INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

RECENT STUDIES from our laboratory using dynamic systems analysis demonstrated a novel finding, i.e., that activation of either the sympathetic or vagal system augmented the dynamic heart rate (HR) response mediated by the other system (10, 11). We referred to this phenomenon as bidirectional augmentation. The mechanism underlying the augmentation, however, remains to be elucidated. With regard to the general phenomenon of sympathetic modulation of the vagally induced chronotropic response, bidirectional augmentation is consistent with the interaction well known as accentuated antagonism; that is, sympathetic stimulation augments the steady-state HR response to vagal stimulation (14-16). As a first step toward revealing the physiological basis for bidirectional augmentation, we investigated the mechanisms for sympathetic augmentation of the dynamic HR response to vagal stimulation.

Sympathetic stimulation leads to a cardiac response by releasing norepinephrine (NE) from sympathetic nerve terminals, which then binds to -adrenoceptors and causes accumulation of intracellular cAMP (8, 9). Possible mechanisms involved in accentuated antagonism include the release of ACh from vagus nerve terminals, thereby inhibiting the actions of sympathetic stimulation at prejunctional (21, 24) and postjunctional effector sites (1-7, 12-17, 19, 20). This study was designed to elucidate whether the sympathetic augmentation of dynamic HR response to vagal stimulation could have its sites of action at the postjunctional effector sites. Because the principal postjunctional interaction appears to operate through the adenylyl cyclase system (14-16), we examined the effects of pharmacological interventions known to modulate the adenylyl cyclase system on the transfer function from vagal stimulation frequency to HR response. The results suggest that the cAMP accumulating at the postjunctional effector sites as a result of sympathetic stimulation could augment the dynamic vagal control of HR.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Surgical preparations. Animal care was in accordance with institutional guidelines. Twenty-one Japanese white rabbits weighing 2.4-3.0 kg were anesthetically induced using an initial dose of urethan (250 mg/kg iv) and -chloralose (40 mg/kg iv) and mechanically ventilated with oxygen-enriched room air. Supplemental doses of anesthetics were given as necessary via the right femoral vein. Aortic pressure was monitored by means of a micromanometer catheter (model PC-340, 3-Fr, Millar Instruments, Houston, TX) inserted via the left femoral artery. Another catheter was inserted into the right femoral vein for the administration of drugs. The carotid sinus nerves and aortic depressor nerves were cut bilaterally to eliminate the effects of the arterial baroreflex systems. We transected the bilateral sympathetic nerves at the level of the stellate ganglion to eliminate the possible interaction between the vagus and sympathetic nerves. Vagus nerves were sectioned bilaterally at the neck, where a pair of bipolar platinum electrodes was attached to the cardiac end of the sectioned right vagus nerve for stimulation. To prevent drying and to provide insulation, the stimulation electrodes and the nerve were immersed in a mixture of white petrolatum (Vaseline) and paraffin. Finally, a pair of bipolar stainless electrodes was sutured to the right atrium to record the electrocardiogram for monitoring of HR. During all experiments, body temperature was maintained at 37°C with a heating pad.

Experimental procedures. The pulse duration of vagal stimulation was set at 2 ms. We adjusted the amplitude of stimulation to yield a HR decrease of ~50 beats/min at 5 Hz. This resulted in an amplitude of stimulation ranging from 2.0 to 5.0 V (3.4 ± 0.3 V). To estimate the transfer function from vagal stimulation frequency to HR response, we stimulated the vagus nerve using a pulse train that was frequency modulated by a band-limited Gaussian white noise (10, 11, 22). The main advantage of such a Gaussian white-noise approach is that it enables estimation of the unbiased linear input-output relation even in the presence of significant nonlinearities in the system (18). The instantaneous stimulation frequency was switched every second. The power spectrum, fairly constant up to 0.5 Hz, decreased gradually to 1/10 at ~0.8 Hz and attenuated sharply as the frequency increased to 1 Hz. We estimated the transfer function only up to 0.8 Hz because the lack of input power above that frequency made the estimation unreliable. The frequency range sufficiently spanned the physiological range of the vagal control of HR (10, 11). We used different perturbation command sequences of Gaussian white noise for different animals.

