Ž .Brain Research 784 1998 37–47

Research report

Production of the Fos protein after contextual fear conditioning ofC57BLr6N mice

Snezana Milanovic a,1,b, Jelena Radulovic a,1, Olgica Laban a, Oliver Stiedl a, Fritz Henn b,Joachim Spiess a,)

a Department of Molecular Neuroendocrinology, Max Planck Institute for Experimental Medicine, Hermann Rein Str. 3, 35075 Goettingen, Germanyb Central Institute for Mental Health, J5, 68159 Mannheim, Germany

Accepted 7 October 1997

Abstract

Male C57BLr6N mice were chosen to determine Fos production during acquisition of context-dependent fear and after re-exposure tothe conditioning context. Fear-conditioning was induced by a single exposure of mice to a context followed by an electric shock. Control

Ž .groups consisted of mice exposed to context only Context group or to an immediate electric shock. When contextual retention wasŽ .measured 24 h after conditioning retention test 1 , significant contextual generalization was observed. However, when animals were

Ž .exposed to a different context from days 2–5 after conditioning and then tested for retention on day 6 retention test 2 , generalizationwas markedly reduced. After the training, the fear-conditioned mice produced higher Fos levels than mice exposed to an immediate shockin the hippocampus, medial amygdaloid nucleus and parietal somatosensory cortex. Both shock groups produced significantly more Fosthan the Context group in the central nucleus of the amygdala. After retention test 1, fear-conditioned mice generated more Fos in thehippocampus and central amygdaloid nucleus than the two control groups. However, all groups exhibited similarly low Fos productionafter retention test 2. The results demonstrated that simultaneous Fos production in the hippocampus, central and medial nuclei ofamygdala and somatosensory parietal cortex closely paralleled the ability of mice to acquire conditioned fear. In contrast, Fos productionafter the retention tests did not correlate with the expression of conditioned fear. q 1998 Elsevier Science B.V.

Keywords: Fear conditioning; Fos production; Stress; Hippocampus; Amygdala; Parietal cortex

1. Introduction

Production of the Fos protein or formation of c-fosmRNA have been extensively investigated with a varietyof experimental manipulations, particularly in models ofstress and learning. The rapid and transient Fos productionw x34,18 evoked by stressful stimuli was mainly investigatedin order to map the brain structures that respond withincreased physiological activity to different stressorsw x8,10,19,39 . Heterodimers formed by Fos and Jun proteinswere found to act as transcriptional regulators and to thusmediate long term neuronal changes in response to envi-

w xronmental stimuli 28 . Therefore, learning studies used the

) Corresponding author. Fax: q49-551-3899-359; E-mail:spiess@mail.mpiem.gwdg.de

1 S.M. and J.R. share first authorship by virtue of their unique contri-butions to this work.

expression of Fos and other immediate early genes asindicators closely linked to learning and memory.

Employment of operant conditioning proceduresdemonstrated that Fos production was increased in definedbrain regions during the acquisition phase as well asfollowing the retention tests. Fos production was elevatedin the intermediate hyperstriatum ventrale of chicks during

w xacquisition of passive avoidance 1 and administration ofFos antisense mRNA completely blocked passive avoid-

w xance learning 25 . Elevated Fos levels were also observedin the motor cortex of rats during acquisition of motor skillperformance, but not in control groups subjected to motor

w xactivity 21 . During the acquisition of a brightness dis-crimination, however, both trained and pseudotrained ratsshowed increased accumulation of Fos mRNA in the hip-

w xpocampus 37 , although protein synthesis was signifi-w xcantly higher only in trained rats 24 .

Other studies suggested that Fos production may beinduced by the retention tests, following memory retrieval.In a study employing the Skinner box escape task, rats thatperformed well during the retention test, in contrast to the

0006-8993r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.Ž .PII S0006-8993 97 01266-3

( )S. MilanoÕic et al.rBrain Research 784 1998 37–4738

yoked ones, showed specific Fos production in the fore-w xlimb motor cortex 7 .

The role of Fos production in experimental modelsemploying classical fear conditioning paradigms has notbeen clearly established. Fos production during acquisitionof a conditioned fear response has not been investigated sofar. In a study focused on the retention phase of thecontextual fear conditioning paradigm, Beck and Fibigerw x2 demonstrated elevated Fos-like immunoreactivity innumerous brain nuclei after the retention test. It wassuggested that some of these changes were linked to theconditional response. Other studies also reported an in-creased Fos level in several brain regions, mainly in theamygdala, of rats exposed to conditioned aversive stimuli,however, similar or even higher Fos production was ob-served in rats exposed to aversive stimuli that induced

w xunconditional fear responses 6,31,36 . Thus far, in allstudies investigating the relationship between conditionedfear and Fos production, multiple-trial multiple shockswere employed for conditioning.

