Osmoregulation and osmoreceptors.
The purpose of this essay is to review the nature of osmoreceptors and their role in the maintenance of body fluid osmolality in the light of some new information.
Water makes up about 60% of the mammalian body weight and the osmotic concentration of solutes dissolved in every body water compartment is a closely regulated invariant at about 0.3 osmol per kg (1 osmol = 6.02 x 1023 discrete particles in solution). The maintenance of body fluid osmolality is related also to the volume and composition of various fluid compartments. These parameters may sometimes make conflicting demands on the regulatory mechanisms; usually the maintenance of osmolality has priority over constancy of volume or composition. 
The following scheme summarises the general nature of the control systems which regulate body fluid osmolality, volume and composition.
The chief mechanisms by which osmoregulation is achieved are: (1) behavioural viz. thirst and salt-appetite and (2) hormonal viz. the antidiuretic hormone (arginine-vasopressin, AVP) and the antinatriuretic hormone (aldosterone). It may be noted that these two hormones conserve water and salt respectively. Homer Smith had postulated also an anti-natriuretic hormone of intracranial origin. Somehow Smith's concept was subsequently twisted by some enthusiastic proponents of the "natriuretic" hormone who quoted him freely without the prefix "anti". Despite several attempts and conflicting claims, a natriuretic hormone has not yet been shown to exist, though sometimes the antidiuretic hormone (ADH) may be natriuretic.
The behavioural regulatory mechanisms are outside the scope of this essay. The aldosterone system is touched briefly. The secretion of the glomerular adrenal cortex is regulated mainly by the renin-angiotensin system and also to some extent by plasma K+ and Na+, corticotrophin and perhaps other neural and hormonal mechanisms. There is a considerable interaction between the renin-angiotensin and the ADH systems; angiotensin releases ADH while ADH is known to suppress renin release., ,  The physiological significance of these interactions remains to be established.
ADH is produced in the neurosecretory cells of the supra-optic nucleus and released from their axon in the neurohypophysis (posterior lobe of the pituitary). The secretion of ADH is regulated by (1) non-osmotic stimuli through afferents from systemic pressure and volume receptors and (2) osmotic stimuli which act directly on the hypothalamus. , , ,,  Fall in blood pressure or blood volume triggers ADH release through neural pathways. Changes in blood pressure or volume also seem to modulate the activity of the ADH system, changing its threshold or sensitivity to osmotic stimuli. The subject of nonosmotic control over the ADH system has been reviewed in detail in the above papers.
Osmotic control over ADH' release is exercised through "osmoreceptors" defined by Verney as "autonomic receptive elements with which the neurohypophysis is functionally linked, and through whose activation the pituitary antidiuretic substance is released". Verney suggested that the total osmotic pressure (irrespective of the ionic composition) of the blood passing through the cerebral branches of the internal carotid, artery provided the activating signal. " Later Jewell and Verney localised the site of the osmoreceptors in the anterior hypothalamus. Verney's''concept remains generally valid even today though his suggestion about the nature of the actuating signal has been questioned.
Swedish workers have performed many experiments in conscious goats, administering hyperosmolal solutions of various substances from both inside and outside the blood-brain barrier and noting the effects on ADH release. According to Andersson "the integrated results of these studies imply that. osmoreceptors are ... juxta-ventricular sodium receptors". This claim has not been received favourably by other workers in the field.,  Resolution of this controversy regarding the signal which triggers osmoreceptors viz. plasma osmotic pressure or plasma sodium was necessary for its own sake as well as for its implications in physiology and pathology. Experiments designed to resolve the controversy were undertaken therefore in our laboratory. The following is a summary of chose experiments and their most plausible interpretation. All the work is cited directly from the thesis by Mr. Sp. Swaminathan., 
The experimental animal was the rhesus monkey because it is the available species closest to man. All experiments were performed on healthy animals which had recovered fully from aseptic surgery for implantation of cannulae etc. Three types of surgical procedures were carried out in different monkeys: (1) implantation of a stainless cannula into the antero-ventral part of the third ventricle by stereotaxic surgery [Fig la], [Fig lb], [Fig 1c] & [Fig ld]; (2) similar implantation of a cannula in the hypothalamus just anterior to the antero-ventral wall of the third ventricle (3) implantation of a silicone rubber cannula into the common carotid artery to infuse solutions into the internal carotid artery via the free end of the tube, which was kept sealed inside a cap implanted on the cranium [Fig 2a] & [Fig 2b]. The positions of the various cannulae were confirmed post mortem after the injection of a dye.
In keeping with the injunctions of Homer Smith, the experiments were performed on pre-operated and trained monkeys in the conscious state to eliminate the spurious effects of anesthesia and physical or psychological trauma. Diet and hydration were standardized and free water clearance was computed by measuring urine volume, urine osmolality and plasma osmolality. The ADH released by various procedures was computed through a very precise bioassay (index of precision 0.1) in the same animals by measuring their graded responses to varying doses of exogenous AVP. Hypertonic solutions of sodium salts and sugars, all iso-osmolar in any given monkey, were administered according to a fixed protocol while the monkeys were undergoing water diuresis (with a constant i.v. infusion of isotonic dextrose). The quantities of ADH released by the various hypertonic solutions administered by the 3 routes are shown in [Table 1].
