INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE KIDNEY RECEIVES A DENSE innervation of sympathetic nerves, and changes in the activity present in these nerves affects renal hemodynamics, sodium excretion, and renin release (9). Early studies, however, suggested that the vasculature was relatively insensitive to small changes in renal sympathetic nerve activity (RSNA). In past experiments RSNA was progressively increased, using either electrical stimulation of renal nerves in anesthetized cats or dogs (3, 10, 23) or reflex activation in conscious dogs (19, 21). It was found that at low RSNA levels only renin secretion was affected, then with slightly higher levels of activity changes in sodium excretion occurred, but only with even greater levels was renal blood flow (RBF) affected. This led to the concept that in daily life changes in RSNA around resting levels would have little impact on RBF (4). However, this concept has not been without challenge; Grady and Bullivant (8) measured RBF in conscious rats in their home cages during various activities ranging from sleep to active movement and found RBF decreased as the activity level of the rats increased, an effect negated by prior blocking of RSNA with local anesthetic. In the conscious rabbit a range of studies has suggested that a modest increase or decrease in RSNA does affect RBF. Air jet stress, noise stress, and hypoxia, each of which increased RSNA between 12 and 31%, produced RBF decreases of ~8-12% from control levels (16). Conversely blood volume expansion that decreased RSNA 25% caused an increase in RBF of 17% (13). In each case the dynamic changes in RSNA with the onset of the stimuli were mirrored by changes in RBF and responses were absent in renally denervated animals.

One criticism of the study by Grady and Bullivant (8) was that because RSNA was not measured directly it was not possible to determine to what degree RSNA was altered during changes in the activity levels of the rats. This is important because the debate is not whether the renal nerves can affect blood flow but rather whether small physiological changes in RSNA around resting levels alter RBF. With regard to previous rabbit studies (13, 16), although experiments have been performed in conscious animals, the experiments have always been conducted in a laboratory environment in which it is likely baseline RSNA was higher than would be found in the home-cage environment. Thus, although resting RBF was higher in renally denervated rabbits compared with innervated animals (17), it could be argued that the baseline RSNA levels were high enough to impact the renal vasculature under the baseline conditions of the experiment. To fully test the hypothesis that physiological changes in RSNA play a role in regulating the resting level of RBF and counter the criticisms of previous studies, it is necessary to develop a new approach that allows continuous recording of RSNA and RBF in animals in their home-cage environment. It is proposed that knowledge of the degree of control of RBF by RSNA is fundamental in understanding how diseases such as hypertension and heart failure may involve the sympathetic nervous system in their pathogenesis. Because relatively long-term recordings of RSNA (up to 4 wk) in the rabbit are possible, this may provide the framework for investigating the link between sympathetic activity and the long-term control of arterial pressure (2).

To characterize the effect of the renal nerves under daily life conditions we have undertaken continuous recordings of RSNA, arterial pressure, and RBF for a 4-wk period in rabbits. Because resting RBF can be quite different between animals, in each animal we have measured RBF to each kidney, one being innervated the other denervated. Thus the arterial pressure and circulating hormone levels seen in each kidney will be the same. We have assessed the spontaneous and circadian changes in each parameter in testing the hypothesis that the renal nerves play a role in regulating RBF in daily life.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal preparation. Experiments were conducted in eight New Zealand White rabbits with initial weights of 2.4-3.5 kg and were approved by the University of Auckland Animal Ethics Committee. For all surgical procedures anesthesia was induced using intravenous administration of propofol (Diprivan, 10 mg/kg) followed by intubation and then maintenance with halothane. Arterial pressure was recorded throughout the study via a radiotelemetry transmitter (model PA-D70, Data Sciences International, St Paul, MN). This was implanted via an abdominal incision, and the area around the iliac bifurcation was exposed. The cannula of the transmitter was inserted into the left iliac artery and advanced so that the tip of the catheter lay in the abdominal aorta ~1 cm above the iliac bifurcation but well below the renal artery. The cannula was fixed in position with a drop of tissue glue, and the body of the transmitter was placed in the abdominal cavity. The incision was closed, and the animal was allowed to recover for 1 wk.

