Neurotransmitter systems regulating behavioral responding in stress

MISSING IMAGE

Material Information

Title:
Neurotransmitter systems regulating behavioral responding in stress
Physical Description:
Book
Creator:
Berridge, Craig W., 1959-
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 20302902
System ID:
AA00020010:00001

Full Text






NEUROTRANSMITTER SYSTEMS REGULATING BEHAVIORAL
RESPONDING IN STRESS


BY

CRAIG W. BERRIDGE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1988










ACKNOWLEDGEMENTS


I wish to thank Dr. Adrian Dunn for his support, energy, and knowledge
that was freely provided, in and out of the laboratory. Also greatly appreciated,
was the support of the other members of my supervisory committee, Drs. Tiana

Leonard, Bill Luttge, and Neil Rowland. Not only did they provide technical
advice that contributed towards the successful completion of my project, but
they also gave much needed moral support and advice on questions and
problems that arose over the years. I am also thankful to the Department of

Neuroscience for providing a positive and stimulating environment in which I
was able to pursue my interests.
I am especially grateful to my friends who made me feel at home during
my stay in Gainesville. Special thanks go to Tom Nutter and Lee Bloomcamp

for getting me involved in diving, showing me Florida and for sharing what was

theirs. Also, I am grateful to Diana and John, Doug, Jim, Dave and Kelly for
keeping life interesting.
Finally, I would not have made it to this point without the constant support
from my family, especially (need it be said?) my parents, to whom I am forever

grateful.







TABLE OF CONTENTS


ACKNOWLEDGEMENTS .

KEY TO ABBREVIATIONS .


. . . . . . . . v i


CHAPTER I.
Background . . . . .
Corticotropin Releasing Factor .
Norepinephrine . . . .
Rationale . . . . .

CHAPTER II. GENERAL METHODS


. . . . . . . 15


CHAPTER IM. INVOLVEMENT OF ENDOGENOUS CRF IN THE
RESTRAINT-INDUCED DECREASE IN EXPLORATORY BEHAVIOR
Introduction . . . . . . . . . . . 2 1
Methods .. . . . ... ........... 23
R results . . . . . . . . . . . . 24
Discussion . . . . . . . . . . . 43


CHAPTER IV. INVOLVEMENT OF NOREPINEPHRINE IN THE
RESTRAINT-INDUCED DECREASE IN EXPLORATORY BEHAVIOR
Introduction . . . . . . . . . . . 48
M ethods . . . . . . . . . . . . 53
R results . . . . . . . . . . . . 54
D discussion . . . . . . . . . . . . 74


CHAPTER V. NORADRENERGIC-CRF INTERACTIONS IN REGULATING
EXPLORATORY BEHAVIOR
Introduction . . . . . . . . . . . 77
M ethods . . . . . . . . . . . . 7 8
Results . . . . . . . . . . . ... 7 8
D discussion . . . . . . . . . . . . 8 5


ABSTRACT


. . . . . . . . ii









CHAPTER VI. GENERAL DISCUSSION . . .

REFERENCES . . . . . . . .

BIOGRAPHICAL SKETCH . . . . .


. . . . 89

. . . . 99

. . . 114










KEY TO ABBREVIATIONS


ahCRF Alpha-helical corticotropin-releasing factor9g-41
CRF Corticotropin-releasing Factor

CNS Central nervous system
CSF Cerebrospinal fluid
DA Dopamine
DOPAC 3,4-Dihydroxyphenylacetic acid
HVA Homovanillic acid
5-HIAA 5-Hydroxyindoleacetic acid
5-HT Serotonin
LC Locus coeruleus
MHPG 3-methoxy,4-hydroxyphenylethylene glycol
NE Norepinephrine










Abstract of Dissertation Presented to the Graduate School of the University of
Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of
Philosophy



NEUROTRANSMITTER SYSTEMS REGULATING BEHAVIORAL
RESPONDING IN STRESS



By

Craig W. Berridge

December 1988


Chairman: Adrian J. Dunn
Major Department: Neuroscience


An activation of cerebral noradrenergic and corticotropin-releasing factor

(CRF) systems has been postulated to be of functional importance in stress.
The relative role of each of these systems in stress-induced changes in
behavior was examined. Exploratory behavior, as measured by the time an
animal spends investigating objects in a novel environment, is decreased by
prior exposure of the animal to a number of stressors. CRF injected

intracerebroventricularly (ICV) elicited a stress-like, dose-dependent decrease
in exploratory behavior in both rats and mice. The CRF-induced decrease in
exploratory behavior, like that observed after restraint-stress, was antagonized
by the opiate-antagonist, naloxone. In rats in which the cerebral aqueduct (CA)
had been blocked, CRF decreased exploratory behavior when injected into the
lateral ventricles, but not when injected into the fourth ventricle, suggesting that










CRF acts at a site anterior to the CA to decrease exploratory behavior. ICV
administration of a CRF-antagonist (alpha-helical CRF9-41; ahCRF) dose-
dependently reversed the restraint-stress-induced decrease in exploratory
behavior. These results suggest that endogenous CRF acts to mediate the
stress-induced decrease in exploratory behavior.
An increased release of NE within the brain is observed during stress

and following ICV administration of CRF. Therefore, the involvement of
noradrenergic systems in mediating the behavioral effect of restraint and CRF
was examined. Stimulation of NE release using the a2-antagonist, idazoxan (1
mg/kg), decreased, and the a2-agonist, clonidine (25 gg/kg) increased

exploratory behavior. Inhibition of NE release using clonidine or the
noradrenergic-selective neurotoxin, DSP-4, antagonized the restraint-induced
decrease in exploratory behavior. The combination of these two treatments
completely prevented the behavioral effect of restraint. The ai-receptor

antagonist, prazosin (200 jg/kg), also prevented the behavioral effect of
restraint, whereas the ai-agonist, phenylephrine (50 or 100 ng) decreased

exploratory behavior. Neither DSP-4 nor prazosin had any effect on the CRF-
induced decrease in exploratory behavior. However, ahCRF (20 jg) ICV
reversed the phenylephrine-induced decrease in this behavior.
These results indicate that both CRF and NE regulate exploratory
behavior and that both CRF and noradrenergic systems may mediate the
behavioral effect of restraint stress in this paradigm. It is concluded that during
stress, these two systems do not act independently, but that NE stimulates the
release of CRF via an ai -receptor, which then influences exploratory behavior

in this novel environment.













CHAPTER I
INTRODUCTION


As commonly used, the term "stress" generally refers to a cognitive or

emotional response, often associated with an increase in arousal and
accompanied by certain physiological responses, such as increases in heart
rate and blood pressure. Although poorly defined, when used in this context,

stress is a common and well recognized concept in modern society. Because

cognitive and emotional responses are subjective in humans, and difficult to
measure in animals, research on stress has primarily involved physiological

measures.
In 1911, Cannon described the stimulation of catecholamines secretion

from the adrenal medulla by both physical (pain and asphyxiation) and

psychological stimuli (exposure of a cat to an aggressively barking dog)
(Cannon & de la Paz, 1911; Cannon etal., 1911; Cannon, 1914). The
increased release of catecholamines in turn increased the availability of blood
glucose, increased the strength of muscular contractions, and, in association
with the sympathetic nervous system, increased blood pressure and
redistributed blood flow to the heart, lungs and skeletal muscles. This
sympatho-adrenal-medullary response, often referred to as the "flight or fight"
response, is important because increasing the efficiency of behavioral

responding facilitates adaptive responding in an emergency situation.

Selye expanded our understanding of the physiological nature of stress
when he described a series of responses elicited by various noxious stimuli








(Selye, 1936). These responses included rapid decreases in the size of the
thymus thymicc involution), spleen, and liver, decreased body temperature,
lesions of the digestive tract, and an enlargement of the adrenal glands,
suggesting an activation of the adrenals. Thus, in his view, a stressor was any
stimulus, either physical or emotional that evoked this response. A key feature
of responding in stress was that it was of a nonspecific nature; the same set of
responses was observed following exposure to any physical or psychological
stressor. As described by Cannon, many of these responses served to increase
the ability of an animal to resist, or cope with, the stressor. Although a series of
biochemical and pathological responses were described by Selye, in his view a
critical component of this general response was the activation of the pituitary-
adrenal axis, involving the release of adrenocorticotropic hormone (ACTH) from
the pituitary, in turn stimulating the secretion of the glucocorticoids (cortisol in
the primate and corticosterone in the rodent) from the adrenal cortex (Selye,
1946).
The importance of Selye's theory was that it defined stress as the
response of the organism and suggested that very different stimuli could be
grouped under a unifying heading stressorr) based on their ability to elicit the

complex of physiological responses. Mason extended the observations of
Selye in a series of experiments characterizing responding of the pituitary-
adrenal axis to psychological stimuli (Mason, 1968a, 1968b). In these studies,
he demonstrated that the stressor-induced activation of the pituitary-adrenal
axis was not observed if the emotional distress associated with a stressor was
eliminated. For example, the increase in plasma cortisol concentrations
normally observed in food-deprived monkeys was absent if fruit-flavored,
nonnutritive cellulose pellets were provided (Mason, 1968a, 1968b).








More recent evidence suggests that different patterns of responding can
be associated with different stressors. In addition to an activation of the
pituitary-adrenal and the sympatho-adrenomedullary axes, a decrease in the
release of the gonadal steroids (Selye, 1946) is also observed in stress.
Although all of these responses often occur simultaneously in response to a
given stressor, this is not always the case. When the plasma cortisol and
epinephrine (Epi) responses of men undergoing paratrooper training were
examined, the cortisol response disappeared after the first jump, whereas the
Epi response, although diminished, was present up to the last jump (Levine,
1985). Further, in the development of dominance relationships in pairs of male
monkeys, an initial period of aggression is observed normally (Coe et al., 1982).
Associated with this period of aggression, all animals (both dominant and
submissive) were observed to have elevated concentrations of plasma cortisol.
However, monkeys that became dominant had increased concentrations of
plasma testosterone, whereas the subordinate monkeys tended to have
decreased concentrations of plasma testosterone (Coe et al., 1982).
Behavioral responses frequently accompany the physiological
responses to a stressor and presumably facilitate coping with the stressor.
These behavioral responses vary depending on the stressor, additional
environmental cues, and the prior experience of an individual animal. For
example, depending on the testing paradigm, increased and decreased
locomotor activity, decreased feeding, decreased sexual behavior and
increased aggressive behavior have all been associated with stress. Not only
do behavioral responses allow an animal to remove itself from the presence of
the stressor, they may directly facilitate coping with a stressor. For example, rats
that are shocked in pairs will often engage in fighting behaviors. These animals
were observed to have lower concentrations of plasma ACTH (Conner et al.,








1971) and lower blood pressure (Williams & Eichelman, 1971) than animals
that were shocked singly. Further, the corticosterone response of rats placed in
a novel environment was diminished by the gnawing of inedible objects, a
response that occurred in preference to the gnawing or eating of edible objects
(Hennessy & Foy, 1987).
To summarize, a complex pattern of responding involving behavioral and
physiological responses is associated with stress, the exact pattern of which
varies depending on the stressor and the environmental conditions. All of these
responses are regulated by the central nervous system (CNS). The
mechanisms by which the CNS initiates and coordinates such responding and
how this responding facilitates adaptation to a stressor are poorly understood.
Neurochemical systems within the brain that have been implicated in regulating
responding in stress include the catecholamines, norepinephrine (NE) and
dopamine (DA), and the opioid peptides. Thus, exposure of an animal to a
stressor elicits an increase in the release of NE throughout the brain, and a
somewhat more regionally specific release of DA, with the greatest sensitivity
for this response observed in the prefrontal cortex (Thierry, et al., 1976: Dunn &
Kramarcy, 1984). The functional significance of these neurochemical
responses is unknown. However, NE has been implicated in regulating
emotional and behavioral responding associated with anxiety (Redmond &
Huang, 1979; Charney et al., 1983). In the frontal cortex of primates, DA
(Brozoski et al., 1979) and NE (Arnsten & Goldman-Rakic, 1985) have been
suggested to regulate responding in a delayed alternation task, a test sensitive
to high-intensity noise stress (Arnsten, personal communication). Opioid-
dependent, stress-induced analgesia has been described frequently (Bodnar et
al., 1980), and a stress-induced change in exploratory behavior was
antagonized by the opiate-antagonist, naloxone, (Arnsten et al., 1985). Thus, it








is likely that the activation of these neurochemical systems may regulate
responding in stress. Increases in the concentration of brain tryptophan and an
enhanced release of serotonin have also been observed, although the
functional significance of these responses is also unknown (Dunn, 1988).

Corticotropin-Releasing Factor
Recently, evidence has suggested that corticotropin-releasing factor
(CRF) might also be released in the brain during stress and participate in
regulating stress-induced responding. CRF is a 41-amino acid peptide isolated
and characterized by Vale et al. (1981). Synthesized in cells of the
paraventricular nucleus of the hypothalamus (PVN) and released from the
median eminence of the hypothalamus into the portal blood supply, CRF
stimulates the release of ACTH from the anterior pituitary which subsequently
stimulates the secretion of glucocorticoids. Elimination of the activation of this
hypothalamo-pituitary-adrenal (HPA) axis seriously compromises the ability of
the animal to survive periods of prolonged stress (Maickel et al., 1967).
In addition to hypothalamic CRF, CRF-like material and specific, high-
affinity binding sites for CRF have been identified in many extrahypothalamic
regions of the CNS. Identification of this CRF-like material within the CNS has
involved both detection of immunoreactive-CRF (irCRF) (Merchenthaler, 1984;
Swanson et al., 1983; Moga & Gray, 1985), and more recently a bioassay
measuring ACTH release from anterior pituitary cells in vitro (Nakane et al.,
1986). Both techniques have indicated a similar pattern of localization of CRF
within the brain. In general, the highest concentration of extra-hypothalamic
irCRF is found in neocortex, areas of the limbic system, and regions involved in
the regulation of the autonomic nervous system (Swanson et al., 1983;
Merchenthaler 1984; Moga & Gray, 1985).








Within the telencephalon, the greatest number of irCRF-containing cell
bodies was found in the prefrontal, insular, and cingulate cortices with cells
localized primarily to layers 11-111, and processes extending through layers I-IV
(Merchenthaler, 1984). Subcortical telencephalic structures containing irCRF
include the hippocampus, the bed nucleus of the stria terminalis (BNST), and
the central amygdaloid nucleus. A dense cluster of irCRF-containing cells is
found in the PVN; however, most hypothalamic nuclei contain some irCRF and
send minor projections to the median eminence (Merchenthaler, 1984). Other
regions containing relatively high concentrations of ir-CRF include certain
thalamic nuclei, the substantial nigra pars compact, the periaqueductal grey
(primarily in the ventral portion), the locus coeruleus (LC), the nucleus of the
solitary tract, the dorsal and ventral parabrachial nuclei, and the cortex and
deep nuclei of the cerebellum (Swanson et al., 1983, Merchenthaler, 1984,
Sakanaka et al., 1987). CRF-containing cells and fibers were also observed
around most sensory nuclei of the cranial nerves and within the spinal cord in
lamina V-VIII and X and in the intermediolateral cell column. Although the
projection patterns of irCRF-containing neurons have not been described in
detail, a projection from the amygdala to the parabrachial nucleus (Moga &
Gray, 1985) and from the inferior olivary nucleus to the cerebellum have been
observed (Powers et al., 1987). In addition, CRF-containing fibers from cells in
the PVN project to various brainstem nuclei (Swanson et al., 1986). The
observation of a K+-stimulated, Ca2+-dependent release of CRF from brain
tissue in vitro, suggests that the release of CRF might be regulated like that of
other neurotransmitters (Smith et al., 1986).
Using both quantitative autoradiography and cell membrane binding
techniques, high-affinity binding sites for CRF have been observed in a pattern
similar to that observed for irCRF (Wynn et al., 1984; De Souza et al., 1985).








