INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENT 1 EXPERIMENT 2 DISCUSSION REFERENCES

VASOPRESSIN (VP) has diverse biological actions and target systems. In addition to the well-known neuroendocrine effects, VP meets the criteria for a brain neurotransmitter. For example, it is synthesized in several brain nuclei, including the paraventricular nucleus of the hypothalamus, medial amygdala, and bed nucleus of the stria terminalis (1, 5). Also, physiological stimuli trigger the release of VP in specific brain regions where it exerts its effects on pre- and postsynaptic membranes (11, 31).

VP exerts its effects through receptors that have been divided into two broad classes, the V₁ and V₂ receptors. The V₁ class can be further divided into the V_1a and V_1b receptors (16, 17, 42). The receptors differ in their distribution and associated second messenger system (17). Neurons that synthesize VP have extensive vasopressinergic projections to sites along the neuraxis, including the septum, hippocampus, amygdala, subfornical organ (SFO), organum vasculosum of the lamina terminalis (OVLT), nucleus of the solitary tract, dorsal motor nucleus of the vagus, and spinal cord (1, 29, 40). Radioligand receptor autoradiography and in situ hybridization results indicate that VP binding sites are distributed in the brain and that the distribution of these sites often overlaps with the terminal distributions of vasopressinergic neurons (22, 32, 46).

Brain release and pituitary release of VP often parallel one another. For example, systemic osmotic and hypovolemic stimuli that increase circulating VP levels also stimulate the release of VP in the lateral septum and lateral ventricle (4). Manipulations that facilitate salt intake can also stimulate the systemic release of VP. The simultaneous administration of furosemide and a low dose of the angiotensin-converting enzyme captopril arouses a sodium appetite and increases plasma VP levels (44). Similarly, adrenalectomized rats and rats treated acutely with DOCA develop a preference for hypertonic NaCl and show increased VP immunoreactivity and mRNA expression in the paraventricular nucleus of the hypothalamus and supraoptic nucleus (15, 37). The correspondence between the peripheral and central release of VP in other situations raises the possibility that during sodium deficiency induced by furosemide and DOCA treatments there is also the central release of VP. A central action of VP may then contribute to the behavioral signature of sodium deficiency, the increased ingestion of concentrated sodium-containing solutions (34). A prediction from this assumption is that the central administration of VP may actually stimulate salt intake. This prediction was not supported in one study where the effect of intraventricular injections of VP on salt intake was measured. Actually, the opposite result was obtained. Sodium-deficient rats were administered intraventricular injections of VP and 20 min later were given access to NaCl. A delay was used because the doses of VP used in the experiment elicited motor disturbance that included rotations along the long axis of the body (barrel rotation) and convulsions that persisted for ~20 min. Under these test conditions, intraventricular injections of VP reduced NaCl intake by sodium-deficient rats (36). This inhibition was quite pronounced, as sodium-deficient rats drank only ~1 ml in 30 min after the VP injection. The authors discount the interpretation that the reduction in intake was secondary to the seizures because the gross motor disturbances had subsided when NaCl was offered to the rats. Although gross motor effects may have subsided, there are long-term postictal changes that endure for 1-2 h after seizurelike behavior has subsided (21, 30). For example, regardless of the type of experimental seizure that is induced, the postictal period contains both increased and decreased motor responsivity (10). As such, features of an enduring postseizure state may have contributed to the observed decrease in NaCl intake.

These observations raise the possibility that the ability of intraventricular injection of VP to reduce intake reflects a pharmacological rather than a physiological effect of exogenous VP. The following experiments readdressed the role of endogenous brain VP in the genesis of salt appetite. To accomplish this goal, sodium-deficient rats were given intraventricular injections of VP and VP receptor antagonists that varied in their selectivity for the V₁ and V₂ receptor subtype, and NaCl intake was measured. Results showed that intraventricular injection of a V₁ receptor antagonist suppressed salt appetite. Depending on the dose, receptor antagonists may have agonistic actions. Therefore, in experiment 2, the receptor action of the VP receptor antagonists was tested by comparing the effects of intraventricular injections of VP with those of the V₁ antagonist on mean arterial blood pressure (MAP).

