Conditioned defeat in male and female Syrian hamsters

Page created by Kenneth Mills
 
CONTINUE READING
Available online at www.sciencedirect.com
                                                                                           R

                                                     Hormones and Behavior 44 (2003) 293–299                       www.elsevier.com/locate/yhbeh

                 Conditioned defeat in male and female Syrian hamsters
    Kim L. Huhman,* Matia B. Solomon, Marcus Janicki, Alvin C. Harmon, Stacie M. Lin,
                          Jeris E. Israel, and Aaron M. Jasnow
                                    Department of Psychology, Georgia State University, Atlanta, GA 30303, USA

                                      Received 1 September 2002; revised 16 April 2003; accepted 4 May 2003

Abstract
    A brief exposure to social defeat in male Syrian hamsters (Mesocricetus auratus) leads to profound changes in the subsequent agonistic
behavior exhibited by the defeated animals. Following defeat in the home cage of an aggressive conspecific, male hamsters will subsequently
fail to defend their home territory even if the intruder is a smaller, nonaggressive male. This phenomenon has been called conditioned defeat.
In Experiment 1, we examined the duration of conditioned defeat by repeatedly testing (every 3–5 days) defeated hamsters with a
nonaggressive intruder. We found that conditioned defeat occurs in all defeated male hamsters and persists for a prolonged period of time
(at least 33 days) in the majority of male hamsters tested despite the fact that these animals are never attacked by the nonaggressive intruders.
In Experiment 2, we examined whether conditioned defeat could be induced in female Syrian hamsters. While conditioned defeat occurred
in some females, they displayed only low levels of submissive/defensive behavior and, in contrast to males, the conditioned defeat response
did not persist beyond the first test. These results suggest that in male hamsters conditioned defeat is a profound, persistent behavioral change
characterized by a total absence of territorial aggression and by the frequent display of submissive and defensive behaviors. Conversely,
social defeat in female hamsters does not appear to induce long-term behavioral changes. Finally, in Experiment 3, we determined that
plasma adrenocorticotropin-like immunoreactivity increases in females following social defeat in a manner similar to that seen in males,
suggesting that the disparate behavioral reactions of males and females are not due to sex differences in the release of, or response to, plasma
adrenocorticotropin.
© 2003 Elsevier Inc. All rights reserved.

Keywords: Aggression; Agonistic behavior; Defense; Sexual dimorphism; Stress; ACTH; Submissive behavior; Social stress

Introduction                                                                other hand, display baseline levels of these hormones. This
                                                                            difference in the response of dominant and subordinate
   Syrian hamsters are believed to be solitary animals that                 hamsters occurs even though both animals usually exhibit
engage in territorial defense against intruding conspecifics                some exertion. Further, previously defeated hamsters will
(Nowack and Paradiso, 1983). In laboratory conditions                       produce a hormonal stress response even when physical
when housed individually, resident hamsters reliably attack                 contact between the conspecifics is prevented indicating that
and defeat intruders that are placed into their home cages.                 the response in the defeated animals is largely psychological
Social defeat appears to be a potent stressor in that defeated              (Huhman et al., 1992). Defeated animals may also exhibit
hamsters exhibit an activated hypothalamic-pituitary-adre-                  increased blood pressure and heart rate and compromised
nal (HPA) axis as evidenced by increases in plasma adre-                    immune function while, again, their dominant opponents are
nocorticotropin (ACTH), ␤-endorphin, and glucocorticoids                    largely unaffected (Blanchard et al., 1995; Bohus et al.,
as well as decreases in plasma testosterone (Huhman et al.,                 1983; Jasnow et al., 2001).
1990, 1991a, 1991b, 1992). Dominant hamsters, on the                           Social defeat can also produce profound changes in the
                                                                            defeated animal’s subsequent social behavior. If a male
    * Corresponding author. Department of Psychology, PO Box 5010, Geor-
                                                                            hamster is defeated in the home cage of another, more
gia State University, Atlanta, GA 30303, USA. Fax: ⫹1-404-651-3929.         aggressive animal, the defeated male subsequently appears
    E-mail address: khuhman@gsu.edu (K.L. Huhman).                          virtually unable to reverse its subordinate social status. This