The highest density of CRF-binding sites was observed throughout neocortex,
especially in layers I and IV. In addition, high-density binding was also
apparent in the cerebellar cortex and deep cerebellar nuclei, BNST, and
various cranial nuclei. Moderate binding was observed in the amygdala,
hippocampus, dorsolateral thalamic nuclei, and the LC. Using CRF and CRF

analogs, it has been demonstrated that the binding characteristics of these sites
are similar to those described for pituitary CRF-binding sites. Binding of CRF or
CRF analogs in both rat and primate brain was saturable, reversible and of a
high affinity with a dissociation constant (Kd) in the nanomolar range (De Souza
et al., 1984; Chen et al., 1986).
CRF-binding in the pituitary stimulates the activation of adenylate cyclase
and the accumulation of cyclic adenosine monophosphate (cAMP), an action
thought to regulate the release of ACTH (Wynn et al., 1984; Giguere et al.,
1982). Similarly, CRF-binding in the brain stimulates activation of adenylate
cyclase (Wynn et al., 1984; Chen et al., 1986). The potency of cyclase
activation by various analogs correlates with their effectiveness in stimulating
ACTH secretion from the pituitary (Chen et al., 1986). Further, the CRF-
antagonist, alpha-helical CRF9-41 (ahCRF), inhibits CRF-stimulated adenylate
cyclase activity in the brain. Thus, it appears that at least some brain CRF-
binding sites are linked to adenylate cyclase activation. Like other receptors
coupled to adenylate cyclase, CRF stimulation of cyclase in the brain is
regulated by guanine nucleotides and divalent cations (Chen et al., 1986).
Adrenalectomy decreases the number of CRF-binding sites in the
pituitary (Wynn et al., 1984). It is thought that the decrease in binding site
number is a consequence of an increase in the rate of receptor internalization
resulting from an increase in release and subsequent binding of CRF at the
pituitary. Adrenalectomy has no effect on the number of cerebral CRF-binding








sites or CRF-stimulated adenylate cyclase activity in the brain (Wynn et al.,
1984). However, in Alzheimer's disease there is a decrease in irCRF in frontal,
temporal and occipital cortices that is accompanied by an increase in the
number of CRF-binding sites (De Souza et al., 1986). In contrast, in depressed
patients there was observed an increase in the concentration of cerebrospinal
fluid CRF (Nemeroff et al., 1984) whereas, a decrease in the number of CRF
binding-sites in the prefrontal cortex of suicide victims was observed (Nemeroff
et al., 1988). In addition, chronic atropine treatments in rats resulted in an
increase in CRF-binding sites in the cerebral cortex (De Souza & Battaglia,
1986). However, no change in the Vmax or the EC50o of CRF-stimulated
adenylate cyclase was observed following this treatment. These results
indicate that the CRF system is a dynamic one, capable of making
compensatory changes to perturbations in the functioning of this system.
A number of behavioral, physiological and neurochemical responses has
been observed following ICV injection of CRF. Many of these responses
resemble those observed in stress, prompting the suggestion that CRF acts to
coordinate a whole body response in stress (Koob & Bloom, 1985). The
association of irCRF-containing cells and processes with brain structures
involved in regulating emotional, behavioral, and autonomic responding is
consistent with this hypothesis. Physiological stress-like effects of ICV CRF
include the activation of sympathetic and adrenomedullary output, resulting in
increases in the concentrations of circulating NE, Epi, glucagon, and glucose
(Brown et al., 1982; Brown and Fisher, 1985). Mean arterial pressure and heart
rate are also increased (Fisher et al., 1982; Fisher et al., 1983; Brown & Fisher,
1985) following ICV CRF. An involvement of the sympathetic nervous system in
mediating these effects is suggested by the ability of the ganglionic blocker,
chlorisondamine, to prevent these CRF-induced responses (Fisher et al., 1982)








and the ability of ICV CRF to increase the electrophysiological activity of the
adrenal sympathetic nerve (Kurosawa et al., 1986).
In addition to an activation of peripheral catecholamine systems, ICV

CRF also stimulates the release of central dopamine and NE systems in a
pattern similar to that observed in stress (Dunn & Berridge, 1987). Activation of
brain electrical activity by CRF has also been observed. Thus, ICV CRF
produced a dose-dependent activation of the electroencephalogram with signs
of both EEG and behavioral arousal predominating at low doses (10-100 ng)
and epileptiform activity and seizures observed at higher doses (1-20 4g)
(Ehlers et al., 1983). When administered either ICV or directly by pressure
microinjection, CRF activates neuronal firing in the LC in anesthetized and
awake rats (Valentino et al., 1983; 1988). Although CRF has a direct excitatory
effect on cortical and hypothalamic neurons, inhibition of cell firing has been
recorded in the thalamus and lateral septum (Eberly et al., 1983). This regional
specificity of CRF on neuronal activity suggests that these effects do not result
from a nonspecific action of the peptide.
Central administration of CRF has been documented to have a number of
behavioral effects in rodents. ICV CRF produced a dose-dependent activation
of locomotor behavior in a familiar environment at 1 and 10 ig (Sutton et al.,
1982). When tested in a novel open field, ICV CRF (1 jg) decreased locomotion
and hearings and increased freezing, results that were interpreted to reflect an
increase in the aversive nature of the novel environment (Sutton et al., 1982).
At the lower dose of 10 ng, CRF increased locomotor activity in this test.
Similarly, prior exposure of rats to a stressor increased locomotor activity in the
open field (Roth & Katz, 1979). Approaches to food in a novel open-field were
decreased by ICV CRF in rats. This effect was opposite to that seen following
benzodiazepine administration, suggesting that CRF acts to enhance the








anxiogenic nature of the novel environment (Britton et al., 1982). Similarly,
when tested in a conflict test, ICV-administered CRF decreased punished
responding, whereas benzodiazepines increased this response (Britton et al.,
1985). In addition, CRF has also been reported to increase grooming in mice
and rats (Morley and Levine, 1982; Veldhuis & de Wied, 1984; Dunn et al.,
1987), and decrease feeding in rats (Morley and Levine, 1982; Krahn et al.,
1986), responses that have been observed following exposure to certain
stressors (Colbern et al., 1978; Dunn et al., 1979; Krahn et al., 1986).

Norepinephrine
Activation of brain noradrenergic systems is one component of the
general response observed in stress (Dunn & Kramarcy, 1984; Glavin, 1985;
Stone, 1975). Experimental evidence supporting an involvement of
noradrenergic systems in regulating behavioral and emotional responding
associated with stress and/or anxiety in humans (Hoehn-Saric et al., 1981;
Charney et al., 1983), non-human primates (Redmond & Huang, 1979; Harris &
Newman, 1987) and rodents (Davis et al., 1977; Handley & Mithani, 1984)
suggests that this increase of NE release during stress is of functional
significance.
In contrast to other neurotransmitter systems, NE is distributed diffusely
throughout the CNS supplied primarily by two major systems, the ventral (VNB)
and dorsal (DNB) noradrenergic bundles, which originates from cell groups
within the pontine and medullary tegmental regions (A1-A7 as described by
Dahlstrom & Fuxe, (1964)). The VNB originates from cells within the dorsal
medullary (A2) and lateral tegmental (A1, A5, A7) noradrenergic cell groups.
The DNB originates from cells of the nucleus locus coeruleus (LC; A4 and A6)
located within the rostral pontine region in the floor of the fourth ventricle.








Through a system of extensive collaterals, the LC innervates most regions of the
brain and provides the sole source of NE to the hippocampus and cerebral and
cerebellar cortices (Lindvall & Bjorklund, 1974; Pickel et al., 1974; Aston-Jones,
1985). Although the NE innervation pattern arising from the LC displays
regional and laminar specificity (Morrison et al., 1982), the widespread
innervation of structures throughout the CNS arising from a small group of
neurons displaying homogeneous firing characteristics (Aston-Jones & Bloom,
1981) suggests a more global action of NE within the brain than is supposed for
other transmitter systems.
A number of functions of brain NE has been proposed including a role in
regulating neuroplasticity (Kasamatsu, 1983), memory (Mason, 1981; Sara,
1985), arousal (Hobson et al., 1975), and selective attention (Mason & Iversen,
1978). An action of NE in selective attention was suggested by
electrophysiological studies in which application of NE or LC stimulation
suppressed spontaneous action potential firing to a greater degree than
stimulus-elicited firing. This results in an increased strength of the elicited
response compared to background firing, increasing the signal to noise ratio.
This effect has been observed in many brain areas including auditory cortex in
response to auditory stimuli (Foote et al., 1975), hippocampus in response to a
conditioned stimulus (Segal & Bloom, 1976), somatosensory cortex in response
to tactile stimulation (Waterhouse et al., 1981), and visual cortex in response to
visual stimuli (Madar & Segal, 1980). Within the visual cortex, the enhancement
of neuronal responding was observed in simple and complex cells and was
more pronounced when the visual stimulus was in the preferred direction.
The increase in the firing rate of LC neurons in response to sensory stimuli that
elicit an orienting response in awake monkey and rats is also consistent with








the hypothesis that NE regulates responding to sensory stimuli (Foote et al.,
1980).
Behavioral studies, however, have provided little support for a role of NE
in selective attention (Britton et al., 1984; Sara, 1985; Robbins et al., 1985),
although, as reviewed by Sara (1985), this might reflect a lack of sensitivity of
the behavioral tests used to examine this question. Given the largely negative
results from behavioral studies testing the selective attention hypothesis, an
alternative view of noradrenergic function was suggested as a regulator of the
vigilance state of an animal, with vigilance being defined as the degree of
attention to or surveillance of the external environment (Aston-Jones, 1985).
Thus, although LC firing is inhibited during automatic, tonic behaviors whether
the animal is active or inactive (e.g. grooming, feeding, sleep), firing becomes
quite vigorous when any of these behaviors is suddenly disrupted and the
animal orients toward the external environment.
In summary, there is evidence supporting the idea that the noradrenergic
system can facilitate processing of sensory information. This evidence includes
that widespread distribution of NE in the brain, the selective firing of the LC to
sensory stimuli of all modalities and the enhancement by the stimulation or the
application of NE to target cells of stimulus-elicited responding of those target
cells. It has been proposed that activation of this system in the presence of a
stressor could facilitate adaptive responding by increasing the vigilance state of
an animal, thereby increasing attention to environmental cues.








Rationale
The time an animal spends investigating objects in a novel,
multicompartment chamber (MCC) has been demonstrated to be a useful index
of exploratory behavior. An advantage of this paradigm is that indices of
locomotor activity (the number of compartment entries and rears) provide an
internal control for changes in exploratory behavior that might result from a
general inhibition or excitation of locomotor activity. In this paradigm the opiate-
agonist, morphine, decreased this exploratory response, whereas the opiate-
antagonist, naloxone, increased this response and blocked the action of
morphine (Arnsten et al., 1981).
The effect of naloxone on exploratory behavior was stereospecific with
the (+)-enantiomer having no significant effects. This form of naloxone has
been demonstrated to be 10,000 times less active in assays of opiate receptor
antagonism than (-)-naloxone (lijima et al., 1978). The effects of both morphine
and naloxone are observed in the absence of changes in locomotor activity.
Various stressors (tail-pinch, restraint, and high-intensity white noise) also
decreased this exploratory response in the absence of significant effects on
locomotor activity (Arnsten et al., 1985). The effect of these stressors was
antagonized by naloxone at a dose that had no behavioral effects in otherwise
untreated animals.
These results are consistent with the hypothesis that opioids released

during stress affect behavior. The involvement of noradrenergic systems in
regulating behavioral responding to a novel environment has also been
examined in this paradigm (Arnsten et al., 1981). Rats in which cortical and
hippocampal NE was depleted by lesions of the DNB using the neurotoxin 6-
hydroxydopamine (6-OHDA) spent less time exploring the objects in the testing

chamber when tested at least 14 days following the lesion.








As described above, CRF and NE have been implicated in regulating

behavioral responding associated with stress and anxiety. Because behavior in
the MCC was sensitive to exposure to stressors, or modulation of opioid and
noradrenergic activity in the absence of changes in locomotor activity,

suggested that this paradigm could be used for examining the involvement of
CRF and NE in mediating behavioral responding in stress.
The following series of experiments examined the possibility that an
increased release of either cerebral CRF or NE has a role in the stressor-
induced decrease in exploratory behavior observed in this paradigm. The
possibility that these two neurochemical systems interact to affect exploratory
behavior was also examined.













CHAPTER II
GENERAL METHODS


Animals
Male CD-1 mice (25-35 g) were obtained from Charles River
(Wilmington, MA) and were group-housed with free access to food and water for
a minimum of one week prior to surgery and testing and maintained on a 12-12
light-dark cycle (lights on 07:00). Male Sprague-Dawley rats (180-200 g) were
obtained from the Animal Care Facility of the University of Florida. They were
housed in pairs for at least 5 days after arrival in the Dept. of Animal Resources
at the University of Florida and were handled daily. They were allowed free
access to food and water and maintained on a light-dark cycle similar to that
used for the mice. The rats were allowed a minimum of 5 days following
surgery for recovery.

Materials
Human/rat CRF was donated by Jean Rivier (Salk Institute) or obtained
from Peninsula Laboratories and was dissolved in 0.15 M NaCI containing
0.001 M HCI. Aliquots of the required concentration were stored frozen and
used within one hour of thawing. Clonidine and phenylephrine were obtained
from Sigma Chemical Co. (St. Louis, MO), and idazoxan was a gift from Reckitt
and Coleman (Hull, England). All were dissolved in normal saline (0.15 M).
Prazosin was a gift from Pfizer (Groton, CT) and was dissolved in water at a
concentration twice the desired concentration and diluted with an equal volume








of 2-fold concentrated normal saline. DSP-4 (Astra) was dissolved in normal
saline immediately prior to use.

Surgeryj
Mice requiring ICV injections were implanted with polyethylene cannulae
(Clay Adams PE-50 tubing) placed for injection into the lateral cerebral
ventricles 3-4 days prior to behavioral testing, following the method described
by Guild and Dunn (1982). Briefly, the tubing was heated and twisted to form a
thickened button. This tubing was then trimmed to a total length of 9 mm, with
one end being 3.0 mm long measured from the button. Mice were anesthetized
using pentobarbital, and the cannulae were implanted bilaterally (3 mm end
down) at 0.6 mm P, 1.6 mm L (estimated) from bregma through drill holes. A
third hole was drilled posterior to the cannulae and a jewelers screw was
screwed into the skull without penetrating the dura, to act as an anchor for the
cement. Cranioplastic cement (Plastic Products Company; Roanoke, VA) was
then applied to form a skull cap around the two cannulae and the screw.
Cannula placement was verified by visual inspection following injection of dye
through the cannulae.
Rats were implanted stereotaxically with stainless steel guide cannulae
(23 ga 10-14 mm) under pentobarbital anesthesia. For injections into the lateral
cerebral ventricles, cannulae were implanted at the following coordinates in
relation to bregma: A-P -0.8, L + 2.5, V 4 mm below dura. Two jewelers screws
were inserted bilaterally into the skull, caudal to the cannulae. A skull cap was
made by applying Cranioplastic cement around the canulae and screws.








Behavioral Testing
The testing chamber was based on that designed by Arnsten and Segal
(1979). For mice this chamber (38 x 38 x 23 cm) consisted of nine
interconnecting compartments within each of which a wire mesh stimulus (3.0
cm sphere) was recessed in a 2.5 cm hole below the floor. The chamber used
for rats was of the same design, enlarged appropriately: 75 x 74 x 46 x cm with
4 cm wire spheres recessed in 3 cm holes. The animal was placed in this
chamber and observed for a 25-30 min period. The MCC was brightly lit, and

background sounds were masked using a white noise generator (70 db).
Testing was conducted between the hours of 09:00 and 15:30. During the
observation period the duration and frequency of a number of behaviors were
recorded using an NEC PC8201A computer modified for use as an event
recorder (Stoelting, Chicago, IL; Model 47250X). Behaviors recorded include:
Measures of contact with the stimuli: total number of contacts; total
duration of contacts; mean time per contact (total duration of contacts/total
number of contacts).
Measures of locomotor activity: total number of compartment entries; total
number of rears (both forefeet off the floor).

Treatments
ICV injection was performed in mice using a 10 gl syringe with a 10 mm
long needle. Unless otherwise stated, ICV injections consisted of a total volume

of 4 g1 divided equally between the two cannulae and given over a 30-45 sec
period. Unless otherwise indicated, peripheral injections were given
intraperitoneally. Restraint was administered by allowing the mouse to enter a

darkened polyethylene tube 10 cm in length and 3 cm in diameter (50-ml

centrifuge tubes in which airholes had been punched), preventing their exit by








taping the ends of the tube with laboratory labeling tape. Restraint was
administered in a room separate from both the housing quarters and testing
room. ICV injections of rats were performed using 30-ga stainless steel tubing
connected via plastic cannula tubing to a 10 gl syringe.

