| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
NEUROHUMORAL CONTROL OF CIRCULATION AND HYPERTENSION
Département de Physiologie et Pharmacologie Clinique, Centre National de la Recherche Scientifique Formation de Recherche en Evolution 2678, Faculté de Pharmacie, Université Claude Bernard Lyon 1, Lyon 69373, France
Submitted 28 February 2003 ; accepted in final form 5 June 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
50% spectral power of both variables in the midfrequency band (MF,
0.27-0.74 Hz) containing the so-called Mayer waves, and induced an eightfold
increase in MAP power without altering RSNA power in the low-frequency band
(0.005-0.27 Hz). In both groups of rats, coherence between RSNA and MAP was
maximal in the MF band and was usually weak at lower frequencies. In SAD rats,
the transfer function from RSNA to MAP showed the characteristics of a
second-order low-pass filter containing a fixed time delay (0.5 s). These
results indicate that arterial baroreceptors are not involved in production of
respiratory-related oscillations of RSNA but play a major role in the genesis
of synchronous oscillations of MAP and RSNA at the frequency of Mayer waves.
The weak coupling between slow fluctuations of RSNA and MAP in sham-operated
and SAD rats points to the interference of noise sources unrelated to RSNA
affecting MAP and of noise sources unrelated to MAP affecting RSNA. Mayer waves; renal sympathetic nerve activity; sinoaortic baroreceptor denervation; spectral analysis; transfer function
90% of the time in conscious
baroreceptor-intact rats and much less frequently in SAD rats
(1). However, it could not be
excluded that this inverse relation might, at least partly, be explained by
the prominent, reciprocal oscillations of RSNA and AP that occur spontaneously
at the frequency of 0.4 Hz in rats
(4,
7). Surprisingly, with the use
of cross-spectral techniques, little, if any, coherence between low-frequency
AP and RSNA fluctuations was reported in the conscious rat
(8,
9). This observation suggested
that the dynamic relations between slow RSNA and AP fluctuations cannot be
explained solely on the basis of simple baroreflex patterns. An obvious experimental approach to the question of the involvement of the baroreceptor reflex in the coupling (or the lack of coupling) between AP and RSNA would be to evaluate the effects of arterial baroreceptor denervation on RSNA and AP variabilities. This has been done in anesthetized rats (12), i.e., in the absence of behavioral influences that make a major contribution to RSNA and AP variabilities (7, 27). To our knowledge, frequency domain analysis of RSNA variability in conscious SAD rats has been performed on one occasion (23). In this study, spectra were calculated over 20-s periods, which precluded the analysis of <0.1-Hz fluctuations.
The main objective of the present study was to describe the effects of arterial baroreceptor denervation on spontaneous fluctuations of RSNA in conscious rats, as well as on the coupling between these fluctuations and those of AP. The renal circulation receives a large fraction of cardiac output, and thus changes in RSNA can affect AP through changes in renal vascular resistance (3, 13, 16, 26). Therefore, the study provided the opportunity to characterize the feedforward link between RSNA and AP under open-loop conditions by modeling the transfer function between these variables in SAD rats.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
SAD. SAD was performed as previously described (33). Rats were anesthetized with a mixture of acepromazine maleate (12 mg/kg ip) and ketamine hydrochloride (120 mg/kg ip). Aortic baroreceptor denervation was achieved by bilaterally removing the superior cervical ganglia and sectioning the superior laryngeal nerves and aortic depressor nerves. Carotid baroreceptor denervation was accomplished by stripping all fibers from the carotid bifurcations and applying 10% phenol (in 95% ethanol). Sham surgery consisted of a midline neck incision and bilateral retraction of the sternohyoideus muscles. Animals were studied 14 days after denervation or sham surgery.
Measurement of AP. Under halothane anesthesia (1.5-2% in oxygen), femoral arterial and venous polyethylene catheters were inserted into the lower abdominal aorta and the inferior vena cava for AP measurement and drug administration, respectively. Both catheters were tunneled subcutaneously and exteriorized between the scapulae. AP was measured by connection of the arterial catheter to a precalibrated pressure transducer (Statham P23 ID, Gould, Cleveland, OH) coupled to an amplifier (model 13-4615-52, Gould)-chart recorder (model 8802, Gould).
