The regulation of ADH release from the posterior pituitary is dependent primarily on two mechanisms involving the osmotic and nonosmotic pathways (Fig. 71.4).¹³ Although magnocellular neurons are themselves osmosensitive, they require input from the lamina terminalis to respond fully to osmotic challenges (Fig. 71.5). Neurons in the lamina terminalis are also osmosensitive and because the subfornical organ (SFO) and the organum vasculosum of the lamina terminalis (OVLT) lie outside the blood-brain barrier, they can integrate this information with endocrine signals borne of circulating hormones, such as angiotensin II (AT-II), relaxin, and atrial natriuretic peptide (ANP). Although circulating AT-II and relaxin excite both oxytocin and vasopressin magnocellular neurons, ANP inhibits vasopressin neurons.¹⁴

The nonosmotic pathways are more physiologically described now as “osmoregulatory gain.”¹⁵ A stable and approximately linear relation is normally observed between vasopressin concentration and plasma osmolality under resting conditions (Fig. 71.4A).¹⁶ The slope of this relation reflects the overall sensitivity of this homeostatic mechanism and the term “osmoregulatory gain” refers to this parameter. Osmoregulatory gain is increased during hypovolemia to help maintain arterial pressure and restore blood volume.¹⁷ Conversely, osmoregulatory gain is attenuated during hypervolemia to promote homeostasis by favoring diuresis (Fig. 71.6). The enhancement of osmosensory gain is mediated by circulating AT-II: neurons in the SFO (Fig. 71.5) contain AT-II and the release of this peptide during hypovolemia or hypotension can lead to an increase in the osmotic activation of magnocellular cells producing vasopressin. Good et al.¹⁸ and Zhang and Bourque¹⁹ found that AT-II enhances osmosensory gain by amplifying mechanosensory transduction, an F actin-dependent phenomenon probably part of a scaffolding complex.

In addition to an angiotensinergic path from the SFO, the OVLT and the median preoptic nucleus provide direct glutaminergic and GABAergic projections to the hypothalamoneurohypophyseal system. Nitric oxide may also modulate neurohormone release.³ The neuropeptide apelin is colocalized with AVP in supraoptic nucleus (SON) magnocellular neurons, and physiologic experiments indicate that AVP and apelin are conversely regulated to facilitate systemic AVP release and to suppress antidiuresis.²⁰

The cellular basis for osmoreceptor potentials has been characterized using patch-clamp recordings and morphometric analysis in magnocellular cells isolated from the supraoptic nucleus of the adult rat. In these cells, stretchinactivating cationic channels transduce osmotically evoked changes in cell volume into functionally relevant changes in membrane potential. In addition, magnocellular neurons also operate as intrinsic Na⁺ detectors. The N-terminal variant of the TRPV1 is an osmotically activated channel expressed in the magnocellular cells producing vasopressin²¹ and in the circumventricular organs, the OVLT, and the SFO.²² Because osmoregulation still operates in Trpv1^–/^– mice, other osmosensitive neurons or pathways must compensate for the loss of central osmoreceptor function.²¹,²²,²³ Afferent neurons expressing the osmotically activated ion channel TRPV4 in the thoracic dorsal root ganglia that innervate hepatic blood
vessels and detect physiologic hypo-osmotic shifts in blood osmolality have recently been identified.²