	EXPERIMENT 1

TOP ABSTRACT INTRODUCTION EXPERIMENT 1 EXPERIMENT 2 DISCUSSION REFERENCES

Methods

Animals and surgery. Male Sprague-Dawley rats were housed in individual suspended steel-mesh cages in a temperature-controlled colony room with a 12:12-h light-dark cycle. All rats were given access to Purina chow and tap water ad libitum.

Screening. To verify the patency of the cannula, nondeprived rats were administered lateral ventricular injections (3 µl) of ANG II (50 ng/rat) and given access to tap water. Intake was measured after 30 min. For rats to be included in the experiment, they had to drink a minimum of 5 ml of water. A total of 48 rats with intraventricular cannulas met this drinking criterion and were used in the following experiments.

Sodium appetite. Rats were adapted to the sodium appetite test procedure before the start of testing. Sodium appetite was aroused by injecting the rats with DOCA (5.0 mg sc) and furosemide (7.5 mg sc). A second furosemide injection (7.5 mg) was administered 2 h later. This procedure reliably induces a salt appetite (3, 13). The rats were then placed in clean cages and given access to a low-sodium diet (ICN, Aurora, OH) and tap water overnight. The next day the low-sodium diet and tap water were removed from the cages, and each rat was administered an intraventricular injection of isotonic saline (3 µl). Rats were then given access to 0.3 M NaCl in 100-ml calibrated glass bottles, and intake was measured (nearest ml) for 60 min. Purina chow and water were then returned to the cages.

Peptides and receptor antagonists. Arginine VP and the VP receptor antagonists were obtained from Bachem (Torrence, CA). The receptor antagonists were selected on the basis of their affinity for the V₁ and/or V₂ receptors (24-27). Peptides were dissolved in sterile 0.9% NaCl. The antagonists were administered in doses of 0 ng (isotonic saline), 10 ng, and 100 ng. The following antagonists were tested: 1) the V₁/V₂ receptor antagonist [d(CH₂)₅¹,O-Et-Tyr²,Val⁴,Arg⁸]VP (TV-AVP), which is potent at the V₁ and V₂ receptors (24); 2) the V₁ receptor antagonist [d(CH₂)₅¹,O-Me-Tyr²,Arg⁸]VP (MeT-AVP; Manning compound), which has ~100-fold greater affinity for V₁ than for the V₂ receptor (24); and 3) the linear V₂ receptor antagonist [propionyl¹,O-Et-D-Tyr²,Val⁴,Abu⁶, Arg^8,9]VP (TVA-AVP), which is one of the most potent V₂ receptor antagonists (25, 26).

Effects of intraventricular VP on NaCl intake. The effects of intraventricular injections of 1 ng, 100 ng, and 1 µg VP were tested in three separate groups of rats. Different groups were used because of the sensitization that occurs to VP with repeated central injections of VP (2, 46) and the potentially harmful effects of VP (e.g., seizures, death). Rats in group 1 (n = 6) were tested with saline and 1 ng VP; group 2 (n = 6) was tested with saline and 100 ng VP; and group 3 (n = 5) was tested with saline and 1 µg VP. A crossover design was used. Sodium-deficient rats were administered an intraventricular injection (3 µl) of either saline or VP and then given access to 0.3 M NaCl in calibrated bottles. Intake was measured to the nearest 1.0 ml at 10, 20, 30, and 60 min. The Purina chow and tap water were returned to the home cages after the test. Sodium deficiency was induced the following week, and the rats were administered the remaining drug injection.

Effects of VP receptor antagonists on NaCl intake. Three groups of rats were used to test the effects of lateral ventricular injections of V₁/V₂ (TV-AVP; n = 8), V₁ (MeT-AVP; n = 6), and V₂ (TVA-AVP; n = 9) VP receptor antagonists on NaCl intake. Rats in each group were assigned a random order of injections of saline and doses (10 and 100 ng) of the receptor antagonist. On the test days, the low-sodium diet and water were removed from the cages, and the sodium-deficient rats were administered their assigned intraventricular injection. Rats were then given access to 0.3 M NaCl. Intake was measured (nearest ml) at 10, 20, 30, and 60 min. Purina chow and water were then returned to the cages, and the test procedure was repeated the following week. The testing continued weekly until the rats had been tested with each of the injection conditions.

