JASN Express. Published on November 14, 2007 as doi: 10.1681/ASN.2007070825

SCIENCE IN RENAL MEDICINE www.jasn.org

How Does the Brain Sense Osmolality?

Joseph G. Verbalis

Professor of Medicine and Physiology, Georgetown University School of Medicine, Washington, DC

ABSTRACT For nearly 60 years, we have known that the brain plays a pivotal role in regulating sponses to hyperosmolality in experi- the osmolality of body fluids. Over this time period, scientists have determined the mental animals3 and in human subjects structure and function of arginine and its receptors, the role of the with brain damage that infarcts the re- posterior pituitary as a storage site, and the determinants of vasopressin release. gion around the OVLT, who typically are The cellular mechanisms by which the responds to vasopressin are also well unable to maintain normal plasma os- understood. One area that remains unclear is the neural mechanisms underlying molalities even under basal conditions.4 osmoreception. New findings have implicated the TRPV family of cation channels as In contrast to the effects of such lesions osmo-mechanoreceptors that may mediate the neuronal responses to changes in to eliminate both osmotically stimulated systemic tonicity. This topic is reviewed here. and AVP secretion, diabetes insip- idus caused by destruction of the magno- J Am Soc Nephrol 18: 3056–3059, 2007. doi: 10.1681/ASN.2007070825 cellular AVP neurons in the supraoptic (SON) and paraventricular (PVN) nu- clei eliminates dehydration-induced Body fluid is directed at WHERE ARE OSMORECEPTORS AVP secretion but not thirst, clearly in- maintaining the stability of the osmo- LOCATED? dicating that osmotically stimulated lality of body fluids (osmotic ho- thirst must be generated proximally to meostasis) and the intravascular blood The pioneering investigations of Verney the AVP-secreting cells themselves (Fig- volume (volume homeostasis). Os- in the 1940s1 found infusion of hyperos- ure 1A). Other regions have also been re- motic regulation serves to minimize motic solutions into blood vessels that ported to contain putative osmorecep- osmotically induced perturbations in perfused the anterior pro- tors, including the hepatic portal cell volume, which has adverse effects duced an antidiuresis in dogs, thereby circulation, leading to the suggestion on multiple cellular functions. Body identifying this area as the site of osmo- that osmoreceptors are widely distribut- fluid osmolality in humans is main- responsive elements in the brain. The ed.5 However, cells in these areas likely tained between 280 and 295 mOsm/kg most parsimonious explanation for these act to modulate the activity of the pri-

H2O, representing one of the most findings would be that the AVP-secret- mary OVLT osmoreceptors because they highly regulated parameters of body ing magnocellular neurons themselves are not able to maintain osmotically physiology. This is accomplished are the osmoreceptors. Although these stimulated AVP secretion or thirst after through an integration of thirst, argi- neurons do display osmoreceptive char- lesions of the OVLT. nine vasopressin (AVP) secretion, and acteristics,2 their location inside the Involvement of the OVLT and sur- renal responsiveness to AVP. To pre- blood–brain barrier does not position rounding areas of the anterior hypothal- serve plasma osmolality within such them to respond quickly to small amus in osmoreception is also supported narrow tolerances, pituitary AVP se- changes in osmolality in the circulation. by studies using immunohistochemical cretion must vary in response to small Subsequent studies strongly implicated changes in plasma osmolality, which is the circumventricular organ named the Published online ahead of print. Publication date achieved through the activation or in- organum vasculosum of the lamina ter- available at www.jasn.org. hibition of central osmoreceptor cells. minalis (OVLT), which lacks a blood– Correspondence: Dr. Joseph G. Verbalis, Division Understanding where and how the brain barrier, as well as areas of the adja- of Endocrinology and Metabolism, 232 Building D, brain senses the osmolality of body flu- cent hypothalamus near the anterior wall Georgetown University Medical Center, 3800 Res- ids and transduces this information of the third cerebral ventricle as the site ervoir Road NW, Washington, DC 20007. Phone: 202-687-2818; Fax: 202-444-7797; E-mail: verbalis@ into mechanisms that regulate AVP se- of the principle brain osmoreceptors. georgetown.edu cretion and thirst is the subject of this Destruction of this area of the brain abol- Copyright © 2007 by the American Society of commentary. ishes both AVP secretion and thirst re- Nephrology

