Systemic signals contribute to induced morphological changes in the hypothalmo-neurohypophysial system

Brain Research Bulletin, Vol. 33, pp. 211-218, Printed in the USA. All rights reserved.

1994

Copyright0

0361-9230/94 $6.00 + MI 1993 Pergamon Press Ltd.

Systemic Signals Contribute to Induced Morphological Changes in the Hypothalmo-neurohypophysial System G. H. BEAGLEY*’

AND

G. I. HA’ITONt

*Psychology Department, Alma College, Alma, MI 48801 j-Department of Neuroscience, University of California, Riverside, CA 92521 Received

27 October

1992; Accepted

19 July 1993

BEAGLEY, G. H. AND G. I. HATTON. Systemic signals contribute to induced morphological changes in the hypothalmosystem. BRAIN RES BULL 33(2) 211-218, 1994.-A systemic injection of hypertonic saline causes morphological changes in the supraoptic nucleus (SON) of the hypothalamus and the posterior pituitary, including withdrawal of astrocytic glial processes from among magnocellular neuroendocrine cells (MNCs) in the SON and increases in neural apposition with the basal lamina (BL) in the posterior pituitary. This study investigated whether systemic adrenalin provides a signal for these changes. Three groups of rats were given intraperitoneal injections of hypertonic (1.5 M) or normal (0.15 M) NaCl and were sacrificed 5 h after injection. One group was given no additional treatment, one group was anesthetized from prior to injection through perfusion, and one group was bilaterally adrenal-medullectomid several days prior to injections. Morphometric analysis of electron micrographs revealed rats given 1.5 M NaCl with no additional treatment showed expected changes in the SON and pituitary. Rats given 1.5 M NaCl and anesthetized showed diminished responses in the SON, and no changes in the pituitary compared to anesthetized rats given 0.15 M NaCl. No changes in withdrawal of glial processes from among MNCs or in neural apposition with the BL were seen in medullectomized animals. Thus, both anesthesia and adrenal-medullectomy interfere with signals for glial retraction and suggest that these responses are mediated by adrenalin. neurohypophysiaf

Supraoptic nucleus Ultrastructure

Posterior pituitary

Neurohypophysis

RESULTS of a study investigating the effects of a single intraperitoneal (IP) injection of 1.5 M NaCl on the ultrastructural appearance of the supraoptic nucleus (SON) and posterior pituitary (4) indicated that a large IP injection of hypertonic saline is an extremely effective stimulus for eliciting rapid structural changes in the SON and posterior pituitary. In the magnocellular neuroendocrine cells (MNCs) of the SON, several morphological changes indicated acute activation of the system. Size of the MNCs, amount of material in the nucleoli, and amount of Golgi apparatus all increased within 5 h of administering injections of 1.5 M NaCl. In the posterior pituitary, injections of 1.5 M NaCl produced large increases in neural apposition with the basal lamina (BL) and a significant decrease in number of axons enclosed by the pituicyte cytoplasm. The basis of many of the changes in the hypothalamo-neurohypophysial system’s (HNS) response to 1.5 M NaCl is thought to be retraction of astrocytic glial processes. Increase in soma-somatic apposition suggests withdrawal of the glial processes normally interposed between the MNC somata, possibly increasing communication among the cells. Increase in the size of synapses contacting MNCs, and increase in number of mul-

Quantitative morphology

Adrenalin

tiple synapses suggest that glial withdrawal in the SON may facilitate synaptic interaction in this system. In the posterior pituitary, retraction of the pituicytes allows for increased neural apposition with the BL, decreased enclosure of axons, and possibly increased access for neuropeptides to enter the bloodstream (15). Early work by Tweedle. and Hatton (28-30) and more recent studies by Smithson, Suarez and Hatton (25) and Luckman and Bicknell(21) provide evidence that glial cells in the neural lobe (pituicytes) actively retract from the BL under some conditions by withdrawing their processes. Withdrawal of glial processes has not been conclusively demonstrated in the SON. Beagley and Hatton (4), however, found high correlation between increased number of multiple synapses and decreased glial apposition with the MNCs, as well as increased size of terminal contact without increased number of terminals contacting the MNC, which argues against the possibility that the increased soma size leads to merely passive exclusion of the glial cells. The present study investigates the nature of the stimuli that signal activation of increased neuropeptide manufacture and secretion in the SON and pituicyte withdrawal in the neural lobe.

’ To whom requests for reprints should be addressed.

