Pathophysiological roles of arginine vasopressin and aquaporin‐2 in impaired water excretion

Impaired water excretion occurs in patients with the syndrome of inappropriate secretion of antidiuretic hormone (SIADH), adrenal insufficiency, congestive heart failure and liver cirrhosis with ascites (Bartter Schrier, 1988a,b). In these clinical settings, there is hyponatraemia to various extents. Nonsuppressible release of arginine vasopressin (AVP, ADH) is found despite hypoosmolality, which should suppress AVP release to undetectable levels. Reversal of hyponatraemia by specific antagonists of AVP provides conclusive evidence for the role of AVP in pathological states of water retention. In response to AVP, concentrated urine is produced by water reabsorption across renal collecting ducts (Ishikawa, 1993; Knepper Fujita et al., 1995; Nielsen et al., 1997; Xu et al., 1997; Saito et al., 2000). Kidney AQP-2 expression has been quantitatively estimated by urinary excretion of AQP-2 (Kanno et al., 1995; Rai et al., 1997). Approximately 3% of AQP-2 in the collecting duct cells is excreted into urine (Rai et al., 1997), and urinary excretion of AQP-2 positively correlates with plasma AVP levels in normal subjects (Saito et al., 1997c). In this review, we focus on the close association of kidney AQP-2 expression with exaggerated release of AVP in pathological states of impaired water excretion. Furthermore, the diagnostic value of urinary excretion of AQP-2 in disorders of water metabolism dependent on AVP is discussed. AVP is a peptide hormone of the posterior pituitary gland that is synthesized in both magnocellular and parvocellular neurosecretory neurones of the hypothalamus (Arnaud et al., 1974; Brimble Poulain Vale et al., 1983; Knepel et al., 1984). Physiological studies suggest that there is independent control of AVP secretion between magnocellular and parvocellular neurones. Osmotic and nonosmotic stimulations are the two major factors that control AVP release. Osmoreceptors reside in the anteroventral third ventricle (AV3V) region of the hypothalamus, particularly in the organum vasculosum of the lamina terminalis (OVLT) (Thrasher et al., 1982; Thrasher Osaka et al., 1988), and are very sensitive to changes in plasma osmolality (Posm). This region is located outside the blood–brain barrier. There are neural inputs from the osmoreceptors to the PVN and SON, probably mediated via a cholinergic pathway (Sladek Day Catelli et al., 1987). Chemical inhibition and lesions of this nucleus increase plasma AVP levels, suggesting an effect attributable to the interruption of tonic baroreceptor inhibition of AVP release (Blessing Head et al., 1987). A series of studies with interruption of the glossopharyngeal and vagal pathways from arterial baroreceptors also demonstrated potent nonosmotic AVP stimulation (Schrier Schrier et al., 1972). The A1 adrenergic cell group of the ventrolateral medulla is suggested to be involved in the afferent pathway from the nucleus of the tractus solitarius to neurosecretory AVP cells of the SON and PVN. AVP is generated from a precursor form of prepro AVP, which is encoded by the AVP gene on chromosome 20 (Simpson, 1988). The AVP gene has three exons. Exon A encodes the signal peptide, AVP and the N-terminal region of NPII. Exon B encodes the highly conserved central part of NPII. Exon C encodes the C-terminal region of NPII, and the glycoprotein domain (Schmale et al., 1983). The hormone precursor, prepro AVP, contains signal peptide, AVP, NPII and glycoprotein domains. This structural organization of prepro AVP and the AVP gene is conserved among various species. Pro AVP is generated by removing the signal peptide from prepro AVP and by adding a carbohydrate chain to the glycoprotein domain. Additional posttranslational processing of prepro AVP, which yields AVP, NPII and glycoprotein, occurs within neurosecretory granules during their transport to axon terminals in the posterior pituitary. AVP, NPII and the glycoprotein are stored in neurosecretory granules in axon terminals of the posterior pituitary, and are released into the bloodstream in response to osmotic or nonosmotic stimulation. The functional properties of NPII and glycoprotein are not completely understood, but intact NPII is at least necessary for the processing of AVP in the endoplasmic reticulum (Kim et al., 1997; Kim Schrier et al., 1979). Linear regression analysis has yielded the osmotic threshold for AVP secretion