It has been reported that contextual fear conditioningw xcan be induced in rats 3,12 by a one trial conditioning

procedure employing a single context presentation fol-lowed by a single electric footshock. This conditioningprocedure may be advantageous for studying the molecularchanges induced by stress and learning, due to betterresolution of molecular changes induced by individualconditioned and unconditioned stimuli. Additionally, theacquisition of fear response can be efficiently controlledby the use of animals exposed to an immediate shock uponplacement into a chamber. These animals do not exhibitunconditional or conditional fear and thus suffer from an

w ximmediate shock freezing deficit 4,12 .The objective of this study was to determine Fos pro-

duction in the brain of mice subjected to one trial contex-tual fear conditioning. Fos production in the hippocampus,amygdala and cortical areas were determined after thetraining and after the retention tests of fear-conditionedmice, mice exposed to an immediate shock and miceexposed to the context only.

2. Materials and methods

2.1. Housing conditions

Inbred C57BLr6N male mice, 8 weeks old, obtainedŽ .from Charles River Germany were used in the experi-

ments. Mice were individually housed under conventionalŽconditions room temperature 22"18C; humidity 55"

.10%; 12 h lightrdark cycle in macrolon cages. Beddingwas changed once a week. All procedures followed therecommendations of the Society for Laboratory Animal

Ž .Science Germany . Standard pelleted diet and water wereoffered ad lib. Animals were kept under constant housingconditions 5–7 days before experiments.

2.2. Apparatus

A computerized fear conditioning system was devel-Ž .oped by the TSE Bad Homburg, Germany . The condi-

Ž . Žtioning context context 1 consisted of a box 58=30=27.cm with a gray interior, a 12 V light at the ceiling and a

Ž .plexiglas chamber 35=20=20 cm , provided by a fanfor ventilation. The chamber was placed at a distinctposition on a removable shock grid made of stainless steel

Ž .rods f 4 mm, spaced 0.9 cm apart . The shock grid wasconnected to a shockerrscrambler unit delivering shocksof defined duration and intensity. A mouse was brought

Ž .into the chamber through a circular opening f 15 cm onthe ceiling. The shock grid, the floor below the shock gridand the chamber were cleaned with 70% ethanol beforeeach mouse entered the chamber. The different contextŽ . Žcontext 2 consisted of a Plexiglas chamber 35=20=20

.cm which was exposed to 20 V light, did not have a rodfloor and was divided in half by inserting a clear plasticpanel connecting the diagonal corners. The chamber waswashed with 1% acetic acid, before each individual mouseentered the chamber.

Activity was measured by an infrared beam systemŽ .detection rate 10 Hz , controlled by the fear conditioningsystem and completely synchronized to the sequence ofstimuli. The infrared beams were spaced 1.3 cm in widthand 2.5 cm in depth. Freezing, defined as the lack ofmovement besides respiration and heart beat, was assessedin 10 s intervals by an observer unaware of the experimen-tal conditions.

2.3. Fear conditioning

The first set of experiments was carried out to establishthe minimal current intensity that was sufficient to inducecontext-dependent fear conditioning in C57BLr6N mice.Animals were placed in the chamber of the apparatus and

Ž .at the end of a 180 s-period, electric shock 2 s wasŽ .applied Context-ES groups . The intensity of the constant

Ž Ž . .electric current was 0.2 Context-ES 0.2 mA group , 0.7Ž Ž . . Ž Ž .Context-ES 0.7 mA group , 1.3 Context-ES 1.3 mA

. Ž Ž . .group , or 2.0 Context-ES 2.0 mA group mA. After theend of this training, the animals were removed from thechamber and returned into their home cages. A controlgroup of animals was subjected to the same procedureexcept that the exposure to the electric current was omittedŽ .Context group . Each group consisted of eight to tenanimals. All procedures were done during the light periodof the dayrnight cycle and animals from each group weretrained in a random order, but each animal was trained andtested at the same time of the day. Contextual retentionwas tested 24 h after removal from the conditioning cham-ber. To this end, the mice were re-exposed for 182 s to thesame conditioning chamber, but without footshock. Freez-ing and locomotor activity were evaluated as retentionparameters.

( )S. MilanoÕic et al.rBrain Research 784 1998 37–47 39

Table 1Mean activity and percentage of freezing in mice exposed to the conditioning context before ES, during ES and during the retention test

Ž . Ž .Group n Mean activity cmrs Freezing %

Before ES During ES Retention test Before ES During ES Retention test

Context 9 6.25"0.39 5.13"3.90 3.65"0.41 0.0"0.0 0.0"0.0 10.5"5.5)Ž .Context-ES 0.2 mA 10 6.42"0.39 9.75"3.43 3.05"0.39 0.5"0.8 0.0"0.0 43.3"5.2

) ) ,b ) ) ,a ) ) ,aŽ .Context-ES 0.7 mA 9 5.84"0.39 27.52"3.43 1.11"0.41 1.9"0.8 0.0"0.0 83.9"5.5) ) ,c ) ) ) ) ,bŽ .Context-ES 1.3 mA 8 6.38"0.41 32.04"3.64 1.62"0.43 0.7"0.8 0.0"0.0 84.0"5.9) ) ,c ) ) ) ) ,cŽ .Context-ES 2 mA 10 6.36"0.37 40.95"3.25 1.36"0.39 0.0"0.74 0.0"0.0 88.9"5.2

Statistically significant differences: ) p-0.01, ) ) p-0.001 vs. 0 mA; a p-0.05 bp-0.01 c p-0.001 vs. 0.2 mA.