211

212

BEAGLEY AND I-IA‘ITON

Harbuz and Lightman (14) found neuropeptide mRNA in SON cells decreased in adrenalectomized animals; therefore, adrenalin might be a blood borne factor contributing to the HNS’s response to the 1.5 M NaCl stimulus. Hatton, Luckman and Bicknell(l6) working with pituicytes in primary culture from adult rats induced a change from amorphous to stellate morphology in response to adrenalin that was significantly larger than that induced by noradrenalin and could be blocked by p2- but not /?l-antagonists. The change in shape of pituicytes is consistent with the shape change seen in electron micrographs of posterior pituitaries from animals stimulated with 1.5 M NaCl. Data from in vitro experiments (16,21) suggest the glial receptors in the neural lobe are /?2-adrenergic receptors which have a higher affinity for adrenalin than noradrenalin. These studies, then, suggest that adrenalin might be a systemic signal causing glial projections to retract and allow increased apposition between MNCs in the SON upon osmotic opening of the blood-brain barrier and increased terminal apposition with the BL in the posterior pituitary where there is no such barrier. Adrenalin release is triggered from the adrenal medulla by the sympathetic nervous system in response to a variety of stressful stimuli and acts on a variety of targets, for example, skeletal muscles, blood vessels, heart and lungs, to enable the animal to engage in the fight or flight response as described by Cannon (27). Researchers such as Lightman and Young (19,20) and Kasting (17) have used large doses of hypertonic saline as a stressor to produce changes in the HNS. It would be expected, therefore, that 1.5 M NaCl at 18 ml/kg would cause a sympathetic release of adrenalin. In the present studies, we attempted to remove the adrenalin signal by two methods to determine the effect its absence would have on the response of the HNS to a hypertonic saline injection. Pentobarbital has been shown to reduce stress-induced release of adrenalin in the brain and periphery and to modulate increases in plasma corticosterone levels (2,26,33), so one group of animals was anesthetized to reduce possible sympathetic adrenal arousal caused by the hypertonic saline stimulus. Other animals underwent bilateral removal of the adrenal medullae to remove that source of adrenalin. Neither manipulation should interfere with the osmotic nature of the stimulus. METHODS

Male Sprague-Dawley rats, 80-100 days old, on a 12:12 reversed light-dark cycle were given 18 ml/kg IP injections of hypertonic (1.5 M) or normal (0.15 M) NaCl solution 1 h after lights on. In addition to the saline injection, animals were also given 1 of 3 treatments in the following conditions. Baseline condition: Ten animals were injected (half received 1.5 M NaCl; half 0.15 M NaCl) and were sacrificed 5 h post injection with no additional treatment. Anesthetized condition: Ten animals were anesthetized with 3.3 ml/kg equithesin prior to injections of either 1.5 M or 0.15 M NaCl. After receiving equithesin doses of this level, animals became immobile and did not respond to an ear or tail pinch. They remained quiescent for the 5 h between saline injection and reanesthetization and perfusion. Medullectomized (MDX) condition: Ten rats were bilaterally adrenal-medullectomized following a widely used procedure (7,31). A 2 cm incision was made in the lateral abdominal wall. The adrenal gland was gently lifted without compromising its vascular supply. A slit was made in the gland with a microscissors. A forceps was used to squeeze the gland; this resulted in the emergence, through the slit, of the medullary core, which was then removed. During this procedure, the

medullae being expressed from the glands can easily be seen either with the naked eye or with the aid of a magnifying glass. Four animals received sham medullectomies, in which the adrenals were exposed but not removed. Incisions were then sutured and the rats were allowed to recover from surgery. One week later, half the animals received 1.5 M NaCl solution injections; half, 0.15 M NaCl. Five hours after the injections, all animals were deeply anesthetized with equithesin and perfused by a brief transcardial rinse of 0.15 M NaCl followed by a fixative of 2.5% glutaraldehyde and 1% paraformaldehyde in a 0.1 M cacodylate buffer. For perfusion, a Masterflex variable speed pump using tubing with 0.06 inch inner diameter was adjusted so the flow rate was approximately 5 ml/min, a flow rate that preserves pituitary ultrastructure. Brains and pituitaries were removed from the skulls and left in fix overnight. Coronal slices 500 pm thick were cut with a tissue chopper through the anterior-posterior extent of the SONS. SONS were recovered from three adjacent slices. Neural lobes were dissected free of anterior and intermediate lobe tissue, oriented in resin blocks and trimmed so that sections could be taken from the center of the broadest part of the neural lobe. This area includes terminal projections from both the SON and the PVN (1). SONS and neural lobes were removed and postfixed in 1% osmium tetroxide for 1 h, en bloc stained with uranyl acetate overnight, dehydrated in alcohol, and embedded in Spurr’s resin. Blocks containing the middle third (rostrocaudally) of the SON were used for data collection. Thin sections (silver-gold interference color) were examined with a Japan Electron Optics Laboratory (JEOL, E. Lansing, MI) 100 CX Electron Microscope. Each animal’s score on each of the various measurements was based on a mean of 10-15 micrographs for both areas, analyzed at a print magnification of 13,500X enlargement. In the SON, the total number of terminals was counted and a Sigma Scan Digitizing program (Jandel Scientific, San Rafael, CA) for an IBM PC was used to measure the extent of terminal contact with the MNCs. Point intersection analysis (32) was used to determine glial, somatic or axon terminal membrane apposition for the soma in the SON and to compute somatic, nuclear and nucleolar size, and amount of Golgi apparatus in cross sectional profiles of MNCs in the SON. The same procedure was performed to compute percentage of neural contact with the BL in the posterior pituitary. In the micrographs of the pituitaries, the number of terminals surrounded by cytoplasm was counted for each animal. Means for SON measurements in the baseline and anesthetized animals were compared using a 2 X 2 Factorial Analysis of Variance (ANOVA), followed by post hoc Tukey tests when appropriate. Values of p < .05 were considered significant. The SON measurements for the two NaCl conditions of MDX animals were compared using Student’s t-test. RESULTS

