J Am Soc Nephrol 13: 27–34, 2002 Preglomerular and Postglomerular Resistance Responses to Different Levels of Sympathetic Activation by Hypoxia

KATE M. DENTON, AMANY SHWETA, and WARWICK P. ANDERSON Department of Physiology, Monash University, Clayton, Victoria, Australia.

Abstract. This study investigated the effects of graded reflex P Ͻ 0.05) and filtration fraction (FF; P Ͻ 0.01) increased. In increases in renal sympathetic nerve activity (RSNA) on renal response to moderate hypoxia, calculated preglomerular preglomerular and postglomerular vascular resistances. With (ϳ20%) and postglomerular (ϳ70%) resistance both in- the use of hypoxia to reflexly elicit increases in RSNA without creased, but only the increase in postglomerular resistance was affecting mean arterial pressure, renal function and stop-flow significant (P Ͻ 0.05). In contrast, severe hypoxia decreased pressures were measured in three groups of rabbits before and ERBF (56 Ϯ 8%; P Ͻ 0.01), GFR (55 Ϯ 9%; P Ͻ 0.001), and Ϯ Ϯ after exposure to room air and moderate (14% O2) or severe glomerular capillary pressure (32 1 versus 29 1 mmHg; Ͻ (10% O2) hypoxia. Moderate and severe hypoxia increased P 0.001), with no change in FF, reflecting similar preglo- RSNA, primarily by increasing the amplitude of the sympa- merular (ϳ240%; P Ͻ 0.05) and postglomerular (ϳ250%; P Ͻ thetic bursts rather than their frequency. RSNA amplitude 0.05) contributions to the vasoconstriction and a decrease in Ϯ Ͻ Ϯ Ͻ Ͻ increased by 20 6% (P 0.05) and 60 16% (P 0.05), calculated Kf (P 0.05). These results provide evidence that respectively. Moderate hypoxia decreased estimated renal reflexly induced increases in RSNA amplitude may differen- blood flow (ERBF; 26 Ϯ 7%; P ϭ 0.07), whereas estimated tially control preglomerular and postglomerular vascular glomerular capillary pressure (32 Ϯ 1 versus 34 Ϯ 1 mmHg; resistances.

Previous renal micropuncture studies have suggested that elec- Physiologically induced graded increases in renal sympathetic trical stimulation of the renal nerves predominantly constrict nerve activity (RSNA) cause progressive renal vasoconstriction the preglomerular vessels (1–4); this is not surprising, because (11,13,23), but the effects of hypoxia on preglomerular and post- the afferent arteriole is much more densely innervated than the glomerular vascular resistances have not been measured. This is efferent arteriole (5). However, electrical stimulation may not of particular interest, given our recent identification of two distinct fully mimic physiologic changes in nerve activity. For exam- nerve types that are differentially distributed to afferent and ef- ple, electrical stimulation at supramaximal voltages activates ferent arterioles (24). Type I axons almost exclusively innervate all nerves simultaneously, and the frequency at which the the afferent arteriole and have larger axon diameters with atypical nerves are fired is altered, whereas it is now clear that the varicosities (5), whereas type II axons are distributed equally to number of nerves activated during a physiologic burst of nerve both afferent and efferent arterioles and resemble those innervat- activity may vary widely (6–18). The bursting pattern of renal ing other blood vessels with typical fusiform varicosities (5). The sympathetic nerve activity can be differentiated into changes in distribution of the two types of nerves to the renal arterioles raises frequency (rate) of nerve firing and changes in amplitude the possibility that preglomerular and postglomerular vascular (number of nerves recruited during the burst) (6,9,19). Physi- resistances may be regulated differentially, depending on the ologic stimuli such as hypoxia, air jet stress, white noise (sound nerve type activated. We hypothesize that increased type I acti- that has a relatively wide continuous range of frequencies of vation would constrict the afferent arteriole, whereas type II nerve uniform intensities), and thermal stimulation have been shown activation would constrict both afferent and efferent arterioles. to reflexly increase the amplitude of the sympathetic bursts to To investigate this hypothesis, we have used two levels of the kidney (i.e., increased recruitment of nerves) without sig- hypoxia (14% and 10% O2) as our stimulus, which is known to nificantly changing the frequency of the discharges produce graded increases in RSNA and renal vascular resistance (11,12,20,21). Speculation has therefore arisen that different (11,23). Importantly, these levels of hypoxia do not alter mean patterns of activation of the renal nerves may produce differ- arterial pressure (MAP), therefore avoiding the potentially con- ential changes in renal function (22). founding influence of renal autoregulation. We have determined the relative changes in preglomerular and postglomerular resis- tances to the hypoxic stimuli, using stop-flow pressure (SFP) Received December 20, 2000. Accepted July 5, 2001. measurements to estimate glomerular capillary pressure. Correspondence to Dr. Kate Denton, Department of Physiology, P.O. Box 13F, VIC, Australia, 3800. Phone: 61-3-9905-2503; Fax: 61-3-9905-2566; E-mail: [email protected] Materials and Methods 1046-6673/1301-0027 Rabbits Journal of the American Society of Nephrology Experiments were performed on male rabbits of a multicolored Copyright © 2001 by the American Society of Nephrology English strain (n ϭ 20; mean, 3.0 Ϯ 0.1 kg; range, 2.5 to 3.5 kg). The 28 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 27–34, 2002 experiments were performed in accordance with the Australian Code gram over 2-s periods was defined as total RSNA. Changes in voltage of Practice for the Care and Use of Animals for Scientific Purposes of the nerve signal above a defined threshold were classified as RSNA and were approved in advance by the Monash University Standing