Neurochemistry
After decapitation (not anesthetized), trunk blood was collected into
heparinized 1.5 ml Eppendorf tubes, and the brain was rapidly removed from
the skull. The brain was placed on a Whatman No.1 filter paper on an inverted
Petri dish in a bucket of ice and flushed with ice-cold isotonic saline to chill it
and rinse it of blood. Excess fluid was absorbed, and the brain was dissected
into medial prefrontal cortex, cingulate cortex, hippocampus, hypothalamus,
brainstem, and cerebellum. The dissected regions were immediately placed in
a tared 1.5 ml plastic centrifuge tube, weighed and placed on dry ice. All tissue
samples were homogenized within 24 hrs by brief ultrasonication (except
brainstem),in 0.4-0.5 ml of 0.1 M HCIO4 containing 0.1 mM EDTA and an
internal standard of N-methyldopamine (NMDA). The tissue was not allowed to
thaw until it was mixed with the ice-cold HCI04. Brainstem was homogenized
using a Tekmar Tissumizer in 1.0 ml of HCIO4-EDTA containing internal
standard. The samples were thawed and centrifuged for 1 min in a Beckman
Microfuge, and left at room temperature for 2-3 hours prior to HPLC analysis..

HPLC
The HPLC system consisted of a Waters WISP 710 OB autosampler, an
LDC Milton-Roy 396 pump with a Bioanalytical Systems (BAS) pulse
dampener, two BAS LC-4B Electrochemical detectors using a BAS TL-5A cell
with a glassy carbon electrode held at potentials of approximately 0.78 and








0.85V with respect to a Ag-AgCI reference electrode. Data were collected using
an LDC Cl-10 and a Hewlett-Packard 3390A integrator. Data were corrected
using an internal standard correction procedure with respect to the NMDA and a
set of standards run immediately prior to each brain region of each experiment,
and were expressed in terms of ng per wet weight of tissue (mg). The
calculations were performed using a commercial spreadsheet program
(Supercalc 3a) on an Apple lie computer.
The chromatography column was a 25 cm reverse phase C-18
(Techsphere 50DS) column from Sep/a/ra/tions Group (Hesperia, CA) held at
37-43 C by a BAS LC-22A column heater. The mobile phase was a 0.05 M

sodium phosphate-0.05 M citrate buffer (about pH 3.2) containing 0.1mM
Na2EDTA, 0.25 mM sodium octyl sulfonate and 8% (v/v) methanol degassed
continuously with helium. The precise pH, column temperature and
concentrations of octylsulfonate and methanol were adjusted to achieve optimal
separation. Compounds assayed were epinephrine, NE and its catabolites,
normetanephrine, 3-methoxy,4-hydroxyphenylethylene glycol (MHPG),
dopamine (DA) and its catabolites: homovanillic acid (HVA), 3,4-
dihydroxyphenylacetic acid (DOPAC) and 3-methoxytyramine, tryptophan; (Trp),
serotonin (5-HT), and its catabolite: 5-hydroxyindoleacetic acid (5-HIAA).

Corticosterone Assay
Plasma corticosterone was determined by radioimmunoassay, using the
procedure described by Gwosdow-Cohen et al. (1982). The antibody was
obtained from Dr. Niswender of Colorado State University. Briefly, the animals
were decapitated and trunk blood was collected into 1.5 ml heparin-coated
plastic centrifuge tubes. These tubes were then centrifuged and the plasma
was removed using a glass pasteur pipette. The steroids were extracted from








the plasma using methylene chloride. Following incubation of the steroids with
antibody and [3H]-corticosterone (overnight at 40 C), the bound and free steroids
were separated by charcoal absorption. Ten min following the addition of the
charcoal, the samples were centrifuged at 1800-2000 rpm for 5 min and 300 pg1

of the supernatant was added to a scintillation counting vial containing 3 ml of
scintillation fluid (Scintiverse II; Fisher Scientific, Pittsburgh, PA). The samples
were then assayed using scintillation spectroscopy.

Statistical Analysis
For comparisons involving multiple groups, statistical analysis was
performed by ANOVA using an Apple lie computer and a statistical software
package (ANOVA II, Human Systems Dynamics). When statistical significance
was indicated by ANOVA (i.e. P < 0.05), individual mean comparisons were
made using Duncan's post-hoc test. When only two groups were being
compared, Student's t-test was used. For comparison of multiple groups to a
single control group, the Dunnett's t-test was used.













CHAPTER III
THE ROLE OF ENDOGENOUS CRF IN REGULATING RESTRAINT-STRESS-
INDUCED CHANGES IN EXPLORATORY BEHAVIOR.


Corticotropin-releasing factor (CRF) released from the hypothalamus
stimulates the secretion of adrenocorticotropic hormone (ACTH) and
P-endorphin from the pituitary (Vale et al., 1981). This action of CRF on the

pituitary plays a major role in regulating peripheral responses in stress. CRF-
containing neurons and high-affinity binding sites for CRF have also been
identified in extrahypothalamic regions of the brain (Swanson et al., 1983; De
Souza et al., 1984; Wynn et al., 1984). Moreover, intracerebral administration of
CRF elicits a variety of physiological (Valentino et al., 1983; Brown et al., 1982),
neurochemical (Dunn & Berridge, 1987), and behavioral (Britton et al., 1982;
Sutton et al., 1982; Morley & Levine, 1982; Kalin et al., 1983; Veldhuis & de
Wied, 1984) responses.
Although a number of CRF-elicited responses has been documented,
little is known concerning the site(s) of action in the brain or interactions with
other transmitter systems in eliciting these effects. When examining the
behavioral effects of treatments that activate the HPA axis, it is important to
determine whether an observed behavioral effect is the result of increased
circulating ACTH and glucocorticoids because both of these products of this
axis can affect behavioral responding. For example, glucocorticoids and ACTH
have been shown to influence locomotor activity (Bohus & Lissak, 1982) and

avoidance (Bohus et al., 1968) and escape (Stone et al., 1988) behaviors. The
effects of CRF on grooming and food intake (Morley & Levine, 1982; Britton et








al., 1986a), and on locomotor activity (Eaves et al., 1985; Britton et al., 1986a;
Britton et al., 1986b) were not attenuated by hypophysectomy. Similarly,
dexamethasone injected at a dose sufficient to block CRF-induced release of
ACTH did not diminish the effect of CRF on punished responding (Britton et al.,
1986a), grooming (Britton et al., 1986a; Dunn et al., 1987), or locomotor activity
(Britton et al., 1986a,b). Thus, these actions of CRF apparently occur
independently of an activation of the pituitary-adrenal axis.
Endogenous opioid peptides are thought to be involved in responding in
stress. Activation of opioid systems may occur with exposure to certain
stressors, and a naloxone-sensitive stress-induced analgesia has frequently
been described (Bodnar et al., 1980). Also, a variety of stressors has been
shown to result in increased concentrations of circulating P-endorphin (Akil &

Watson, 1983). Finally, alterations in opiate binding in the brain have been
reported following exposure to stressful stimuli (Nabeshimi et al., 1985). Thus, it
is possible that if CRF regulates certain stress-related responses, endogenous
opioids might also be involved in such an action. Consistent with this
hypothesis, an involvement of opioid systems in CRF-induced responding has
been suggested by a number of studies in which the opiate antagonist,
naloxone, prevented the effects of CRF. For example, naloxone blocked the
ICV CRF-induced decrease in sexual behavior (Sirinathinghji, 1985, 1987), the
CRF-induced increase in grooming behavior (Dunn et al., 1987), and the CRF-
induced increase in the mean arterial pressure and heart rate (Kiang & Wei,
1985; Saunders & Thornhill, 1986). However, naloxone failed to inhibit the
increase in locomotor activity induced by 1 gg of ICV CRF when tested in a

familiar environment (Koob et al., 1984). Therefore, some of the effects of CRF
may involve an activation of endogenous opioid systems, whereas others may
not.








The following studies were designed to test further the hypothesis that

endogenous CRF participates in regulating behavioral responding in stress
using the exploratory response to the MCC. Additional studies examined
whether an activation of endogenous opioid systems might be involved in any
such action and where within the ventricular system ICV-administered CRF
might act to elicit its behavioral effects.

Materials and Methods

Materials
CRF was obtained and prepared as described in Chapter 2. Morphine-
HCI was obtained from Eli Lilly & Co. and was diluted in 0.15 M NaCl to the
appropriate concentration. Naloxone was a gift from Endo Laboratories. The
CRF antagonist, alpha-helical CRF9-41 (ahCRF), was obtained from Peninsula
Laboratories and was dissolved in 0.15 M NaCI containing 0.001 M HCI.
Aliquots of the required concentration were stored frozen and used within an
hour of thawing.

Animals
The mice and rats used were obtained and housed as described in

Chapter 2. Hypophysectomized male mice (CD-1) were obtained from Charles
Rivers (Wilmington, MA) weighing 24-28 g and were group-housed with free
access to food and a 9% sucrose solution for at least 5 days prior to testing or
surgery. Following surgery the hypophysectomized mice were group-housed to
minimize the hypothermia that occurs in these animals. Verification of the

hypophysectomy was determined by visual inspection, and by measurement of

plasma corticosterone concentrations using an radioimunnoassay (RIA)
following exposure of the animals to 30 min of restraint.










Effects of morphine and naloxone
First, we examined whether the effects of morphine, naloxone and stress
on investigatory behavior observed in rats by Arnsten et al. (1979, 1981, 1985)
also occurred in mice. Pilot studies were conducted to determine the
approximate minimum doses of morphine and naloxone necessary to elicit
behavioral effects. These studies indicated that doses of greater than 1 mg/kg
of both naloxone and morphine were necessary to elicit obvious behavioral
effects.
For the studies examining the effect of morphine or naloxone, mice were
transported from the housing quarters to a room separate from the testing room
where they were injected SC with either saline (n=7), naloxone (0.75 mg/kg;
n=7), naloxone (1.20 mg/kg; n=8), morphine (1.75 mg/kg; n=7), or a mixed
solution containing morphine (1.75 mg/kg) and naloxone (0.75 mg/kg) (n=8).
Immediately following the injection, the animals were transferred to the testing
room and 10 min later placed in the chamber for observation.
Morphine and naloxone significantly affected the mean time per contact
with the stimuli (F4,32 = 22.1, P < 0.001; Fig. 3-1). Morphine decreased the
mean time per contact compared to saline, whereas naloxone (0.75 mg/kg)
reversed the effect of morphine, and at the higher dose (1.25 mg/kg),
significantly increased the mean time per contact. The total number of contacts
was slightly, but significantly decreased by both doses of naloxone (F4,32 = 3.0,
P < 0.05). Morphine did not significantly affect the total number of contacts. The
number of compartment entries was significantly decreased in the naloxone
(0.75 and 1.25 mg/kg) and naloxone + morphine-treated groups (F4,32 = 8.2,





























Sal 0.75 1.2 1.75 Morph
Nal Nal Morph +
Nal


Sal 0.75 1.2 1.75 Morph
Nal Nal Morph +


Fig 3-1. Effect of morphine and naloxone on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were
injected SC with either saline, 1.75 mg/kg morphine, 0.75 mg/kg naloxone, 1.25
mg/kg naloxone, or a combination of 1.75 mg/kg morphine and 0.75 mg/kg
naloxone 10 min before placement in the chamber.
*Significantly different from saline-controls (P < 0.05, **P < 0.01).
++Significantly different from morphine-treated animals (P < 0.01).








P < 0.001; Fig. 3-1). There were no statistically significant effects of morphine or
naloxone on the number of rears (Fig. 3-1).

Effects of restraint and naloxone
In the same experiment, two additional groups were subjected to 30 min
of restraint after which they were injected with either saline or naloxone (0.75
mg/kg). Mice were restrained as described in Chapter 2. Restraint was
administered for 30 min in a room separate from both the housing quarters and

the testing room. Animals were removed from the restraining apparatus,
immediately injected with either saline (n=8) or naloxone (0.7 mg/kg; n=8) and
transported to the testing room. Ten min following the injection the animals
were placed in the testing chamber for observation.
As shown in Fig. 3-2, restraint significantly decreased the mean time per
contact with the stimuli (F1,26 = 7.5, P < 0.01) in the absence of significant effects
on locomotor activity. Naloxone (0.75 mg/kg) completely reversed the effect of
restraint on the mean time per contact. In addition, restraint appeared to
increase the sensitivity of the animals this dose of naloxone, such that the
restraint + naloxone group exhibited a significantly increased mean time per
contact relative to saline and naloxone controls. A similar tendency of restraint
to produce a sensitization to this behavioral effect of naloxone was also
observed in rats tested in the MCC (Arnsten et al., 1985).
The major effect of restraint on stimulus interaction appeared to be an
inhibition of habituation to the novel environment. During the first 5 min of the

observation period the mean time per contact in the restrained group did not
differ significantly from the unrestrained controls (Fig. 3-3). In contrast, in the 5-
10 min interval restrained animals exhibited a significantly shorter mean time

per contact compared to controls. The effects of restraint stress appeared to










MEAN TIME PER CONTACT (sec)


Saline Restraint Nal Restraint
+


600
500
400
300
200
100

500
400
300
200
100


Saline Restraint Nal Restraint


Fig. 3-2. Effect of restraint and naloxone on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were
restrained for 30 min. Immediately following the restraint the animals were
given an SC injection of either saline or naloxone (0.75 mg/kg) and placed in
the chamber 10 min later.
*Significantly different from saline-unrestrained mice (P < 0.05).
++Significantly different from saline-restrained mice. (P < 0.01).


3.0


2.0


1.0










- QUIET
IRESTRAINI


*iT


2.0h


1.O%


5 10 15 20 25 30


TIME tmin)

Fig. 3-3. Effect of restraint stress on the mean time per contact in 5-min
intervals. Shaded boxes represent animals receiving 30 min restraint stress.
Open boxes represent unrestrained-controls.
*P < 0.05 (Student's t-test).


4.0


3.01-








dissipate later in the observation period such that by the 20-25 min interval the
animals behavior was similar to that of unstressed controls. A similar effect was
observed with morphine administration, except that the initial mean time per
contact was maintained throughout the testing period (data not shown).

Effects of CRF in mice.
To determine whether ICV CRF would elicit stress-like effects on
behavior in the MCC, mice were injected with either saline (n=9) or 1 (n=7), 5
(n=7), 10 (n=7), 20 (n=8), 50 (n=7) 100 (n=7), or 150 (n=10) ng CRF injected 15
min prior to testing. The mice were cannulated 3-5 days prior to testing and
housed singly following the surgery. The cannulation and injection procedures
were as described in Chapter 2.
A significant effect of CRF on stimulus-contact time was observed (F7,54 =
8.2, P < 0.002) in the absence of significant effects on compartment entries (Fig.
3-4). Rears were significantly decreased at the 150 ng dose (F7,54 = 2.2, P <
0.045). In preliminary studies, higher doses (> 200 ng) produced a significant
inhibition of both compartment entries and rears, suggesting a nonspecific
depression of locomotor activity (data not shown).

Effects of CRF and naloxone.
To compare further the effect of restraint and CRF on exploratory
behavior in the MCC, the effect of naloxone on the CRF-induced decrease in
this behavior was examined using four groups of mice (n=8). Mice were
removed from the housing quarters and given an ICV injection of either saline
or 75 ng CRF. Immediately following this, they were injected with either saline
or 0.7 mg/kg naloxone and transported to the testing room. Ten min following
the injections, the animals were placed in the testing chamber.



























Saline 1 5 10 20 50 100 150


Saline 1 5 10 20 50 100 150


ICV CRF (ng) ICV CRF (ng)



Fig. 3-4. Effect of varying doses of CRF injected ICV on the mean time per
contact, the total number of rears and the total number of compartment entries.
Mice were injected ICV with either saline or 1, 5, 10, 20, 50,100, or 150 ng CRF
15 min prior to testing.
Significantly different from saline-controls (P < 0.05, **P < 0.01).