Morphological

Changes in the SON

Baseline conditions. Animals that received 1.5 M NaCl in the baseline condition showed a significant decrease in glial apposition and a concomitant increase in somatic and terminal apposition with the membrane of MNCs in the SON compared to the isotonic saline controls (see Tables 1 and 2). Cell body and nuclear size, amount of nucleolar material, amount of Golgi apparatus in the MNC cytoplasm and number of multiple synapses also all increased. The results essentially replicated those of our earlier study (4). Anesthesia conditions. The effects of anesthesia on the HNS of animals receiving 1.5 M or 0.15 M NaCl injections were eval-

213

SYSTEMIC SIGNALS TO HNS

TABLE 1 EFFECTS OF ANESTHESIA AND SALINE TREAWENT ON MEMBRANE APPOSITION WITH MNCS AND ON THE SIZES OF hfNC CELL BODIES, NUCLEI AND NUC~OLI % Apposition w/MNC Sizes (m fhm2)

Condition

Glisl

Tem-kd

Somatic and/or Dendiitic

1.5 M NaCl (n = 5) 0.15 M NaCl (n = 5) 1.5 M NaCl + anesthesia (n = 4) 0.15 M NaCl + anesthesia (n = 4)

63 -+ l*$ 792 1

13 If 1st 81t .7

18 rt l$II 8ir 1

293.4 rf; 8.9.t 245.0 If: 18.7

102.6 + 5.6t 86.0 ” 2.9

8.6 ?I .6$ 7.3 2 .8

722

10 2

.9

lOI?; 1

259.2 t 11.1

91.0 ;c 8.1

8.0 k .9#

9t

.8

10 r?r5

230.0 It 23.0

89.3 ” 6.7

5.3 Y!Y .6

2

75 c 5

Cell Body

Nucleus

Nucleoli

* <0.15 M NaCI,p < .Ol. i >0.15 M NaCI,p < .OS. $ >O.lS M NaCI,p < .Ol. Qcl.5 M NaCl anesthetized,p < .05. //>1.5 M NaCl anesthetized,p < .OS. ##>0.15 M NaCI anesthetized,p < .05.

uated with a 2 X 2 Factorial ANOVA (see Tables 1 and 2). There was a significant interaction effect between saline treatment and anesthesia on the values of all the SON parameters measured except amount of nucleoiar material (Fs = 4.6-17.4, ps < .OS.oOl). Post hoc comparisons showed that there were significant differences in the SON between groups of animals receiving injections of 1.5 M NaCl and 0.15 M NaCl in the nonanesthetized condition for all elements in apposition with MNC membrane, cell body and nucleus size, amount of nucleolar material, percent of MNCs contacted by multiple synapses and amount of Golgi appara~s in the MNC cytoplasm. In the anesthetized condition, differences between 1.5 M NaCl and 0.15 M NaCl injected animals were significant only for total number of terminals contacting the MNC membrane and amount of Golgi apparatus in the cytoplasm; for all other SON elements measured, the anesthesia interfered with the changes that are usually induced by hypertonic saline. On measures of total number of terminals contacting the MNC membrane and amount of Golgi apparatus in the cytoplasm, as well as for glial and somatic apposition, and percent of MNCs contacted by multiple synapses, the 1.5 M NaCl injected unanesthetized animals measured significantly above the 0.15 M NaCl injected anesthetized animals; thus the anesthesia was shown to significantly mitigate the effect of the hypertonic saline injection for these measures. There were no other signifi-

cant differences in the two saline injection conditions in anesthetized animals. Mean amount of nucleolar material was significantly affected by the saline condition only (and not by anesthesia or intera~ion between saline and anesthesia), suggesting that for this element only, the anesthetized animals responded to the injections of 1.5 M NaCl in a manner similar to the unanesthetized animals. The mean values for SON elements for each group are presented in Tables 1 and 2. MDX conditions. Although MDX animals were given the option of drinking 1% NaCl solution or tap water, no MDX animal drank a measurable quantity of the 1% NaCl solution; all consumed the tap water. After the animal was perfused, the adrenal glands were removed, cut in half and subjected to visual examination to confirm that adrenal cortices were intact and free of lesions or gross degeneration, and that the adrenal medullae were absent. Sham operated animals were found to be similar on all measures to animals who had received no surgery (i.e., baseline condition). SON measurements from MDX animals receiving either 1.5 M NaCl or 0.15 M NaCl were compared. Amount of Golgi apparatus in SON neuronal cytoplasm was significantly greater in the 1.5 M NaCl injection condition @ < .Ol, t-test). This was the only SON measure to have significantly different values in the