Sucrose intake test. The results showed that intraventricular injections of VP, MeT-AVP, and TV-AVP suppressed NaCl intake, while the V₂ antagonist had no effect on drinking NaCl. It was of interest to determine if VP and MeT-AVP were causing an overall inhibition of ingestive behavior, regardless of the taste.

Histological analysis. The placement of the lateral ventricle cannulas was confirmed postmortem in all rats. Rats were deeply anesthetized, and 3 µl of India ink was injected through the cannula. The rat was perfused successively with saline and 10% formalin solution. The brains were extracted and cut in the coronal plane, and the placement of the cannula tract and the deposition of ink were recorded.

Results

Histology. Rats that served in the experiment had met the drinking criterion after intraventricular injection of angiotensin. At the conclusion of the experiment, ink injections showed that the cannulas were similarly positioned in all of the rats. The cannulas were positioned in the anterior portion of the lateral ventricle, and ink was deposited into the lateral ventricle of all of the rats.

VP. Intake was analyzed using separate repeated-measures 2 (injection) × 4 (time) ANOVAs, and the data are shown in Fig. 1. At the low dose (1 ng), one rat displayed mild barrel rotation that subsided 2-3 min after the injection. Intraventricular injections of 1 ng VP had no significant effect on NaCl intake by sodium-deficient rats. Intraventricular injections of 100 ng VP elicited barrel rotation in two of the six rats, and the duration and rate of rotation were greater than those observed after the lower dose. Injections of 100 ng VP significantly reduced NaCl intake [F(1,5) = 9.9, P < 0.02]. The significant injection × time interaction reflected that 100 ng VP produced the greatest inhibition in NaCl intake early during the test and that intake then increased and approached the amount consumed by saline-treated rats [F(3,15) = 3.9, P < 0.03]. Intraventricular injections of 1 µg VP produced severe barrel rotation and seizurelike activity in three of five rats. Two of these rats drank 0 ml and the third drank 3 ml during the first 30 min of the test. This resulted in there being a significant difference in NaCl intake after injections of 1 µg VP and saline [F(1,4) = 41.3, P < 0.003]. The inhibitory effect of 1 µg VP on intake was most pronounced at 10 min, and by 1 h intake increased, but intake was still 30% below that of saline-treated rats (Fig. 1). The results show that as the dose of VP was increased, the magnitude of the barrel rotation increased and NaCl intake decreased. This behavioral profile raises the possibility that the reduction in NaCl intake was linked to the motor disruption caused by VP.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Mean intake (ml) of 0.3 M NaCl (±SE) by sodium-deficient rats after intraventricular injections of isotonic saline and 1 ng (A), 100 ng (B), or 1 µg vasopressin (VP; C). * P < 0.02 and ** P < 0.003.

V₁/V₂ receptor antagonist. NaCl intake after injections of saline and the V₁/V₂ receptor antagonist was analyzed using a two-way (injection × time) ANOVA. The significant main effect of injection indicated that intraventricular injections of TV-AVP suppressed NaCl intake by sodium-deficient rats [F(2,14) = 6.3, P < 0.01 (Fig. 2)]. The significant drug × time interaction reflected that early during the test, rats in the three injection conditions drank a similar amount of NaCl. However, as the test progressed, saline-treated rats continued to ingest NaCl, whereas rats injected with 10 and 100 ng TV-AVP stopped drinking NaCl [F(6,42) = 4.5, P < 0.001]. Behavioral observations revealed that there was nothing that distinguished rats treated with saline from those treated with TV-AVP. Early in the test, drinking behavior predominated for both groups, and resting (inactivity) and grooming the body in a normal rostral-caudal direction often appeared after drinking subsided. As such, the behaviors displayed by rats injected with saline and TV-AVP were similar.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   NaCl intake (ml) after intraventricular injections of V1/V2 receptor antagonist [d(CH2)51,O-Et-Tyr2,Val4, Arg8]VP (TV-AVP; A), V1 receptor antagonist [d(CH2)51,O-Me-Tyr2,Arg8]VP (MeT-AVP; Manning compound; B), and V2 receptor antagonist [propionyl1,O-Et-D-Tyr2,Val4,Abu6,Arg8,9]VP (TVA-AVP; C). Significant difference compared with NaCl intake after isotonic saline injection for that group: * P < 0.01, ** P < 0.001.

