Mediobasal hypothalamic neurons are excited by the iontophoretic application of sodium

Brain Research, 273 (1983) 35-44

35

Elsevier

Mediobasal Hypothalamic Neurons are Excited by the Iontophoretic Application of Sodium I. MANDELBROD, S. FELDMAN and R. WERMAN*

Laboratory of Neurophysiology, Department of Neurology, Hadassah University Hospital and Neurobiology Unit, Institute of Life Sciences, Hebrew University of Jerusalem (Israel) (Accepted January 4th, 1983)

Key words: Sodium - rat mediobasal hypothalamus - - iontophoresis - - extracellular recording - - cortisol - - osmoreception

In the course of studies on the responsiveness of mediobasal hypothalamic neurons to the iontophoretic application of cortisol, it was found that positive currents applied to a sodium chloride (1 M) barrel alone, but not to a choline chloride (1 M) barrel, frequently increased the firing of these neurons. Subsequently, systematic examination demonstrated that out of 102 MBH neurons 52 (51%) increased their firing by at least 30% with application of NaCI, using currents no greater than 10 nA. No such effect was obtained in response to Na application from a dilute solution (0.05 or 0.1 M). When glutamate was absent from the electrodes, the incidence of Na ÷ sensitivity fell to 17%, despite the routine use of backing currents to the glutamate barrel. K÷ ions were more active than Na + ions in producing excitation. When Na + sensitivity was found, however, Na ÷ effects were produced by currents greater than K÷ currents producing equivalent excitation. Like glutamate, K÷ ions were capable of greatly enhancing responses to Na ÷. Comparison was made between cortisol and Na ÷ sensitivity in 70 MBH neurons; 28 cells responded to both, and 24 of them were inhibited by cortisol. Thus Na ÷ sensitivity is a frequent characteristic of MBH neurons inhibited by cortisol, and was present in 83% of cortisol-sensitive cells in this region. Iontophoresis of Na ÷ is commonly used as a control in pharmacological studies of the nervous system. Even more common is the case of concentrated NaCI solutions for recording. These procedures may not be as inert as previously thought, particularly in the hypothalamus. INTRODUCTION The site of neurohypophyseal osmoreceptors related to the secretion of vasopressin has been a topic of some controversy. Though the supraoptic nucleus has been primarily implicated, there is also evidence that osmoreceptors may be situated outside this region 27. In one series of experiments the induction of phasic activity in neurosecretory cells by intraperitoneal injection of hypertonic saline was abolished after complete or anterior hypothalamic deafferentation a. O n the other hand, in in vivo experiments, direct application of hypertonic saline solution by microtap to antidromically identified supraoptic neurons evoked a characteristic phasic pattern of firing 14. F u r t h e r m o r e , in in vitro studies, in which intracellular recordings were obtained from supraoptic neurons in a rat hypothalamic slice preparation, it * To whom correspondence should be sent, 0006-8993/83/$03.00 © 1983 Elsevier Science Publishers B.V.

was found that increasing the extracellular osmotic pressure produced m e m b r a n e depolarization, increase in the frequency of excitatory postsynaptic potentials, and increased firing 15,24. Thus, it is evident that supraoptic n e u r o n s respond directly to changes in the osmolarity of their immediate environment. Since these cells also control the rate of secretion of vasopressin, they are clearly part of the central osmoreceptive mechanism 15. The earliest studies on the responsiveness of individual hypothalamic n e u r o n s to osmotic stimuli were conducted more than 20 years ago 6. It was found that the firing of supraoptic n e u r o n s in rabbits was generally accelerated by intracarotid injections of hypertonic sodium chloride. Many cells in the vicinity of the supraoptic and paraventricular nuclei - - the preoptic, anterior and dorsal hypothalamic areas - - also responded to hypertonic stimuli. In subsequent stud-