A significant main effect of CRF (F1,29 = 6.9, P < 0.02; Fig. 3-5) and a
significant CRF-naloxone interaction (F1,29 = 11.8, P = 0.002) was observed for
the mean time per contact. CRF (75 ng) injected into the lateral ventricles
decreased the mean time per contact with the stimuli. This effect of CRF on
stimulus-contact time was antagonized by naloxone (0.7 mg/kg) to levels
indistinguishable from controls. Naloxone at this dose had no effect on the
mean time per contact in animals receiving ICV saline. Neither CRF nor
naloxone significantly affected the number of rears or compartment entries (Fig.

3-5).
As observed with the restraint- and morphine-treated animals, the major
effect of CRF was not on the initial mean time per contact with the stimuli; rather,
the shorter mean time per contact observed early in the testing session was
maintained throughout the entire 30 min testing session (Fig. 3-6). Naloxone
reversed this apparent inhibition of habituation so that the animals showed a
behavioral time course similar to that of controls (data not shown).

Effects of CRF in Rats.
The effect of CRF on the behavioral response to the MCC was examined
in rats to test whether the behavioral effect of CRF could be observed in a
species other than the mouse. Rats were implanted with bilateral cannulae for
injection into the cerebral lateral ventricles as described in Chapter 2. They
were housed singly and were allowed a minimum of 5 days for recovery. One

day before testing they were moved from the colony room to a room separate
from the testing room. On the day of testing, animals were injected ICV with

either saline (n=9) or 25 ng CRF (n=7) in the colony room and replaced into
their home cage. Fifteen min following the injection, they were transferred to the

testing room and immediately placed in the MCC. As observed with mice, CRF
















3.0


2.0 -
"- B 400
1 3 300
1.0 H 200
100

Saline CRF Nal CRF Saline CRF Nal CRF
+ +
Nal Nal



Fig. 3-5. Effect of CRF and naloxone on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were given
an ICV injection of either saline or 75 ng CRF, immediately followed by a SC
injection of either saline or 0.75 mg/kg naloxone.
**Significantly different from saline-controls (P < 0.01). ++Significantly
different from CRF-saline (P < 0.01).











-SALINE
*CRF


1.01-O


5 10 15 20 25 30


TIME tmin)


Fig. 3-6. Effect of 75 ng CRF on the mean time per contact in 5-min intervals.
Mice were injected ICV with saline (open bars) or CRF (solid bars) 10 min prior
to testing.
*Significantly different from saline-controls (P < 0.05, **P < 0.01;
student's t-test).


4.0-


3.01-


2.0F







Table 3-1. Effect of ICV CRF on exploratory behavior of rats


Mean time per Rears Compartment
contact entries


Saline 2.99 0.36 128 13 197 19


CRF (25ng) 1.94 0.07* 146 14 184 21


Mean time per contact and the mean total number of rears and compartment
entries of rats injected ICV with either saline or CRF. *Significantly different
from saline-controls (P < 0.025, Student's I-test) .

decreased the mean time per contact compared to saline-injected animals (t =
2.57, P < 0.025; Table 3-1). CRF did not significantly affect the number of

compartment entries, or rears (Table 3-1).

The Effects of Restraint and CRF in Hypophysectomized Mice.
To test whether the effects of restraint and ICV CRF on behavior in the
MCC occurred independently of an activation of the HPA axis, the effects of

these two treatments were examined in hypophysectomized mice. For the
experiments testing the effects of restraint, restrained mice (n=9) were removed
from their home cage, transferred to a room separate from both the colony and
testing rooms and immediately restrained. Following the 40 min period of
restraint, the animal was removed from the restraining device, transferred to the
testing room and immediately placed in the testing chamber and observed.
Unrestrained mice (n=9) were removed from their home cage and transferred to
the testing room where they were immediately placed in the testing chamber
and observed. Restraint significantly decreased the mean time per contact with

the stimuli (t = -6.16, P < 0.001; Table 3-2) in the hypophysectomized mice.
There were no significant effects on the number of compartment entries or rears.










Table 3-2. Effect of restraint and CRF on exploratory behavior in
hypophysectomized mice.


Mean time per Rears Compartment
Contact Entries


U n restrain ne

Restrained


Saline


CRF (50 ng)


3.01 0.26

1.4 0.09**


2.56 0.15

1.19 0.08**


The mean time per contact and the mean total number of rears and
compartment entries ( SEM) of hypophysectomized that were either restrained
or injected with ICV CRF (50 ng). The data are from two separate experiments.
Statistical significance was determined using Student's t-test.
**Significantly different compared to the respective control group (P < 0.01)


196 28

168 16


148 24

133 16


376 35

359 29


243 25

290 48








To test the effect of CRF in hypophysectomized mice, the mice were
implanted with bilateral polyethylene cannulae for injection into the lateral

ventricles. Testing was conducted 3 days following surgery. A number of the
hypophysectomized mice did not tolerate the surgery well. These animals were
hypoactive and had lost a significant amount of weight between surgery and
testing. Animals that appeared ill and that had lost more than 4 g were
excluded from testing. Mice were injected with either 50 ng of saline or CRF 15
min prior to testing. The injections were divided equally between the left and
right cannulae consisting of a total volume of 2 ;J given over 45 sec. CRF
significantly decreased the mean time per contact with the stimuli in the
hypophysectomized mice (t = -8.55, P < 0.001; Table 3-2). It did not significantly

affect the number of compartment entries or the number of rears.

Effect of a CRF-antagonist on the restraint-induced change in stimulus-directed
behavior.
The above results indicate similar effects of restraint and CRF on

exploratory behavior in both rats and mice. These results are consistent with
the hypothesis that endogenous CRF participates in regulating behavioral
responding in stress. To test this hypothesis more directly, the effect of a CRF-
receptor antagonist (ahCRF) on the restraint-induced decrease in exploratory

behavior was examined.
Mice were cannulated for injection into the lateral cerebral ventricles as

described in Chapter 2. The animals were housed singly following surgery.
Restrained mice were transported individually in their home cage to a room
separate from both the housing quarters and the testing room, and injected ICV
with either saline or 5,10, 20, or 50 gg ahCRF divided equally between the two

cannulae. The mice were placed in the restraining devices 5 min later for a 40








min period. Immediately following restraint, the animals were carried into the

testing room and 5 min later placed in the experimental chamber. Unrestrained
controls were injected in the colony room 45 min before transport in their home
cage to the testing room and placed in the testing chamber 5 min following
transport.
The results for the mean time per contact, rears and compartment entries

are shown in Fig. 3-7. As observed in the previous experiments, restraint
significantly decreased the mean time per contact with the stimuli (F1,72 = 20.5,
P < 0.001). ICV administration of ahCRF 5 min prior to restraint reversed the
restraint-induced decrease in the mean time per contact in a dose-dependent

manner (F4,72 = 8.6, P < 0.001). The 10, 20, and 50 jg doses significantly

antagonized the restraint-induced decrease in the mean time per contact. The
effect of the 10 jg dose was significantly less than that of the 20 ig dose (P <
0.05), but was not significantly different from the unrestrained-saline control
group. Although the restrained-5 jg group did not differ significantly from the
restrained-saline group, neither did it differ significantly from the unrestrained-

saline group, suggesting a slight effect of the antagonist at this dose. The 50 jg
dose appeared to be less effective than the 20 jg dose although this difference
was not statistically significant. There were no statistically significant effects of

the antagonist on the mean stimulus-contact time in unrestrained mice.
Restraint did not significantly affect the number of contacts made with the
stimuli. There was a small but significant effect of the antagonist on the number
of contacts made with the stimuli (F4,72 = 2.5, P = 0.05) However, this was only
statistically significant in the 5 jg-restrained and 20 jg-unrestrained groups

(data not shown).
Neither restraint stress nor the antagonist significantly affected locomotor
activity as measured by compartment entries. Even though there was a


























Sal 5 10 20 50


Sal 5 10 20 50


DOSE OF ANTAGONIST (pg)



Fig. 3-7. Effect of ahCRF and restraint on the mean time per contact, the total
number of rears and the total number of compartment entries. Unrestrained
mice (open bars) were injected ICV 45 min prior to testing with either saline or
varying doses of ahCRF and returned to their home cages. Restrained mice
(solid bars) were injected with either saline or varying doses of ahCRF 5 min
prior to being restrained for 40 min.
*Significantly different from saline-unrestrained mice (P < 0.05,
**P < 0.01). +Significantly different from saline-restrained mice. (P < 0.05, ++P
<0.01).








tendency for an increase in this measure in some of the ahCRF-injected groups,
there was not a statistically significant correlation between compartment entries
and stimulus-contact time (r(43) = -0.09, P = 0.57 for restrained; r(38) = 0.17, P =
0.31 for unrestrained mice). Restraint significantly decreased rears in saline-
injected animals (F1,72 = 6.53,.P < 0.01), an effect not observed previously. This
effect was apparent in those animals that were treated with either 5 or 50 jg,
ahCRF but not those treated with 10 or 20 jg of the antagonist. ICV ahCRF had
no significant effect on the total number of rears.

Ventricular blocks & CRF
To assess where in the brain CRF might act to regulate exploratory
behavior in the MCC, the ability of CRF (20 ng) to affect exploratory behavior
was examined in rats in which the cerebral aqueduct (CA) was blocked to
prevent diffusion of the peptide throughout the ventricular system. The
blockade of the CA was achieved by injecting cold cream (Nivea) either
immediately rostral or caudal of the CA. These plugs remained patent for at
least 24 hours. CRF was then injected into either the lateral or fourth ventricles
Rats were implanted with cannulae as described in Chapter 2. To
determine the effect of a CA plug on the behavioral effect of CRF injected into
the lateral ventricles, one cannula was inserted for injection into the lateral
ventricle (A -0.8, L 2.5, V -4.0) and one for injection of cold cream just caudal
of the CA (A -7.0, L +1.0, V -4.5 from dura). To determine the effect of a CA plug
on the behavioral effect of CRF when injected into the fourth ventricle, cannulae
were implanted in the fourth ventricle (A -10.0, L 0.0, V -6.0 from dura) and just
rostral of the CA (A -3.8, L 1.0, V -6.5 from dura). Surgery was performed at
least 5 days prior to testing to allow sufficient recovery. On the day before
testing, the animals were weighed and transported from the colony room and








housed individually in the testing room. Two hours following transport from the

testing room, the animals were injected with 6-8 pl of the cold cream into the CA.
Injections were performed by drawing the cold cream into a section of
polyethylene tubing (PE-50) calibrated to hold 10 il via a short section of 23-
gauge stainless steel tubing. This stainless steel tubing was then attached with
a small piece of polyethylene (PE-50) tubing to the appropriate cannula in the
skull. The cold cream was then injected manually over a 45-60 sec period.
On the day of testing each animal was removed and injected with either

saline or CRF (2 gil) into either the lateral or fourth ventricles and placed in the
MCC 10 min later. The animals were observed for 25 min. Animals that
appeared ill and that had lost more than 15 g during the interval between
surgery and testing were not tested. Some of the animals that appeared normal
on the day prior to testing appeared ill (hypoactive or hyperreactive to handling)
on the day of testing, most likely related to damage resulting from injection of

the cold cream. These animals were excluded from testing. Some animals that
appeared normal prior to testing were markedly hypoactive when observed in
the MCC. Because this behavior was never observed in cannulated animals
receiving only saline or 20 ng CRF, these animals were excluded from the

analysis.
After the behavioral testing of all the animals was completed on a given
day, 2-4 gl of Indian ink was injected through the cannula through which CRF or
saline was injected. Five min later the rat was injected with a lethal dose of
pentobarbital. When the animal was deeply anesthetized, it was decapitated
and the brain excised and fixed in 9% formaldehyde. After a minimum of 24
hours the formaldehyde-fixed brain was frozen and sectioned on a freezing
microtome. The location of the cold cream and the spread of the dye throughout
the ventricular system was recorded. The assessment of an adequate plug was








determined from observing the entire 3-dimensional configuration of the plug as
sections of the brain were cut. Based on the completeness of the plug, animals
were grouped into the following categories: 1) Incomplete plug + lateral
ventricular CRF (n=9), 2) Complete plug + lateral ventricular CRF (n=8), 3)
Incomplete plug + fourth ventricular CRF (n=7), 4) Complete plug + fourth
ventricular CRF (n=9), 5) Animals with injections of saline into the lateral (n=5)
and fourth (n=5) ventricles with complete or incomplete CA plugs that received
saline. Nivea-injected animals that were not hypoactive did not differ in
measures of stimulus-contact or from the few cannulated animals receiving no
injections that were run at the beginning of the experiment as additional
controls (data not shown).
One-way analysis of variance of the five groups receiving cold cream
indicated a significant main effect of CRF on the mean time per contact with the
stimuli (F4,38 = 9.5, P < 0.001; Fig. 3-8). CRF injected into the lateral or fourth
ventricles decreased the mean time per contact with the stimuli in animals with
an incomplete block of the CA compared to saline-injected controls (group 5).
Blocking the CA had no effect on the behavioral effect of CRF when injected into
the lateral ventricles. In contrast, the CA block did prevent the CRF-induced
decrease in stimulus-contact time when injected into the fourth ventricle. The
mean time per contact of this group was not significantly different from the cold
cream-injected controls, but was significantly higher than animals receiving
CRF in the fourth ventricle with an incomplete block of the CA. There were no
significant effects of CRF on compartment entries or rears in any of these groups
(Fig 3-8).















3.0 100 1 M Mi
3.0 REARS
2.0 300


1.0- 200-
100

Sal L/In L/Com IV/ln IV/Com Sal L/Inc L/Com IV/ln IV/Com




Fig. 3-8. The mean time per contact, the total number of rears and the total
number of compartment entries of rats injected with CRF into either the lateral or
fourth ventricles. Rats were injected with cold cream into the CA the day before
testing. On the day of testing, they received an injection of either saline (open
bar) or 20 ng CRF into either the lateral (hatched bars) or fourth ventricles (solid
bars) 10 min prior to testing. Bars represent rats injected with saline (Sal) into
either the lateral or fourth ventricles, CRF injected into the lateral ventricles with
either incomplete (L/ln) or complete (L/Com) blocks of the CA or CRF injected
into the fourth ventricle with either incomplete (IV/ln) or complete (IV/Com)
blocks of the CA.
** Significantly different from saline + CA plug (P < 0.01). ++Significantly
different from IV ventricular CRF + incomplete CA plug (P < 0.01).








