Studies of Cerebral Osmoreceptors in Anesthetized Dogs: The Effect of Intravenous and Intracarotid Infusion of Hyper-Osmolar Sodium Chloride Solutions during Sustained Water Diuresis

The function of the suggested hypothalamic osmoreceptors was investigated in dogs during light chloralose anesthesia. The dogs were subjected to an i.v. load of 40 ml/kg b.wt. of a hypo-osmolar solution of sodium chloride and glucose. This degree of hydration was kept constant by a specially constructed servo system based on the weight of the animal. During water diuresis the renal free water clearance remained essentially constant for several hours (mean about 0.2 ml/kg b.wt.-min). Renal sodium excretion was low (mean 0.82 μmol/kg b.wt. · min) and decreased continuously throughout the experimental period. I.v. infusion of hyper-osmolar sodium chloride solution (1.33 mmol/kg b.wt. in 30 min) was followed by prolonged parallel increases in free water clearance and sodium excretion, without any detectable change in the excretion of osmoles and potassium. The renal response to bilateral infusion of hyper-osmolar NaCl (1.33 mmol/kg b.wt. in 30 min) into the common carotid arteries was identical to the response to i.v. infusion. The estimated increase in the osmolality of the carotid blood was 2.2 %. In seven out of eight experiments intracarotid infusion of NaCl (1.33 mmol/kg b.wt. in 8 min) did not elicit any reduction in free water clearance. On the contrary, an increase was found similar to that obtained after i.v. infusion. The estimated increase in the osmolality of the carotid blood was 8.4%. The present results question the validity of the currently held view that hypothalamic osmoreceptors play an important role in the control of the osmolality of plasma.     

Natriuresis induced by intracarotid infusion of hypertonic NaCl.

The intracarotid infusion of hypertonic sodium chloride in anesthetized dogs pretreated with vasopressin and mineralocorticoid resulted in an increase in sodium excretion as well as an increase in the percent of the filtered load of sodium excreted. This increase was not due to changes in renal hemodynamics or arterial blood pressure. The change in fractional sodium excretion was approximately 1%. Intracarotid infusion of isotonic sodium chloride did not result in a natriuretic response. The intracarotid infusion of hyperosmotic glucose did not evoke an increase in sodium excretion. Femoral vein infusions of hypertonic sodium chloride failed to evoke any natriuretic response. These data indicate that a sodium-sensing mechanism may exist in the brain that can alter the renal handling of sodium. The efferent mechanism of the response is discussed.     

Effect of hypertonic intracarotid infusions on plasma vasopressin concentration

The effect of short-term bilateral intracarotid infusions of hypertonic saline on plasma vasopressin concentration (pAVP) was evaluated in five dogs. Intracarotid infusion of saline at 90 mumol . kg-1 . min-1 . artery-1 significantly (P less than 0.05) increased jugular vein osmolality (pOsm) and sodium concentration (pNa+) within 2 min. Saphenous vein pOsm was not altered during the 6 min of infusion, whereas pNa+ was increased (P less than 0.05) from 0.8 +/- 0.1 to 2.3 +/- 0.3 pg/ml. Subsequent experiments using hypertonic saline infusions of 90 and 180 mumol . kg-1 . min-1 administered intracarotidly and intravenously for 6 min were performed. Intracarotid isotonic infusions and intravenous hypertonic infusions did not significantly alter pAVP. Hypertonic intracarotid saline increased jugular vein pOsm and pNa+ in a dose-related fashion, whereas saphenous vein pOsm and pNa+ were not significantly changed after 6 min of infusion. Plasma vasopressin, compared with the isotonic intracarotid infusion (1.5 +/- 0.3 pg/ml), was increased (P less than 0.05) after hypertonic saline to 3.2 +/- 0.6 and 4.8 +/- 0.2 pg/ml for the 90 and 180 mumol . kg-1 . min-1 infusions, respectively. The cerebral osmolality indicated by jugular vein pOsm was therefore increased in the absence of changes in systemic pOsm during intracarotid hypertonic infusions. The increase in pAVP in response to these changes in pOsm supports the presence of central osmoreceptors regulating vasopressin release in the area of distribution of the common carotid arteries.     