36 ies in cats, it was found that supraoptic neurons augmented their rate of firing following the intracarotid injection of hypertonic sodium chloride s,l~. More recently it was found that antidromically identified supraoptic neurons in the rat and m o n k e y were activated following the intracarotid injection of hypertonic solutions of NaCI 4,7,jl. Many of the osmosensitive cells in cats, however, were in the region of the anterior and lateral hypothalamic areas, superior to the supraoptic nucleus I3. Also, in experiments in the rat, lateral preoptic, lateral hypothalamic and dorsal midbrain neurons r e s p o n d e d by an increase in firing to the infusion or subcutaneous injection of hypertonic saline solutions 3.17.31 . Thus it is evident that there are neurons that respond to hypertonic NaC1 outside the supraoptic nucleus. The purpose of this p a p e r is to describe the presence of sodium-sensitive neurons in the mediobasal hypothalamus ( M B H ) , and their possible relation to cortisol-sensitive cells in this area, which plays a m a j o r role in corticosteroid regulation. A preliminary report of this work has appearedel. MATERIALS AND METHODS Experiments were p e r f o r m e d on male albino rats of the H e b r e w University Strain, weighing 220-270 g and anesthetized with urethane, 150 mg/100 g. A f t e r exposure of the cerebral cortex, a 5-barrelled micropipette with a tip d i a m e t e r of 1-2/~m was introduced stereotactically into the tuberal hypothalamus, 0.2--0.4 m m from the midline. The micropipette was driven by a hydraulic microdrive in micron steps and controlled outside the F a r a d a y cage. The small size of the electrode tip a p p a r e n t l y allowed relatively close approximation of the electrode tip to the neurons studied; this is reflected in the low iontophoretic currents used throughout (see below).

The central barrel of the electrode contained 3 M NaCI and was used for extracellutar recording, while the other barrels contained 1 M NaCI, as well as other compounds including 1 M choline chloride, 0.1 M glutamate, p H 8.0, 1 M KCI, 0.05 M cortisol sodium succinate (Solu-Cortef), pH 7.1/. The cortisol was always administered as an anion. The materials were applied by electronically controlled constant-current sources. Current was measured separately for each barrel as the voltage drop on a series resistor. The recording barrel was connected to a P 5 l l Grass preamplifier through a cathode follower. Single cell firing was displayed on an oscilloscope before, during and after the iontophoretic application of the agents tested. In addition, the amplified potentials were also fed through a window discriminator ( C A T model 607) to a C A T 1000 c o m p u t e r for counting and through an integrator especially designed for these experiments 16. The figures in this p a p e r show the firing each second as a function of time, derived from that integrator. All agents were tested for periods Of 5 (very obvious actions) to 60 s and each agent tested was applied 4 or 5 times to a given cell in o r d e r to determine sensitivity. Changes in firing rate of 30% or greater were considered significant*. In general, iontophoretic currents less than 10 n A were used; only rarely were currents as large as 15 n A (considered a small current in most CNS iontophoretic work) used. A large n u m b e r of the neurons in this region are inactive in the absence of stimulation. All of these cells and many of the cells with low rates of spontaneous activity ( < 3 Hz) were tested in the presence of background glutamate currents. This allowed us to detect inhibitory influences of the substances tested. We have already shownlg, 20 that such use of glutamate does not affect the nature of the response to cortisol. Similarly, except in the case of inhibition of

Fig. 1. Sodium ions but not choline ions increase the firing of a mediobasal hypothalamic neuron. A: oscilloscope record. B: ratemeter record. This silent cell was tested with a background of glutamate iontophoresis (2 nA), producing a firing rate of 4-5 Hz. Sodium ions from a sodium chloride barrel (1 M; 5 nA) produced, within 4-5 s, an increase in firing of 7-8 Hz. When a larger positive current (8 nA) was applied to a choline chloride barrel (1 M), a small and insignificant decrease in firing was seen. In this and subsequent figures, the ratemeter shows firing rate each second, each line (or dot) above the baseline representing 1 spike/s. * The stringent critical ratio (CR) = (E-T)/(E+T) ~ used by us in previous papers2°,22, can be applied to these data. The critical value of (CR) > I 1.96 I is obtained for a 30% change in firing when the control sample contains 99 spikes in the case of excitation. The requirement for significance falls rapidly as the percentage of change increases; thus for a 100% excitation, only 12 control spikes suffice.

3'7

G (2)

Na (,5)

Ch ( 8 )

I

i'

~

~,.~, . ~'

..

,

r

,~ , ~ ' , ' . ~

t.

~4,1,..e~'

! ~ t .~, ,'q ,~, ~ I~, W'~.,,,. ,.'~ .~".. , , " ~ , "

2 0 sec

II

• ~,

, .!,. If

g'e

.P,

.

~ ,

.t~,o, lr. . . . .

r

~

,1~

,,~II'~

~,

.~.