In mice lacking the osmotically activated ion channel TRPV4, hepatic sensory neurons no longer exhibit osmosensitive inward currents and the activation of peripheral osmoreceptors in vivo is abolished. In a large cohort of human liver transplantees, who presumably have denervated livers, plasma osmolality is significantly elevated compared to healthy controls, suggesting the presence of an inhibitory vasopressin effect of hyponatremia perceived in the portal vein from hepatic afferents.² TRPV1 (expressed in central neurons) and TRPV4 (expressed in peripheral neurons) thus appear to play entirely complementary roles in osmoreception. McHugh et al.²⁴ have therefore identified the primary afferent neurons that constitute the afferent arc of a well-characterized reflex in man, which was recently identified in rodents. This reflex engages the sympathetic nervous system to raise blood pressure and stimulate metabolism.²⁵,²⁶ Of clinical interest, it has already been demonstrated that orthostatic hypotension and postprandial hypotension respond to water drinking.²⁷,²⁸,²⁹ Moreover, water drinking in man can prevent neutrally mediated syncope during blood donation or after prolonged standing.³⁰ Finally, water drinking is also associated with weight loss in overweight individuals.³¹ Other peripheral sensory neurons expressing other mechanosensitive proteins may also be involved in osmosensitivity.³²

The osmotic stimulation of AVP release by dehydration, hypertonic saline infusion, or both, is regularly used to determine the vasopressin secretory capacity of the posterior pituitary. This secretory capacity can be assessed directly by comparing the plasma AVP concentrations measured sequentially during the dehydration procedure with the normal values³³ and then correlating the plasma AVP values with the urine osmolality measurements obtained simultaneously (Fig. 71.7).

The AVP release can also be assessed indirectly by measuring plasma and urine osmolalities at regular intervals during the dehydration test.³⁴ The maximal urine osmolality obtained during dehydration is compared with the maximal urine osmolality obtained after the administration of vasopressin (Pitressin, 5 U subcutaneously in adults, 1 U subcutaneously in children) or 1-desamino-8-D-arginine vasopressin (desmopressin [dDAVP], 1 to 4 µg s.c. or intravenously [IV], over 5 to 10 minutes).

The nonosmotic stimulation of AVP release can be used to assess the vasopressin secretory capacity of the posterior pituitary in a rare group of patients with the essential hyponatremia and hypodipsia syndrome.³⁵ Although some of these patients may have partial central diabetes insipidus, they respond normally to nonosmolar AVP release signals such as hypotension, emesis, and hypoglycemia.³⁵ In all other cases of suspected central diabetes insipidus, these nonosmotic stimulation tests will not give additional clinical information.³⁶

Angiotensin is a well-known dipsogen and has been shown to increase thirst in all the species tested.³⁷ However, knockout models for angiotensinogen³⁸ or for angiotensin-1A (AT-IA) receptor³⁹,⁴⁰ did not alter thirst or water balance. Disruption of the AT-II receptor induced only mild abnormalities of thirst postdehydration.⁴¹,⁴² However, as described earlier, AT-II enhances osmosensory gain. Earlier reports suggested that the intravenous administration of atrial peptides inhibits the release of vasopressin,⁴³ but this was not confirmed by Goetz et al.⁴⁴ Furthermore, Ogawa et al.⁴⁵ found no evidence that ANP, administered centrally or peripherally, was important in the physiologic regulation of plasma AVP release in conscious rats. A very rapid and robust release of AVP is seen in humans after a cholecystokinin (CCK) injection.⁴⁶ Nitric oxide is an inhibitory modulator of the hypothalamoneurohypophyseal system in response to osmotic stimuli.⁴⁷,⁴⁸,⁴⁹,⁵⁰ Vasopressin secretion is under the influence of a glucocorticoid-negative feedback system,⁵¹ and the vasopressin responses to a variety of stimuli (hemorrhage, hypoxia, hypertonic saline) in normal humans and animals appear to be attenuated or eliminated by pretreatment with glucocorticoids. Finally, nausea and emesis are potent stimuli of AVP release in humans and seem to involve dopaminergic neurotransmission.⁵²

The neurohypophyseal hormone AVP has multiple actions, including the inhibition of diuresis, the contraction of smooth muscle, the aggregation of platelets, the stimulation of liver glycogenolysis, the modulation of adrenocorticotropic hormone release from the pituitary, and the central regulation of somatic functions (thermoregulation and blood pressure) and the modulation of social and reproductive behavior.⁵³ These multiple actions of AVP can be explained by the interaction of AVP with at least three types of G protein-coupled receptors: the V_1a (vascular, hepatic, and brain) and V_1b (anterior pituitary) receptors act through phosphatidylinositol hydrolysis to mobilize calcium, and the V₂ (kidney) receptor is coupled to adenylate cyclase.⁵⁴,⁵⁵,⁵⁶

