Continuously measured renal blood flow does not increase in diabetes if nitric oxide synthesis is blocked

Abstract

This study used 16 h/day measurement of renal blood flow (RBF) and arterial pressure (AP) to determine the role of nitric oxide (NO) in mediating the renal vasodilation caused by onset of type 1 diabetes. The AP and RBF power spectra were used to determine the autoregulatory efficiency of the renal vasculature. Rats were instrumented with artery and vein catheters and a Transonic flow probe on the left renal artery and were divided randomly into four groups: control (C), diabetes (D), control plus nitro-l-arginine methyl ester (l-NAME; CL), and diabetes plus l-NAME (DL). Mean AP averaged 90 ± 1 and 121 ± 1 mmHg in the D and DL groups, respectively, during the control period, and RBF averaged 5.9 ± 1.2 and 5.7 ± 0.7 ml/min, respectively. Respective C and CL groups were not different. Onset of diabetes (streptozotocin 40 mg/kg iv) in D rats increased RBF gradually, but it averaged 55% above control by day 14. In DL rats, on the other hand, RBF remained essentially constant, tracking with RBF in the nondiabetic C and CL groups for the 2-wk period. Diabetes did not change mean AP in any group. Transfer function analysis revealed impaired dynamic autoregulation of RBF overall, including the frequency range of tubuloglomerular feedback (TGF), and l-NAME completely prevented those changes as well. These data strongly support a role for NO in causing renal vasodilation in diabetes and suggest that an effect of NO to blunt RBF autoregulation may play an important role.

the mechanism for renal vasodilation in early diabetes mellitus has generated considerable interest because of its effect to allow transmission of systemic arterial pressure into the glomerulus, thus facilitating hemodynamic-mediated renal injury (32, 33, 55). However, despite the general renal vasodilatory effect of nitric oxide (NO), there is evidence both for (5, 8, 12, 17, 28) and against (1, 37, 38, 42) NO as a primary determinant of the increased renal blood flow (RBF) and glomerular filtration rate (GFR) in early experimental diabetes. The explanation for why there are opposing findings is not clear, but it could be due to different time points in diabetes at which GFR is measured, the degree of hyperglycemia, anesthetized versus awake preparations, or the level of reactive oxygen species. Another potential explanation, however, could be the methods used to measure renal hemodynamics.

All studies to date, including many previous studies from our laboratory (5, 8, 17), have used acute measurements of RBF and GFR primarily with clearance techniques. One limitation is that those measurements are a snapshot of renal function over a period of minutes, and in a chronic disease such as diabetes, all the factors noted above can come into play and contribute to the potential for wide variation in results among studies that measure renal function from the first several days to several months after induction of diabetes. We recently developed the technique for measuring RBF and arterial pressure (AP) for up to 24 h/day in undisturbed conscious rats during control conditions and through the first 2 wk of type 1 diabetes (2). We measured a sustained increase in RBF immediately following induction of diabetes, and in the present study, we used that same technique to test the hypothesis that the increase in RBF is dependent on nitric oxide (NO).

In addition, we used transfer function analysis of the daily AP and RBF signals as a means to evaluate the role of NO in RBF autoregulation in this chronic experimental animal model. An advantage of using dynamic autoregulation to study autoregulation of RBF is that repeated measurements can be made in the same animal in the conscious, undisturbed state over the time course of the entire study (4, 18). Previously we used this method to show that dynamic autoregulation in the tubuloglomerular feedback (TGF) frequency range was impaired (2). This is a novel approach in a chronic experimental animal model for gaining insight into the potential intrarenal mechanisms through which NO may control RBF in diabetes, particularly because we analyzed 16-h RBF and AP records each day in each rat.

METHODS

The experiments were conducted in male Sprague-Dawley rats (350–375 g; Harlan Sprague-Dawley, Madison, WI), and protocols were approved by the Institutional Animal Care and Use Committee. Anesthesia was induced with pentobarbital sodium (50 mg/kg), and atropine was administered (40 μg ip per rat) to minimize airway secretions. Under aseptic conditions, a laparotomy was performed, and a 1-mm Transonic flow probe (Transonic Systems) was placed on the artery of the left kidney. Although care was taken to avoid denervation, fat was cleared from the renal artery using a cotton swab to enable an unimpeded ultrasonic signal. The right kidney was not removed. A nonocclusive polyvinyl catheter then was implanted in the abdominal aorta, below the renal artery, by inserting the catheter tip through a puncture wound in the aortic wall made with the tip of an L-shaped 18-gauge needle. The insertion point was sealed with cyanoacrylate adhesive, and the catheter and flow probe cable were exteriorized through the lateral abdominal wall. A femoral vein catheter was implanted through a separate incision. The flow probe leads and catheters were routed subcutaneously to the scapular region and exteriorized through a Dacron-covered plastic button sutured subcutaneously over the scapulae.