3056 ISSN : 1046-6673/1812-3056 J Am Soc Nephrol 18: 3056–3059, 2007 www.jasn.org SCIENCE IN RENAL MEDICINE

secretion and thirst, this has not been de- 10 A + Na THIRST finitively confirmed. Separate but paral- 9 sodium chloride Na+ Primary osmo- lel pathways for these complementary + receptors 8 Na functions remain possible (Figure 1B) mannitol Na+ AVP and could account for the lower osmotic 7 threshold for activation of AVP secretion 6 B Primary Na+ 8 thirst compared with thirst. 5 osmo- THIRST + receptor urea Na (pg/mL) 4 + Na Primary

AVP Plasma Vasopressin 3 Na+ osmo- AVP WHAT DO BRAIN receptor OSMORECEPTORS RESPOND TO? 2

Figure 1. Brain osmoreceptor pathways. 1 glucose The primary brain osmoreceptors lie out- Neither AVP secretion nor thirst is 0 side the blood–brain barrier in the OVLT. equally sensitive to all plasma solutes. So- 285 295 305 315 Different neural projections connect the dium and its anions, which normally Plasma Osmolality primary osmoreceptors to brain areas re- contribute Ͼ95% of the osmotic pres- (mOsm/kg H2O) sponsible for AVP secretion and thirst. sure of plasma, are the most potent sol- Figure 2. Solute specificity of brain osmo- Whether the same (A) or different subsets utes in terms of their capacity to stimu- receptors. The lines represent the relation- (B) of osmoreceptors project to both areas late AVP secretion and thirst, although ship of plasma AVP to plasma osmolality in is presently unknown. Although osmore- some sugars such as mannitol and su- healthy adults during intravenous infusion ceptors can both stimulate as well as in- crose are also equally effective when in- of hypertonic solutions of different solutes. hibit AVP secretion and thirst in response fused intravenously.8 In contrast, in- Note that effective solutes, i.e., those com- to systemic hyper-and hypotonicity, re- partmentalized to the extracellular fluid spectively, it is also not known whether creases in plasma osmolality caused by (NaCl and mannitol), are much more effec- there are separate subsets of excitatory solutes such as urea or glucose cause little tive at eliciting AVP secretion than the non- and inhibitory osmoreceptor cells, or or no increase in plasma AVP levels in effective solutes, urea and glucose, that whether this is a property of single osmo- humans or animals (Figure 2).8,9 These distribute across cell membranes into the receptive cells. differences in response to various plasma intracellular fluid as well (adapted from solutes are independent of any recog- Zerbe and Robertson GL.8) nized nonosmotic influence, which indi- techniques to detect early gene products cates they are an intrinsic property of the in rats, which serve as markers of cell ac- osmoregulatory mechanism itself. Thus, sensitive neurons has been found to acti- tivation after dehydration. Intense ex- it is clear that osmoreceptor cells in the vate membrane nonselective cationic pression of the cFos protein in and brain primarily respond to plasma tonic- conductances that generate inward cur- around the OVLT confirms this area is ity rather than to total plasma osmolality. rent; if of sufficient magnitude, the re- strongly activated by induced dehydra- The physiological relevance of this find- sulting depolarization of the osmorecep- tion, and retrograde tracing studies ver- ing is that osmoreceptors function pri- tor neuron then produces an action ify that a subset of the activated neurons marily to preserve cell volume; elevations potential.10 Conversely, “ineffective” sol- send projections to the magnocellular of solutes such as urea, unlike elevations utes that penetrate cells readily create no AVP neurons in the hypothalamus.6 Al- of sodium, do not cause cellular dehy- osmotic gradient and thus have little to though many of the neural pathways dration and consequently do not activate no effect on the cell volume of the osmo- connecting the OVLT and other circum- the mechanisms that defend body fluid receptors. Electrophysiological studies of ventricular organs with the magnocellu- homeostasis by preserving or increasing neurons in the OVLT show they display lar AVP-secreting cells in the SON and body water stores. changes in action potential firing rate PVN have been identified, the neural cir- that vary in proportion to the tonicity of cuits in the forebrain that stimulate thirst extracellular fluid, supporting the likeli- after osmoreceptor activation are still hood that these cells represent osmosen- largely unknown. Recent studies using WHAT ARE THE CELLULAR sory neurons.5 Osmotically evoked functional magnetic resonance imaging MECHANISMS UNDERLYING changes in the firing rate of the OVLT in humans have shown that the anterior OSMORECEPTION? neurons in turn synaptically regulate the cingulate area of the cortex is reliably ac- electrical activity of downstream effector tivated in conjunction with the sensation “Effective” solutes are those that pene- neurons, importantly including the mag- of thirst.7 Although data from lesion trate cells slowly, or not at all, thereby nocellular AVP neurons in the SON and studies in both animals and man support creating an osmotic gradient that causes PVN, through graded changes in release the concept of a single group of osmore- an efflux of water from osmoreceptor of the excitatory neurotransmitter gluta- ceptive neurons that control both AVP cells. The resultant shrinkage of osmo- mate. This mechanism agrees well with

