Renal vasoconstriction by vasopressin V_1a receptors is modulated by nitric oxide, prostanoids, and superoxide but not the ADP ribosyl cyclase CD38

Abstract

Renal blood flow (RBF) responses to arginine vasopressin (AVP) were tested in anesthetized wild-type (WT) and CD38^−/− mice that lack the major calcium-mobilizing second messenger cyclic ADP ribose. AVP (3–25 ng) injected intravenously produced dose-dependent decreases in RBF, reaching a maximum of 25 ± 2% below basal RBF in WT and 27 ± 2% in CD38^−/− mice with 25 ng of AVP. Renal vascular resistance (RVR) increased 75 ± 6% and 78 ± 6% in WT and CD38^−/− mice. Inhibition of nitric oxide (NO) synthase with nitro-l-arginine methyl ester (l-NAME) increased the maximum RVR response to AVP to 308 ± 76% in WT and 388 ± 81% in CD38^−/− (P < 0.001 for both). Cyclooxygenase inhibition with indomethacin increased the maximum RVR response to 125 ± 15% in WT and 120 ± 14% in CD38^−/− mice (P < 0.001, <0.05). Superoxide suppression with tempol inhibited the maximum RVR response to AVP by 38% in both strains (P < 0.005) but was ineffective when administered after l-NAME. The rate of RBF recovery (relaxation) after AVP was slowed by l-NAME and indomethacin (P < 0.001, <0.005) but was unchanged by tempol. All vascular responses to AVP were abolished by an AVP V_1a receptor antagonist. A V₂ receptor agonist or antagonist had no effect on AVP-induced renal vasoconstriction. Taken together, the results indicate that renal vasoconstriction by AVP in the mouse is strongly buffered by vasodilatory actions of NO and prostanoids. The vasoconstriction depends on V_1a receptor activation without involvement of CD38 or concomitant vasodilatation by V₂ receptors. The role of superoxide is to enhance the contractile response to AVP, most likely by reducing the availability of NO rather than directly stimulating intracellular contraction signaling pathways.

arginine vasopressin (AVP) exerts diverse actions on the kidney and the cardiovascular system. In the kidney, AVP promotes water reabsorption by stimulating V₂ receptors that regulate aquaporin 2 in the distal tubule and collecting duct (50) and increase sodium reabsorption in distal segments beginning with the ascending limb of Henle's loop (72, 78). Cardiovascular actions of AVP are primarily mediated by V_1a receptors that induce vasoconstriction through activation of G_q11/12-coupled receptors (51). In some species, such as dog and rabbit, AVP also stimulates V₂ receptors in the vasculature to produce vasodilation (2, 55, 67, 82). V₂ receptors may inhibit V_1a-mediated constriction of rat descending vasa recta (86). The receptors mediating the effects of AVP on the renal vasculature of the mouse are not known.

Mice with V_1a receptors genetically deleted are characterized by attenuated vasoconstrictor responses to AVP, hypotension, reduced blood volume, decreased sympathetic nerve activity, and reduced activity of the renin-angiotensin-aldosterone system (4, 51). These mice also show decreased susceptibility to salt-induced hypertension (69). Intravenous or intrarenal administration of a V₁ receptor agonist produces hypertension in previously normotensive rats (21, 22, 79, 80). In models of genetic hypertension, such as the spontaneously hypertensive rat (SHR) and the SHR-stroke-prone rat (SHR-SP), AVP causes exaggerated renal vasoconstriction compared with normotensive controls (11, 30, 32) and contributes to the elevated arterial pressure (17, 64, 77). AVP has been implicated in the pathophysiology of hypertension in humans and appears to contribute more to essential hypertension in African-Americans than Caucasian subjects (7).

In common with other vasoconstrictor G protein-coupled receptors (GPCRs), the response to AVP V_1a receptor activation in vascular smooth muscle cells (VSMC) usually involves enhanced Ca²⁺ entry across the plasma membrane in combination with Ca²⁺ mobilization from internal sarcoplasmic reticulum stores (25, 26, 32, 42, 81). In addition to these mechanisms, many GPCRs that initiate renal vasoconstriction, e.g., AT₁ receptors for angiotensin II and ET_A endothelin-1 receptors, also stimulate the ADP ribosyl cyclase CD38 to activate Ca²⁺ mobilization via ryanodine receptors (RyR) in VSMC (28, 29). At present, it is not known whether renal vascular responsiveness to AVP also involves the CD38/RyR/Ca²⁺ signaling pathway.

Vasodilators including nitric oxide (NO) and cyclooxygenase (COX)-derived prostanoids usually blunt the magnitude of AVP-induced vasoconstriction (8, 36, 41, 57, 70), and the arterial baroreflex strongly counteracts AVP-mediated increases in arterial pressure (AP) (37). The renal circulation appears to be particularly well buffered against the vasoconstrictor actions of AVP (36), with participation of either V_1a or both V_1a and V₂ receptors, depending on species (2, 55, 67, 82, 86).

To our knowledge, no study has evaluated mechanisms mediating AVP-induced renal vasoconstriction in the mouse. The purpose of our study was to test whether the renal vascular actions of AVP are mediated by V₁ or a combination of V₁ and V₂ receptors. A second aim was to determine the extent to which the ADP ribosyl cyclase CD38 contributes to AVP-induced renal vasoconstriction by testing responses in wild-type (WT) and CD38-null mice. We also tested the effects of NO synthase (NOS) and COX inhibition to determine the extent to which NO or vasodilator prostanoids buffer the magnitude of acute renal vasoconstriction produced by AVP. Other experiments assessed the effects of superoxide (O₂^·−) dismutation with tempol to determine the contribution of reactive oxygen species to the renal vascular actions of AVP.

