Osmotic activation of the hypothalamo-neurohypophysial system reversibly downregulates the NMDA receptor subunit, NR2B, in the supraoptic nucleus of the hypothalamus

Molecular Brain Research 70 Ž1999. 187–196 www.elsevier.comrlocaterbres

Research report

Osmotic activation of the hypothalamo-neurohypophysial system reversibly downregulates the NMDA receptor subunit, NR2B, in the supraoptic nucleus of the hypothalamus ) Margarita C. Curras-Collazo , Jennifer Dao ´ Department of Neuroscience, UniÕersity of California at RiÕerside, RiÕerside, CA 92521, USA Accepted 30 March 1999

Abstract NMDA receptor activation produces a characteristic pattern of neuronal firing in magnocellular neuroendocrine cells ŽMNCs. of the supraoptic nucleus of the hypothalamus ŽSON. which has been associated with greater hormone release in vivo and in vitro. In addition, i.c.v. administered NMDA receptor blockers suppress the dehydration-induced rise in plasma vasopressin and drinking. To investigate the role of NMDA receptor subunits in the neuroendocrine functions of the magnocellular neuroendocrine cells of the hypothalamus, we examined the effects of osmotic stimulation on the protein expression of the NMDA receptor subunits, NR1 and NR2B, important in binding glycine and glutamate, respectively. Homogenates of SON, paraventricular nucleus of the hypothalamus ŽPVN., cortex and lateral hypothalamus from control rats and rats given 2% saline water to drink for 4–10 days were subjected to SDS–PAGE and Western blot analysis. This saline water drinking regimen produced a significant rise in plasma osmolality levels. NR1 and NR2B immunoreactivity was detected in SON, PVN, lateral hypothalamus and cortex but not in liver homogenates using subunit-specific polyclonal antibodies and quantified using computer-assisted densitometry. Mean NR2B immunoreactivity was significantly lower in SON Ž29%. and PVN homogenates Ž23%. from saline-treated rats than in those from control rats. In addition, the effect of dehydration on NR2B was regionally specific since no significant changes in NR2B expression were observed in homogenates of cortex and lateral hypothalamus. Rehydration allowed recovery of plasma osmolality as well as NR2B protein levels in the SON. These results suggest that changes in NMDA receptor subunit expression contribute to the plasticity manifested by in magnocellular neuroendocrine cells in response to osmotic activation of the hypothalamo-neurohypophysial system. In addition, our results indicate that NMDA receptors on SON and PVN MNCs may contribute to neuroendocrinological functions associated with body fluid homeostasis. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Excitatory amino acid receptors; Glutamate receptors; Receptor modulation; Receptor down-regulation; Osmotic regulation; Vasopressin; Magnocellular neuroendocrine cells

1. Introduction The supraoptic ŽSON. and paraventricular nuclei of the hypothalamus ŽPVN. contain two populations of magnocellular neuroendocrine cells ŽMNCs. which synthesize the peptide hormones, vasopressin or oxytocin, which are transported to axon terminals located in the neurohypophysis. Secretion of these hormones from the hypothalamoneurohypophysial system ŽHNS. occurs in response to hemorrhage, dehydration, suckling and parturition w20x. In particular, vasopressin has an antidiuretic action and is

) Corresponding author. [email protected]

Fax:

q 1-909-787-2966;

E-mail:

important in body water balance. Osmotic stimulation in vivo and in vitro increases the electrical activity of MNCs in the SON and PVN via synaptic or intrinsic mechanisms w6,16,19,56x. Glutamate is believed to mediate the majority of excitatory synaptic activity in the PVN and SON w34,50x. Functional studies suggest that NMDA-type glutamate receptors are present on SON MNCs w11,22,33x and that they may mediate, in part, the excitatory synaptic transmission in the SON w17,56,57x. In support of these studies glutamatergic synapses have been located in the region dorsal to the SON w34x. Interestingly, activation of NMDA but not non-NMDA glutamate receptors on SON MNCs produces a clustered pattern of neuronal firing w22x which has been associated with greater release of peptide hormones in vivo

0169-328Xr99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 3 2 8 X Ž 9 9 . 0 0 1 2 9 - 1

188

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´

and in vitro w4,5,15,16x. It has been suggested that glutamate receptors mediate basal release as well as osmotically-induced vasopressin release w43x. We have recently shown that NMDA-type and metabotropic-type glutamate receptor activation evokes vasopressin release in vitro w33x. In contrast, little is known about the role of NMDA receptors in MNCs during osmotic activation. An increase in NMDA receptor density has been demonstrated in response to water deprivation w35x and dehydration w14x. A recent report has also shown i.c.v. injections of an NMDA receptor antagonist suppresses the osmotic-induced rise in plasma vasopressin and drinking w39x. However, more work is needed to determine whether NMDA receptors expressed by magnocellular neuroendocrine cells are involved in their synaptic and intrinsic responses to osmotic activation. NMDA receptors are probably assembled from two classes of protein subunits: NR1 and NR2 subunits ŽNR2A–D. w29,36,46x. A recent gene expression study has found that NR1 is the most abundant subunit in the adult SON followed by NR2B, NR2C and NR2D subunits in relatively equal proportions w1x. However, functional and gene expression studies disagree on whether NR2B is expressed in the SON w1,23,52x. We have recently shown that the SON and PVN express NR1 and NR2B proteins and that this expression is dynamic depending on developmental age, a finding which may explain the discrepancies reported previously w9x. Given the importance of NR1 and NR2B subunits in binding the coagonists, glycine and glutamate, respectively w10,27,28,31x, we have designed experiments to explore the possibility that these subunits are involved in the response of SON and PVN MNCs to osmotic activation of the HNS in vivo. Our present findings indicate that osmotic activation evoked by saline water drinking results in the downregulation of NR2B expression. Dehydration-induced alterations in NR2B levels are specific to the SON and PVN and, at least in the SON, are reversible upon rehydration. In addition, we have found that NR1 and NR2B subunits in MNCs are differentially regulated in the adult rat in response to osmotic stimulation. These data provide further evidence for an important role of NMDA receptors in the neuroendocrinological responses of MNCs. Portions of these data have been presented in preliminary form w13x.