J Am Soc Nephrol 18: 3056–3059, 2007 How Does the Brain Sense Osmolality? 3057 SCIENCE IN RENAL MEDICINE www.jasn.org the observed relationship between the ef- the function of baroreceptors at the cell membrane stretch, although osmo- fect of specific solutes such as sodium, vascular level. sensitive inhibitory neurons have not yet mannitol, and glucose on AVP secretion The cellular osmosensing mechanism been identified in the OVLT.5 (Figure 2). utilized by the OVLT cells is an intrinsic The combined studies to date there- The presumption that the cell vol- depolarizing receptor potential, which fore strongly support the characteriza- ume of the osmoreceptor cells repre- these cells generate through a molecular tion of TRPV1, TRPV2, and TRPV4 as sents the primary signaling event by transduction complex. Recent results osmomechano-TRPs.15 However, de- which osmoreceptors detect changes in suggest this likely includes members of spite the very promising nature of these the tonicity of the extracellular fluid the transient receptor potential vanilloid findings, several dilemmas are evident raises some interesting dilemmas. (TRPV) family of cation channel pro- with regard to their involvement in brain First, most cells in the body regulate teins. These channels are generally acti- osmoreception. First, it is striking that their volume to prevent or minimize vated by cell membrane stretch to cause a animals with gene deletions of individual the detrimental effects of cell swelling nonselective conductance of cations, members of the TRPV family manifest ϩ or shrinkage on cellular functions. with a preference for Ca2 . Multiple blunted AVP secretion and thirst but However, if osmoreceptors displayed studies have characterized various mem- have a normal basal plasma osmolality. volume-regulatory increases or de- bers of the TRPV family as cellular mech- These results stand in marked contrast to creases in response to changes in extra- anoreceptors in different tissues.14 animals with lesions that destroy the cellular tonicity, this would not allow Both in vitro and in vivo studies of the OVLT and surrounding hypothalamus, for an absolute plasma osmolality TRPV family of cation channel proteins in which osmotically stimulated AVP se- around which body fluid homeostasis provides evidence supporting roles for cretion and thirst are virtually abolished, is maintained; that is, chronic hyperos- TRPV1, TRPV2, and TRPV4 proteins in leading to chronically elevated plasma molality would not elicit sustained the transduction of osmotic stimuli in osmolality. This raises the likelihood that stimuli to AVP secretion and thirst. Re- mammals.15 An N-terminal trpv1 variant different ion channels, or possibly com- sults using OVLT neurons in short- is expressed in OVLT cells, and trpv1- binations of subunits from different term dispersed cultures indeed suggest null mice have defects in osmotically channels, mediate osmoresponsivity in that these cells do not volume-regulate, stimulated AVP secretion and thirst.5 the brain and compensate for the ab- consistent with their putative function Heterologous expression of the trpv2 sence of individual