discharges (11). Committee on Ethics in Animal Experimentation. Micropressure measurements were taken during the baseline and gas periods; micropressures were not measured during the recovery Preparation period. The micropressures were pooled for each period, with 3 to 6 On the experimental day, catheters were placed in the ear central SFP measurements and 3 to 5 proximal tubular pressures being taken arteries and marginal veins under local anesthetic (Xylocaine, 0.5% in each period. The micropressures were measured by use of a wt/vol lidocaine; Astra Pharmaceuticals, North Ryde, NSW, Austra- servo-null pressure measuring device (Micropressure System, model lia). Conscious arterial pressure was measured for 20 min. Anesthesia 900A; WPI, Sarasota, Florida). was then induced by intravenous administration of pentobarbitone MAP, heart rate, total RSNA, RSNA frequency, and RSNA am- sodium (90 to 150 mg plus 0.1 to 0.2 mg/kg per min intravenously; plitude were all continuously recorded throughout the experiment and Nembutal, Boehringer Ingelheim, Artarmon, NSW, Australia) and the were sampled by use of an analog-to-digital data acquisition card ϩ rabbits ventilated (small animal ventilator, Model 683; Harvard Ap- (Lab-PC , National Instruments, Melbourne, Australia). Calibrated paratus, Holliston, MA). During surgery and throughout the experi- signals were displayed on screen and saved to disk as 2-s averages of ment, fluid was infused intravenously at 0.17 ml/min per kg (Hart- each variable by use of a program written in the LabVIEW graphical mann’s solution [compound sodium lactate]; Baxter Healthcare, programming language (National Instruments) (11). Toongabbie, NSW, Australia) to replace fluid losses. Surgery was performed on a heated table, and esophageal temperature was main- Measurements and Calculations tained between 36°C and 38°C throughout the experiment via a MAP, heart rate, total RSNA, frequency of RSNA, and RSNA servocontrolled infrared lamp (Digi-Sense Temperature Controller; amplitude were averaged over each clearance period. GFR was esti- Cole Palmer, Vernon Hills, IL). The rabbit was placed in an upright mated via the clearance of [3H]-inulin, and estimated renal blood flow crouching position, and a left flank incision made to expose the left (ERBF) as the clearance of [14C]-para-aminohippurate corrected for kidney. Silastic catheters were inserted into both ureters for the hematocrit. Renal vascular conductance was calculated as ERBF collection of urine, and the left kidney was placed in a micropuncture divided by MAP. Filtration fraction was calculated as GFR divided by cup. Before the kidney was prepared for micropuncture, the renal estimated renal plasma flow. Plasma renin activity (PRA) was deter- nerves were identified and a 3- to 5-mm length freed from the mined by RIA (26). Urinary sodium concentrations were determined surrounding tissue and carefully placed across a pair of hooked silver by flame photometry (Model 943; Instrumentation Laboratory, Milan, recording electrodes. The nerve and electrode were insulated from the ␲ Italy). Plasma oncotic pressure ( a) was calculated from arterial surrounding tissue and effluent from the micropuncture bath by use of plasma protein concentration (Ca), determined by use of the Lowry Wacker Sil-Gel (Wacker-Chemie, Munich, Germany). The kidney assay (25). Estimated glomerular capillary pressure (Pgc) was calcu- was then prepared for micropuncture as described elsewhere (25). ϩ ␲ lated as SFP a. Preglomerular resistance was estimated as (MAP After completion of surgery, [3H] inulin (4-␮Ci bolus plus 300 Ϫ Ϫ Pgc)/ERBF and postglomerular resistance as Pgc/(ERBF GFR). 14 nCi/ml; NEN Research Products, Sydney, Australia) and [ C] para- The preglomerular to postglomerular resistance ratio was determined aminohippurate (1-␮Ci bolus plus 83 nCi/ml; NEN Research Prod- as the preglomerular resistance divided by the postglomerular resis- ucts) were administered intravenously by addition to the infusion of ϭ Ϫ Ϫ ␲ tance. The mean net filtration pressure (Puf) Pgc PT gc, and Hartmann’s solution. During a 90-min equilibration period, the arti- ϭ ␲ the glomerular ultrafiltration coefficient (Kf) GFR/Puf, where gc ficial ventilation rate (tidal volume, 14 to 18 ml/min; rate, 30 to 50 ␲ ϩ equals the mean glomerular oncotic pressure calculated as ( a breaths/min) was adjusted such that, during the baseline period (room ␲ ␲ e)/2. The efferent oncotic pressure ( e) was derived from the effer- air), the arterial partial pressure for oxygen (PaO ) was between 90 Ϫ 2 ent protein concentration, which was calculated as Ca/(1 filtration and 110 mmHg. fraction). However, Kf can only be calculated by the above formula if there is a net positive pressure at the efferent end of the glomerular Experimental Protocol capillaries (27). If there is not, the differential equation of Deen et al. Measurements were made over three periods, a 60-min baseline (27) must be used, and then only a minimum value for Kf can be period, a 60-min gas period, and a 10-min recovery period. There was derived. a 20-min washout between each period (samples discarded). Each animal was randomly assigned to receive one of three gas mixtures Statistical Analyses (room air, 14% O2, or 10% O2). Each period consisted of a timed All data are reported as mean Ϯ SEM. P Յ 0.05 was considered to urine collection, with an arterial blood sample (2 ml total) taken at the be statistically significant. One-way ANOVA was used to test for midpoint for clearance, blood gas, hematocrit, and plasma renin differences between the groups during the baseline period. One-way activity measurements. At the conclusion of the experiment, the repeated measures ANOVA was performed to compare the baseline animal was killed with an intravenous overdose of pentobarbitone and gas periods in each group. (300 mg).