,! ~ ~ ¢~,,,

,

I~

38 silent cells which cannot otherwise be detected, small doses of glutamate did not affect the nature of responses to putative neurotransmitters tested 23. At the end of the experiments, a stainless steel electrode was lowered stereotactically to the highest point of the region studied and a DC electrolytic lesion was placed there. The brain was fixed in formalin and, following histological examination, the location of each recording and stimulating site was determined by calculating the calibrated movements of the micromanipulator as related to this point. RESULTS When positive currents through a barrel containing NaCI (1 M) were simultaneously applied to neutralize the negative currents through another barrel containing cortisol sodium succinate (0,05 M), the inhibitory action of the cortisol applied alone 20 was seen to be suppressed in 3 consecutive MBH neurons. The suspicion that the effects related to cortisol application might be produced by negative current was removed when it was demonstrated that these effects could not be reproduced by negative currents through either the NaCI barrel or through a barrel containing sodium succinate (0.1 M) 20. When positive currents were applied to the sodium chloride (1 M) barrel alone, increases in firing of MBH neurons were frequently seen. In the cell illustrated in Fig. 1A, application of 5 nA of positive current through the NaCI barrel, thereby expelling Na ÷ ions, produced a moderate but definite increase in frequency of firing. In the ratemeter record of Fig. 1B, it can be seen that the background rate of firing, produced by the constant application of glutamate (2 nA) was 4--5 Hz and was increased within 4-5 s to 7-8 Hz by the iontophoresis of Na ÷ ions. When a larger dose of choline ions was applied with a higher positive current (8 nA) from a choline chloride barrel (1 M), no apparent increase in firing was seen (Fig. 1A), and the ratemeter record (Fig. 1B) shows a small and insignificant decrease in firing. Subsequent to this finding, that sodium ions but not choline ions can increase the firing of MBH neurons, which was repeated a number of times, 102 neurons in this region were systemically examined for sodium sensitivity. Of these cells, 52 (51.0%) increased their firing by at least 30% with applications

of positive currents through an NaC1 (1 M) barrel no greater than 10 nA. The rise in frequency generally appeared after a few seconds of current application and was never immediate. In general, the increase in frequency outlasted the current application, but never by more than 10 s (Figs. 1 and 3). These findings, together with the absence of response to injection of choline ions, indicate that current was not responsible for the increased firing. Table I shows the response of these neurons when the cells are separated into silent, slow (~<3 Hz) and fast ( > 3 Hz) units according to their behavior in the absence of glutamate application. The highest percentage of responsive neurons was found in the slow group (57.8%), but more than 40% of the neurons in each of the other two groups, silent (46.7%) and fast (41.7%), also exhibited sensitivity to sodium iontophoresis. It is known from the Goldman equation10,12, 3e that increase in monovalent cation concentrations on the outside of a neuron should produce depolarization of the cell and increased firing. Since neurons at rest are far more permeable to potassium ions than to sodium ions, the ability of high external concentrations of potassium ions to produce depolarization is marked, while that of sodium ions is generally considered to be almost negligible. In order to evaluate the relative effects of potassium and sodium ions on the sodium-sensitive neurons of the MBH, a number of experiments were carried out comparing the effects and interactions of potassium and sodium iontophoresis on the same cell (Figs. 2 and 3). The silent cell of Fig. 2 was found to respond to equivalent doses of Na ÷ about one-third as well as to K + ions. It is not likely, however, that this relationship represents the ratio of Na ÷ to K ÷ permeability in these cells; for such permeability ratio TABLE I lncreased firing in MBH neurons produced by iontophoresis of sodium ions and classified according to spontaneous firing rate of the cells Initial rate (Hz )

Number of cells

Responded to sodium (%)

Silent (0) Slow (~<3) Fast (>3)

45 45 12

21 (46.7) 26 (57.8) 5 (41.7)

Total

102

52 (51.0)

39 G(-2} Con(t0) Na (O)

(5)

!

!

f

t.~

I

(-2)

K (O)

(5)

i

(-2)

