0013-7227/90/1263-1709S02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society

Vol. 126, No. 3 Printed in U.S.A.

Induction of fos-Like Immunoreactivity in Hypothalamic Corticotropin-Releasing Factor Neurons after Adrenalectomy in the Rat* LAUREN JACOBSONt, FRANK R. SHARPS AND MARY F. DALLMAN Departments of Physiology (L.J., F.R.S., M.F.D.) and Neurology (F.R.S.), University of California, San Francisco, California 94143-0444

ABSTRACT. To identify brain sites responding to the removal of corticosterone feedback by adrenalectomy (ADX), rat brains were processed for fos immunocytochemistry 1, 3, and 7 days after ADX, sham-ADX, or no surgery using a polyclonal antiserum to fos residues 132-154. Compared to SHAM, ADX rats exhibited strong fos-like immunoreactivity (FLI) only in the parvocellular neurons of the paraventricular hypothalamic nuclei (PVN) 1, 3, and 7 days after surgery. Replacement with a corticosterone pellet at the time of adrenalectomy (ADX + B) prevented this increase in PVN FLI in three of four rats at 1 day, all rats at 3 days, and two of seven rats 7 days after surgery; 100 Mg/ml corticosterone in the drinking water for 2 days before perfusion reversed ADX-induced increases in PVN FLI in 7-day

ADX rats. Providing 25 Mg/ml corticosterone in the drinking water to ADX rats for 5 days after surgery did not prevent expression of PVN FLI, even though this dose has been shown to normalize morning basal ACTH levels in ADX rats. Virtually all parvocellular PVN neurons expressing FLI after ADX costained for CRF. Some parvocellular neurons also expressed both fos and vasopressin. In all rats, many brain regions expressed FLI that was not related to adrenalectomy. We conclude that the changes in neuronal FLI correlate with demonstrated changes in neuroendocrine activity after ADX; however, suppression of ADX-induced FLI may require higher replacement levels of corticosterone than inhibition of ADX-induced ACTH secretion. {Endocrinology 126: 1709-1719, 1990)

N

EURONAL expression of the c-fos proto-oncogene may serve as a marker of neural activity (1, 2). The c-fos protein product fos has been found to be induced in PC12 cells by membrane depolarization (3, 4). Pharmacologically or electrically induced seizures stimulate rapid and transient expression of fos mRNA (1) and protein (1, 5) in the limbic system and cerebral cortex. Specific electrical, sensory, or noxious stimuli have also been shown to induce fos expression in discrete brain regions corresponding to known neuroanatomical targets for such stimuli. Cutaneous nociceptive and proprioceptive stimuli induce fos in the dorsal horn of the spinal cord (6). Fos induction occurs in thalamus, pontine nuclei, and cerebellum after electrical stimulation of Received August 7, 1989. Address all correspondence and requests for reprints to: Mary F. Dallman, Department of Physiology, S-762, University of California, San Francisco, California 94143-0444. * This work was supported in part by USPHS Grants KD-28172 (to M.F.D. and NS-24666 (to F.R.S.). t Recipient of an Achievement Rewards for College Scientists (ARCS) fellowship. Work performed during predoctoral studies in the interdisciplinary Program in Endocrinology, University of California, San Francisco. Present address: Department of Biological Sciences, Stanford University, Stanford, California 94305. $ Present address: Neurology Service V127, Veterans Administration Medical Center, 4150 Clement Street, San Francisco, California 94121.

motor/sensory cortex (2, 7). Dehydration induces fos expression in the hypothalamic paraventricular (PVN) and supraoptic nuclei (2, 8). The induction of fos in brain areas not immediately adjacent to sites of cortical lesion, stimulation, or nerve growth factor injection suggests that neuronal fos expression may be regulated transsynaptically (2, 7, 9). Consistent with this possibility, electrical stimulation of motor/sensory cortex has been shown to induce fos expression at least two synapses away in areas of the cerebellum similar to those exhibiting increases in metabolic activity (2, 7). Neuronal activity has been reported to change in several brain regions in response to adrenalectomy (ADX). The hypothalamic PVN, median eminence, hippocampus, and locus ceruleus were found to exhibit increases in cerebral glucose metabolism 5 h after adrenalectomy that could be prevented by dexamethasone administration (10). The electrical activity of PVN neurons in hypothalamic slices has also been found to increase after adrenalectomy (11). Several well characterized neuroendocrine changes occur in the hypothalamus, including increased expression of CRF and vasopressin (AVP) in the parvocellular division of the PVN (12-19) and increased secretion of both CRF and AVP into the hypophysial-portal blood (19, 20).

1709

CRF NEURONS EXPRESS fos AFTER ADX

1710

The purpose of this study was 2-fold: 1) to identify, using fos immunocytochemistry, the neurons that respond to removal of glucocorticoid feedback after adrenalectomy, and 2) to determine whether PVN CRF neurons express fos after adrenalectomy. Because there is little information on the identity of neurons expressing fos after a given stimulus, the correlation of fos induction with changes in neuroendocrine activity in the PVN will confirm the utility of fos expression as a marker of neural activity. Portions of this work have been presented in abstract form (21).