In the second set of experiments, the specificity ofinduction and expression of conditioned fear were investi-gated. Electric shock of 0.7 mA current intensity was

Ž Ž .employed for fear conditioning Context-ES 0.7 mA.group . Control groups consisted of mice exposed to the

Ž .context only for 182 s Context group and mice thatŽ .received an immediate 2 s, 0.7 mA footshock and then

Ž Žcontinued to be exposed to the context for 180 s ES 0.7. .mA -Context group . The retention test was performed 24

Žh later, by exposure of the mice of each group ns8–. Ž .10rgroup to contexts 1 and 2 retention test 1 . The

freezing and mean activity were recorded as described.In the third set of experiments, separate groups of mice

were subjected to a procedure similar to the extinctionw xparadigm described by Bouton and Bolles 5 , in order to

reduce generalization without affecting the conditionedfear response. Groups of mice were exposed to the same

Žtraining conditions as described above Context group,Ž . Žns14; Context-ES 0.7 mA group, ns18; and ES 0.7

. .mA -Context group, ns14 and then exposed to context 2Ž .for 4 consecutive days days 2–5, 3 minrday . Freezing

and locomotor activity were recorded daily. On day 6,each experimental group was subdivided in two additionalgroups consisting of mice that were exposed to context 2,

Ž .or re-exposed to context 1 retention test 2 .

2.4. Fos production

ŽSeparate mice were subjected to the training Context,Ž . Ž . .Context-ES 0.7 mA and ES 0.7 mA -Context groups

trials described above and sacrificed 60 min after theŽ .training ns6rgroup . The mice employed in the behav-

Ž .ioral experiments ns6rgroup were sacrificed 60 minafter the retention test 1 and 2, respectively. Their brainswere processed for immunohistochemical staining withanti-Fos antibodies and the number of Fos-positive nucleiwas determined.

2.5. Immunohistochemical analysis

Mice were anesthetized with ketaminerxylasineŽ . Ž .130r13 mgrg b.wt. in saline 0.01 mlrg, i.p. and then

Žtranscardially perfused with ice cold phosphate buffer 0.1.M, pH 7.4, 150 ml , followed by 4% paraformaldehyde in

Ž .Fig. 1. Freezing and activity in mice of the Context, Context-ES 0.7 mAŽ .and ES 0.7 mA -Context groups in the conditioning and different

contexts during the retention tests 1 and 2. The number of mice per groupwas 7–10. Statistically significant differences: ) p-0.001 vs. Context

a Ž .group; p-0.001 vs. ES 0.7 mA -Context group.

( )S. MilanoÕic et al.rBrain Research 784 1998 37–4740

Ž .phosphate buffer pH 7.4, 150 ml . Brains were postfixedfor 48 h in the same fixative and then immersed for 24 hŽ .each time in 10%, 20% and 30% sucrose in phosphatebuffered saline. After the tissue was frozen by liquidnitrogen, 50 mm thick coronal sections were cut on thecryostat. Every fifth section was collected, starting fromthe fimbriae hippocampi to the end of the hippocampus.

Ž .Twelve well plates Corning were used for performingimmunocytochemical staining of free-floating sections. Tostandardize the binding conditions, a set of sections wastaken from the brain of a naive control mouse anesthetizedas described above and sacrificed immediately after re-moval from the home cage. These sections were run oneach plate. Elimination of endogenous peroxidase activity

was accomplished using 1% H O in methanol for 15 min,2 2

followed by rinses in 0.01 M sodium phosphate-bufferŽ .PBS, pH 7.4 containing 0.2% Triton-X 100. Five percentgoat serum and 0.3% Triton-X 100 in 0.01 M PBS wereused for preincubation. Sections were then incubated for

Ž48 h at q48C with rabbit anti-Fos antibody Oncogene.Science, 1:20,000 dilution . Specificity of immunostaining

was confirmed on sections that were incubated with Fosantibody preabsorbed overnight at q48C with appropriatesynthetic antigenic peptide in tenfold excess over the

Ž .amount of antibody Oncogene Science . Subsequently, thesections were washed and incubated at room temperaturewith biotinylated goat anti-rabbit antibody followed by the

Ž .ABC complex Vector ABC kit . For visualization, DAB

ŽFig. 2. Number of Fos-positive nuclei in the CA1 region of the hippocampus and in different amygdaloid nuclei of mice of the context, Context-ES 0.7. Ž .mA and ES 0.7 mA -Context groups 60 min after the training session or after the retention test. The number of mice per group was 6. Statistically

Ž .) a Ž .significant differences p-0.05 or greater vs. Context group vs. ES 0.7 mA -Context group.

