INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

WITH USE OF PERIPHERAL thermal receptor stimulation and measurement of single nerve fiber activity, a unique group of rat renal sympathetic nerve fibers has been identified (5). These fibers are not spontaneously active and are not activated by stimulation of arterial baroreceptors or central or peripheral chemoreceptors but are activated by peripheral thermal receptor stimulation. These results indicate that rat renal sympathetic nerve fibers are heterogeneous and suggest the possibility that there are functionally specific groups of renal sympathetic nerve fibers.

Peripheral thermal receptor stimulation increases multifiber renal sympathetic nerve activity (RSNA) and produces renal vasoconstriction, which is prevented by renal denervation (19).

We sought to determine if increasing RSNA by peripheral thermal receptor stimulation (known to activate a unique group of rat renal sympathetic nerve fibers) produced a different renal vasoconstrictor response than that produced when RSNA is increased to the same magnitude by another stimulus, somatic receptor stimulation, which activates renal sympathetic nerve fibers uniformly, producing renal functional effects (2, 6, 7).

We studied anesthetized rats instrumented with an RSNA recording electrode on a left renal nerve branch and an electromagnetic flow probe on the left renal artery. We compared the responses of RSNA and renal blood flow (RBF) to peripheral thermal receptor stimulation and somatic receptor stimulation. Spectral analysis techniques were employed to examine responses in the frequency domain and to examine the impact of oscillations in RSNA on oscillations in RBF.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Sprague-Dawley rats of either sex, weighing between 250 and 300 g, were used for all experiments. They were maintained in individual cages for a minimum of 1 wk before experimentation and were given normal rat chow and tap water as drinking fluid.

Rats were anesthetized with pentobarbital sodium (50 mg/kg ip), an endotracheal tube was placed, and the rats were allowed to breath spontaneously. The right jugular vein and carotid artery were catheterized for administration of supplemental doses of anesthetic and isotonic saline at 50 µl/min and measurement of arterial pressure, respectively.

Via a midline abdominal incision the left renal nerves were identified and isolated. A small branch was dissected and placed on a bipolar platinum wire electrode. The recorded signal was led to a high impedance input and a band pass (30-1,000 Hz) amplifier (×20,000-50,000). The resulting signal was led to a Tektronix oscilloscope and a loudspeaker for visual and audio monitoring, respectively. The quality of the signal was evaluated by its signal to noise ratio (2-5/1), its pulse synchronous rhythmicity (bursting in diastole, relative silence in systole), and its response to both increases (norepinephrine iv) and decreases (nitroprusside iv) in arterial pressure. Then the nerve-electrode preparation was embedded in Wacker Sil-Gel, and the nerve was cut between the electrode and the kidney, assuring the recording of efferent RSNA. The remaining left renal nerves were left intact. Then a noncannulating electromagnetic flow probe was placed around the left renal artery. The bladder was drained via a catheter inserted in the bladder dome. One hour was allowed for equilibration after completion of surgery.

In another group of similarly prepared rats the remaining left renal nerves were sectioned to denervate the left kidney. This was verified at the end of the experiment by placing a bipolar stimulating electrode on the ipsilateral suprarenal portion of the lumbar sympathetic chain and stimulating at 20 V, 0.1 ms, and 5 Hz (Grass S88 stimulator). This produces a marked decrease in left RBF in innervated kidneys; complete renal denervation abolished this response.

Arterial pressure was measured with an electronic pressure transducer, yielding both pulsatile arterial pressure (PAP) and mean arterial pressure (MAP). Heart rate (HR) was measured with a cardiotachometer driven by the PAP wave form. RBF, both pulsatile (PRBF) and mean (MRBF), was measured with an electromagnetic flow meter. Renal vascular resistance (RVR) was electronically calculated as PAP/PRBF. The amplified filtered renal neurogram signal was full-wave rectified and integrated with a 20-ms time constant. The integrated RSNA (IRSNA) signal consisted of voltage peaks whose amplitude, duration, and frequency reflected the bursting characteristics of the amplified filtered renal neurogram. HR, PAP, MAP, PRBF, MRBF, and IRSNA signals were continuously recorded on a direct-writing recorder and on videotape via a pulse code modulation adapter throughout the duration of the experiment.