Effects on renal sodium and potassium excretion of vasopressin and oxytocin in conscious dogs

Renal effects of arginine vasopressin and oxytocin were studied in conscious dogs, made water-diuretic by a waterload equivalent to 2% of body weight. Body water and content of sodium were maintained by separate servo-controlled infusions. Peptides were infused for 60 min at rates of 50 pg kg^-1 min^-1 (arginine vasopressin) or 1 ng kg^-1 min^-1 (oxytocin), either separately or combined. Infusions increased plasma arginine vasopressin to 1.9 ± 0.2 (arginine vasopressin alone) and 1.8 ± 0.3 pg kg^-1 (arginine vasopressin plus oxytocin and plasma oxytocin to 72 ± 5 (oxytocin alone) and 77 ± 8 pg ml^-1 (oxytocin plus arginine vasopressin). Arginine vasopressin or arginine vasopressin plus oxytocin increased urine osmolality similarly by a factor of 13, decreased urine flow to between 5 and 7% of control and decreased free water clearance. Oxytocin reduced urine flow and free water clearance and increased urine osmolality by a factor of 2. Oxytocin and arginine vasopressin separately increased excretion of sodium from 4 ± 2 to 15 ± 6 μmol min^-1 and from 7 ± 4 to 25 ± 13 μmol min^-1, respectively. Arginine vasopressin plus oxytocin led to a pronounced natriuresis (13 ± 4 to 101 ± 27 μmol min^-1). Arginine vasopressin and arginine vasopressin plus oxytocin increased the excretion of potassium by a factor of 2.5. Oxytocin and arginine vasopressin plus oxytocin increased urinary Na⁺/K⁺ ratio by a factor of 3.7. It is concluded, that oxytocin at plasma concentrations of 70–80 pg ml^-1 has modest antidiuretic and natriuretic effects and that the combined action of arginine vasopressin oxytocin may elicit supra-additive natriuretic effects.     

Studies of cerebral osmoreceptors in anesthetized dogs: the effect of intravenous and intracarotid infusion of hyper-osmolar sodium chloride solutions during sustained water diuresis.

The function of the suggested hypothalamic osmoreceptors was investigated in dogs during light chloralose anesthesia. The dogs were subjected to an i.v. load of 40 ml/kg b.w.t. of a hypo-osmolar solution of sodium chloride and glucose. This degree of hydration was kept constantly by a specially constructed servo system based on the weight of the animal. During water diuresis the renal free water clearance remained essentially constant for several hours (mean about 0.2 ml/kg b.w.t. min). Renal sodium excretion was low (mean 0.82 mumol/kg b.w.t. min) and decreased continuously throughout the experimental period. I.v. infusion of hyperosmolar sodium chloride solution (1.33 mmol/kg b.w.t. in 30 min) was followed by prolonged parallel increases in free water clearance and sodium excretion, without any detectable change in the excretion of osmoles and potassium. The renal response to bilateral infusion of hyper-osmolar NaCl (1.33 mmol/kg b.w.t. in 30 min) into the common carotid arteries was identical to the response to i.v. infusion. The estimated increase in the osmolality of the carotid blood was 2.2%. In seven out of eight experiments intracarotid infusion of NaCl (1.33 mmol/kg b.w.t. in 8 min) did not elicit any reduction in free water clearance. On the contrary, an increase was found similar to that obtained after i.v. infusion. The estimated increase in the osmolality of the carotid blood was 8.4%. The present results question the validity of the currently held view that hypothalamic osmoreceptors play an important role in the control of the osmolality of plasma.     

A central osmosensitive receptor for renal sodium excretion.

1. The effect on renal Na and water excretion of increasing the NaCl concentration of blood supplying the brain was investigated in conscious water-loaded sheep. Intracarotid infusion ot 4 M-NACl at 0-8 ml./min for 60 min was compared with equivalent intrajugular infusion. 2. A more rapid increase in renal Na excretion and urine osmolality occurred with the intracarotid infusions than with intrajugular infusions. 3. Intracarotid infusions of 2 M sucrose or fructose at 1-6 ml./min for greater increase in renal Na excretion, urine osmolality and a decrease in urine flow rate. 4. The results suggest that there are receptors in the brain sensitive to changes in extracellular tonicity which influence renal Na excretion. It is possible that changes in ADH secretion alone mediate the early natriuresis seen with intracarotid hypertonic infusions although an alternative concurrent mechanism cannot be ruled out.     

Effect of angiotensin II on renal water excretion.