The rats were allowed to recover from surgery and then were placed in individual metabolic cages in a quiet, air-conditioned room with a 12:12-h light-dark cycle. The catheters and flow probe leads were passed through a stainless steel spring that was attached to the button, and the opposite end of the spring was connected to a customized adapter on an electrical swivel (Airflyte Electronics) mounted above the cage. The four flow probe leads were soldered to the swivel wires, and the artery and vein catheters were passed through a hole in the center of the electrical swivel to a dual-channel hydraulic swivel (Instech) above. A customized connector joined the two swivels so that both would turn synchronously with rat movement. The venous catheter was connected, via the hydraulic swivel, to a syringe pump (Harvard Apparatus) that ran continuously throughout the study. All solutions were infused through a Millipore filter (0.22 μm; Cathivex; Millipore). The arterial catheter was filled with heparin solution (1,000 USP U/ml) and connected, also via the hydraulic swivel, to a pressure transducer for continuous measurement of arterial pressure. The flow probe connector was soldered to the electrical swivel wires and connected to a Transonic model T206 flowmeter (Transonic Systems) for continuous measurement of RBF. The pulsatile flow signals from the flowmeter and the amplified pulsatile arterial pressure signals (CB Sciences) were sampled at 100 Hz, 16 h/day throughout the experiment, using PowerLab and a Macintosh computer.

Total sodium intake throughout the experiment was maintained constant at ∼3.1 mmol/day by continuous intravenous infusion of 20 ml/day sterile 0.9% saline combined with sodium-deficient rat chow (0.006 mmol sodium/g; Teklad). A sodium-deficient diet ensured that the daily sodium intake could be controlled precisely at normal levels by the infusion. This infusion was begun immediately after placement of the rat in the metabolic cage, and 3–5 days were allowed for acclimation.

Experimental Protocol

After recovery from surgery and acclimation, the rats were divided randomly into four groups: control (C; n = 7), diabetes (D; n = 6), control plus nitro-l-arginine methyl ester (l-NAME; CL; n = 7), and diabetes plus l-NAME (DL; n = 8). After baseline measurements, l-NAME (10 μg·kg⁻¹·min⁻¹ iv) (5, 17, 21, 30) was added to the infusate of the CL and DL rats and maintained throughout the remainder of the study. Six days later, type 1 diabetes was induced in the two diabetic groups with streptozotocin (STZ; 40 mg/kg iv). The diabetes experimental period lasted for 2 wk. Insulin was added to the daily saline infusion syringe if needed to keep blood glucose from exceeding 450 mg/dl. On day 4 of the control period and once per week during the 2-wk diabetes period (days 3 or 4 and day 12), 1.3 ml of arterial blood was collected from the arterial catheter for measurement of GFR, plasma renin activity (PRA), hematocrit, glucose concentration, and plasma protein and electrolyte concentrations. Samples were replaced with an equal volume of 0.9% saline.

Analytical Methods

GFR was measured after a 24-h intravenous infusion of [¹²⁵I]iothalamate (Glofil; ∼20 μCi). Because steady state is achieved during the 24-h infusion, the isotope infusion rate was substituted for urinary isotope excretion rate to calculate clearance (3). Urinary sodium and potassium concentrations were determined with ion-sensitive electrodes (Synchron El-ise; Beckman Coulter), and urine glucose was documented with a Bayer Clinitek 50 dipstick analyzer. PRA was measured by radioimmunoassay (Diasorin). Data were analyzed using a two-factor analysis of variance (ANOVA) with repeated measures, with Dunnett's test used for within-group comparisons over time. For between-group comparisons, completely randomized ANOVA tests were run for each day and supplemented with Scheffé's test to determine specific group differences. P < 0.05 was considered statistically significant, and data are means ± SE.

Transfer function analysis.