2. Materials and methods 2.1. Osmotic stimulation of animals Twenty eight pairs of male Holtzman rats Ž250–400 g. were matched according to their weight and age. Rats were randomly assigned to two experimental groups: Ž1. a saline-treated group which received 2% saline to drink for 4–10 days and Ž2. a control group which was allowed free

access to tap water. Each rat was given rat chow ad libitum. Saline water drinking for 3–7 days causes magnocellular neuroendocrine cells to release oxytocin and vasopressin w16,25x. Only rats which showed a 2.5% or greater increase in plasma osmolality at the end of treatment relative to control rats were used to study the effects on NR1 and NR2B subunits. Some animals from each group were used to study the time course of hyperosmolality with saline water drinking. In these rats blood was collected from the tail before and during the course of saline treatment Ždays 0, 3, 4 and 7.. On the day of sacrifice blood was collected transcardially under deep anesthesia prior to euthanasia. Blood samples were centrifuged at 3800 = g for 5 min and plasma in the supernatant was collected. An average plasma osmolality was calculated from triplicate measurements made for each sample using a vapor pressure osmometer ŽWescor.. Mean values for saline-treated and control groups were compared using a Student’s t-test with a 5% experiment-wide error rate. Statistical comparisons were considered significant at the 0.05 level. 2.2. Rehydration of animals The reversibility of saline drinking effects were examined in additional experiments using 22 rats separated into two weight-matched groups which were both given saline water to drink for 7–10 days. At the end of the treatment one group was given tap water for 4–12 days of rehydration Žrecovery. while the other group was sacrificed at the end of treatment Žsaline.. In some recovery animals blood samples were taken before the start of rehydration Žon the last day of saline treatment. and at days 1, 2, 4, 7, 10 and 12 days during rehydration. All procedures were conducted with the approval, and in strict accordance to the guidelines of, the UC Riverside Chancellor’s Committee on Laboratory Care. 2.3. Tissue harÕest Brain and liver tissue were collected from each rat and placed in ice-cold 0.01 M phosphate-buffered saline ŽPBS.. The frontal cortex, cerebellum and liver were removed free-handedly using a clean razor blade. Hypothalamic regions of the brain were obtained by placing the brains into an ice-cold rodent brain matrix with coronal channels spaced 1 mm apart ŽActivational Systems. and cutting 10 = 15 mm blocks containing the SON, PVN and lateral hypothalamic area ŽLH.. Using a vibratome blocks were cut into 700 mm sections in ice-cold PBS and transferred to moist glass slides placed on a plastic-covered ice block. Using the third ventricle, fornix and optic tract as landmarks, SON, PVN, and LH tissue were microdissected with iris scissors and a syringe attached to a modified 23-gauge needle. Samples were collected in ice-cold protease inhibitor cocktail ŽPIC. consisting of 10 mM Tris–OH

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´

ŽpH 8.2. plus 320 mM sucrose containing the protease inhibitors: EDTA Ž5 mM., benzamidine Ž1 mM., aprotinin Ž2.3 mgrml., bacitracin Ž200 mgrml., phenylmethylsulfonyl fluoride Ž0.2 mM., and leupeptin Ž10 mgrml.. Samples were subsequently homogenized using a sonic dismembrator ŽFisher. and centrifuged at 21,000 = g for 20 min at 48C. Tissue pellets were resuspended in 100 mM Tris ŽpH 6.8. containing 1 mM EDTA and assayed for protein using the bicinchoninic acid method ŽPierce. and bovine serum albumin as a standard. 2.4. SDS–PAGE and immunoblotting Protein samples obtained from different brain regions were resuspended and denatured in an equal volume of 2 = sample buffer Ž800 mM Tris, pH 6.8, 0.2% bromophenol blue, 10% beta-mercaptoethanol, 40% glycerol and 6% SDS. for 20 min and stored at y208C until use. Since our previous results suggest that cortex is enriched with NR2B protein relative to SON, PVN and LH w9x, lanes were loaded with either 5 mg protein for cortex or 20 mg protein for the other brain regions. Each SON sample contained the nuclei from both sides of individual adult brains. Hiver samples were used as negative controls. Samples were loaded into individual lanes of 6% polyacrylamide gels and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis ŽSDS–PAGE.. Molecular weights were estimated using prestained protein standards obtained from BioRad ŽRichmond, CA.. Proteins were electrotransferred onto polyvinylidene fluoride ŽPVDF. blotting membranes

189

in modified Towbin’s buffer. Subsequently, blots were blocked in PBS containing 0.05% Tween ŽT–PBS. and 5% non-fat milk overnight at 48C. Blots were then incubated for 1.5 h at room temperature with either a polyclonal rabbit anti-rat NR2B antibody w44x Ž1:600, 10 mgr50 ml; Chemicon. or a polyclonal rabbit anti-rat NR1 antibody w41x Ž1:100, 50 mgr500 ml; Chemicon. which recognize a portion of the putative C-terminal regions of the NR2B or NR1 subunits, respectively. Data on the specificity of these two antibodies has been published previously w9,40,44x. After washes in T–PBS, blots were incubated with a peroxidase-labeled goat anti-rabbit secondary antibody Ž1:10,000; BioRad. for 1 h at room temperature. After more washes, NR1 and NR2B immunoreactivity was detected as dark bands on photographic film ŽHyperfilm; Amersham . using enhanced chemiluminescence ŽAmersham.. 2.5. QuantitatiÕe and statistical analysis Immunoreactive bands were scanned with 8-bit accuracy and 1024 dpi spatial resolution using a scanner with a transparency illuminator ŽScanjet 4CrT; Hewlett-Packard. and quantified using computer-assisted densitometry ŽGel Pro, Media Cybernetics.. This system calculates an integrated optical density ŽIOD. for each immunoreactive band using all pixels in the band which are above the film background. IODs represent both the intensity and area of the band. IOD values for NR1 and NR2B protein levels were expressed as the mean " S.E.M.= 10 4 for each group.

Fig. 1. Time course of plasma osmolality changes during saline treatment. Plasma osmolality was measured in adult male Holtzman rats given tap water ŽCONTROL. or 2 g% saline to drink ŽSALINE.. Measurements were taken immediately before treatment Žday 0. or during treatment Ždays 3, 4, 7, and 10.. In the SALINE group, plasma osmolality increases dramatically after 3 days, remains elevated until day 7, and then increases again by day 10. These results show that 3–4 days of saline treatment is sufficient to induce a significant osmolality increase of 10–11% increase in plasma osmolality. Asterisks indicate that plasma osmolality values of saline-treated rats were significantly elevated relative to those of control rats Ž p - 0.05..