Materials and Methods Animal preparation Male Sprague-Dawley rats (275-350 g) from Holtzman (Madison, WI) were kept in a temperature-, humidity-, and light (12 h on)-controlled room and allowed food and water ad libitum. The experimental paradigm was approved by the University of California-San Francisco committee on animal care. ADX or sham-adrenalectomy (Sham) was performed under ether anesthesia. All surgeries were performed in the morning, except for rats killed 12 h after ADX, which underwent surgery at lights-off. Except for 3- and 7-day adrenalectomized (ADX) rats used for colocalization of fos and CRF or AVP immunoreactivity, all rats undergoing surgery were also implanted at the time of surgery with an indwelling ip catheter to permit anesthetic administration without handling the rat before perfusion. The catheter was protected by a metal spring attached to the rat by a sc dacron tab and leading out the top of the cage (22). For the time course of brain fos expression after ADX, the following treatment groups were studied at the indicated times. ADX. Rats were ADX and given 0.5% NaCl to drink (12 h; 1, 3, and 7 days). Sham. Rats were sham-ADX and given tap water to drink (12 h; 1, 3, and 7 days). ADX + B. Rats were ADX, replaced at the time of surgery with a sc pellet (23) of 35-60% corticosterone:cholesterol, and given 0.5% NaCl to drink (1, 3, and 7 days). Steroid in the 35% pellets was purchased from Steraloids (Wilton, NH), while pellets with higher proportions of corticosterone were made from Sigma corticosterone (St. Louis, MO). Untouched. Rats were undisturbed until injected with pentobarbital (ip) before perfusion (12 h; 1, 3, and 7 days). ADX + 5W2o-Rats were ADX and given 0.5% NaCl to drink until day 5, when they were given 100 fig/ml corticosterone (B) in 0.5% NaCl-0.5% ethanol to drink until perfusion on day 7. +B and -B. To determine the effects on FLI of removal of corticosterone feedback without simultaneous surgery, some rats were given 25 tig/ml in 0.5% NaCl-0.2% ethanol to drink after ADX. At lights-out on day 5 after surgery, water bottles were changed for all rats in this treatment group; half of these rats again were given the corticosterone-containing fluid (+B),

Endo • 1990 Vol 126 • No 3

while the other half was switched to the saline-ethanol vehicle (—B). We have previously shown that such stressless removal of corticosterone does not result in significant elevations in plasma ACTH relative to that in +B rats for 18-21 h (24). At the same time that water bottles were changed for -I-B and —B rats, additional rats were ADX, sham-ADX (Sham), or left undisturbed (Untouched). ADX rats were given 0.5% NaCl, while Sham and Untouched rats were given tap water to drink. All rats were perfused 12 h later within 1.5 h of lights-on (day 6 after surgery in +B and - B rats; 12 h after surgery in ADX and Sham rats). All ADX rats replaced with corticosterone (ADX + B, ADX + BH2O, and +B groups) were always killed with ADX (ADX unreplaced) rats. When used, Untouched rats were always killed with rats from other treatment groups, although they were not included with every set of rats. Perfusion and immunocytochemistry Rats were deeply anesthetized in the animal room with 80 mg/kg sodium pentobarbital ip (Antony Products, Arcadia, CA) containing 300 U sodium heparin (Elkins-Sinn, Cherry Hill, NJ). For most experiments, animals were anesthetized within 3 h of lights-on, although for more accurate timing of the 12-h time point, rats were anesthetized within 1-1.5 h of lights-on. Within 15 min of anesthetic administration, rats were perfused through the ascending aorta with 200 ml 0.9% saline (37 C), followed by 300 ml iced periodate-lysine-paraformaldehyde fixative (25). Brains were removed, blocked, and postfixed in periodate-lysine-paraformaldehyde for a maximum of 5 h. In most cases brains were cut in 100-/um sections on a Vibratome immediately after postfixation. Occasionally, brains were transferred to 0.1 M phosphate buffer (PB) for 1-2 h after postfixation, but were always cut within the same day. Brainstems were embedded in 4% agarose in 0.1 M PB before slicing. Sections were collected into 0.1 M PB and kept in buffer overnight before being processed for immunocytochemistry. Fos immunocytochemistry was performed using a polyclonal rabbit antiserum against a synthetic peptide corresponding to fos residues 132-154 (Microbiological Associates, Bethesda, MD). These residues are part of the M-peptide identical to both v-fos and c-fos (26). All procedures for generating the antiserum were performed under contract by Berkeley Antibody Co., Inc. (Berkeley, CA). The antiserum was affinity purified and used at a final dilution of 1:50 in 2% goat serum0.1% BSA-0.2% Triton X-100 in 0.1 M PB. For preadsorption controls, the primary antiserum in a 1:50 dilution was incubated with 1 Mg/ml of the original synthetic peptide antigen for 12 h at 4 C before addition of the brain sections. Before incubation with each primary antiserum, floating tissue sections were sequentially incubated for 10 min each at room temperature in avidin and biotin in 0.1 M PB (Vector Blocking Kit, Vector Laboratories, Burlingame, CA), followed by 2% goat serum-0.1% BSA-0.2% Triton X-100 in 0.1 M PB (30 min), to block nonspecific binding. Sections were incubated with the primary fos antiserum for 48-72 h at 4 C on a slow shaker platform. Sections were then washed three times (5-10 min) in PB and stained by the avidin-biotin-peroxidase technique using a Vectastain kit according to kit instructions (Vec-