and the sensitivity of osmoreceptors. The osmotic threshold for AVP secretion is the point of the interception on the horizontal axis, that is approximately 280 mmol/kg. Several factors potentially affect the osmotic threshold (Dunn et al., 1973). There seems likely to be a species difference in osmotic threshold for AVP secretion; the osmotic threshold ranges from 285 to 292 mmol/kg in the rat, dog and monkey, values higher than the 280 mmol/kg threshold in humans (Dunn et al., 1973; Robertson et al., 1973). Posm decreases by 8–10 mmol/kg during pregnancy, and this decrease is followed by a decrement in the osmotic threshold for AVP release (Davison et al., 1988). The osmotic threshold is also influenced by nonosmotic stimuli (Robertson et al., 1977). Decreases in circulatory blood volume and blood pressure enhance the secretion of AVP by the osmotic stimulus, which shifts (increases) the osmotic threshold to the left in the absence of any change in the sensitivity (Dunn et al., 1973). The sensitivity of osmoreceptors is rather extraordinary. A 1-mmol/kg change in Posm will alter AVP release; this sensitivity, however, is influenced by the nature of the solute. Increases in Posm produced by sodium, sucrose and mannitol exert comparable osmotic effects, but this is not the case with urea or glucose (McKinley et al., 1978; Ishikawa et al., 1980). Other factors, such as the rate of change in Posm, age and drinking behaviour, can also affect osmotic sensitivity. Decreases in arterial blood pressure and circulating blood volume are potent nonosmotic stimuli for AVP secretion, mediated via the high-pressure and low-pressure (left atrial) baroreceptors. It has been generally accepted that a decrement in blood pressure or blood volume of the order of 8–10% is necessary to stimulate AVP secretion. Baylis (1983) demonstrated that the relationships between plasma AVP concentrations and the percentage fall in mean arterial blood pressure is exponential. Several factors, including low cardiac output, left atrial distention, atrial tachycardia, nicotine and hypoxia, are also nonosmotic stimulants for AVP release (Schrier et al., 1979). It is of value to note the separate osmotic and nonosmotic control of AVP release. Electrophysiological studies verified that the osmotic and nonosmotic pathways independently enter the same magnocellular neurosecretory neurones of the PVN and SON (Kannan these plasma AVP concentrations are, however, increased with respective to the low Posm (Saito et al., 1997a). These issues have been dealt with in other reviews. As described initially by Verbalis and Drutarosky (1988), model animals of SIADH were made by the subcutaneous administration of the V2 agonist 1-deamino-8-D-AVP (dDAVP) by osmotic minipumps, and offering a liquid diet. In this model a serum sodium concentration less than 120 mmol/l occurred in association with a concentrated urine throughout the 14-day observation period (Fujisawa et al., 1993). The oral administration of the nonpeptide V2 receptor antagonist OPC-31260 was started on day 7 and continued once a day during the rest of experiment. This manoeuvre promptly normalized serum sodium levels in 12 h in association with an increase in urine volume and a decrease in urinary osmolality. The normalization of serum sodium concentration was maintained during the rest of the experimental period (Fujisawa et al., 1993). Similar results were obtained in experimental SIADH animals with the peptide AVP V2 receptor antagonists d(CH2)5Tyr(Et)VAVP or d(CH2)5DTyr(Et)VAVP (Laszolo et al., 1984; Kinter et al., 1986). In these other studies, the SIADH model was made by subcutaneous injection of pitressin or dDAVP, and the V2 AVP antagonists were administered intravenously or intraperitoneally. The in vivo use of AVP V2 receptor antagonists directly verifies the AVP dependency of impaired water excretion, because both peptide and nonpeptide antagonists are known to specifically antagonize the V2 receptor binding of AVP (Sawyer et al., 1981; Yamamura et al., 1992; Serradeil-Le Gal et al., 1996). These results indicate that the dilutional hyponatraemia in the rats with the experimental model SIADH can be reversed by the AVP V2 receptor antagonists. The efficacy of the nonpeptide AVP V2 receptor antagonist OPC-31260 has also been demonstrated in patients with SIADH (Saito et al., 1997a). A single intravenous injection of OPC-31260 increased urine volume, decreased urinary osmolality (Uosm) and