The experimental protocol commenced with a 30-min control period without any intervention. Then the first 5-min experimental period consisted of tail compression (pinch). Tail compression was standardized by using a standard force applied to the tail of the rat. A 30-min recovery period followed the first 5-min experimental period. Then the second 5-min experimental period consisted of external heat applied to the tail of the rat by inserting the tail in 53°C water (5, 11). A 30-min recovery period followed the second 5-min experimental period. In one-half the rats in each group the order of interventions was control, tail compression, external heat, and in the other one-half the order of interventions was reversed, i.e., control, external heat, tail compression. Respiratory frequency was not notably altered by either intervention. At the conclusion of the experiment the rat was killed with an overdose of pentobarbital sodium, and 30 min thereafter death signals were recorded for all variables.

Tape-recorded signals were sampled continuously at 1 kHz using an analog-to-digital converter in a Pentium IBM-compatible computer. All signals were corrected by subtraction of the death signals. Beat-to-beat values for HR, PAP, PRBF, and IRSNA were calculated for each entire period. For determination of mean values in each period the beat-to-beat values from the last 10 min of the control and the two postexperimental intervention recovery periods and the last 2 min of the tail compression and external heating periods were averaged (Tables 1 and 2).

The dynamic interactions of the fluctuations in IRSNA, PAP, and PRBF were investigated in the frequency domain using spectral analysis techniques (10, 17, 18; adapted for LabVIEW 4.1, National Instruments). The initial step involved analysis of 8,192 ms of original data to characterize oscillations over the frequency range of 0-10 Hz. The next step involved analysis of the oscillations that were slower than cardiac frequency and contained the majority of the spectral density power. Here, to facilitate a comparison between rats, the beat-to-beat values for PAP, PRBF, and IRSNA were expressed as percent of the absolute value in the control period (normalized units). Then, to generate equidistantly sampled data, the beat-to-beat data were converted by extracting the maximum value of each parameter from sequential 250-ms time periods. The resulting 4.0-Hz time series was divided into half-overlapping sequential blocks of 256 data points (frequency resolution 0.016 Hz). Each block was subjected to linear trend removal and cosine tapering of the first and last 60 data points before calculation of spectral density power. For each parameter spectral density power was calculated as the average over the sequential data sets in each rat.

The spectra over the frequency range below the cardiac frequency were further divided into three different frequency bands (8, 14). First, a respiration-related-frequency band containing oscillations related to respiration. The range of this band was 0.75-1.40 Hz to include respiration-related fluctuations during the various experimental conditions. Second, a mid-frequency band of 0.25-0.40 Hz containing oscillations related to sympathetic activity. Third, a low-frequency band of 0.04-0.25 Hz containing all undefined low-frequency oscillations.

To analyze the influence of fluctuations in IRSNA on the fluctuations in PRBF, the transfer function between IRSNA (input signal) and PRBF (output signal) was calculated as the quotient of the cross spectrum and the input spectrum. For each segment of 256 data points, the magnitude (gain) and phase of the transfer function were calculated. The values for transfer gain were normalized by the ratio of the steady-state values of the output (PRBF) and input signal (IRSNA). The normalized transfer gains reflect the ratio of the fractional variation in IRSNA and PRBF. A value of 1 indicates that an alteration in IRSNA of some fraction of the mean value is associated with (i.e., is transferred to) an alteration in PRBF of the same magnitude. A value <1 indicates that the magnitude of the alterations in IRSNA are transferred to a lesser extent to the alterations in PRBF. A value close to 0 indicates that fluctuations in PRBF are not influenced by fluctuations in IRSNA. A value of >1 suggests that the fluctuations in IRSNA are amplified within the renal vascular bed or that the fluctuations in PRBF are generated within the renal vascular bed itself. The phase of the transfer function reflects the temporal relationship between the input and output signals in the frequency domain with a difference of 0° (0 radians) indicating synchrony of the two signals and a difference of 180° ( radians) indicating a reciprocal relationship between the two signals. The cross correlation indicates the optimal shift between the input and output signals in the time domain. The coherence function yields a value that varies between 0 and 1 and is a frequency domain estimate of a linear regression coefficient, indicating the degree to which variance in one signal can be explained by a linear operation on the variance in the other signal. For each rat results from these analyses were averaged over the sequential data segments to reduce variance.