CRF NEURONS EXPRESS fos AFTER ADX

1711

tor). Secondary antiserum was applied for 2 h at room temperature, followed by incubation with the avidin-biotin reagent for 3 h at room temperature. Peroxidase activity was revealed using 0.25% diaminobenzidine and 0.02% hydrogen peroxide (Sigma). For double label immunocytochemical detection of fos and CRF or AVP, 50-fim brain sections were collected and stained first for fos as described above. They were then put through a second round of blocking steps to reduce nonspecific binding and incubated for 48-72 h (4 C) with rabbit antiserum to either CRF or AVP at a final dilution of 1:500 in 2% goat serum-0.1% BSA-0.2% Triton X-100-0.1 M PB. Rabbit antiserum to ovine CRF was the generous gift of Dr. Wylie Vale (27). Rabbit antiAVP was obtained from Peninsula Laboratories (Belmont, CA). CRF- or AVP-immunoreactive neurons were revealed using a fluorescein-conjugated goat antirabbit antiserum (Vector) at a dilution of 1:50 in 2% goat serum-0.1% BSA-0.2% Triton X100-0.1 M PB. Incubations with the fluorescently tagged secondary antibody were performed over 3 h in foil-wrapped cell culture plates at room temperature. Sections for fluorescence microscopy were mounted in 0.1% paraphenyldiamine-10% 0.01 M PBS-90% glycerol. Unless otherwise indicated, all chemicals were obtained from Sigma.

those in Sham rats (data not shown). With the exception of one rat at the 1 day time point, corticosterone pellet implantation consistently prevented the appearance of FLI in the parvocellular PVN 1 and 3 days after ADX (Fig. 1). Seven days after surgery, parvocellular FLI was still low in two of seven rats. In the remaining 7-day ADX + B rats, /os-like staining was detectable, but the intensity was less than that in ADX rats (Fig. 1). The 1and 7-day ADX + B rats expressing PVN FLI had plasma corticosterone levels within the range of those in 1- and 7-day ADX + B rats exhibiting low levels of immunoreactivity. To determine if corticosterone could reverse as well as prevent ADX-induced increases in PVN FLI, 5-day ADX rats were given 100 ixg/m\ corticosterone in their drinking water for 2 days. This treatment reduced parvocellular /os-like staining to levels observed in Sham animals by day 7 (Fig. 1, ADX +

Hormone assay

We have previously shown that if corticosterone is removed stresslessly from ADX rats by changing their drinking fluid (—B), plasma ACTH does not differ for 18-21 h from that in rats continuing to drink steroid (+B). In contrast, plasma ACTH increases rapidly after surgical removal of corticosterone by ADX (24). To determine if a similar discrepancy might be observed for /os-like staining in the PVN, rats either underwent surgery or had their water bottles changed at lights-off. Rats were perfused 12 h later, a time when plasma ACTH had been observed to be relatively low and not to differ significantly between +B and —B rats (24). As in the foregoing experiments, ADX induced high levels of FLI in the parvocellular PVN by 12 h, whereas little to no staining was evident in Untouched rats. Parvocellular /os-like staining in 12-h Sham rats was more pronounced than at 1, 3, and 7 days, but was lighter than that in ADX rats. However, both —B and +B rats exhibited strong FLI comparable to that in ADX rats (Fig. 2).

Corticosterone was measured, as previously described (28), in plasma from tail nick samples taken after anesthesia and immediately before perfusion. The n for hormone values and immunocytochemistry differs because some sample tubes broke during centrifugation.

Results ADX-induced changes in FLI in the hypothalamic parvocellular PVN The hypothalamic PVN was the only brain area exhibiting strong and consistent differences in fos immunostaining between ADX rats and other treatment groups. Numerous /os-immunoreactive nuclei were found in the parvocellular PVN of ADX rats at all times after surgery, whereas the magnocellular PVN lacked fos immunoreactivity. Preadsorption of the primary fos antibody with the fos 132-154 peptide also abolished all nuclear FLI in the parvocellular PVN (Fig. 1, adsorption control) and throughout the brain (not shown). In contrast, /os-like staining was virtually absent from the parvocellular PVN of Sham and Untouched rats at 1, 3, and 7 days. Some staining was present in the dorsal cap and ventral margin of the parvocellular division, but this expression was not related to ADX, as it was found even in Untouched rats. Replacement of ADX rats at surgery with a sc corticosterone pellet produced corticosterone levels of 8.4 ± 1.6 (n = 4), 5.5 ± 0.9 (n = 6), and 4.8 ± 0.9 (n = 5) /ugM at 1, 3, and 7 days, respectively. The concentration of corticosterone in the pellets was determined in preliminary experiments to normalize morning plasma ACTH and thymus weight relative to

Effects of surgical and stressless removal of corticosteroids on expression of FLI in the parvocellular PVN

Coexpression of CRF or A VP with FLI in the parvocellular PVN Fos expression in ADX rats occurred throughout the medial division of the parvocellular PVN, suggesting that the distribution of FLI might coincide with that demonstrated for CRF (27, 29). To identify the PVN neurons expressing fos after ADX, sections were double stained for either CRF or AVP. Figure 3 shows sections through the PVN of 7-day ADX rats. Neurons expressing nuclear FLI (visible under light illumination and as black dots under fluorescent illumination) and either CRF or AVP (visible under fluorescent illumination) were localized by

CRF NEURONS EXPRESS fos AFTER ADX

1712

ADX

Sham •.'::•'''••••'•

ADX + B '"'

•

;

' ' ' • ' ' - '

Endo • 1990 Voll26«No3

Untouched

V

-

< - c '

\

••

1d

100 jjm

(6)

(3)

(5)

3d

•\

ADX Adsorption Control

(8)

(7).