The transfer of water across the principal cells of the collecting ducts is now known at such a detailed level that billions of molecules of water traversing the membrane can be represented; see useful teaching tools at http://www.mpibpc.gwdg.de/abteilungen/073/gallery.html and http:// www.ks.uiuc.edu/research/aquaporins. The 2003 Nobel Prize in Chemistry was awarded to Peter Agre and Roderick MacKinnon, who solved two complementary problems presented by the cell membrane: How does a cell let one type of ion through the lipid membrane to the exclusion of other ions? And, how does it permeate water without ions? See Figure 71.8. This contributed to a momentum and renewed interest in basic discoveries related to the transport of water and indirectly to diabetes insipidus.⁵⁷,⁵⁸ The first step in the action of AVP on water excretion is its binding to arginine vasopressin type 2 receptors (hereafter referred to as V₂ receptors) on the basolateral membrane of the collecting duct cells (Fig. 71.9). The human AVPR2 gene that codes for the V₂ receptor is located in chromosome region Xq28 and has three exons and two small introns.⁵⁹,⁶⁰ The sequence of the cDNA predicts a polypeptide of 371 amino acids with seven transmembrane, four extracellular, and four cytoplasmic domains (Fig. 71.10). The activation of the V₂ receptor on renal collecting tubules stimulates adenylyl cyclase via the stimulatory G protein (Gs) and
promotes the cyclic adenosine monophosphate (cAMP)-mediated incorporation of water channels into the luminal surface of these cells. There are two ubiquitously expressed intracellular cAMP receptors: (1) the classical protein kinase A (PKA) that is a cAMP-dependent protein kinase, and (2) the recently discovered exchange protein directly activated by cAMP that is a cAMP-regulated guanine nucleotide exchange factor. Both of these receptors contain an evolutionally conserved cAMPbinding domain that acts as a molecular switch for sensing intracellular cAMP levels to control diverse biologic functions.⁶¹ Several proteins participating in the control of cAMP-dependent aquaporin-2 AQP2 trafficking have been identifieD (e.g., A-kinase anchoring proteins tethering PKA to cellular compartments; phosphodiesterases regulating the local cAMP
level; cytoskeletal components such as F-actin and microtubules; small GTPases of the Rho family controlling cytoskeletal dynamics; motor proteins transporting AQP2-bearing vesicles to and from the plasma membrane for exocytic insertion and endocytic retrieval; soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptors (SNAREs) inducing membrane fusions, hsc70, a chaperone important for endocytic retrieval). These processes are the molecular basis of the vasopressin-induced increase in the osmotic water permeability of the apical membrane of the collecting tubule.⁶²,⁶³,⁶⁴

AVP also increases the water reabsorptive capacity of the kidney by regulating the urea transporter UT-A1 that is present in the inner medullary collecting duct, predominantly in its terminal part.⁶⁵,⁶⁶ AVP also increases the permeability of principal collecting duct cells to sodium.⁶⁷ Finally, the vasopressin V₂ receptor has been located in the primary cilium and Bardet-Biedl syndrome-derived unciliated renal epithelial cells were unable to respond to luminal AVP and to activate luminal AQP2.⁶⁸

In summary, in the absence of AVP stimulation, collecting duct epithelia exhibit very low permeabilities to sodium urea and water. These specialized permeability properties permit the excretion of large volumes of hypotonic urine formed during intervals of water diuresis. By contrast, AVP stimulation of the principal cells of the collecting ducts leads to selective increases in the permeability of the apical membrane to water, urea, and sodium.