increased serum sodium levels by approximately 3 mmol/l during the 4-h observation period (Fig. 2). Following the initial study, chronic oral administration of OPC-31260 was shown to be very effective in the patients with SIADH (unpublished observation). Alteration in urine volume (a) and Uosm (b) after intravenous administration of the nonpeptide AVP antagonist OPC-31260 in patients with SIADH. Control (▪), 0·1 mg/kg OPC-31260 (○), 0·25 mg/kg OPC-31260 (□), and 0·5 mg/kg OPC-31260 (•). * P < 0·05 and ** P < 0·01 vs. the respective values at the first urine collection. (Data from Saito et al., 1997a.) Hyponatraemia in hypopituitarism and glucocorticoid deficiency is closely related to the nonsuppressible release of AVP despite hypoosmolality (Slessor, 1951; Amhed et al., 1967; Ishikawa Pyo et al., 1993). Hypothalamic AVP mRNA expression was increased in adrenalectomized rats (Hermann, 1995; Ma Saito et al., 2000). Plasma AVP levels remained in the normal ranges in the glucocorticoid-deficient rats despite hypoosmolality, which should suppress plasma AVP to an undetectable level. There were no alterations in Kd and Bmax in 3HAVP receptor binding between the glucocorticoid-deficient and the control rats (Saito et al., 2000). Also, no change in the expression of AVP V2 receptor mRNA was found. Therefore, no downregulation of AVP receptor binding was found in glucocorticoid deficiency under this condition of nonsuppressible AVP release. Intrarenal factors could also participate in the impairment of water excretion in glucocorticoid deficiency; this is discussed later. The replacement of hydrocortisone in glucocorticoid-deficient rats normalized renal water excretion, plasma AVP levels and serum sodium concentration. Thus, glucocorticoid deficiency per se appears to be the primary factor for the exaggerated release of AVP from posterior pituitary gland with hypopituitarism. Liver cirrhosis. The pathogenesis of the nonosmotic stimulation of AVP is similar in oedematous disorders. Anderson et al. (1976) demonstrated that removal of the endogenous source of AVP by acute hypophysectomy improved renal water excretion in glucocorticoid-replaced dogs with acute portal vein constriction. However, a residual factor in the impaired water excretion persisted, thus implicating intrarenal factors. It has also been shown that plasma AVP levels are elevated in cirrhotic rats and humans, and AVP secretion was not sufficiently suppressed by an acute water load (Linas et al., 1981; Bichet et al., 1982a, 1983; Reznick et al., 1983; Tsuboi et al., 1994). Additionally, the expression of AVP mRNA in the hypothalamus was significantly increased in cirrhotic rats (Kim et al., 1993). Bichet et al. (1982a) further studied the mechanisms of impaired water excretion in patients with liver cirrhosis. Decompensated cirrhotic patients exhibited hyponatraemia, ascites and peripheral oedema. There was a significant increase in basal concentration of plasma AVP in the decompensated cirrhotic patients compared to compensated cirrhotic patients without ascites. An acute water load maximally suppressed plasma AVP to below 0·5 pmol/l in the compensated cirrhotic patients, whereas the decompensated cirrhotic patients did not suppress plasma AVP and their hyponatraemia persisted. Furthermore, activation of the renin–angiotensin–aldosterone system and the sympathetic nervous system was present in the cirrhotic patients, and the degree of activation correlated directly with the degree of water and sodium retention (Bichet et al., 1982b; Arroyo et al., 1983). Several mechanisms for the exaggerated secretion of AVP and other neurohumoral hormones in patients with liver cirrhosis have been proposed (Anderson et al., 1976; Linas et al., 1981; Bichet et al., 1982a,b, 1983; Arroyo et al., 1983; Reznick et al., 1983; Schrier, 1988b; Kim et al., 1993; Tsuboi et al., 1994). It is evident that V2 receptor antagonists reversed the impairment in renal water excretion in the experimental model of cirrhotic rats (Fig. 3) (Claria et al., 1989; Tsuboi et al., 1994). Greater activation of the sympathetic nervous system, the renin–angiotensin–aldosterone system and the nonosmotic release of AVP in decompensated compared with compensated cirrhotic patients suggests that as cirrhosis progresses there is more evidence of arterial underfilling with neurohumoral stimulation. Effect of the nonpeptide AVP antagonist OPC-31260 on renal water excretion after an acute water