Discussion

The major behavioral effects of the treatments used in this series of
experiments were changes in the mean time per contact with the stimuli in the
multicompartment chamber. This measure was significantly reduced by 30 min
restraint, morphine (1.75 mg/kg SC), and by CRF (5-150 ng ICV). In general,
these effects resulted from a decreased total time spent in contact with the
stimuli without significant effects on the number of contacts. In contrast,
naloxone at a dose of 1.25 mg/kg, significantly increased the mean time per
contact, and the lower dose (0.70-0.75 mg/kg) which had no significant effect on
this measure by itself, significantly reversed the effects of morphine, restraint,
and CRF. The results with morphine and naloxone on investigatory behavior in
mice are qualitatively similar to those previously reported in rats by Arnsten et

al. (1979, 1981). The effect of restraint was also similar to that reported in rats in
which the mean time per contact with environmental stimuli was decreased
following exposure to a number of different stressors (Arnsten et al., 1985).
The ability of CRF to elicit stress-like responding has prompted the
hypothesis that this peptide might play a role in regulating behavioral and
physiological responding in stress (Gold et al., 1984; Koob & Bloom, 1985).
The fact that CRF concentrations are decreased in several brain regions
following stress suggests that CRF may be released during stress (Chappell et
al., 1986). Consistent with this is the preliminary report of a stress-induced
increase in the concentration of CRF in the cerebrospinal fluid (Britton et al.,
1984).
The effect of CRF on exploratory behavior in the MCC is consistent with

the hypothesis that endogenous CRF regulates behavioral responding in stress.
More direct support for this hypothesis is the ability of ahCRF to block the








restraint-induced decrease in exploratory behavior in a dose-dependent
manner (10-50 jg). Similarly, Brown et al. (1985) reported that 100 gjg ahCRF
ICV prevented the ether-induced increases in plasma glucose and Epi (but not
NE), and Rivier et al. (1986) reported that a similar dose of ahCRF prevented
the stress-induced inhibition of luteinizing hormone secretion in castrated male
rats. Further, 50 gg ahCRF antagonized the restraint-induced decrease in food
intake in 24-hr food-deprived rats (Krahn et al., 1986), and blocked the stress-
induced fighting in rats (Tazi et al., 1987).
Taken together, these results suggest that CRF within the brain may
mediate certain stress-induced changes in autonomic, endocrine, and
behavioral responses. However, the possibility that ahCRF acted at sites other
than CRF-receptors cannot be excluded in any of these studies. The need for
such comparatively high doses of the antagonist may be because ahCRF is a
relatively weak antagonist at CRF receptors. Thus, inhibition of the ether-
induced ACTH secretion required a dose of ahCRF 400 times greater than the
dose at which CRF stimulates ACTH secretion (Rivier et al., 1984).
Although the antagonist had no statistically significant quantitative effects

on locomotor activity in the MCC, certain qualitative effects were observed in
both restrained and unrestrained mice. Many of the animals in the groups
receiving 10-50 jg ahCRF displayed a more rounded posture, with a shuffling
and hesitant manner of locomotion. This behavior resembles that we have
observed with high doses of CRF, morphine or naloxone. Therefore, it is
possible that in some of the animals the peptide produced effects unrelated to
action at CRF-receptors.
It is unlikely that the consistent action of the antagonist on the restraint-
induced decrease in stimulus-contact time can be explained simply by

nonspecific effects on locomotor activity. First, not all restrained animals that








received the antagonist and showed longer mean contact durations also
displayed an abnormal pattern of locomotion. Second, in none of the
experiments, including those with the antagonist, was there a statistically
significant relationship between measures of locomotor activity and stimulus-
contact times. The possibility has been raised that the specificity of the
behavioral and physiological actions of CRF could in part be regulated by
further processing of this 41-amino acid peptide, or by interactions with other
brain or pituitary peptides (Veldhuis & de Wied, 1984). Additional studies
examining neurochemical and behavioral responses to CRF and potential
metabolites are needed to assess how and where CRF exerts its actions.
If endogenous CRF regulates behavioral responding in stress, the site(s)
of this action is (are) unknown. CRF was effective in decreasing stimulus-
contact times when injected into either the lateral or fourth ventricles.
Preventing diffusion of CRF into the fourth ventricle by blocking the CA did not
diminish the behavioral effect of CRF when injected into the lateral ventricle. In
contrast, blocking the CA did prevent the behavioral effect of CRF injected into
the fourth ventricle. These results suggest that CRF acts to influence behavior
in this paradigm at a site anterior to the CA. In examining the effect of CA plugs

on the locomotor activating effects of CRF, a similar conclusion regarding the
site of action for CRF was reached (Tazi et al., 1987). From additional studies
examining the effect of CRF injected directly into various sites within the brain,
these authors concluded that the ventral forebrain was the general region
where CRF acted to elicit this effect. Using a similar, but more extensive,
approach to study ACTH-induced grooming (Dunn & Hurd, 1987), angiotensin-
induced drinking (Hoffman & Phillips, 1976), and the bradykinin-induced

increase in blood pressure (Phillips, 1984), the organum vasculosom region of
the third ventricle (AV3V) was determined to be the critical site along the








ventricular surface at which these peptides acted to elicit their respective effects.
Thus, one question of interest is whether the critical site for the behavioral

action of exogenous CRF on exploratory behavior is the AV3V region.
In an attempt to localize the site where CRF acts to elevate plasma NE
concentration, Brown (1986) microinjected CRF into various brain areas.
Although injection of CRF into many sites elicited a plasma NE response in
none of these CRF-sensitive sites did CRF injection elicit a response larger than
that observed with injection of CRF into the third ventricle. Thus, it remains
unclear whether CRF acts at a single site or multiple sites to elevate plasma NE
concentrations. Krahn et al. (1988) demonstrated that CRF inhibited food intake
when injected into the PVN but not when injected into the lateral hypothalamus,

the ventromedial hypothalamus, the globus pallidus or the striatum of rats. CRF
injected into the PVN also increased grooming and locomotor activity.
Although centrally administered CRF has certain behavioral effects, the
mechanisms by which these effects are mediated are unknown. The fact that
CRF and restraint decreased exploratory behavior in hypophysectomized mice
indicates that an activation of the HPA axis by these two treatments is not an
essential component in mediating their behavioral effects. Further, the dose
range within which CRF affects behavioral responding in this paradigm is well
below that required to observe stimulatory effects on heart rate, blood pressure
and adrenal catecholamine secretion (Brown et al., 1982). Therefore, it is
unlikely that CRF affects exploratory behavior in this paradigm through any of
the peptide's documented peripheral effects.
The ability of naloxone to antagonize the CRF-induced decrease in

exploratory behavior contrasts with a report that found no effect of naloxone on
CRF-induced alterations in the behavioral response to a familiar environment

(Koob et al., 1984). It is likely that this discrepancy is due to a combination of








differences in the testing paradigms used, the duration of observation, and the
dose of CRF. Naloxone antagonized the CRF-induced increase in grooming
(Dunn et al., 1987) and the CRF-induced decrease in both male and female

sexual behavior (Sirinathinghji, 1985, 1987).
The antagonistic effect of naloxone on the effects of CRF in the MCC
suggests that CRF influences exploratory behavior by stimulating the release of
endogenous opioids. However, it does not prove the involvement of such a
mechanism and the possibility remains that these two systems are acting in
parallel to affect exploratory behavior. For example, restraint sensitized the
mice to the behavioral effects of naloxone, such that the mean time per contact
was significantly increased in the restrained + 0.75 mg/kg naloxone group

compared to saline-controls and 0.75 mg/kg naloxone controls. Thus, it is
possible that CRF increases the sensitivity of the animals to the behavioral
effects of naloxone, without directly interacting with opioid systems. In either
case, these results suggest that endogenous CRF affects exploratory behavior
in rats and mice by acting at a site anterior to the cerebral aqueduct, and that
this action involves either direct or indirect interactions with endogenous opioid
systems.












CHAPTER IV
THE ROLE OF CENTRAL NORADRENERGIC SYSTEMS IN REGULATING
RESTRAINT-STRESS-INDUCED CHANGES IN EXPLORATORY BEHAVIOR


An activation of central noradrenergic systems is a general response
observed in stress (Stone, 1975; Dunn & Kramarcy, 1984; Glavin, 1985).
Reports that activation of noradrenergic systems elicit stress-like and anxiety-
like behavioral and emotional responding suggested that the stress-induced
increase in NE release may regulate certain aspects of behavioral responding
in stress. Therefore, the possible involvement of NE in regulating the restraint-
induced decrease in exploratory behavior in the MCC was examined using two
different approaches. In the first, the behavioral effect of the acute stimulation
and inhibition of noradrenergic systems was observed in restrained and
unrestrained mice. In other experiments, the effect of noradrenergic-specific
lesions was examined. The use of both lesion and pharmacological inhibition
of noradrenergic activity was chosen to control for questions that arise in the
interpretation of results associated with each individual approach. Analysis of
brain biogenic amines and their catabolites using HPLC-EC was performed to
verify that the treatments had the desired effect on NE systems and to provide
an indication of the magnitude of these effects.

Noradrenergic-specific lesions.
The noradrenergic-specific toxin, N-(2-chloroethyl)-N-2-

bromobenzylamine (DSP-4) primarily affects neurons of the DNB to produce a
long-term and selective depletion of NE in certain brain structures within hours
after IP administration (Ross, 1976; Jaim-Etcheverry & Ziehr, 1980; Jonsson et








al., 1981). Although the exact mechanism of its toxicity is not known, this action
is dependent on uptake of DSP-4 into noradrenergic neurons (Hallman &
Jonsson, 1984). This toxin also depletes NE in certain peripheral tissues,
including iris, heart, and salivary gland but not the superior cervical ganglion
(Jaim-Etcheverry & Ziehr, 1980; Jonsson et al, 1981). The effect in the
periphery is not permanent and an almost complete recovery is observed within
1 week following administration of DSP-4. The greatest depletion of NE within
the brain is observed in the neocortex and the hippocampus. In these regions
DSP-4 administration (50 mg/kg, IP) results in a NE depletion of 80-90%. A less
dramatic depletion of approximately 60% is observed in the cerebellum,
whereas depletions of only 20-40 % are observed in the brainstem and
hypothalamus. The action of this toxin is highly specific, affecting primarily
neurons arising from the LC with little effects on DA, 5-HT, acetylcholine or
gamma-aminobutyric acid (GABA) concentrations (Jaim-Etcheverry, 1983).
Other than questions of specificity, the greatest concern with the use of
lesions is the possibility of compensatory responses initiated by the lesion that
act to minimize the effect of that lesion. For example, lesions of NE systems
within the brain result in alterations in the number of P3- and al-binding sites that

are negatively correlated with the NE concentration in those regions (U'Prichard
et al., 1980). Thus, 4-6 weeks following a 6-OHDA lesion of the DNB, cortical
and hippocampal NE was significantly decreased whereas P-receptor binding
was increased. In contrast, the number of al- and P-binding sites was

decreased in the cerebellum, where NE concentrations were increased.
Binding of the selective a2-agonist, clonidine, was also elevated in frontal

cortex but decreased in the amygdala and septum in these animals. Similar
results are seen following 6-OHDA lesions of the LC or treatment with DSP-4
(Harik et al., 1981; Mogilnicka, 1986). In addition, P3-receptor-mediated








stimulation of cyclic adenosine monophosphate (cAMP) formation was also
enhanced following 6-OHDA lesion of the LC (Dismukes et al., 1975; Sporn
1977; Skolnick, 1978; Dolphin et al., 1979; Harik et al., 1981).
Increases in NE turnover as measured by MHPG:NE ratios have also

been described in cortex and hippocampus 7 days following IP injection of
DSP-4 (Hallman & Jonsson, 1984; Logue et al., 1985). In addition, changes in
other transmitter systems have been noted. For example, a2-mediated release

of serotonin from hippocampal slices of DSP-4 treated rats was increased 10
days after administration of the toxin (Benkirane et al., 1985). Further,
concentrations of DA and its catabolites were increased in the ipsilateral cortex
following unilateral lesion of the LC (Harik, 1984). Therefore, lesions of NE
systems result in a number of compensatory responses throughout the brain, all
of which are present within 2 weeks after the lesion.
In addition to changes in receptor functioning and transmitter release,
behavioral studies also indicate an increased sensitivity of noradrenergic
responding following noradrenergic lesions. For example, a supersensitivity to
the depressive effects of clonidine and clenbuterol (a 13-agonist) on exploratory

behavior (head-dipping) was observed 10 days following treatment with DSP-4
(Dooley et al., 1983). Thus, although rats treated with DSP-4 showed no
difference in this behavior compared to untreated rats, the effect of clonidine
was enhanced in the DSP-4 treated rats. Similarly, Mogilnicka (1986) reported
an increased sensitivity to the ai-selective agonist, phenylephrine, on

exploratory behavior in an open field in DSP-4 treated rats.
These compensatory responses to NE depletion indicate the necessity of

carefully choosing a time point after a NE lesion to conduct behavioral testing
so that the activity of this system is in fact decreased. The general practice of
waiting at least 2 weeks to allow for sufficient depletion of NE and recovery of








the animal may in fact prevent the observation of functional consequences of
these depletions. Therefore, in the following experiments, the effects of lesions
of noradrenergic systems in the brain were examined in animals 3 days
following treatment with DSP-4. This toxin was chosen to avoid the recovery
period required following the use of 6-OHDA.

Pharmacology of noradrenergic systems.

Pharmacological inhibition and stimulation of noradrenergic systems has
been argued to be a more appropriate approach to study an involvement of NE
in regulating behavioral responding than the use of lesions of these systems
(Sara, 1985). For example, when the effect of pharmacologically stimulating NE
release was examined on acquisition of a linear maze task, an increase in the

rate of acquisition and retention of the task was observed, whereas no effect of
NE lesions could be observed (Sara, 1985). However, pharmacological
approaches also have inherent problems. One such problem is that the drugs
used to act at specific receptors rarely display an absolute specificity. One way
to partially overcome this problem is to use more than one drug known to act on
that transmitter system. It is unlikely that all drugs tested will have the exact
same actions at other receptor subtypes or on other transmitter systems.
The pharmacology of noradrenergic systems has been extensively
studied. Among the four defined subclasses of noradrenergic receptors (pi, P2,
al and 0a2), a2-receptors were originally proposed to be presynaptic
autoreceptors acting to inhibit NE-release, whereas ai-receptors were thought

to be postsynaptic. However, it now appears that in certain tissues in the
periphery, a2-receptors also exist postsynaptically (see Starke & Docherty,

1980 for review). Further, because lesions of central noradrenergic systems
resulted in an upregulation of a2-binding sites, it is thought that a large








proportion of these receptors in the brain exist postsynaptically (U'Prichard et
al., 1977; U'Prichard et al., 1980).
Clonidine is an a2-agonist that has been shown to inhibit spontaneous

LC firing in anesthetized rats when administered either directly in the LC or
systemically at doses above 5 jg/kg (Svensson et al., 1975; Marwaha &
Aghajanian, 1982). Inhibition of NE activity by clonidine is also indicated from
studies measuring NE metabolites as an index of NE turnover. Thus, clonidine
was observed to inhibit the accumulation of LC DOPAC measured by in vivo
voltammetry (Quintin et al., 1986), DHPG measured by multiple ion detection
mass fragmentography (Warsh et al., 1981), and DOPEG measured using a
radioenzymatic assay (Scatton, 1982). An inhibition of noxious stimulus-
induced LC firing by clonidine has also been reported (O'Neill & Haigler, 1984;
Quintin et al., 1986).
The a2-antagonist, idazoxan, antagonizes the actions of clonidine and

independently stimulates noradrenergic activity in a dose-dependent manner
as indicated by electrophysiological and neurochemical measurements
(Goldstein et al., 1983; Scheinin & Virtanen, 1986). Similar results using other
a2-agonists and antagonists have been demonstrated (Warsh et al., 1981;
Marwaha & Aghajanian, 1982; Scatton, 1982). In contrast, ai-antagonists have
no effect on noradrenergic firing. Thus, a2-agonists and antagonists provide a

mechanism for noninvasively modulating the rate of NE release within the brain.







Methods

Materials

Clonidine, phenylephrine, prazosin, and DSP-4 were obtained and
prepared as described in Chapter 2.

Procedures
Male mice were used in all experiments and were obtained and housed
as described in Chapter 2. In experiments requiring ICV injections, cannulae
were implanted for injection into the lateral ventricles as described in Chapter 2.
In experiments in which cannula implantation was not required, animals were
group-housed throughout the experiment. Cannulated animals were housed
singly immediately following surgery. Behavioral testing was conducted

between 08:30 and 16:00. The testing chambers and procedures were those
described in Chapter 2.
Restrained animals were transferred to a room separate from both the

colony and testing rooms and restrained for 40 min. Restraint was administered
as described in Chapter 2. Immediately following restraint, the animals were
carried into the testing room and placed in the testing chamber. Unrestrained
animals were transferred from the colony room to the testing room and
immediately placed in the testing chamber.
All injections were administered in the colony room. In experiments
involving administration of a drug to restrained animals, the injection was given
5 min prior to restraint. Unrestrained animals were injected and placed back in
their home cage where they remained until testing. All drugs were injected
intraperitoneally (IP) except CRF, ahCRF and phenylephrine which were
administered ICV. DSP-4 (50 mg/kg) was injected three days prior to testing.








Animals with cortical concentrations of NE greater than 50% of the mean of the
saline-controls were excluded from further analysis.

Statistical Analysis
Analysis of the behavioral data was as described in Chapter 2. For
analysis involving only the comparison of experimental groups to one control
group, Dunnett's t-test was used.