(7)

(10)

(7)

(2)

(2)

7d

(3)

(2)

"IS .;v: ; :vM

(5)

j

FIG. 1. Expression of FLI in the PVN after ADX and corticosterone replacement. Rats were ADX, Sham, or ADX and immediately replaced with a sc corticosterone pellet (ADX + B). Rats were perfused for immunocytochemistry 1, 3, or 7 days later. Additional rats that had not undergone surgery were perfused at the same time (Untouched). ADX + BH2o, ADX rats given 100 ng/m\ corticosterone in the water from day 5 until perfusion on day 7; Adsorption Control, Section from a 7-day ADX rat in which the primary fos antiserum had been preincubated with the peptide corresponding to fos residues 132-154. Two photographs are provided for 7-day ADX + B rats to represent rats that did and did not exhibit PVN FLI; one ADX + B rat at 1 day also exhibited a high level of FLI similar to that in the lower 7-day ADX + B photo (not shown). The number of rats from which hypothalamic sections were obtained is indicated under each photograph.

overlay tracing of light and fluorescence photographs of each section. Virtually all /os-immunoreactive neurons were positive for CRF as well (Fig. 3, A, B, E, F). Few /os-immunoreactive nuclei are visible in the region of the magnocellular AVP neurons; however, many parvocellular neurons expressing FLI were also found to stain for AVP (Fig. 3, C, D, G, H). The colocalization of FLI with that of CRF and AVP was also observed in 3-day ADX rats (data not shown). Basal expression of FLI in rat brain in the morning Many brain areas exhibited nuclear FLI in all rats, regardless of treatment. This basal staining was abol-

ished by preadsorption of the primary antiserum (data not shown). Representative drawings adapted from the atlas of Paxinos and Watson (30) depict the distribution of FLI in Fig. 4. Telencephalon Staining was observed in the cerebral cortex (layers II, III, and V), piriform cortex, lateral septum, accumbens nuclei, most divisions of the amygdala, bed nucleus of the stria terminalis, caudate putamen, and in scattered cells in the hippocampus and dentate gyrus.

CRF NEURONS EXPRESS fos AFTER ADX

Untouched

Sham

ADX

1713

mk I

-B

+B . -•

. S*

100 jjm

•

.

:

.

.

.

•

•

•

•

.

'

*

.

•

•

•

•

•

.

-

•

.

,

:

FIG. 2. Comparison of the effects of surgical and stressless removal of corticosterone feedback on PVN expression of FLI at 12 h. ADX rats maintained on 25 (ig/m\ corticosterone in the drinking fluid had their water bottles changed at lights-off so that some rats continued to drink steroid (+B) while others were given only vehicle fluid (—B). At the same time as water bottles were changed, other rats were ADX, Sham, or left undisturbed (Untouched), n = 3 for all groups.

Diencephalon FLI was found in the organum vasculosum of the lamina terminalis and throughout much of the hypothalamus. With the exception of a dense aggregation of immunoreactive neurons in the ventrolateral suprachiasmatic nuclei and a cluster of neurons in the posterior hypothalamus, hypothalamic FLI tended to be scattered. Although several /os-immunoreactive cells were frequently observed in the supraoptic nuclei, FLI was rarely present in the magnocellular division of the PVN. Other hypothalamic areas expressing FLI were the preoptic area (including the median preoptic nucleus), anterior hypothalamus, dorsomedial nucleus, lateral hypothalamus, arcuate nucleus, and neurons along the ventral edge of the hypothalamus. Staining was also observed in the supramammillary nuclei. In the thalamus and neighboring structures, FLI was identified in midline nuclei (including the paraventricular, centromedian, xiphoid,

rhomboid, and mediodorsal thalamic nuclei), the intergeniculate leaflet, the posterior nucleus, the subparafascicular and prerubral regions of the thalamus, the zona incerta, pretectal nuclei, and a few cells in the lateral habenula. Brainstem A few cells in each of the dorsal, lateral, and ventral divisions of the central gray expressed FLI, as did cells within the optic nerve and intermediate white layers of the superior colliculus. Staining was also evident in the brachium of the inferior conjunctivum and throughout the inferior colliculus. The pedunculopontine, subpenduncular, microcellular, and lateral tegmental nuclei exhibited FLI, as did scattered cells in the raphe pontis and the deep mesencephalic and pontine reticular formations. FLI was strong in the pontine nuclei and extended into the reticulotegmental nucleus and ventral periolivary region. The cuneiform nucleus, parabrachial

1714

CRF NEURONS EXPRESS fos AFTER ADX

Fluorescence

FlG. 3. Coexpression of FLI in the parvocellular PVN with CRF and AVP after ADX. Sections through the PVN of 7day ADX rats were stained first for fos using diaminobenzidine as the chromogen and then for CRF (A, B, E, F; n = 5) or AVP (C, D, G, H; n = 4) using a fluorescein-conjugated second antibody. Arrowheads indicate representative double labeled neurons that were identified by overlay tracing of the light and fluorescence photos. A and B, Fluorescence and light photographs, respectively, of a section stained for fos and CRF. C and D, Fluorescence and light photographs of a section stained for fos and AVP. E and F, Higher power photographs under light and fluorescent illumination, respectively, of a different section from the same rat in A and B, stained for fos and CRF. G and H, Higher power photographs of a different section from the same rat in C and D, stained for fos and AVP under fluorescent and combined light and fluorescent illumination, respectively.

