerves play a critical role in the progression of heart failure. As noted above, cardiac nerves mediate catecholamine desensitization. We might then speculate that desensitization mechanisms are potentially protective and that continued sympathetic stimulation might be deleterious in the progression of heart failure. Thus, although cardiac nerves may play an important role in the response to stress, in the chronic setting, these compensatory mechanisms may actually exacerbate the condition of cardiac dysfunction. In support of this concept are the current experiments in dogs with VD, which tolerated the development of heart failure better as evidenced by less severe hemodynamic perturbations. Further evidence can be extrapolated from recent experiments in our laboratory with a transgenic mouse model with overexpressed Gsα, in which chronically enhanced sympathetic stimulation during the life of the animals resulted in a picture of cardiomyopathy.31 32 Whether these concepts are unique to the specific models studied or can be extrapolated to human heart failure remains to be determined.
Selected Abbreviations and Acronyms
|LV||=||left ventricular, left ventricle|
|RV||=||right ventricular, right ventricle|
|Vcfc||=||mean velocity of circumferential fiber shortening corrected by heart rate|
Dr Sato is the recipient of the Samuel A. Levine Fellowship grant from the American Heart Association, Massachusetts Affiliate. Dr Shannon is the recipient of a Clinician-Scientist Award from the American Heart Association. This study was supported in part by USPHS grants HL-38070, HL-33107, HL-37404, and RR-00168.
- Received July 29, 1996.
- Revision received November 13, 1996.
- Accepted November 25, 1996.
- Copyright © 1997 by American Heart Association
Leimbach WN, Wallin BG, Victor RG, Aylward PE, Sundlöf G, Mark AL. Direct evidence from intraneural recordings for increased central sympathetic outflow in patients with heart failure. Circulation. 1986;73:913-919.
Chidsey CA, Harrison DC, Braunwald E. Augmentation of plasma norepinephrine response to exercise in patients with congestive heart failure. N Engl J Med. 1962;267:650-654.
Engelmeier RS, O’Connell JB, Walsh R, Rad N, Scanlon PJ, Gunnar RM. Improvement in symptoms and exercise tolerance by metoprolol in patients with dilated cardiomyopathy: a double-blind, randomized, placebo-controlled trial. Circulation. 1985;72:536-546.
Waagstein F, Bristow MR, Swedberg K, Camerini F, Fowler MB, Silver MA, Gilbert EM, Johnson MR, Goss FG, Hjalmarson Å, for the Metoprolol in Dilated Cardiomyopathy (MDC) Trial Study Group. Beneficial effects of metoprolol in idiopathic dilated cardiomyopathy. Lancet. 1993;342:1441-1446.
Krum H, Sackner-Bernstein JD, Goldsmith RL, Kukin ML, Schwartz B, Penn J, Medina N, Yushak M, Horn E, Katz SD, Levin HR, Neuberg GW, DeLong G, Packer M. Double-blind, placebo-controlled study of the long-term efficacy of carvedilol in patients with severe chronic heart failure. Circulation. 1995;92:1499-1506.
Metra M, Nardi M, Giubbini R, Cas LD. Effects of short- and long-term carvedilol administration on rest and exercise hemodynamics variables, exercise capacity and clinical conditions in patients with idiopathic dilated cardiomyopathy. J Am Coll Cardiol. 1994;24:1678-1687.
Tsutsui H, Spinale FG, Nagatsu M, Schmid PG, Ishihara K, DeFreyte G, Cooper G IV, Carabello BA. Effects of chronic β-adrenergic blockade on the left ventricular and cardiocyte abnormalities of chronic canine mitral regurgitation. J Clin Invest. 1994;93:2639-2648.
Warner AL, Bellah KL, Raya TE, Roeske WR, Goldman S. Effects of β-adrenergic blockade of papillary muscle function and the β-adrenergic receptor system in noninfarcted myocardium in compensated ischemic left ventricular dysfunction. Circulation. 1992;86:1584-1595.
Fowler MB, Laser JA, Hopkins GL, Minobe W, Bristow MR. Assessment of the β-adrenergic receptor pathway in the intact failing human heart: progressive receptor down-regulation and subsensitivity to agonist response. Circulation. 1986;74:1290-1302.