Data collection. Once rabbits were returned to their home cages, the flow probe and RSNA recording wires were immediately connected to the relevant amplifiers via a swivel system (Airflyte Electronics, Bayonne, NJ). This system allowed rabbits a complete range of movement. Initially locomotor activity was recorded by displacement of the blood pressure transmitter located in the rabbit's abdomen relative to the blood pressure receiver. In preliminary experiments we found this system recorded only large movements of the rabbit. Therefore we later developed a purpose-built activity monitor that used an infrared detector (Optex RX-40QZ, Otsu, Japan) to detect all animal movements. Data collection began within 3 h of completion of surgery. RSNA was further amplified outside the animal between 20 and 200 times to give a final gain between 20,000 and 200,000, filtered between 50 and 2,000 Hz, full-wave rectified and integrated using a low-pass filter with a time constant of 20 ms. With chronic continuous recording conditions it is not possible to experimentally determine a "zero" baseline RSNA level, e.g., using ganglionic blockade. However, the approach of rectification and integration of the original RSNA signal allowed us to automatically subtract the baseline noise. The integration of RSNA over 20 ms reflects the bursts of RSNA as a series of waves or peaks, and between these bursts there are silent periods where the integrated level drops to baseline. Our data acquisition program automatically set the background noise, which is of a constant magnitude, to zero. This integrated RSNA signal, the RBF waveforms via transit time flowmetry (Transonic Systems), and arterial pressure were sampled at 500 Hz using an analog-to-digital data acquisition card (AT-MIO64E-3 National Instruments, Austin, TX). All subsequent data collection and analysis were performed using a purpose-written data acquisition program developed using the LabVIEW graphical programming language (National Instruments). In brief, calibrated signals were continuously displayed on a computer screen both as the pulsatile waveforms and as averaged data. Previously collected data could be viewed online from minutes to days. Although data was sampled at 500 Hz, the length of the experiment (28 days) did not permit continuous saving of all of these data. Two-second averages of all parameters were saved continuously to disk. In addition 10-min periods of the 500-Hz data were automatically saved every 2 h to record the quality of the RSNA and RBF signals. Because absolute microvolt levels of RSNA cannot be compared between animals, we converted RSNA to a percentage of the average microvolt values obtained from a 4- to 10-day period at least 11 days after surgery. The mean microvolt value for this period was defined as 100%. We regularly validated our RSNA signal by triggering off the systolic pressure and constructing averaged RSNA for 1-s periods (Fig. 1). In general this cardiac-related signal was apparent for at least 21 days postoperatively (maximum 32 days).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Averaged 1-s sweeps of integrated renal sympathetic nerve activity (RSNA, dark line) and arterial pressure (light line) starting from each consecutive systolic pressure for a 76-s period of recording. Top: 3 days after surgery. Bottom: 12 days after surgery in a single rabbit. To construct the sweeps the systolic pressure for each heart beat was detected and the average RSNA signal for the following second was recorded (22). In this animal the bursts of RSNA remained related to the cardiac cycle until day 24.