Fluorescence

CRF

100 pm

AVP

B

Endo • 1990 Vol 126 • No 3

Light

Light

CRF NEURONS EXPRESS fos AFTER ADX

1715

B FlG. 4. Basal expression of FLI in rat brain in the morning. Common areas containing /bs-immunoreactive nuclei were plotted on representative brain atlas sections (30) using the camera lucida technique. Plots were constructed from comparison of ADX, Sham, and Untouched rats at the 12 h and 7 day points (n = 3/group). Rat brain atlas drawings and structure name abbreviations were adapted from the atlas of Paxinos and Watson (30) with the authors' and publisher's permission.

nucleus, locus ceruleus, and areas coinciding with the A5 and A7 catecholaminergic cell groups were also fos immunoreactive. Staining was further observed in the medullary reticular formation, the lateral reticular nucleus, the dorsomedial and ventromedial aspects of the spinal trigeminal nucleus (predominantly interpolaris and cau-

dalis), and throughout the rostrocaudal extent of the nucleus of the solitary tract.

Discussion We have shown that ADX induces a sustained high level of FLI in CRF- and AVP-staining neurons of the

1716

CRF NEURONS EXPRESS fos AFTER ADX

parvocellular PVN, and that PVN expression of /os-like proteins can be both prevented and reversed by corticosterone. The induction of FLI in the medial parvocellular PVN originally suggested that these neurons might be CRF neurons, which are localized in this nuclear subdivision and exhibit specific increases in gene expression and secretory activity after ADX (12-20, 27, 29). ADXinduced increases in CRF or AVP expression are normalized relative to those in sham-operated animals by glucocorticoid replacement (12-14, 17, 31). Likewise, we have shown that ADX-induced expression of FLI in the parvocellular PVN could be prevented by corticosterone replacement at ADX and reversed by corticosterone replacement 5 days after surgery. The colocalization of FLI with that of CRF or AVP in parvocellular neurons strengthens the correlation of fos induction with changes in neuroendocrine activity after ADX.

Endo • 1990 Vol 126 • No 3

The responsiveness of PVN FLI to removal and replacement of corticosterone strongly suggests that glucocorticoids directly or indirectly regulate expression of c-fos and related genes in CRF neurons. Most CRF neurons express glucocorticoid receptor immunoreactivity (32, 33); the inhibition of ADX-induced increases in CRF and AVP immunoreactivity by dexamethasone implants over the PVN (34, 35) indicates that glucocorticoids may directly affect CRF neurons. However, it is also likely that corticosterone effects are mediated by afferent inputs from other glucocorticoid-sensitive brain areas. Extrahypothalamic feedback regulation of the hypothalamic-pituitary-adrenal axis has been inferred from the presence of corticosteroid receptors and corticosteroid-sensitive neurons, most notably in the hippocampus and lateral septum, and from the effects of localized corticosterone implants on ACTH secretion

CRF NEURONS EXPRESS fos AFTER ADX

(36-41). Any one or a combination of these extrahypothalamic sites might also account for the effects of corticosterone on PVN FLI. There were two instances in which the expression of FLI in the parvocellular PVN deviated from the known changes in plasma ACTH that are mediated by corticosterone. First, in ADX rats implanted with a corticosterone pellet (ADX + B), five of seven rats at 7 days and one of four rats at 1 day exhibited FLI in the parvocellular PVN, even though plasma corticosterone concentrations in these rats were within the range observed at the respective time points. In contrast to the frequently high levels of FLI at 7 days, morning plasma ACTH has been found to be normalized even by the lower plasma corticosterone levels of approximately 5 ixg/d\ measured in the 7-day ADX + B rats (23). Second, provision of 25 /ug/ml corticosterone in the drinking water did not prevent the expression of FLI even though this treatment also normalizes morning plasma ACTH up to 19 days after surgery (24). We had expected that like morning plasma ACTH, PVN expression of /os-like proteins in ADX rats maintained on corticosterone in their water (+B) would be low. Instead, the high level of PVN FLI in +B rats as well as in rats switched to drinking steroidfree fluid (—B) obscured any differences in the effects of surgical us. stressless removal of corticosterone on FLI. The discrepancy between PVN FLI and the probable levels of plasma ACTH in ADX rats replaced with corticosterone pellets or 25 Mg/ml corticosterone in the drinking water may indicate that the plasma corticosterone levels required to normalize CRF mRNA and plasma ACTH (13, 23) are not sufficient to normalize PVN FLI after ADX. The level of corticosterone in both of these models represents a physiological, rather than supraphysiological, replacement because it normalizes morning plasma ACTH, pituitary ACTH content, and thymus weight relative to those in Sham rats (23, 24) (our unpublished data). However, neither form of replacement normalizes circadian peak (evening) levels of plasma ACTH, which are amplified almost 4-fold relative to those in Sham rats. These circadian increases in plasma ACTH are only inhibited at much higher corticosterone levels that suppress ACTH responses to stress and produce thymic atrophy (23, 31, 42). Such higher supraphysiological corticosterone concentrations may be required for consistent inhibition of ADX-induced FLI. We suspect, therefore, that the high levels of immunoreactivity observed in the ADX -I- B rat at 1 day, in the five ADX + B rats at 7 days, and in the ADX rats given 25 Mg/ml in the drinking fluid (+B) may reflect the greater circadian hypothalamic-pituitary activity remaining at corticosterone levels that normalize morning plasma ACTH. Preliminary results from a circadian study of fos expression after stressless removal of corticosterone also indi-