TABLE 2 EFFECTS OF ANESTIIESIA AN’D SALINE TREATMENT ON SYNAFTIC ELEMENTS AND GOLGI APPARATUS OF MNCS

# synapse/ Condition 1.5 M NaCl (n = 5)

0.15 M NaCl (n = 5) 1.5 M NaCl anesthesia (n = 4) 0.15 M NaCl anesthesia (n = 4) * >0.15 M NaCI, p < .Ol. t >0.15 M NaCI, p < .05. $ ~1.5 M NaCl anesthetized,p < .05. 5 >0.15 M NaCl anesthetized,p < .05.

%MNcS w/m~tipl~

I@Qfl

.082 + .083 2 .12 2 .12 ?

.012 .004 .004 .007

.028 2 .018 2 .023 ? ,015 +

.004*$ .003 .005~ .002

69 + 4*$ 242 8 48 +- 11 492 2

% Golgi Appsmt=

8.4 2 .4t$ 4 2 .Ol 6.5 5 .59 4.3 -c .4

214

BEAGLEY

TABLE EFFECTS

OF ADRENAL-MEDULLECKIMY

AND SALINE

% Apposition

AND HATTON

3

TREATMENT ON MEMBRANE APPOSITION BODIES, NUCLEI AND NUCLEOLI

WITH MNCS AND ON THE SIZES OF CELL

wfMNC Somatic

Sizes (in km’)

ZUld/0r

Condition

1.5 M NaCl baseline

Tehllal

Dendritic

Cell Body

Nucleus

63 i 1 79 ? 1

13 k 1.5 82 .7

18 + 1 8fl

293.4 + 8.9 245.0 k 18.7

102.6 + 5.6 86.0 + 2.9

60 64

17 11

291.8 319.7

103.0 123.6

8.8 8.5

79 75 73 t 2 74 + 2

6 8

207.0 226.2 252.6 + 12.8 248.2 f 9.4

79.0 82.8 116.4 + 9.2 106.4 ? 5.2

7.4 6.6 9.3 -t 1.0 8.2 k .9

Glial (n = 5)

M NaCl baseline (n = 5) 1.5 M NaCl Sham A Sham B 0.15 M NaCl Sham C Sham D 1.5 M NaCl + MDX (n = 5) 0.15

0.15 M NaCl + MDX (n = 5)

82 6k1

20 21 10 8 12 t 2 12 2 2

.4

Changes in the Posterior Pituitary

Baseline conditions. In intact rats, decreased pituicyte and increased neural apposition with the BL in the posterior pituitary resulted from 1.5 M as compared to 0.15 M NaCl injections. The number of axons/terminals enclosed by pituicyte cytoplasm was

TABLE EFFECTS

4

OF ADRENAL-MEDULLECTOMY AND SALINE AND GOLGI APPARATUS

Condition

# Synapse/100

1.5 M NaCl baseline (n = 5) 0.15 M NaCl baseline (n = 5) 1.5 M NaCl Sham A Sham B 0.15 M NaCl Sham C Sham D 1.5 M NaCl + MDX (n = 5) 0.15 M NaCl + MDX (n = 5)

.082 2 ,012 .083 2 ,004

* Greater than 0.15 M NaCl MDX,

.08 .lO

pm

8.6 5 7.3 2

TREATMENT OF MNCS

ON SYNAPTIC

ELEMENTS

# Mult/lOO jm

% MNCs w/Multiples

% Golgi Apparatus

,028 5 ,004 ,018 ? ,003

69 -e 4 24 k 8

8.4 I? .4 4 ? .01

.07 .08

,021 ,030

57 64

9 8

.08 .09 ? .Ol ? .Ol

,016 ,012 ,022 t ,003 ,013 t .002

29 25

4 4 8.4 -e .7* 3.8 2 .7

p < .Ol.

.6 .8

significantly smaller in animals that had received the 1.5 M NaCl injections. Anesthesia conditions. As in prior studies, hypertonic as compared to isotonic injections resulted in more neural apposition with the BL and fewer axons enclosed by pituicyte cytoplasm in nonanesthetized animals. In anesthetized animals, however, hypertonic injections failed to produce increased neural apposition with the BL, but did result in fewer enclosed axons (see Table 5). MDX conditions. No differences were observed between the MDX animals in the two injection conditions for neural or pituicyte apposition with the BL, or number of axons/terminals enclosed by pituicytes (Mean percent neural apposition with BL: 1.5 M NaCl = 33% 2 3, 0.15 M NaCl = 36% 2 2, p < .44; mean percent pituicyte apposition with BL: 1.5 M NaCl = 66% 2 3, 0.15 M NaCl = 63% ? 2, p c.47; mean number axons enclosed by pituicyte cytoplasm: 1.5 M NaCl = 1.2 +- .3, 0.15 M NaCl = 1.8 2 .3,p < .24). Both anesthesia and adrenal medullectomy interfered with increased neural apposition with the BL normally seen in an intact, awake animal in response to an injection of 1.5 M NaCl. Animals that received 1.5 M NaCl injections and no further treatment (baseline condition) were the only group to show a significant increase in neural (and decrease in pituicyte) apposition with the