Daily data files covering the time period from ∼1300 to 0500 (16 h) were processed off-line using previously developed software routines (16) written for Matlab (version 7, release 14). Data from day 8 were used for the day 7 data in some cases if the day 7 file was truncated due to catheter clotting or probe malfunction. The 100-Hz data files were digitally low-pass filtered (3.5-Hz cut-off frequency, finite-impulse response, order 50) and decimated to 5 Hz. These 5-Hz data files were split into blocks of 4,096 data points, yielding a frequency discrimination of 0.001 Hz. Power spectral density of AP and RBF was calculated using Welch's method [with units of mmHg²/Hz and (ml/min)²/Hz, respectively]. The transfer function spectra were calculated from AP (input) and RBF (output). The transfer function was taken as the quotient of the cross spectrum of input and output divided by the power spectrum of the input. These algorithms involved mean detrending and a Hanning window with 50% overlap of the blocks. To permit comparison among rats, we normalized the transfer function gain (magnitude) values over the frequency range to the mean renal vascular conductance over the duration of the data set. After conversion of the normalized transfer function gain values into decibels [20 log(gain)], a mean spectrum was calculated from the consecutive spectra in each rat, and these were subsequently averaged for all rats. Curves were compared using ANOVA followed by unpaired t-tests.

RESULTS

Renal blood flow averaged 6.8 ± 0.8 and 6.2 ± 0.8 ml/min in the L and DL groups, respectively, during baseline measurements made before beginning the control period (data not shown) and averaged 5.5 ± 0.7 and 5.7 ± 0.7 ml/min, respectively, under the influence of the chronic l-NAME infusion during the control period [days 4–6 of the control period (C4–C6) are plotted in Fig. 1 as a percentage of control]. RBF averaged 6.1 ± 0.7 and 5.9 ± 1.2 ml/min in the C and D groups, respectively, during the control period and also is plotted as percent control values for C4–C6 in Fig. 1. Mean arterial pressure (MAP) was higher during the control period in the l-NAME-infused rats, averaging 118 ± 1 and 121 ± 1 mmHg in CL and DL compared with 93 ± 1 and 90 ± 1 mmHg in the C and D groups, respectively (Fig. 1). Thus, although chronic l-NAME administration did not cause pronounced reductions in daily RBF, there were approximate 57 and 48% increases in renal vascular resistance caused by chronic l-NAME infusion in the L and DL groups, respectively, at baseline.

Fig. 1.

Mean arterial pressure (AP; top) and renal blood flow (RBF; bottom) in control (○), diabetic (•), control + nitro-l-arginine methyl ester (l-NAME; □), and diabetic + l-NAME rats (▪) during the last 3 control period days (C4–C6) and the 14-day diabetes period (D1–D14).

Administration of STZ increased blood glucose significantly in the D and DL groups (Table 1). RBF averaged 111 ± 3% of control in the D rats by the second day of diabetes, which, although not statistically significant, was similar to the initial response we reported previously with these methods (2). The plateau in RBF from days 3–6 of diabetes (D3–D6), however, differed from the continued rise we reported previously during week 1 (2), but RBF was increased significantly by D7 and was ∼55% above control levels by D14 (Fig. 1). In the DL rats, on the other hand, the renal vasodilatory response was eliminated completely, with RBF essentially remaining constant over the entire experimental period and tracking with RBF in the nondiabetic C and CL control groups (Table 1). Likewise, Fig. 2 shows that the increase in GFR caused by diabetes also was prevented in the DL rats, similar to our previous reports (5, 17). Diabetes did not change MAP in any group.

Table 1.

Blood glucose, PRA, hematocrit, plasma protein concentration, and urinary Na⁺ excretion during each study period in control, diabetic, control + l-NAME, and diabetes + l-NAME rats

Variable/Group	Experimental Period
	Control	DW1	DW2
Blood Glucose, mg/dl
C	131±4.0	123±1.7	124±5.0
D	131±3.2	477±22.2*†	475±17.5*†
CL	123±2.8†	132±9.1	124±1.8
DL	120±3.5†	429±46.2*†	391±47.1*†
PRA, ng ANG I·ml−1·h−1
C	1.6±0.7	1.1±0.7	1.2±0.1
D	0.8±0.4	1.5±1.2	2.1±0.8
CL	0.8±0.3	2.2±0.8*	1.9±0.8
DL	0.6±0.2#	1.8±1.2*	2.0±1.7
Hematocrit, %
C	41±0.7	42±0.6	40±2.7
D	40±2.7	42±0.9	39±3.1
CL	42±0.8	42±0.9	42±0.7
DL	43±0.7	43±0.6	42±0.7
Plasma Protein, g/dl
C	6.8±0.4	6.8±0.2	6.7±0.1
D	6.4±0.2	7.5±0.4*†	6.8±0.2
CL	6.7±0.1	6.4±0.1	7.0±0.2
DL	6.5±0.1	7.5±0.3*†	6.7±0.2
UNaV, mEq/day
C	2.3±0.5	2.8±0.8	2.8±0.1
D	2.9±0.1	3.4±0.1*†	3.6±0.1*†
CL	2.4±0.2	2.6±0.1	2.7±0.1*
DL	2.9±0.1	3.4±0.1*†	3.3±0.2*†