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´

190

Fig. 2. Effect of dehydration on the protein expression of NR1 and NR2B subunits in various brain regions. Protein aliquots of SON, PVN, LH and CBM homogenates Ž20 mg. and of COR Ž5 mg. obtained from control Ž C . and rats given saline Ž S . were loaded in individual lanes of a 6% acrylamide gel and subjected to SDS–PAGE. Blots were immunoprobed with a polyclonal NR1 Žtop. or NR2B antibody Žbottom.. The estimated molecular size of the major immunoreactive bands for NR1 Ž116 kDa. and NR2B Ž178 kDa. approaches the values predicted from their cDNAs w24,41x. Saline treatment results in upregulation of NR1 in SON Žtop. and downregulation of NR2B in SON and PVN Žbottom.. In comparison, LH and COR do not show significant changes in NR2B levels. COR, frontal cerebral cortex; CBM, cerebellar cortex; LH, lateral hypothalamic area: PVN, paraventricular nucleus of the hypothalamus; SON, supraoptic nucleus of the hypothalamus.

We have previously shown that IOD values obtained using the conditions described are linear over the range of protein amount used throughout this study, i.e., 0.5 to 30 mg and at 1–3 min exposure times w9x. NR1 and NR2B IOD values of samples from individual saline rats were expressed as a percent of the IOD values that of their weight-matched control samples loaded on the same blot in order to control for minor differences in immunoblotting and film exposure time. A mean percent change was then calculated for each brain region in order to determine the degree of change produced by dehydration. For valid statistical comparisons between the groups, IODs were expressed in natural log units in order to standardize them to a normal distribution. The ln IOD value of samples from individual saline rats was then subtracted from that of their weight-matched control rats. Using a one-sample t-test the difference in these values was compared to a hypothesized mean difference of zero.

P-values less than 0.05 were required to declare a significant difference in the means.

3. Results In order to determine the efficacy of 2% saline water drinking in Holtzman rats, we measured the degree and time course of plasma osmolality changes resulting from treatment. Fig. 1 shows that the average plasma osmolality of untreated rats Žrange of 296 " 3–300 " 2 mOsMrkg. remained constant throughout the duration of the experiment and was no different from that of saline rats before the start of treatment Žday 0; 296 " 2, mOsMrkg.. During treatment, however, saline rats had significantly elevated plasma osmolality compared to control rats: 325 " 13 mOsMrkg at day 3, 329 " 10 mosMrkg at day 4, 329 " 7 mosMrkg at day 7 and 345 " 8 mosMrkg at day 10

Table 1 Summary of NR1 and NR2B protein expression in selected brain regions from rats given 2% saline to drink Ž S . or control rats Ž C . allowed free access to tap water Region

SON PVN LH COR

NR2B

NR1

C

S

Mean % change

Mean Žln C y ln S .

n

C

S

Mean % difference

mean Žln C y ln S .

n

0.87 " 0.31 2.78 " 0.75 0.73 " 0.30 2.05 " 0.42

0.71 " 0.32 1.77 " 0.57 0.53 " 0.24 2.25 " 0.57

29.09 23.04 4.87 y14.74

0.64 " 0.13U 0.41 " 0.14U 0.26 " 0.16 0.00 " 0.11

17 18 15 24

0.47 " 0.14 1.61 " 0.50 0.52 " 0.13 5.08 " 0.67

0.79 " 0.31 1.46 " 0.51 0.30 " 0.09 4.78 " 0.82

y61.27 y14.89 36.77 10.14

0.20 " 0.21 y0.02 " 0.14 0.53 " 0.11U 0.17 " 0.09

15 15 13 16

Values are expressed as integrated optical densities Žmean " S.E.M.= 10 4 . unless otherwise noted. Percent change relative to control was calculated for each pair of weight-matched rats and the results pooled to obtain a mean for each brain region. Positive percent values indicate a decrease in subunit expression relative to control values. Asterisks indicate statistical significance at p - 0.01.

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´

Fig. 3. Dehydration results in downregulation of the NR2B protein in the SON and PVN. NR2B IOD values for each saline rat treated for 4–10 days were expressed as a percent of that measured for its weight-matched control Žset at 100%. and were pooled for each brain region. Mean NR2B immunoreactivity in the SON and PVN of saline rats was 29.1% and 23% lower relative to control indicating a dehydration-induced downregulation of the NR2B protein. In comparison, LH and COR do not show significant changes in NR2B levels. Only data were included from saline-treated rats with plasma osmolality values at least 2.5% greater than control values Žmean values were 11%–16% greater in saline rats.. Asterisks represent statistical significance at the 0.05 level. COR, frontal cerebral cortex; CBM, cerebellar cortex; LH, lateral hypothalamic area; PVN, paraventricular nucleus of the hypothalamus; SON, supraoptic nucleus of the hypothalamus.

Ž p - 0.05, n s 8–23.. Therefore, saline treatment produced an average increase in plasma osmolality of 10% by day 3 and an increase of 16% by day 10 as well as the characteristic symptoms of osmotic stimulation such as

191

polydipsia and polyuria. After seven days of the same protocol Zerbe and Palkovits w58x observed an identical mean increase of 11% in plasma osmolality which was associated with a rise in plasma vasopressin from 2.0 to 13.7 pgrml. Nineteen of 25 rats tested responded to saline water drinking with a 2.5% or greater increase in plasma osmolality compared to their weight-matched control. Immunoblots of brain homogenates performed using affinity-purified NR1 and NR2B-specific antibodies revealed differences in the protein expression levels of these NMDA receptor subunits during dehydration ŽFig. 2.. NR1 and NR2B blots show major immunoreactive bands at 116 kDa and 178 kDa, respectively, which are slightly greater than that predicted from the cDNA of these subunits. However, these values agree well with the molecular mass of the glycosylated forms of these subunits and that of the phencyclidine-binding polypeptide associated with the NMDA receptor w7,24,45,51x. Robust NR2B expression was observed in homogenates of adult SON, PVN as well as in those of cortex and the lateral hypothalamic area ŽLH., but only slight expression in those of cerebellum as previously reported w51x. NR1 expression was robust in SON, PVN, cortex and LH homogenates. Regional comparisons of NR2B and NR1 in these regions have been previously reported by our laboratory w9,26x. More importantly, Fig. 2 shows that saline treatment resulted in decreased NR2B protein levels in the SON and PVN. Conversely, dehydration appeared to upregulate NR1 levels in these brain regions, especially in the SON. By comparison, NR2B levels in cortex and LH homogenates did not change markedly.