V₁ receptor antagonist. Injections of the V₁ receptor antagonist had a significant effect on NaCl intake [F(2,10) = 10.5, P < 0.003]. Further comparisons showed that intraventricular injections of 10 ng MeT-AVP had no significant effect on NaCl intake compared with saline injection in sodium-deficient rats. The analysis comparing NaCl intake after injections of saline and 100 ng MeT-AVP revealed that this dose of the V₁ receptor antagonist significantly reduced NaCl intake in sodium-deficient rats [F(1,5) = 12.3, P < 0.01]. The significant injection × time interaction showed that while 100 ng MeT-AVP suppressed intake during the entire test, the reduction in NaCl intake was greatest at 10 min after the injection [F(3,15) = 3.5, P < 0.04 (Fig. 2)]. Intraventricular injections of the V₁ antagonist had no adverse effect on the rat's motor function.

V₂ receptor antagonist. Intraventricular injections of TVA-AVP had no significant effect on NaCl intake by sodium-deficient rats (P > 0.1). The intake of NaCl by saline- and TVA-AVP-treated rats increased during the test period [F(3,24) = 23.6, P < 0.001]. The injection × time interaction was not significant [P > 0.2 (Fig. 2)].

Sucrose intake. Intraventricular injections of VP significantly reduced sucrose intake (Fig. 3). The significant group × time interaction reflects that VP suppressed sucrose intake early in the test and that intake increased to that of saline-treated rats at 30 min [F(3,18) = 5.4, P < 0.008].

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Sucrose intake (ml) after intraventricular injection of isotonic saline and 100 ng VP. * P < 0.01.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Sucrose intake (ml) after intraventricular injection of isotonic saline and 100 ng MeT-AVP. (Intake was not significantly different after the 2 treatments.)

	EXPERIMENT 2

TOP ABSTRACT INTRODUCTION EXPERIMENT 1 EXPERIMENT 2 DISCUSSION REFERENCES

The results of experiment 1 showed that intraventricular injections of VP reduced NaCl intake. The reduction in NaCl intake after VP was accompanied by motor disturbances and the suppression in intake appeared secondary to this motor disruption. The motor activity triggered by VP may also have caused the reduction in sucrose intake. NaCl intake was also reduced after intraventricular injections of a V₁/V₂ receptor antagonist and a selective V₁ receptor antagonist, but not a V₂ receptor antagonist. In contrast to the motor disruption induced by VP, intraventricular injections of the VP receptor antagonists did not induce motor disturbances. Also, whereas intraventricular injections of VP reduced the intake of sucrose, injections of MeT-AVP had no effect on sucrose intake. These results lead to the conclusion that the reduction in NaCl intake is a primary consequence of the intraventricular injections of MeT-AVP and that endogenous VP acting at V₁ receptors plays a role in arousing salt intake by sodium-deficient rats.

However, given the similar effect of VP and MeT-AVP on reducing NaCl intake, another possible explanation must be considered. Because intraventricular injections of MeT-AVP mimicked the action of VP on NaCl intake, the suppressive effect of the receptor antagonists could possibly be due to their action as a receptor agonist rather than an antagonist. Indeed, intraventricular injections of V₁ receptor "antagonists" at doses of 100 ng and higher may or may not show agonist-like properties (7, 8, 28). To test whether 100 ng MeT-AVP was acting as a receptor antagonist or agonist, the effects of intraventricular injection of VP and MeT-AVP on MAP were compared. If MeT-AVP were acting as a receptor antagonist, then its effects on MAP should be opposite to the pressor action of centrally administered VP. Second, as an antagonist, intraventricular pretreatment with MeT-AVP should block the pressor effect of VP.