Values are means ± SE for blood glucose, plasma renin activity (PRA), hematocrit, plasma protein concentration, and urinary Na⁺ excretion (U_NaV) in control (C), diabetic (D), control + nitro-l-arginine methyl ester (l-NAME; CL), and diabetic + l-NAME (DL) rats during control period and weeks 1 (DW1) and 2 (DW2) of diabetes period.

^*P < 0.05 for within-group comparisons vs. control period value.

^†P < 0.05 for between-group comparison with control group during that period.

Fig. 2.

Glomerular filtration rate (GFR) in control, diabetic, control + l-NAME, and diabetic + l-NAME rats during the control period and weeks 1 (DW1) and 2 (DW2) of the diabetes period. *P < 0.05 for within-group comparison vs. control.

Effects of Chronic NO Blockade on Spectral Analysis of AP and RBF

Power spectra, coherence, gain, and phase for the C group are shown in Fig. 3. Peak coherence occurred at 0.3–0.5 Hz, consistent with pressure-passive dynamics in this region of the spectrum, and coherence was reduced at frequencies below 0.2 and 0.04 Hz, consistent with operation of autoregulation. Gain was positive above 0.2 Hz due to renal arterial compliance. There was a discrete reduction in gain below 0.2 Hz and a local maximum in gain between 0.04 and 0.06 Hz. Phase approached zero at 0.3–0.5 Hz, and there was a strong positive phase peak coincident with the gain reduction below 0.2 Hz and another phase peak centered near 0.03 Hz. The pattern of gain reduction in association with a positive phase peak indicates an autoregulatory mechanism that operates to attenuate the AP-induced RBF fluctuations within its operating frequency band. A zero phase angle indicates simultaneous AP and RBF fluctuations so that a pressure-passive behavior of the vasculature is reflected by a slightly positive but near-zero phase angle.

Fig. 3.

Power spectra for arterial pressure (AP) and renal blood flow (RBF), coherence, gain, and phase in the control rats. In each panel, the solid black line represents day 5 of the control period (C5), the dashed black line represents day 6 of the control period (C6), the solid red line represents day 1 of the diabetes period (D1), and the dashed red line represents day 7 of the diabetes period (D7).

Effect of diabetes.

After STZ administration, neither AP nor RBF were significantly affected on D1 (Fig. 1), and the spectral parameters were unchanged from those observed on C5 and C6 (Fig. 4). On D7, RBF was increased significantly and RBF power near 0.2 Hz was approximately twice as large as on D1; however, that difference did not achieve statistical significance. Instead of the normal decrease in coherence below 0.3 Hz seen on C5, C6, and D1, coherence remained above 0.5 Hz below 0.3 Hz; this close relationship between AP and RBF fluctuations results from the absence of autoregulation. In the gain spectrum, instead of the steep reversal from positive to negative gain values between 0.1 and 0.3 Hz (myogenic) as well as the local gain maximum at 0.02–0.05 Hz (TGF) observed on C5, C6, and D1, gain remained positive and declined slowly to zero, reflecting a pressure-passive behavior dependent on arterial compliance (i.e., lack of autoregulatory behavior). In contrast to the normal presence of positive phase peaks at frequencies where gain reduction is maximal (0.1–0.3 and 0.02–0.06 Hz) observed on C5, C6, and D1, phase was significantly different for D7, remaining at low positive values consistent with the absence of autoregulatory behavior.

Fig. 4.

Power spectra for AP and RBF, coherence, gain, and phase in the diabetes mellitus (DM) rats.

Effect of l-NAME.

Commensurate with the hypertension, l-NAME increased AP spectral power in both L and DL rats (from 0.001 to 0.01 Hz; data not shown). In addition, compared with control rats (Fig. 3), gain above 0.2 Hz was increased in control/l-NAME rats (Fig. 5), and gain below 0.2 Hz was significantly more negative compared with control over the frequency range 0.04–0.1 Hz. Phase peak from 0.08 to 0.3 Hz also increased. Similar observations have been made in previous transfer function analysis studies (26, 53) and are taken to indicate a strengthening of the myogenic mechanism following NO synthase (NOS) inhibition. When diabetes was induced in l-NAME-treated rats, not only was the increase in RBF prevented but also the abnormal coherence, gain, and phase spectra measured on D7 in the diabetic rats (Fig. 4) were eliminated (Fig. 6).