These actions of vasopressin in the distal nephron are possibly modulated by prostaglandin E2 (PGE2), nitric oxide,⁶⁹ and by luminal calcium concentration. High levels of E-prostanoid-3 receptors are expressed in the kidney.⁷⁰ However, mice lacking E-prostanoid-3 receptors for PGE2 were found to have quasinormal regulation of urine volume and osmolality in response to various physiologic stimuli.⁷⁰ PGE2 is synthesized and released in the collecting duct, which expresses all E-prostanoid receptors.⁷¹,⁷²,⁷³ An apical calcium/polycation receptor protein expressed in the terminal portion of the inner medullary collecting duct of the rat
has been shown to reduce AVP-elicited osmotic water permeability when luminal calcium concentration rises.⁷⁴ This possible link between calcium and water metabolism may play a role in the pathogenesis of renal stone formation.⁷⁴

A useful strategy to establish the physiologic function of a protein is to determine the phenotype produced by pharmacologic inhibition of protein function or by gene disruption. Transgenic knockout mice deficient in AQP1, AQP2, AQP3, AQP4, AQP3, and AQP4; CLCNK1; NKCC2; AVPR2; AGT; or adenylyl cyclase 6 (AC6) have been engineered.⁷⁵,⁷⁶,⁷⁷,⁷⁸,⁷⁹,⁸⁰,⁸¹,⁸²,⁸³,⁸⁴ Angiotensinogen (AGT)-deficient mice are characterized by both concentrating and diluting defects secondary to a defective renal papillary architecture.⁷⁶

As reviewed by Rao and Verkman,⁸⁵ the extrapolation of data in mice to humans must be made with caution. For example, the maximum osmolality of mice (greater than 3,000 mOsmol per kilogram of H₂O) is much greater than that of human urine (1,000 mOsmol per kilogram of H₂O), and normal serum osmolality in mice is 330 to 345 mOsmol per kilogram of H₂O, substantially greater than that in humans (280 to 290 mOsmol per kilogram of H₂O). These differences are related to renal anatomy and metabolic rate: (1) the mouse kidney exhibits much larger loops of Henle in the inner medulla and papillae and (2) because of the large difference in body size (30 g to 70 kg) the mouse metabolic rate (and thus the food intake) per gram of body mass is 20 times greater than that of humans. The amount of osmoles that need to be excreted by the kidney is also disproportionately larger when expressed per gram of body mass or kidney mass. These two features account for the much higher urine concentrating ability of the mouse as compared to humans. Protein expression patterns, and thus the interpretation of phenotype studies, may also be species dependent. For example, AQP4 is expressed in both the proximal tubule and the collecting duct in mice but only in the collecting duct in rats and humans.⁸⁵

The Aqp3, Aqp4, Clcnk-1, and Agt knockout mice have no identified human counterparts. Of interest, AQP1-null individuals have no obvious symptoms.⁸⁶ Yang et al.⁸⁷ have generated an AQP2-T126M conditional knockin model of nephrogenic diabetes insipidus (NDI), to recapitulate the clinical features of the naturally occurring human AQP2 mutation T126M.⁸⁸ The conditional knockin adult mice showed polyuria, urinary hypo-osmolality, and endoplasmic reticulum (ER) retention of AQP2-T126M in the collecting duct. The screening of candidate protein folding correctors in AQP2-T126M-transfected kidney cells showed increased AQP2-T126M plasma membrane expression with the Hsp90 inhibitor 17-allylamino-17-demethoxygeldanamycin (17-AAG), a compound currently in clinical trials for tumor therapy. 17-AAG increased urine osmolality in the AQP2-T126M mice (without effect in AQP2 null mice) and partially rescued defective AQP2-T126M cellular processing. These proof-of-concept findings suggest the possibility of using existing drugs for therapy of some forms of NDI. Ethylnitrosourea-mutagenized mice heterozygous for the F204V mutation in the Aqp2 gene have been described.⁸⁹ The homozygous mice are viable because of a moderate phenotype with a possibility to concentrate urine from 161 to 470 m Osmol per kilogram of H₂O in response to dDAVP. Cell biology experiments performed on renal tissue from Aqp2F^204V/+ animals suggest that the mutant protein is being rescued by the wild-type protein.