MATERIALS AND METHODS

CD38 knockout mice and C57BL6J WT mice were bred in our animal facility from original stock provided by Dr. F. Lund (84). Animals were housed and studied in accordance with the NIH guidelines for care and use of laboratory animals, and all experimental protocols were approved by the Animal Care and Use Committee at the University of North Carolina at Chapel Hill. Tap water and chow containing 0.4% sodium were provided ad libitum. Animals were not fasted before study.

Animal preparation.

Animals were anesthetized with pentobarbital sodium (80 mg/kg ip) and placed on a thermostatically controlled operating table to maintained body temperature at 37°C (65). A tracheotomy tube (PE90) was installed to maintain a clear airway and the urinary bladder cannulated at the dome with PE50 tubing for urine drainage. Maintenance infusions of 0.9% NaCl at 0.3 μl/min per g body wt and a 2% solution of BSA in 0.9% NaCl at 3 μl/min were provided through indwelling cannulas in the right jugular vein. Anesthesia was sustained with intermittent supplemental intravenous doses of pentobarbital. The left carotid artery was cannulated with a narrowed PE50 cannula for measurement of AP with a pressure transducer (Statham Gould P23Db) connected to a carrier amplifier (HP 8805B). The right kidney was exposed in the retroperitoneal space through a flank incision and the renal artery separated from the renal vein and surrounding adventitia. As part of the renal artery dissection, the renal nerves were severed to isolate the renal circulation from sympathetic nerve activity. A transit time blood flow transducer (Transonic MA0.5PSB) was positioned around the renal artery and connected to a flow meter (Transonic TS420). AP and RBF signals were digitized at 120 Hz with an A/D converter (DT9816) for continuous display on a PC-based data acquisition system (DtEz software). Data were averaged into 1-s bins and stored on a disk for offline analysis.

Precise 10-μl boluses of AVP at doses of 3, 5, 10, and 25 ng (equivalent to 3, 5, 10, and 25 pmol, or ∼120, 200, 400, and 1,000 pmol/kg) were injected into the jugular vein with a high-performance liquid chromatography injection valve positioned in line with the 2% BSA infusion. For each injection, the infusion rate was increased from 3 μl/min to 60 μl/min simultaneously with activation of the injection valve and then returned to 3 μl/min 30 s later. Each AVP injection was complete 20 s after activation of the valve. The time between individual injections was based on the return of RBF and AP to preinjection levels but was at least 10 min for all injections. At the end of each experiment, an animal was euthanized with an overdose of anesthetic, and the right kidney was removed, decapsulated, and weighed.

Experimental protocols.

After completion of surgery, 30–45 min were allowed for stabilization before AVP injections commenced. All doses were tested in duplicate in a randomized sequence during a control period followed by an experimental period that evaluated the effects of NO depletion with l-NAME (25 mg/kg iv), COX inhibition with indomethacin (170 μg/kg per min), or O₂^·− suppression with tempol (4 mg/kg per min). These agents were infused for 20–30 min before AVP injections were recommenced. In a subgroup of nine mice receiving l-NAME, a third period of AVP injections was included to test the effects of tempol after inhibition of NOS with l-NAME. Acute inhibition of normally coupled NOS with l-NAME effectively reduces renal production of NO and urinary nitrite/nitrate excretion and causes renal vasoconstriction in healthy animals (60, 61, 92). In vitro studies have shown that NOS inhibition reduces endothelium-dependent NO production by isolated renal vessels (71, 99, 101).

The contribution of V_1a receptor activation to renal vascular responses to AVP was tested in a separate group of WT mice in which the effects of AVP were determined before and during an intravenous infusion of the selective V_1a receptor antagonist Manning compound [β-Mercapto-β,β-cyclopentamethylenepropionyl1, O-me-Tyr2, Arg8]-Vasopressin (MC, Sigma) (62). In accordance with published reports, MC was infused at 100 ng/min iv to achieve effective blockade of peripheral V_1a receptors in mice (1). In some animals, antagonism of V₂ receptors [Adamantaneacetyl1, O-Et-D-Tyr2, Val4, Aminobutyryl6, Arg8,9]-Vasopressin (62) was tested in combination with V_1a receptor blockade. In other animals, V₂ receptor activation was tested in a group of five WT mice in which RBF and AP were monitored during intravenous bolus injections (50 and 100 ng) of the selective V₂ receptor agonist dDAVP (desmopressin) (91).

Analysis of results.

A 300-s sampling interval was allocated to each AVP injection, beginning with a 60-s baseline period immediately before the onset of each vasoconstrictor response. Values for mean arterial pressure (MAP), RBF, and RVR were converted to a percentage change from the mean values obtained for each parameter during the 60-s baseline period. The magnitude of vasoconstrictor responses to AVP were evaluated as the maximum percent change from baseline and from the integrated area of each response curve over a 120-s period beginning at the onset of the response. Dynamic characteristics of responses to AVP were estimated from the downslope (vasoconstrictor phase) and upslope (relaxation phase) of the RBF response. These rates were obtained from the first derivative of the RBF data trace over linear segments of the contractile and relaxation phases of the response. A linear portion in the downslopes consistently occurred between 5 and 8 s after the beginning of each response, and a straight segment in the upslope occurred between 10 and 20 s after the maximum reduction in RBF. The response slopes were expressed as %RBFchange/s compared with the baseline RBF and plotted against the corresponding mean maximum percent reduction in RBF for each dose.