Measurement of RSNA. RSNA was measured using a previously described technique (1). Briefly, under pentobarbital sodium anesthesia (60 mg/kg ip), the left renal nerve was exposed via a flank incision. After careful isolation, a major branch of the renal nerve was placed on a bipolar platinum-iridium electrode and insulated with a silicone gel (Wacker Chemie, Munich, Germany). The electrical signal from the nerve was amplified (50,000 times), band-pass filtered (300-3,000 Hz; model P-511J, Grass, Quincy, MA), and rectified (analog home-made rectifier including a low-pass filter with a cutoff frequency of 5 Hz). The rectified RSNA was then simultaneously recorded with the AP signal on the chart recorder.
Experimental protocol. Twelve days after baroreceptor denervation or sham surgery, the arterial and venous catheters were implanted. One day later, the renal electrode was positioned, and the rats were then allowed 12-14 h for recovery from anesthesia before the experiments were started.
On the day of the study, i.e., 14 days after denervation, while the rats
were unrestrained, AP and RSNA were simultaneously and continuously recorded
for 
3 h. Then the sensitivity of the baroreceptor reflex was evaluated
using intravenous injections of phenylephrine hydrochloride (1.5 µg/kg;
Sigma Chemical, St. Louis, MO) and sodium nitroprusside (8 µg/kg; Sigma
Chemical). Finally, at the end of the experiment, the background noise level
of RSNA was measured as the residual electrical activity obtained after
administration of the short-acting ganglionic blocker trimethaphan camsylate
(10 mg/kg iv; Hoffmann-LaRoche, Basel, Switzerland).
Data acquisition and analysis. With the use of a personal computer with an analog-to-digital converter (model ATMIO-16, National Instruments, Austin, TX) and LabVIEW 5.0 software (National Instruments), the AP and RSNA data were sampled at 500 Hz and stored on CD-ROM.
From the total recording, one continuous 60-min period free of artifacts was selected for further analysis. Off-line processing of data was performed on a workstation (Sparc1, Sun Microsystems, Mountain View, CA). For each cardiac cycle, the computer calculated mean AP (MAP) and heart rate (HR). Beat-to-beat time series of MAP were then resampled at 10 Hz after linear interpolation. RSNA data were averaged over consecutive 100-ms periods, the background noise was subtracted, and all values were normalized by the mean value calculated over the 1-h period. For MAP and RSNA, the 10-Hz time series were segmented into 34 data sets of 2,048 points (204.8 s) overlapping by one-half. For each data set, power spectral density was calculated using a fast Fourier transform algorithm after linear trend removal and application of a Hanning window (10). The spectra obtained for the different data sets were averaged. The frequency resolution was 0.005 Hz, and the upper frequency limit was 5 Hz. Spectral powers were calculated by integration within three frequency bands: a low-frequency (LF) band, extending from 0.005 to 0.269 Hz, a midfrequency (MF) band, extending from 0.273 to 0.742 Hz, and a high-frequency (HF) band, extending from 0.747 to 5 Hz.
The coherence function between RSNA and MAP was assessed by cross-spectral analysis as previously described (10). Coherence values exceeding 0.2 indicate a statistically (P < 0.01) reliable linear relation between fluctuations of the two signals (22). In addition, cross-spectral analysis was used to calculate the gain and phase of the transfer function (4, 5) from RSNA to MAP. In each SAD rat, gain and phase functions were subjected to linear modeling analysis (see RESULTS).
The baroreceptor reflex sensitivity was estimated as the ratio of the peak change in HR (beats/min) or RSNA (percent change from predrug value) to the peak change in MAP (mmHg) after phenylephrine and nitroprusside administrations.
Statistics. Values are means ± SE. Statistical comparisons between sham-operated and SAD rats were performed using the nonparametric Mann-Whitney U test.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Baroreflex sensitivities, mean levels, and overall variabilities of MAP, HR, and RSNA. As indicated in Table 1, SAD markedly reduced the reflex HR and RSNA responses to phenylephrine-induced increases in MAP and nitroprusside-induced decreases in MAP. The 1-h mean levels of MAP and HR did not differ significantly between sham-operated and SAD rats. After denervation, the overall variabilities (variation coefficients) of MAP and HR were increased and decreased, respectively, whereas the RSNA variability was not significantly altered (Table 1). The cardiovascular consequences of baroreceptor denervation, especially the exaggerated MAP variability, are illustrated in Fig. 1.
|
|
Spectral analysis of MAP and RSNA time series. For each group of
rats, the 1-h average spectra of MAP and RSNA and the corresponding coherence
function are shown in Fig. 2.