( )S. MilanoÕic et al.rBrain Research 784 1998 37–47 41

Ž .was used as chromogen Sigma fast tablet set . The sec-tions were mounted, dehydrated and coverslipped withEukitt.

2.6. Quantification and statistical analysis

Fos-positive cells were counted with a Macintosh-basedŽ .image analysis system NIH Image , as described by Chen

w xand Herbert 9 . Nuclei were counted individually andexpressed as number of Fos-positive nuclei per 0.1 mm2.The AP coordinates of sections included for detailed analy-sis were AP q0.26, cingulate and frontal cortex; APy1.22, medial amygdaloid nucleus; AP y1.34 central,cortical, basolateral amygdala, CA1 region of the hip-

Ž .pocampus and parietal somatosensory cortex; AP y2.54,

w xtemporal and occipital cortex 15 . The counting was per-formed in an area of the same shape and size for eachbrain region. Statistical analysis of behavioral and im-munohistochemical data was performed by one-wayANOVA followed by Tukey–Kramer’s test for multiplecomparisons. The results are presented as mean"S.E.

3. Results

3.1. Fear conditioning

Prior to electric shock administration, the groups ofmice exposed to the context only or to electric shock ofdifferent current intensities did not differ significantly in

Ž . Ž .Fig. 3. Number of Fos-positive nuclei in different cortical areas of mice of the context, Context-ES 0.7 mA and ES 0.7 mA -Context groups 60 min afterŽ .)the training session or after the retention test. The number of mice per group was 6. Statistically significant differences p-0.05 or greater vs. Context

a Ž .group vs. ES 0.7 mA -Context group.

( )S. MilanoÕic et al.rBrain Research 784 1998 37–4742

Ž .Fig. 4. Baseline Fos staining in coronal sections of the CA1 region of the hippocampus of naive mice and mice of the Context, Context-ES 0.7 mA andŽ .ES 0.7 mA -Context groups 60 min after the training. Scale bars50 mm.

Ž .Fig. 5. Baseline Fos staining in coronal sections of the of the medial nucleus of the amygdala of naive mice and mice of the Context, Context-ES 0.7 mAŽ .and ES 0.7 mA -Context groups 60 min after the training. Scale bars50 mm.

( )S. MilanoÕic et al.rBrain Research 784 1998 37–47 43

Ž .Fig. 6. Baseline Fos staining in coronal sections of the of the parietal somatosensory cortex of naive mice and mice of the Context, Context-ES 0.7 mAŽ .and ES 0.7 mA -Context groups 60 min after the training. Scale bars50 mm.

Ž .Fig. 7. Baseline Fos staining in coronal sections of the central nucleus of the amygdala of naive mice and mice of the Context, Context-ES 0.7 mA andŽ .ES 0.7 mA -Context groups 60 min after the training. Scale bars50 mm.

( )S. MilanoÕic et al.rBrain Research 784 1998 37–4744

Žbasal behavioral recordings locomotor activity and freez-.ing . The mean activity during shock administration was

Ž .significantly higher in the Context-ES 0.7 mA , Context-Ž . Ž .ES 1.3 mA and Context-ES 2.0 mA groups than in the

Ž . Ž .Context and Context-ES 0.2 mA groups Table 1 . Theincrease in behavioral activity during the 2 s exposure tofootshock demonstrated significant group differences, asdemonstrated by ANOVA using current intensity as a

Ž .factor, F 4,45 s18.44, p-0.001. Significant group dif-ferences were also recorded in the percentage of freezing

Ž .during the retention test, F 4,45 s38.78, p-0.001.Ž . ŽFreezing of the Context-ES 0.7 mA , Context-ES 1.3

. Ž .mA and Context-ES 2.0 mA groups was markedly in-Žcreased statistical significance: p-0.05, p-0.01 and

.p-0.001 respectively in comparison to the Context-ESŽ . Ž .0.2 mA group Table 1 . Four out of 10 mice of the

Ž .Context-ES 0.2 mA group exhibited low or no freezingŽ .0–33% , whereas the remaining six animals of the group

Ž .showed high freezing scores 56–78% . Activity scores inthe retention test also indicated significant differencesŽ .F 4,45 s8.64, p-0.001 between groups. Whereas the

Ž . Žactivities of the Context-ES 0.7 mA , Context-ES 1.3. Ž .mA and Context-ES 2.0 mA groups did not differ

significantly from one another, the activity of these groupsŽ .was significantly lower p-0.001 than the activity of the

Context group. Additionally, the activity of the Context-ESŽ . Ž .0.7 mA group was significantly lower p-0.05 than

Ž .the activity of the Context-ES 0.2 mA group.ŽDuring retention test 1, mice of the Context and ES 0.7

.mA -Context groups did not exhibit any fear-related be-havior, as indicated by low freezing and relatively high

Ž .activity scores after re-exposure to context 1 Fig. 1A .Ž .However, mice of the Context-ES 0.7 mA group hadŽ .significantly higher freezing scores, F 2,24 s18.07, p-

Ž .0.001 and reduced mean activity F 2,24 s8.17, p-0.001Žthan mice of the Context p-0.001, freezing; p-0.01,

. Ž . Žactivity or ES 0.7 mA -Context p-0.001, freezing;.p-0.05, activity groups upon re-exposure to context 1.