two MDX saline injection conditions. Percentage of Golgi apparatus in the SON cell cytoplasm of MDX animals was similar to that of intact animals that had received comparable injections. As shown in Tables 3 and 4, for all other SON elements, values obtained from MDX animals injected with either 1.5 M or 0.15 M NaCl did not differ from those obtained from intact animals receiving 0.15 M NaCl injections. For all MDX animals (treated with either 1.5 M or 0.15 M NaCl) these elements were significantly different from intact, 1.5 M NaCl injected animals; but there were no differences between the two MDX saline injection conditions. Both 1.5 and 0.15 M NaCl injected MDX animals had enlarged nuclei and nucleoli compared to intact rats given 0.15 M NaCl. This suggests the adrenal medullectomy procedure itself may increase the size of the nuclei of the SON and no further activation as the result of the hypertonic injection can be measured.

Morphological

Nucleoli

49 rt 7 35 5 5

21s

SYSTEMIC SIGNALS TO HNS

TABLE 5 NEURAL LOBE MEASURES FOR ANESTHETIZED AND NONANESTHETIZED ANIMALS INJECTED WITH 1.5 M OR 0.15 M NaCl 3%Neural Membrane/ BL Apposition

Condition

1.5 M NaCl baseline n=6 0.15 M NaCl baseline n=6 1.5 M NaCl anesthetized n=4 0.15 M NaCl anesthetized Factorials

59 -+ 2t 35 k 3 39 k 4 38 2 4 Saline: F = 15.7, p < SKI1 Anesthesia: F = 6.95, p < .017 Saline X anesthesia: F = 14.4,~ < .OOl

Axons Enclosed by Pituicyte

.92 ? .20$ 1.8 2 .32 .74 k .36# 2.2 k 44 Saline: F = 17.3, p < 0307 Anesthesia: F = .093, p < .76 Saline X anesthesia: F=.M,p<.37

* >0.15 M NaCl,p < .Ol. t >1.5 M NaCl anesthetized, p < .05. $ co.15 M NaCl, p < .05. 5 <0.15 M NaCl anesthetized, p < .05

BL. Fig. 1 compares mean neural apposition with the BL for baseline, anesthetized, and medullectomized animals receiving 1.5 M and 0.15 M NaCl injections.

Fig. 2 is a micrograph from the neural lobe illustrating morphological changes typically produced by an IP injection of 1.5 M NaCl. The stellate shape is typical of an activated pituicyte

Baseline 1.5 M NaCl

Anesthesia EEA

and contrasts with the amorphous shape of the pituicytes in Figs. 3 and 4. Fig. 3 is a micrograph from the neural lobe of an animal medullectomized prior to receiving the 1.5 M NaCl injection, and Fig. 4 is a micrograph from the neural lobe of an animal which received an injection of 0.15 M NaCl. The shape of pituicyte and amount of neural apposition with the BL was similar for these conditions.

MDX

O.lSMNaCl

FIG. 1. Neural apposition with the basal lamina in the posterior pituitary for (intact) baseline, anesthetized and adrenal medullectomized animals. (* = significantly different from corresponding 0.15 M NaCL control at p < .OOl, r-test.)

FIG. 2. Electron micrograph of neural lobe of an intact rat injected with 1.5M NaCl. The pituicytes acquire a characteristic stellate form. Pituicyte cytoplasm (p) is retracted, and most terminals (t) abut the basal lamina (BL). N = pituicyte nucleus; 1 = lipid; Bar = lpm.

BEAGLEY

216

FIG. 3. Electron micrograph of neural lobe of medullectomized rat injected with 1.5 M NaCI. Axons (ax) are enclosed by pituicyte cytoplasm (p), and the pituicyte is interposed between axons and the basal lamina (BL) and fenestrated capillary (fc). N = pituicyte nucleus; 1 = lipid; Bar = lpm.