1717

cate that parvocellular expression of FLI is high in both +B and - B rats in the morning (Plotsky, P. M., and P. E. Sawchenko, personal communication). However, PVN expression of/bs-like proteins is clearly responsive to corticosteroid inhibition, even if the levels required appear to be higher than those necessary to normalize morning plasma ACTH. Provision of 100 ixg/ ml corticosterone in the drinking water to 5-day ADX rats was sufficient to reverse expression of /os-like proteins within 2 days. It is likely that the plasma corticosterone concentrations were at least twice as high in rats treated with 100 fig/ml corticosterone as in rats treated with 25 ixg/m\, since previous experiments in our laboratory indicate that quadrupling the concentration of corticosterone in the drinking water doubles the circulating steroid levels (42). The inhibition of PVN FLI in most 1- and 3-day and in some 7-day ADX + B rats also demonstrates that corticosteroids can inhibit ADX-induced changes in /os-like proteins. Because FLI was always high in the ADX unreplaced rats killed with the ADX + B rats, the frequent low levels of FLI in ADX + B rats are attributable to corticosterone replacement rather than to inherent variability in FLI in ADX rats. The more consistent inhibition of PVN FLI in 1- and 3day ADX + B rats may be due to the higher corticosterone levels produced 1 day after pellet implantation and the slow response of the hypothalamic-pituitary axis (24) to the decline of those levels. It is interesting to note that ADX induced prolonged expression of/os-like proteins in the PVN, whereas many stimuli have been reported to evoke only transient expression of the c-/os gene (1, 43). This difference may be due to either differences in the nature of the stimulus or the possibility that our antiserum may recognize nuclear proteins related to fos. ADX may represent a more sustained stimulus than many of the manipulations previously used (1, 5-7) to induce c-/os gene expression. Water deprivation, which is also a sustained stimulus, has been found to induce expression of fos immunoreactivity in the magnocellular PVN and supraoptic nuclei 24 h later (2, 8); infraorbital nerve section induces FLI in the spinal trigeminal nucleus for at least 2 weeks (44). Thus, although fos has been shown to inhibit expression of its own gene (45) and although stimuli repeated within a few hours have been reported to be less effective in inducing fos expression (1), it is possible that prolonging a stimulus over a longer time frame may overcome such negative regulation. It is also highly likely that our antibody detects more than one /os-like protein. Three /os-related antigens (fra) have been identified using an antibody to fos residues 127-152 (46). Because our antiserum was raised against similar residues (132-154), it may recognize some or all of these proteins. One/os-related antigen, fra-1, has been

1718

CRF NEURONS EXPRESS fos AFTER ADX

shown to have a slightly delayed time course of induction relative to that of authentic fos in serum-stimulated fibroblasts (46). The sustained FLI in the PVN after ADX may, therefore, represent expression of fos and a succession of /os-related proteins. Nevertheless, the physiological significance of these proteins is clearly demonstrated by their responsiveness to corticosteroid inhibition. We observed a widespread distribution of FLI that was not affected by ADX. We are confident that this staining represents an unstressed level of expression, since rats were rapidly anesthetized and perfused, well within the time necessary for significant fos synthesis (1). The comparison of staining patterns in Sham and Untouched rats also indicated that the handling necessary to administer anesthesia (Untouched rats) and, at 1, 3, and 7 days, the effect of surgical incisions (Sham rats) had little impact on fos expression. The light parvocellular fos immunoreactivity in 12-h Sham animals may be attributable to the recent surgery, but this effect subsided by 1 day. Some of the observed staining, particularly that in the suprachiasmatic nucleus, might depend on the time of day that the rats were killed. However, at least some of the /os-like staining observed may represent a constitutive level of expression, since the hippocampal complex, bed nucleus of the stria terminalis, and cerebral cortex have been reported to express fos when no specified attempt was made to control for stressful or circadian influences (1, 2). We were surprised to find that of all the brain sites potentially responsive to glucocorticoids, only the PVN exhibited changes in /os-like proteins after ADX. Several extrahypothalamic areas suggested to be corticosteroid feedback sites, including the limbic system, cerebral cortex, and certain cranial nerve nuclei (38-41), expressed FLI that was unaffected by ADX. It has been suggested (Sagar, S. M., personal communication) that fairly strong stimuli or changes relative to baseline stimulation may be required to observe a change in fos expression. Another interpretation of this view is that the PVN is the convergence of inputs from a number of brain sites responding to the loss of corticosteroid feedback, and that, at least within the time frame examined here, only this convergence provides a sufficiently strong signal to induce fos expression. Alternatively, it is possible that brain fos expression is regulated only by certain neurotransmitters or second messenger systems, such that the signal reflected in increased expression of FLI at the parvocellular PVN may be qualitatively different from that occurring at other brain feedback sites. The demonstration of FLI in the PVN, which has been shown to increase metabolic (10) and electrical (11) activity after ADX, supports the suggestion that fos expression may be a marker of neural activity (1,2). Fos

Endo • 1990 Vol 126 • No 3

has been proposed to participate in transcriptional regulation via interaction with the c-jun/APl protein at specific DNA response elements (47). Thus, fos could conceivably function as a transcriptional activator in enhancing CRF and AVP expression after ADX. However, fos expression is probably not a general marker of neuronal activity. The patterns of FLI found after motor/sensory cortex stimulation were similar but not identical to the patterns of [14C]2-deoxyglucose uptake (7). In addition to the PVN, both the locus ceruleus and hippocampus were reported to increase glucose utilization after ADX (10). While the discrepancy between these results and the present data may be attributed to differences in the time points examined (5 h in the former study us. 12 h to 7 days), it is equally likely that fos expression may only identify certain classes of neurons responding to a given stimulus. Nevertheless, the ability to identify such neurons individually by double label immunocytochemistry still makes this approach very powerful. These experiments have shown that expression of FLI in brain after ADX is consistent with characteristic changes in neuroendocrine activity of the adrenocortical system. In doing so, they further confirm the role of the brain in the ACTH response to ADX (48), provide support for the use of fos expression as a marker of activity in certain classes of neurons, and are one of the first demonstrations of the identity of neurons expressing fos after a specific stimulus.