Fig. 5.

Power spectra for AP and RBF, coherence, gain, and phase in the control/l-NAME rats.

Fig. 6.

Power spectra for AP and RBF, coherence, gain, and phase in the DM/l-NAME rats.

DISCUSSION

The main finding from this study is that the increase in RBF caused by the onset of diabetes was prevented by chronic blockade of NO synthesis. In addition, blocking NO synthesis prevented diabetes from impairing dynamic autoregulation of RBF. These results not only strongly suggest that NO is required for diabetes-induced renal vasodilation but also suggest that the mechanism may involve attenuation of RBF autoregulation.

Understanding the mechanism for renal vasodilation early in diabetes is important because of its effect to allow transmission of systemic arterial pressure into the glomerulus, thus facilitating hemodynamic-mediated renal injury (32, 33, 55). NO is an attractive candidate to mediate the increase in GFR, but its potential role in that regard has been questioned. For example, Pflueger et al. (37, 38) reported that NO-dependent renal vasodilation actually was defective in STZ diabetic rats, and blockade of NO synthesis has been reported to have similar effects on GFR in diabetic and normal rats (1, 42). In addition, Keynan et al. (27) reported decreased neuronal NOS (nNOS) activity and protein in the renal cortex with no change in endothelial cell NOS (eNOS) activity and protein during the first week of STZ diabetes in rats. Thus there is evidence against NO-mediated renal vasodilation in diabetes, which has been attributed to decreased, or no change in, NOS activity or expression.

On the other hand, there is evidence that NO may play a significant role in mediating diabetic hyperfiltration and renal vasodilation. Increased NO metabolites have been reported in diabetic animals (12, 28, 31, 45, 50) and in human subjects (11, 20), and inhibition of NO synthesis has been reported to selectively block or attenuate diabetic hyperfiltration (12, 15, 28, 31, 45, 50). The reason for the disparate findings is not known, but it could be due to factors such as the level of hyperglycemia (44), the stage of diabetes at which measurements are made (39), or the relative balance between superoxide (13, 19, 34, 35, 41, 46) and NO under various experimental conditions. An alternative, but related, explanation could be the method for evaluating renal hemodynamics.

Until our recent report (2), all measurements of RBF and/or GFR in diabetes were acute measurements, made at isolated time points, primarily using clearance techniques, in either the awake or anesthetized state. However, we measured RBF 18 h/day using a Transonic flow probe and recorded RBF during control conditions continuing through the first 2 wk of diabetes in the same animal (2). That study revealed a rapid increase in RBF when diabetes was initiated that was sustained for the 2-wk study period and was accompanied by significant increases in GFR (2). The present study used the same experimental approach, confirmed the renal vasodilatory response to onset of diabetes, and, in addition, showed that the renal vasodilatory response was blocked completely when diabetes was induced in rats with chronic blockade of NO synthesis. The increase in GFR also was blocked. Continuous measurement of RBF in conscious, unrestrained, and undisturbed animals during control and experimental conditions is the most accurate and sensitive method for determining true minute-to-minute RBF throughout the day and from day to day. It removes questions related to the potential influence of anesthesia, stress from transport to the laboratory for measurements, and time of day at which measurements are made (most acute laboratory studies in rats are conducted during the time when rats normally would be asleep). In addition, by measuring RBF continuously for 2 wk, the method addresses questions about whether experiments at the 2-wk time point reflect renal function during the first several days of hyperglycemia, and vice versa. Thus the methods used in the present study provide very strong evidence that NO synthesis is required for diabetes to cause renal vasodilation.

One potential mechanism for that relationship could be through interaction of NO with reactive oxygen species. Reactive oxygen species are increased in diabetes and have been shown to decrease NO levels and impair NO activity in the kidney (13, 19, 34, 35, 41, 46). On the other hand, an important action of NO may be to blunt the increase in superoxide and thereby protect against its potential vasoconstrictor influence (5). Another mechanism through which NO could induce, or mediate, renal vasodilation in diabetes could be through direct effects on renal blood vessels. Although some studies have not measured increases in renal eNOS in diabetes (23, 42), others have measured increased eNOS in pre- (15, 36, 45) and postglomerular vessels (15, 36) from diabetic animals and in human glomerular endothelial cells from type 2 diabetic patients with increased GFR (20). However, the effect of nonspecific NOS inhibition with l-NAME to block increases in RBF and GFR in diabetes, as shown in the present study and previously (5, 8, 12, 17, 28), is not necessarily limited to effects on eNOS, because there also is strong evidence that nNOS may play a major role in controlling renal vascular resistance and GFR in diabetes (24, 29, 47).