The Prism statistical and graphing software program was used for data analyses and presentation (v.6.03 GraphPad). Overall responses are reported as means ± SE for each dose. Comparisons within groups (control vs. experimental periods) and between groups (WT vs. CD38^−/−) were performed with paired and unpaired t-tests or two-way analysis of variance. Dose-response curves for maximum percent changes in AP, RBF, RVR, and rates of contraction or relaxation were analyzed with nonlinear regression using best-fit curves, which were either straight lines or one-phase exponentials, depending on the data.

RESULTS

Baseline values during control conditions for 30 WT and 30 CD38-deficient mice are presented in Table 1. MAP, RBF, and RVR were similar in the two groups, as were body weight and left kidney weight. Hematocrit was stable throughout each experiment.

Table 1.

Basal values for WT and CD38^−/− mice

	WT	CD38−/−
Arterial blood pressure, mmHg	109 ± 1	107 ± 2
Heart rate, beats/min	465 ± 11	441 ± 14
Renal blood flow, ml·min−1·g kidney wt−1	8.42 ± 0.47	9.23 ± 0.57
Renal vascular resistance, mmHg·ml−1·min−1·g kidney wt	14.7 ± 1.3	12.7 ± 0.9
Body weight, g	26.4 ± 0.4	25.1 ± 0.6
Right kidney weight, g	0.192 ± 0.003	0.189 ± 0.004
HCT-1, %	43.5 ± 1.0	41.6 ± 1.0
HCT-2, %	44.0 ± 1.3	43.9 ± 1.3
Number of animals	30	30

Values are means ± SE. There were no significant differences between strains. HCT-1, start hematocrit; HCT-2, end hematocrit; WT, wild-type.

Vascular responses to bolus injections of AVP in WT mice are shown in Fig. 1, top. After a 60-s baseline recording, AVP was injected intravenously, and the hemodynamic responses were recorded for the following 240 s. A dose-response relationship was apparent for MAP and RVR at all doses of AVP and for RBF at the two higher doses (10 and 25 ng). All AVP doses resulted in a small increase in RBF after 120 s, which was accompanied by a persistent increase in RVR. This combination made it most likely that the increase in RBF is the result of a relatively prolonged systemic vasopressor action of AVP compared with a more rapid decline in renal vasoconstriction. Figure 1, bottom, shows the maximum responses produced by AVP during control conditions (○). RBF was reduced in a dose-dependent manner to a maximum of 28% below baseline. RVR progressively increased to 75% above control, whereas MAP increased 35%.

Fig. 1.

Arginine vasopressin (AVP)-induced changes in renal blood flow (RBF), renal vascular resistance (RVR), and mean arterial pressure (MAP) in the control period in wild-type (WT) mice (n = 20). Top: mean responses to 4 doses of AVP (3–25 ng) administered at 60 s. For clarity, symbols are shown at 10-s intervals; SE is shown by dashed lines. Bottom: maximum changes recorded during the control period (○) for each dose of AVP. ■, maximum responses to the same doses of AVP following V_1a receptor antagonism with Manning compound [β-Mercapto-β,β-cyclopentamethylenepropionyl¹, O-me-Tyr², Arg⁸]-Vasopressin (n = 6). ▲, responses to V₂ receptor activation with dDAVP (desmopressin, 25 and 50 ng, n = 5).

Renal and systemic vascular responses to bolus injections of the V₂ receptor agonist are also shown in Fig. 1 (bottom, ▲). The small changes in RBF and RVR (<5% of baseline) were not different from vehicle injections. These data show no sign of a vasodilatory response to dDAVP at the renal or systemic level.

The V_1a receptor antagonist MC was administered to WT mice (n = 8) to evaluate the contribution of this receptor to the action of endogenous AVP and to the acute hemodynamic responses to administered AVP. Baseline MAP was not affected by MC, but heart rate was increased (Table 2). RBF rose by 15% (P < 0.05), and RVR fell by 7% (P < 0.05), which is consistent with a small to modest effect of endogenous AVP on the renal vasculature during anesthesia. Importantly, the renal vasoconstriction normally produced by administered AVP was abolished by V_1a receptor blockade with MC (Fig. 1, bottom, ■). MC also almost completely eliminated the pressor response to AVP. These results suggest that the renal vascular effects of AVP are mediated exclusively by V_1a receptors in the mouse. This conclusion is reinforced by the finding that antagonism of V₂ receptors with [Adamantaneacetyl1, O-Et-D-Tyr2, Val4, Aminobutyryl6, Arg8,9]-Vasopressin (dose 2–4 nmol/kg) (62) during continued blockade of V_1a receptors had no effect on RBF or RVR (data not shown).

Table 2.