Estimation of the transfer function. After applying an anti-aliasing filter, we resampled the input (nerve stimulation frequency)-output (HR) data pairs at 10 Hz, then segmented the data into eight 50%-overlapping segments of 1,024 data points each. For each segment, the linear trend was subtracted and a Hanning window was applied. We then performed the fast Fourier transformation to obtain the frequency spectrum of nerve stimulation frequency, N( f), and that of HR response, HR( f). The resolution of the frequency was 0.01 Hz.

1

1

Chemicals. The following drugs were used: dl-isoproterenol hydrochloride, forskolin, dl-propranolol hydrochloride, and theophylline (Wako Pure Chemical Industries, Osaka, Japan). Isoproterenol was dissolved in 0.9% NaCl solution containing 0.01% L-ascorbic acid. Forskolin was prepared in 0.9% NaCl solution containing 1% ethanol. Theophylline was prepared in 0.1 N KOH solution.

Statistical analysis. Statistical significance was assessed within groups using the Student's paired t-test and between groups by Dunnett's multiple comparisons after one-way analysis of variance. Differences were considered statis- tically significant if P < 0.05. All values are presented as means ± SE.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of pharmacological interventions on HR and aortic pressure. Figure 1 shows typical recordings of the vagal stimulation frequency and the associated changes in HR in the absence and presence of isoproterenol infusion. HR changed in a manner roughly reciprocal to the stimulation pattern. Isoproterenol elevated the mean level of HR and augmented HR response to the dynamic vagal stimulation. As summarized in Table 1, all of the interventions studied increased the mean level of HR before and during dynamic vagal stimulation (P < 0.05). In the experiments with forskolin and with theophylline, we pretreated rabbits with propranolol (0.5 mg/kg iv). Therefore, the control HR levels of these groups tended to be lower than those of the isoproterenol group. These differences, however, did not reach a statistically significant level. Moreover, this dose of propranolol did not affect mean aortic pressure. Whereas neither isoproterenol nor theophylline significantly changed mean aortic pressure, forskolin decreased mean aortic pressure from 91 ± 4 to 69 ± 6 mmHg (P < 0.05).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Representative recordings of vagus nerve stimulation frequency (VNS) via band-limited Gaussian white noise (top) and the associated heart rate (HR) responses (bottom) before (control; left) and during infusion of isoproterenol (0.3 µg · kg1 · min1 iv; right).

Effect of isoproterenol on the transfer function. Figure 2A shows the transfer function from vagal stimulation frequency to HR response in the absence and presence of isoproterenol. The gain plots, phase plots, and coherences are shown. As already shown in our previous report (10), characteristics of the gain and phase plots match what is known as a first-order low-pass filter with time lag. The phase shift was nearly out-of-phase (i.e., radians) at its lowest frequency and further delayed with increases in frequency. The coherence was above 0.8 in the frequency range from 0.01 to 0.3 Hz. The high coherence values were similar to those reported previously (10, 11), indicating that the HR response linearly depends on the vagal stimulation frequency in this frequency range. Isoproterenol increased the gain of the transfer function. However, neither the phase shift nor the coherence was changed by the intervention. These observations are summarized in Table 2. Isoproterenol did not change the parameters of the transfer function, except for increasing the dynamic gain from 7.12 ± 0.67 to 12.4 ± 1.21 beats · min¹ · Hz¹ (P < 0.05).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   A: transfer functions from vagal stimulation frequency to HR change before (control; left) and during infusions of isoproterenol (0.3 µg · kg1 · min1 iv; right) (pooled data; n = 7). Gains (top), phase shifts (middle), and coherence functions (Coh; bottom) are shown. Gain and frequency axes are logarithmically scaled. Solid lines indicate means, and dashed lines indicate mean + SE values. bpm, Beats/min. B: calculated step responses to 1-Hz tonic vagal stimulations shown for corresponding transfer functions. Solid lines indicate means, and dashed lines indicate mean  SE values.

