1 Introduction
THE KIDNEYand its neural control have been under investigation for more than a hundred years (BERNARD, 1859). The kidney is richly innervated, with terminals in the arteries, the afferent and efferent arterioles, the arterioles of the vasa recta, the proximal and distal tubules and the juxtaglomerular apparatus (McKENNA and ANGELAKOS, 1968; BARAJAS,1978; 1981). The innervation is complex. Multiple varicosities of the same axon terminate on or near smooth muscle cells of the afferent and efferent arterioles, near the distal tubule, and near the macula densa (BARAJAS, 1978). Blood flow, filtration and renin release may be controlled simultaneously by the same nerve fibre (BARAJAS, 1978). Stimulation of the renal sympathetic nerves of rats with low-frequency trains of pulses (2 Hz) produced little or no variation in glomerular pressure although blood flow Correspondence should be addressed to Dr Spelman at the University of Washington. First received 3rd July 1989 and in final form 1st March 1990 9 IFMBE: 1991
Medical & Biological Engineering & Computing
decreased (HERMANSSONet al., 1981), suggesting that the afferent and efferent arterioles constrict together. In some cases, both glomerular pressure and blood flow decreased with stimulation and, about ! s after stimulation, the pressure was maintained slightly below the prestimulus level. This finding, and similar results in another study (KeN and ICHIKAWA, 1983), suggest that the efferent arteriole constricts after the afferent arteriole. Neither study showed whether the increased vasoconstriction resulted directly from stimulation of the nerve or indirectly from renin release, which leads to change in vascular resistance. We do know, however, that the change in renal excretion during stimulation of the renal nerve is not decreased by inhibiting the production of angiotensin II (ZAMBRASKI and DIBONA, 1976; DIBONA et al., 1977). A number of investigators have measured the effects of sympathetic nerve stimulation on blood flow both in the renal artery and within the kidney. DISALVO and FELL (1971) stimulated the renal sympathetic nerves in dogs and found that renal arterial flow decreased most significantly at frequencies of 20 Hz and less so at frequencies above and below 20 Hz. HERMANSSON et al. (1984) stimulated the
March 1991
121
renal nerve in rats and found that total renal flow, cortical flow and medullary flow all decreased with increasing frequency of stimulation, although medullary flow decreased more rapidly at frequencies of > 2 H z than at 2-10Hz. Total renal blood flow decreased linearly with increasing frequency of stimulation. Hermansson et al. used the rubidium extraction method to estimate flow, and were limited to making a single observation at a single time during the stimulation period. To our knowledge, no one has measured blood flow simultaneously and continuously in the renal artery and cortex while stimulating the renal nerve. STERN et al. (1979) used laser Doppler spectroscopy to estimate blood flow in the renal cortex and medulla of the rat. While they did not show continuous measurements of flow, the technique correlated well with electromagnetic flowmetry and radioactive microspheres during aortic constriction or the administration of vasoactive drugs. There were some indications of nonlinearities in the measurement (see Fig. 2 in STERX et al. 1979), which NILSSON (1984) discussed and eliminated. It appears that laser Doppler flowmetry should be useful to discover the effects of renal nerve stimulation on regional flows within the kidney. Because the temporal action of segments of the renal circulation in response to renal sympathetic nerve stimulation is unclear, we attempted to answer the following questions:
nerve was usually imbedded in adipose tissue, making it difficult to see. Moving from the diaphragm caudally near the aorta, we found the renal nerve distal to the aorticorenal ganglion. To verify that it was the renal nerve, we placed it on a pair of platinum hook electrodes and stimulated it; if the kidney blanched, we knew we had the renal nerve. Then we exposed the renal artery and fitted it with an electromagnetic flow transducer (2.5mmID, Zepeda Instruments, Seattle, Washington, USA); we also placed a snare occluder on the artery, distal to the flow transducer, to obtain flow zeros. To avoid respiratory movement artefacts, especially in the cortical flow measurement, we placed the kidney on tapes and supported it from a ringstand. We placed the fibre-optic probe of the laser Doppler flowmeter (Periflux, Perimed, Stockholm, Sweden) in a small plastic fitting that rested on the surface of the exposed kidney. The fitting was held in place by a segment of latex, which was wetted with saline and draped over the surface of the kidney (Fig. 1). Then we covered the incision with damp surgical sponges and allowed the animal to rest for at least an hour before making measurements.
(a) What is the temporal change of renal arterial and cortical blood flow during stimulation at 5, 10 and 50 Hz? (b) Do the afferent and efferent arterioles act simultaneously or at different times during stimulation ? (c) Does stimulation cause a direct or an indirect vasoconstriction within the cortex?