Basal values before (control) and during AVP V_1a receptor blockade with MC in WT mice

	Control	V1a Receptor Blockade
Arterial blood pressure, mmHg	113 ± 4	115 ± 3
Heart rate, beats/min	474 ± 37	560 ± 53†
Renal blood flow, ml·min−1·g kidney wt−1	7.93 ± 1.20	9.13 ± 1.5*
Renal vascular resistance, mmHg·ml−1·min−1·g kidney wt	19.5 ± 4.8	18.2 ± 4.7*
Number of animals	8	8

Values are means ± SE. Control and Manning compound (MC) data were compared by paired t-test.

^*P < 0.05,

^†P < 0.01. AVP, arginine vasopressin.

Similar RBF studies were conducted on CD38^−/− mice to determine the importance, if any, of this type of ADP ribosyl cyclase in the acute hemodynamic responses to AVP in the mouse. As is shown in Fig. 2, the maximum responses of RBF, RVR, and MAP to AVP in CD38-null mice (■) did not differ from those in WT mice (○). Thus we conclude that CD38 and its downstream Ca²⁺ signaling pathway are not involved in the renal vascular actions of AVP mediated by vascular V_1a receptor stimulation in the mouse.

Fig. 2.

AVP-induced changes in RBF, RVR, and MAP in WT (○, n = 20) and CD38^−/− (■, n = 20) mice. Top: maximum responses produced by AVP (3–25 ng). Bottom: area under curves for the same parameters in WT and CD38^−/− mice following AVP injections. There are no differences between WT and CD38^−/− mice.

Analysis of the maximum hemodynamic responses to AVP is somewhat limited in that it does not provide insight into the overall magnitude of the response that includes duration as well as the maximum response. To assess this, we calculated the integrated area of the response curves to provide a more comprehensive index of the overall response. These data are summarized in Fig. 2, bottom. The integrated responses of RBF, RVR, and MAP produced by AVP were similar for WT and CD38^−/− mice. These results reinforce the conclusion that AVP responses in CD38^−/− and WT mice are not different. Because the two approaches to data analysis yielded similar conclusions for group comparisons, subsequent analyses of the magnitude of measured responses are based solely on maximum changes produced by AVP.

Other experiments tested whether NO buffers part of the vasoconstrictor actions of AVP in the mouse. To this end, l-NAME was administered to inhibit NOS in the experimental period. Basal data for these animals are presented in Table 3. l-NAME reduced RBF and increased RVR and MAP in both WT and CD38^−/− mice. The maximum vasoconstrictor responses to NOS inhibition were similar in both groups of mice, with changes averaging 50% for RBF, 230% for RVR, and 12% for MAP. Thus resting, tonic formation of NO produces appreciable renal vasodilation that is independent of CD38 under resting conditions. The renal vasoconstrictor actions of AVP were magnified during inhibition of NO production by l-NAME. As is shown in Fig. 3, the reductions in RBF produced by the three lowest doses of AVP were increased more than fourfold during l-NAME treatment, and the highest dose response was increased two- to threefold. The increases in RVR due to AVP were amplified two- to fourfold by NOS inhibition. The systemic pressor response to AVP was less pronounced in the absence of NO. The effect of l-NAME on AVP responses was similar in WT (Fig. 3, top) and CD38^−/− mice (Fig. 3, bottom) with no significant differences between the two groups. These results indicate that NO blunts a fraction of AVP-induced renal vasoconstriction and that the contribution of NO is independent of the ADP ribosyl cyclase CD38.

Table 3.

Basal values for l-NAME effects on AVP responses in WT and CD38^−/− mice

	WT		CD38−/−
	Control	l -NAME	Control	l -NAME
Arterial blood pressure, mmHg	107 ± 1	119 ± 3*	103 ± 2	117 ± 2†
Heart rate, beats/min	474 ± 10	391 ± 42	428 ± 28	379 ± 46
Renal blood flow, ml·min−1·g kidney wt−1	6.61 ± 0.44	3.32 ± 0.37†	7.09 ± 0.46	3.54 ± 0.26†
Renal vascular resistance, mmHg·ml−1·min−1·g kidney wt	17.1 ± 1.2	40.6 ± 4.2†	15.6 ± 1.0	35.7 ± 2.9†
Hematocrit, %	43.4 ± 1.6	43.2 ± 2.67	40.2 ± 1.4	43.0 ± 1.9
Body weight, g	26.9 ± 0.7		25.9 ± 1
Right kidney weight, g	0.190 ± 0.004		0.177 ± 0.006
Number of animals	13		13

Values are means ± SE.

^*P < 0.005,

^†P < 0.001 for differences between control and N-nitro-l-arginine methyl ester (l-NAME) periods in each strain.

Fig. 3.

Effects of nitric oxide (NO) synthase inhibition with nitro-l-arginine methyl ester (l-NAME) on vascular responses to AVP in WT mice (top: n = 13) and CD38^−/− mice (bottom: n = 13). ○, Control responses to AVP (3–25 ng). ■, Responses to AVP after NO synthase inhibition with l-NAME. #P < 0.05 for WT vs. CD38^−/− *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001 for difference between control and l-NAME periods.

We next tested for possible buffering of the renal vascular actions of AVP by COX-derived vasodilator prostanoids. Indomethacin was administered in the experimental period to inhibit COX. Indomethacin had a negligible effect on basal renal hemodynamics and MAP in either group of mice (Table 4). COX inhibition magnified the AVP-induced reductions in RBF by 1.5–2.0-fold and the increases in RVR to a lesser extent (Fig. 4), again with no significant differences between responses in WT and CD38^−/− mice. The pressor response to AVP was reduced during COX inhibition by ∼30% on the average in both groups of animals. Thus COX-generated vasodilator prostanoids also buffered some of the AVP-induced renal vasoconstriction in a manner independent of CD38 activity.