Methods

Animals and surgery. Male Sprague-Dawley rats (n = 8; Charles River Laboratories) were housed in hanging wire-mesh cages and maintained on a 12:12-h light-dark cycle. Rats were maintained on ad libitum access to Purina chow and water.

Procedure. Rats (n = 5) were removed from their home cages and moved to a clear Plexiglas chamber (23 × 24 × 22 cm). The arterial catheter was flushed with heparin-saline to ensure its patency and then connected to the pressure transducer. The output from the pressure transducer was fed to a PowerLab System, and data were stored on the computer. Blood pressure was recorded at a rate of 1,000 samples/s. Rats were left undisturbed for 30 min. The obturator was then removed, and an injector was inserted into the intraventricular cannula. Blood pressure data were collected for another 5 min (baseline), and then the rat was administered an intraventricular injection (3 µl) of isotonic saline. MAP was monitored for the following 30 min. After the saline test, an injector filled with MeT-AVP was inserted into the intraventricular cannula. Baseline blood pressure was recorded for 5 min, and then the rats were administered 100 ng MeT-AVP. MAP was recorded for the following 30 min. Rats were allowed additional time for MAP to return to baseline levels. This procedure was repeated another two times. On the third test, after the baseline recording period, rats were administered an intraventricular injection of 100 ng VP. Again, MAP was allowed to return to baseline levels. On the final test, rats were administered an intraventricular injection of 100 ng MeT-AVP just before the intraventricular injection of 100 ng VP. In all cases, the blood pressure response was measured for 30 min after the intraventricular injection.

Results and Discussion

View larger version (36K): [in this window] [in a new window]   Fig. 5.   A: change in mean arterial blood pressure (MAP) in awake, behaving rats before (baseline) and after intraventricular injection of saline, 100 ng VP, or 100 ng MeT-AVP or after pretreatment with MeT-AVP before VP injection. B: change in MAP from baseline of conscious rats after intraventricular injections of 500 ng MeT-AVP compared with the effects of 100 ng VP and 100 ng MeT-AVP (replotted from A).

The present results are consistent with reports showing that intraventricular injection of VP increases MAP (33). The structure-activity relation for the V₁ receptor antagonist was opposite to that of VP. Intraventricular injection of 100 ng MeT-AVP caused a decrease in MAP. This result suggests that, in the present test conditions, 100 ng MeT-AVP shows VP receptor antagonistic activity. Moreover, pretreatment with 100 ng MeT-AVP blocked the effect of VP on MAP. Had the intraventricular injection of MeT-AVP had agonistic activity then there should have been an enhancement of the pressor action of VP. The results show that 100 ng MeT-AVP did indeed act as a VP receptor antagonist. Furthermore, the results support a role of endogenous brain VP in the control of cardiovascular function. Diamant and De Weid (8) reported that intraventricular injection of the V₁ receptor antagonist increases heart rate. Our findings that a selective V₁ receptor antagonist reduces MAP are consistent with the suggestion (8) that endogenous VP acting at V₁ receptors plays a role in the maintenance of tonic autonomic control over cardiovascular function.

Intraventricular injections of 500 ng MeT-AVP produced effects opposite to 100 ng MeT-AVP and mimicked the effects of 100 ng VP on MAP (Fig. 5B). Unlike the decline in MAP after 100 ng MeT-AVP, intraventricular injections of 500 ng MeT-AVP and VP both induced an elevation in MAP. The ability of high doses (500 ng) of MeT-AVP to elevate MAP suggests that at this dose, MeT-AVP may be a partial agonist at the receptor sites underlying the pressor effect. While the pressor action of MeT-AVP is consistent with the interpretation that the receptor antagonist can have agonistic activity, it is also possible that at high doses, effects on MAP identical to VP result from MeT-AVP acting on distinct, non-VP receptors.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENT 1 EXPERIMENT 2 DISCUSSION REFERENCES