2 Methods
2.1 Subjects Subjects were ten weanling white pigs of the breed 'Swedish landrace.' They weighed between 13.5 and 19kg (158 + 1"51 kg). 2.2 Surgery The pigs were fasted 24 h before the acute experiments. Before surgery they received a preanaesthetic dose of azaperone (60mg) with atropine sulphate (0.5mg). When the azaperone was maximally effective (about 30min), they recived a bolus injection of metomidate chloride (4 mg kg-1) via an ear vein. Azaperone and atropine were supplemented about once per hour. After exposing the right carotid artery and jugular vein, we cannulated the artery for measurement of systemic arterial pressure (AME840-2!, AME, Horten, Norway) and the vein for continuous infusion of metomidate chloride ( 7 . 5 m g k g - l h -1 in a concentration of lmgm1-1 in sterile saline) and injection of other drugs. We exposed the trachea and incised it for insertion of an endotracheal tube to facilitate spontaneous breathing. Then we closed the neck incision and placed the animal on its right side for exposure of the kidney. We exposed the kidney via a flank approach (RuSSeLL et al., 1981), removing the lowermost rib and incising the flank anteriorly. A rostral-caudal incision intersected the anterior-posterior incision at its most posterior point. We opened Gerota's fascia and reflected the kidney anteriorly for exposure of the renal nerve. The renal sympathetic 122
Fig. 1 Fibre-optics of the laser Doppler flowmeter. The fibres are held on the surface of the kidney by a plastic holder and a piece of latex
2.3 Renal stimuli The stimuli used during the recording sessions were biphasic square pulses of constant current (800/1s per phase with a separation of lOOps between phases) with amplitudes of 3 mA and pulse repetition rates of 5, 10 and 50Hz. The duration of the stimulus was 12s; six stimuli were delivered at each frequency, with a 3 min rest between each stimulus. 2.4 Measurement of cardiovascular variables We measured the following variables: renal cortical blood flow, renal arterial blood flow, systemic arterial pressure, heart rate, mean renal arterial blood flow, mean systemic arterial pressure and red cell concentration. All variables were recorded on a stripchart recorder (Brush Instrument Co., Cleveland, Ohio, USA) and on a magnetic
Medical & Biological Engineering & Computing
March 1991
tape recorder (SE7000, SE Labs (EMI) Ltd., Feitham, UK) at a speed of 2 . 3 7 c m s - 1 and a frequency bandpass of DC-312Hz. Renal cortical blood flow was measured by laser Doppler flowmetry (STERN et al., 1979; NILSSON et al., 1980a; b) with a commercially available flowmeter (PF2, Perimed, Stockholm, Sweden); renal arterial blood flow was measured with an electromagnetic flowmeter (Zepeda Instruments, Seattle, WA, USA); systemic arterial pressure was measured from the carotid artery with a cathetermanometer (AME, Norten, Norway); and heart rate was derived from the electrocardiogram by use of a beat-bybeat ratemeter (Studsvik Atomenergi AB, Sweden). The two mean signals were derived from the instantaneous flow and pressure with single-pole, low-pass active filters whose time constants of 0.2s were selected to matoh the filter time constant of the laser Doppler flowmeter. Red cell concentration was derived from the scattered light that returned to the flowmeter from the kidney (NILSSON, 1984). Time delay was measured between the arterial flow and the cortical flow response. The half-amplitude point was measured at the onset of the flow decrease; the difference in time between the half-amplitude points of the two flows in a particular experiment was taken as the measured time delay. 2.5 Stimulus conditions Cardiovascular variables were measured under several conditions (Table 1). In nine pigs, we tied and crushed the renal nerve proximal to the location of the stimulating electrodes. In two pigs, a renal nerve was intact for all stimuli; a third pig was considered to be denervated unsuccessfully because its pressure responses resembled those of the two innervated animals but its flow responses were like those recorded from the denervated animals. In four pigs we stimulated the renal nerve at a frequency of 50Hz before and after IV administration of labetalol hydrochloride (5-7mgkg-1). This neural blocking agent helped us determine whether flow responses depended on the activation of the renal nerve or on other factors. In
above the rate that can produce sampling errors (GARDENSmRE, 1964). Digital filtering eliminated all frequency components above 1.2 Hz. The filtered signals were recorded in blocks of 4096 samples per channel (a total recording time of 40.95s). Each recording began at the onset of the stimulus. The responses were averaged within a given animal and the mean responses were averaged across all animals that were tested in the same way. 3 Results
3.1 Effect o f renal sympathetic nerve stimulation on cardiovascular variables
Typical results of stimulation in a pig with a crushed renal nerve are shown in Fig. 2. With 50Hz stimulation, both cortical and arterial flow decreased rapidly to about half of their resting rates, began to recover slowly before the stimulus train ended, and returned to their baseline levels within 30-40 s after the end of the stimulus train. With 10 Hz stimulation the flows had a more trapezoidal shape, and at 5 Hz they decreased only slightly. The cortical blood flow signal had rapid repetitive temporal variations whose frequency occurred at the heart rate, and occasional artefacts that resulted from gross body movements. The base line of the cortical flow signal was stable, and the response to the stimulus was much greater than the noise level. Blood pressure increased slowly and had a maximum change of about 6 mm Hg. Heart rate decreased to a minimum of about 5 beats m i n - 1 below baseline.