Table 4.

Basal values for indomethacin effects on AVP responses in WT and CD38^−/− mice

	WT		CD38−/−
	Control	Indomethacin	Control	Indomethacin
Arterial blood pressure, mmHg	111 ± 3	111 ± 4	110 ± 4	110 ± 4
Heart rate, beats/min	466 ± 18	400 ± 9*	459 ± 11	392 ± 8†
Renal blood flow, ml·min−1·g kidney wt−1	11.14 ± 0.63	14.10 ± 1.10	11.46 ± 0.48	13.42 ± 0.93
Renal vascular resistance, mmHg·ml−1·min−1·g kidney wt	10.4 ± 0.6	8.3 ± 0.8	10.7 ± 0.6	8.5 ± 0.6
Hematocrit, %	45.1 ± 1.1	48.6 ± 1.6	40.7 ± 0.7	42.1 ± 2.3‡
Body weight, g	25.5 ± 0.6		25.0 ± 0.7
Right kidney weight, g	0.200 ± 0.005		0.179 ± 0.006
Number of animals	8		7

Values are means ± SE.

^*P < 0.01,

^†P < 0.005 for differences between control and indomethacin period.

^‡P < 0.005 for difference between WT and CD38^−/− strain.

Fig. 4.

Effects of cyclooxygenase inhibition with indomethacin on vascular responses to AVP in WT mice (top: n = 8) and CD38^−/− mice (bottom: n = 7). ○, control responses to AVP (3–25 ng). ■, Responses to the same doses after indomethacin. There were no differences between CD38^−/− and WT mice at any dose during control or indomethacin periods. *P < 0.05, ***P < 0.005, ****P < 0.001 for difference between control and indomethacin treatment in same strain.

In other experiments, we determined the participation of O₂^·− in AVP-induced renal vasoconstriction by administering tempol in the experimental period to scavenge O₂^·−. Tempol had a negligible effect on basal renal hemodynamics and MAP in either group of mice (Table 5). As Fig. 5 shows, the antioxidant attenuated the magnitude of the reduction in RBF caused by AVP by 25–50%, whereas the increase in RVR was reduced by 30% in WT and CD38-null mice. Tempol had little effect on the MAP response to AVP in either strain. The similar degree of attenuation by tempol on AVP-mediated vascular responses in WT and CD38^−/− mice supports the conclusion that the potentiating action of O₂^·− is also independent from CD38 and its downstream Ca²⁺-signaling pathway.

Table 5.

Basal values for tempol effects on AVP responses in WT and CD38^−/− mice

	WT		CD38−/−
	Control	Tempol	Control	Tempol
Arterial blood pressure, mmHg	110 ± 3	110 ± 3	110 ± 2	105 ± 3
Heart rate, beats/min	454 ± 28	481 ± 23	447 ± 23	448 ± 16
Renal blood flow, ml·min−1·g kidney wt−1	8.25 ± 0.71	10.64 ± 1.11	10.76 ± 0.78	13.37 ± 1.19
Renal vascular resistance, mmHg·ml−1·min−1·g kidney wt	15.4 ± 2.2	12.0 ± 2.0	10.7 ± 0.6	8.5 ± 0.9
Hematocrit, %	42.4 ± 0.9	41.6 ± 2.1	39.8 ± 1.3	45.3 ± 1.2
Body weight, g	26.5 ± 0.6		23.9 ± 0.6
Right kidney weight, g	0.187 ± 0.005		0.176 ± 0.007
Number of animals	11		10

Values are means ± SE. There are no significant differences within or between strains.

Fig. 5.

Effects of superoxide quenching with tempol on vascular responses to AVP in WT mice (top: n = 11) and CD38^−/− mice (bottom: n = 10). ○, Control responses to AVP (3–25 ng). ■, Responses to the same doses after superoxide inhibition with tempol. #P < 0.05 for WT vs. CD38^−/− in control period $P < 0.05 for WT vs CD38^−/− after tempol, **P < 0.01, ***P < 0.005, ****P < 0.001 for difference between control and tempol in same strain.

To further investigate the influence of O₂^·− on vascular responses to AVP, we tested the effects of tempol after NOS inhibition with l-NAME in WT mice. These experiments showed no inhibitory action of tempol after l-NAME treatment; the 25 ng dose of AVP increased the RVR response to 283 ± 81% above control after l-NAME and to 241 ± 96% after tempol combined with l-NAME (NS, n = 9). This demonstrates that O₂^·− does not contribute to the vasoconstrictor response to AVP in the absence of NO. Our interpretation is that the effect of O₂^·− on AVP-mediated renal vasoconstriction is primarily due to the neutralization of NO rather than a direct action of O₂^·− on the excitation-contraction process per se.