ion channels. Second, as the primary brain osmoreceptors.11 gene in Chinese hamster ovary (CHO) it is surprising that all of the TRPV chan- ϩ Whether this is also true after longer cells causes an activation of Ca2 influx nels appear to be activated by membrane periods of sustained changes in tonicity in response to hypotonicity, a response stretch, including the cell swelling in- has not been studied. Second, in re- that can be mimicked by cell membrane duced by extracellular hypotonicity, sponse to chronic changes in tonicity, stretch.15 trpv4-Transfected cells re- whereas in vitro studies of putative the magnocellular AVP neurons un- spond similarly to hypotonicity and me- OVLT osmoreceptors have indicated dergo effects opposite of those ex- chanical stretch, and they display defi- that the mechanism responsible for hy- pected. These neurons enlarge in re- cient volume-regulatory decreases in perosmolar activation of these cells is ac- sponse to chronic hypertonicity12 and response to hypoosmolality.16 But in vivo tivation of a stretch-inactivated cationic shrink in response to chronic hypoto- studies have yielded inconsistent find- conductance that responds to cell nicity.13 This is postulated to be a result ings. trpv4-Null mice have a potentiated shrinkage.10 These and other questions of changes in cell synthetic machinery; AVP response to a combined hypertonic remain to be answered before we fully upregulation of the many proteins re- and hypovolemic stimulus in one study17 understand brain osmoreceptors and quired for increased AVP synthesis but blunted responses of both AVP se- how they function. during chronic hypertonicity causes cretion and thirst to a selective hyper- cell hypertrophy, and downregulation tonic stimulus in another.18 These find- UNRESOLVED QUESTIONS of these proteins during chronic hypo- ings are not necessarily contradictory tonicity produces the opposite effects. because both AVP secretion and thirst Although the details of exactly how and Thus, the true determinant of osmore- are likely under bimodal control; that is, where various members of the TRPV ceptor activity must be the degree of they are stimulated by hypertonicity and family of cation channel proteins partic- stretch of the osmoreceptor cell mem- inhibited by hypotonicity.19 In support ipate in in different spe- brane, with subsequent effects on of this possibility, treatment with desmo- cies remains to be ascertained by addi- stretch-activated or stretch-inactivated pressin leads to hyponatremia in trpv4- tional studies, a strong case can be made channels, rather than the absolute size null mice but not wild-type controls, in- for their involvement in the transduction of the neurons.10 In this sense, osmore- dicating a failure of osmotic inhibition of of osmotic stimuli in the neural cells in ceptors function as mechanoreceptors drinking.18 Thus, different channels the OVLT and surrounding hypothala- that detect the degree of membrane and/or different sets of osmoreceptor mus that regulate osmotic homeostasis, stretch at the cellular level, similar to cells may mediate opposite responses to which appears to be highly conserved