Table 1 Pressure increases with frequency
Pressure increase, mm Hg Stimulus frequency, Nerve Intact Nerve crushed Hz 50 24-7 + 0.81 5.3 _+ 1.44 10 30-6 _ 0.10 3.7 + 1.12 5 34-1 _+ 0.46 2.5 _ 0'78 All measures are mean ___SD. Intact nerve data are from six tests in one pig; crushed nerve data are from five pigs. four other pigs we stimulated the nerve at a frequency of 50Hz before and after IV administration of captopril (0.3mgkg-1), a drug that aided in determining whether flow responses were mediated by the renal nerves or indirectly by the renin-angiotensin system. To test the effectiveness of captopril as an inhibitor of angiotensinconverting enzyme, we administered angiotensin I (300ngkg -1 IV; VOLLMER and BOCCAGNO, 1977) before and after administration of captopril. 2.6 Data analysis Renal cortical flow, mean renal arterial flow, mean arterial pressure, heart rate and red cell concentration were analysed by a digital computer system (Model 545lC, Hewlett Packard Co., Palo Alto, California, USA). Each channel was digitised at a rate of 100 samples s-1, well Medical & Biological Engineering & Computing
Fig. 2 Data collected followiny 12s stimulation of a denervated renal nerve at frequencies of 50, 10, and 5 Hz in one pig. Heavy lines indicate duration of the stimulus
In the four pigs injected with labetalol, the responses to renal stimulation were reduced markedly. The change in cortical flow was reduced by 70 _+ 12 per cent and that in the arterial flow by 77 __+ 14 per cent. The flow, pressure, and heart rate responses averaged across five animals (Fig. 3) were similar to those seen in the instantaneous recording. At stimulus frequencies of 50 Hz (Fig. 3, panel a), the flows were triangular in shape: they decreased from baseline to the minimum within 5 s of stimulus onset, remained there for less than 5s, then
March 1991
123
returned toward baseline, reaching a value about 30 per cent below baseline at the end of the 12s stimulus and increasing more slowly thereafter. The entire response lasted 40--50 s. The cortical flow signal occurred after the arterial flow signal (Figs. 2-4). Blood pressure increased slowly at the onset of the stimulus, reached a maximum increase of 5.3__+ 1 . 4 m m H g during the stimulus, and showed another increase after the stimulus ended. Heart rate decreased slightly with the onset of the stimulus and increased slightly at its end.
five animals after stimulation at 50 Hz (Fig. 4a) and 10 Hz (Fig. 4b). The delays observed during a variety of conditions are shown in Fig. 5. During occlusion of the renal artery, the delay averaged 0.47 + 0.1 s in seven animals. With the stimulus frequency at 50 Hz, the delays averaged 1.53 + 0 . 4 s in five animals with afferent denervation, 1.59 __+0.2s in two animals with afferent innervation, and 1-74 + 0.4s in three animals after administration of captopril. As there was no significant difference between the innervated and denervated animals, one denervated and b
c
120[ o _
120
d
o.~
40 O~
'T_c 120 I "~ d E
80
'- us 40 50Hz
e~
o
Io
20
30
40
5Hz
lOHz 0
10
20
40
30
10
20
30
40
60
time,s
Fig. 3
Averaged responses in five pigs to stimulation of the denervated renal nerve at frequencies of 50, I0 and 5 Hz. The responses are means and means + SD. The signals have been digitally filtered (see text). Heavy lines indicate duration of stimulus
When the stimulus frequency was decreased to 10Hz, the responses differed from those obtained at 50 Hz (Fig. 3, panel b). The shape of the cortical flow was triangular, but the slope was gradual at the onset of the stimulus and more rapid at its end. The arterial flow response was trapezoidal, with rapid changes occurring within 5s of the stimulus onset and offset, and a slow decrease from 5 s after the stimulus onset to the time of its offset. As at 50 Hz, the cortical flow signals occurred after the arterial flow signals. Blood pressure increased at stimulus onset, but not as much as at 50Hz. It increased by 3.7 _+ 1.1 m m H g during the stimulus and showed another slight increase after the stimulus ended. Heart rate decreased slightly during the stimulus. When the stimulus frequency was decreased to 5 Hz, arterial flow decreased slowly to a minimum and increased after the stimulus ceased (Fig. 3, panel c). (This panel summarises the data from four animals because stimulus artefact in the arterial flow signal of the fifth animal so distorted the signal that the data were not included in the average.) Any possible decrease in cortical flow was masked in the noise level. Because the cortical flow response was so small, a delay between the cortical and the arterial responses was not evident and was not measured directly. Pressure increased by 2.5 + 0.8 mm Hg during the stimulus. Heart rate did not change significantly.