Mice lacking the AVPR2 receptor failed to thrive and died within the first week after birth due to hypernatremic dehydration.⁷⁸ Li et al.⁹⁰ generated mice in which the Avpr2 gene could be conditionally deleted during adulthood by the administration of 4-OH-tamoxifen. Adult mice displayed all characteristic symptoms of X-linked NDI (XNDI), including polyuria, polydipsia, and resistance to the antidiuretic actions of vasopressin. Gene expression analysis suggested that the activation of renal EP4 PGE₂ receptors might compensate for the lack of renal V2R activity in X-linked NDI mice, and both acute and chronic treatment of the mutant mice with a selective EP4 receptor agonist greatly reduced all major manifestations of XNDI. This beneficial effect is likely secondary to the intracellular generation of cAMP at the principal cell level by EP4 PGE₂ receptors.

The absence of the gene coding for the NaK2Cl cotransport (NKCC2) in the luminal membrane of the thick ascending loop of Henle in the mouse also caused polyuria that was not compensated elsewhere in the nephron and recapitulated many features of the human classical Bartter syndrome.⁷⁹ The absence of transcellular NaCl transport via NKCC2 probably abolished the lumen-positive transepithelial voltage that enables paracellular reabsorption of Na and K across the wall of the thick ascending tubule. The combined absence of transcellular and paracellular transport of salt across the thick ascending limb cells prevents the establishment of the normal osmotic gradient necessary for urine concentration.

The animal model of diabetes insipidus that has been most extensively studied is the Brattleboro rat. Discovered in 1961, the rat lacks vasopressin and its neurophysin, whereas the synthesis of the structurally related hormone oxytocin is not affected by the mutation.⁹¹ Its inability to synthesize vasopressin is inherited as an autosomal recessive trait. Schmale and Richter⁹² isolated and sequenced the vasopressin gene from homozygous Brattleboro rats, and found that the defect is due to a single nucleotide deletion of a G residue within the second exon encoding the carrier protein neurophysin (Fig. 71.11). The shift in the reading frame caused by this deletion predicts a precursor with an
entirely different C terminus. The messenger RNA (mRNA) produced by the mutated gene encodes a normal AVP but an abnormal NPII moiety,⁹² which impairs transport and processing of the AVP-NPII precursor and its retention in the endoplasmic reticulum of the magnocellular neurons where it is produced.⁹³,⁹⁴ Homozygous Brattleboro rats may still demonstrate some V2 (vide infra) antidiuretic effects since the administration of a selective nonpeptide V2 antagonist (SR 121463A, 10 mg per kilogram i.p.) induced a further increase in urine flow rate (200 to 354 ± 42 mL per 24 hours) and a decline in urinary osmolality (170 to 92 ± 8 mmol per kilogram).⁹⁵ This decline in urine osmolality following the administration of a nonpeptide V2r antagonist could also be secondary to the “inverse agonist” properties of SR121463A: the intrinsic activity, or tone, of the V2R would be deactivated by the SR121463A compound (for the inverse agonist properties of SR121463A (see reference 96). There is also an alternative explanation to this relatively high urine osmolality of 170 because, in Brattleboro rats, low levels of hormonally active AVP are produced from alternate forms of AVP preprohormone. Due to a process called molecular misreading, one transcript contains a 2-bp deletion downstream from the single nucleotide deletion that restores the reading frame and produces a variant AVP preprohormone that is smaller in length by one amino acid and differs from the normal product by only 13 amino acids in the neurophysin II moiety.⁹⁷ Oxytocin, which is present at enhanced plasma concentrations in Brattleboro rats, may be responsible for the antidiuretic activity observed.⁹⁸,⁹⁹ Oxytocin is not stimulated by increased plasma osmolality in humans.