~Pf~f~,f

I

ri

|

' ,

,

,.,-,...,,.,,,

20 sec

Fig. 2. An MBH neurons is more sensitive to K+ than to Na ÷ ions. Silent cell examined with braking current (--2 nA) applied to glutamate barrel. Na ÷ ions (5 nA) produce a firing rate of approximately 1 Hz, while the same dose of K÷ ions produces a rate of approximately 3 Hz. A larger positive current (I0 nA) through a dilute Na ÷ ion solution (Con = sodium cortisolsuccinate, 0.05 M) fails to produce firing. the Goldman equation predicts that the neuron at rest would be depolarized by about 40 mV compared to cells with more normal ratios (closer to one-tenth of that found). Only the presence of an unusually powerful electrogenic pump, hyperpolarizing the cell, would allow the neuron to function normally. Thus the effects of Na + ions cannot be attributed only to an increased permeability to them. Another important finding is illustrated in Fig. 2, the absence of response to Na ÷ ions when applied from a dilute solution. Although clear firing was produced in this cell by only 5 nA from a 1 M solution, twice that dose from a 0.05 M solution, one-twentieth the concentration, failed to produce firing (Con (10)). This finding was consistent (Fig. 3); the transport number of Na ÷ in the dilute solution is probably appreciably smaller than one-half of that in the concentrated solution. Indeed the cortisol compound used in the dilute barrel may be poorly ionized, further lowering the transport number of Na +. TABLE II

Comparison of sodium ion and cortisol sensitivities of MBH neurons Response to cortisol Number of cells Responsive to Na+(%) Inhibition Excitation None

24 4 42

20 (83.3) 1 (25.0) 17 (40.5)

Total

70

38 (54.3)

A comparison was made between sodium ion and cortisol sensitivities of M B H neurons in 70 cells (Table II). Of these neurons (54.3%), only slightly more than the larger group in Table I, were sensitive to Na ÷ ion iontophoresis. Thus, the group can be considered to be representative. Of the 28 cortisol-sensitive neurons, 21 (75.0%) were also excited by Na ÷ ions. Particularly noteworthy was the sodium sensitivity of 20 of the 24 neurons that were inhibited by cortisol (83.3%). Since the percentage of MBH neurons inhibited by cortisol is only 40.7% 20 (40.0% in this smaller sample), the incidence of sodium sensitivity among cortisol-sensitive cells expected by chance alone is only 20.8%, less than one-fourth the percentage actually found. Thus it is clear that Na ÷sensitivity is an important characteristic of MBH neurons inhibited by cortisol; alternatively, 20 of 38 Na+-sensitive MBH neurons (52.6%) were inhibited by cortisol. It is of interest that 25 out of 32 (78.1%) of neurons that did not respond to sodium also did not respond to cortisoi. Glutamate is known to produce depolarization in MBH neurons19,23; among the probable mechanisms of its action is to increase the neuron membrane permeability to Na + ions 33. It is therefore possible that the excitation mediated by Na + ions seen by us was mediated by the action of small amounts of glutamate leaking from the electrodes even in the face of backing currents. We then examined the effect of iontophoresis of Na ÷ ions from multi-barrelled electrodes that were prepared without glutamate. Of 12 MBH

40

A

neurons studied under these conditions, only 2 (16.7%) were excited by the application of Na + ions. The cells examined in the absence of glutamate did not include silent cells which could only be identified

K(-2) Na(5), (10)

by use of the amino acid. It can be seen (Table 1), however, that elimination of silent cells in the presence of glutamate should actually increase the percentage of Na+-sensitive n e u r o n s from 51.0% to 54.4% (31 of 57 cells). It is thus clear from the small percentage of Na+-sensitive n e u r o n s found in the ab-

f II

sence of glutamate, that, while not necessary to dem-

o

onstrate Na + sensitivity, presence of the amino acid augments the response to Na + ions.

K(5) Na(5)~ (I0) •

f , .

Glutamate is not the only depolarizing agent to augment the excitatory action of Na + ions; K + ions also share this property. This can be clearly seen in the experiments of Fig. 3, on the two Na+-sensitive cells found in the absence of glutamate (the effect was also seen in the presence of glutamate). In the

,o

IIOp,aW,J~o~fe, F l m ~ e e m J m f

very slow n e u r o n of Fig. 3A (spontaneous firing < 0.1 Hz), when a braking current ( - - 2 n A ) was applied to the KCI barrel, 5 n A of positive current applied to the NaCI barrel failed to produce any firing and 10 n A only produced two spikes. In this n e u r o n , 5 n A positive current applied to the K + barrel pro-

N S). (e)

duced occasional spiking; now, however, a superim~ W~,l f l l a . ~ . ~ e l l a R , / . . ' . , , . . ~ , ~

~ I

posed positive current of 5 n A to the NaCI barrel produced steady firing at about 2 Hz. W h e n the current on the NaC1 barrel was increased to 10 n A , a further increase in firing was seen. W h e n the backgroud K +

I,

current was increased to 10 n A , continuous firing ap-

B oo)

boo)

,,.'~.

w

, , I

f

,,l~l~,pl'~,'!

f

q

, m ..