two innervated animals were grouped together for calculation of the response to captopriI. There was no significant difference between the animals that received captopril and those that did not. At a stimulus frequency of 10Hz, the average delays were 1.99 + 0 . 6 s for denervated animals and 1.79 + 0.4 s for intact animals. The difference between the delays of the five denervated animals at 50 Hz and 10Hz was significant at the 10 per cent level (Student's t-test), suggesting that there may be a frequency-dependent 0-6
1-2
0.5
1"0
0,4
0'6
T 06 c E 0"4 E 0"2 ~"
0-3
0-2 >. 0-1
50Hz i
,
n
n
i
stimulation I
0
-~ 0.6
o '
0
1-2 ~o 1.0 -~
u
:.= 0-5
o41
0-8
o-3i
0.6 b
0-2 1OHz stimulation
0.1
o
,'o
3'5
0-4 102
Zo O
time,s
3.2 Delay between arterial and cortical f l o w responses The delay of renal cortical flow with respect to renal arterial flow is evident in the flow responses averaged over 124
Fig. 4
Mean arterial (solid curves) and cortical (broken curves) flow signals from five pigs durin 9 stimulation of the denervoted renal nerve at frequencies of 50 and I0 Hz
Medical & Biological Engineering & Computing
November 1990
delay between the cortical flow and the arterial flow. In one additional animal, stimulation of the crushed sympathetic nerve produced no delay between the renal arterial flow response and the renal cortical flow response. Data from that animal are not included in the five-animal average.
o
!~iiiii~iii
crushed
occlusion ~
(Table 1). The pressure increase was greater by a factor of at least two in the animal with intact nerves than in the animals with crushed nerves. 3.4 Effects of captopril on systemic arterial pressure In three animals stimulated at 50Hz before and after administration of captopril, the effect of renin-angiotensin on pressure appeared 20-25 s after the beginning of the stimulus. With captopril, the maximum pressure decreased after the end of the stimulus; when captopril was not present or was rendered ineffective (i.e. angiotensin I produced a response equal to that of a precaptopril response), pressure increased. The change in pressure was always less when captopril was given (Table 2; results in one animal are not shown because only one pressure response was observed after the captopril injection).
Table 2 Comparative sizes of late pressure changes (mm Hg) Pig Control Captopril 7 12.9 _ 0.8 6.3 + 4.27 10 -0.5 + 0.26 -3.0 + 0.7 All responses are mean _ SD.
intact captopril crushed intact ~. ,~ ~ angiotensin
50Hz lOHz 1 Fig. 5 Delays between arterial and cortical blood flows following renal stimulation under several conditions. Means are large bars; standard deviations are shown as vertical lines with cross bars
3.3 Effects of stimulus frequency on flow and pressure Comparison of the changes in both arterial and cortical flow at all stimulus frequencies (Fig. 6) revealed a significant difference between the changes in both arterial and cortical flows at 10Hz and at 5Hz (p < 1 per cent, Student's t-test). The responses to 10 Hz and to 50 Hz were not significantly different for arterial flow, but the change in cortical flow was significantly different at the 1 per cent level. 1"50
75
....J'I //tt
"7
c_ 50 > E
1'00
/ / t"
9-6 c
L
1 1/
/
arterial flow
o 25
0-50
0
0
0
-
1;
2'0
3'0
The laser Doppler flowmeter provides the means to measure renal cortical blood flow continuously. While instantaneous flow signals are damped (the 3 dB frequency is 0.8 Hz), the flowmeter responds rapidly enough to show changes in the mean blood flow. STERN et al. (1979) demonstrated that the technique is useful to measure flows in the kidney, but that the response to cortical flow may not be linear at high rates of flow. NILSSON'S (1984) modifications of the signal processing in the flowmeter corrected the nonlinearity. The technique still cannot provide the absolute value of flow, but it does provide continuous, linear and rapid indications of changes in the renal cortex.
cortical flow
/ /
l!