Statistical Analysis

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Rats With Intact Renal Innervation

Steady-state responses. Both tail compression and external heat increased MAP, HR, IRSNA, and RVR and decreased MRBF (Table 1). These changes were reversible, and measurements made during the recovery periods (recoveries 1 and 2) following each experimental intervention were not significantly different from those during the initial control period. The changes in all variables except IRSNA were significantly greater with external heat than with tail compression. Tail compression increased IRSNA by 77 ± 5%, decreased MRBF by 16 ± 2%, and increased RVR by 33 ± 3%. Although external heat increased IRSNA to the same extent as tail compression (83 ± 5%) the decrease in MRBF (24 ± 2%) and the increase in RVR (61 ± 3%) were greater.

Dynamic responses. Figure 1A shows 5-s traces of PAP, PRBF, and IRSNA sampled at 1 kHz in an anesthetized rat during control, tail compression, and external heating, and Fig. 1B shows the corresponding spectral density power. During the control period IRSNA bursts occurred synchronously with cardiac rhythm. Tail compression and external heating produced similar increases in IRSNA. This sympathetic activation was associated with increases in PAP and decreases in PRBF, which were greater during external heating than during tail compression.

View larger version (48K): [in this window] [in a new window]   View larger version (21K): [in this window] [in a new window]   Fig. 1.   A: records (5 s) of pulsatile arterial pressure (PAP), pulsatile renal blood flow (PRBF), and integrated renal sympathetic nerve activity (IRSNA) sampled at 1 kHz in anesthetized rat with intact renal innervation during control, tail compression (Pinch), and external heat (Heat). B: corresponding spectral density power calculated from records in A.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Mean cross correlation between IRSNA and PRBF during control, tail compression, and external heat in rats with intact renal innervation. SE bars omitted for clarity; cross correlation becomes significant at values more than 0.06 and less than 0.06; n = 10.

View larger version (22K): [in this window] [in a new window]   View larger version (14K): [in this window] [in a new window]   Fig. 3.   A: mean spectral density power in PAP, PRBF, and IRSNA during control, tail compression, and external heat in rats with intact renal innervation. SE bars omitted for clarity; statistically significant changes in spectral density power in frequency bands of interest are shown in Table 3; n = 10. B: mean spectral density power in IRSNA during control, tail compression, and external heat in rats with intact renal innervation. SE bars omitted for clarity; frequency range limited to those below respiratory frequency; n = 10.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Mean transfer gain in normalized units, phase, and coherence between IRSNA as input and PRBF as output during control, tail compression, and external heat in rats with intact renal innervation. SE bars omitted for clarity; n = 10.

Rats With Renal Denervation

Steady-state responses. Both tail compression and external heat increased MAP, HR, and IRSNA (Table 3). The magnitude of the increases in MAP, HR, and IRSNA in the rats with renal denervation was similar to those seen in rats with intact renal innervation. However, tail compression and external heat did not affect MRBF or RVR in rats with renal denervation. These changes were reversible, and measurements made during the recovery period after each experimental intervention (recoveries 1 and 2) were not significantly different from those during the initial control period.

Dynamic responses. Figure 5A shows 5-s traces of PAP, PRBF, and IRSNA sampled at 1 kHz in an anesthetized rat during control, tail compression, and external heating periods, and Fig. 5B shows the corresponding spectral density power. During the control period IRSNA bursts occurred synchronously with cardiac rhythm. Tail compression and external heating produced similar increases in IRSNA. This sympathetic activation was associated with increases in PAP, whereas PRBF was unchanged in the denervated kidney.