~,.

,,,.'~e',.,~,./~,,.,.,.,ee~w....,eleO

,.

Fig. 3. lontophoresis of K÷ ions augments the excitatory response to Na + ions in MBH neurons. A: very slow neuron (<0.1 Hz). Upper trace: when a braking current (--2 nA) was applied to the KCI barrel, no response was seen to 5 nA of current through the Na barrel and only a minimal response to 10 hA. Middle trace: with 5 nA positive current to the KCI barrel, occasional firing was seen; now both 5 and 10 nA applied to the NaCI barrel produced increased firing. Lower trace: 10 nA positive current applied to the KCI barrel produces steady background firing; again Na + ion iontophoresis increases firing. B: slow cell (about 2 Hz). In the absence of K+ ion iontophoresis, 10 nA positive current applied to the Na + barrel fails to produce any change in firing. With 1 nA applied to the K+ barrel, no increased firing is seen; now, however, the same current that was ineffective in the absence of K+ ions produces a marked excitation. In both cells, electrodes did not carry glutamate. NaC1and KCI, 1 M in both cases.

41 peared, but once again additional Na ÷ ion iontophoresis increased the firing rate. Similarly, in the neuron of Fig. 3B, responsiveness to Na ÷ ions was augmented by K ÷ iontophoresis. This slow neuron (about 2 Hz) was not affected by currents in the Na ÷ barrel up to 10 nA in the absence of current applied to the KCI barrel. When, however, only 1 nA positive current was applied to the KCI barrel, 10 nA of Na ÷ current now produced brisk firing (up to 8 Hz) even ihohgh the K ~ current by itself produced no obvious change. From these experiments, we can conclude that K ÷ ion iontophoresis augments the excitatory response to Na ÷ ions in MBH neurons. The results of experiments like those shown in Fig. 3 were suspiciously reminiscent of effects produced by current. We have already shown that large positive currents through choline chloride (1 M; Fig. 1) or cortisol sodium succinate (0.05 M; Fig. 2) failed to produce increased firing. We now repeated these experiments with positive current through glutamate barrels (monosodium salt, 0.1 M) and once again failed to reproduce the Na ÷ response. A further set of controls was undertaken with barrels containing 0.1 M NaC1, close to normal saline. No cell that responded to positive current through the 1 M NaCI barrel (20 cells) responded to the same or larger currents through the 0.1 M NaCI barrel of the same electrode assembly. We did find, however, that it was possible, with large (10-15 nA) positive currents applied to either 0.1 M NaCI or 0.05 M cortisol barrels during the waning of the response to currents through the 1 M NaCI barrel, to sometimes obtain a definite but weak augmentation of the Na ÷ ion response. These augmentations could not be obtained using similar currents through a choline chloride electrode nor after periods greater than 3 s following offset of current through the 1 M NaCI barrel. It is thus probable that positive currents up to 15 nA do not produce significant depolarizations in the neurons examined. Moreover, the effects described cannot be attributed to current flow. Since the common feature of glutamate and K + action on neurons is depolarization and since both agents produce augmentation of the effect of Na + ion iontophoresis, it is fair to conclude that Na+ ions actions are augmented by depolarization. It is also likely that the action of Na ÷ ions is to produce depolarization of MBH neurons and thereby increase firing.