Additionally, the fear response of mice of the Context-ESŽ .0.7 mA group was significantly generalized to context 2as well.

During retention test 2, preceded by the extinctionŽ .procedure described above, the Context-ES 0.7 mA groupŽ .also froze significantly more in context 1, F 2,20 s17.34,

Ž . Ž .p-0.001, than the Context p-0.001 or ES 0.7 mA -Ž .Context p-0.001 groups. However, at this point, the

Ž .fear response of the Context-ES 0.7 mA group was notaccompanied by generalization, as demonstrated by signifi-

Ž .cant contextual discrimination Fig. 1B .

3.2. Fos production during acquisition of contextual fearand after the retention tests

Mice subjected to different experimental proceduresexhibited Fos production in the amygdaloid nuclei, hip-

Ž .pocampus and different cortical areas Figs. 2 and 3 .

These brain regions did not show Fos staining in naiveŽ .mice Figs. 4–7 . A common feature in all experimental

groups was stronger Fos production after the training thanŽ .after the retention tests Figs. 2 and 3 . Analysis of Fos

production in individual brain areas revealed several differ-ences depending on the training procedure and learning

Ž .phase acquisitionrretention .After training, significant group differences in Fos pro-

duction were determined in the hippocampal CA1 area,central amygdaloid nucleus, medial amygdaloid nucleus

Žand parietal somatosensory cortex. The Context-ES 0.7.mA and Context groups did not differ significantly from

one another in Fos production in all tissues except for theŽ .central amygdaloid nucleus Figs. 2 and 7 . In this nucleus,

Ž .the Context-ES 0.7 mA group was found to produceŽ .significantly more Fos p-0.01 than the Context group.

Ž .In contrast, the ES 0.7 mA -Context group which pro-Ž .duced as much Fos as the Context-ES 0.7 mA group in

Ž .the central amygdaloid nucleus Fig. 2 , generated less FosŽ .than the Context-ES 0.7 mA group in the hippocampus

Ž .p-0.01; Figs. 2 and 4 , medial amygdaloid nucleusŽ .p-0.05; Figs. 2 and 4 and parietal somatosensory

Ž .cortex p-0.01; Figs. 3 and 5 .After retention test 1, the Fos production of the Con-

Ž . Ž .text-ES 0.7 mA group was significantly higher p-0.05in the hippocampus, central amygdaloid nucleus, occipitalcortex and temporal cortex than Fos production of the

Ž .Context group in these brain areas Figs. 2 and 3 . Simi-Ž .larly, the Context-ES 0.7 mA group produced signifi-

Ž .cantly more Fos than the ES 0.7 mA -Context group inŽthe hippocampus and central amygdaloid nucleus p-

. Ž .0.05 Fig. 2 .After retention test 2, Fos production in the brain of all

Ž .groups was low Figs. 2 and 3 and no differences betweenthe experimental groups were observed.

4. Discussion

The behavioral experiments presented here demon-strated that a current of 0.7 mA was necessary and suffi-cient to induce one trial fear conditioning in C57BLr6Nmice. In contrast to fear-conditioned mice, it was observedthat mice exposed to the context only or to an immediateshock, did not freeze during the retention tests. Thisobservation was in full agreement with previous studiesw x4,11 and indicated that under entirely non-associativeconditions, the shock itself did not produce a freezingresponse upon subsequent re-exposure of the mice to thechamber used in the training phase. The finding that miceexposed to context paired with shock exhibited signifi-cantly more freezing than mice exposed to immediateshock or to context only, demonstrated that the freezingbehavior was induced by associative learning and did notrepresent an unconditional or non-associative response to

w xthe stress employed 11–13 .

( )S. MilanoÕic et al.rBrain Research 784 1998 37–47 45

Generalization of the fear response of mice of theŽ .Context-ES 0.7 mA group observed during retention test

1 was also observed in rats subjected to foreground but notw xbackground fear conditioning 4,11,20 . In agreement with

w xprevious findings 5 , application of an extinction proce-dure before retention test 2 significantly enhanced thecontextual discrimination of fear-conditioned mice.