In the present study the effect of angiotensin II (AII) on renal water excretion was evaluated. In dogs undergoing a water diuresis, neither the intravenous (IV) (40ng/kg per min) nor intracarotid (5-10 ng/kg per min) infusion of AII significantly altered urinary osmolality (Uosm) or free-water clearance (CH2O). Intravenous infusion of a competitive inhibitor of AII (1-sarcosine,8-glycine AII) into hydropenic dogs also failed to alter Uosm and CH2O significantly. To examine whether AII might suppress, rather than stimulate, vasopressin release, AII was also infused into hydropenic animals. No effect on Uosm and CH2O was observed during the intracarotid infusion. A significant fall in Uosm and rise in CH2O occurred during the intravenous AII infusion, but reversal after cessation of the infusion was incomplete and statistically not significant. Some suppression of antidiuretic hormone (ADH) release during the intravenous infusion of AII, however, was suggested since no similar alteration in renal water excretion was observed during an intravenous AII infusion in hypophysectomized animals receiving a constant infusion of ADH. Taken together, the present results provide no evidence for a direct effect of AII to alter ADH release or to interfere with the peripheral action of ADH. Suppression of ADH release may sometimes occur with pressor doses of intravenous angiotensin, but this effect is clearly less consistent than previously observed with intravenous norepinephrine.     

The role of antidiuretic hormone in cold-induced diuresis in the anaesthetized rat

The aim of this study was to investigate whether the increased diuresis in consequence of hypothermia is due to a depression of the hypothalamic release of antidiuretic hormone (ADH). The plasma concentration of antidiuretic hormone and the effect of intravenous (i.v.) administration of 65 ng kg^?1 desmopressin (selective V₂-receptor agonist) were determined in the anaesthetized rat. In spite of a 50% (P < 0.001) decrease in glomerular filtration rate, urine flow increased sixfold (P < 0.01) and urine sodium excretion increased sevenfold (P < 0.05), whereas urine osmolality decreased (P < 0.001). At the same time plasma antidiuretic hormone decreased from 7.5 ± 1.1 to 3.8 ± 0.4 pg mL^?1 (P = 0.01). After injection of desmopressin urine flow was completely restored, whereas urine osmolality and sodium excretion were only partially normalized. Since tubular conservation of water and fractional water reabsorption decreased during hypothermia, the diuresis must have resulted from an augmented loss of water. This is further supported by the fact that osmolal excretion was not influenced either by hypothermia or by desmopressin. It is concluded that the diuresis in consequence to hypothermia is due both to a decrease in the release of ADH and to a reduction of renal medullary hypertonicity.     

Renal response to AVP and DDAVP after angiotensin II blockade in the rat

Inhibition of the renin-angiotensin system with saralasin or with SQ 20,881 was used to examine a possible interaction between endogenous angiotensin II and antidiuretic hormone in renal water excretion. In sodium-depleted anesthetized rats undergoing water diuresis, neither saralasin nor SQ 20,881 had any significant effect on water excretion. Both drugs, however, markedly potentiated the antidiuretic response to arginine vasopressin. A similarly enhanced antidiuresis was also observed when the renal response to 1-desamino-8-D-arginine vasopressin was investigated in the presence or absence of saralasin. There were no significant changes in sodium or potassium excretion or in glomerular filtration rate in any of the groups studied. Neither was there any interaction of the drugs with relation to blood pressure. These results seem to indicate that endogenous angiotensin II is capable of antagonizing the hydroosmotic effect of antidiuretic hormone.     

Left atrial receptors and the antidiuretic hormone

1. Stimulation of left atrial receptors by distending balloons located at the junction between the pulmonary veins and the left atrium and in the body of the left atrium resulted in a diuresis in anaesthetized dogs.2. In the same dogs the diuresis was not consistently associated with a reduction in the antidiuretic activity of plasma as compared with that during the initial control period. The antidiuretic activity of dog plasma was assayed in water-loaded rats anaesthetized with ethanol.3. Intravenous injections of vasopressin suppressed a water diuresis in water-loaded dogs anaesthetized with chloralose. The changes in antidiuretic activity of dog plasma which resulted from these injections were reliably detected by the bio-assay in the rat.4. It is concluded that the diuretic response to the stimulation of left atrial receptors is not accompanied by a reduction in the plasma antidiuretic hormone level.     

Lithium-induced diuretic effect of antidiuretic hormone in rats.