4 Discussion
20 5'o
f,Hz
Fig. 6 Amplitudes (Fm.x Fmin) of the changes of the cortical (x) and arterial ( + ) flow signals as functions of stimulus frequency. The signals are presented as mean and SD
In all animals, pressure increased from a minimum value at the onset of stimulus to a maximum value by the end of the stimulus. The maximum value was a function of the stimulus frequency: in the animals with denervated afferent connections, the change in pressure increased with frequency, while in the intact animal tested with all stimulus frequencies, pressure decreased with increasing frequency Medical & Biological Engineering & Computing
4.1 Effects of neural stimulation Stimulation of the renal nerve produced repeatable decreases in blood flow in the renal artery and in the renal cortex. Both responses to 50 Hz stimuli were qualitatively similar to those obtained in dogs (DISALVO and FELL, 1971) and in baboons (SPELMAN et al., 1986). The flows decreased rapidly to minima and increased slowly during the time of the stimulus current. Unlike DISALVO and FELL, however, we found that blood flow decreased more at 50 Hz than at 10 Hz (Fig. 6). This result was not caused by sequencing, because the same effect occurred when stimuli were delivered in the order 50-10-5Hz, 50-5-10Hz, and long sequences of 50-, 10-, and 5-Hz stimuli. The 3 min rest between stimuli should have allowed complete recovery before a new stimulus was given. Although 50 Hz is a higher frequency than that normally applied acutely to the renal nerve, there is evidence that burst activity in sympathetic nerves is higher than the 10Hz often applied ( A N D E R S S O N , 1983). Certainly, high-frequency activity is necessary to achieve renal constriction (DISALVO and FELL, 1971), and such constriction has been observed in unanaesthetised primates during exposure to an unfamiliar cagemate (SMITH et al., 1988). We suggest that a highfrequency stimulus is necessary to achieve rapid vasoconstriction and to minimise the delays observed in the present study.
March 1991
125
Our observation that the flow decrease was greater at higher frequencies of stimulation is consistent with observations that the resistances of both afferent and efferent arterioles in rats increase with increasing frequency of stimulation (KON and ICHIKAWA 1983; HERMANSSONet al., 1981; 1984). Quantitative comparisons are invalid, however, as we used short- (12 s) rather than long-duration stimuli. Our stimuli are more directly comparable with those of DISALVO and FELL (1971), although theirs were of variable duration (20-30 s). The measured responses appear to have been the direct effect of stimulating the renal nerves and not an indirect action mediated through another mechanism. Labetalol, an alpha and beta sympathetic blocker, reduced the response to neural stimulation by more than 90 per cent in one animal. As the renal innervation is alpha adrenergic (DIBONA, 1982), labetalol should affect the flow response to stimulation. However, blockade of the response with labetalol is not sufficient evidence to support the hypothesis that the flow responses result from direct neural activation. There remains the possibility that the renal nerves affect the renin-angiotensin system, which, in turn, causes vasoconstriction in the renal vessels. The addition of captopril in doses sufficient to block the effects of angiotensin I produced no change in the shape of the flow responses, nor were significant changes observed in the delays between the decreases in arterial flow and cortical flow. The magnitude of the responses was different with and without captopril, but the first response was always larger, with or without captopril. Hence, the difference is likely to be related to fatigue of the animal rather than to the presence or absence of the renin-angiotensin system. 4.2 Pressure changes The pressure responses demonstrate an effect of the renin-angiotensin system: when the system was intact, there was a late increase in the pressure response to stimulation; when the system was blocked, the increase disappeared. The time to achieve the maximum pressure change, about 20s, was consistent with circulation time from the kidney to the lungs in weanling pigs (MOUNT and INGRAM, 1971 ; RUCH and PATTON, 1974). The increased blood pressure during stimulation appears to have been the result of two effects. The animals with crushed nerves exhibited a passive increase in pressure when flow to the kidney decreased. The flow decreased by 60 per cent and, assuming that flow in one kidney is 5-10 per cent of the cardiac output (MouNT and INGRAM, 1971), the increase of 5 per cent in blood pressure is consistent with a passive effect. In the animals with intact nerves, blood pressure increased by 10-15 per cent during stimulation at 50 Hz. 4.3 Estimate of laser Doppler flow linearity As the kidneys were not weighed in this study, we could not calculate the renal flow per gram of tissue. We did obtain an estimate, however, by removing and weighing both kidneys from six pigs that weighed about the same as our subjects. After the kidneys were rinsed with normal saline and allowed to drain for 30 min, they weighed 55 _+ 7g. This weight combined with the range of flows obtained in our study yielded a specific flow of 118226mlmin -1 100g - t with an average of 160mlmin -1 100g-t. Although this value is slow compared with 300400 ml min-1 100 g-1 for the rat (HERMANSSONet al., 1984) and 400mlmin -1 100g -1 for the dog (POMERANZet al., 1968; ROSIVALLand NAVAR,1983), the laser Doppler flow126