The observation that Fos protein was produced in manybrain regions including the cortical areas, hippocampusand amygdala of the Context group was consistent with

w xprevious mapping studies 16,29 and may indicate a neu-ral response to the multiple sensory input elicited by anarray of novel visual, tactile, olfactory and spatial stimuli

w xof the context 28,23 .The intensity and regional distribution of Fos produc-

tion in most of the brain regions, except for the centralamygdaloid nucleus, did not differ between the Context-ESŽ .0.7 mA and Context groups 60 min after the training.This finding contrasted with the results obtained afteroperant conditioning, that demonstrated increased Fos pro-duction in different brain regions specifically involved in

w xacquisition of a particular task 1,21 . We assume that thisdiscrepancy was primarily due to the different learningprocesses underlying operant and classical conditioning.During operant conditioning, animals learn to perform aparticular task probably requiring activation of additionalneuronal structures whereas classical conditioning requiresonly association between external stimuli. For this associa-tion, convergence of processing pathways of conditionedand unconditioned stimuli within a brain structure appearsto be an essential and sufficient prerequisite for the forma-

w xtion of conditional responses 30 . In this study, the brainregions activated 60 min after the training by shock only,except for the central amygdaloid nucleus, could not bedifferentiated from the regions activated by contextualexposure. Thus, it seems likely that in most of the brainregions Fos production could have been induced by con-vergence of contextual and stressful stimuli.

Ž .The high Fos production of the Context-ES 0.7 mAŽ .and ES 0.7 mA -Context groups but not the Context

group in the central amygdaloid nucleus, as observed 60min after the training, was most probably an effect of theelectric footshock only and may represent an index of theunconditioned fear response. Accordingly, Lamprecht and

w xDudai 22 demonstrated in the model of conditioned tasteaversion that the formation of Fos mRNA in the centralamygdaloid nucleus was increased by the unconditionedstimulus but not by the non-associated conditioned stimu-lus. It is well established that the central nucleus of the

w xamygdala plays a crucial role in acquisition 33 of fearresponses. It appears however, that the high Fos produc-

Žtion in the central amygdaloid nucleus of the ES 0.7.mA -Context group was not sufficient for the acquisition

of conditioned freezing behavior without appropriate in-crease of Fos levels in other brain regions, such as theparietal cortex, hippocampus and medial amygdaloid nu-

Ž .cleus, areas where the ES 0.7 mA -Context group pro-Ž .duced significantly less Fos than the Context-ES 0.7 mA

group. This interpretation is in agreement with previousw xreports on the role of the associative parietal cortex 38 ,

w x w xhippocampus 20,32 and amygdala 17 in the acquisitionof context-dependent fear.

It is thought that fear conditioning occurs when pro-cessed contextual stimuli from the hippocampus arrivesimultaneously with the shock input in the lateral, basolat-

w xeral andror central amygdaloid nuclei 22,34 . Accord-ingly, it has been proposed that the immediate shockfreezing deficit results from insufficient time for process-ing of contextual information and activation of the amyg-

w xdala before the arrival of the shock input 13 . Only limitedinformation is available on the molecular mechanisms ofthe immediate shock freezing deficit. It has to be taken

Ž .into consideration that the ES 0.7 mA -Context groupproduced significantly less Fos than the Context group inseveral brain areas, although both groups spent the sametime in context 1 during the training. On this basis, it maybe speculated that the immediate shock input in the centralamygdaloid nucleus prevented subsequent processing ofcontextual information in the hippocampus, medial amyg-daloid nucleus and parietal somatosensory cortex. Thus, itseems likely that the absence of associative learning in theES-Context group may not be due only to temporal mis-match in the processing of stimuli, but also to a shock-in-duced suppression of neuronal activity. Similar effects ofstress, i.e., suppression of hippocampal long-term potentia-

w xtion has also been previously described in rats 14 . Itseems, however, that the shock could not interfere withFos production once the processing of contextual stimulihad already been initiated.

The low Fos production observed after exposure of themice to an immediate shock appeared paradoxical in viewof numerous reports mentioning an increased Fos produc-tion after exposure of rats to inescapable electric stressw x6,31,36 . In those studies, however, multiple shock admin-istrations resulted in learning of the contextual and appara-

w xtus cues 26,27 , so that the role of non-associative stressand contextual conditioning could not be easily distin-guished.

After both retention tests, the Context group producedsignificantly less Fos than after the training. This observa-tion was in agreement with reduced Fos production ob-

w xserved in rats repeatedly exposed to spatial cues 29 andsupported the interpretation that the reduced Fos levelsmay reflect non-associative learning, i.e., habituation. The

Ž .finding that the Context-ES 0.7 mA group producedsignificantly more Fos in the hippocampus and the centralnucleus of the amygdala after retention test 1 than both

Ž Ž . .control groups ES 0.7 mA -Context; Context was con-sistent with increased levels of Fos mRNA or Fos protein

w xafter the memory tests of other operant 7 or classicalw xconditioning 2,6,31 paradigms. In the latter studies, it has

been suggested that Fos production was elicited by the

( )S. MilanoÕic et al.rBrain Research 784 1998 37–4746

conditioned emotional response. Taking into considerationthat after retention test 1 the fear response of the Context-

Ž .ES 0.7 mA group was markedly generalized it is alsopossible that the Fos production was elicited by fearaccompanied by anxiety. Alternatively, higher Fos levelsafter retention test 1 could represent slower habituation ofFos production in mice conditioned to aversive stimulithan in non-conditioned controls. The latter possibilityseems more likely in view of the observation that habitua-tion and reduced generalization preceding retention test 2,resulted in low Fos production of all groups, although theconditioned mice still exhibited a strong fear response inthe conditioning context. In contrast to former suggestionsw x35 , it appears that the Fos production does not necessarilycorrelate with the emotional behavior of animals, in partic-ular not after habituation, reduced anxiety andror delaybetween training and retention.