DISCUSSION The excitatory action of Na + ions on MBH neuron is very prominent; Na ÷ ions excited 51.0% (52 of 102 cells) of MBH neurons tested. In a study of 8 putative neurotransmitters carried out in this laboratory 23, only glutamate, which excited all neurons tested, was more potent as an excitatory agent. Less active were histamine which excited 47.9% (34 of 71 cells), norepinephrine 11.1% (3 of 27 cells) and acetylcholine 10.5% (8 of 76 cells). Furthermore, Na ÷ ions never inhibited the firing of MBH neurons, unlike these other 3 excitatory agents which all showed prominent inhibitory actions in the MBH: histamine, 31.0% of neurons tested; norepinephrine, 66.7% and acetylcholine, 43.4%. Although it is true that the Na + ion effects were much less prominent in the absence of glutamate, with only 2 of 12 cells (16.7%) responding, it should be pointed out that all of the agents were tested in the presence of glutamate. It is probable that any augmentation caused by leakage of glutamate would similarly enhance the excitatory actions of the other agents. Not including the non-selective action of glutamate, Na + ions are the most potent excitatory agent found in the MBH. When cortisol sensitivity of M B H neurons was compared to Na ÷ sensitivity, 21 of 28 cortisol-sensitive neurons (75.0%) were found to be excited by Na ÷ ions. In our study of neurotransmitter candidates 23, histamine was found to be less effective than Na + ions: 18 out of 31 cortisol-sensitive neurons (58.1%) were excited by histamine. Acetylcholine excited 8 of 32 neurons (25.0%) sensitive to cortisol, while norepinephrine excited 3 of 13 (23.1%) cortisol-sensitive neurons. Once again, unlike Na ÷ ions, the 3 putative excitatory neurotransmitters also produced prominent inhibitory actions: histamine inhibited 25.8% of cortisol-sensitive neurons; acetylcholine 50.0%; and norepinephrine 61.5%. In our previous study 23, we singled out a small group, including about 10% of MBH neurons, which were felt to be likely candidates for CRF-releasing cells. These MBH neurons were characterized by their responses to cortisol, acetylcholine and histamine; they were inhibited by cortisol and excited both by acetylcholine and histamine. Of particular interest was the observation that all of the neurons excited by acetylcholine were members of this group.

42 Unfortunately we do not have direct information on the Na + sensitivity of M B H neurons excited by ACh. It should be noted, however, that 20 of 24 (83.3%!) of neurons inhibited by cortisol were excited by Na + ions. It is thus very likely that our C R F candidate neurons are also excited by Na + ions. Although the results indicate a clear response to Na + ions in many (51%) M B H neurons, they do not tell us whether we are examining Na + sensitivity, per se, or osmoreceptivity. While the iontophoresis of choline ions from a 1 M choline chloride barrel provides a clear control for the effects of current, it does not provide as good a control for osmoreceptive effects. First, it is likely that the transport number of choline ions is lower than that of Na + ions. Thus, from solutions of equal concentration, a larger fraction of the current will be carried by Na + ions, the rest being carried by hydronium ions leaving the barrel or small anions entering the barrel. Furthermore, choline is known to be taken up actively by brain cells 30, although this process is probably too slow to account for a major reduction in local choline concentration during the iontophoresis. Finally, it should be pointed out that osmorecep' tion is rather selective and that such agents as glucose and galactose are ineffective osmostimulators, while sucrose and mannitol are effective 24,25,28,29. We can find no report of the effect of hypertonic choline on osmoreception. Thus, while we can report the presence of Na + ion sensitivity in M B H neurons, we can neither confirm nor reject the presence of non-ionic osmoreceptive properties in these cells. The physiological significance of Na + ion excitation of M B H neurons is not clear. Strong augmentation of iontophoretic effects of Na + ions by both glutamate and K + ions does not argue against a physiological role for the Na + sensitivity. The cortisol sensitivity of most of Na+-sensitive neurons (52.6%) is striking, but the significance of cortisol sensitivity in the M B H is not clear in most cases. If we accept the contention that about 10% of the M B H population consists of cortisol-inhibited neurons responsible for CRF release 23, and that another 9%, those cells excited by cortisol, are inhibitory interneurons to C R F cells 20, we are still left with the majority of cortisolsensitive M B H neurons, 30% of M B H neurons20 whose role is not at all clear. It would be of interest to determine by additional tests, particularly the use of