Data Acquisition Results No significant differences were observed in any variable Arterial pressure was continuously measured by connecting the ear artery catheter to a pressure transducer (Cobe, Arvarda, CO) and the among the three treatment groups during the baseline (pre–gas signal amplified (Model 7D, Grass Instruments, Quincey, MA). mixture) period. Table 1 shows the baseline measurements in RSNA was amplified, filtered between 50 and 5000 Hz, full-wave each group: the P values refer to an ANOVA for each variable. rectified, and integrated by use of a low-pass filter with a time These results demonstrate that the groups were under similar constant of 20 ms. The average voltage from the sympathetic neuro- resting conditions before administration of the gas mixture. J Am Soc Nephrol 13: 27–34, 2002 Renal Vascular Responses to Graded RSNA 29

Table 1. Baseline variables in the three groups prior to gas administrationa

Control Moderate Parameter (Prior to Severe P (Prior to 14% O ) (Prior to 10% O ) Room Air) 2 2

MAP (mmHg) 83 Ϯ 477Ϯ 677Ϯ 4 0.6 HR (beats/min) 221 Ϯ 8 224 Ϯ 14 231 Ϯ 10 0.8 Ϯ Ϯ Ϯ PaO2 (mmHg) 90 2923934 0.8 Ϯ Ϯ Ϯ PaCO2 (mmHg) 22 2263222 0.5 Total RSNA (␮V) 3.7 Ϯ 0.8 2.1 Ϯ 0.6 2.5 Ϯ 0.7 0.3 RSNA amplitude (␮V) 10.0 Ϯ 2.0 9.1 Ϯ 2.2 10.8 Ϯ 2.8 0.8 Frequency of RSNA (Hz) 9.5 Ϯ 0.5 8.0 Ϯ 1.1 8.7 Ϯ 0.7 0.5 Hematocrit (%) 30 Ϯ 233Ϯ 230Ϯ 1 0.5 ERBF (ml/min) 17 Ϯ 122Ϯ 420Ϯ 3 0.5 GFR (ml/min) 4.4 Ϯ 0.5 4.1 Ϯ 0.5 4.4 Ϯ 0.5 0.9 Filtration fraction (%) 35 Ϯ 328Ϯ 333Ϯ 4 0.4 UFR (ml/min) 0.11 Ϯ 0.05 0.17 Ϯ 0.07 0.16 Ϯ 0.04 0.7 Na excretion (␮mol/min) 11 Ϯ 313Ϯ 320Ϯ 7 0.6 PRA (ng AngI/ml per h) 7.7 Ϯ 1.9 5.8 Ϯ 0.8 5.9 Ϯ 1.1 0.5

a Ϯ Values are mean SEM. MAP, mean arterial pressure; HR, heart rate; PaO2, arterial partial pressure of oxygen; PaCO2, arterial partial pressure of carbon dioxide; RSNA, renal sympathetic nerve activity; EBRF, estimated renal blood flow; GFR, glomerular filtration rate; UFR, urine flow rate; Na excretion, sodium excretion rate; PRA, plasma renin activity. P values show the outcomes of one-way ANOVA, comparing levels in the three groups prior to exposure to the gas mixtures.

Responses to Graded Hypoxia ERBF decreased by 26 Ϯ 7% (P ϭ 0.05; Figure 2), and Room Air (Time Control). No significant time-related RVC decreased by 30 Ϯ 5% (P ϭ 0.04; Figure 2) in response Ϯ changes in MAP, PaO2, RSNA (total, frequency, or ampli- to moderate hypoxia. GFR did not change significantly (4.1 tude), renal vascular resistance (RVC), ERBF, GFR, filtration 0.5 versus 3.9 Ϯ 0.9 ml/min; P ϭ 0.5; Figure 2), but filtration fraction, urine flow rate, or sodium excretion rate were seen fraction rose significantly from 28 Ϯ 3to39Ϯ 5% (P ϭ 0.006; throughout the duration of the experiment (Figures 1 and 2). Figure 2) in response to moderate hypoxia. Urine flow rate Heart rate (221 Ϯ 8 versus 218 Ϯ 8 beats/min), arterial partial (P ϭ 0.7), sodium excretion (P ϭ 0.2), and fractional sodium Ϯ Ϯ ϭ pressure of carbon dioxide (PaCO2) (20 1.3 versus 20 1.7 excretion (P 0.9) did not change significantly (Figure 2). mmHg) and arterial pH (7.41 Ϯ 0.02 versus 7.41 Ϯ 0.015) SFP was 21.4 Ϯ 0.2 mmHg during the baseline period and were unaffected by time. PRA decreased by 16 Ϯ 7% (P ϭ rose significantly to 23.4 Ϯ 0.3 mmHg during moderate hy- 0.04; Figure 3) with time. Estimated glomerular capillary pres- poxia (P ϭ 0.007; Figure 4). Plasma protein concentration was Ϯ Ϯ Ϯ sure was 32.2 0.5 mmHg during baseline and 32.5 0.4 not different during baseline and 14% O2 (3.8 0.1 versus mmHg during room air (P ϭ 0.4). Plasma protein concentra- 3.8 Ϯ 0.1 g%, P ϭ 0.1). Estimated glomerular capillary pres- Ϯ Ϯ Ϯ ϭ tions were 4.0 0.1 g% during baseline and 3.9 0.1 g% sure increased by 2.1 0.5 mmHg during 14% O2 (P 0.007; during room air (P ϭ 0.4). Estimated glomerular capillary Figure 4). Proximal tubule pressure was 10.7 Ϯ 0.3 mmHg Ϯ ϭ pressure, preglomerular and postglomerular vascular resis- during baseline and 11.5 0.6 mmHg during 14% O2 (P tances, and Kf were unaffected by time (Figure 4). 0.4). Estimated preglomerular resistance increased in response 14% Oxygen (Moderate Hypoxia). Moderate hypoxia to moderate hypoxia, being 2.7 Ϯ 0.6 mmHg/ml per min and Ϯ ϭ Ϯ caused PaO2 to decrease to 48 7 mmHg (P 0.003; Figure 3.7 1.3 mmHg/ml per min, respectively, but this did not 1) and total RSNA to increase by 89 Ϯ 22% (P ϭ 0.03; Figure reach statistical significance (P ϭ 0.4; Figure 4). Postglomeru- 1). This increase in total RSNA was associated with a signif- lar resistance increased significantly, being 2.1 Ϯ 0.4 icant increase in the RSNA amplitude (20 Ϯ 6%; P ϭ 0.03; mmHg/ml per min during baseline and 3.4 Ϯ 0.5 mmHg/ml per Figure 1) but no significant change in the frequency of RSNA min during moderate hypoxia (P ϭ 0.016; Figure 4). The (7 Ϯ 4%; P ϭ 0.2; Figure 1). MAP and heart rate did not preglomerular to postglomerular resistance ratio decreased by Ϯ ϭ change significantly after moderate hypoxia (Figure 1). PaCO2 25 13% during 