Acknowledgements

This work was conducted in the Department of Molecu-lar Neuroendocrinology at the Max Planck Institute forExperimental Medicine in Goettingen, Germany where Dr.Milanovic worked as a guest scientist of the CentralInstitute for Mental Health, Mannheim, Germany. Theauthors gratefully acknowledge the technical assistance ofKarin Birkenfeld and Markki Palve.

References

w x1 K.V. Anokhin, S.P.R. Rose, Learning-induced increase of immediateearly gene messenger RNA in the chick forebrain, Eur. J. Neurosci.Ž .3 1991 162–167.

w x2 C.H.M. Beck, H.C. Fibiger, Conditioned fear-induced changes inbehaviour and in the expression of the immediate early gene c-fos:

Ž .with and without diazepam treatment, J. Neurosci. 15 1995 709–722.

w x3 R.J. Blanchard, D.C. Blanchard, Crouching as an index of fear, J.Ž .Comp. Physiol. Psychol. 67 1969 370–375.

w x4 R.J. Blanchard, K.K. Fukunaga, D.C. Blanchard, Environmentalcontrol of defensive reactions to footshock, Bull. Psychonom. Soc. 8Ž .1976 129–130.

w x5 M.E. Bouton, R.C. Bolles, Contextual control of the extinction ofŽ .conditioned fear, Learn. and Motivation 10 1979 445–466.

w x6 P. Campeau, M.D. Hayward, B.T. Hope, J.B. Rosen, E.J. Nestler,M. Davis, Induction of the c-fos proto-oncogene in the rat amygdala

Ž .during unconditioned and conditioned fear, Brain Res. 565 1991349–352.

w x7 M.A. Castro Alamancos, J. Borrell, L.M. Garcia Segura, Perfor-mance in an escape task induces Fos-like immunoreactivity in a

Ž .specific area of the motor cortex of the rat, Neuroscience 49 1992157–162.

w x8 S. Ceccatelli, M.J. Villar, M. Goldstein, H. Hokfelt, Expression ofc-fos immunoreactivity in transmitter-characterized neurons after

Ž .stress, Proc. Natl. Acad. Sci. USA 86 1989 9569–9573.w x9 X. Chen, J. Herbert, Regional changes in c-fos expression in the

basal forebrain and brainstem during adaptation to repeated stress:

correlations with cardiovascular, hypothermic and endocrine re-Ž .sponses, Neuroscience 64 1995 675–685.

w x10 W.E. Cullinan, J.P. Herman, D.F. Battaglia, H. Akil, S.J. Watson,Pattern and time course of immediate early gene expression in rat

Ž .brain following acute stress, Neuroscience 64 1995 477–505.w x11 M.S. Fanselow, Conditional and unconditional components of post-

Ž .shock freezing, Pavlovian J. Biol. Sci. 15 1980 177–182.w x12 M.S. Fanselow, Factors governing one-trial contextual conditioning,

Ž .Anim. Learn. Behav. 18 1990 264–270.w x13 M.S. Fanselow, J.P. DeCola, S.L. Young, Mechanisms responsible

for reduced contextual conditioning with massed unsignaled uncon-Ž . Ž .ditional stimuli, J. Exp. Psychol. Anim. Behav. 19 1993 121–137.

w x14 M.R. Foy, M.E. Stanton, S. Levine, R.F. Thompson, Behavioralstress impairs long-term potentiation in rodent hippocampus, Behav.

Ž .Neurol. Biol. 48 1987 138–149.w x Ž .15 K.B.J. Franklin, G. Paxinos Eds. , The Mouse Brain in Stereotaxic

Coordinates, Academic Press, San Diego, 1997.w x16 R.J. Handa, K.M. Nunley, M.R. Bollnow, Induction of c-fos mRNA

in the brain and anterior pituitary gland by a novel environment,Ž .Neuroreport 4 1993 1079–1082.

w x17 F.J. Helmstetter, Contribution of the amygdala to learning andŽ .performance of conditional fear, Physiol. Behav. 51 1992 1271–

1276.w x18 P. Hughes, M. Dragunow, Induction of immediate-early genes and

the control of neurotransmitter-regulated gene expression within theŽ .nervous system, Pharmacol. Rev. 47 1995 133–178.

w x19 T. Imaki, T. Shibasaki, M. Hotta, H. Demura, Intracerebroventricu-lar administration of corticotropin-releasing factor induces c-fosmRNA expression in brain regions related to stress responses:comparison with pattern of c-fos mRNA induction after stress, Brain