Fig. 4. Time course of plasma osmolality normalization during recovery from dehydration. Plasma osmolality was measured in adult male Holtzman rats given 2 g% saline to drink for 7–10 days and subsequently allowed free access to tap water ŽRECOVERY.. The mean value measured for these rats on the last day of saline treatment Žday 0 of recovery. is significantly greater than that measured for control rats ŽCONTROL.. It can be noted that the increased plasma osmolality values produced by saline treatment are still elevated relative to that of control rats after 1–2 days of rehydration but recover completely by day 4. These results show that 4 days of rehydration is sufficient to recover from osmotic loading lasting for 7–10 days. Asterisks indicate statistical significance at 0.05 level.

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´

192

NR2B expression from each saline rat was then subtracted from that of its weight-matched control and differences were pooled to obtain a mean which was compared to a hypothesized mean of 0 using a one-sample t-test. Differences in NR2B levels between saline and control rats were significantly different from 0 in the SON and PVN Žasterisks.. On the other hand, dehydration-induced differences in NR2B expression did not reach statistical significance in the COR Ž p s 0.99. or LH Ž p s 0.13.. In comparison, dehydration had a tendency to increase NR1 protein levels in the SON and PVN although this was not statistically significant. Interestingly, a significant change in NR1 levels was calculated in the LH during dehydration. In a separate set of experiments we examined if dehydration-evoked downregulation of NR2B was reversed by allowing osmotic homeostasis. Rats given saline water to drink were allowed to recover with free access to tap water for 4–12 days post treatment. Although the mean plasma osmolality decreased dramatically from 354 " 3 Ž n s 14. to 307 " 3 mosMrkg Ž n s 5. within the first day of recovery, the mean value on this day and at 2 days post treatment were still significantly elevated as compared to that of untreated control rats ŽFig. 4.. The mean value decreased more gradually with successive days of recovery and was comparable to that of control rats by post treatment day 4. Western blot analysis was performed to determine if rehydration after saline treatment resulted in recovery of normal NR2B protein levels. Pairs of weight-matched saline and recovery rats Ž7–11 pairs. were dehydrated for 7–10 days but only recovery rats were given tap water to drink during a rehydration period of 4–12 days. In Fig. 5 representative NR2B blots from a saline and recovery rat treated with saline for 7 days demonstrate downregulated NR2B levels in the SON and PVN during dehydration and a return back to normal levels with rehydration for 6 days. In comparison, NR2B expression in cerebral cortex homogenates was constant at all hydration states. These results indicate that osmotic activation evokes a reversible downregulation in NR2B protein levels of MNCs.

Fig. 5. Dynamic expression of NR2B in the SON is dependent on osmotic state. Homogenates of SON and PVN Ž20 mg. and COR Ž5 mg. obtained from control Ž C . and saline-treated rats Ž S . were loaded in individual lanes of a 6% acrylamide gel and subjected to SDS–PAGE. Immunoblots were probed with an anti-rat NR2B antibody. Saline water drinking for 7 days produces a decrease in NR2B immunoreactivity at 180 and 186 kDa in the SON and PVN but not in the COR of the same rat. NR2B levels are back to normal in SON after 6 days of rehydration Ž R .. In the PVN, NR2B expression appears to recover partially during this period. COR, frontal cerebral cortex; SON, supraoptic nucleus of the hypothalamus; PVN, paraventricular nucleus of the hypothalamus.

NR2B IOD values for each brain region obtained from each saline rat were expressed as a percent of its weightmatched control in order to estimate the degree of change evoked by dehydration. Data were pooled to obtain a mean percent change due to dehydration for each brain region ŽTable 1.. This data is summarized in Fig. 3 which shows that NR2B levels in the SON and PVN from saline-treated rats were reduced by an average of 29.1% and 23.0%, respectively, relative to controls. Fig. 3 also demonstrates that the effects of saline treatment on NR2B protein levels are specific to the magnocellular neuroendocrine nuclei. In order to perform reliable statistical comparisons, NR2B IOD values were converted to natural log units to normalize the distribution of IOD values. The ln IOD for

Table 2 Summary of NR1 and NR2B protein expression in selected brain regions from rats given 2% saline to drink Ž S . or allowed to recover with free access to tap water following similar saline treatment Ž R . Region

SON PVN LH COR

NR2B

NR1

S

R

Mean % change

Mean Žln S y ln R .

n

S

R

Mean % change

Mean Žln S y ln R .

n

0.19 " 0.04 0.55 " 0.20 0.23 " 0.05 2.64 " 1.15

0.34 " 0.08 0.80 " 0.41 0.27 " 0.07 2.68 " 1.07

y85.6 y20.4 y4.09 y12.09

y0.51 " 0.19U y0.12 " 0.12 y0.27 " 0.32 0.04 " 0.09

7 10 7 11

1.14 " 0.42 1.09 " 0.26 0.13 " 0.04 2.46 " 0.89

0.60 " 0.20 1.14 " 0.31 0.11 " 0.01 2.48 " 0.89

26.96 y17.69 26.8 y0.68

0.53 " 0.30 y0.04 " 0.19 y0.15 " 0.34 y0.06 " 0.05

7 8 7 9

Values are expressed as integrated optical densities Žmean " S.E.M.= 10 4 . unless otherwise noted. Percent change relative to levels in saline rats was calculated for each pair of weight-matched rats and the results pooled to obtain a mean for each brain region. Negative percent values indicate an increase in subunit expression relative to saline values. Asterisks indicate statistical significance at p - 0.01.