Bolus injections of AVP provide transient, dynamic hemodynamic changes with consistent onset and recovery slopes. Response slopes for RBF were directly correlated with the magnitude of the vasoconstriction, and this relationship was used to characterize the dynamics of AVP responses. The rates of decline in RBF following AVP injections (vasoconstrictor response) were not different between the control and experimental periods in any protocol (Fig. 6 and Table 6). Thus, although the magnitude of the vasoconstrictor response to AVP is buffered by NO and prostanoids and augmented by O₂^·−, these agents do not influence the rate of contraction induced by AVP within the renal microcirculation. Similarly, there were no differences in the rates of vasoconstriction between WT and CD38^−/− mice in the control period or during any experimental period (i.e., NO, prostanoid, or O₂^·− inhibition, Table 6). Pooled values of the vasoconstriction phase provide an overall average rate of vasoconstriction of −0.065 ± 0.004% of basal RBF/s. Unlike the rate of vasoconstriction, the relaxation phase was more complex and strongly influenced by the buffering action of NO and prostanoids. These data are shown in Fig. 6, right, and in Table 6. Both l-NAME and indomethacin slowed the rate of recovery or relaxation from AVP-induced vasoconstriction compared with their respective control responses. This difference was statistically significant for both groups. The rate of relaxation was unaffected by tempol.

Fig. 6.

Slopes of RBF changes in WT mice as a function of the maximum reduction in RBF produced by AVP at each dose. Left: rate of constrictions (downslope). Right: rate of relaxations (upslope). Each horizontal pair shows the effects of l-NAME, indomethacin, or tempol. Rate of change in RBF is calculated as percent change from the basal level per second. ***P < 0.005, ****P < 0.001 for difference between control recovery rate and treatment recovery rate.

Table 6.

Rates of contraction (downslope) and relaxation (upslope) as a function of maximum reduction in RBF following AVP injections in WT and CD38^−/− mice

	WT				CD38−/−
	Contraction		Relaxation		Contraction		Relaxation
	Control	Exp	Control	Exp	Control	Exp	Control	Exp
l-NAME	−75.1 ± 8.4	−72.7 ± 10.1	46.4 ± 4.8	16.6 ± 3.3†	−81.3 ± 6.7	−57.9 ± 9.4	33.6 ± 3.9	9.1 ± 2.8†
Indomethacin	−41.5 ± 7.8	−56.5 ± 4.4	44.9 ± 5.9	18.3 ± 2.0*	−54.2 ± 5.8	−60.5 ± 5.2	40.2 ± 4.0	6.3 ± 2.1*
Tempol	−53.4 ± 7.8	−66.0 ± 10.7	43.7 ± 4.2	32.7 ± 7.8	−67.9 ± 10.4	−81.5 ± 11.1	41.7 ± 4.7	33.6 ± 5.5

Values are mean % change in basal renal blood flow (RBF)/s maximum reduction in RBF⁻¹ × 1,000 ± SE. Actual slopes are 1/1,000 of values shown.

^*P < 0.005,

^†P < 0.001 for difference between control and experimental periods (Exp) in each strain.

The RBF recovery rate after indomethacin treatment was notable as the only slope that showed a difference between CD38^−/− and WT mice. The difference was small and of unknown physiological significance. In all other cases, CD38^−/− recovery rates were indistinguishable from the corresponding WT values.

DISCUSSION

Our results provide the first assessment of vascular reactivity to AVP in the mouse kidney. We observed that the vascular responses are restricted to vasoconstriction due to AVP V_1a receptor activation. We found no evidence for a renal vasodilatory action of AVP at the doses tested, and the AVP V₂ receptor agonist dDAVP did not influence RBF or AP. V₂ receptor blockade was also ineffective. The magnitude of the V_1a receptor-mediated vasoconstriction is modulated significantly by vasodilator systems involving release of NO and COX-generated prostanoids and by the vasoconstrictor actions of reactive oxygen species, probably O₂^·− anion, as revealed by a quenching action of the superoxide dismutase (SOD) mimetic tempol. The inhibitory action of tempol was eliminated by NOS inactivation with l-NAME, suggesting that the primary role of O₂^·− on AVP-induced renal vasoconstriction is an indirect one involving reduced availability of NO. A major new finding was that, unlike other vasoconstrictor GPCR agonists tested to date, AVP-induced renal vasoconstriction does not involve the ADP ribosyl cyclase enzyme CD38 (27–29). Moreover, inhibition of O₂^·− generation with tempol produces the same inhibitory effect on renal vasoconstriction in response to AVP in WT and CD38^−/− mice, indicating an action of O₂^·− that is also independent of CD38. This contrasts with previous reports showing that O₂^·− induces Ca²⁺ mobilization and vasoconstriction, in part, by stimulation of cADP ribose production (28, 29, 96, 100). It seems that, unlike other vasoconstrictor GPCRs, the V_1a AVP receptor does not engage the CD38/O₂^·− pathway as part of the contractile response.

Previous studies have provided evidence for predominant V_1a receptor mediation of AVP actions on the renal vasculature of the rat (8, 20, 30). V₂ receptors in the rat kidney localized to the distal tubule and collecting duct regulate water permeability, but no V₂ receptor mRNA or protein expression is found in the rat renal vasculature (20). Consistent with this, pharmacological studies have failed to reveal V₂ receptor-mediated renal vasodilation in the rat (8, 30, 57). In contrast, however, AVP activates V₂ receptors to produce NO-dependent renal vasodilation in the dog (2, 55, 67) and rabbit (49).