Ž .Res. 616 1993 114–125.w x20 J.J. Kim, M.S. Fanselow, Modality-specific retrograde amnesia of

Ž .fear, Science 256 1992 675–677.w x21 J.A. Kleim, E. Lussnig, E.R. Schwarz, T.A. Comery, W.T. Gree-

nough, Synaptogenesis and Fos expression in the motor cortex of theŽ .adult rat after motor skill learning, J. Neurosci. 16 1996 4529–4535.

w x22 R. Lamprecht, Y. Dudai, Differential modulation of brain immediateŽ .early genes by intraperitoneal LiCl, Neuroreport 7 1995 289–293.

w x23 K.J. Mack, P.A. Mack, Induction of transcription factors in so-matosensory cortex after tactile stimulation, Mol. Brain Res. 12Ž .1992 141–147.

w x24 H. Matthies, In search of cellular mechanisms of memory, Prog.Ž .Neurobiol. 32 1989 277–349.

w x25 R. Mileusnic, K. Anokhin, S.P. Rose, Antisense oligodeoxynu-cleotides to c-fos are amnestic for passive avoidance in the chick,

Ž .Neuroreport 7 1996 1269–1272.w x26 T.R. Minor, R.L. Jackson, S.F. Maier, Effects of task-irrelevant cues

and reinforcement delay on choice-escape learning following in-escapable shock: evidence for a deficit in selective attention, J. Exp.

Ž . Ž .Psychol. Anim. Behav. Proc. 10 1986 543–556.w x27 T.R. Minor, V.M. LoLordo, Escape deficits following inescapable

Žshock: the role of contextual odor, J. Exp. Psychol. Anim. Behav.. Ž .Proc. 10 1984 168–181.

w x28 J.I. Morgan, T. Curran, Stimulus-transcription coupling in the ner-vous system: involvement of the inducible proto-oncogenes Fos and

Ž .Jun, Annu. Rev. Neurosci. 14 1991 421–451.w x29 M. Papa, M.P. Pellicano, H. Welzl, A.G. Sadile, Distributed changes

in c-fos and c-jun immunoreactivity in the rat brain associated withŽ .arousal and habituation to novelty, Brain Res. Bull. 32 1993

509–515.w x30 J.P. Pascoe, W.F. Supple, B.S. Kapp, Learning and memory: verte-

Ž .brate models, in: J.L. Martinez Jr., R.P. Kesner Eds. , Learning andMemory: A Biological View, Academic Press, New York, 1991, pp.359–407.

w x31 M.A. Pezzone, W.S. Lee, G.E. Hoffman, B.S. Rabin, Induction ofc-fos immunoreactivity in the rat forebrain by conditioned and

Ž .unconditioned aversive stimuli, Brain Res. 597 1992 41–50.

( )S. MilanoÕic et al.rBrain Research 784 1998 37–47 47

w x32 R.G. Phillips, J.E. LeDoux, Differential contribution of amygdalaand hippocampus to cued and contextual fear conditioning, Behav.

Ž .Neurosci. 106 1992 274–285.w x33 B. Rozendaal, J.M. Koolhaas, B. Bohus, Central amygdala lesions

affect behavioral and autonomic balance during stress in rats, Phys-Ž .iol. Behav. 50 1991 777–781.

w x34 S.M. Sagar, F.R. Sharp, T. Curran, Expression of c-fos protein in theŽ .brain: metabolic mapping at the cellular level, Science 240 1988

1328–1331.w x35 G. Sandner, P. Oberling, M.C. Silveira, G. Di Scala, B. Rocha, A.

Bagri, R. Depoortere, What brain structures are active during emo-tions? Effect of brain stimulation elicited aversion on c-fos immuno-

Ž .reactivity and behavior, Behav. Brain Res. 58 1993 9–18.w x36 M.A. Smith, S. Banerjee, P.W. Gold, J. Glowa, Induction of c-fos

mRNA in rat brain by conditioned and unconditioned stressors,Ž .Brain Res. 578 1992 135–141.

w x37 W. Tischmeyer, L. Kaczmarek, M. Strauss, R. Jork, H. Matthies,Accumulation of c-fos mRNA in rat hippocampus during acquisition

Ž .of a brightness discrimination, Behav. Neurol. Biol. 54 1990165–171.

w x38 B.C. Thinus, E. Save, A.C. Rossi, A. Tozzik, T.M. Ammassari, Thedifferences shown by C57BLr6 and DBAr2 inbred mice in detect-ing spatial novelty are subserved by a different hippocampal and

Ž .parietal cortex interplay, Behav. Brain Res. 80 1996 33–40.w x39 R. Titze de Almeida, H. Shida, F.S. Guimaraes, E.A. Del Bel,

Stress-induced expression of the c-fos proto-oncogene in the hip-Ž .pocampal formation, Brazil J. Med. Biol. Res. 27 1994 1083–1088.