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´

Pairwise comparisons of NR2B IOD values from saline and weight-matched recovery rats indicated a marked recovery in SON NR2B levels during rehydration ŽTable 2.. A similar trend, albeit with a lesser magnitude, was observed in PVN homogenates. Statistical comparisons of normalized NR2B IODs yielded significant differences only for the SON. Unlike NR2B protein levels in the SON, those in the PVN Ž p s 0.33., LH Ž p s 0.44. or cerebral cortex Ž p s 0.44. did not recover during rehydration. By comparison, NR1 expression remained relatively constant in all brain regions examined although the tendency for NR1 upregulation which was found in the SON during dehydration ŽTable 1. appeared to reverse upon rehydration Ž p s 0.12.. No reversal of dehydration-induced NR1 downregulation in the LH was observed Ž p s 0.68.. 4. Discussion Our findings provide the first demonstration that NR2B levels in the adult SON and PVN are actively downregulated by osmotic activation of the hypothalamo-neurohypophysial axis ŽHNS.. NR2B levels return to normal during recovery from salt loading suggesting that NMDA receptors are important in body fluid balance, a process essential for homeostasis and environmental adaptation. That NR2B levels in the SON are reversibly regulated by osmotic stimuli is relevant since neurons in this region have been shown to respond electrophysiologically, functionally and morphologically to osmotic perturbations w6,16,19x and since neuroendocrine cells in this region produce the majority of vasopressin released into the systemic circulation w2x. Moreover, these results provide further evidence for the plastic nature of the magnocellular neuroendocrine cells in the adult. Using an NR2B-specific antibody we have confirmed that the NR2B protein subunit is expressed by the magnocellular nuclei, SON and PVN. Functional and gene expression studies disagree on whether the NR2B subunit is expressed in the SON w1,23,52x. In fact, Al-Ghoul et al. w1x reported that in addition to NR1, NR2B and NR2C are the most abundant subunits in the adult SON. We have recently shown that NR1 and NR2B protein expression in the SON is dynamically upregulated during postnatal development, a finding which contrasts with that of a previous functional study which concluded that NR2B-containing NMDA receptors are not present on SON neurons w23x. Similarly, a previous in situ hybridization report failed to detect NR2B mRNA in the PVN w52x. Our findings suggest that NR2B is synthesized in the PVN as well as SON and supplement those of a previous immunohistochemical study showing immunoreactivity in the adult SON and PVN to a non-specific NR2ArNR2B antibody w40x. In the present study we used an NR2B-specific antibody which has been previously tested for specificity to NR2B w9,44x. Our results show that saline water drinking significantly decreases protein levels of NR2B by 23–29% in the SON

193

and PVN, respectively. Dehydration-induced downregulation of NR2B was specific to the area of the SON and PVN since neither cortex nor LH showed significant changes in NR2B expression. In parallel, plasma osmolality levels were significantly elevated above normal during the time course of this regimen indicating correlative changes in subunit expression and osmotic state. The selective impact of the 2% saline regimen on SON and PVN NR2B levels is probably not a secondary effect of the increase in MNC size induced by osmotic stimulation w20x. Cellular hypertrophy provoked by dehydration cannot explain the decreased levels of NR2B since NR1 levels in the same homogenates from dehydrated rats do not decrease concomitantly. Instead, the specificity of the dehydration effect on NR2B in the SON may be related to the intrinsic osmosensitivity of MNCs in this region w5,6,19x which has been associated with stretch-sensitive channels w37x. The mechanosensitivity displayed by NMDA receptors in diencephalic neurons w38x may also regulate NMDA receptor subunit levels. Alternatively, alterations in NMDA receptor subunit expression in the SON and PVN may be a secondary effect of increased synaptic inputs from other osmosensitive brain areas. For example, NMDA receptormediated synaptic responses in the SON have been demonstrated by stimulating afferents in the organum vasculosum lamina terminalis ŽOVLT. w56x. In the PVN only a minor NMDA receptor-mediated contribution to the response to perifornical stimulation has been demonstrated w50,53x. However, our protocol does not allow us to distinguish between intrinsic and synaptic mechanisms underlying the changes in NMDA receptor subunit levels. Nevertheless, it appears that NMDA receptors may contribute significantly to body water hemeostasis since application of NMDA to SON neurons evokes a firing pattern similar to that evoked by osmotic activation in vivo and which has been associated with enhanced vasopressin release w4,15,16,22,42x. In addition, Xu and Herbert w54x have shown that general blockade of NMDA receptors suppresses drinking by 81% and suppresses the rise in plasma vasopressin by 75% evoked by water deprivation or injection of hypertonic saline. Non-NMDA receptor antagonists, on the other hand, failed to inhibit water-deprivation-induced drinking. Taken together, these findings suggest that NMDA-type glutamate receptors are involved in normal synaptic processing within the SON and PVN and that they may contribute to the neuroendocrinological responses of MNCs to dehydration. The reversible and selective changes in NR2B expression which occur in the supraoptic nucleus in response to dehydration and rehydration suggest that NR2B-containing NMDA receptors in this region are particularly important for osmotic homeostasis. Significant and consistent downregulation of NR2B was also detected in the PVN during dehydration. However, NR2B levels did not normalize within the rehydration period examined in this study Ž4–12 days.. This period of time was chosen since plasma osmo-

194

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´

lality had reached normal levels by day 2 of post-treatment and since vasopressin levels measured by others using immunohistochemistry recover with a similar timecourse in the SON and PVN. For example, it has been previously shown that vasopressin levels normalize between 4 and 7 days of rehydration following 7 days of dehydration w55x. Failure of NR2B levels in the PVN to normalize may be related to its heterogeneous physiological functions, in addition to body fluid balance. Indeed, osmotic stimulation does not evoke vasopressinergic changes in autonomic nuclei innervated by the caudal PVN such as the solitary tract nucleus and septum w2x. Other physiological stimuli such as dehydration-associated stress, which affects PVN function and which may take longer to dissipate than the recovery period chosen here, may be influencing NMDA subunit changes in the PVN. However, unlike the changes evoked by dehydration, a 24-h immobilization protocol for inducing stress has been shown to cause an increase in NR2B gene expression in the SON w3x. In addition, studies addressing short-term stress induced by hypertonic saline injection or restraint suggest that stress does not seem to cause changes in vasopressin mRNA in SON or PVN magnocellular neurons w18x. The results of other studies have also downplayed the effects of dehydration on NMDA receptors in the PVN. Indeed, water deprivation, an osmotic stimulus which produces additional changes in blood volume, evokes significant changes in NMDA receptor density in the SON but not PVN w35x. Indeed, the large size of the SON relative to the magnocellular PVN makes it the most important site of production for vasopressin w2x. Therefore, the reversible changes in NR2B levels occurring in the SON may be more directly associated with osmotic homeostasis than those changes occurring in the PVN. A possible confounding variable in our study is weight loss produced by the saline water drinking regimen. The possibility that weight loss influences NR2B expression is being examined in ongoing experiments by comparing satiated and food-deprived animals. In addition, NR2B expression is being compared in saline rats and foodmatched control rats that are fed identical amounts of food and whose weights are similar throughout saline treatment. Our initial observations are that NR2B expression in the SON is not consistently affected by food deprivation ŽCurras-Collazo, Khan, Dao and Stanley, unpublished ob´ servations.. This finding, in combination with the lack of effect of dehydration on the NR2B protein levels in the lateral hypothalamus, obtained in the present study, suggests that weight loss probably does not mediate the results shown in the present study. In addition to MNCs, our sample homogenates from SON and PVN most likely contain interneurons from the SON proper or perinuclear zone w2x and non-neuronal cells which may also synthesize the NR2B subunit. In particular, glia associated with the SON may be involved in the molecular plasticity of NMDA receptor subunits since they