It is evident from the condensed results shown in [Table 1] that: (1) Via the third ventricle CSF, Na salts triggered ADH release, the sugars by contrast were rather ineffective. Furthermore, in 4 monkeys the dose effect relationship between log dose Na infused into the CSF and ADH release was determined; it showed linear relationship over 4 doses (10 to 100 µM) of total Na. (2) Via the carotid artery, all hypertonic solutions released ADH. However, Na salts were significantly more effective than the sugars. The most economical unitary hypothesis which explains these findings is: the osmoreceptors of Verney are Na sensitive receptors on dendrites known to occur on the ependymal lining of the anterior third ventricle. The dendrites may be functional units by themselves or in conjunction with specialised ependymal cells., ,  It follows from the hypothesis that through the CSF, only Na salts should trigger the osmoreceptors, while through blood and to some extent the external surface of the ependyma, hypertonic solutions of all substances relatively impermeable through the blood brain barrier should stimulate by drawing out water, thus raising (Na+) in the immediate vicinity of the osmoreceptors.
Certain other facts are also explained by the same hypothesis. (1) The difference in response to hyperosmolar solutions of sugars and NaCl injected into the carotid artery: while NaCl and the sugars are both only slightly permeable through the blood-brain barrier, the small amount of Na which does permeate into the CSF may continue to activate the osmoreceptors, the sugars being ineffective; thus intracarotid injection of hypertonic NaCl released significantly more ADH than comparable solutions of mannitol or sucrose. (2) The difference in the relative responses to sodium acetate via the third ventricle and the carotid artery: Third ventricle infusions of sodium acetate and chloride were both far more effective than the sugars in releasing ADH. The acetate effect seemed greater than the chloride effect but the difference was statistically not significant. The important finding was that Na + and not Cl- provided the relevant signal. On the other hand, via the carotid artery, the acetate was less effective than the chloride. This findings was consistent with the vasodilator effect of even small amounts of sodium acetate.,  The vasodilation could be expected to dilute the signal reaching the osmoreceptors. (3) The precisely graded response to NaCl: the linear relationship between log (Na) infused into the CSF and ADH release suggested a physiologically relevant mechanism. (4) In physiological and at least some pathological conditions, changes in plasma osmolality coincide with corresponding changes in plasma Na (the most abundant plasmacation) and therefore it seems quite natural that the osmoreceptors are in fact Na sensitive receptors.
The findings of Swaminathan do not detract much from Verney's original concept of the osmoreceptor and in any case do not warrant a change in terminology. The ependymal Na sensitive receptors function as osmoreceptors by virtue of their anatomical relations. The question remains as to whether the dendritic nerve endings which constitute osmoreceptors belong to the neurosecretory cell itself or to a "sensory" neurone which sends efferent impulses to the neurosecretory cell .. The neurosecretory cell which secretes ADH belongs to the large family of endocrine cells of neuroectodermal origin and known as "APUD" cells.,  Other APUD endocrine cells respond directly to the controlled parameter, e.g. those regulating plasma Ca or blood glucose respond to the circulating level of Ca or glucose respectively. It is also known that at least in lower vertebrates the pre-optic magnocellular neuroendocrine cells have dendrites reaching into the ependyma of the third ventricle.,  Therefore it is likely that neurosecretory cells of the pre-optic or supraoptic nucleus have dendrites ending in the specialised ependyma of the anteroventral part of the third ventricle and that these dendrites respond directly to (Na + )in the CSF bathing them. Synapses on the body of the neurosecretory cell may of course determine responses to non-osmotic neural inputs. A combination of modern histological techniques applied to the same sample, e.g. intracellular horseradish peroxidase or procion yellow staining to trace nerve connections together with immunocytochemistry to identify neurosecretory granules should help further investigation of the above hypothesis.
Some human patients of diabetes insipidus show no osmoreceptor-thirst response to hypertonicity but show normal ADH release in response to non-osmotic stimuli. It has been claimed that such a dissociation between osmotic and nonosmotic stimuli establishes the existence of two types of synapses on the neurosecretory cell viz. (1) with efferents from pressure-volume receptor pathways and (2) with efferents from the osmoreceptor cells wherein the cause of such cases of diabetes insipidus lies.,  However, it is equally possible that the defect in osmoregulation is due to a functional anomaly in the neurosecretory cell dendrite in the ependyma, postulated above as the osmoreceptor. Hypersensitivity and/or low threshold of the same receptors may be responsible for at least some cases of the syndrome of inappropriate secretion of ADH. The proposed hypothesis about the nature of osmoreceptors is summarised diagramatically in [Fig 3]
Attention must be drawn to two old papers which had postulated on morphological grounds, an osmoreceptor function for the neurosecretory cell dendrites in the third ventricle of the frog, but have been ignored in recent reviews (7, 15): (i) Dierick, K.: The dendrites of the preoptic neurosecretory nucleus of the Rana temporaria and the osmoreceptors. Arch. Int. Pharmacodyn. Ther. 140: 708-725, 1962. (ii) Smoller, C. G.. Neurosecretory processes extending into the third ventricle: secretory or sensory Science, 147: 882-884, 1965.
The author is indebted to Mr. Sp. Swaminathan for permission to reproduce material from his Ph.D. thesis. His work with all technical details will be published elsewhere in due course. Ms. K. Gopalakrishnan and S. Mazumdar provided valuable assistance in preparation of the manuscript.