non-permeable non-ionic hypertonic stimulants, whether Na+-sensitive neurons are indeed osmoreceptors. As such they may indeed represent part of the population of osmoreceptors outside of the supraoptic nucleus postulated by others 3.6,~.~3.w,27,31 . It is interesting in this context that Dyball and Prilusky s found that synaptic phasic activity following intraperitoneal hypertonic saline was abolished by hypothalamic deafferentation. Furthermore, intracellular recording in vitro from supraoptic neurons exposed to hypertonic solutions showed, in addition to direct effects on the neurons studied, signs of increased presynaptic activity 15,24. It is not clear whether the presynaptic activity seen reflects the well known sensitivity to hypertonic stimuli of presynaptic terminals 9, but this seems hardly likely as the effect is already prominent with as little as 8 mosmol.1 -I hypertonicity2q far less than is necessary to stimulate neuromuscular terminals (Werman, unpublished observations). Alternatively, the presynaptic activity may reflect a specialized osmosensitive property of cells presynaptic to the supraoptic neurons, some of which may also be the Na+-sensitive neurons found in the M B H in the present study. In a survey of the literature on iontophoretic studies in the hypothalamus, we have found few reports which actually denote responses or failure of responses to Na + ions or illustrate Na + ion iontophoresis. O o m u r a et al. 26 reported that none of 15 cells tested in the ventromedial hypothalamus responded to Na + ions. In the lateral hypothalamic region, however, they found that 5 of 24 neurons (20.8%) were excited by Na + ions, while one cell was inhibited. Na + ions were ejected from 0.2 M NaCI barrels. They do not report use of glutamate but the recording barrel contained 5 M NaCI. Beckman and Eisenman 2, in a study of thermosensitive neurons in the hypothalamus, illustrate Na + ion iontophoresis in two neurons: in one, 100 nA (from 0.2 M NaCI solution) clearly produced a current effect; in the other, the same dose failed to produce any change. NaCI (4 M) was used for recording. Barone et al. 1 found one of 25 neurons (4.0%) in the lateral preoptic area and 4 of 20 (20%) in the lateral hypothalamic area excited by Na ÷ ions; no neurons were inhibited by Na+; glutamate was present and Na + was ejected from 0.4 M NaC1 solutions. Recording was carried out from a 2 M NaCi barrel. No other studies provide direct infor-

43 marion on the Na + sensitivities of hypothalamic neu-

be less inert than previously thought, particularly

rons. The m a n y reports in which Na ÷ influence is not m e n t i o n e d , although presumably tested, are of ques-

when using glutamate, or another strong depolari-

tionable help. As we have pointed out, we cannot differentiate between Na ÷ sensitivity and osmoreceptive effects in

correct for investigations in the hypothalamus, and may also be relevant elsewhere in the nervous sys-

zing agent in another barrel. This statement is clearly

tem.

our studies. Although specific Na ÷ receptors, in addition to the more generally accepted brain osmoreceptors have been postulated in brain 25, this has been

ACKNOWLEDGEMENTS

strongly challenged29. In all events, our finding of Na + sensitivity introduces some important caveats

Supported in part by the Lena P. Harvey Endow-

that should be brought to the attention of investigators recording with concentrated NaC1 solutions or

ment F u n d for Neurological Research (to S.F.) and

using Na ÷ ions as a 'control'. These electrodes may

F o u n d a t i o n (to R.W.).

REFERENCES

584-590. 14 Leng, G., Rat supraoptic neurones: the effects of locally applied hypertonic saline, J. Physiol. (Lond.), 304 (1980) 405-414. 15 Leng, G., Mason, W. T. and Dyer, R. G., The supraoptic nucleus as an osmoreceptor, Neuroendocrinology, 34 (1982) 75-82. 16 Mahler, Y., Leng, M. and Mandelbrod, I., Analogue counter integrator for brain-cell unit activity counting, Med. Biol. Eng., 14 (1976) 51-55. 17 Malmo, R. B., Osmosensitive neurons in the rat's dorsal midbrain, Brain Research, 105 (1976) 105-120. 18 Malmo, R. B. and Mundl, W. J., Osmosensitive neurons in the rat's preoptic area: medial-lateral comparison, J. comp. Physiol., 88 (1975) 161-175. 19 Mandelbrod, I., Feldman, S. and Werman, R., Effects of microelectrophoretically administered cortisol on single cell activity in the tuberal hypothalamus of the rat, Isr. J. Med. Sci., 8 (1973) 1058--1061. 20 Mandelbrod, I., Feidman, S. and Werman, R., Inhibition of firing is the primary effect of microelectrophoresis of cortisol to units in the rat tuberal hypothalamus, Brain Research, 80 (1974) 303-315. 21 Mandelbrod, I., Feldman, S. and Werman, R., Effect of iontophoretically applied Na ÷ on neuronal firing rates in the tuberal hypothalamus of the rat, Isr. J. Med. Sci., 11 (1975) 70. 22 Mandelbrod, I., Feldman, S. and Werman, R., Modification of responses to sensory and hippocampal stimuli in neurons of the rat mediobasal hypothalamus in the presence of iontophoretically applied cortisol, Brain Research, 218 (1981) H5-130. 23 Mandelbrod, I., Feldman, S. and Werman, R., The effects of iontophoretic application of suspected neurotransmitters on the electrical activity of rat mediobasal hypothalamic neurons in relation to their steroid sensitivity, Brain Research, in press. 24 Mason, W. T., Supraoptic neurones of rat hypothalamus are osmosensitive, Nature (Lond.), 287 (1980) 154-157. 25 McKinley, J. J., Denton, D. A. and Weisinger, R. S., Sensors for antidiuresis and thirst - - osmoreceptors or CSF sodium detectors, Brain Research, 141 (1978) 98-103.