have been shown to respond robustly to osmotic signals. An active retraction of astrocytic processes over the neuronal surfaces has been postulated to help explain the increase in the amount of directly apposed membranes between MNC somata occurring during conditions which stimulate hormone release w39,47–49x. In support of this, Hawrylak et al. w21x have recently shown a dehydration-associated reduction in immunoreactive glial fibrillary acidic protein in the SON and subjacent glia limitans. If SON glial processes express NR2B, as do cortical astrocytes and Bergman glia w12,32x, then decreased astrocytic coverage within the SON Žand PVN. would contribute to the NR2B downregulation observed during dehydration. However, while Conti et al. w12x observed NR2B immunoreactive labeling on distal processes of cortical astrocytes, it was only found on a minor proportion of astrocytes. Moreover, recent immunohistochemical results obtained in our laboratory suggest that astrocytes do not express the NR2B subunit w7x. Other changes in ligand-gated receptors in SON–VGL astrocytes have been previously documented; upregulated beta-adrenergic receptors have been measured in this area during dehydration w30x. We expect that the cellular resolution afforded by NR2B immunohistochemistry on SON and PVN sections, currently being performed in our laboratory, will detect changes in NR2B expression specifically associated with MNCs and help identify other cell types, if any, responding to osmotic activation with changes in NR2B expression. Recently, we reported changes in NR1 protein levels in the SON and PVN after subjecting rats to a dehydration protocol identical to the one used here. Using quantitative immunohistochemistry on hypothalamic sections containing the SON and PVN we detected a significant increase in NR1 immunoreactivity in MNCs double-labeled with vasopressin- or oxytocin-associated neurophysin antibodies w14x. As in the present study, our previous study failed to detect consistent changes in NR1 expression in SON and PVN homogenates when using Western blot analysis. One reason for this finding may be dilution of MNC proteins by other cellular components such as glia and non-MNC neurons present in homogenates from these regions. Nevertheless, the present immunoblotting experiments were able to detect robust and consistent dehydration-induced changes in NR2B expression in the magnocellular nuclei. As for the subunit changes observed in the LH, its involvement in feeding-related drinking may explain why its NR1 levels were significantly changed during dehydration. In agreement with this finding, Meeker et al. w35x have also observed changes in NMDA receptor binding in the lateral hypothalamic area during dehydration. Our data suggests that NR1 and NR2B may contribute differently to the physiological compensatory processes implemented during osmotic thirst. While the mechanisms by which NMDA receptor subunits are regulated remain to be defined, these mechanisms must explain the differential and independent regulation of NR1 and NR2B subunits in

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´

brain regions associated with osmoregulation. We speculate that enhanced activity of NMDA receptors provoked by increased synaptic activity arriving from osmosensitive regions such as OVLT w56x may subsequently lead to NR2B downregulation. In support of this, our in vitro data indicate that subunit expression of NR2B but not NR1 is strongly regulated by extracellular concentrations of NMDA receptor agonists ŽM.C. Curras-Collazo and Hayes, ´ unpublished observations.. It is likely that changes in the expression of other NR2 subunits present within the SON may be altered concomitantly with NR2B downregulation since recent autoradiographic data obtained for the SON suggests that NR1 and total NR2 subunit mRNAs are present in equimolar ratios under normal conditions w1x. Further work is needed to understand the functional significance of state-dependent and region-specific NMDA receptor subunit plasticity in the neuroendocrine hypothalamus.

Acknowledgements This study was supported by NSF grant IBN-9604233 ŽM.C.C.-C.. and Sigma Xi Grant-in-Aid ŽJ.D... We are grateful to Dr. Robert Beaver, Chair of the Statistics Department at UC Riverside for help with statistical analysis and to Mr. C.M. Collazo for assistance with the graphics.

References w1x W. Al-Ghoul, R. Meeker, R. Greenwood, Differential expression of five N-methyl-D-aspartate receptor subunit mRNAs in vasopressin and oxytocin neuroendocrine cells, Mol. Brain Res. 44 Ž1997. 262–272. w2x W.E. Armstrong, Hypothalamic supraoptic and paraventricular nuclei, in: G. Paxinos ŽEd.., The Rat Nervous System, Academic Press, New York, 1995, pp. 377–390. w3x V. Bartanusz, J.-M. Aubry, S. Pagliusi, D. Jezova, J. Baffie, J.Z. Kiss, Stress-induced changes in messenger RNA levels of N-methylD-aspartate and AMPA receptor subunits in selected regions of the rat hippocampus and hypothalamus, Neuroscience 66 Ž1995. 247– 252. w4x R.J. Bicknell, G. Leng, Relative efficiency of neural firing patterns for vasopressin release in vitro, Neuroendocrinology 33 Ž1981. 295–299. w5x M.J. Brimble, R.E.J. Dyball, M.L. Forsling, Oxytocin release following osmotic activation of oxytocin neurones in the paraventricular and supraoptic nuclei, J. Physiol. 278 Ž1978. 69–78. w6x C.W. Bourque, L.P. Renaud, Activity patterns and osmosensitivity of rat supraoptic neurones in perfused hypothalamic explants, J. Physiol. 349 Ž1984. 631–642. w7x L. Bozzetti, M.C. Curras, ´ Vasopressin-immunoreactive neuroendocrine cells of the supraoptic and paraventricular nuclei of the hypothalamus express the NMDAR2B subunit, Soc. Neurosci. Abstr. 24 Ž1998. 120. w9x M.C. Curras, ´ J. Dao, Developmental plasticity of the expression of the NR1 and NMDAR2B subunit in the supraoptic nucleus of the rat hypothalamus, Dev. Brain Res. 109 Ž1. Ž1998. 1–12.