In the HF band containing respiratory fluctuations
(30), peaks were observed in
the MAP and RSNA spectra from sham-operated and SAD rats. For both variables,
spectral power in the HF band was not altered after denervation
(Table 2) and maximum coherence
did not differ between sham-operated (0.55 ± 0.02 at 1.54 ± 0.09
Hz) and SAD (0.60 ± 0.04 at 1.46 ± 0.07 Hz) rats. In
sham-operated rats, spectral peaks for MAP and RSNA were also present in the
MF band (Fig. 2). Coherence
between RSNA and MAP reached a maximum (0.89 ± 0.03) at 0.42 ±
0.01 Hz. After SAD, MF spectral power was reduced by 50% for MAP and RSNA
(Table 2) and residual power in
the band was not organized with a clear periodicity
(Fig. 2). Maximum coherence
between RSNA and MAP (0.69 ± 0.06 at 0.39 ± 0.02 Hz) was
significantly (P < 0.001) decreased compared with sham-operated
rats in the MF band. These effects of baroreceptor denervation on Mayer waves
and corresponding RSNA oscillations are illustrated in
Fig. 3. Finally, SAD induced a
large increase in the LF component of MAP variability, especially below 0.1
Hz, without altering the corresponding LF component of RSNA variability
(Fig. 2,
Table 2). In both groups of
rats, coherence was usually weak in the LF band. Average coherence calculated
below 0.1 Hz did not differ statistically between sham-operated (0.21 ±
0.02) and SAD (0.27 ± 0.04) rats. Transfer function analysis and
modeling. In sham-operated and SAD rats, the transfer function from RSNA
to MAP was studied up to 1 Hz, inasmuch as previous studies indicated that SNA
fluctuations are not translated into MAP fluctuations at frequencies beyond 1
Hz (18,
25,
32). In both groups of rats,
gain values declined with increasing frequency, especially beyond 0.1 Hz
(Fig. 4A). Below 0.1
Hz, gain values were significantly (P < 0.001) higher in SAD than
in sham-operated rats (0.243 ± 0.036 vs. 0.063 ± 0.004 mmHg/NU,
where NU is normalized units). At higher frequencies, the gain functions were
similar in both groups of rats. In particular, the gain values measured at the
peak coherence frequency were almost identical in sham-operated (0.045
± 0.004 mmHg/NU at 0.42 ± 0.01 Hz) and SAD (0.046 ± 0.004
mmHg/NU at 0.39 ± 0.02 Hz) rats.
|
|
|
|
In both groups of rats, phase values were negative and tended to decline,
especially starting from 0.2 Hz (Fig.
4B). Interestingly, the phase values measured at the peak
coherence frequency in the MF band were similar and close to 
rad (-2.82
± 0.07 and -2.85 ± 0.06 rad in sham-operated and SAD rats,
respectively), which indicates that, at this frequency, RSNA and MAP tended to
vary in opposite directions.
Assuming a simple linear relation between RSNA and MAP in SAD rats and in view of the previously reported low-pass filter properties of the vasculature (18, 21, 25), we attempted to fit equations of first-order (Eq. 1) and second-order (Eq. 2) low-pass filters to experimental gain values after transformation [G = 20 log(gain), expressed in dB]
 | (1) |
where K is the static gain (mmHg/NU) and fc is the corner frequency (Hz), and
 | (2) |
where 
is the damping coefficient and fn is the
natural frequency (Hz).
The fitting of equations used an iterative least-squares procedure (SYSTAT 8.0, SPSS, Chicago, IL). Only gain values associated with a significant coherence were used (n = 130 ± 11 in the 0.005- to 1-Hz frequency band).
Both models provided satisfactory fittings of the experimental data.