DISCUSSION

AND HATTON

There is evidence, however, that animals without their adrenal medullae are capable of responding to the hypertonic stimulus, as the amount of Golgi apparatus in the MNC cytoplasm of animals receiving 1.5 M NaCl was elevated to the level of the intact, awake animal receiving 1.5 M NaCl. Increased amount of Golgi apparatus reflects increased production of secretory vesicles for release (6). This suggests that the MNCs are responding to the 1.5 M NaCl stimulus in a manner independent of glial retraction. Glial retraction does not appear to be necessary for early increases in hormone synthesis. Activity of the HNS in response to a systemic adrenalin increase may involve glial retraction from between MNCs in the SON permitting increased communication between the cell bodies, and pituicyte retraction from the BL facilitating release of neuropeptides from the neurohypophysis. That the signalling agent may be adrenalin is borne out by the failure of the system to respond in animals which have had the medullary portions of their adrenal glands removed. The neural lobe is located outside the blood-brain barrier and might respond to catecholaminergic signals originating from the adrenal gland and carried via the blood stream. There is long standing evidence that peptides released from the neural lobes act upon the adrenal glands (12), and it is equally possible that signals from the adrenal gland act directly or indirectly on the HNS. Glial cells in the posterior pituitary have direct access to the circulatory system and could be directly influenced by a blood borne signal such as adrenalin. The ventral glial lamina of the SON is in direct contact with the pial surface of the brain, and although the blood-brain barrier is more difficult to penetrate in this area of the brain than in the neural lobe, it is easier for some substances to pass through the barrier in the SON than in less vascularized areas (13). SON glial cells are in close proximity to many blood vessels which could allow for an effective blood borne signal. Also, like the pituicytes, SON glia have now been

Results of this investigation suggest that adrenal activation plays an important role in signalling some of the activities involved in increased neuropeptide release in the HNS. Both the manipulations of keeping the animal anesthetized prior to injection of saline through perfusion, and removal of the adrenal medullae interfered with responses of the HNS to the hypertonic stimulus. Many of the responses interfered with by anesthesia or medullectomy could be the result of reduced glial retraction from between the MNC somata. For animals that were anesthetized or adrenal medullectomized prior to the 1.5 M NaCl injection, there was little or no difference in any of the apposition measures which typically characterize an animal 5 h after being injected with hypertonic vs. isotonic saline. Nor was there a large increase in MNC cell body size, which could be related to glial retraction (22). For some SON parameters, hypertonic saline did effect a change on the ultrastructure in the anesthetized animals, but the magnitude of these effects was smaller than in the nonanesthetized animals. In the anesthetized animals, only the amount of nucleolar material was increased by the hypertonic saline injection without significant interference from the anesthesia. As previous evidence suggests (3), increases in amount of nucleolar material may precede glial retraction from the MNC membrane, and this may be an early and very sensitive response to the hypertonic NaCl injection stimulus. Large systemic injections of hypertonic saline produced no change in most SON measures in MDX animals. When compared to intact animals, HNS morphological measures for MDX animals in both saline conditions appear similar to those of the baseline animals injected with isotonic saline.

4. Electron micrograph of neural lobe of rat injected with 0.15 M NaCl. Axons (ax) are enclosed by pituicyte cytoplasm (p), and the pituicyte is interposed between axons and the basal lamina BL. N = pituicyte nucleus; Bar = lpm. FIG.

SYSTEMIC SIGNALS TO HNS

217

shown to be rich in P2-adrenoceptors which increase in numbers relative to @l-receptors in response to osmotic stimulation (18). Furthermore, perfusion of the brain with hyperosmolar solutions has been shown to increase barrier permeability (23). Thus, the stimulus of hypertonic saline could increase the penetrability of the blood-brain barrier, allowing the entry of systemic adrenalin to signal retraction in the glial cells in the SON. Without the adrenal medullae, the animal may be incapable of releasing sufficient quantities of adrenalin to stimulate glial retraction in the SON in the 5 h time interval between injection and sacrifice. In the neural lobe both anesthesia and adrenal medullectomy prevented the changes in neural/BL apposition that typically occur within 5 h in response to a hypertonic injection in intact animals. Pituicyte coverage of the BL remained at levels found for animals receiving isotonic injections in both cases. This suggests that anesthesia either inhibits the retraction of the glial cells directly, or suppresses a peripheral response (such as a stimulusinduced increased adrenalin output from the adrenal medullae) that would normally activate the system in response to the 1.5 M NaCl injection. Since removal of systemic adrenalin also cancels the responses of the neural lobe to the 1.5 M NaCl injection, it is probable that blood borne adrenalin is the agent responsible for acting on the pituicytes causing both their withdrawal from the BL and their release of formerly enclosed axons. The response of the HNS to 1.5 M NaCl for all the elements measured in the anesthetized animal appear to be somewhat suppressed, suggesting a more general inhibition of activity by the anesthesia. Dyball(l1) found that pentobarbitone decreased oxytocin (OX) release from incubated neural lobes by interfering with calcium uptake (presumably by the terminals) from the incubation media. Direct interference with calcium uptake might explain the inhibitory effect of the anesthesia on the HNS, but this would be in the neural as opposed to glial compartment. Evidence against substantial anesthetic interference with HNS neural activity is provided by Brimble and Dyball(5) who found that cell bodies in the SON responded to an intracarotid injection of NaCl with similar increases in firing rate in anesthetized and unanesthetized animals. Cheng and North (9) determined the sodium pentobarbital did not interfere with the basal neuropeptide release (measured by plasma levels of OX and VP) but did reduce the responsiveness of VP but not OX neurons to acute osmotic stimulation, perhaps by modifying activity at the glutamate receptors on VP neurons. Neither study rules out the possibility of direct suppression of glial activity by anesthesia. An alternative interpretation, however, is that an anesthetized animal may be insufficiently stressed to mobilize increased release of adrenalin