meter used in this study was well within its linear range (AnN et al., 1987). Calibration of the electromagnetic flow system in six animals showed that the gain was constant within 10 per cent in all of them (2.1 _ 0 . 2 m V m l - 1 min-1). The flow transducers fit the vessels well: the lumen of the flow transducers was 2.5 mm in diameter, while the vessels appeared to be no greater than 3.0 mm in diameter before implantation of the transducer. Blood pressure did not increase after implantation of the transducer. It is unlikely that the renal artery was constricted severely by the transducer or that such an effect was the reason that total flow was low. Likewise, it is unlikely that the anaesthetic suppressed renal flow since cardiovascular function is unimpaired in animals anaesthetised with metomidate chloride (Janssen Pharmaceutica, 1973). The level of the measured renal flow did not change from the beginning to the end of the experiment, and the kidneys showed no change in colour from the time of first exposure to the end of the experiment. The low specific flows suggest the possibility that the kidney was injured, but that seems unlikely, as neither pressure nor flow changed during the study. Either the relative weight of the kidneys is greater in weanling pigs than in adult pigs, or the flow per gram of tissue is less in pigs than it is in other animals. 4.4 Suggested mechanism for time delays The delays measured between the decreases in arterial and cortical flow suggest that the afferent arteriole constricts before the efferent arteriole. The flow measured in the renal cortex is probably sensed from the surface to a depth of 0.5-1.0mm (STERN et al., 1977; 1979). While the superficial renal cortex includes the afferent and efferent arterioles, the glomerular capillaries, the peritubular capillaries, and some vesssels of the vasa recta, more than 50 per cent of the drop in arteriovascular pressure occurs across the afferent arteriole and about 40 per cent across the efferent arteriole (KNox and SPIELMAN,1983). Both arterioles are innervated, as are the inlets to the vasa recta. As flow decreases in the renal artery when a stimulus is applied, and continues unabated in the renal cortex for a short time thereafter, it seems likely that blood is stored in the glomerular capillaries, and that the afferent arteriole constricts before the efferent arteriole and any other downstream vessels. Because the efferent arterioles comprise controlled resistance vessels distal to the glomerular capillaries, their constriction probably has a greater effect than constriction of the inlet to the vasa recta. Fig. 7 shows a simple passive model of the renal circulation to illustrate the effect. If the two resistances to flow, Raa and Rea, represent the afferent and efferent arterioles, respectively, and Cg represents the fluid capacitance of the glomerular capillaries, then constriction of the upstream resistance would produce the flows shown in the figure. Fluid stored in the glomerular capillaries can flow out of them into the venous circulation until the volume is depleted or until the efferent arteriole decreases flow out of the capillaries. The time constant varies with changes in the resistance, increasing as the resistances increase. Figs. 2-4 clearly demonstrate that there was a delay between the cortical and arterial flows in our pigs. The delay during stimulation was greater than that during occlusion, suggesting that the effect was not purely passive. Furthermore, the delay depended on the frequency of stimulation, at least as the frequency was varied from 10 to 50Hz. It was possible to show a further increase in the delay, to about 6 s, by analysing the phase characteristic curves of the flow responses to 5 Hz stimuli in the fre-
Medical & Biological Engineering & Computing
March 1991
quency domain (JAvID and BRENNER, 1963). That information, which was less clear than the data produced with the 50 Hz and 10Hz stimuli, is offered merely to suggest that delay increases as stimulation frequency decreases. The delayed decrease in flow in the renal cortex was probably caused by the slower action of smooth muscle cells in the efferent arteriole. Fibres that terminate on the cells of the afferent arteriole, the efferent arteriole, and the juxtaglomerular apparatus (BARAJAS, 1978) should act simultaneously on the smooth muscle cells of both arterioles, and the observed delay must be caused by later activity in the efferent arteriole. This hypothesis is supported by HERMANSSON et al.'s (1981) report of a decrease in glomerular pressure during 2 Hz stimulation (see their Fig. 3). The decrease began at the onset of stimulation and reached a minimum 3s later: the pressure remained constant throughout the rest of the stimulus. That observation is consistent with the hypothesis that the efferent arteriole constricts later than the afferent arteriole. It is possible that vasoconstriction occurs successively along the vascular tree from the proximal to the distal vessels, but that possibility was not addressed in this study. RQQ