Results

The effect of clonidine and idazoxan on exploratory behavior.
We first tested the effects of clonidine and idazoxan on the behavioral

response of mice to the MCC. Significant main effects were observed on the
mean time per contact (F5,41 = 6.65, p < 0.001; Fig. 4-1). Clonidine, injected at a
dose of 25 gg/kg, but not 10 jg/kg significantly increased the mean time per
contact with the stimuli. This effect was principally due to an increase in the
duration of contacts, with no change in the total number of contacts, but neither
of these parameters were altered significantly (data not shown). Clonidine did
not alter locomotor activity as measured by either the number of compartment

entries or rears, suggesting that thr effect effect of clonidine on exploratory
behavior was not due to a nonspecific effect on locomotor activity (Fig. 4-1). At
a dose of 50 jg/kg (data not shown), variable data were obtained with many
animals showing a pronounced sedation (increased inactivity and decreased
compartment entries) and no change in stimulus-contact time, whereas others
showed stimulus-contact times similar to those treated with the 25 ig/kg dose.
The effect of clonidine at the 50 jg/kg dose is consistent with its known sedative
action.








Idazoxan had an effect opposite to that of clonidine. At a dose of 1 mg/kg
it significantly decreased stimulus-contact times (Fig. 4-1). At a dose of 3 mg/kg,
idazoxan had a similar effect on the mean time per contact, but which was not
quite statistically significant (0.05 < 2P < 0.1). A significant decrease in total
duration of contact with the stimuli was observed at the 3.0 mg/kg dose (F5,41 =
5.43, P < 0.001); data not shown). Idazoxan at 0.1 mg/kg had no significant
effects on any measure of stimulus contact. Like clonidine, idazoxan had no
statistically significant effects on compartment entries, or rears (Fig. 4-1). In pilot
studies, the effect of 25 jg/kg clonidine was reversed by the a2-antagonist,

yohimbine (1 mg/kg), suggesting that the effect of clonidine was indeed due to
its action on a2-receptors.

The effect of NE depletion on the restraint-induced decrease in exploratory
behavior.
Mice were injected with either saline or DSP-4 3 days prior to testing. On
the day of testing, saline- and DSP-4-treated mice were either removed from
their home cages and immediately tested, or restrained for 40 min prior to
testing. DSP-4 depleted NE in cortex by approximately 75% compared to quiet-
controls. Significant main effects of DSP-4 (Fi,26 = 45.5, P < 0.001) and
restraint (F1,26 = 23.3, P < 0.001) on the mean time per contact were observed
(Fig. 4-2). DSP-4 treatment increased and restraint decreased the mean time
per contact. The restraint-induced decrease in the mean time per contact was
significantly antagonized by DSP-4, but restraint still had a significant effect on
this behavior in DSP-4-treated animals. However, the magnitude of the
restraint-induced decrease appeared to be less in DSP-4-treated animals
(21%) than in saline-injected animals (48%). Neither restraint nor DSP-4 had
any significant effects on the number of rears or compartment entries (Fig. 4-2).











MEAN TIME PER CONTACT (sec)







T


Sal 10 25
Clonidine


0.1 1.0
Idazoxan


Sal 10 25 0.1 1.0 3.0
Clonidine Idazoxan


Fig. 4-1. Effect of clonidine and idazoxan on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were
injected with either saline, 10 or 25 gg/kg clonidine, or 0.1, 1.0. or 3.0 mg/kg
idazoxan 10 min prior to testing.
*Significantly different from saline-controls (P < 0.05).


3.0



2.0



1.0















3.0









Saline Restraint DSP-4 Restraint Sailne Restraint DSP-4 Restraint
+ +00





DSP-4 DSP-4
Fig. 4-2. Effect of DSP-4 and restraint on the mean time per contact, the total






number of rears and the total number of compartment entries. Mice injected IP
with either saline or DSP-4 3 days before testing were either removed from their
home cage and immediately tested or restrained for 40 mmn immediately prior to
testing.
**Significantly different from saline-unrestrained mice (P < 0.01).
40Significantly different from saline-restrained mice. (P < 0.05).
1.030
200-

Saline Restraint DSP-4 Restraint Saline Restraint OSP-4 Restraint
DSP-4 DSP-4




Fig. 4-2. Effect of DSP-4 and restraint on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice injected IP
with either saline or DSP-4 3 days before testing were either removed from their
home cage and immediately tested or restrained for 40 min immediately prior to
**ting'Significantly different from saline-unrestrained mice (P < 0.01).
+Significantly different from saline-restrained mice. (P < 0.05).







The effect of clonidine on the restraint-induced decrease in exploratory
behavior.
Unrestrained and restrained mice were injected with either saline or 25
jg/kg clonidine 45 min prior to testing. Significant main effects for restraint (F1,28
= 41.7, P < 0.001) and clonidine (Fi,28 = 36.3, P < 0.001) were observed for the
mean time per contact (Fig. 4-3). Restraint decreased the mean time per
contact in saline-injected animals, whereas clonidine increased this response
in unrestrained animals and antagonized the restraint-induced decrease in this
response. Restraint significantly decreased the mean time per contact in
clonidine-injected animals. However, as was observed with DSP-4, the
magnitude of this decrease appeared to be less in the clonidine-treated mice
(26% vs 41%). In this experiment, clonidine significantly decreased the number
of compartment entries in unrestrained animals (F1,28 = 6.03, P = 0.02), an effect
of clonidine not observed previously (see Fig. 4-1). Restraint had no effect on
compartment entries. Neither restraint nor clonidine had any significant effects
on rearing (Fig. 4-3).
The concentration of MHPG in mice tested in the MCC was increased in
the nucleus accumbens (NAS), hypothalamus (HTH) and hippocampus (HPC)
in saline-unrestrained animals and in the NAS, HTH, HPC and brainstem (BST)
in saline-restrained mice as compared with quiet-controls (Table 4-1). In
animals tested in the MCC, ANOVA indicated significant main effects of restraint
on the concentrations of MHPG in NAS (F1,26 = 12.5, P = 0.001), HTH (F1,27 =
16.8, P < 0.001), and a marginally significant effect in HPC (F1,26 = 3.8, P =
0.058) such that restraint increased MHPG concentrations in these regions
compared to unrestrained mice. These results suggest that exposure to the
MCC increases the release of NE, an effect exacerbated by prior restraint. Main

effects of clonidine were observed in NAS (F1,26 = 9.8, P = 0.004), HTH (F1,27 =











MEAN TIME PER CONTACT (sec)


Saline Restraint Clon Restraint Saline Restraint Clon Restraint
+ 4.
Clon Clon



Fig. 4-3. Effect of clonidine and restraint on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were
injected with either saline or clonidine (25 gg/kg) 45 min prior to testing.
Restrained mice were injected with either saline or clonidine (25 ig/kg) 5 min
prior to restraint.
*Significantly different from saline-unrestrained mice (P < 0.05,
**P < 0.01). ++Significantly different from saline-restrained mice. (P < 0.01).


3.0 1


2.0 F


1.0I









Table 4-1. Effects of exposure to the MCC and clonidine on the concentration of
MHPG in nucleus accumbens (NAS), hypothalamus (HTH), hippocampus
(HPC) and brainstem (BST).


NAS HTH HPC BST


Saline-unrestrained 121 8* 126 5* 131 4* 120 4


Saline-restrained 147 4** 168 11** 162 18** 148 20*


Clonidine-unrestrained 96 9 85 9 96 12 118 14


Clonidine-restrained 12410*+ 12311*+ 11513+ 137 30


Mean concentrations of MHPG expressed as percentage of quiet controls (
SEM). Animals were removed immediately following behavioral testing and
NAS, HTH, HPC and BST were excised.
*Significantly different compared to quiet-controls (P < 0.05, **P < 0.01;
Dunnett's t-test). +Significantly different compared to saline-restrained mice (P
< 0.05; Duncan's multiple-range test).








18.8, P < 0.001), and HPC (F1,26 = 10.2, P = 0.003). In these three regions,
clonidine blocked the increase in MHPG in unrestrained animals and
significantly attenuated this increase in restrained animals (Table 4-1).

Biogenic amines response in select brain regions and concentration of plasma
corticosterone following exposure to the MCC.
Analysis of plasma and brain tissue of animals exposed to the MCC
indicated that like other stressors, exposure to the MCC elicited an increase in

the concentration of plasma corticosterone and an increase in release of brain
biogenic amines. However, in these experiments the variability of the
monoamines and their catabolites was high, probably because of the collection
of tissue over a number of days and at different times throughout the day,
necessitated by the experimental design. Therefore, the effect of exposure to
the MCC on the concentrations of plasma corticosterone and brain biogenic
amines and their catabolites was examined under temporally controlled
conditions. A group exposed to 15 min footshock was included to provide a
comparison with the effects of a well characterized stressor. Animals were
housed singly 3 days prior to testing and randomly assigned to one of 3 groups;
Quiet (n=7), a 15 min period of footshock (n=8), or a 15 min period of exposure
to the MCC (n=8). Footshock consisted of 20-22 1-sec footshocks of 0.2 mA
over the 15 min period. Quiet animals were left in the colony room until they
were transported for dissection of the brain. Immediately following the period of

footshock or exposure to the MCC, the animals were decapitated, trunk blood
was collected and the brain was excised and PFM, NAS, HTH and BST were
dissected as described in Chapter 2. HPLC analysis of the tissue and RIA
measurement of plasma corticosterone were as described in Chapter 2. One-
way ANOVA was used to determine statistical significance with Duncan's







Table 4-2. Concentrations of plasma corticosterone from quiet, footshocked-
treated and mice exposed to the MCC.

Quiet Footshock MCC

Corticosterone (ng/ml) 29 4 236 21 ** 231 28**

Mean ( SEM) concentrations of plasma corticosterone. Quiet mice remained in
their home cage until trunk blood was collected. Footshock-treated mice
received 20-22 0.15 mA footshocks over a 15 min period. MCC-exposed mice
were placed in the MCC for 15 min. **Significantly different from quiet-controls
(P < 0.01 compared).

multiple range test used for comparison between groups.
Shown in Table 4-2, both exposure to the MCC and footshock
significantly increased the concentrations of plasma corticosterone (F2,2o = 30.6,
P < 0.001). The concentrations of the NE catabolite, MHPG, and the MHPG:NE
ratio for PFM, NAS, HTH, and BST expressed as percentage of quiet-controls
are shown in Fig. 4-4. Significant main effects on MHPG were observed in PFM
(F2,18 = 7.5, P = 0.004), HTH (F2,2o = 20.9, P < 0.001) and BST (F2,19 = 7.4, P =
0.004). In these three regions MHPG was increased in the footshock-treated
animals as compared to quiet-controls. MHPG was significantly increased in
the HTH and BST of the MCC-exposed animals. A similar pattern was
observed for the MHPG:NE ratio. This ratio was significantly increased in both
the footshocked and MCC-exposed mice in the HTH (F2,2o = 43.4, P < 0.001)
and BST (F2,19 = 12.44, P < 0.001) and in the PFM of footshocked mice (F2,2o =
11.0, P < 0.001). The increase in the MHPG:NE ratio was significantly greater in
the footshock-treated group than the MCC-exposed group in the PFM and HTH.
Both the dopamine catabolite, DOPAC, and the DOPAC:DA ratio were
significantly affected by these treatments (Fig. 4-5). In the PFM (F2,18 = 5.5, P =
0.013) and BST (F2,20 = 7.0, P = 0.005), DOPAC was significantly increased in






63





300 7 ---------------
250E300 SHOCK
S250-* E MCC
D
0 200 .- ++
LL *
300 -- **
Z
.. III
65 150 ++-




''":.
z 100 @

CL 50-

300 . .

250 **
5
( 200+
0 r- **
q. 150
"r +
z
w 100.

C. 50.

0
PFM NAS HTH BST
REGION



Fig. 4-4. The effect of footshock and exposure to the MCC on the
concentrations of MHPG and the MHPG:NE ratios in various brain regions
(prefrontal cortex, medial subdivision, PFM; nucleus accumbens, NAS;
hypothalamus, HTH; brainstem, BST). Bars represent the mean SEM
expressed as percentage of quiet controls. Mice either received 15 min of
footshock (20-22 0.15 mA footshocks; open bars) or were placed in the MCC for
15 min (shaded bars).
*Significantly different from quiet-controls (P < 0.05, **P < 0.01).
+Significantly different from footshock-treated mice (P < 0.05, ++P < 0.01).






64







250 -
l SH-KCK

a ** **
wI 200 T-


0 15 7- ..
ow
a. <
O F- 100
o Z
w

W 50
a.

250 -O--------


1. 200
5 +
0
o w 150 -


oz
w 100
UJ
w
a. 50

Q. -. ___ ,
0- -------------
PFM NAS HTH BST
REGION


Fig. 4-5. The effect of footshock and exposure to the MCC on the
concentrations of DOPAC and the DOPAC:DA ratios in various brain regions
Data are from the same experiment as in Fig Fig. 4-4.
*Significantly different from quiet-controls (P < 0.05, **P < 0.01).

+Significantly different from footshock-treated mice (P < 0.05, ++P < 0.01).








the footshocked mice. No significant effect on DOPAC was observed in the
MCC-exposed mice. The concentration of DOPAC in the PFM of footshock-
treated mice was significantly greater than that of the MCC-exposed mice.
However, there was no significant difference between the DOPAC
concentrations of footshocked and MCC-exposed mice in the BST. Statistically
significant effects of the treatments on DOPAC:DA ratios were observed in PFM

(F2,20 = 8.7, P = 0.002), NAS (F2,19 = 5.9, P = 0.01), HTH (F2,2o = 4.2, P = 0.03)
and BST (F2,2o = 10.5, P < 0.001). A statistically significant increase was
observed in all regions of both groups except in the PFM of the MCC-exposed
mice. The concentrations of another DA catabolite, HVA, were increased only
in the NAS of the MCC exposed-mice (F2,2o = 4.0, P = 0.034; Fig 4-6). The
HVA:DA ratio was increased in NAS in both the footshocked and MCC-exposed
mice (F2,2o = 8.8, P = 0.002).
The serotonin (5-HT) catabolite, 5-HIAA (F2,2o = 4.5, P = 0.024) and the 5-
HIAA:5-HT ratio (F2,2o = 6.6, P = 0.006) were significantly affected only in the
NAS (Fig 4-7). In this region, 5-HIAA was increased in the MCC-exposed mice.
Although the concentration of this catabolite was not significantly increased in
footshock-treated mice compared to quiet-controls, it was not significantly
different from that of the MCC-exposed mice. The 5-HIAA:5-HT ratio was
significantly increased in this region in both experimental groups. Tryptophan
was significantly increased in the NAS (F2,2o = 6.3, P = 0.007) of both the
footshocked and MCC-exposed mice and in the MCC-exposed mice in the HTH
(F2,2o = 4.4, P = 0.026; Fig 4-8). However, there was no significant difference in
the concentrations of Trp in the HTH between the two experimental groups





66






200

El SHOCK
w
5 150 .
0T
< 0 _
S5 100-

z
w




< w --- T *

ILl
C
c 50 -



200- -- -- -- -- -- --I-


F-
| 150-

0 _
> 100 -


w

0
I 50- '




PFM NAS HTH BST
REGION




Fig. 4-6. The effect of footshock and exposure to the MCC on the
concentrations of HVA and the HVA:DA ratios in various brain regions. Data
are from the same experiment as in Fig. 4-4 and are expressed in terms of
percentage of quiet-controls.
*Significantly different from quiet-controls (P < 0.05, **P < 0.01).





67





200 -
D SHOCK
** O[mO
w l MCC
D 150-
I- L.L

t"o |w
c 5 I 0 |- ._._._ ..:


.:. . .. ...

,250 -


F- 200
< -' 100 ^
w
D
o




< 5 0_









PFM NAS HTH BST
REGION




Fig. 4-7. The effect of footshock and exposure to the MCC on the
concentrations of 5-HIAA and the 5-HIAA:5-HT ratios in various brain regions.
0J





















Data are from the same experiment as Fig. 4-4 and are expressed in terms of
percentage of quiet-controls.
Significantly different from quiet-controls (P < 0.05, **P < 0.01).
2-50 f

tii 20
5F NA* T S
REIO


Fi.4-.Th ffc o oosok n epsue150e C o h
cocnrtoso<-IAan hw-IA5H ais nvrosbanrgos
Dat 100rmtesm eprmn sFg.44adaeeprse ntrso
pecntg of quietcontrols
*Sgiiatydfeetfo0ue-otos( .5 .1)





68









250-
J SHOCK
200- El MCC
Lu 200 OO
5
z O
< L.L **
= O 150- **
o

0U 50
'" T T. -j ^i r|
I.-

0~0
0- -- 100aia ^_ ~fs;_^ wKm __ ^ffl


PFM NAS HTH BST
REGION


Fig. 4-8. The effect of footshock and exposure to the MCC on the
concentrations of tryptophan in various brain regions. Data are from the same
experiment as in Fig. 4-4 and are expressed in terms of percentage of quiet-
controls.
*Significantly different from quiet-controls (P < 0.05, **P < 0.01).