load (30 ml/kg body weight) in rats with CCl 4 -induced liver cirrhosis (LC). (a) Percent excretion of water load. (b) Minimal Uosm. (Data from Tsuboi et al., 1994.) Peripheral arterial vasodilatation, primarily in the splanchinic vascular bed, has been proposed to account for the neurohumoral stimulation and the initiation of water and sodium retention in patients with cirrhosis (Gines et al., 1996). Nitric oxide contributes to the vasodilatation, in addition to the anatomical shunting (Martin et al., 1998). The peripheral arterial vasodilatation in cirrhosis produces renal sodium and water retention, and leads to an increase in total blood volume. However, this volume expansion is not sufficient to refill the enlarged arterial vascular compartment. A reduction in 'effective arterial blood is closely related to the elevation of plasma AVP, and (Schrier, 1988a,b). The arterial underfilling that occurs as a of the arterial vasodilatation plays a crucial role in the secretion of these hormones in cirrhotic patients (Bichet et al., 1983). Nitric oxide appears to be a of the vasodilatation in cirrhosis, and an inhibition of oxide for 7 days the circulation, and plasma AVP, and levels and renal water and sodium retention in experimental cirrhosis (Martin et al., 1998). Nonsuppressible AVP release in cirrhosis water reabsorption in collecting duct cells, and results in impaired water excretion. In the circulation, elevated plasma AVP might to the of arterial pressure by stimulation of the AVP receptor (Claria et al., 1991). water induces an increase in central blood volume et al., and this manoeuvre has been in patients with liver cirrhosis to the and renal water excretion were (Bichet et al., 1983). water suppressed plasma AVP and in the decompensated cirrhotic patients with ascites, in an increase in renal sodium and water excretion. showed that water increased cardiac output, atrial pressure and and decreased vascular et al., 1986). It should be however, that water but did not renal water and sodium excretion in cirrhotic patients with ascites. There are at least two for the in renal sodium and water excretion. any that might be by increased pressure not be to to a by water et al., 1985; et al., 1984; Bichet et al., Ishikawa et al., 1986). nonosmotic release of AVP was closely associated with increased expression of AVP mRNA in the hypothalamus of rats with congestive heart failure to (Kim et al., 1990). The administration of AVP V2 receptor antagonists the of AVP in the impairment in renal water excretion in rats with heart failure et al., et al., 1994). The peptide V2 receptor antagonist reversed the in water excretion in the model of cardiac failure to of the et al., 1986). nonpeptide antagonists of the AVP V2 receptor Xu et al. and et al. have shown that these nonpeptide antagonists improved water excretion in heart failure in rats with of and humans with congestive heart failure, Bichet et al. the effect of the and the in the in water retention in patients with and congestive heart The administration of these cardiac and increased cardiac output, improved renal water excretion and suppressed plasma AVP in response to an acute water load. The that a decrease in volume and cardiac output, as sensed by arterial appears to be the primary for the nonosmotic release of AVP in cardiac failure, an effect that can be reversed by cardiac In impaired water excretion occurs in patients with SIADH, adrenal insufficiency, liver cirrhosis with ascites and congestive heart These patients are into euvolaemic or hypervolaemic Nonsuppressible release of AVP is found despite hypoosmolality, which should suppress AVP release to undetectable levels. The exaggerated release of AVP is stimulated by a decrease in 'effective circulatory blood in liver cirrhosis with ascites and congestive heart Reversal of hyponatraemia by the AVP V2 receptor antagonists provides conclusive evidence for the role of AVP in pathological states of impaired water excretion, and a use of the AVP receptor antagonists per AVP on the of renal collecting duct cells are to the to the activation of (Ishikawa, 1993; Knepper Sasaki et al., the of and a chain This is in the of the major intrinsic et al., 1984; Nielsen et al., 1995; et al., et al., 1995; et al., 1995). An reduction in plasma AVP concentrations or the AVP V2 receptor antagonist causes of AQP-2 into the et al., Saito et al., et al., 1998). This observation AQP-2 as the AVP-regulated water channel and