14% O2, but this was not significant (P levels were 26 Ϯ 3 mmHg during baseline and 26 Ϯ 2 mmHg 0.14). The rabbits were all in a state of filtration disequilibrium ϭ ⌬ Ͼ ␲ during 14% O2 (P 0.6). pH levels were unchanged between (i.e. P e) during both periods of the experiments. The Ϯ Ϯ Ϯ baseline (7.36 0.033) and 14% O2 (7.36 0.034). PRA calculated glomerular ultrafiltration coefficient was 0.97 levels were not significantly different between baseline and 0.24 ml/min per mmHg during baseline and 0.84 Ϯ 0.13 Ϯ ϭ moderate hypoxia, 5.8 0.8 ng angiotensin I (AngI)/ml per h ml/min per mmHg during 14% O2 (P 0.6; Figure 4). and 7.6 Ϯ 2.0 ng AngI/ml per h, respectively (P ϭ 0.3; Figure 10% Oxygen (Severe Hypoxia). Severe hypoxia caused Ϯ ϭ 3). PaO2 to decrease to 36 4 mmHg (P 0.001; Figure 1), 30 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 27–34, 2002

Figure 1. (Left panels) Variables measured during the baseline period (room air), administration of the gas mixture (room air; ●), moderate ⅜ (14% O2; ) and severe (10% O2; ) hypoxia, and recovery period (room air). Mean arterial pressure and the arterial partial pressure for oxygen are expressed as absolute values (mean Ϯ SEM). Renal sympathetic nerve activity (RSNA) (total, amplitude, and frequency) is expressed as the percentage of change from baseline levels (mean Figure 2. (Left panels) Values (mean Ϯ SEM) for the renal variables Ϯ SEM). (Right panels) For each adjacent variable in the left panel, measured during the baseline period (room air), administration of the the percentage of change from baseline during the administration of ● ⅙ gas mixture (room air; ), moderate (14% O2; ) and severe (10% O2; each gas mixture (room air [e], moderate [u], and severe [f] hyp- ) hypoxia, and recovery period (room air). (Right panels) For each oxia) for each variable (mean Ϯ SEM). *, P Ͻ 0.05; **, P Ͻ 0.01. adjacent variable in the left panel, the percentage of change from Both values indicate a significant change from the respective baseline baseline during the administration of each gas mixture (room air [e], value. moderate [u], and severe [f] hypoxia) for each variable (mean Ϯ SEM). *, P Ͻ 0.05; **, P Ͻ 0.01. Both values indicate significant change from the respective baseline value. whereas total RSNA increased by 162 Ϯ 34% (P ϭ 0.04; Figure 1). This increase in total RSNA was associated with a significant increase in the RSNA amplitude (60 Ϯ 16%; P ϭ not change significantly after severe hypoxia, being 77 Ϯ 4 and 0.05; Figure 1) and with a small but significant increase in the 74 Ϯ 5 mmHg during baseline and severe hypoxia, respec- frequency of RSNA (10 Ϯ 2%; P ϭ 0.002; Figure 1). MAP did tively (Figure 1). Heart rate increased significantly during J Am Soc Nephrol 13: 27–34, 2002 Renal Vascular Responses to Graded RSNA 31

Figure 3. Plasma renin activity in the individual rabbits (●) and for group mean (mean Ϯ SEM; ●) during the baseline period and during the administration of the gas mixture (room air, 14% O2 [moderate Ͻ hypoxia], or 10% O2 [severe hypoxia]) in the three groups. *, P 0.05 (significant change from baseline during the gas period). severe hypoxia, from 230 Ϯ 10 beats/min to 242 Ϯ 12 beats/ ϭ Ϯ min (P 0.04). PaCO2 levels were 22 2 mmHg during Ϯ ϭ baseline and 24 2 mmHg during 10% O2 (P 0.2). pH levels were 7.37 Ϯ 0.021 during baseline and 7.35 Ϯ 0.024 ϭ during 10% O2 (P 0.3). PRA rose in response to severe hypoxia, levels being 5.9 Ϯ 1.1 ng AngI/ml per h and 9.42 Ϯ 1.4 ng AngI/ml per h during baseline and severe hypoxia, respectively (P ϭ 0.017; Figure 3). ERBF decreased by 56 Ϯ 8% (P ϭ 0.005; Figure 2), and renal vascular conductance decreased by 60 Ϯ 10% (P ϭ 0.04; Figure 2) in response to severe hypoxia. GFR decreased sig- nificantly in response to severe hypoxia, falling from 4.7 Ϯ 0.9 ml/min during the baseline period to 2.0 Ϯ 0.5 ml/min during severe hypoxia (P ϭ 0.001; Figure 2), but filtration fraction was unchanged (33 Ϯ 4 versus 32 Ϯ 4%, respectively; P ϭ 0.6; Figure 2). Urine flow rate decreased significantly in re- sponse to severe hypoxia, falling from 0.16 Ϯ 0.04 to 0.11 Ϯ ϭ 0.02 ml/min (P 0.008; Figure 2), as did sodium excretion Figure 4. (Left panels) Values (mean Ϯ SEM) for the renal variables rate, falling from 20 Ϯ 7to7Ϯ 2 ␮mol/min (P ϭ 0.022). The measured during the baseline period (room air) and administration of ● ⅜ decrease in fractional sodium excretion in response to severe the gas mixture (room air; ) and moderate (14% O2; ) and severe Ϯ hypoxia was not statistically significant (45 12% versus 25 (10% O2; ) hypoxia. (Right panels) For each adjacent variable in the Ϯ 5%; P ϭ 0.09). left panel, the percentage of change from baseline during the admin- SFP was 21.1 Ϯ 0.2 mmHg during the baseline period and istration of each gas mixture (room air [e], moderate [u], and severe fell to 18.5 Ϯ 0.3 mmHg during severe hypoxia (P ϭ 0.001; [f] hypoxia) for each variable (mean Ϯ SEM). *, P Ͻ 0.05; and **, Ͻ Figure 4). Plasma protein