The effect of a low dose of lithium (1 meq/kg per day) on renal function and its response to antidiuretic hormone (ADH) was studied in unanesthetized rats. This dose of lithium itself had no influence on renal water and electrolyte excretion, but lithium-treated rats responded paradoxically to exogenous ADH by increases in urinary volume, excretion of total solutes, sodium, potassium, and phosphate. Administration of ADH in the presence of lithium led to a lowering of urine osmolality, but free water clearance was not significantly reduced. Adenylate cyclase from the renal medulla of animals treated with ADH and lithium had a lower response to synthetic vasopressin in vitro than in animals treated with lithium alone. The results suggest that exogenous ADHis diuretic in the presence of a low concentration of lithilm. The predominant mechanism for this diuresis is probably inhibition of electrolyte and isomotic water reabbsorption in various nephron segments, including those proximal to the collecting ducts. ADH also markedly increased urinary excretion of lithium and appears to promote accumulation of lithium in the renal medulla.     

Mechanism of cold diuresis in the rat

The effect of cold exposure (CE) on renal water excretion has not been clearly delineated. Conscious rats were exposed to decreased ambient temperature (15 degrees C). Forty-five minutes of CE resulted in reversible increases in urine flow and decreases in urine osmolality. The diuresis was not due to a diminished response to vasopressin (VP), as the antidiuresis associated with 500 microU of Pitressin given to water-diuresing rats was comparable at 15 and 30 degrees C. To determine whether the diuresis was due to intrarenal factors, glomerular filtration rate, renal blood flow, sodium excretion, and osmolar clearances were measured and found to be equivalent during control and cold conditions. To determine whether the observed diuresis was due to suppression of endogenous VP, VP-free Brattleboro rats undergoing a constant VP infusion were cold exposed. In these rats, CE was not associated with a change in either urine flow or urinary osmolality. This antidiuretic hormone-mediated mechanism was corroborated by a decrease in immunoassayable VP levels. To determine the mechanism whereby CE suppresses endogenous VP, plasma osmolality and hemodynamic parameters were measured. Although CE was not associated with a change in plasma osmolality, it did result in a significant increase in both mean arterial pressure and cardiac index. Pretreatment of rats with 6-hydroxydopamine prevented both the increase in mean arterial pressure and cold diuresis. We conclude that the diuresis observed upon exposure to 15 degrees C results from nonosmotic suppression of endogenous VP, as a consequence of the increase in mean arterial pressure.     

Serum arginine-vasotocin (AVT) and afferent and central control of osmoregulation in conscious Pekin ducks

In conscious Pekin ducks adapted either to fresh water or to hypertonic saline (1.9%) as drinking fluid, urinary excretion, salt gland secretion and the serum concentration of radioimmunoassayable arginine-vasotocin (AVT) were examined with regard to their afferent and central control. The experiments were carried out under conditions of water diuresis, osmotic diuresis or supraorbital salt gland secretion, which were induced by continuous infusions of appropriate solutions.Temporary bilateral vagus blockade caused rises in AVT serum concentration accompanied by antidiuresis in hydrated ducks and by inhibition of salt gland secretion in salt-stressed ducks.Rostral brainstem cooling caused decreases of AVT serum concentration and water diuresis in ducks under osmotic diuresis and reduction of AVT serum concentration and inhibition of salt gland secretion in saltstressed ducks.Cerebral osmotic stimulation in hydrated ducks by intracarotid injection or by intracerebroventricular microinfusion of hypertonic NaCl solutions caused antidiuretic reactions associated with rises of AVT serum concentration.Supported by Deutsche Forschungsgemeinschaft (Si 230/2)     

Effects of fluid intake on basal and vasopressin-responsive urinary prostaglandin E