1 Barone, F. C., Wayner, M. J., Tsai, W. H. and Barash, F. E., Effects of microiontophoretically applied chemicals on hypothalamic neural activity, Brain Res. Bull., 5 (1980) 325-332. 2 Beckman, A. L. and Eisenman, J. S., Microelectrophoresis of biogenic amines on hypothalamic thermosensitive cells, Science, 170 (1970) 334-336. 3 Blank, D. L. and Wayner, M. J., Lateral preoptic single unit activity: effects of various solutions, Physiol. Behav., 15 (1975) 723-730. 4 Brimble, K. J. and Dyball, R. E. J., Characterization of the responses of oxytocin- and vasopressin-secreting neurones in the supraoptic nucleus to osmotic stimulation, J. Physiol. (Lond.), 271 (1977) 253-271. 5 Brooks, C. McC., Ushiyama, J. and Lange, G., Reactions of neurons in or near the supraoptic nuclei, Amer. J. Physiol., 202 (1962) 487--490. 6 Cross, B. A. and Green, J. D., Activity of single neurones in the hypothalamus: effect of osmotic and other stimuli, J. Physiol (Lond.), 148 (1959) 554-569. 7 Dyball, R. E. J., Oxytocin and ADH secretion in relation to electrical activity in antidromically identified supraoptic and paraventricular units, J. Physiol. (Lond.), 214 (1971) 245-256. 8 Dyball, R. E. J. and Prilusky, J., Responses of supraoptic neurones in the intact and deafferented rat hypothalamus to injections of hypertonic NaCI, J. Physiol. (Lond.), 311 (1981) 443-452. 9 Fatt, P. and Katz, B., Spontaneous subthreshold activity at motor nerve endings, J. Physiol. (Lond.), 117 (1952) 109-128. 10 Goldman, D. E., Potential, impedance and rectification in membrane, J. gen. Physiol., 27 (1943) 37-60. 11 Hayward, J. N., Functional and morphological aspects of hypothalamic neurons, Physiol. Rev., 57 (1977) 574-658. 12 Hodgkin, A. C. and Katz, B., The effect of sodium ions on the electrical activity of the giant axon of the squid, J. Physiol. (Lond.), 108 (1949) 37-77. 13 Joynt, R. J., Functional significance of osmosensitive units in the anterior hypothalamus, Neurology, 14 (1964)

by a grant from the U.S.-Israel Binational Science

44 26 Oomura, Y. T., Ono, T., Ooyama, H. and Wayner, M. J,, Glucose and osmosensitive neurons in the rat hypothalamus, Nature (Lond.), 222 (1969) 282-284. 27 Poulain, D. A, and Wakerley, J. B., Electrophysiology of hypothalamic magnocellular neurones secreting oxytocin and vasopressin, Neuroscience, 7 (1982) 773-808. 28 Robertson, G. L., Athar, S. and Shelton, R. [_., Osmotic control of vasopressin function, In T. E. Andreoli, J. J. Grantham and F. C. Rector (Eds,), Disturbances in Body Fluid Osmolali(v, Am. Physiol. Soc., Washington, 1977, pp. 125-248. 29 Schrier, R. W., Berl, T. and Anderson, R. J., Osmotic and nonosmotic control of vasopressin release, Amer. J. Physiol., 236 (1979) F321-332.

30 Simon, J. R., Oderfeld-Nowak, B., Felton, D. C. and Aprison, M. H., Distribution of choline acetyltransferase, acctylcholinesterase, muscarinic binding, and choline uptake in discrete areas of the rat medulla oblongata, Ne~rochem. Res., 6 (1981) 497-505. 31 Weiss, C. S. and Almli, R., Lateral preoptic and lateral hypothalamic units: in search of the osmoreceptors for thirst, Physiol. Behav., 15 (1975) 713-722. 32 Werman, R., The reversal potential as a diagnostic tool in transmitter identification. In G. Pepeu, M. J. Kuhar and S. J. Enna (Eds.), Receptors for Neurotransmitter and Peptide Hormones, Raven Press, New York, 1980, pp. 21-31. 33 Zieglgansberger, W. and Puil, E. A., Actions of glutamic acid on spinal neurones. Exp. Brain Res.. 17 (1973) 35-49.