Effect of forskolin on the transfer function. As shown in Fig. 3A, the effects of forskolin on the transfer function were similar to those of isoproterenol. Table 2 shows that forskolin did not change the parameters of the transfer function, except for increasing the dynamic gain from 7.20 ± 1.15 to 10.4 ± 1.97 beats · min¹ · Hz¹ (P < 0.05).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   A: transfer functions from vagal stimulation frequency to HR change before (control; left) and during infusions of forskolin (5 µg · kg1 · min1 iv; right) (pooled data; n = 7). Gains (top), phase shifts (middle), and coherence functions (bottom) are shown. Solid lines indicate means, and dashed lines indicate mean + SE values. B: calculated step responses to 1-Hz tonic vagal stimulations shown for corresponding transfer functions. Solid lines indicate means, and dashed lines indicate mean  SE values.

Effect of theophylline on the transfer function. As shown in Fig. 4A, theophylline increased the gain of the transfer function without altering coherence. Table 2 shows that theophylline increased the dynamic gain from 7.96 ± 1.09 to 12.5 ± 1.27 beats · min¹ · Hz¹ (P < 0.01) and slightly decreased the corner frequency from 0.13 ± 0.02 to 0.11 ± 0.02 Hz (P < 0.01). However, theophylline did not change the lag time.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   A: transfer functions from vagal stimulation frequency to HR change before (left) and after injections of theophylline (20 mg/kg iv; right) (pooled data; n = 7). Gains (top), phase shifts (middle), and coherence functions (bottom) are shown. Solid lines indicate means, and dashed lines indicate mean + SE values. B: calculated step responses to 1-Hz tonic vagal stimulations shown for corresponding transfer functions. Solid lines indicate means, and dashed lines indicate mean  SE values.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

We have shown that all pharmacological interventions to increase the level of intracellular cAMP examined in this study augmented the gain of the transfer function from vagal stimulation frequency to HR response. Isoproterenol and forskolin augment cAMP synthesis via the indirect and direct activation, respectively, of adenylyl cyclase, whereas theophylline prevents cAMP hydrolysis by inhibiting PDE. The fact that the increase in gain of the transfer function was observed irrespective of the pharmacological mechanisms to increase cAMP suggested that it is nonspecifically coupled with the level of cAMP. Because the effects of these interventions on the transfer function were similar to those of simultaneous sympathetic stimulation in our previous study (10, 11), we conjecture that the accumulation of cAMP at postjunctional effector sites contributes, at least in part, to the sympathetic augmentation of the dynamic HR response to vagal stimulation.

In this study, we evaluated the influences of all pharmacological interventions on both the transfer function in the frequency domain and the calculated step response in the time domain. We derived the time domain representation of the dynamic characteristics of the transfer function in order to facilitate a somewhat intuitive interpretation of the system properties, even though the frequency domain representation faithfully provides the opportunity to understand the functional mechanisms of autonomic interaction involved in the control of HR. The gain and the corner frequency of the transfer function reflect the steady-state amplitude and time constant of the step response, respectively. Hence, the fact that all pharmacological interventions increased the gain of the transfer function means that the amplitude of the steady-state HR response to vagal stimulation was increased in all cases. Unlike isoproterenol and forskolin, theophylline alone slightly (~15%) decreased the corner frequency. Although the alteration in corner frequency was not reflected in the time constant of the calculated step response, the current data suggested that slowing of the degradation of cAMP might somewhat decelerate the HR response to vagal stimulation. Needless to say, however, this study indicated that the effect of theophylline was by far larger on the gain than on the corner frequency. Thus the augmented response of HR to vagal stimulation appears mainly attributable to accumulation of cAMP.