Po
I
ql
-
Rea
Cg
q2
T
t
ry
ql
of the concentration signal in the renal cortex (M. Wolgast, personal communication, 1985). The concentration signal decreased later than did the flow signal (the time delay was < 0.2 s), suggesting a change in the velocity of the red cells that took place before the change in the total number of cells. Such a finding is consistent with storage of blood in the vessels of the kidney, an early decrease in the supply to that volume and a later depletion of the volume. If blood is stored in the kidney, renal volume should change more slowly than renal arterial flow. Measures of volume change during stimulation of the renal nerve would clarify this point. 5 Conclusion In summary, this study revealed that transient changes occur in blood flow in the renal artery and the renal cortex as functions of stimulation of the renal efferent nerve. The decrease in cortical flow occurred after that in arterial flow, consistent with earlier data suggesting that the afferent arteriole constricts before the efferent arteriole (HERMANSSON et al., 1981). These findings suggest that there should be transient changes in glomerular filtration rate and urine production during the early activation of the renal nerves, but that such changes should be complete within a few seconds. Acknowledgments--We acknowledge the technical assistance of
Lars Eriksson, Per Torngren and Per Sveider for the design and construction of the stimulator and the stimulus electrodes. G6ran Salerud and Toshiyo Tamura helped with surgical techniques. Squibb provided us with captopril, and Pal van Szokolay of Leo AB loaned us hypnodil during a desperate time when there was none available. Mats Wolgast offered helpful suggestions for the manuscript. This project was supported by US National Institutes of Health grants RR00166 and HL16910 and by grants from the National Swedish Board for Technical Development, Project No. 84-3790. References
I
~t
I I q2 ----
4--
I I
stimulation Raa changes, Rea does not
~t
Fig. 7 Simple model of the renal cortical circulation and its response to a square stimulus R,= = resistance of afferent arteriole Re= = resistance of efferent arteriole Cg = capacitance of glomerular capillaries R~ = resistance of venules and veins ql , q2 are flows; Pa = SAP
Further evidence to support the idea that efferent arterioles constrict after afferent arterioles is available from the red cell concentration signal (NILSSON, 1984). In our study, this signal was proportional to the total number of red cells moving in the sample volume observed by the laser Doppler flowmeter. The change in the concentration signal was small (Fig. 3), and its relation to the cortical flow signal became clear after averaging within and across animals. This experience is not unusual for measurements Medical & Biological Engineering & Computing
AI-IN, H.C., JOHANSSON,K., LUNDGREN, O. and NILSSON, G. E. (1987) In vivo evaluation of signal processors for laser Doppler tissue flowmeters. M ed. & Biol. Eng. & Comput., 25, 207-211. ANDERSSON, P. O. (1983) Comparative vascular effects of stimulation continuously and in bursts of the sympathetic nerves to cat skeletal muscle. Acta. Physiol. Scand., 118, 343-348. BARAJAS,L. (1978) Innervation of the renal cortex. Fed. Proc., 37, 1192-1201. BARAJAS, L. (1981) The juxtaglomerular apparatus: anatomical considerations in feedback control of glomerular filtration rate. Ibid., 40, 78-86. BERNARD, C. (1859) Leqons sur les propribtbs des liquides de l'organisme. Balliere, Paris, 172-173. DIBONA, G. F., ZAMBRISK[,E. J., AGUILERA,A. J. and KALOVANIDIES, G. J. (1977) Neurogenic control of renal tubule sodium reabsorption in the dog. Circ. Res., 40, 1127-1130. DIBoNA, G. F. (1982) The functions of the renal nerves. Rev. Physiol. Biochem. Pharmacol., 94, 75-181. DISALVO, J. and FELL, C. (1971) Changes in renal blood flow during renal nerve stimulation. Proc. Soc. Exp. Biol. Med., 136, 150-153. GARDENSHIRE, L. W. (1964) Selecting sample rates. ISA J., 11, 59-64. HERMANSSON,K., LARSON,M., KXLLSKOG,O. and WOLGAST,M. (1981) Influence of renal nerve activity on arteriolar resistance, ultrafiltration dynamics and fluid reabsorption. PfliJgers Arch., 389, 85-91. HER~ANSSON,K., OJTEG, G. and WOLGAST,M. (1984) The cortical and medullary blood flow at different levels of renal activity. Acta Physiol. Scand., 120, 161-169. Janssen Pharmacetica (1973) Stresnil : a new sedative for pigs. Veterinary Development Department, Beerse, Belgium.