The objective of the experiment was to reevaluate the role of brain VP in the control of NaCl intake by sodium-deficient rats. Like a previous report (36), intraventricular injections of VP significantly reduced NaCl intake. Behavioral observations and the results obtained using VP receptor antagonists, however, call into question whether this effect of VP is physiological. First, there was a correspondence between the ability of the dose of VP to elicit barrel rotations and the ability of VP to reduce NaCl intake. The low dose (1 ng) had a minimal effect on motor function and no significant effect on NaCl intake. As the dose of VP was increased, the magnitude of the motor disturbance increased, and NaCl intake was suppressed. This motor effect is consistent with previous findings showing that central administration of 10 ng VP causes barrel rotations (7). Second, the issue of whether the behavioral action of VP was selective for NaCl or generalized to unrelated tastes was addressed by measuring sucrose intake after intraventricular injection of VP. Injections of VP, at a dose that reduced NaCl intake, also reduced sucrose intake in rats.

The role of endogenous VP during sodium appetite was tested using VP receptor antagonists. Pretreatment with the V₁/V₂ receptor antagonist suppressed salt appetite. Furthermore, the decrease in intake was not accompanied by abnormal motor behaviors. This receptor antagonist does not discriminate between the V₁ and V₂ receptor subtypes (24). Therefore, more selective VP receptor antagonists were tested. Intraventricular injections of the V₁ receptor antagonist, but not V₂ antagonist, reliably suppressed NaCl intake by sodium-deficient rats. The ability of a selective V₁ receptor antagonist to suppress salt appetite supports the view that endogenous VP has a significant role in the genesis of salt intake. The reduced NaCl intake caused by the V₁ antagonist did not reflect a generalized inhibition of ingestive behavior because, unlike VP, intraventricular injections of the V₁ antagonist had no effect on sucrose intake.

Depending on the dose, intraventricular injections of V₁ receptor "antagonists" may show agonist-like properties (8, 28). This issue was addressed in experiment 2. Inferences about whether the drug is having agonist and/or antagonistic actions can be made by comparing the effects of the putative receptor antagonist with those of VP and by determining if the receptor antagonist blocks the effects of VP. First, regarding the antagonist's blocking action, intraventricular injections of VP induce barrel rotations, and intraventricular injections of 100 ng V₁ receptor antagonist before the intraventricular injection of VP prevent the barrel rotations (7). Second, the present results are consistent with the finding that central injection of VP stimulates a rise in MAP (33). We show that intraventricular injection of 100 ng MeT-AVP had the opposite effect and produced a decline in MAP. Moreover, pretreatment with 100 ng MeT-AVP before VP injection blocked the pressor effect of VP. The structure-activity relation of VP and 100 ng MeT-AVP was opposite to one another, and on the basis of these observations, we attribute the effects of 100 ng MeT-AVP to its receptor antagonistic activity.

It is important to note that this interpretation is warranted for the 100-ng dose of MeT-AVP. Dose and the specific V₁ receptor antagonist used determine whether the receptor antagonist shows agonist activity. For example, injections of high doses (5 µg) of MeT-AVP potentiate the barrel rotation and hypothermia induced by intraventricular injection of VP, and 1.5 µg MeT-AVP can itself induce hypothermia (28). Furthermore, we show that intraventricular injections of 500 ng MeT-AVP do indeed show partial agonist properties and mimicked the action of 100 ng VP on MAP.

Central injections of the V₁ receptor antagonist that lowered MAP also reduced NaCl intake. It is unlikely that the reduced salt intake is secondary to the lowered MAP for several reasons. Hypotension is present after various treatments that induce a salt appetite. For example, the administration of furosemide and captopril produces a modest decrease in MAP and stimulates the ingestion of hypertonic NaCl. Also, salt appetite is stimulated, not reduced, after administration of a vasodilator (minoxidil) and captopril, which causes a modest decrease in MAP. Finally, the magnitude of the salt appetite is reduced when phenylephrine is used to block the decrease in MAP after injections of furosemide and captopril treatment (43). Thus the small decrease in MAP after intraventricular injection of the V₁ receptor antagonist probably has no role in reducing salt intake by sodium-deficient rats in the present study.