The effects of fluid intake on basal and vasopressin-responsive urinary PGE excretion (UPGEV) were examined in conscious rats under conditions of 1) ad libitum water intake, 2) water deprivation, and 3) water diuresis induced by ad libitum intake of 5% dextrose in water. UPGEV fell progressively during 40 h of water deprivation. Water diuresis after water deprivation increased UPGEV transiently (8 h). Vasopressin (Pitressin tannate in oil, 5 U/kg subcutaneously) increased UPGEV and decreased urine volume (V) in rats on ad libitum water intake but did not alter UPGEV during water deprivation. Indomethacin suppressed UPGEV (70-90%), increased basal urine osmolality (Uosmol), and potentiated the antidiuretic response to Pitressin in rats on ad libitum water intake. Indomethacin accelerated by 8 h the onset of maximal antidiuresis in water-deprived rats but did not significantly alter water balance. During water diuresis, UPGEV declined in the first 8 h after Pitressin. Thereafter, UPGEV increased markedly, concurrent with early vasopressin escape. Indomethacin or meclofenamate inhibited the rise in UPGEV, the decline in Uosmol, and the increase in V of the escape phase. Indomethacin or meclofenamate also impaired the excretion of an acute water load (5% body wt) given during escape. The spontaneous decline in UPGEV during hydropenia may serve to maximize physiologic antidiuresis. Conversely, the marked increase in UPGEV induced by administration of vasopressin during water diuresis may serve to suppress the antidiuretic response and thus play a role in the mediation of escape from physiologically inappropriate antidiuresis.     

Osmotic control of plasma vasopressin in anesthetized dogs

The relation between plasma osmolality (pOsm) and the concentration of immunoreactive vasopressin in plasma from an external jugular vein (pAVP) was studied in dogs prepared with carotid loops and anesthetized with chloralose-pentobarbital. Control meail pAVP was 0.6–1.0 μU/ml, after 24 h of dehydration pAVP was tripled. Isosmotic volume expansion for 10 min elicited a decrease in pAVP in all cases not associated with low control values of pAVP. I.v. hyperosmotic infusions (A pOsm: 9 or 18 mOsm/kg in 10 min) increased pAVP. No significant alterations in pAVP occurred in relation to infusions during which the head or the remainder of the body was selectively supplied by hyperosmotic blood. Statistically, the distribution of the pAVP values suggests the existence of two populations, in euhydrated animals high concentrations (>8 μU/ml) were found in a small fraction of the samples. It is concluded that (i) in chloralose-anesthetized, non-traumatized dogs pAVP is very similar to values found in venous plasma from conscious dogs, (ii) the pAVP measurements support the concept of an episodic secretion of vasopressin, and (iii) under the present circumstances, an ubiquitous increase in pOsm is a more effective stimulus for vasopressin release than a similar but selective increase in the osmolality of the blood flowing towards the head.     

Potentiation of antidiuretic response to systemic angiotensin II by elevation of CSF NaCI concentration

The antidiuretic effect of the simultaneous intracerebroventricular (ICV) infusion of 0.24 M NaCl (0.02 ml/min) and intravenous (i. v.) infusion of angiotensin II (12 ng/kg. min) was studied in hydrated goats, and was compared to the antidiuretic effects of the separate infusions. The combined infusions inhibited the water diuresis for 30 min, whereas the separate infusions only reduced the water diuresis by 25% (ICV NaCl) and by 50% (i. v. angiotensin). The combined infusions increased the urine osmolality on the average by 415%. Corresponding increases induced by ICV NaCl and by i. v. angiotensin were 100 and 160%. The results suggest that systemic angiotensin II and elevated CSF NaCl concentration interact and potentiate each other as stimuli for antidiuretic hormone secretion. It is postulated that this synergism may help to preserve body water in hypovolemic conditions associated with hyperactivity of the renal renin-angiotensin system.     

An attempt to quantitate the contribution of antidiuretic hormone to the diuresis of left atrial distension in conscious dogs

In order to quantitate the contribution of the antidiuretic hormone (ADH) to the diuresis of left atrial distension 52 experiments have been performed in 12 conscious, chronically instrumented beagle dogs. Left atrial pressure was increased by a reversible mitral stenosis by about 10 cm H₂O (1.0 kPa) for 60 min. Plasma ADH concentration (range between 1.3 and 6.0 pg·ml^?1) (radioimmunoassay) decreased in every experiment, the average decrease being about 50%. An i. v. infusion of vasopressin (0.05 mU ·min^?1·kg^?1) abolished the diuretic effect of left atrial distension or decreased the urine volume below control values; natriuresis was not affected. The magnitude of the vasopressin effect was dependent on the concurrent sodium excretion: when sodium excretion was low during left atrial distension, vasopressin was more effective in reducing the urine volume than when sodium excretion was high. It is concluded that the diuresis of left atrial distension is mediated (a) by a suppression of ADH and (b) by factors controlling sodium excretion, the contribution of these two mechanisms being dependent on the concurrent sodium excretion.     