Statistical analysis. Data are presented as mean ± SE. Unless otherwise stated data were compared using paired two-tailed t-tests. To examine relationships between different parameters a coefficient of linear correlation was calculated. To compare the regression lines between intact and denervated kidneys a linear mixed model was used because it allows the simultaneous modeling of the fixed effects (differences between slopes or intercepts) and random effects (estimates of variances). In this case the fixed factor is either the innervated or the denervated kidney, while the random factor is the rabbit. The subroutine PROC MIXED in SAS version 8 was used to fit the models. Preliminary visual examination of the data showed consistent heterogeneity of residual variances between regression lines for the innervated and denervated RBF and renal resistance within rabbits. This was accommodated by allowing separate error variances for innervated and denervated kidneys. A significance test of the null hypothesis of no difference between the residual variances was performed using a likelihood ratio test with a single degree of freedom (14).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Circadian profile. In each rabbit blood pressure, heart rate activity, RSNA, and RBF to one innervated and one denervated kidney were recorded continuously for up to 32 days. In six of the eight rabbits the heart rate and blood pressure gradually fell over the first 5-6 days after surgery. In all of the animals where RSNA was recorded (n = 4) a gradual reduction also was seen during the 5-day period after surgery. Although RBF often took a few days to reach a stable level, no reliable pattern was observed; in some animals RBF fell over the first few days, in others it increased, and no consistent differences in the time taken for RBF to settle between innervated and denervated kidneys were observed either. In general the circadian variation in the various parameters did not develop until 3-6 days after surgery (Fig. 2). In every rabbit a circadian rhythm was present in both RBFs and in heart rate. In four of eight rabbits a rhythm was also clearly seen in blood pressure (Fig. 3A). In the other four rabbits, although a circadian variation was seen bilaterally in RBF and a smaller one in heart rate, no clear pattern could be identified in arterial pressure (Fig. 3B). The presence of a circadian rhythm in mean arterial pressure (MAP) appeared to be correlated with the presence of a circadian rhythm in RSNA. In the four rabbits in which RSNA was directly recorded, two displayed a strong circadian rhythm in both RSNA and blood pressure; the other two had no discernible rhythm in either RSNA or blood pressure. Locomotor activity recorded via the infrared detector displayed a circadian rhythm that mirrored the pattern in the blood flows. The mean circadian profile for the eight rabbits, with the first 10 days of recording in each rabbit eliminated, is shown in Fig. 3C. The circadian profile was characterized by an increase in each parameter beginning at 0900 (Fig. 3C), with maximum values in each parameter occurring at approximately midday (Table 1), whereupon slow decreases were seen over the next 12 h. The period from midnight until feeding the next day was characterized by relatively stable levels in each parameter. Relative to the mean levels of each parameter, the range of values recorded throughout the day was extensive (Table 1); for example, the range for heart rate was 79 ± 10 beats/min. Each rabbit generally consumed the food within a 120-min period. To confirm the origin of the circadian rhythm, the feeding time was shifted to 1500 in three rabbits for a 5-day period. This caused an abrupt shift in the circadian profile in all parameters with the increase now occurring at 1500 (Fig. 4).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Data from a single rabbit for each measured parameter for the 5 days immediately after surgery (A) and between days 23 and 26 after surgery (B). On average the circadian rhythm in various parameters did not develop until 4 days after surgery. HR, heart rate; MAP, mean arterial pressure; RSNA, expressed as a percentage of the 24-h average from that animal between days 21 and 27 (the same days were taken for determining the 100% level in the circadian analysis); RBF intact, renal blood flow of the innervated kidney; RBF denerv, RBF of the denervated kidney; activity, locomotor activity, arbitrary values; bpm, beats/min.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Average changes in each parameter over 24 h taken from a period of 4-11 consecutive days at least 10 days after surgery. A: from a single rabbit with a prominent rhythm in all parameters. B: from a single animal with a weak rhythm in HR and an absence of rhythms in MAP and RSNA. C: group response for all 8 rabbits except for RSNA where n = 4. Values are means ± SD presented as the thin lines on either side of the mean line. The dark bar above time scale shows dark (night) period with the light section of the bar the light period (0600-1800 h). The vertical dotted line indicates the feeding time (0900 h).

View larger version (25K): [in this window] [in a new window]   Fig. 4.   The effect of shifting the feeding time on the circadian rhythm in a single rabbit. , data obtained with feeding at 0900 (vertical dashed line); , feeding at 1500 (vertical dotted line). Unit abbreviations are same as for Fig. 2. Similar responses were obtained from the other 2 rabbits that also underwent a shift in feeding time.

Comparison between RBF to innervated and denervated kidneys. When looking at the group data, we found that the RBF to innervated and denervated kidneys was not significantly different (Table 1). Immediately after feeding, an increase in both RBF and MAP was observed, and at the same time renal resistance decreased. A decrease in resistance at the same time as an increase in pressure suggests that the increase in RBF must have been proportionally greater than the increase in MAP. Surprisingly, this decrease in resistance coincided with the increase in RSNA (Fig. 3). Renal resistance in the innervated kidney and the denervated kidney was not significantly different (Fig. 5).

View larger version (17K): [in this window] [in a new window]   Fig. 5.   The circadian variation in renal resistance of the innervated and denervated kidneys (n = 8). , resistance of the innervated kidney and SE (dashed line). , resistance of the denervated kidney and SE (solid line). Resistance was not significantly different between the groups. Vertical dotted line indicates the feeding time.

View larger version (37K): [in this window] [in a new window]   Fig. 6.   The effect of episodic elevations on cardiovascular parameters between 0030 and 0330 in a single animal. Black boxes beneath HR indicate the automatically detected periods of episodic elevations. Resist intact, resistance in the innervated kidney; resist denerv, resistance in the denervated kidney. Other abbreviations as in Fig. 2.

