INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

FUNCTIONAL LATERALITY is known to exist in the sympathetic efferent nervous system regulating the heart. The positive chronotropic effect is stronger during electrical stimulation of the right than the left cardiac sympathetic efferent nerves (15, 24). The inotropic response may be most accurately assessed by a slope of the end-systolic pressure-volume relationship (E_es), a load-insensitive index of ventricular contractility (25). Electrical stimulation of the right cardiac sympathetic nerve increases E_es by a direct inotropic effect through the ventricular innervation and by an indirect inotropic effect through changes in heart rate (HR) (17, 20). In contrast, electrical stimulation of the left cardiac sympathetic nerve increases E_es mainly through direct inotropic effect. The functional laterality of the sympathetic effects on the heart is evident in not only static but also dynamic regulation of HR and E_es (18). Complex innervation patterns of the left and right cardiac sympathetic nerves, which would underlie the functional laterality, have been described in detail (23). The putative laterality of cardiac innervation patterns, however, becomes less evident when decentralized stellate or middle cervical ganglia are investigated (22).

Regional differences in sympathetic nerve activity (SNA) are considered to have physiological significance for the individualized regulation of target organs (21). Dynamic baroreflex regulation of SNA shows derivative characteristics (7). In other words, the magnitude of SNA response becomes greater as the input frequency of baroreceptor pressure perturbation increases. In our previous study, cardiac SNA and renal SNA showed different derivative characteristics, suggesting that dynamic baroreflex regulation differed among sympathetic nerves (11). As the functional laterality of cardiac sympathetic nerves resulted in different ventricular performance in response to electrical stimulation of the left- and right-sided neuronal structures (18), it is possible that the dynamic baroreflex regulation of cardiac sympathetic nerves differs between the left and right sides to effectively control cardiac function.

Although Ninomiya et al. (21) reported no detectable difference in left cardiac SNA (LSNA) and right cardiac SNA (RSNA) in response to epinephrine-induced arterial pressure (AP) input, we think that the study was incomplete with respect to the following points. First, the quantitative description was not made to what extent LSNA and RSNA were similar. Second, because the rate of pressure changes critically affects the baroreflex responses (26), the epinephrine-induced AP input, where the rate of pressure change is uncontrollable exactly, is not a proper method for analyzing the input-output relationship of the baroreflex system. Finally, despite the importance of the dynamic characteristics of the baroreflex system in stabilizing AP (7), the study did not deal with dynamic baroreflex regulation of SNA. To extend the study by Ninomiya et al. (21), more sophisticated analysis of the baroreflex regulation of LSNA and RSNA is mandatory. Accordingly, the present study was designed to test the hypothesis that the dynamic baroreflex regulation differs between LSNA and RSNA. We performed a baroreflex open-loop experiment while recording LSNA and RSNA simultaneously in anesthetized rabbits. The results obtained indicated that no laterality existed in dynamic or static baroreflex regulation of cardiac sympathetic nerves as far as grouped axonal activity was concerned.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Surgical Preparations

Multifiber Preparations

Because we recorded grouped axonal activity, varied activities from individual axons in the nerve bundle could not be determined (2). Although we analyzed synchronous phasic activity generated by the majority of efferent axons, nonphasic activity or asynchornous activity may be buried in the phasic activity. In addition, the cardiac sympathetic nerve examined may have contained axons innervating pulmonary tissues. Such axonal components may be irresponsive to the baroreceptor pressure input. Therefore, the recorded LSNA and RSNA could not represent all of the individual axonal activities in the nerve bundle. Notwithstanding the abovementioned limitation relating multifiber preparations, we think that examining the laterality of grouped axonal activity in cardiac sympathetic efferent nerves is useful for better understanding the baroreflex control over the heart.

Protocols

Dynamic protocol. CSP in both carotid sinuses was servo-controlled to follow AP until steady-state pressure was reached. After obtaining operating pressure (OP) at the steady state, we assigned CSP randomly to either high (OP + 20 mmHg) or low (OP  20 mmHg) pressure every 500 ms according to a binary white noise sequence for 10 min (10, 11, 13). The input power spectra of CSP were reasonably flat up to 1 Hz. The upper frequency bound of the input perturbation was determined from the dynamic characteristics of the carotid sinus baroreflex determined from previous studies (7, 10, 11, 13). Because the total baroreflex showed low-pass characteristics and its dynamic gain at 1 Hz was less than one-tenth than at 0.01 Hz in rabbits, we focused on the dynamic characteristics in the frequencies below 1 Hz. Conversely, normal baroreceptor afferent activity has a dynamic pattern associated with the rising and falling phases of the beat-by-beat pressure pulse itself. Possible differences between LSNA and RSNA relating such pulsatile pressure input could not be assessed by the present protocol.