View larger version (55K): [in this window] [in a new window]   View larger version (19K): [in this window] [in a new window]   Fig. 5.   A: records (5 s) of PAP, PRBF, and IRSNA sampled at 1 kHz in anesthetized rat with renal denervation during control, tail compression, and external heat. B: corresponding spectral density power calculated from records in A.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Mean spectral density power in PAP and PRBF during control, tail compression, and external heat in rats with renal denervation. SE bars omitted for clarity; statistically significant changes in spectral density power in frequency bands of interest are shown in Table 4; n = 10.

Comparison of Rats With Intact Renal Innervation and Rats With Renal Denervation

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Mean spectral density power in PRBF during control, tail compression, and external heat in rats with intact renal innervation (thick solid lines) and in rats with renal denervation (thin solid line). SE bars omitted for clarity; n = 10.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Despite similar increases in overall IRSNA in response to the two different stimuli, the renal vasoconstrictor response was significantly greater during external heat than during tail compression. Peripheral thermal receptor stimulation is known to activate a unique group of rat renal sympathetic nerve fibers (5), whereas somatic receptor stimulation activates renal sympathetic nerve fibers more uniformly (2, 6, 7). This is reflected in the spectral density power of IRSNA (Fig. 3B). Compared with control, tail compression was associated with the appearance of oscillations in the frequency range of 0.3-0.4 Hz. During external heat these oscillations at 0.3-0.4 Hz persisted and there was the appearance of an additional separate oscillation centered at 0.2 Hz. Therefore at similar increases in overall IRSNA (Tables 1 and 3), spectral density power analysis disclosed different patterns, which distinguished tail compression from external heat.

These findings are also seen in PRBF spectral density power (Fig. 7). In the rats with intact renal innervation, compared with control, the renal vasoconstrictor response to tail compression was associated with the appearance of an oscillation at 0.3-0.4 Hz in PRBF spectral density power. This oscillation was also present during the renal vasoconstrictor response to external heat and there was the appearance of an additional separate oscillation centered at 0.2 Hz. Thus the greater renal vasoconstrictor response to external heat was associated with the appearance of separate oscillations at 0.2 and 0.3-0.4 Hz, whereas the lesser renal vasoconstrictor response to tail compression was associated with the appearance of a single oscillation only at 0.3-0.4 Hz. During tail compression and external heat, these newly appearing oscillations (as well as the progressive increase in the oscillations centered at 0.1 Hz) were dependent on IRSNA as they were not present after renal denervation. There was little effect of renal denervation on PRBF spectral density power during the control period.

This is further demonstrated in the gain of the transfer function (Fig. 4). During tail compression there was a significant increase in the gain at 0.3-0.4 Hz. During external heating there was a further significant increase in gain at 0.3-0.4 Hz and a significant increase in gain at 0.2 Hz. Thus the separate additional oscillations in IRSNA spectral power density (Fig. 3B), which appeared during tail compression and external heat, were substantially transferred into oscillations in PRBF spectral power density. The greater renal vasoconstriction produced by external heating, compared with tail compression, despite similar increases in overall IRSNA, was related to two factors: a twofold greater transfer of IRSNA power spectral density into PRBF power spectral density at 0.3-0.4 Hz and an additional transfer of IRSNA power spectral density into PRBF power spectral density centered near 0.2 Hz. Both of these changes in PRBF power spectral density were mediated by IRSNA, as revealed by their disappearance following renal denervation (Fig. 7). As the increases in IRSNA and PRBF power spectral density centered near 0.2 Hz were observed during external heating and not during tail compression, it is likely that this is the major factor contributing to the greater renal vasoconstriction observed during external heating.

Although the two interventions produced similar increases in overall IRSNA, the increases in HR and MAP were greater with external heating than with tail compression. Although efferent sympathetic nerve activity to the heart or other vascular beds was not measured, these results suggest that there may have been greater increases in sympathetic nerve activity to these areas during external heating than during tail compression. This would suggest that whereas the afferent pathways of the two stimuli centrally engaged efferent renal neuronal pathways in such a way as to produce a similar increase in IRSNA, this was not the case for efferent pathways to the heart and other vascular beds. This would allow a greater increase in efferent sympathetic nerve activity to the heart and other vascular beds with external heating, resulting in greater increases in HR and MAP with external heating than with tail compression.