Colucci WS, Denniss AR, Leatherman GF, Quigg RJ, Ludmer PL, Marsh JD, Gauthier DF. Intracoronary infusion of dobutamine to patients with and without severe congestive heart failure. J Clin Invest. 1988;81:1103-1110.
Marzo KP, Frey MJ, Wilson JR, Liang BT, Manning DR, Lanoce V, Molinoff PB. β-Adrenergic receptor–G protein–adenylate cyclase complex in experimental canine congestive heart failure produced by rapid ventricular pacing. Circ Res. 1991;69:1546-1556.
Roth DA, Urasawa K, Helmer GA, Hammond HK. Downregulation of cardiac guanosine 5′-triphosphate-binding proteins in right atrium and left ventricle in pacing-induced congestive heart failure. J Clin Invest. 1993;91:939-949.
Colucci WS, Riberio JP, Rocco MB, Quigg RJ, Creager MA, Marsh JD, Gauthier DF, Hartley LH. Impaired chronotropic response to exercise in patients with congestive heart failure. Circulation. 1989;80:314-323.
Parker JD, Landzberg JS, Bittl JA, Mirsky I, Colucci WS. Effects of β-adrenergic stimulation with dobutamine on isovolumic relaxation in the normal and failing human left ventricle. Circulation. 1991;84:1040-1048.
Kiuchi K, Shannon RP, Komamura K, Cohen DJ, Bianchi C, Homcy CJ, Vatner SF, Vatner DE. Myocardial β-adrenergic receptor function during the development of pacing induced heart failure. J Clin Invest. 1993;91:907-914.
Liang C-S, Frantz RP, Suematsu M, Sakamoto S, Sullebarger JT, Fan T-HM, Guthinger L. Chronic β-adrenoceptor blockade prevents the development of β-adrenergic subsensitivity in experimental right-sided congestive heart failure in dogs. Circulation. 1991;84:245-266.
Harding SE, Poole-Wilson PA. Beta-adrenoceptors in heart failure: a mechanism or a cAMP follower? In: Lewis BS, Kimchi A, eds. Heart Failure Mechanisms and Management. Berlin, Germany: Springer-Verlag; 1991:15-20.
Randall WC, Thomas JX Jr, Barber MJ, Rinkema LE. Selective denervation of the heart. Am J Physiol. 1983;244(Heart Circ Physiol 13):H607-H613.
Levett JM, Marinelli CC, Lund DD, Pardini BJ, Nader S, Scott BD, Augelli NV, Kerber RE, Schmid PG Jr. Effects of β-blockade on neurohormonal responses and neurochemical markers in pacing-induced heart failure. Am J Physiol. 1994;266(Heart Circ Physiol 35):H468-H475.
Vatner DE, Lavallee M, Amano J, Finizola A, Homcy CJ, Vatner SF. Mechanisms of supersensitivity to sympathomimetic amines in the chronically denervated heart of the conscious dog. Circ Res. 1985;57:55-64.
Delehanty JM, Himura Y, Elam H, Hood WB, Liang C-S. β-Adrenoceptor downregulation in pacing-induced heart failure is associated with increased interstitial NE content. Am J Physiol. 1994;266(Heart Circ Physiol 35):H930-H935.
Young MA, Hintze TH, Vatner SF. Correlation between cardiac performance and plasma catecholamine levels in conscious dogs. Am J Physiol. 1985;248(Heart Circ Physiol 17):H82-H88.
Vatner DE, Sato N, Kiuchi K, Shannon RP, Vatner SF. Decrease in myocardial ryanodine receptors and altered excitation-contraction coupling early in the development of heart failure. Circulation. 1994;90:1423-1430.
Iwase M, Bishop SP, Uechi M, Vatner DE, Shannon RP, Kudej RK, Wight DC, Wagner TE, Ishikawa Y, Homcy CJ, Vatner SF. Adverse effects of chronic endogenous sympathetic drive induced by cardiac Gsα overexpression. Circ Res. 1996;78:517-524.
Iwase M, Uechi M, Vatner DE, Asai K, Shannon RP, Kudej RK, Wagner TE, Patrick TA, Ishikawa Y, Homcy CJ, Vatner SF. Cardiomyopathy induced by cardiac Gsα overexpression. Am J Physiol. 1997;272:H585-H589.