Static protocol. CSP was first decreased to 40 mmHg. After LSNA and RSNA responses reached steady state, CSP was increased stepwise from 40 to 160 mmHg at increments of 20 mmHg. Each pressure step was maintained for 60 s. We recorded CSP, LSNA, RSNA, and AP at a sampling rate of 200 Hz using a 12-bit analog-to-digital converter. The data were stored on the hard disk of a dedicated laboratory computer system for later analysis.

Data Analysis

(1)

(2)

To compare the dynamic responses to CSP perturbation between LSNA and RSNA directly, the transfer function from LSNA to RSNA was calculated. In this context, the transfer function represented the amplitude ratio and the phase difference between LSNA and RSNA in the frequency domain.

In the static protocol, we performed a regression analysis for the four-parameter logistic function using input-output data pairs averaged from the last 10 s of each CSP level as follows (14).

(3)

where y indicates the LSNA or RSNA value corresponding to each CSP level. P₁ is a response range (i.e., the difference between the maximum and minimum values of y). P₂ is a coefficient of gain. P₃ is a midpoint of the operating range in the CSP axis. P₄ is a minimum value of y.

To compare the static responses to CSP perturbation between LSNA and RSNA directly, we performed a linear regression analysis on LSNA-RSNA data pairs. A coefficient of determination (r²) was calculated from the following equation.

(4)

where LSNA(i) and RSNA(i) indicate the LSNA and RSNA values averaged from the last 10 s of each CSP level, respectively. n Represents the number of CSP steps in the static protocol. and indicate the mean values of LSNA(i) and RSNA(i), respectively. The goodness of fit (q) was assessed by the standard error of estimate divided by the range of RSNA using the following equation.

(5)

where RSNA_est(i) is a linear estimate of RSNA(i) from LSNA(i). RSNA_max and RSNA_min represent the maximum and minimum values of RSNA(i), respectively. q Approaches zero with increasing accuracy of estimation.

Statistical Analysis

In the static protocol, because the P₁ and P₄ values could be matched between CSP-LSNA and CSP-RSNA curves by arbitrarily scaling of LSNA and RSNA, we compared only the P₂ and P₃ values by paired t-test (6). Note that the P₂ value has a reciprocal relationship with the width of operating range (P_range) in the CSP axis independent of the other parameters as follows.

(6)

where P_th and P_sat denote the threshold and saturation pressure, respectively. k Is a constant for calculating the threshold and saturation points. We used k = 1.317 according to the model by Kent et al. (14).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Figure 1 shows the typical time series of CSP, LSNA, RSNA, and AP. CSP was perturbed with a binary white noise sequence. When CSP was increased, both LSNA and RSNA decreased. When CSP was decreased, the opposite responses were observed. Although each discharge pattern between LSNA and RSNA did not match exactly, general dynamic responses were similar between the two activities. Changes in AP did not affect CSP in the present experimental settings, validating the application of a conventional open-loop transfer function analysis (9, 12, 16).

View larger version (45K): [in this window] [in a new window]   Fig. 1.   Representative recordings of intracarotid sinus pressure (CSP), left cardiac sympathetic nerve activity (LSNA), right cardiac SNA (RSNA), and aortic pressure (AP) during the dynamic protocol. CSP was changed according to a binary white noise signal. Data were resampled at 10 Hz for this panel. a.u., Arbitrary units.

Figure 2 depicts averaged H_L and H_R from all animals. Gain plots (top), phase plots (middle), and coherence functions (bottom) are shown. The gain value approximated unity at the lowest frequency in both H_L and H_R because of the normalization procedure. The gain values increased as the input frequency of CSP increased between 0.03 and 0.3 Hz, indicating derivative characteristics of the neural arc. The slope of the increasing transfer gain was similar between H_L and H_R. The phase plots indicated an out-of-phase relationship between the input and output signals below 0.2 Hz, reflecting negative feedback attained by the baroreflex neural arc. The coherence functions associated with H_L and H_R showed no marked difference.

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Transfer function from CSP to LSNA (HL) and from CSP to RSNA (HR). The gain plots, phase plots, and coherence functions (Coh) are shown. Solid and dashed lines represent mean and [means ± SD] values, respectively.

Table 1 summarizes the parameters of the neural arc transfer functions. Neither gain nor phase values differed between H_L and H_R at 0.01, 0.1, 0.5, and 1 Hz. The slope of increasing dynamic gain showed no difference between H_L and H_R.

Figure 3 shows the amplitude ratio and phase difference between LSNA and RSNA during the dynamic protocol averaged from all animals. The amplitude ratio of RSNA to LSNA approximated unity in the frequencies between 0.01 and 1 Hz. The phase difference between LSNA and RSNA approximated zero radians over the frequency range under study. The coherence values between LSNA and RSNA were above 0.9 in the frequencies below 0.1 Hz and near unity above 0.1 Hz.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Amplitude ratio and phase difference between LSNA and RSNA in the frequency domain. Solid and dashed lines represent mean and [means ± SD] values, respectively.