However, the regression coefficients (r2, observed vs.
predicted values) were always higher in the case of a second-order, low-pass
filter (0.931 ± 0.013 and 0.912 ± 0.014 for Eqs. 2 and
1, respectively). An individual example is shown in
Fig. 5A. The estimated
parameters obtained from the 10 SAD rats were 0.439 ± 0.086 mmHg/NU,
0.186 ± 0.023 Hz, and 2.25 ± 0.26 for K, fn,
and , respectively.
|
Theoretically, in a system including a low-pass filter with a fixed time
delay, the linear portion of the phase function mainly reflects the influence
of the time delay. Therefore, to estimate the time delay between RSNA and MAP
fluctuations, individual linear fittings of phase functions were performed in
the group of SAD rats. Because at low frequencies the influence of the time
delay is weak, phase values below 0.2 Hz were excluded. Linear regression
analysis (SYSTAT 8.0) was thus performed between 0.2 and 1 Hz, with
consideration of only phase values associated with a significant coherence
(r2 = 0.809 ± 0.038, n = 107 ± 8).
An individual example is presented in Fig.
5B. The time delay (
, s) was calculated from the
slope (rad/Hz) of the linear regression: = slope/2. The time delay
estimated in the group of SAD rats was 0.46 ± 0.03 s.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
SAD rats exhibited HF respiratory-related fluctuations of MAP and RSNA that did not differ in amplitude from those measured in sham-operated rats. This observation was not unexpected with respect to HF oscillations of MAP, which are essentially of mechanical origin in the rat (17). Regarding HF oscillations of RSNA, the results of the present study suggest that the baroreceptor reflex contributes little to their genesis and/or modulation. It has been shown in the anesthetized, vagotomized rat that the respiratory-related oscillations of RSNA have a mixed central and baroreflex origin (14). In the conscious SAD rat, it has been reported that HF oscillations of RSNA are absent most of the time but can appear sporadically (23). Inasmuch as we did not examine the stability over time of the HF RSNA fluctuations, we cannot exclude that, in some animals, this component was not always present and that sporadic large-amplitude fluctuations produced spectral power in the HF band.
In sham-operated rats, 0.4-Hz oscillations of AP (the so-called Mayer waves) were associated with large, coherent oscillations of RSNA, which confirms previous observations (4, 7). These oscillations were strongly attenuated in SAD rats, which is consistent with a recent report (23). The latter observation further supports the hypothesis that Mayer waves and accompanying RSNA oscillations are resonant oscillations within the baroreceptor reflex loop (2, 5). There was significant residual power in the MF band for MAP and RSNA in SAD rats, although no clear peaks could be discerned. This finding is in keeping with the previous observation that acute ganglionic blockade abolishes MF power of MAP in conscious SAD rats (10). The origin of this residual sympathetic rhythmicity is unclear. One possibility is that the surgical procedure used to produce SAD spared some baroreceptor fibers, which is indeed suggested by the persistence of reflex HR and RSNA responses to drug-induced changes in AP (31). Another possibility is that, in the MF band, endogenous RSNA rhythms of small amplitude and seemingly random frequency are generated by central nervous structures. Whatever the mechanism of their production may be, calculation of the coherence function indicated that these RSNA and MAP rhythms were linearly correlated in SAD rats. This finding is consistent with the proposal that MF fluctuations of AP are mainly, if not solely, mediated by the sympathetic nervous system (18, 25), with a possible modulation by the cyclic release of nitric oxide (28). Other factors known to alter AP, such as myogenic reactivity of resistance vessels or humoral systems activation, act with rather slow time constants and, thus, induce AP fluctuations at frequencies below the MF band (6, 11, 24).
The RSNA of SAD rats clearly exhibited slow fluctuations, such that LF
power contributed 10% of the total power. Such slow changes in RSNA can
be evoked as part of the response to environmental stressors
(29) or can accompany the
performance of natural behaviors
(27). The intriguing finding
of the present study is that LF power of RSNA did not significantly differ in
sham-operated and SAD rats. This can be taken to indicate that, in the
conscious baroreceptor-intact rat, baroreflex-induced changes in RSNA
contribute little to its slow variability. Alternatively, it is possible that
RSNA fluctuations evoked by the central command would be attenuated by the
baroreceptor reflex in such a way that it would then compensate for the
increase in RSNA variability directly caused by the reflex. An exaggeration of
vasoconstrictor responses to acute environmental stress has indeed been
demonstrated in the conscious SAD rat
(34). Finally, it is possible
that afferent renal nerves contributed to slow RSNA variability, which could
not be evaluated in the present study because the renal nerve was not cut
distal from the electrode.