from the adrenal medulla. This view is supported by the work of Taborsky, Halter, Bau, Best, and Porte (26) and of Chaouloff, Baudrie and Laude (8) who suggest pentobarbital’s blunting of stress induced activation of the sympathetic adrenal system occurs at the level of the adrenal medulla. The absence of adrenalin may then be at least partially responsible for lack of glial retraction (and neuropeptide release) from the neural lobe. Sladek and Armstrong (24) reviewed several studies examining the possible location of osmoreceptors and sodium receptors that result in vasopressin (VP) release. Because of the large osmolality changes required to elicit significant changes in neuronal firing in in vitro studies, they concluded that it is unlikely that osmotic depolarization is the sole mechanism responsible for somatic regulation of VP release in vivo. The osmoreceptor role of the VP neuron may be to maintain responsiveness of the system to chronic stimulation. Our findings are consistent with this hypothesis. While we cannot discount the importance of the osmotic character of the stimulus, the decreased response in anesthetized and MDX animals suggests at least one additional factor is signalling activity in the HNS. The HNS increases release of neuropeptides in response to many of the same signals (such as pain, fever, nausea, and a variety of other stressors) that also increase release of adrenalin by the sympathetic nervous system. Many actions, such as VP induced vasoconstriction and increased hepatic glycogenolysis or OX induced constriction of arteriolar smooth muscle, are actions similar to those of adrenalin in an aroused animal (10). This suggests the HNS neuropeptides may potentiate or maintain the effects of adrenalin when the organism is required to undergo stress for an extended period of time. Adrenalin release from the adrenal medulla and its action on various targets in the periphery have long been established as a characteristic response to stress (27). The findings of this study suggest that adrenalin plays a crucial role in the HNS by acting on astrocytic glia, causing retraction and increasing communication among MNCs in the SON and facilitating neuropeptide release in the posterior pituitary. The neuropeptides, in turn, act in the periphery to enable the organism to better deal with a long lasting stressful stimulus. ACKNOWLEDGEMENTS

Thanks are due to Dr. J. Kiong, Dr. Nilah On, and Dr. Mark Weiss for comments on early drafts of this manuscript. The technical support of P. Rusch and L. Koran is greatly appreciated. Supported by NIH Research Grant NS 09140 and Training Grant NS 07279.

REFERENCES 1. Alonso, G.; Assenmacher, I. Radioautographic studies on the neurohypophysial projections of the supraoptic and paraventricular nuclei in the rat. Cell Tiss. Res. 219525-534; 1981. 2. Anishchenko, T. G.; Igosheva, N. B. Sex differences in stress response in conscious and anesthetized rats during surgical stress. Bvull. Eksn. Biol. Med. 113:26-28: 1992. Beagley, d. H.; Neuroplasticity in’ adult hypothalamo-neurohypophysialsystem: Mechanisms involved in responding to a single systemic hypertonic saline injection. PhD dissertation UMI microfilm; 1991. Beagley, G. H.; Hatton, G. I. Rapid morphological changes in supraoptic nucleus and posterior pituitary induced by a single hypertonic saline injection. Brain Res. Bull. 28:61%18; 1992. Brimble, M. J.; Dyball, R. E. J. Characterization of the responses of oxytocin- and vasopressin-secreting neurones in the supraoptic nucleus to osmotic stimulation. J. Physiol. (London) 271253-271; 1977.

6 Broadwell, R. D.; Oliver, C. Golgi apparatus, GERL, and secretory granule formation within neurons of the hypothalmo-neurohypophysial system of control and hyperosmotically stressed mice. J. Cell Biol. 90:474A84; 1981. 7. Buu, N. T.; Lussier, C. Origin of dopamine in the rat adrenal cortex. Am. J. Physiol. 258:F287-291; 1990. 8. Chaouloff, F.; Baudrie, V.; Laude, D. Pentobarbital anaesthesia prevents the adrenaline-releasing effect of the S-HTiA receptor agonist, 8-hydroxy-2-(di-n-propylamino) tetralin. Euro. J. Pharm. 180:175178; 1990. 9. Cheng, S. W. T.; North, W. G. Inlluence of pentobarbital and urethane on release from magnocellular neurons. Neuroendocrinology. 54:482-487; 1991. 10. Cunningham, E. T.; Sawchenko, P. E. Reflex control of magnocellular vasopressin and oxytocin secretion. Trends in Neurosci. 14:407-411; 1991.