View larger version (59K): [in this window] [in a new window]   Fig. 7.   The frequency of coinciding HR and RBF values over a 24-h period in one rabbit. The RBF to the innervated kidney is shown. A: 2-dimensional density diagram; B: same data presented in 3-dimensional fashion. Four distinct preferred levels were observed; these peaks corresponded with the baseline values recorded during night (HR, 220 beats/min; RBF, 45 ml/min) and day (HR, 270 beats/min; RBF, 47 ml/min), and during the episodic elevations during the night (HR, 250 beats/min; RBF, 59 ml/min) and day (HR, 300 beats/min; RBF, 58 ml/min).

View larger version (28K): [in this window] [in a new window]   Fig. 8.   Change in RBF to the innervated and denervated kidney during episodic elevations. Data obtained from the innervated kidney () and the denervated kidney (). Data includes all elevations for a 6-day-period in a single animal, n = 321 elevations. A: change in flow vs. the change in arterial pressure during each elevation. B: distribution of the change in flow in response to each episodic elevation. The change in flow to the denervated kidney showed a significantly narrower distribution than the change in flow to the innervated kidney (P < 0.001 using Brandt-Snedecor 2 test for comparison of arbitrary distributions). This same pattern was seen in all 8 animals. Thus in response to an episodic elevation the change in RBF in the presence of RSNA was more variable than without RSNA.

View larger version (22K): [in this window] [in a new window]   Fig. 9.   The change in renal resistance vs. change in MAP during each episodic elevation in a single rabbit. Data obtained from the innervated kidney () and the denervated kidney (). Data includes all elevations for a 6-day-period, n = 321 elevations. In the absence of RSNA renal resistance changes during the elevations showed a stronger correlation to the changes in arterial pressure. This effect was seen in every rabbit.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have developed a system for the long-term continuous monitoring of cardiovascular parameters in rabbits living in home cages that allowed us to test the hypothesis that the renal nerves play a role in regulating RBF in daily life. By comparing RBF to an innervated and denervated kidney in the same rabbit we found that mean levels of RBF and renal resistance were not different but that the control around this mean level was significantly different. In particular we observed episodic elevations in renal resistance, heart rate, and MAP. In the denervated kidney the increase in renal resistance that occurred with these episodic elevations closely followed the increase in arterial pressure. In contrast the renal resistance to the innervated kidney displayed a more variable response, indicating an interaction of the sympathetic nervous system.

In previous experiments in which we have recorded RBF in conscious rabbits in a laboratory setting, we have observed a significant difference in the baseline blood flows to innervated and denervated kidneys, although in these studies within-animal comparisons were not made (16, 17). In those studies we presented a case for the resting levels of RSNA modulating the level of RBF. Although RSNA was recorded directly and changes in RSNA and RBF during a variety of stimuli reported, it was not possible to ascertain whether the resting level of RSNA was higher than would be found in the home-cage environment. In the present study RBF to the innervated kidney was higher than previously reported (16), especially during daytime hours and there was no difference in RBF to the innervated and denervated side. These results suggest that in previous studies the "resting" RSNA levels were indeed higher than that found in the home-cage environment. Furthermore, the similar RBF to the innervated and denervated kidneys under resting conditions in this study suggests that baseline RSNA is usually very low.

We believe an important strength of our experimental design was recording RBF bilaterally with one kidney having the renal nerves intact and the other denervated. As we observed sizable differences in mean RBF between animals (range 23-57 ml/min) a between-animal design may have failed to detect any effect of RSNA on the control of RBF. This within- animal design ensured that each kidney was subjected to the same perfusion pressure and circulating levels of hormones, thus any difference in RBF or renal resistance between the kidneys was likely due to the action of the renal nerves.

We observed a decrease in renal resistance after feeding, although both RBF and arterial pressure increased at this time. Thus the increase in RBF was proportionally greater than the increase in arterial pressure. Interestingly this time was also when the maximum RSNA was recorded. Because increased RSNA would be expected to be associated with increased renal resistance, doesn't this suggest that the renal nerves were not affecting RBF? We suggest that RSNA is simply one of a host of factors that regulate RBF and that other factors such as circulating hormonal and metabolite levels clearly dominate in controlling mean RBF at this time. For example, it has been shown that the hormone secretin, released in the small intestine in response to an increase in acidity, can cause renal vasodilation (18). It might be suggested that because we did not find differences in mean RBF between kidneys, RSNA was not affecting the renal vasculature but may have been selectively affecting renin secretion or sodium excretion. Again we do not believe that this is so. It must be remembered that the calculation of the circadian rhythm in all parameters was determined from averaging data collected over many days. Thus, unless RSNA was playing a major role in setting the mean level of RBF, we would be unlikely to observe its effect in the circadian analysis because of the range of additional factors e.g., circulating hormones, that also affect RBF. Rather to discern the effect of RSNA we believe that it is necessary to examine the control around the mean level.