Expression of nNOS is increased in the macula densa in diabetes (29), and selective nNOS inhibition has been shown to blunt the increase in GFR in diabetes (24, 29, 47). A particularly intriguing aspect of an nNOS effect on GFR and RBF is the likely involvement of TGF (47). However, the mechanism linking NO and TGF to the control of renal vascular resistance in diabetes is not clear. One hypothesis is that glucose-driven sodium transport in the proximal tubule essentially causes the macula densa to sense decreased sodium concentration despite an increase in total distal delivery, thus initiating an nNOS-dependent signal in the macula densa cells that decreases TGF-mediated vasoconstriction and thereby causes afferent arteriolar dilation (48, 51). That hypothesis invokes a normal TGF reflex that responds to increases and decreases in sodium chloride sensing by the macula densa, with respective increases and decreases in afferent arteriolar resistance. However, an alternative hypothesis invokes the reported effect of nNOS-derived NO to blunt the sensitivity of TGF (22, 49, 52, 54). Blunted TGF sensitivity means that there is less afferent arteriolar resistance for any given level of distal sodium chloride delivery, thus leading to increased RBF and GFR. Our dynamic autoregulation analyses reported previously (2) and in the present study suggest that the TGF reflex is blunted at the onset of diabetes, and our present data with l-NAME suggest that NO mediates that effect.

In addition, there is evidence that NO strongly modulates myogenic autoregulation and that TGF is involved in that modulation (26, 43), and an apparent effect of diabetes shown in Fig. 4 is the increase in gain over the TGF and myogenic frequency ranges. That effect was measured on D7, but not D1, of diabetes, and it is interesting that RBF was not increased significantly until D7. In our previous study (2), increased gain was measured on D1, and RBF also increased much sooner. The explanation for the different RBF responses is not apparent, but the delayed vasodilation that occurred in the present study actually strengthens the associations between increased RBF in diabetes and impairment in dynamic autoregulation, because it shows that one response is not present without the other. Moreover, Figs. 2 and 6 show that when NO synthesis was inhibited, none of the effects of diabetes on RBF or dynamic autoregulation occurred.

Because there are no other reports on the effect of chronic NO synthesis inhibition on daily RBF and dynamic autoregulation, it is important that the effect of l-NAME alone to enhance the myogenic mechanism that we measured (gain plots in Figs. 5 and 6) is similar to what has been reported in acute dynamic autoregulation studies (26, 53). This builds confidence in our findings on the effect of NO on dynamic autoregulation in diabetes. In addition, our transfer function data in this study were normalized to mean renal vascular conductance, rather than the value at 0-Hz frequency as in our previous study (2). Both methods revealed impaired dynamic autoregulation of daily RBF in diabetes, which strengthens that fundamental observation, but normalizing by the more conventional method using renal vascular conductance (4, 14, 18, 25, 53) in the present study also builds confidence in the new findings regarding the role of NO.

One effect of NO synthesis inhibition in this study that would seem to contradict our previous reports (5, 6, 9, 17) is the finding that MAP did not increase in the DL rats, even though the increase in GFR was prevented. However, that hypertensive response is critically dependent on increased renin-angiotensin system activity (6, 10, 40), and there was not a marked or sustained increase in PRA in the DL rats in this study. These results are, however, consistent with previous studies that have shown an effect of diabetes to increase RBF and GFR, and the use of 16 h/day measurement of RBF strongly supports studies that suggest NO is required for that effect. In addition, our dynamic autoregulation analysis suggests that an effect of NO to blunt RBF autoregulation may contribute to the renal vasodilatory action. The unique roles of TGF and nNOS in RBF control may explain why RBF is increased in diabetes at the same time that skeletal muscle flow is decreased (7). However, although the links among NOS inhibition, RBF, and the TGF mechanism in this study are consistent with a predominant nNOS effect, neither the expression or activity of those NOS isoforms were measured in this study.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-56259 and HL-75625. G. F. DiBona was supported by research grants from the Adlerbertska Forskningsstiftelsen, The Royal Society of Arts and Sciences, Göteborg, Sweden, and Karolinska Institute, Stockholm, Sweden.