Impaired ability to increase water excretion in mice lacking the taurine transporter gene TAUT

Summary Cellular taurine uptake or release counteracts alterations of cell volume. Na⁺-coupled taurine transporter TAUT mediates concentrative cellular uptake of taurine. Inhibition of vasopressin secretion by hypotonicity may involve taurine release from glial cells of supraoptic nucleus. We compared renal function of mice lacking TAUT (taut^–/–) and wild-type littermates (taut^+/+). We observed renal taurine loss and subsequent hypotaurinemia in taut^–/– mice. With free access to water, plasma and urine osmolality, urinary flow rate as well as urinary excretion and plasma concentrations of Na⁺ and K⁺ were similar in taut^–/– and taut^+/+ mice, whereas plasma concentrations of urea were enhanced in taut^–/– mice. An oral water load (1 ml/16 g body weight) induced a similar diuresis in both genotypes. Repeating the oral water load immediately after normalization of urine flow rate, however, resulted in delayed diuresis and higher urinary vasopressin/creatinine ratios in taut^–/– mice. In comparison, the repeated diuretic response to vasopressin V₂ receptor blockade was not different between genotypes. Water deprivation for 36 h led to similar antidiuresis and increases of urinary osmolality in both genotypes. Upon free access to water after deprivation, taut^–/– mice continued to concentrate urine up to 6 days, while taut^+/+ mice rapidly returned to normal urinary osmolality. Urinary vasopressin/creatinine ratios and plasma aldosterone concentrations were not different under basal conditions but were significantly higher in taut^–/– mice than in taut^+/+ mice at 6 days after water deprivation. In conclusion, taut^–/– mice suffer from renal taurine loss and impaired ability to lower urine osmolality and to increase urinary water excretion. The latter defect could reside extrarenally and result from a role of taurine in the suppression of vasopressin release which may be attenuated in taut^–/– mice.     

Inhibition of vasopressin-release during developing hypernatremia and plasma hyperosmolality: an effect of intracerebroventricular glycerol

In non-hydrated goats prolonged (3 h, 0.02 ml/min) intracerebroventricular (IVT) infusion of 0.35 M glycerol depressed the plasma vasopressin level during the entire infusion period which resulted in a conspicuous water diuresis outlasting the infusion by about 20 min. Since no compensatory drinking occurred during this sustained water diuresis it gradually induced pronounced dehydration (loss of greater than 1 liter of total body water causing 5% increase in plasma [Na+] and osmolality). The same degree of dehydration was in other experiments induced by water deprivation. It then caused a 5-fold increase in plasma vasopressin level. Corresponding IVT infusions of 0.35 M d-glucose depressed plasma vasopressin level only during the first half of the 3 h infusion period. Consequently, the resulting water diuresis was transient and subsided before the glucose infusion was finished. Plasma renin activity increased during the IVT glycerol infusion and during water deprivation, but was largely unaffected by IVT glucose. Both IVT glycerol and glucose decreased renal sodium excretion. The possibility is discussed that the pronounced ability of IVT glycerol to depress the vasopressin release and thirst is not only due to dilution induced reduction of CSF [Na+], but also to an influence of glycerol on choroidal and/or transependymal Na+-transporting mechanisms.     

Reducing brain sodium concentration prevents post-prandial and dehydration-induced natriuresis in sheep

Renal Na excretion during the 24 h following feeding was studied in sheep. A pronounced natriuresis occured 3.5-5.5 h after feeding. Na excretion then fell to low levels in animals allowed to drink water, but was significantly elevated above this level in water-deprived sheep for most of the remaining period. Both the post-prandial and dehydration-induced natriuresis were prevented by intracerebroventricular (icv) infusions of low Na concentration 0.3 mol 1^-1 mannitol at 1 ml h^_1, and a water diuresis also occurred. These effects were not caused by icv infusion of artificial cerebrospinal fluid (Na concentration = 150 mmol l^-1). As a result, there was a much greater increase in plasma Na concentration and osmolality in the sheep given icv mannitol. Intravenous infusion of vasopressin prevented the water diuresis induced by icv mannitol, but the inhibition of natriuresis was still observed and plasma Na concentration increased by 8 mmol l^-1 over 24 h compared with an increase of 3 mmol l^-1 in dehydrated sheep infused icv with artificial cerebrospinal fluid. The results show that the ambient Na concentration in the brain plays an important role in the normal homeostatic regulation of Na balance by the kidney in sheep.