The effect of IRSNA on RBF (linear) velocity (RBFV) has been analyzed in conscious rabbits using graded hypoxia to reflexly increase RSNA (8). With moderate hypoxia increasing IRSNA by 90% and decreasing RBFV by 18%, 0.3-Hz coherent oscillations of IRSNA and RBFV were found; these were prevented by renal denervation. The 0.3-Hz coherent oscillations were not found with mild hypoxia, which only increased IRSNA by 13% and decreased RBFV by 9%. During moderate hypoxia the new oscillation in IRSNA was monomorphic and spanned the frequency range from 0.2 to slightly <0.5 Hz, whereas the coherent oscillation in RBFV (also monomorphic) spanned the frequency range from 0.1 to 0.4 Hz. Transfer gain showed a distinct peak between 0.2 and 0.3 Hz. The frequency ranges of these findings in rabbits overlap with those found herein in rats. However, a major difference is that the interventions used to increase IRSNA produced different responses in IRSNA and RBFV power spectral density. In the rabbit hypoxia resulted in a monomorphic oscillation. However, in the rat, whereas tail compression resulted in a monomorphic oscillation (0.3-0.4 Hz), external heat elicited two separate oscillations with peaks at 0.2 Hz and at 0.3-0.4 Hz. It is likely that this difference is related to the activation of a unique group of renal sympathetic nerve fibers in the rat by external heat, which in turn accounts for the differences in renal hemodynamic responses.

Values for transfer gain decreased rapidly from values of 0.2-0.3 at frequencies <0.2 Hz to values 0.1 at frequencies above 0.2 Hz. Thus at <0.2 Hz, only 20-30% of the amplitude in IRSNA oscillations was transmitted to the corresponding PRBF oscillations; at >0.2 Hz it was ~10%. During tail compression there were 0.3- to 0.4-Hz coherent oscillations in IRSNA and PRBF; the PRBF oscillation was eliminated by renal denervation. Although there was a threefold increase over control in transfer gain at 0.3-0.4 Hz, transfer gain at >0.5 Hz was 0.2. During external heat there were two separate coherent oscillations in IRSNA and PRBF at 0.2 Hz and at 0.3-0.4 Hz; the PRBF oscillations at these frequencies were eliminated by renal denervation. Transfer gain was substantially elevated in both these frequency ranges but was 0.25 at >0.5 Hz.

Therefore, these studies indicate that rat renal (as well as mesenteric; 20, 21) resistance vessels, like those in the rabbit (8, 14), act as a low-pass filter for oscillations in IRSNA. At frequencies >0.5 Hz, the transfer gain is relatively constant at 0.1-0.2, indicating that the renal resistance vessels dampen these oscillations by a factor of 5-10. At frequencies <0.5 Hz, the transfer gain increases with decreasing frequency, allowing a greater transmission of the oscillations in IRSNA to PRBF. In response to specific interventions that increase IRSNA and decrease RBF, oscillations in IRSNA at <0.5 Hz appear that are passed to PRBF. These IRSNA oscillations can be intervention specific (tail compression at 0.3-0.4 Hz; external heat at both 0.2 and at 0.3-0.4 Hz) and contribute to different magnitudes of renal vasoconstriction despite similar increases in overall IRSNA.