218

11. Dyball, R. E. J. Potentiation by urethane and ambition by pentobarbitone of oxytocin released in vitro. J. Endocrinol. 67453-458; 1975. 12. Gaunt, R.; Lloyd, C.; Chart, I. The adrenal-neurohypophysial interrelationship. In Heller, H., ed. The neurohypophysis. London: Butterworth; 1957233-251. 13. Gross, P. M.; Sposito, N. M.; Pettersen, S. E.; Fenstermacher, J. D. Differences in function and structure of the capillary endothelium in the supraoptic nucleuses and pituitary neural lobe of rats. Neuroendocriuology. 44401-407; 1986. 14. Harbuz, M.; Lightman, S. Responses of hypothalamic and pituitary mRNA to physical and psychological stress in the rat. J. Endocrinol. 122:705-711; 1989. 15. Hatton, G.I. Emerging concepts of structure-function dynamics in adult brain: The hy~~al~o-neuroh~physial system. Prop. Neurobiol. 34:437+X& 1990. 16. Hatton, G. I.; Luckman, S. M,; Bicknell, R. J. Adrenalin activation of betaZadrenoceptors stimulates morphological changes in astrocytes (pituicytes) cultured from adult rat neurohypophyses. Brain Res. Bull. 39:701-709; 1991. 17. Kasting, N. W. Simultaneous and independent release of vasopressin and oxytucin in the rat. Can. J. Physiol. Pharm. 6622-26; 1988. 18. Lafarga, M.; Berciano, M. T.; Del Olmo, E.; Andres, M. A.; Pazos, A. Osmotic simulation changes in the expression of P-adrenergic receptors and nuclear volume of astrocytes in supraoptic nucleus of the rat. Brain Res. 588:311-316; 1992. 19. Lightman, S. L.; Young, W. S. Vasopressin, oxytocin, dynorphin, enkephalin and corticotrophin-releasing factor mRNA stimulation in the rat. J. Physiol. (London) 394:%39; 1987. 20. Lightman, S. L.; Young, W. S. Corticotrophin releasing factor, vasopressin, and pro-opiomelan~~in mRNA responses to stress and opiates in the rat. J. Physiol. (London) 403:511-523; 1988. 21. Luckman, S. M.; Bicknell, R. J. Morphological plasticity that occurs in the neurohypophysis following activation of the magnocellular neurosecretory system can he mimicked in vitro by beta adrenergic stimulation. Neuroscience 39:701-709; 1990.

BEAGLEY

AND

HATTON

22. Modney, B. K.; Hatton, G. 1. Multiple synapse formation: A possible compensatory mechanism for increased cell size in rat supraopticnucleus. J. Neuroendocrinol. 1:21-27, 1989. 23. Rapoport, S. I. Effect of concentrated solutions on the blood-brain barrier. Am. I. of Physiol. 9:270-274; 1970. 24. Sladek, C.; Armstrong, W. Osmotic control of vasopressin release. Trends in Neurosci. 8:166-168; 1985. 25. Smithson, K.: Suarez. I.; Hatton, G. I. Beta-adrenersic stimulation decreases’giial and increases neural contact with the basal larnina in rat neurointermediate lobes incubated in vitro. J. Neuroendocrinol. 2:693-699; 1990. 26. Taborsky, G. J.; Halter, J. B.; Bau, D.; Best, J. D.; Porte, D. Pentobarbital anesthesia suppresses basal and 2-deoxy-D-glucose-stimulated plasma catecholamines. Am. J. Physiol. 247:R905; 1984. 27 Tepperman, J. Metabolic and endocrine physiology year book. Chicago: Medical Publishers Inc.; 1968. 28. Tweedle, C. D.; Hatton, G. I. Pituicytes and neurosecretory axons; dynamic interactions and the control of hormone release in the rat. Neuroscience 5:661-667; 1980. 29. Tweedle, C. D.; Hatton, G. I. Glial cell enclosure of neurosecretory endings in the neurohypophysis of the rat. Brain Res. 192555-559; 1980. 30. Tweedle, C. D.; Hatton, G. I. Mo~hological adaptability at neurosecretory axonal endings on the neurovascular contact zone of the rat neurohypophysis. Neuroscience Z&241-246; 1987. 31. Wilkinson, C.; Shinsako, J.; Dallman, M. F. Return of pituitaryadrenal function after adrenal enucleation or transplantation: Diurnal rhythms and responses to ether. Endocrinology 109:162-169; 1981. 32. Williams, M. A. Quantitative methods in biology. in: Glauert, A. M., ed. Practical methods in electron micros~py, vof. 6. New York: Elsevier/North Holland Biomedical Press; 1981:4-84. 33. Yoshishige, I.; Tsuda, A.; Tsujimaru, S.; Satoh, M.; Masatoshi, T. Pentobarbital attenuates stress-induced increases in noradrenaline release in specific brain regions of rats. Pharm. Bio. Behav. 36:953956; 1990.