Acknowledgments We are indebted to Dr. Susan F. Akana for valuable advice, to Dr. Stephen M. Sagar for providing affinity-purified fos antiserum and helpful suggestions, to Dr. Wylie W. Vale for his gift of ovine CRF antiserum, to Margaret Bradbury for assistance in determining corticosterone pellet replacement levels, and to Martin P. Sorette for dedicated assistance with the graphics.

References 1. Morgan JI, Cohen DR, Hempstead JL, Curran T 1987 Mapping patterns of c-fos expression in the central nervous system after seizure. Science 237:192 2. Sagar SM, Sharp FR, Curran T 1988 Expression of c-fos protein in brain: metabolic mapping at the cellular level. Science 240:1328 3. Morgan JI, Curran T 1986 Role of ion flux in the control of c-fos expression. Nature 322:552 4. Bartel DP, Sheng M, Lau LF, Greenberg ME 1989 Growth factors and membrane depolarization activate distinct patterns of early response gene expression: dissociation of fos and jun expression. Genes Dev 3:304 5. Dragunow M, Robertson HA 1987 Kindling stimulation induces cfos protein(s) in granule cells of the rat dentate gyrus. Nature 329:441 6. Hunt SP, Pini A, Evan G 1987 Induction of c-/os-like protein in spinal cord neurons following sensory stimulation. Nature 328:632 7. Sharp FR, Gonzalez MF, Sharp JW, Sagar SM 1989 c-fos expression and [14C]2-deoxyglucose uptake in caudal cerebellum during motor/sensory cortex stimulation in the rat. J Comp Neurol 284:621 8. Sagar SM, Sharp FR 1988 Dehydration induces fos immunostaining

CRF NEURONS EXPRESS fos AFTER ADX

9. 10. 11. 12.

13.

14.

15.

16.

17.

18.

19.

20. 21. 22. 23. 24.

25. 26. 27.

in hypothalamic magnocellular neurons. 18th Annual Meeting of the Society for Neuroscience, Toronto, Ontario, Canada, part 1, p 307 Sharp FR, Gonzalez MF, Hisanaga K, Mobley WC, Sagar SM 1989 Induction of the c-fos gene product in rat forebrain following cortical lesions and NGF injections. Neurosci Lett 100:117 Kadekaro M, Ito M, Gross PM 1988 Local cerebral glucose utilization is increased in acutely adrenalectomized rats. Neuroendocrinology 47:329 Kasai M, Yamashita H 1988 Inhibition by cortisol of neurons in the paraventricular nucleus of the hypothalamus in adrenalectomized rats; an in vitro study. Neurosci Lett 91:59 Jingami H, Matsukura S, Numa S, Imura H 1985 Effects of adrenalectomy and dexamethasone administration on the level of preprocorticotropin-releasing factor messenger ribonucleic acid (mRNA) in the hypothalamus and adrenocorticotropin//3-lipotropin precursor mRNA in the pituitary in rats. Endocrinology 117:1314 Beyer HS, Matta SG, Sharp BM 1988 Regulation of the messenger ribonucleic acid for corticotropin-releasing factor in the paraventricular nucleus and other brain sites of the rat. Endocrinology 123:2117 Davis LG, Arentzen R, Reid JM, Manning RW, Wolfson B, Lawrence KL, Baldino Jr F 1986 Glucocorticoid sensitivity of vasopressin mRNA levels in the paraventricular nucleus of the rat. Proc Natl Acad Sci USA 83:1145 Swanson LW, Simmons DM 1989 Differential steroid hormone and neural influences on peptide mRNA levels in CRH cells of the paraventricular nucleus: a hybridization histochemical study in the rat. J Comp Neurol 285:413 Kiss JZ, Mezey E, Skirboll L 1984 Corticotropin-releasing factorimmunoreactive neurons of the paraventricular nucleus become vasopressin positive after adrenalectomy. Proc Natl Acad Sci USA 81:1854 Sawchenko PE 1987 Adrenalectomy-induced enhancement of CRF and vasopressin immunoreactivity in parvocellular neurosecretory neurons: anatomic, peptide and steroid specificity. J Neurosci 7:1093 Whitnall MH 1988 Distributions of pro-vasopressin expressing and pro-vasopressin deficient CRH neurons in the paraventricular hypothalamic nucleus of colchicine-treated normal and adrenalectomized rats. J Comp Neurol 275:13 Plotsky PM, Sawchenko PE 1987 Hypophysial-portal plasma levels, median eminence content and immunohistochemical staining of corticotropin-releasing factor, arginine vasopressin, and oxytocin after pharmacological adrenalectomy. Endocrinology 120:1361 Fink G, Robinson ICAF, Tannahill LA 1988 Effects of adrenalectomy and glucocorticoids on the peptides CRF-41, AVP and oxytocin in rat hypophysial portal blood. J Physiol 401:329 Jacobson L, Sharp FR, Adrenalectomy induces prolonged c-fos expression in the paraventricular nucleus. 71st Annual Meeting of The Endocrine Society, Seattle WA, 1989, p 458 Jacobson L, Dallman MF 1988 ACTH and ventilation increase at similar arterial PO2 in conscious rats. J Appl Physiol 66:2245 Akana SF, Cascio CS, Shinsako J, Dallman MF 1985 Corticosterone: narrow range required for normal body and thymus weight and ACTH. Am J Physiol 249:R527 Jacobson L, Akana SF, Cascio CS, Scribner K, Shinsako J, Dallman MF 1989 The adrenocortical system responds slowly to removal of corticosterone in the absence of concurrent stress. Endocrinology 124:2144 McLean IW, Nakane PK 1974 Periodate-lysine-paraformaldehyde fixative: a new fixative for immunoelectron microscopy. J Histochem Cytochem 22:1077 Curran T, Van Beveren C, Ling N, Verma IM 1985 Viral and cellular proteins are complexed with a 39,000-dalton cellular protein. Mol Cell Biol 5:167 Swanson LW, Sawchenko PE, Rivier J, Vale WW 1983 Organization of ovine corticotropin-releasing factor immunoreactive cells and fibers in the rat brain: an immunohistochemical study. Neuroendocrinology 36:165