Figure 4 presents typical results obtained from the static protocol. Figure 4A shows the time series of CSP, LSNA, and RSNA. LSNA and RSNA decreased in response to the increments in CSP. Figure 4B illustrates the CSP-LSNA and CSP-RSNA relationships obtained from the last 10 s of each CSP level. The fitted logistic functions were superimposable by arbitrary scaling of the ordinate. The group averaged parameters of the coefficient of gain and the midpoint of the operating range did not differ between the CSP-LSNA and CSP-RSNA curves (P₂: 0.125 ± 0.049 vs. 0.125 ± 0.039 AU/mmHg; P₃: 110.2 ± 13.1 vs. 109.9 ± 13.8 mmHg). The width of the operating range derived from the P₂ values was 24.4 ± 10.5 mmHg for the CSP-LSNA curve and 23.2 ± 7.7 mmHg for the CSP-RSNA curve. Figure 4C illustrates the scattergram of RSNA vs. LSNA obtained from the same data as in Fig. 4B. The static LSNA-RSNA relationship approximated a straight line, indicating no significant difference between LSNA and RSNA. r² Value of the linear regression analysis was 0.998 in Fig. 4C and ranged from 0.969 to 0.999 across all animals. The standard error of estimation divided by the range of RSNA (Eq. 5) was 0.017 in Fig. 4C and ranged from 0.005 to 0.074 across all animals.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   A: typical time series of CSP, LSNA, and RSNA during the static protocol. Increments in CSP decreased LSNA and RSNA. Data were resampled at 2 Hz. B: sigmoidal curves representing the steady-state CSP-LSNA and CSP-RSNA relationships. The 2 sigmoidal curves were superimposable by arbitrarily scaling of the ordinate. C: scattergram of RSNA vs. LSNA. The LSNA-RSNA relationship approximated a straight line (r2 = 0.998).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

To our knowledge, the present study is the first to demonstrate that the arterial baroreflex regulation was indistinguishable between LSNA and RSNA in dynamic characteristics when grouped activity of whole nerve slips was examined. Despite the significant laterality in the chronotropic and inotropic responses to electrical stimulation of the cardiac sympathetic nerves, the carotid sinus baroreflex may not use the functional laterality of the sympathetic effects on the heart to optimize cardiac performance for dynamic AP regulation.

Dynamic Baroreflex Regulation of LSNA and RSNA

If the derivative characteristics of the baroreflex neural arc are totally ascribable to the baroreceptor transduction properties and what we observed in H_L and H_R are the simple reflection of the baroreceptor transduction properties, similarity between H_L and H_R might be a matter of course. However, the derivative characteristics of the neural arc from pressure input to sympathetic efferent nerve activity are more enhanced than those of the baroreceptor transduction alone, suggesting that the central processing plays an important role in determining the derivative characteristics (11, ²⁷).

In a previous study, Ikeda et al. (8) demonstrated that instantaneous HR was successfully decoded from sympathetic efferent nerve activity in anesthetized and vagotomized rabbits. In that study, LSNA was used as the input for the HR prediction. Because HR could be predicted from LSNA using a transfer function analysis as long as HR was explained by linear dynamics with LSNA, the uniformity between LSNA and RSNA is not a prerequisite for HR prediction. However, the uniformity between LSNA and RSNA may have contributed to the successful prediction of HR in respect to the following. If the frequency bandwidth of the LSNA response had been much narrower than that of the RSNA response, the precise prediction of HR from LSNA alone might have been impossible.

Limitations

Second, we sectioned the vagal nerves to simplify our system identification. Because the vagal efferent nerves participate in the chronotropic and inotropic effects on the heart (19), further studies are required to clarify whether the baroreflex control differs between the left and right vagal nerve activities.

Third, most of the previous descriptions of functional laterality of cardiac sympathetic nerves have been obtained from the dog. In the cat, activation of the right rostral ventral lateral medulla (RVLM) caused a HR increase greater than activation of the left RVLM (5). Although we usually observed a marked HR decrease only after sectioning the right inferior cardiac sympathetic nerve in the present study, we did not directly assess the functional laterality of the cardiac sympathetic nerves. Further studies are required to elucidate to what degree the rabbit shows functional laterality of cardiac sympathetic nerves as compared with the more commonly studied dog.

Finally, we filled the isolated carotid sinuses with warmed physiological saline. As the ionic content affects the sensitivity of the baroreceptors (2), the absolute gain values of the carotid sinus baroreflex might have differed from normal physiological values. However, because we perturbed CSP without changing intravascular ion content, LSNA and RSNA responses should mainly result from changes in CSP.

In conclusion, uniformity in the dynamic as well as static baroreflex regulation of LSNA and RSNA was identified. The carotid sinus baroreflex may not use the functional laterality of the cardiac sympathetic efferent nerves to optimize cardiac performance for AP regulation. The significance of the neural regulation on the heart compared with humoral regulation appears to be the speed of its action rather than the differential regulation of the chronotropic and inotropic effects on the heart.

Perspectives