concentration was not different dur- P 0.01. Both values indicate a significant change from the respec- Ϯ Ϯ ϭ tive baseline value. ing baseline and 10% O2 (3.9 0.1 versus 3.8 0.1 g%, P 0.3). Estimated glomerular capillary pressure decreased by 3.0 Ϯ ϭ 0.5 mmHg during 10% O2 (P 0.001; Figure 4). Proximal Ϯ Ϯ tubule pressure was 11.0 0.7 mmHg during baseline and icantly decreased during 10% O2, being 0.85 0.20 ml/min Ϯ ϭ Ϯ 10.9 0.3 mmHg during 10% O2 (P 0.9). Estimated per mmHg during baseline and 0.40 0.10 ml/min per mmHg ϭ preglomerular resistance was increased in response to severe during 10% O2 (P 0.034; Figure 4). hypoxia, being 2.7 Ϯ 0.4 mmHg/ml per min and 7.0 Ϯ 1.8 mmHg/ml per min, respectively (P ϭ 0.05; Figure 4). Post- Recovery from Hypoxia glomerular resistance also increased significantly, being 2.5 Ϯ After a return to breathing room air, the rabbits were fol- 0.3 mmHg/ml per min during baseline and 6.9 Ϯ 1.5 mmHg/ml lowed for 30 min. During the final 10 min of this period, per min during moderate hypoxia (P ϭ 0.03; Figure 4). The measurements were taken. Measurements of MAP, heart rate, preglomerular to postglomerular resistance ratio was not dif- PaO2, PaCO2, and RSNA (total, frequency, or amplitude) were Ϯ ferent from baseline during 10% O2 (1.15 0.07 versus 1.19 made in all rabbits, but clearance measurements were only Ϯ ϭ 0.17; P 0.3). The rabbits were all in a state of filtration made in five rabbits exposed to room air or 14% O2 and four ⌬ Ͼ ␲ disequilibrium (i.e. P e) during both periods of the rabbits exposed to 10% O2. All variables, MAP, heart rate, experiments; therefore, Kf could be calculated. Kf was signif- PaO2, PaCO2, RSNA (total, frequency, or amplitude), RVR, 32 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 27–34, 2002

ERBF, GFR, filtration fraction, urine flow rate, and sodium increase in preglomerular vascular resistance and decreased excretion rate had returned toward baseline values in all three filtration (1–3,28,29). This is in contrast to the present study, in groups (Figures 1 and 2). which a 25% decrease in renal blood flow evoked reflexly in response to moderate hypoxia caused greater postglomerular Discussion vasoconstriction and maintenance of GFR. We suggest that an The two levels of hypoxia caused graded increases in total explanation lies in the way in which the increases in renal RSNA and decreases in renal blood flow in these experiments, nerve activity in the two circumstances are evoked. Electrical with no change in MAP (11,13,23). The increases in total stimulation at supramaximal voltages results in the synchro- RSNA evoked by both moderate and severe hypoxia were due nous firing of all renal sympathetic nerves, with the frequency almost entirely to increases in the amplitude of RSNA rather of firing being varied. This does not appear to be the case than to increases in the frequency of the bursts (11,13,23). The physiologically, where it is an increase in the amplitude of amplitude of the sympathetic nerve activity signal reflects the RSNA (recruitment) that occurs in response to a wide range of number of activated fibers (6–17); therefore, an increase in the physiologic stimuli, including hypoxia (22). In this regard, it is amplitude of the signal reflects recruitment of additional nerve interesting to note that some recent studies that have attempted fibers in response to hypoxia (11). The conditions were there- to mimic the bursting pattern of renal nerve activity have fore established to test the aim of this study of how changes in elicited different renal responses to that of standard square renal nerve recruitment affect preglomerular and postglomeru- wave stimuli (15,17,30). lar vascular resistances. The glomerular ultrafiltration coefficient (Kf) calculated in This study demonstrated that graded physiologic activation this study by use of whole-kidney GFR values should be of the renal nerves evokes different preglomerular and post- regarded with caution. Although significant changes in Kf glomerular responses. In response to moderate hypoxia, the when calculated in this manner are likely to be real, small

RSNA amplitude increased by 20% and renal blood flow fell changes in Kf are unlikely to be detected. The fall in Kf in by 25%. Despite this marked renal vasoconstriction, GFR was response to severe hypoxia is in accord with reports elsewhere well maintained; therefore, filtration fraction was increased. of decreased Kf in response to electrical stimulation (1–4). The renal vasoconstriction was due to increases in both pre- Whether this effect is due to a direct neural action on the glomerular (ϳ20%) and postglomerular (ϳ70%) resistance, mesangium, as suggested by Kon and Ichikawa (2), or the and a small increase in glomerular capillary pressure, as esti- result of an indirect action of a vasoactive agent remains mated via stop-flow pressure measurement, was seen. During unanswered. However, in the three-dimensional ultrastructral severe hypoxia (10% O2), when RSNA amplitude increased by analysis of the innervation of the juxtaglomerular region by 60%, there was even greater renal vasoconstriction (56%) and Luff et al. (5), no axons were seen close to the extraglomerular now GFR also fell, but filtration fraction was unchanged. Both mesangium, making the decrease in Kf unlikely to be due to a estimated preglomerular (ϳ240%) and postglomerular direct neural action. (ϳ250%) resistance increased, but the glomerular ultrafiltra- The results of this study support our hypothesis that reflex tion coefficient decreased by almost 50%. These results dem- changes in renal nerve activity may differentially evoke onstrate that lower levels of increased renal nerve activation, changes in preglomerular and postglomerular vascular resis- evoked reflexly, produced a greater increase in postglomerular tances by activating different populations of nerves (22). As vascular resistance than in preglomerular resistance. This re- indicated in the introduction, we have previously shown that sponse would tend to conserve GFR, despite renal vasocon- two types of axons, types I and II, innervate the afferent and striction. Higher levels of RSNA, evoked reflexly by severe efferent arterioles, with type I axons almost exclusively inner- hypoxia, resulted in a greater preglomerular component to the vating the afferent arteriole, whereas type II axons are distrib- renal vasoconstriction and a reduction in the glomerular ultra- uted at similar innervation densities to both afferent and effer- filtration coefficient, with consequent reductions in GFR. ent arterioles (24). Thus, activation of type I nerves would Thus, in severe hypoxia, there is both marked renal vasocon- cause almost exclusive preglomerular resistance increases. In striction and loss of GFR. contrast, we suggest that type II axon stimulation would be The renal responses to reflexly induced increases in nerve expected to constrict both the afferent and efferent arterioles. activity via hypoxia were different from those seen previously In turn, this would be predicted to result in a greater increase in response to electrical stimulation of the renal nerves (1–4). in postglomerular resistance than preglomerular resistance, on Postglomerular increases in resistance were as great or greater the basis of a consideration of Poiseuille’s relationship, as we for both levels of hypoxia than preglomerular responses, unlike have argued elsewhere (31). Briefly, the resting diameter of the the situation with electrical stimulation, in which greater pre- efferent arteriole is less than that of the afferent arteriole; glomerular effects have been documented to all but the lowest therefore, similar changes in afferent and efferent arteriole stimulation frequencies (1–4). Particularly, the responses to radius in response to increased RSNA would increase resis- moderate hypoxia were at odds with previous whole kidney tance to a greater extent in the efferent arteriole (resistance and micropuncture studies in which the renal nerves were being proportional to the fourth power of the radius [31]). electrically stimulated (1–4,28,29). Electrical stimulation stud- It is important to emphasize that although the present results ies have demonstrated that a stimulus sufficient (ϳ3 Hz) to are compatible with our hypothesis, there are a number of reduce renal blood flow by 20% to 30% causes a predominant important considerations that require further work. First, it is J Am Soc Nephrol 13: 27–34, 2002 Renal Vascular Responses to Graded RSNA 33 possible that some of the actions within the kidney are due to rabbit with superficial glomeruli, directly measured glomerular the hypoxia itself. However, we have demonstrated elsewhere capillary pressures were 30 mmHg in response to room air, 35 that in conscious and anesthetized rabbits, the renal response to to 38 mmHg in response to 14% hypoxia, and 25 mmHg to similar levels of hypoxia are due to the reflex activation of the 10% hypoxia, in complete agreement with the stop-flow esti- sympathetic nerves, because renal denervation abolishes the mates of glomerular capillary pressure. Importantly, arterial renal responses (11,23,32). We are therefore confident that pressure was unaffected by the hypoxic stimuli; therefore, changes in preglomerular and postglomerular vascular resis- autoregulatory changes in renal blood flow and GFR have not tances are due to reflex increases in sympathetic nerve activity confounded the interpretation of the results. As reported else- and that there are no significant direct effects of hypoxia on the where (12), a further methodological consideration is that renal circulation. The rapid return of all variables to control setting the threshold for the determination of nerve activity has levels during the recovery phase of the experiment strengthens the potential to bias the results. Therefore, as indeed we saw in this conclusion. this study, an increase in the amplitude of