Cyclic ADP ribose in afferent arterioles has emerged as an important component in Ca²⁺ signaling activated by angiotensin-II (Ang II), endothelin-1 (ET-1), and norepinephrine (NE) (27–29). The direct consequence of this in vivo is evident as a diminished renal vasoconstrictor response to Ang II, ET-1, NE, phenylephrine, and thromboxane in mice lacking the major mammalian ADP ribosyl cyclase CD38 (65, 83–85). These publications raised the possibility that cyclic ADP ribose is utilized as a Ca²⁺-mobilizing agent by all vasoconstrictor agents that activate G_q/11 protein-coupled receptors in the renal microcirculation. However, we now show that AVP is an exception to this rule. This may be the result of differences in Ca²⁺ signaling or Ca²⁺ sensitivity between the excitation/contraction pathways used by AVP and other GPCR vasoconstrictors, an issue worthy of further investigation.

In common with other G_q/11-coupled receptors that activate phospholipase C, the vasoconstrictor action of AVP V_1a receptors is mediated in part by Ca²⁺ entry through l-type Ca²⁺ channels (9, 25, 26, 32, 42) and mobilization of Ca²⁺ from internal VSMC stores in response to inositol 1,4,5-trisphosphate (IP₃) (32, 81). In vitro studies on afferent arterioles have shown additional Ca²⁺ entry through store-operated Ca²⁺ channels triggered by Ca²⁺ mobilization (26). Several studies into the vasoconstrictor action of AVP on nonrenal arterial VSMC maintained in culture conditions have implicated inhibition of K⁺ channels (90, 93), resulting in a depolarizing stimulus that promotes Ca²⁺ entry through voltage-sensitive L-type Ca²⁺ channels. Others have found that physiological levels of AVP may preferentially inhibit K⁺ channels and activate transient receptor potential channel 6 (TRPC6) cation channels in mesenteric VSMC independently from the IP₃ signaling pathway (38, 59). Such a mechanism independent of Ca²⁺ mobilization might distinguish the signaling pathways for AVP from other G_q/11-coupled vasoconstrictors.

A strong buffering influence of NO against AVP-mediated vasoconstriction in the mouse kidney is readily apparent from the marked potentiating effect of l-NAME inhibition of NOS. Early studies on isolated-perfused kidneys suggested that NO buffering of AVP-mediated renal vasoconstriction is the result of endothelial NOS activation by increased shear stress (8). However, in vivo vasoconstrictor responses to Ang II in the rat (23, 45) and the rabbit (74) and thromboxane agonists in the mouse (N. Moss and W, Arendshorst, unpublished observations) are weakly or not at all potentiated by l-NAME, suggesting that V_1a receptor-mediated AVP-induced vasoconstriction is more actively buffered by NO compared with other vasoconstrictor agents. This is supported by a direct renal vasodilatory response to subpressor doses of AVP in rats that is inhibited by l-NAME (75). Other reports have shown that selective activation of V₁ receptors in the rat promote production of NO and vasodilator prostaglandins (PGs) in the renal vasculature that offset the vasoconstrictor action of AVP (19, 57). This is not limited to V₁ receptors, as NO-dependent vasodilatation is stimulated by V₂ receptor activation in the dog (2, 55, 67) and rabbit (82). In the present study, we observed no vasodilatory responses to AVP. The lowest tested dose of AVP reduced RBF by 3%, increased RVR by 15%, and increased MAP by 10–20%. Our data in the mouse support direct stimulation of NO production by AVP to buffer the vasoconstrictor response involving V_1a, but not V₂, receptor activation.

In addition to l-NAME, COX inhibition magnified AVP-induced renal vasoconstriction, implicating primary inhibition of production of a vasodilatory PG, such as PGE₂ and/or PGI₂. Several reports have implicated PGs as a buffering agent against the systemic and renal vascular response to AVP in the rat (19, 30, 39, 94), dog (70), rabbit (41, 76), and humans (66). We cannot exclude the possibility that synthesis of the vasoconstrictor thromboxane (TxA) was also attenuated, but previous studies have shown that basal production of TxA₂ is quite low in healthy kidneys, and it normally exerts little basal renal vasoconstriction in vivo (10, 12, 35, 44, 89). Moreover, intrarenal infusion of AVP causes COX-dependent release of PGE₂ and PGI₂ (19, 58, 63) but not TxA₂ (63). Overall, the net effect of COX inhibition in our mice was removal of vasodilator prostanoids that normally buffer renal vasoconstriction produced by AVP.

Superoxide anions have not been shown previously to participate in AVP-mediated vasoconstriction in the kidney. The potentiating role of O₂^·− in renal vasoconstrictor responses to angiotensin AT₁ and thromboxane prostanoid receptor activation are well documented (47, 65, 73), but little is known for AVP. Previous studies on the interactions between AVP and O₂^·− in nonrenal arteries and VSMC have shown upregulation of the V_1a receptor by O₂^·− (52) and an action of AVP to stimulate O₂^·− production (54). AVP stimulation of O₂^·− production by NADPH oxidase is reported to increase TRPC6 channel activity and Ca²⁺ entry in the A7r5 line of cultured VSMC and mouse aortic VSMC (24). In our study, a potentiating action of O₂^·− on the renal vasoconstrictor response to AVP is clearly supported by an inhibitory action of the SOD mimetic tempol on the vascular response, but the mechanisms responsible for this action of O₂^·− are uncertain. Possibilities include the suppression of NO-mediated buffering due to the mutually destructive reaction between O₂^·− and NO and a direct vasoconstrictor action of O₂^·− through activation of Ca²⁺ signaling pathways. The lack of effect of tempol treatment after NOS inhibition provides insight into the relative importance of these two possibilities. The effect of O₂^·− to reduce NO availability would be nullified by prior inhibition of NO production and its reduced bioavailability, whereas a direct action of O₂^·− on the contractile response of VSMC would be expected to persist after l-NAME. Because the inhibitory action of tempol on vascular activity of AVP was not evident after NOS inhibition with l-NAME, we conclude that the principle potentiating effect of O₂^·− is to reduce the availability of NO. This also implies that O₂^·− generated by AVP does not participate in cellular Ca²⁺ signaling, such as the ADP ribosyl cyclase/RyR/Ca²⁺ pathway (28, 29). The similarity between renal vasoconstrictor responses to AVP in WT and CD38^−/− mice is quite consistent with this conclusion and represents a clear distinction between AVP V_1a receptors and other vasoconstrictor GPCRs. This could result in different susceptibilities to the effects of oxidative stress and inflammatory reactions during the development of disease states.