The effect of combined NE depletion and clonidine on the restraint-induced
decrease in exploratory behavior.
Because clonidine does not completely block NE release (Table 4-1) and
DSP-4 does not completely deplete NE, these two treatments were combined to
produce a more effective inhibition of NE release. Animals were injected with
either saline or 50 mg/kg DSP-4 3 days prior to testing. On the day of testing,
mice previously injected with saline were injected with saline either 45 min
before testing or 5 min before being restrained. DSP-4-injected animals were
injected with 25 pg/kg clonidine either 45 min before testing or 5 min before
being restrained.
DSP-4 depleted cortical NE by 65% compared to saline-unrestrained
mice (data not shown). The less effective depletion observed in this experiment
can likely be explained by the fact that exposure to the MCC alone results in a
slight decrease (10-20%) in cortical NE. Main effects of DSP-4 (Fi,26 = 32.5, P <
0.001), restraint (F1,26 = 6.65, P < 0.015), and a significant interaction between
the two (F1,26 =4.3, P < 0.05) were observed for the mean time per contact (Fig.
4-9). DSP-4 + clonidine increased the mean time per contact in unrestrained
animals and completely blocked the effect of restraint on this response. The
significant interaction term apparently results from the fact that the mean time
per contact of the restraint plus DSP-4/clonidine-treated animals did not
significantly differ from the unrestrained-saline controls. Restraint significantly
decreased the number of rears in the DSP-4 + clonidine-treated animals (F1,26
= 6.2, P < 0.02). Neither restraint nor DSP-4 + clonidine significantly affected
compartment entries (Fig. 4-9).

The effect of the al-antagonist. prazosin. on the restraint-induced decrease in
exploratory behavior.
The above experiments indicate that an activation of noradrenergic

systems mediate the behavioral effect of restraint. To determine the subtype of












MEAN TIME PER CONTACT (sec)


T ++
T T


Saline Restraint DIC Restraint


Saline Restraint D/C Restraint


D/C


D/C


Fig. 4-9. Effect of DSP-4 plus clonidine and restraint on the mean time per
contact, the total number of rears and the total number of compartment entries.
Mice were injected IP with either saline or DSP-4 three days before testing. On
the day of testing, saline-injected mice were injected with saline 45 min before
testing or 5 min prior to restraint. DSP-4-injected animals were injected with 25
jig/kg clonidine 45 min before testing or 5 min prior to restraint.
*Significantly different from saline-unrestrained mice (P < 0.05,
**P < 0.01). ++Significantly different from saline-restrained mice. (P < 0.01).


3.0-


2.0 F


1.0-








postsynaptic receptor involved in mediating the restraint-induced change in
stimulus contact, saline or the al -specific antagonist, prazosin (50, 100 or 200

gg/kg) was administered 45 min before testing to restrained and unrestrained
mice.
Significant main effects on the mean time per contact (Fig. 4-10) were
observed for both restraint (F1,49 = 8.9, P <0.005) and prazosin (F3,49 = 11.4, P <
0.001). Restraint significantly decreased the mean time per contact in the saline

and 50 jg/kg prazosin groups. In unrestrained mice, prazosin increased the
mean time per contact in a dose-dependent manner with a maximum effect
observed at the 100 jg/kg dose. Prazosin antagonized the restraint-induced
decrease in mean time per contact in a dose-dependent manner, with a
complete blockade at the 200 gg/kg dose. Main effects of restraint on rears
were observed (F1,49 = 20.0, P < 0.001) such that restraint significantly
decreased the number of rears in the saline-, 100 gg/kg and 200 ig/kg prazosin-
treated animals (Fig. 4-10). A significant main effect of restraint on compartment
entries was also observed (F1,49 = 4.2, P < 0.05; Fig. 4-10). However, post-hoc
analysis demonstrated that the only significant difference between groups was
a decrease in compartment entries in the restrained + 200 jg/kg prazosin group
compared to the unrestrained + 200 jg/kg group. Prazosin did not significantly
affect either compartment entries or rears.

The effect of the al-agonist. phenyleDhrine (ICV). on behavior in the MCC.

The reversal of the restraint-induced decrease in stimulus-contact time by
an al-antagonist suggested the involvement of ai-receptors in regulating

restraint-induced responding. To test this hypothesis further, mice were injected
ICV with either saline or the al -selective agonist, phenylephrine, at 50 or 100

ng 10 min prior to testing. Both doses of phenylephrine significantly decreased














4.0- + UU
200

3.0 100
REARS
2.0 400
300-
1.0, 200
100

Saline 50 100 200 Saline 50 100 200

PRAZOSIN (pg/kg) PRAZOSIN (pg/kg)


Fig. 4-10. Effect of prazosin and restraint on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were
injected with either saline or 50, 100 or 200 pg/kg prazosin 45 min prior to
testing. In relevant groups, mice were restrained 5 min later. Open bars
represent unrestrained mice. Solid bars represent restrained mice.
*Significantly different from saline-unrestrained mice (P < 0.05).
++Significantly different from saline-restrained mice. (P < 0.01).










3.0 ,.U
400:
300:
T 200
2.0 100
REARS
400
1.0- 300 T
200:
100:
Id N I M I I o
Saline 50 100 Saline 50 100
ICV PHENYLEPHRINE (ng) ICV PHENYLEPHRINE (ng)


Fig. 4-11. Effect of phenylephrine on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were
injected ICV with either saline or 50 or 100 ng phenylephrine 10 min prior to
testing.
**Significantly different from saline-controls (P < 0.01).








the mean time per contact with the stimuli (F2,18 = 24.3, P < 0.001; Fig. 4-11).
Neither dose of phenylephrine had significant effects on the total number of
rears or compartment entries (Fig. 4-11).

Discussion

As described previously, restraint stress decreased the time mice spent
investigating the wire stimuli in this novel environment without consistent effects
on locomotor activity. This effect was mimicked by the ai -selective agonist,
phenylephrine, injected ICV (Fig. 4-11), and the a2-antagonist, idazoxan (Fig. 4-

1), at a dose that has been observed to increase locus coeruleus firing
(Goldstein et al., 1983) and NE release (Scheinin & Virtanen, 1986). In
contrast, treatments that inhibit NE release (DSP-4, clonidine; Figs. 4-1, 4-2, 4-
3, 4-9), or block al -receptors (prazosin; Fig. 4-10) increased this exploratory

activity, and diminished or blocked the restraint-induced decrease in this
behavior. The ability of phenylephrine to mimic the restraint-induced change in
behavior and that of prazosin to prevent this change suggests that this effect of
restraint involves an activation of ai-receptors. These results are consistent

with work suggesting that noradrenergic systems mediate behavioral
responding associated with anxiety in a variety of experimental paradigms
(Davis et al., 1977; Redmond & Huang, 1979; Charney et al., 1983; Handley &
Mithani, 1984)
Untreated animals exposed to the MCC displayed an increased release
of NE in a number of brain regions as measured by the accumulation of the NE
catabolite, MHPG, and the MHPG:NE ratio. This observation, together with the
conclusion that an activation of NE release results in decreased stimulus-
contact times would explain why treatments that inhibit noradrenergic activity








increased this behavioral response in unrestrained animals. In addition to an
increased release of brain NE, the concentrations of plasma corticosterone and
the catabolites of dopamine (DOPAC, HVA) and serotonin (5-HIAA) or the ratios
of the catabolite to the parent amine were increased in a number of brain
regions in animals exposed to the MCC. Because an increased release of both
corticosterone and brain monoamines are typically associated with stress,
exposure to the MCC can be considered as a stressor (Dunn & Kramarcy,
1984).
Phenylephrine did not affect stimulus-interaction when administered

peripherally (200 ig/kg) but did significantly decrease the mean time per contact
when injected directly into the brain at a dose of 50 ng (approximately 1.5
jg/kg). Thus, if the behavioral action of phenylephrine involves an action of

peripheral systems, this action is centrally mediated. Therefore, although most
of the treatments used in the above series of experiments directly affect
peripheral noradrenergic systems, such an action cannot be the sole
mechanism by which these treatments affected stimulus-directed behavior.
Further, the increase in stimulus-contact time produced by clonidine and
prazosin is unlikely to result from any sedative properties of these two drugs
because this effect on stimulus-directed behavior was observed at doses that
did not inhibit locomotor activity.
To summarize, these results are consistent with the hypothesis that the
restraint-induced decrease in exploratory behavior involves the activation of
noradrenergic systems in the brain. Arnsten et al., (1981) examined the effect of
6-OHDA lesions of the noradrenergic bundle on exploratory behavior of rats in
the MCC. They observed a significant decrease in the mean time per contact in
lesioned animals, an effect opposite to that we observed using DSP-4.

However, these rats were tested at least 14 days after the 6-OHDA injection, at








a time when compensatory responses within the noradrenergic system are
known to occur (U'Prichard et al., 1980; Harik et al., 1981; Hallman & Jonsson,

1984; Logue et al., 1985). When the 6-OHDA-treated rats were tested 2-4 days

post-lesion in the MCC, an increased mean time per contact was observed.
Although this increase in the mean time per contact was not statistically
significant, it was similar in magnitude to that we observed in DSP-4-treated
mice. Further, when we observed mice 10 days after DSP-4 treatment, they

displayed a significantly decreased mean time per contact (1.5 0.2 vs 2.3
0.2 sec), i.e. opposite to the effect observed 3 days after DSP-4 treatment and
consistent with that observed by Arnsten et al. Similarly, rats injected ICV with
6-OHDA demonstrated a pattern of responding at 2 days following the lesion
that was opposite to that observed at 14 days following the lesion, when tested
in an open field (Diaz et al., 1978). Thus, we believe that the effects observed
14 days following 6-OHDA treatment reflects a lesion-induced
hyperresponsivity of noradrenergic systems. This hypothesis is consistent with
the ability of noradrenergic-specific lesions to enhance the behavioral response
to noradrenergic agonists as described above.













CHAPTER V
NORADRENERGIC-CRF INTERACTIONS IN REGULATING EXPLORATORY
BEHAVIOR


The previous two chapters have presented evidence suggesting that

cerebral CRF and noradrenergic systems participate in regulating the restraint-
induced decrease in exploratory behavior observed in the MCC. A question
unanswered by these studies is whether these two neurochemical systems act
independently as parallel systems to influence behavior or whether they
interact. CRF could stimulate NE release or vice versa. Both types of
interactions have been observed within the brain. For example, centrally
administered CRF has been demonstrated to increase the concentration of
MHPG in various brain regions (Dunn & Berridge, 1987), and to cause a dose-
dependent increase in LC spontaneous discharge rates when injected either
locally or ICV in anesthetized and unanesthetized rats (Valentino et al., 1983;
Valentino & Foote, 1988). In contrast to this stimulatory action of CRF on NE
release, NE has been demonstrated to stimulate the release of CRF into the
portal blood supply in both rats and man (Plotsky, 1987; Szafarczyk et al., 1987;
AI-Damluji et al., 1987). Therefore, the following series of experiments was
designed to determine whether either type of interactions might occur in the
regulation of exploratory behavior in the MCC.








Methods

Materials
Clonidine, phenylephrine, prazosin, DSP-4, and CRF were obtained and
prepared as described in Chapter 2. ahCRF was a generous gift from Dr. Jean

Rivier and was dissolved in water and brought to 50% normal saline

concentration by the addition of the appropriate volume of 25-fold concentrated
normal saline.

Procedures
Male mice, obtained and housed under conditions as described in

Chapter 2 were used in all experiments. Mice were implanted with cannulae for
injection into the lateral ventricles as described in Chapter 2. Cannulated
animals were housed singly following surgery. Behavioral testing was
conducted between 08:30 and 16:00.
All injections were administered in the colony room. All drugs were
injected IP except CRF, ahCRF and phenylephrine which were administered
ICV. DSP-4 (50 mg/kg) was injected three days prior to testing. Animals with
cortical concentrations of NE greater than 50% of the mean of the saline-
controls were excluded from further analysis.

Results

The effect of DSP-4 on the CRF-induced decrease in exploratory behavior.
Mice that had been injected with either saline or DSP-4 3 days

previously were injected ICV with either saline or 20 ng CRF 10 min prior to

testing. DSP-4 significantly decreased cortical NE by approximately 75%
compared to quiet controls. Significant main effects of CRF (F1i,24 = 52.3, P <








0.001) and DSP-4 (F1,24 = 5.8, P < 0.03) and a significant interaction between
the two (F1,24 = 10.6, P = 0.003) were observed for the mean time per contact
(Fig. 5-1). DSP-4 increased the mean time per contact with the stimuli in saline-
injected animals. CRF significantly decreased the mean time per contact in

animals injected with either saline or DSP-4. There was no observable effect of
DSP-4 in CRF-injected animals on this response, explaining the significant
interaction term. Neither CRF nor DSP-4 significantly affected the number of
rears (Fig. 5-1). In this experiment, DSP-4 significantly decreased the number
of compartment entries made in both the saline- and CRF-injected groups (Fig.
5-1) an effect not observed previously (see Fig 4-2).

The effect of clonidine on the CRF-induced decrease in exploratory behavior.
Mice were tested 5-7 days following cannulation. On the day of testing,
the animals were injected IP with either saline or 25 pg/kg clonidine 5 min prior
to an ICV injection of either saline or 20 ng CRF. The mice were tested 10 min
following the ICV injection. A significant main effect of CRF on the mean time
per contact (F1,23 = 15.5, P < 0.001) was observed, with significant decreases in
both the saline- and clonidine-treated animals (Table 5-1). In contrast to the
effect of clonidine in uncannulated mice, in cannulated animals, clonidine did
not significantly affect the mean time per contact in saline or CRF-injected mice.
There were no significant effects of CRF or clonidine on compartment entries,
rears, or the number of contacts with the stimuli (Table 5-1).

The effect of prazosin on the CRF-induced decrease in exploratory behavior.
To determine whether the CRF-induced decrease in stimulus contact
time could be reversed by blocking al-receptors, mice were injected with saline

or 200 jg/kg prazosin 5 min prior to an ICV injection of either saline or 20 ng











MEAN TIME PER CONTACT (sec)


3.0 F


Saline CRF DSP-4


CRF
+
DSP-4


Saline CRF DSP-4 CRF
+
DSP-4


Fig. 5-1. Effect of DSP-4 and CRF on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were
injected with either saline or DSP-4 3 days before testing. On the day of testing,
saline- and DSP-4-injected mice were injected ICV with either saline or 20 ng
CRF 10 min prior to testing.
**Significantly different from saline-controls (P < 0.01).


2.01


1.0 I










Table 5-1. Effect of CRF and clonidine on behavior in the MCC


Mean Time Per Rears Compartment
Contact Entries


Saline-Saline 1.82 0.16 268 28 437 43


Saline-CRF 1.27 0.08* 241 23 359 30


Clonidine-Saline 1.94 0.21 218 29 327 52


Clonidine-CRF 1.31 0.14* 226 31 348 40



Numbers represent the mean time per contact and the mean total number of
rears and compartment entries SEM. Mice were injected with either saline or
clonidine (25 ig/kg) 5 min prior to receiving an ICV injection of either saline or
20 ng CRF. Testing was begun 10 min following the ICV injection.
*Significantly different compared to saline-saline controls (P < 0.05).








CRF. Mice were tested 10 min following the ICV injection. Again, significant
main effects of prazosin (F1,28 = 7.9, P = 0.008) and CRF (F1,28 = 86.3, P <

0.001) and a significant interaction between the two (F1,28 = 11.2, P = 0.002)
were observed for the mean time per contact (Fig. 5-2). Prazosin significantly
increased the mean time per contact in ICV saline-injected mice. CRF
significantly decreased the mean time per contact in both the saline- and
prazosin-injected groups. Prazosin had no effect on the CRF-induced decrease
in the mean time per contact, explaining the significant interaction term. Neither

CRF nor prazosin had any effect on the number of rears, compartment entries or
the number of contacts with the stimuli (Fig. 5-2).