195

w10x M.C. Curras, ´ B.S. Pallotta, Single-channel evidence for glycine and NMDA requirement in NMDA receptor activation, Brain Res. 740 Ž1996. 27–40. w11x M.C. Curras, ´ P. Rack, R.B. Meeker, Channel properties of NMDA glutamate receptors on magnocellular neuroendocrine cells cultured from the rat supraoptic nucleus, Brain Res. 789 Ž1998. 181–193. w12x F. Conti, S. DeBiasi, A. Minelli, M. Melone, Expression of NR1 and NR2ArB subunits of the NMDA receptor in cortical astrocytes, Glia 17 Ž1996. 254–258. w13x J. Dao, M.C. Curras, ´ Downregulation of NMDAR2B in the supraoptic and paraventricular nuclei of the adult rat hypothalamus following dehydration, Soc. Neurosci. Abstr. 23 Ž1997. 1248. w14x C. Decavel, M.C. Curras, ´ Increased expression of the NMDA receptor subunit, NR1, in immunohistochemically identified magnocellular hypothalamic neurons during dehydration, Neuroscience 78 Ž1. Ž1997. 191–202. w15x A. Dutton, R.E.J. Dyball, Phasic firing enhances vasopressin release from the rat neurohypophysis, J. Physiol. 290 Ž1979. 433–440. w16x R.E.J. Dyball, P.S. Poutney, Discharge patterns of supraoptic and paraventricular neurones in rats given 2% NaCl solution instead of drinking water, J. Endocrinol. 56 Ž1973. 91–98. w17x V.K. Gribkoff, Electrophysiological evidence for N-methyl-D-aspartate excitatory amino acid receptors in the rat supraoptic nucleus in vitro, Neurosci. Lett. 131 Ž1991. 260–262. w18x M.S. Harbuz, D.S. Jessop, S.L. Lightman, H.S. Chowdrey, The effects of restraint or hypertonic saline stress on corticotrophin-releasing factor, arginine vasopressin, and proenkephalin A mRNAs in the CFY, Sprague–Dawley and Wistar strains of rat, Brain Res. 667 Ž1994. 6–12. w19x G.I. Hatton, W.E. Armstrong, W.A. Gregory, Spontaneous and osmotically-stimulated activity in slices of rat hypothalamus, Brain Res. Bull. 3 Ž1978. 497–508. w20x G.I. Hatton, Emerging concepts of structure–function dynamics in adult brain: the hypothalamo-neurohypophysial system, Prog. Neurobiol. 34 Ž1990. 437–504. w21x N. Hawrylak, J.C. Fleming, A.K. Salm, Dehydration and rehydration selectively and reversibly alter glial fibrillary acidic protein immunoreactivity in the rat supraoptic nucleus and subjacent glial limitans, Glia 22 Ž1998. 260–271. w22x B. Hu, C.W. Bourque, NMDA receptor-mediated rhythmic bursting activity in rat supraoptic nucleus neurones in vitro, J. Physiol. 458 Ž1992. 667–687. w23x N. Hussy, L. Boissinagasse, P. Richard, M. Desarmenien, NMDA receptor properties in rat supraoptic magnocellular neurons: characterization and postnatal development, Eur. J. Neurosci. 9 Ž7. Ž1997. 1439–1449. w24x T. Ishii, K. Moriyoshi, H. Sugihara, K. Sakurada, H. Kadotani, M. Yokoi, C. Akazawa, R. Shigemoto, N. Mizuno, M. Masu, Molecular characterization of the family of the N-methyl-D-aspartate receptor subunits, J. Biol. Chem. 268 Ž1993. 2836–2843. w25x C.W. Jones, B.T. Pickering, Comparison of the effects of water deprivation and sodium chloride imbibition on the hormone content of the neurohypophysis of the rat, J. Physiol. 203 Ž1969. 449–458. w26x A.M. Khan, M.C. Curras, J. Dao, F.A. Jamal, C. Turkowski, R.K. Goel, E.R. Gillard, S.D. Wolfsohn, B.G. Stanley, Lateral hypothalamic NMDA receptor subunits NR2A andror NR2B mediate eating: immunohistochemicalrbehavioral evidence, Am. J. Physiol., in press. w27x N.W. Kleckner, R. Dingledine, Requirement for glycine in activation of NMDA-receptors expressed in Xenopus oocytes, Science 241 Ž1988. 835–837. w28x A. Kuryatov, B. Laube, H. Betz, J. Kuhse, Mutational analysis of the glycine-binding site of the NMDA receptor: structural similarity with bacterial amino acid-binding proteins, Neuron 11 Ž1994. 1049– 1056. w29x T. Kutsuwada, N. Kashiwabuchi, H. Mori, K. Sakimura, E. Kushiya, K. Araki, H. Meguro, H. Masaki, T. Kumanishi, M. Arakawa, M.