3058 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 3056–3059, 2007 www.jasn.org SCIENCE IN RENAL MEDICINE throughout evolution.15 Future studies compartments. Prog Brain Res 92: 267–276, tic neurons. Prog Brain Res 139: 85–94, 2002 will be necessary to address still unan- 1992 11. Oliet SH, Bourque CW: Mechanosensitive 3. Johnson AK, Thunhorst RL: The neuroendo- channels transduce osmosensitivity in su- swered questions, including the exact crinology of thirst and salt appetite: Visceral praoptic neurons. Nature 364: 341–343, structure of the molecular transduction sensory signals and mechanisms of central 1993 complex that regulates the opening of integration. Front Neuroendocrinol 18: 292– 12. Armstrong WE, Gregory WA, Hatton GI: Nu- cationic channel(s) in response to 353, 1997 cleolar proliferation and cell size changes in changes in tonicity; whether different 4. Baylis PH, Thompson CJ: Osmoregulation of rat supraoptic neurons following osmotic vasopressin secretion and thirst in health and volemic challenges. Brain Res Bull 2: heteromultimeric combinations of and disease. Clin Endocrinol (Oxf) 29: 549– 7–14, 1977 TRPV1, TRPV2, and TRPV4 and possi- 576, 1988 13. Zhang B, Glasgow E, Murase T, Verbalis JG, bly other cationic channels, mediate dif- 5. Bourque CW, Ciura S, Trudel E, Stachniak TJ, Gainer H: Chronic hypoosmolality induces a ferential responses to changes in tonicity; Sharif-Naeini R: Neurophysiological character- selective decrease in magnocellular neurone whether separate excitatory and inhibi- ization of mammalian osmosensitive neu- soma and nuclear size in the rat hypotha- rones. Exp Physiol 92: 499–505, 2007 lamic supraoptic nucleus. J Neuroendocri- tory osmoreceptors control AVP secre- 6. McKinley MJ, Hards DK, Oldfield BJ: Iden- nol 13: 29–36, 2001 tion and thirst; and the potential involve- tification of neural pathways activated in de- 14. Liedtke W, Kim C: Functionality of the TRPV ment of the TRPV family of ion channels hydrated rats by means of Fos-immunohis- subfamily of TRP ion channels: add in the responses of different tissues to tochemistry and neural tracing. Brain Res mechano-TRP and osmo-TRP to the lexicon! changes in tonicity, particularly the kid- 653: 305–314, 1994 Cell Mol Life Sci 62: 2985–3001, 2005 7. Egan G, Silk T, Zamarripa F, Williams J, Fe- 15. Liedtke W: Role of TRPV ion channels in ney and the vasculature. But with prom- derico P, Cunnington R, Carabott L, Blair- sensory transduction of osmotic stimuli in ising candidate cells and gene products West J, Shade R, McKinley M, Farrell M, mammals. Exp Physiol 92: 507–512, 2007 now clearly identified, answers to these Lancaster J, Jackson G, Fox P, Denton D: 16. Becker D, Blase C, Bereiter-Hahn J, Jen- questions should be forthcoming. Neural correlates of the emergence of con- drach M: TRPV4 exhibits a functional role in sciousness of thirst. Proc Natl Acad Sci U S A cell-volume regulation. J Cell Sci 118: 2435– 100: 15241–15246, 2003 2440, 2005 DISCLOSURES 8. Zerbe RL, Robertson GL: Osmoregulation of 17. Mizuno A, Matsumoto N, Imai M, Suzuki M: None. thirst and vasopressin secretion in human Impaired osmotic sensation in mice lacking subjects: effect of various solutes. Am J TRPV4. Am J Physiol Cell Physiol 285: C96– REFERENCES Physiol 244: E607–E614, 1983 C101, 2003 9. Thrasher TN: Osmoreceptor mediation of 18. Liedtke W, Friedman JM: Abnormal osmotic 1. Verney EB: The antidiuretic hormone and thirst and vasopressin secretion in the dog. regulation in trpv4Ϫ/Ϫ mice. Proc Natl Acad the factors which determine its release. Proc Fed Proc 41: 2528–2532, 1982 Sci U S A 100: 13698–13703, 2003 R Soc London (Ser B) 136: 25–106, 1947 10. Bourque CW, Voisin DL, Chakfe Y: Stretch- 19. Verbalis JG: Osmotic inhibition of neurohy- 2. Mason WT, Hatton GI, Kato M, Bicknell RJ: inactivated cation channels: Cellular targets pophysial secretion. Ann N Y Acad Sci 689: Signal transduction in the neurohypophyseal for modulation of osmosensitivity in supraop- 146–160, 1993

J Am Soc Nephrol 18: 3056–3059, 2007 How Does the Brain Sense Osmolality? 3059