Effect of ahCRF on the phenylephrine-induced decrease in exploratory
behavior.
The above results suggest that a CRF-stimulated release of NE is not
involved in the behavioral effect of CRF in this paradigm. To test whether NE
stimulates CRF release to affect exploratory behavior, we examined the effect of
the CRF-antagonist, ahCRF, on the phenylephrine-induced decrease in the
mean time per contact.
Mice were injected ICV (2 gl) with either saline or 20 gg ahCRF 5 min
prior to a second ICV injection (2 jil1) of either saline or 50 or 100 ng
phenylephrine. The animals were tested 10 min following the second injection.
Phenylephrine significantly decreased the mean time per contact in animals
that were first injected with saline (F2,4o = 6.0, P = 0.005; Fig. 5-3). A main effect
of ahCRF on this behavior was also observed (F1,40 = 17.9, P < 0.001). The
CRF antagonist produced a slight but nonsignificant increase in the mean time
per contact in animals receiving saline on the second injection. This effect of

the antagonist in otherwise untreated animals is similar to that observed











MEAN TIME PER CONTACT (sec)


3.0 -


2.0 1-


1.01"


Saline


CRF Prazosin CRF
+


600
500
400
300
200
100


400
300
200
100


Saline CRF Prazosin CRF
+


Prazosin Prazosin



Fig. 5-2. Effect of prazosin and CRF on the mean time per contact, the total
number of rears and the total number of compartment entries. Mice were
injected ICV with either saline or 20 ng CRF 5 min after receiving an IP injection
of either saline or 200 gg/kg prazosin. Animals were tested 10 min following the
ICV injection.
**Significantly different from saline-controls (P < 0.01).








4.0
MEAN TIME PER CONTACT (sec) 600 COMPARTMENT ENTRIES
500
400
3.0 300:
S 200
100
2.0 -.
REARS
400 I
1.0, 300
200
100
Saline 50 100 Saline 50 100

ICV PHENYLEPHRINE (ng) ICV PHENYLEPHRINE (ng)


Fig 5-3. Effect of ICV phenylephrine and ahCRF on the mean time per contact,
the total number of rears and the total number of compartment entries. One half
of the mice were injected ICV with saline (open bars) 5 min after receiving an
ICV injection of either saline, 50, or 100 ng phenylephrine. The remaining
animals were injected ICV with 20 pIg ahCRF (shaded bars) 5 min after receiving
an ICV injection of either saline, or 50, or 100 ng phenylephrine. Animals were
tested 10 min following the second injection.
*Significantly different from saline-controls (P < 0.05, **P < 0.01).








previously (see Chapter 3, Fig. 3-7). ahCRF completely blocked the
phenylephrine-induced decrease in the mean time per contact. Neither
phenylephrine nor ahCRF had any significant effects on the number of rears or
compartment entries (Fig. 5-3).

Discussion

As observed previously, CRF and phenylephrine decreased and
prazosin and DSP-4 increased exploratory behavior in the MCC. Neither
prazosin nor DSP-4 altered the CRF-induced decrease in exploratory behavior,
in contrast to their effect on the restraint-induced decrease in this behavior.
Lastly, a CRF-antagonist reversed the phenylephrine-induced decrease in
stimulus-contact time. The simplest explanation of these results is that CRF and
noradrenergic systems interact, such that NE stimulates the release of CRF to
influence behavior in this novel environment. If these two systems functioned

as parallel and independent systems, then the combined treatment with CRF
and prazosin (or DSP-4) should result in a response intermediate between that
observed with either treatment alone. Therefore, the fact that there was no
additivity between the behavioral effect of either prazosin or DSP-4 and that of
CRF suggests that CRF and NE do not act as two parallel systems.
Clonidine had no effect on stimulus-directed behavior in cannulated
mice, in contrast to its effects in uncannulated mice. In follow-up studies of this
phenomenon, clonidine had no effect on stimulus-directed behavior in animals
tested up to 14 days following cannulation. This effect of the surgery did not
simply reflect a shift in the sensitivity of the animals to the behavioral effect of
clonidine because neither higher (50-100 gg/kg) nor lower (10 jg/kg) doses of
the drug had any effect on stimulus-contact times. Clonidine increased








stimulus-contact times in animals treated with anesthetic, but not cannulated,
just as observed in otherwise untreated animals.
It has been documented that chronic stress induces a number of changes
within noradrenergic systems. Thus, chronic footshock decreased the

sensitivity of NE-stimulated accumulation of cAMP (Stone, 1981). This effect of
footshock might be related to the down-regulation of P-receptors observed in

cortex and cerebellum following repeated sessions of immobilization
(U'Prichard & Kvetnansky, 1980, Stone & Platt, 1982). Repeated restraint-
stress also decreased the a-potentiation of P-stimulated cAMP accumulation.
This decreased a-potentiation appears to involve ai- and not a2-receptors

(Stone, 1987). An increase in tyrosine hydroxylase activity (Stone & McCarty,

1982) and an increase in NE synthesis (Dunn et al., 1981) was also observed
following repeated stress (footshock), whereas the release of NE in response to
a subsequent footshock was lower than observed in animals receiving acute
footshock (Stone & McCarty, 1982). Thus, the lack of activity on exploratory
behavior in cannulated animals might be related to changes within
noradrenergic systems, stimulated by the stress of the surgery.
However, the simple explanation that the stress of the surgery evoked a
desensitization of presynaptic a2-receptors is unlikely because MHPG
concentrations in various brain regions, suggested that clonidine inhibited NE
release in both cannulated and uncannulated animals tested in the MCC.

Obviously, the change in receptor function could be limited to a small region of
the brain or a subset of neurons within a region that is critical in the regulation of
the behavior, and that would not be detected by the relatively large tissue
sections we collected. A second possibility is that clonidine acts to regulate
stimulus-directed behavior through postsynaptic a2-receptors. An involvement
of postsynaptic a2-receptors in regulating responding in a delayed-response








task in primates, a stressor-sensitive response, has been suggested (Arnsten &
Goldman-Rakic, 1985). The ability of DSP-4, prazosin and phenylephrine to

alter stimulus-directed behavior in cannulated animals indicates that NE
continues to regulate this behavior in these animals and that ai-receptors are
involved. If postsynaptic a2-receptors are involved, they function in an
opposing manner to that of the al -receptors, because stimulating a2-receptors

(clonidine) increased this behavior. However, if this is the case, the function of
these a2-receptors is unclear, because the effects of restraint on exploratory

behavior in the MCC in cannulated and uncannulated animals is similar.
Therefore, for the moment, the explanation for the loss of the behavioral activity
of clonidine remains unclear.
Our results suggesting that noradrenergic neurons stimulate CRF release
to influence behavior is opposite to what would be predicted based on the

ability of CRF to increase locus coeruleus firing (Valentino et al., 1983) and NE
release in the brain (Dunn & Berridge, 1987, Butler et al., 1988). However,

these stimulatory effects of CRF on noradrenergic activity were only observed at
doses significantly higher than those needed to elicit a behavioral effect in the
MCC. At these higher doses, mice tested in the MCC show a very different

response profile including a pronounced hypoactivity and associated postural

changes. Thus, although both types of interactions may occur (i.e., NE-
mediated release of CRF and vice versa), they may regulate different
responses.
The ability of ICV ahCRF to block the nitroprusside-induced increase in
LC firing suggests that endogenous CRF does have a stimulatory effect on

noradrenergic function in stress (Valentino, personal communication).

Therefore, it is possible that in stress, certain behaviors are regulated through a
CRF-mediated activation of cerebral noradrenergic systems. Consistent with






88

this hypothesis, is the ability of the noradrenergic 3-antagonist, propranolol, to

block the fear-enhancing effects of CRF in rats tested in a conditioned emotional
response paradigm (CER; Cole & Koob, 1988).
The stimulatory effect of NE on CRF release suggested by the present
studies is similar to that reported in the regulation of CRF release into the portal
blood supply in both the rat (Plotsky, 1987; Szafarczyk et al., 1987) and man
(AI-Damluji & Rees, 1987). However, we do not know whether the presumed
release of CRF during stress that alters behavior in the MCC derives from the
hypothalamus or from other sites within the brain.













CHAPTER VI
GENERAL DISCUSSION


The central mechanisms regulating behavioral and physiological
responding in stress are poorly understood. It has been proposed that both
CRF and NE have significant roles in regulating behavioral responding
associated with stress. The studies described in the previous chapters tested

the possible involvement of brain NE and CRF in regulating the stress-induced
decrease in exploratory behavior of a novel environment. It is concluded that
an increase in the release of both CRF and NE are involved in mediating the
restraint-induced decrease in exploratory behavior. ICV CRF elicited a dose-
dependent, stress-like decrease in the mean time per contact with a minimum

effective dose of 5 ng. This effect was observed in the absence of changes in
locomotor activity. Further, the CRF-antagonist, ahCRF, blocked the restraint-
induced decrease in the mean time per contact, suggesting a role of
endogenous CRF in this behavioral effect of restraint. These results are
consistent with previous studies suggesting that endogenous CRF mediates
behavioral and physiological responding in stress (Sutton et al., 1982; Brown et
al., 1982; Koob & Bloom, 1985; Rivier et al., 1986; Dunn & Berridge, 1987).
The ai-selective agonist, phenylephrine, injected ICV and the
a2-antagonist, idazoxan at a dose that increases NE release, also decreased

exploratory behavior. Inhibition of NE release using DSP-4 or clonidine, or
blockade of ali-receptors (prazosin) increased this exploratory response, and

diminished or prevented the effect of restraint on this behavior. Thus, these








results support the hypothesis that an increased release of NE modulates
behavioral responding in stress and suggests the involvement of a,-receptors

in this action of NE. In contrast, neither prazosin nor DSP-4 altered the CRF-
induced change in this behavior. However, a CRF-antagonist reversed the

phenylephrine-induced decrease in stimulus-contact time. Thus, it is concluded

that these two neurochemical systems interact, such that NE stimulates the

release of CRF to influence behavior in this novel environment.
The opiate antagonist, naloxone, prevented the effect of CRF and
restraint on exploratory behavior. Therefore, one possible mechanism is that

during stress, the NE-stimulated release of CRF activates endogenous opioid

systems, which in turn affect exploratory behavior. However, it is unclear
whether CRF and the opioids act in such a tandem arrangement or whether the
opioids modulate the sensitivity of this exploratory response to the effects of
stress and CRF.
During stress, an increase in the release of cerebral NE, DA, and 5-HT
and an increase in the concentration of plasma corticosterone are observed.
Exposure to the MCC also increased the release of the biogenic amines in
various brain regions, as determined by the increased accumulation of their

catabolites, and increased the concentration of plasma corticosterone. Thus, as
determined by these physiological responses, exposure to the MCC can be
considered a stressor.
The involvement of NE in mediating stress-induced behavior in this

paradigm is consistent with the reported involvement of NE in regulating
emotional and behavioral responding associated with anxiety in humans and

non-human primates. Thus, monkeys in which the LC was electrically
stimulated, or that received the a2-antagonists, piperoxan or yohimbine which

act to increase NE release, displayed fear-associated behaviors (Redmond &








Huang, 1979). Further, separation-induced vocalization in monkeys was
decreased by clonidine and increased by yohimbine (Harris & Newman, 1987).
Clinical studies have provided similar results; yohimbine increased the
subjective rating of anxiety (Charney et al., 1983, 1984), whereas clonidine
antagonized this action of yohimbine and appeared to have anxiolytic
properties in anxious subjects (Hoen-Saric et al., 1981). In rats, clonidine
decreased the potentiated startle response (Davis et al., 1977) and increased
punished responding in the Geller-Seifter conflict test (Kruse et al., 1981).
Because anxiolytics affect responding in these two tests in the same way as
clonidine, these results are consistent with clonidine having anxiolytic
properties. The activation of noradrenergic systems by idazoxan or yohimbine
decreased the time rats spent exploring the open arms of the X-maze (Handley
& Mithani, 1984); an effect opposite to that of the benzodiazepines (Pellow &
File, 1986). It remains to be determined whether the behavioral effects
associated with an increase in noradrenergic activity observed in these various
paradigms involves the release of CRF.
Although the activation of noradrenergic systems appears to be involved
in mediating the stress- or anxiety-like effects in these paradigms, it is unclear
what the precise role of NE is in anxiety or stress. For example, NE could be
involved in regulating an affective or emotional component of anxiety with the
behavioral effects of activating noradrenergic systems resulting from the animal
experiencing a greater sensation of anxiety or fear. Alternatively, noradrenergic
systems could regulate a general component of behavioral responding that is
also affected in anxiety or stress, such as vigilance or arousal. In this case, the
behavioral effects of the activation of noradrenergic systems and that of the
benzodiazepines would be related to changes in the vigilance or arousal state
of the animal. Clearly, a better understanding of the behavioral responses








studied in the different experimental paradigms is needed to understand the
neurochemistry of stress and anxiety.
The stimulatory effect of NE on CRF release suggested by the present
studies is consistent with the known mechanisms regulating CRF release into
the portal blood supply in both the rat (Plotsky, 1987; Szafarczyck et al., 1987)
and man (AI-Dumaluji and Rees, 1987). Thus, noradrenergic fibers innervating
the CRF-containing region of the PVN have been described, and knife cuts
interrupting these fibers increased the density of immunostaining for CRF
(Swanson et al., 1986). Further, stimulation of the ventral noradrenergic bundle
(Plotsky, 1987) or injection of NE into the lateral ventricles or directly into the
PVN stimulated the release of CRF into the portal blood supply (Plotsky, 1987;
Szafarczyck et al., 1987). The stimulatory effect of these treatments on CRF
release was prevented by the ail-antagonists prazosin (Szafarczyck et al.,
1987) and coryanthine, but not the P-antagonist, propranolol (Plotsky, 1987).

Thus, the pharmacology of this system resembles that involved in stress-related
changes in exploratory behavior.
Therefore, it is possible that collaterals from CRF-containing cells within
the PVN are involved in regulating this behavioral response. In support of such
an hypothesis, the PVN has been shown to be a critical site in regulating the
CRF-induced decrease in feeding behavior and injection of CRF into the PVN
increased grooming and locomotor activity (Krahn et al., 1988). Alternatively,
noradrenergic fibers from the LC that project to other brain structures could
regulate the release of CRF within these structures. For example, certain
telencephalic structures that receive a particularly dense innervation from the
LC (Nieuwenhuys, 1985) also contain high densities of CRF-containing cells
(e.g., the central nucleus of the amygdala, the bed nucleus of the stria
terminalis, and portions of the neocortex (Merchenthaler, 1984; Sakanaka et al.,








1987)). Thus, NE may control the release of CRF in a region separate from the
PVN through a mechanism similar to that observed in the PVN. The fact that

DSP-4, which primarily affects fibers of the DNB and thus has the greatest

observable effect on cortical and hippocampal NE-containing cells, affects this
behavior, might suggest the involvement of CRF in these regions. However,
because the LC contributes to the NE innervation of the PVN and other non-
telencephalic structures, the critical structure within the brain at which CRF acts
need not be limited to the telencephalon. Immunohistochemical and

biochemical studies will be of importance in determining whether a
noradrenergic regulation of CRF release is likely to occur in these brain regions,
and where CRF might act to affect exploratory behavior.
Although a number of the behavioral effects of CRF suggest an

involvement of endogenous CRF in stress, the functional significance of these
responses, and whether they serve similar functions in stress remains unclear.
CRF could act on a single system regulating a general component of behavior,
such as vigilance or motivation, that affects different behaviors depending on
the environmental cues present. Alternatively, CRF could act at multiple sites
within the brain responsible for regulating individual behaviors. Based on the
sensitivity of some of the CRF-induced behaviors to opiate and noradrenergic
antagonists, the former hypothesis is unlikely to explain all the behavioral
effects of CRF. For example, NE appears to stimulate the release of CRF to
affect exploratory behavior in the MCC, whereas, in a conditioned emotional
response paradigm, CRF appears to influence behavior by stimulating the
release of NE (Cole & Koob, 1988). Thus, if CRF influences behavioral

responding by affecting a general motivational or arousal state, it does so

through two different neurochemical mechanisms in these two paradigms.
However, both of these mechanisms are consistent with the proposed role of