small bursts of renal Second, apart from the direct actions of renal nerves on the nerve activity, previously just below the threshold to above renal vasculature, indirect vasoconstriction via activation of threshold, can result in an apparent increase in the frequency of vasoactive agents may occur. In particular, the renin-angioten- nerve activity. sin system plays an important role in modulating the effects of In conclusion, we have shown that reflex activation of the the renal sympathetic nerves (33) and may be directly respon- sympathetic nerves by hypoxia produces different preglomeru- sible in part for the changes seen in preglomerular and post- lar and postglomerular effects, depending on the severity of the glomerular resistances in response to hypoxia. Studies that hypoxia and extent of increase in RSNA amplitude. Moderate have examined the effect of AngII inhibition on the response to hypoxia increased postglomerular resistance predominantly, electrical stimulation of the renal nerve have shown conflicting whereas more severe hypoxia markedly increased preglomeru- findings. Pelayo et al. (1) found that both afferent and efferent lar and postglomerular resistance. These results provide evi- vasoconstrictor responses to electrical stimulation were atten- dence that the renal sympathetic nerves can differentially reg- uated during AngII inhibition, whereas Handa and Johns (34) ulate preglomerular and postglomerular resistances. Thus, inferred from whole kidney studies that AngII was predomi- glomerular pressure and filtration may be differentially regu- nantly responsible for the constriction of the efferent arteriole lated, depending on the nature of change in renal nerve activity. in response to electrical stimulation. Neylon et al. (35,36) examined the renal responses to hypoxia in rats; however, RSNA was not measured, and marked changes in arterial Acknowledgments pressure, PaCO , and pH occurred in response to hypoxia, We acknowledge and wish to thank Rebecca Flower for technical 2 assistance with the animal experiments and Katrina Worthy for per- confounding the results. In this study, moderate hypoxia did forming the RIA. This study was supported by National Health and not result in a significant change in plasma renin activity. On Medical Research Council Australia Grant 97/0426. this basis, the observed predominant postglomerular effects of increased RSNA in response to moderate hypoxia do not appear to be due to the actions of AngII. However, changes in References plasma renin levels may not reflect intrarenal AngII levels; 1. Pelayo JC, Ziegler MG, Blantz RC: Angiotensin II in adrenergic- thus, this conclusion requires caution. Further experiments in induced alterations in glomerular hemodynamics. Am J Physiol AngII-inhibited animals to elucidate the contribution of the 247: F799–F807, 1984 renin-angiotensin system to the effects of reflex activation of 2. Kon V, Ichikawa I: Effector loci for renal nerve control of the renal sympathetic nerves should be conducted. Similarly, it cortical microcirculation. Am J Physiol 245: F545–F553, 1983 3. Kon V: Neural control of renal circulation. Miner Electrolyte is possible that hypoxia is altering catecholamine or adenosine Metab 15: 33–43, 1989 release, which may indirectly influence the renal circulation 4. Tucker BJ, Peterson OW, Munger KA, Bird JE, Mitchell M, (33). Pelayo JC, Blantz RC: Glomerular hemodynamic alterations Finally, the experimental setting is an important consider- during renal nerve stimulation in rats on high- and low-salt diets. ation and potentially confounding. In this anesthetized rabbit Am J Physiol 258: F133–F143, 1990 preparation, arterial BP and kidney function are closely 5. Luff SE, Hengstberger SG, McLachlan EM, Anderson WP: matched to that in the conscious rabbit (25). Estimated glo- Distribution of sympathetic neuroeffector junctions in the juxta- merular capillary pressures in this study were comparable to glomerular region of the rabbit kidney. J Auton Nerv Syst 40: those measured previously in rabbits, by use of both stop-flow 239–254, 1992 estimates and direct capillary puncture in superficial glomeruli 6. Ninomiya I, Malpas SC, Matsukawa K, Shindo T, Akiyama T: (25,37,38). It is recognized that the use of SFP to estimate The amplitude of synchronized cardiac sympathetic nerve activ- ity reflects the number of activated pre- and postganglionic fibers glomerular capillary pressure theoretically has inherent prob- in anesthetized cats. J Auton Nerv Syst 45: 139–147, 1993 lems, in that flow is interrupted to the macula densa, which 7. Malpas SC, Ninomiya I: Effect of asphyxia on the frequency and may cause afferent arteriole vasodilatation and reduce myo- amplitude modulation of synchronized renal nerve activity in the genic responsiveness. This may have led to underestimation of cat. J Auton Nerv Syst 40: 199–205, 1992 preglomerular resistance. However, a detailed study has shown 8. Malpas SC, Ninomiya I: The amplitude and periodicity of syn- that this effect is small (39). Also, in this study in a single chronized renal sympathetic nerve discharges in anesthetized 34 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 27–34, 2002