The effect of IRSNA on volumetric PRBF has been analyzed in conscious rabbits subjected to hemorrhage to reflexly increase RSNA (14). In the control state before hemorrhage, an oscillation at 0.3 Hz was observed in IRSNA but this was not passed to PRBF; however, during hemorrhage before the onset of hypotension, the strength of oscillations at 0.3 Hz in IRSNA increased and this was passed to PRBF at the same frequency. In the control state, when analyzing the relationship between PAP and PRBF, both the transfer function gain and coherence were lower in innervated than denervated kidneys. These results were interpreted to mean that the control level of IRSNA, even though its 0.3-Hz oscillation was not passed to PRBF, was of sufficient magnitude to dampen the effect of oscillations in PAP on PRBF and that there was tighter coupling between PAP and PRBF in the absence of IRSNA. These interpretations are heavily dependent on the absolute magnitude of the control level of IRSNA in these conscious chronically instrumented rabbits. Because control MRBF was 55% greater in renal denervated compared with innervated rabbit kidneys, this would indicate that the control level of IRSNA was relatively high. In fact, in the previous conscious rabbit studies (8), because a modest degree of hypoxia that increased IRSNA by 90% reduced RBFV by only 18%, this suggests that the control level of IRSNA must have been substantially greater to account for an increase of 55% in MRBF following renal denervation. Thus it would appear that the absolute magnitude of the control level of IRSNA was relatively high in these conscious rabbits, sufficient to dampen the effect of oscillations in PAP on PRBF and diminish the coupling between PAP and PRBF.

In contrast with the clear effects of renal denervation on power spectral density of PRBF during tail compression and external heat in these rat studies, there was little effect of renal denervation on power spectral density of PRBF during control. Thus, whereas the control levels of IRSNA contain oscillations at distinct frequencies, these do not exert effects on PRBF at similar frequencies. After the initial transient increase in PRBF and decrease in RVR immediately following the acute renal denervation, the steady-state values of PRBF and RVR were slightly higher but not significantly different from those in the innervated kidneys. These several observations would suggest that the absolute magnitude of the control level of IRSNA in these rat studies was not markedly elevated because its removal by renal denervation had no steady-state effect on the renal vasculature. When the relationship between PAP and PRBF was analyzed, both the transfer function gain and coherence were slightly but not significantly lower in the renal-denervated compared with the -innervated state. That these results differ from those in rabbit studies noted above likely relates to differences in the absolute magnitude of the control level of IRSNA. Based on the steady-state renal vascular response to renal denervation, the 55% increase in MRBF in the rabbit studies (14) compared with no significant change in the rat studies indicates a greater absolute magnitude of the control level of IRSNA in the rabbit studies compared with the rat studies.

Despite similar increases in IRSNA during tail compression and external heating, power spectral analysis disclosed differences in the inherent patterns of IRSNA. These differences were of physiological importance as reflected in the difference in the magnitudes of the renal vasoconstrictor responses. It is likely that these findings reflect functional differences in renal sympathetic nerve fiber groups, e.g., relative vasoconstrictor activity. Thus external heat activated a renal sympathetic nerve fiber group that was not activated during tail compression and that mediated an enhanced renal vasoconstrictor response. This could be a manifestation of recruitment of a group of renal sympathetic nerve fibers with preferential or greater density of innervation of certain segments of the renal resistance vasculature (13). In this regard our previous studies (3) analyzing synchronized renal sympathetic discharges demonstrated that external heat increases IRSNA by increasing the amplitude of discharges without increasing the frequency of discharges. Increasing amplitude reflects the recruitment of additional (previously silent) nerve fibers (16), whereas frequency reflects the inherent generation of discharges by the central nervous system and their modulation by baroreceptor inputs (12, 15).

The strategy of measuring IRSNA in a renal nerve branch and PRBF of the entire kidney provided two advantages: 1) elimination of any contribution of afferent renal nerve activity to the overall IRSNA (efferent RSNA) signal that was recorded and 2) retained ability to compare the IRSNA responses in rats with intact renal innervation and renal denervation. However, this approach of using less than the entire population of renal sympathetic nerves to measure IRSNA may have underestimated the magnitude of the dynamic interaction between IRSNA and PRBF.

In conclusion, peripheral thermal receptor stimulation activates a unique group of previously silent renal sympathetic nerve fibers whose activity contains oscillations at distinct frequencies, which are transmitted to RBF at these same frequencies. The functional effect is manifest as a greater degree of renal vasoconstriction than that observed with another stimulus that, while increasing overall RSNA to the same degree, does not activate this unique group of renal sympathetic nerve fibers.

Perspectives