The mechanism by which the accumulated cAMP augments the dynamic HR response to vagal stimulation is not entirely clear. It is well known that the stimulation of muscarinic receptors directly alters the characteristics of the ACh-sensitive K⁺ channel (8), which, in turn, slows HR regardless of the level of cAMP. If intracellular cAMP accumulates, it is conceivable that the stimulation of muscarinic receptors may augment the negative chronotropic response by decreasing the level of cAMP through the inhibition of adenylyl cyclase (4, 7, 19, 20) and through the acceleration of cAMP hydrolysis (5, 6, 12). Increases in mean HR resulting from the accumulation of cAMP might also affect the HR response to vagal stimulation. Additionally, a muscarinic antagonism of signal transduction with cAMP accumulation may also be operative (1, 17, 25).

Prejunctional interaction might also have been responsible for the augmentation observed with concomitant sympathetic stimulation of the dynamic HR response to vagal stimulation. In prejunctional interaction, ACh released from vagus nerve terminals inhibits the release of NE from sympathetic nerve terminals, whereas NE and neuropeptide Y released from sympathetic nerve terminals inhibit the release of ACh from vagus nerve terminals (16). Our previous study has shown that simultaneous sympathetic stimulation did not affect the parameters of the transfer function from vagal stimulation frequency to HR response, except for increases in dynamic gain, despite the fact that vagal stimulation changed HR more quickly than did sympathetic stimulation (10). The data suggest that the reduction of NE release from sympathetic nerve terminals does not contribute substantially to the augmentation of the HR response to dynamic vagal stimulation. Moreover, because NE and neuropeptide Y inhibit the release of ACh from vagus nerve terminals (21, 24), if the inhibitory mechanism predominated, then the gain of the transfer function from vagal stimulation frequency to HR response would decrease. Therefore, it seems that prejunctional interaction may be less important in the sympathetic augmentation of the dynamic HR response to vagal stimulation.

Our previous report also suggested that the operating point of HR regulation would be determined through a sigmoidal relation between autonomic nervous activity and HR (10). This framework could explain bidirectional augmentation. Briefly, whenever we stimulated the sympathetic system or the vagal system alone, the operating point of the HR deviated from the steepest region of the sigmoidal curve, resulting in a loss of gain in the HR response. Simultaneous stimulation of the sympathetic and vagal systems moved the operating point back to the steepest region of the sigmoidal curve, thereby increasing the HR response to dynamic stimulation. In this study, we selected doses of each drug so as to mimic the HR response to sympathetic stimulation observed in our previous studies (10); therefore, these interventions should also shift the operating point and increase the gains of the transfer functions. However, the framework predicts that extremely low or high levels of sympathetic stimulation would result in a decrease in the dynamic HR response to vagal stimulation. Hence, the smaller and the larger doses of each drug used in this study may fail to increase the gain of transfer function of vagal stimulation frequency to HR response. Indeed, the presence of an effective stimulation range necessary for eliciting observable interactions between the sympathetic and the vagal system in regulating cardiac function have been shown previously (7, 23). Obviously, further studies are required to validate this framework.

In summary, we found that pharmacological interventions, which mimicked the sympathetic-mediated cardiac response, increased the gain of the transfer function from vagal stimulation frequency to HR response. These results suggest that the accumulation of cAMP at postjunctional effector sites contributes, at least in part, to the sympathetic augmentation of the dynamic HR response to vagal stimulation. Because the data presented here were obtained from experiments using anesthetized rabbits, the relative significance of cAMP under normal physiological conditions must be evaluated in careful investigations using conscious rabbits. Furthermore, to understand further the mechanisms of bidirectional augmentation, we should elucidate whether the vagal augmentation of HR response to dynamic sympathetic stimulation could be mediated through the adenylyl cyclase system. Nevertheless, we would like to stress here that this is the first report that demonstrates the importance of the postjunctional interaction in the sympathetic augmentation of vagally mediated dynamic changes in HR.