cats: Differential effect of baroreceptor activity. J Auton Nerv rioles of the rabbit and rat kidney differ structurally from those Syst 40: 189–198, 1992 supplying other arteries. J Neurocytol 20: 781–795, 1991 9. Malpas SC, Ninomiya I: A new approach to analysis of synchro- 25. Denton KM, Anderson WP: Glomerular ultrafiltration in rabbits nized sympathetic nerve activity. Am J Physiol 263: H1311– with superficial glomeruli. Pflugers Arch 419: 235–242, 1991 H1317, 1992 26. Oliver J, Korner P, Woods R, Zhu J: Reflex release of vasopres- 10. Malpas SC, Ninomiya I: Effect of chemoreceptor stimulation on sin and renin in hemorrhage is enhanced by autonomic blockade. the periodicity of renal sympathetic nerve activity in anesthetized Am J Physiol 258: H221–H228, 1990 cats. J Auton Nerv Syst 37: 19–28, 1992 27. Deen WM, Robertson CR, Brenner BM: A model of glomerular 11. Malpas SC, Shweta A, Anderson WP, Head GA: Functional ultrafiltration in the rat. Am J Physiol 223: 1178–1183, 1972 response to graded increases in renal nerve activity during hyp- 28. Hesse IF, Johns EJ: The effect of graded renal nerve stimulation oxia in conscious rabbits. Am J Physiol 271: R1489–R1499, on renal function in the anaesthetized rabbit. Comp Biochem 1996 Physiol A 79: 409–414, 1984 12. Malpas SC, Bendle RD, Head GA, Ricketts JH: Frequency and 29. Hermansson K, Ojteg G, Wolgast M: The cortical and medullary amplitude of sympathetic discharges by baroreflexes during hyp- blood flow at different levels of renal nerve activity. Acta Physiol oxia in conscious rabbits. Am J Physiol 271: H2563–H2574, Scand 120: 161–169, 1984 1996 30. Malpas SC, Hore TA, Navakatikyan M, Lukoshkova EV, 13. Malpas SC, Evans RG: Do different levels and patterns of Nguang SK, Austin PC: Resonance in the renal vasculature sympathetic activation all provoke renal vasoconstriction? J Au- evoked by activation of the sympathetic nerves. Am J Physiol ton Nerv Syst 69: 72–82, 1998 276: R1311–R1319, 1999 14. Malpas SC, Evans RG, Head GA, Lukoshkova EV: Contribution 31. Denton KM, Anderson WP, Sinniah R: Effects of angiotensin II of renal nerves to renal blood flow variability during hemor- on regional afferent and effernet arteriole dimensions and the rhage. Am J Physiol 274: R1283–R1294, 1998 glomerular pole. Am J Physiol 279: R629–R638, 2000 15. Malpas SC: The rhythmicity of sympathetic nerve activity. Prog 32. Leonard BL, Malpas SC, Denton KM, Madden AC, Evans RG: Neurobiol 56: 65–96, 1998 Differential control of intrarenal blood flow during reflex in- 16. Dibona GF: Differentiation of vasoactive renal sympathetic creases in sympathetic nerve activity. Am J Physiol 280: R62– nerve fibres. Acta Physiol Scand 168: 195–200, 2000 R68, 2001 17. Dibona GF, Sawin LL: Functional significance of the pattern of 33. DiBona GF, Kopp UC: Neural control of renal function. Physiol renal sympathetic nerve activation. Am J Physiol 277: R346– R353, 1999 Rev 77: 75–197, 1997 18. DiBona GF: Neural control of the kidney: Functionally specific 34. Handa RK, Johns EJ: Interaction of the renin-angiotensin system renal sympathetic nerve fibers. Am J Physiol 279: R1517–R1524, and the renal nerves in the regulation of rat kidney function. 2000 J Physiol (Lond) 369: 311–321, 1985 19. Gebber GL, Barman SM: Basis for 2–6 cycle/s rhythm in sym- 35. Neylon M, Marshall J, Johns EJ: The role of the renin-angioten- pathetic nerve discharge. Am J Physiol 239: R48–R56, 1980 sin system in the renal response to moderate hypoxia in the rat. 20. DiBona GF, Jones SY: Analysis of renal sympathetic nerve J Physiol 491: 479–488, 1996 responses to stress. Hypertension 25: 531–538, 1995 36. Neylon M, Marshall JM, Johns EJ: The effects of systemic 21. DiBona GF, Sawin LL: Renal hemodynamic effects of activation hypoxia on renal function in the anaesthetized rat. J Physiol of specific renal sympathetic nerve fiber groups. Am J Physiol (Lond) 487: 497–511, 1995 276: R539–R549, 1999 37. Denton KM, Anderson WP: Intrarenal haemodynamic and glo- 22. Malpas S, Leonard BL: Neural regulation of renal blood flow: A merular responses to inhibition of nitric oxide formation in re-examination. Clin Exp Pharmacol Physiol 27: 956–964, 2000 rabbits. J Physiol (Lond) 475: 159–167, 1994 23. Leonard BL, Evans RG, Navakatikyan MA, Malpas SC: Differ- 38. Denton KM, Anderson WP: Vascular actions of endothelin in the ential neural control of intrarenal blood flow. Am J Physiol 279: rabbit kidney. Clin Exp Pharmacol Physiol 17: 861–872, 1990 R907–R916, 2000 39. Romano G, Favret G, Federico E, Bartoli E: Measurement of 24. Luff SE, Hengstberger SG, McLachlan EM, Anderson WP: Two nephron filtration from distal and proximal tubules. J Nephrol types of sympathetic axon innervating the juxtaglomerular arte- 11: 211–215, 1998