M a r c h 1991
127
JAVID, M. and BRENNER, E. (1963) Analysis, transmission, and filtering of signals. McGraw-Hill, New York. KNOX, F. G. and SPIELMAN,W. S. (1983) Renal circulation. In Handbook of Physiology. GEIGER, S. R. (Ed.). Section 2: The cardiovascular system, vol. 3, Part 1. SHEPHERD, J. T. and AaBOUD, F. M, (Vol. Eds), American Physiological Society, Bethesda, Maryland, 183 217. KoN, V. and [CHIKAWA, |. (1983) Effector loci for renal nerve control of cortical microcirculation. Am. J. Physiol, 245, F545F553. MCKENNA, O. C. and ANGELAKOS, E. T. (1968) Acetylcholinesterase containing nerve fibers in the canine kidney. Circ. Res., 23, 645-651. MOUNT, L. E. and INGRAM,D. L. (1971) The pig as a laboratory animal. Academic Press, New York. NILSSON, G. E., TENLAND, T. and OBERG, P. A. (1980a) A new instrument for continuous measurement of tissue blood flow by light beating spectroscopy. IEEE Trans., BME-27, 12-19. NILSSON, G. E., TENLAND, T. and OaERG, P. A. (1980b) Evaluation of a laser Doppler flowmeter for measurement of tissue blood flow. Ibid., BME-27, 597-604. NILSSON, G. E. (1984) Signal processor for laser Doppler tissue flowmeters. Med. & Biol. Eng. & Comput., 22, 343-348. POMERANZ,B. H., BIRTCH,A. G. and BARGER,A. C. (1968) Neural control of intrarenal blood flow. Am. J. Physiol., 215, 10671081. ROSIVALL, L. and NAVAR, L. G. (1983) Effects on renal hemodynamics of intraarterial infusions of angiotensins I and II. Ibid., 245, FI8I-F187. RUCH, T. C. and PATTON, H. D. (Eds.) (1974) Physiology and biophysics. 11: Circulation, respiration and fluid balance. Saunders, Philadelphia. RUSSELL, J. M., WEBB, R. T. and BOYCE, W. H. (1981) Intrarenal surgery. Invest. Urol., 19, 123-125. SMITH, O. A., GOLANOV, E., ASTLEY, C. A., CHALYAN, V., SPELMAN, F. A., URMANCHEEVA, T., BOWDEN, D. M. and CHESNEY, M. A. (1988) Behavioral stress and the cardiovascular system. In Neurocardiology. KULBER~S, H. E. and FRANCK,G. (Eds.), Futura, Mt. Misco, New York. SPELMAN, F. A., OBERG, P. A. and ASTLEY, C. (1986) Localized neural control of blood flow in the renal cortex of the anesthetized baboon. Acta Physiol. Scand., 127, 437-441. STERN, M. D., LAPPE, D. L., BOWEN, P. D., CHIMOSKEY,J. E., HOLLOWAY, G. A., KEISER, H. R. and BOWMAN, R. L. (1977) Continuous measurement of tissue blood flow by laserDoppler spectroscopy. Am. J. Physiol., 232, H441-H448.
128
STERN, M. D., BOWEN, P. D., PARMA, R., OSGOOD, R. W., BOWMAN, R. L. and STEIN, J. H. (1979) Measurement of renal cortical and medullary blood flow by laser-Doppler spectroscopy in the rat. Ibid., 236, F80-F87. VOLLMER, R. R. and BOCCAGNO, J. A. (1977) Central cardiovascular effects of SQ 14,225, an angiotensin-converting enzyme inhibitor in chloralose-anesthetized cats. Eur. J. Pharm., 45, 117-125. ZAMBRASKI, E. J. and DIBONA, G. F. (1976) Angiotensin II in antidiuresis of low-level renal nerve stimulation. Am. J. Physiol., 231, 1105-1110.
A u t h o r s ' biographies Francis A. Spelman was born in San Francisco, California, USA, in 1937. He received the BSEE degree from Stanford University in 1959, the MSEE degree from the University of Washington, Seattle, in 1968, and the Ph.D. degree from the University of Washington in 1975. He has headed the Bioengineering and Computer Divisions of the Regional Primate Research Center at the University of Washington since 1965. He was a Visiting Researcher at the Department of Bioengineering, Link6ping University, Sweden, from 1985 to 1986. He is a member of US/USSR joint research team studying hypertension in baboons, and also studies the inner ear.
P. ,~ke Oberg received the MSEE degree from the Chalmers University of Technology, G6teborg, Sweden, in 1964, and the Ph.D. in Biomedical Engineering from Uppsala University in 1971. He is currently Professor of Biomedical Engineering at Link6ping University and Director of the Department of Clinical Engineering, University Hospital. His research interests include biomedical instrumentation, transducers and clinical engineering. Dr Oberg is a Past President of the Swedish Society of Medical Physics & Medical Engineering and a board member of the IFMBE's Division of Clinical Engineering. He is a Fellow of the Swedish Academy of Engineering Sciences and the Royal Swedish Academy of Sciences.
Medical & Biological Engineering & Computing
March 1991