Despite the unchanged LF power of RSNA in SAD rats, there was an eightfold
increase in the LF power of MAP, which indicates that sympathetic influences
were not responsible for the major part of the increased AP variability after
baroreceptor denervation. Accordingly, rank correlation analysis has revealed
that MAP and RSNA are positively correlated for only 25% of the time in
conscious chronically SAD rats
(1), whereas linear regression
analysis did not disclose any consistent relation between the two signals in
this model (15). In the
present study, although coherence values were usually low in the LF band, they
did reach significance in some instances. Therefore, using transfer gain
values that were associated with a significant coherence, we attempted to
characterize in SAD rats the feedforward effects of RSNA on MAP by means of
linear modeling analysis. In all cases, the equation of a second-order
low-pass filter could be satisfactorily fitted to experimental gain values.
Model parameters were in good agreement with those obtained for the transfer
function relating stimulation of the lumbar sympathetic chain and hindlimb
vascular conductance in the urethane-anesthetized rat
(4). In addition, with the use
of the linear portion of the phase function, it was possible to reliably
estimate a fixed time delay between RSNA and MAP changes. The 0.46-s value we
report here is close to that reported previously in the rat with use of
time-(29) or frequency-domain
(4) methods. Therefore, despite
a relatively weak coupling in the LF range, the modeling analysis was robust
enough to extract linear causality between RSNA and MAP after baroreceptor
denervation. However, 70% of the variance of MAP below 0.15 Hz could not
be explained by the variance of RSNA. This indicates that internal noise
sources (hemodynamic perturbations) unrelated to the sympathetic nervous
system powerfully affect AP in this frequency range. We and others have
provided evidence that transient changes in stroke volume, autoregulatory
responses of regional circulations, and active muscular vasodilatations are
important contributors to these slow hemodynamic perturbations
(20,
24,
35).
Linear systems analysis predicts that the input-output relations (i.e., the transfer function) between two signals in a control loop would be identical in the open- and closed-loop configurations, provided there are no independent noise sources affecting either signal (21). This theoretical assumption was verified in the MF band, where the gain functions were almost indistinguishable in sham-operated and SAD rats. As discussed above, this can be interpreted as sympathetic influences being predominant in this frequency range. Additionally, this indicates that vascular responsiveness to sympathetic stimulation is essentially unaltered in the SAD rat. On the contrary, the gain functions clearly differed at frequencies below 0.15 Hz, which can be explained by the strong interference of noise sources unrelated to RSNA in this frequency range. In the SAD rat, any slow hemodynamic perturbation increases MAP power and the gain of the RSNAMAP transfer function in the LF band. This is because the transfer gain at a given frequency is proportional to the square root of the ratio of MAP to RSNA powers, i.e., the ratio of the amplitude of MAP to RSNA fluctuations. In the baroreceptor-intact rat, the same hemodynamic perturbation induces an opposite change in RSNA through the baroreflex transfer function, which increases RSNA power. The RSNA variation then attenuates the impact of the hemodynamic perturbation on MAP power. Both effects result in a lowered transfer gain.
Perspectives
Although the feedforward influence of SNA on AP cannot be neglected, it explains only a small portion of slow AP variability after baroreceptor denervation in the conscious rat. The baroreceptor reflex limits AP variability in the LF range without increasing the SNA variability, which suggests that the reflex generates SNA fluctuations in response to AP perturbations but also limits the amplitude of SNA changes originating from the central command. We speculate that, in baroreceptor-intact rats, the weak coherence observed between SNA and AP in the LF range results from the continuous mixing of these two types of events.
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
-adrenoceptor blockade on the 0.4-Hz sympathetic rhythm in conscious
rats. Clin Exp Pharmacol Physiol
28: 983-985,
2001.[ISI][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