196

w30x

w31x

w32x w33x

w34x

w35x

w36x

w37x w38x w39x

w40x

w41x

w42x

w43x

M.C. Curras-Collazo, J. Dao r Molecular Brain Research 70 (1999) 187–196 ´ Mishina, Molecular diversity of the NMDA channel, Nature 358 Ž1992. 36–41. M. Lafarga, M.T. Berciaqno, E. del Olmo, M.A. Andres, A. Pazos, Osmotic stimulation induces changes in the expression of betaadrenergic receptors and nuclear volume of astrocytes in supraoptic nucleus of the rat, Brain Res. 588 Ž1992. 311–316. B. Laube, H. Hirai, M. Sturgess, H. Betz, J. Kuhse, Molecular determinants of agonist discrimination by NMDA receptor subunits: analysis of the glutamate binding site on the NR2B subunit, Neuron 18 Ž1997. 493–503. J.M. Luque, J.G. Richards, Expression of NMDA 2B receptor subunit mRNA in Bergman glia, Glia 13 Ž1995. 228–232. R.B. Meeker, M.C. Curras, ´ J. Stewart, A. Serje, W. Al-Ghoul, Functional activation of punch-cultured magnocellular neuroendocrine cells by glutamate receptor subtypes, J. Neurosci. Methods, Žin press.. R.B. Meeker, R.S. Greenwood, J.N. Hayward, Glutamate is the major excitatory transmitter in the supraoptic nuclei. The Neurohypophysis: a window on brain function, Ann. N.Y. Acad. Sci. 689 Ž1993. 636–639. R.B. Meeker, S. McGinnis, R.S. Greenwood, J.N. Hayward, Increased hypothalamic glutamate receptors induced by water deprivation, Neuroendocrinology 60 Ž1994. 477–485. H. Monyer, R. Sprengel, R. Sshoepfer, A. Herb, M. Higuchi, H. Lomeli, N. Burnashev, B. Sakmann, P.H. Seeburg, Heteromeric NMDA receptors: molecular and functional distinction of subtypes, Science 256 Ž1992. 1217–1221. S.H.R. Oliet, C.W. Bourque, Mechanosensitive channels transduce osmosensitivity in supraoptic neurons, Nature 364 Ž1993. 341–343. P. Paoletti, P. Ascher, Mechanosensitivity of NMDA receptors in cultured mouse central neurons, Neuron 13 Ž1994. 645–655. L.S. Perlmutter, C.D. Tweedle, G.I. Hatton, Neuronrglial plasticity in the supraoptic dendritic zone in response to acute and chronic dehydration, Brain Res. 361 Ž1985. 225–231. R. Petralia, Y. Wang, R. Wenthold, The NMDA receptor subunits NR2A and NR2B show histological and ultrastructural localization patterns similar to those of NR1, J. Neurosci. 14 Ž10. Ž1994. 6102–6120. R. Petralia, N. Yokotani, R. Wenthold, Light and electron microscope distribution of the NMDA receptor subunit NR1 in the rat nervous system using a selective anti-peptide antibody, J. Neurosci. 14 Ž1994. 667–696. D.A. Poulain, D. Brown, J.M. Wakerley, Statistical analysis of patterns of electrical activity in vasopressin and oxytocin-secreting neurones, in: G. Leng ŽEd.., Pulsatility in Neuroendocrine Systems, CRC Press, FL, 1988, pp. 119–154. C.D. Sladek, K.Y. Fisher, H.E. Sidorowicz, J.R. Mathiasen, Osmotic stimulation of vasopressin mRNA content in the supraoptic nucleus requires synaptic activation, Am. J. Physiol. 268 Ž1995. R1034– R1039.

w44x L. Snell, K. Nunley, R. Lickteig, M. Browning, B. Tabakoff, P. Hoffman, Regional and subunit specific changes in NMDA receptor mRNA and immunoreactivity in mouse brain following chronic ethanol ingestion, Mol. Brain Res. 40 Ž1. Ž1996. 71–78. w45x M. Sonders, P. Bartmettler, J. Lee, Y. Kitahara, J. Keana, E. Weber, A novel photoaffinity ligand for the phencyclidine site of the N-methyl-D-aspartate receptor labels a Mr 120,000 polypeptide, J. Biol. Chem. 205 Ž1990. 6776–6781. w46x N.J. Sucher, M. Awobuluyi, Y.-B. Choi, S.A. Lipton, NMDA receptors: from genes to channels, TIPS 17 Ž1996. 348–355. w47x D.T. Theodosis, D.A. Poulain, J.D. Vincent, Possible morphological bases for synchronisation of neuronal firing in the rat supraoptic nucleus during lactation, Neuroscience 6 Ž1981. 919–929. w48x C.D. Tweedle, G.I. Hatton, Ultrastructural comparisons of neurons of supraoptic and circularis nuclei in normal and dehydrated rats, Brain Res. Bull. 1 Ž1976. 103–121. w49x C.D. Tweedle, G.I. Hatton, Ultrastructural changes in rat hypothalamic neurosecretory cells and their associated glia during minimal dehydration and rehydration, Cell Tissue Res. 181 Ž1977. 59–72. w50x A.N. van den Pol, J.-P. Wuarin, F.E. Dudek, Glutamate, the dominant excitatory transmitter in neuroendocrine regulation, Science 250 Ž1990. 1276–1278. w51x Y. Wang, T. Bosy, R. Yasuda, D. Grayson, S. Vicini, T. Pizzorusso, B. Wolfe, Characterization of NMDA receptor subunit-specific antibodies: distribution of NR2A and NR2B receptor subunits in rat brain and ontogenic profile in the cerebellum, J. Neurochem. 65 Ž1. Ž1995. 176–183. w52x M. Watanabe, Y. Inoue, K. Sakimura, M. Mishina, Distinct distributions of 5 N-methyl-D-aspartate receptor channel subunit messenger RNAs in the forebrain, J. Comp. Neurol. 338 Ž3. Ž1993. 377–390. w53x J.-P. Wuarin, F.E. Dudek, Excitatory amino acid antagonists inhibit synaptic responses in the guinea pig hypothalamic paraventricular nucleus, J. Neurophysiol. 65 Ž1991. 946–951. w54x Z. Xu, J. Herbert, Effects of intracerebroventricular dizocilpine ŽMK801. on dehydration-induced dipsogenic responses, plasma vasopressin and c-fos expression in the rat forebrain, Brain Res. 784 Ž1998. 91–99. w55x K. Yagita, H. Okamura, Y. Ibata, Rehydration process from saltloading: recovery of vasopressin and its coexisting galanin, dynorphin, and tyrosine hydroxylase immuno-reactivities in the supraoptic and paraventricular nuclei, Brain Res. 667 Ž1994. 13–23. w56x C.R. Yang, V.V. Senatorov, L.P. Renaud, Organum vasculosum lamina terminalis-evoked postsynaptic responses in rat supraoptic neurones in vitro, J. Physiol. 477 Ž1994. 59–74. w57x Q.Z. Yang, K.G. Smithson, G.I. Hatton, NMDA and non-NMDA receptors on rat supraoptic nucleus neurons activated monosynaptically by olfactory afferents, Brain Res. 680 Ž1995. 207–216. w58x R.L. Zerbe, M. Palkovits, Changes in the vasopressin content of discrete brain regions in response to stimuli for vasopressin secretion, Neuroendocrinology 38 Ž1984. 285–289.