Oxytocin is another neurohypophysial peptide that is synthesized at multiple sites in the brain and projects to a variety of brain areas. Stricker and Verbalis (39) have shown that brain oxytocin is involved in the inhibitory control of sodium appetite. Oxytocin receptors bind VP, and many VP receptor antagonists are also potent at the oxytocin receptor (18, 27). These observations raise the question of whether the effects reported in the present study are due to an antioxytocinergic action of the VP receptor antagonist. This is unlikely for several reasons. First, MeT-AVP is 100 times more active at blocking the V₁ receptor than the oxytocin receptor (27). Second, others have shown that pretreatment with MeT-AVP neither blocked nor enhanced the behavioral actions of oxytocin (9). This result shows that the V₁ receptor antagonist was neither an oxytocin receptor antagonist nor an agonist. Third, central oxytocin acts to inhibit sodium appetite (39), whereas the present results suggest that endogenous brain VP acts to facilitate NaCl intake. Had the V₁ receptor antagonist had antioxytocin properties, then the prediction would be that NaCl intake would be enhanced, not suppressed. For these reasons, it would appear that the reduction in sodium appetite after intraventricular injection of MeT-AVP is due to an action at VP and not oxytocin receptors. The finding that brain VP plays an excitatory role and oxytocin an inhibitory role in the control of NaCl is one more example that VP and oxytocin exert opposite central nervous system effects (6).

The present results strongly implicate brain VP transmitter activity at V₁ receptors in the arousal of salt appetite. There is other evidence suggesting that brain VP release correlates with elevated salt intake. Adrenalectomized rats show an elevated preference for hypertonic NaCl, and hypothalamic levels of VP are elevated in adrenalectomized rats without changes in the neurohypophysial levels of VP (14, 37). This elevation in brain VP was not simply a stress response because immobilization did not result in an increase in VP levels. These results suggest that VP is released centrally to modulate adaptive responses stemming from adrenalectomy (14).

The proposed central role for VP is consistent with a broadly defined role of VP in the maintenance of fluid balance, particularly as it relates to the regulation of extracellular fluid volume. Extracellular dehydration is a factor in sodium deficiency, and volume reduction is corrected through behavioral and endocrine processes. It appears that VP acts on both processes. Hypovolemia triggers the central and peripheral release of VP (4). One central action of VP is to stimulate water intake (41), and plasma VP reduces urine loss. We show that a second central action of VP is to stimulate salt intake. The ingestion of NaCl would serve to increase extracellular sodium and shift intracellular water into the extracellular space (12) and prevent the osmotic dilution that might otherwise result for the sole ingestion of water (40). These combined actions of VP would serve to restore extracellular fluid volume. In these ways, central VP has actions like those of brain ANG II in stimulating the ingestion of both water and NaCl (19, 43).

The site of action of the intraventricularly injected VP receptor antagonists is not specified in this study. The correspondence between the location of VP binding sites, their proximity to the ventricular system, and the relationship of the receptor distribution to structures controlling salt intake suggests two possible target sites. The antagonists would have relatively direct access to the OVLT and SFO. These structures surround the third ventricle and play an important role in the control of body fluid osmolarity (19). Also, VP terminals and V₁ receptors are present in both of these structures (1, 5, 20, 38).

In the present experiment, sodium appetite was aroused by injections of furosemide and DOCA. These treatments produce intense c-Fos expression in the OVLT. Moreover, even when different treatments are used to arouse a sodium appetite, a common feature is the induction of c-Fos expression in the OVLT. This result suggests a key role for the OVLT in the genesis of sodium appetite (23, 35, 45). The presence of VP in neuronal projections to the OVLT raises the possibility that VP synaptic action contributes to the induction of c-Fos expression in this area after sodium deficiency and the arousal of salt ingestion. Regardless of the central site of action, the present results show that the sodium appetite aroused by injections of DOCA and furosemide involves brain VP acting at V₁ receptors.