Analysis of dynamic components in the vasoconstrictor and relaxation responses to AVP is a novel approach to characterize acute, transient RBF responses. Our data indicate that, after AVP injection, RBF declined at a rate that was directly proportional to the maximal reduction in RBF. Interestingly, this rate of vasoconstriction was not different between WT and CD38^−/− animals and was unaffected by treatment with l-NAME, indomethacin, or tempol. Future studies are required to determine whether the constancy of this relationship is a function of the intrinsic contractility of the renal vasculature or whether it varies with different vasoconstrictor agents and their respective GPCR. In contrast, the relaxation phase following AVP-mediated constriction was clearly prolonged by l-NAME and indomethacin. This is readily explained by the suppression of vasodilatory stimuli from NO and prostanoids that would normally accelerate the rate of relaxation of VSMC. Previous studies in the rat have shown that PGE₂ and PGI₂ and their analogs accelerate the rate of relaxation following Ang II-, norepinephrine-, and TxA₂-induced renal vasoconstriction (18). It is notable that the relaxation rate was slowed to the same extent by l-NAME and indomethacin, even though the RVR response to the highest dose of AVP was increased more by l-NAME (308%) than indomethacin (125%). This is consistent with the notion that NO and prostanoids engage a similar vasodilatory mechanism to buffer the vasoconstrictor effects of AVP. Unlike vasodilator inhibition, suppression of O₂^·− with tempol did not alter the recovery rate from AVP-induced vasoconstriction. The extremely short half-life of O₂^·− may limit its biological activity to the contractile phase of the vascular response.

Our new information documents that AVP is a potent renal vasoconstrictor in the mouse, an action that is normally strongly counteracted in vivo by vasodilatory actions of NO and prostanoids. A modest increase in RBF and decrease in RVR following inhibition of V_1a receptor activity indicate a small tonic vasoconstrictor effect of circulating endogenous AVP on the renal circulation. Previous studies have reported that physiological plasma concentrations of AVP (<10 pg/ml) or V₁ receptor antagonism does not appreciably affect MAP, RBF, or RVR in conscious or pentobarbital-anesthetized rats (15, 34, 36, 48, 68) or conscious dogs (16). Perhaps endogenous AVP has weaker renal vasoconstrictor actions in other species in which buffering of vasoconstriction by vasodilator systems involving NO and prostanoids exert a larger effect. On the other hand, V_1a antagonists can produce a systemic hypotensive effect during severe dehydration in most species when circulating AVP levels approach maximum endogenous levels (13, 95). Similarly, vasopressor and renal vasoconstrictor effects of exogenous AVP administered at more physiological levels are typically more readily apparent when normal buffering or offsetting vasodilator mechanisms are suppressed experimentally (94) or in pathological states such as hypertension (17, 64, 77, 98). AVP produces exaggerated Ca²⁺ signaling and vasoconstriction of renal microvessels of SHR with genetic hypertension (30, 32, 43, 88), largely due to upregulation of V_1a receptor mRNA and protein levels (87). These models likely represent situations when plasma AVP is elevated and vasodilatory buffering actions are suppressed, most notably by oxidative stress and excessive O₂^·− levels in VSMC (5).

Under normal circumstances, the renal medullary microcirculation is more responsive than the cortical circulation to changes in plasma AVP concentration, with a sensitivity that parallels that of the collecting duct (20, 33). In normal rats, localized AVP infusions into the renal medulla to activate V_1a receptors results in increased AP when not offset by NO secondary to V₂ receptor activation (21, 79). Nevertheless, from a systemic standpoint, the mesenteric and hindquarter circulations appear to be more sensitive to AVP than the renal microvasculature (14, 36, 40, 97) perhaps due to more effective buffering of the renal vasomotor actions of AVP. This may contribute to better maintenance of RBF and protection of kidney function when AVP is compared with other vasoconstrictor agents used to support blood pressure during hemorrhagic or septic shock (3, 46, 53).

GRANTS

This work was supported by NIH grant RO1 HL-02334.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: N.G.M. and W.J.A. conception and design of research; N.G.M. and T.E.K. performed experiments; N.G.M. analyzed data; N.G.M. and W.J.A. interpreted results of experiments; N.G.M. prepared figures; N.G.M. and W.J.A. drafted manuscript; N.G.M. and W.J.A. edited and revised manuscript; N.G.M., T.E.K., and W.J.A. approved final version of manuscript.