1719

28. Logsdon CD, Akana SF, Meyer C, Dallman MF, Williams JA 1987 Pancreatic acinar cell amylase gene expression: selective effects of adrenalectomy and corticosterone replacement. Endocrinology 121:1242 29. Antoni F, Palkovits M, Makara GB, Linton EA, Lowry PJ, Kiss JZ 1983 Immunoreactive corticotropin-releasing hormone in the hypothalamoinfundibular tract. Neuroendocrinology 36:415 30. Paxinos G, Watson C 1986 The Rat Brain in Stereotaxic Coordinates, ed 2. Academic Press, Sydney 31. Akana SF, Cascio CS, Du J-Z, Levin N, Dallman MF 1986 Reset of feedback in the adrenocortical system: an apparent shift in sensitivity of adrenocorticotropin to inhibition by corticosterone between morning and evening. Endocrinology 119:2325 32. Uht RM, McKelvy JF, Harrison RW, Bohn MC 1988 Demonstration of glucocorticoid receptor-like immunoreactivity in glucocorticoid-sensitive vasopressin and corticotropin-releasing factor neurons in the hypothalamic paraventricular nucleus. J Neurosci Res 19:405 33. Cintra A, Fuxe K, Harfstrand A, Agnati LF, Wikstrom A-C, Okret S, Vale W, Gustafsson J-A 1987 Presence of glucocorticoid receptor immunoreactivity in corticotrophin releasing factor and in growth hormone releasing factor immunoreactive neurons of the rat diand telencephalon. Neurosci Lett 77:25 34. Kovacs K, Kiss JZ, Makara GB 1986 Glucocorticoid implants around the hypothalamic paraventricular nucleus prevent the increase of corticotropin-releasing factor and arginine vasopressin immunostaining induced by adrenalectomy. Neuroendocrinology 44:229 35. Sawchenko PE 1987 Evidence for a local site of action for glucocorticoids in inhibiting CRF and vasopressin expression in the paraventricular nucleus. Brain Res 403:213 36. De Kloet ER, Ratka A, Reul JMHM, Sutanto W, Van Eekelen JAM 1987 Corticosteroid receptor types in brain: regulation and putative function. In: Ganong WF, Dallman MF, Roberts JL (eds) The Hypothalamic-Pituitary-Adrenal Axis Revisited. New York Academy of Sciences, New York, p 351 37. Fuxe K, Cintra A, Harfstrand A, Agnati LF, Kalia M, Zoli M, Wikstrom A-C, Okret S, Aronsson M, Gustafsson J-A 1987 Central glucocorticoid receptor immunoreactive neurons: new insights into the endocrine regulation of the brain. In: Ganong WF, Dallman MF, Roberts JL (eds) The Hypothalamic-Pituitary-Adrenal Axis Revisited. New York Academy of Sciences, New York, p 362 38. Keller-Wood M, Dallman MF 1984 Corticosteroid inhibition of ACTH secretion. Endocr Rev 5:1 39. Steiner FA, Ruf K, Akert K 1969 Steroid-sensitive neurones in rat brain: anatomical localization and responses to neurohormones and ACTH. Brain Res 12:74 40. Vidal C, Jordan W, Zieglgansberger W 1986 Corticosterone reduces the excitability of hippocampal pyramidal cells in vitro. Brain Res 383:54 41. Kovacs K, Makara GB 1988 Corticosterone and dexamethasone act at different brain sites to inhibit adrenalectomy-induced adrenocorticotropin hypersecretion. Brain Res 474:205 42. Wilkinson CW, Engeland WC, Shinsako J, Dallman MF 1981 Nonsteroidal adrenal feedback demarcates two types of pathways to CRF-ACTH release. Am J Physiol 24O:E136 43. Curran T 1988 The fos oncogene. In: Reddy EP (ed) The Oncogene Handbook. Elsevier, New York, p 307 44. Sharp FR, Griffith J, Gonzalez MF, Sagar SM 1989 Trigeminal nerve section induces/os-like immunoreactivity (FLI) in brainstem and decreases FLI in sensory cortex. Mol Brain Res 6:217 45. Sassone-Corsi P, Sisson JC, Verma IM 1988 Transcriptional autoregulation of the proto-oncogene fos. Nature 334:314 46. Cohen DR, Curran T 1988 Fra-1: a serum-inducible, cellular immediate-early gene that encodes a /os-related antigen. Mol Cell Biol 8:2063 47. Chiu R, Boyle WJ, Meek J, Smeal T, Hunter T, Karin M 1988 The c-fos protein interacts with c-jun/A.P-1 to stimulate transcription of AP-1 responsive genes. Cell 54:541 48. Levin N, Shinsako J, Dallman MF 1988 Corticosterone acts on v,he brain to inhibit adrenalectomy-induced ACTH secretion. Endocrinology 122:694