In each rabbit we observed episodic elevations in a number of parameters, most noticeably heart rate. These elevations occurred throughout the day and night and were associated with movement. With regard to heart rate we observed almost two states, low and high, with rapid shifts between these levels. These episodic elevations appeared similar to the cardiovascular changes occurring with exercise; that is, heart rate, blood pressure, and sympathetic activity all increased. Although we did not measure cardiac output we suggest it is likely that the increase in blood pressure was predominantly due to an increase in cardiac output. Why heart rate should "jump" between such states is unknown. It would appear that these episodic elevations are not unique to the rabbit as there are reports of similar spontaneous changes in RBF in the conscious dog (24).

The episodic elevations allowed us to examine neural control over the renal vasculature under spontaneous conditions. We believe that the data collected under such conditions strengthen our hypothesis that changes in RSNA from resting levels do regulate RBF. Although RSNA can be measured during a range of stimuli, it is always referred to as a percentage change. This makes it difficult to reference the level of activation obtained during any stimuli to the home-cage environment and the changes that occur in daily life. Analysis of RBF control during the episodic elevations indicated that RSNA was affecting RBF under these conditions. In particular in the presence of RSNA the RBF response was more variable than that in the denervated side during the episodic elevations. This larger variability indicates the presence of an extra input in the control of RBF (17). Thus we suggest that the renal nerves are actively regulating the renal vasculature in daily life. It will be noted that the effect of RSNA on RBF is small. However, we do not believe that the magnitude of the change in RBF is the important aspect but rather that we were actually able to observe any effect of RSNA. Previously it had been proposed that the renal vasculature was relatively insensitive to changes in RSNA (4). Electrical stimulation of the nerves in anesthetized dogs (10, 23) or reflex activation using hemorrhage in conscious dogs failed to see any change in RBF during low levels of sympathetic activation (21). We believe our observations indicate that with any increase in RSNA levels, the renal vasculature will be affected. Such information is important when considering that a number of disease states are associated with increased levels of sympathetic nerve activity (1, 6, 7, 12, 20) and may be important in understanding the pathogenesis of altered renal function in these conditions.

It is likely that not only the home-cage environment but also the time for recovery after surgery is important when examining resting RSNA. In the present experiment we observed elevated RSNA and heart rate over the first 4- to 5-day-period following surgery; a similar time was also required before the circadian rhythm became evident. These results suggest that in rabbits any experiments performed immediately after surgery are not representative of the resting condition and tend to overestimate RSNA and heart rate.

Limitations. We used renal denervation to remove the effect of RSNA on the kidney. Previously we have shown that 3 wk after denervation the norepinephrine content of the denervated kidneys was only 2% that of the innervated kidneys (11). Furthermore the process of implanting the flow probe and nerve recording electrode did not change norepinephrine content compared with nonexperimented animals. Thus we suggest that it is unlikely for functional reinnervation of the kidney to have occurred during the experiment. We cannot discount the possibility that after denervation the kidney became supersensitive to circulating norepinephrine levels, although other research has indicated that this does not occur (15). If this was a factor, it would tend to have reduced our ability to determine the effect of the renal nerves, and the true effect on the renal vasculature might be larger than actually recorded. As we outlined previously we believe that our unilateral renal denervation preparation was the most appropriate control because this meant that each kidney saw the same perfusion pressure and circulating hormonal levels. However, it should be acknowledged that other factors may have further reduced the differences in RBF between the innervated and denervated kidneys. These factors include the neurally mediated renin release from the innervated kidney and the ensuing angiotensin II level, which could have altered RBF in the denervated kidney. Thus the effect of the renal nerves may indeed be larger than we have shown. It should also be noted that in denervating a kidney, not only have we removed the efferent sympathetic nerve fibers but also any afferent nerves fibers. Although it is clear that renal afferents play a role in sensing the composition and volume of urine (4), it is unclear as to their effect for the long-term control of RBF around the resting levels and thus they cannot be ignored at this stage.

Perspective