0018-506X/$ – see front matter © 2003 Elsevier Inc. All rights reserved.
doi:10.1016/j.yhbeh.2003.05.001
294                                 K.L. Huhman et al. / Hormones and Behavior 44 (2003) 293–299

is such a profound behavioral change that the defeated               as the larger adrenal size is dependent on estrogen in female
hamster subsequently fails to defend its territory (its own          rats (Gaskin and Kitay, 1970). Given that the response of
home cage) even when a smaller, nonaggressive hamster is             the hypothalamic-pituitary-adrenocortical (HPA) axis has
introduced. This behavior is characterized by a total absence        been hypothesized to affect ongoing and future agonistic
of territorial aggression accompanied by the repeated dis-           behavior in rodents (Leshner, 1975, 1983), it is possible that
play of submissive and defensive behaviors (Potegal et al.,          a differential responsiveness of the HPA axis in male and
1993; Jasnow et al., 1999) and has been called conditioned           female hamsters following social defeat might account for
defeat. A response similar to conditioned defeat has also            possible differences in the behavioral responses. In Exper-
been demonstrated in male rats and mice (Corrigan and                iment 3, we determined whether exposure to a defeat stres-
Flannelly, 1979; Frishknecht et al., 1982).                          sor stimulates the HPA axis in female hamsters as it does in
    We have begun to explore the neurobiological basis for           males (Huhman et al., 1990) by measuring plasma adreno-
the profound behavioral changes exhibited by defeated male           corticotropin-like immunoreactivity following social defeat.
hamsters. We have demonstrated that intracerebroventricu-
lar administration of a corticotropin releasing factor (CRF)
receptor antagonist, D-Phe CRF, significantly reduces the            Materials and methods
submissive and defensive behavior exhibited by previously
defeated hamsters (Jasnow et al., 1999). This effect can also        Subjects
be produced by infusing D-Phe CRF into the bed nucleus of
the stria terminalis but not into the central amygdala (Jas-            Subjects were adult, male and female Syrian hamsters
now A.M., et al., unpublished data). In the basolateral              (Mesocricetus auratus, Charles River Laboratories) that
amygdala, infusion of either a GABA agonist or an NMDA               weighed 100 –130 g and were 9 weeks old at the beginning
glutamate antagonist virtually abolishes the submissive/de-          of the experiment. All hamsters to be tested were individ-
fensive behaviors (Jasnow et al., 1999) (Jasnow, A.M., et            ually housed for 2 weeks prior to testing in a temperature-
al., in press). In all of these studies, we have tested the          controlled (20 ⫾ 2°C) colony room on a 14:10 h light:dark
subjects for conditioned defeat only once. Little is known           (LD) cycle with lights off at 1100 h. Each animal was
about the duration of the behavioral changes following the           briefly handled daily for 1 week before the beginning of
initial defeat. In Experiment 1, we examined the duration of         each experiment. Hamsters were not prescreened for base-
conditioned defeat in male Syrian hamsters tested repeat-            line aggressiveness or submissiveness before the initial so-
edly with a nonaggressive intruder. In addition, we com-             cial defeat. Additional singly-housed male and female ham-
pared a group of defeated males with a group of nonde-               sters weighing 130 –180 g (⬎6 months old) at the beginning
feated males to assess the extent of the behavioral change           of the experiment were used as resident aggressors for
produced by the initial defeat experience.                           defeat training. These animals were paired with a smaller
    In most mammalian species, including laboratory ro-              hamster several times before the beginning of each experi-
dents, males are larger and more aggressive than are fe-             ment to verify that they would reliably defeat an intruder
males. In contrast, female hamsters are larger than are males        placed into their home cage and that they also would not
(Payne and Swanson, 1970; Swanson, 1967; Marques and                 produce tissue damage in their opponents. We are interested
Valenstein, 1977), and both male and female Syrian ham-              primarily in the psychological effects of social defeat; there-
sters are highly aggressive (Payne and Swanson, 1970;                fore we immediately terminate any pairing in which a ham-
Lerwill and Makings, 1971). In fact, when not sexually               ster is bitten such that it bleeds. Group-housed male and
receptive, female hamsters tend to be dominant over male             female hamsters (five/cage) weighing 90 –120 g (7 weeks
conspecifics even if their weights are the same (Payne and           old) were used as nonaggressive intruders (NAI) during
Swanson 1970; Brain, 1972). Little is known about the                behavioral testing. These smaller, group-housed hamsters
response of female hamsters to a social defeat stressor, and         generally do not display agonistic behavior during a social
all our previous work has been done with males. It has been          encounter but will instead spend most of the test time
shown that group housing in female hamsters induces a                exploring the cage. In the case of the female NAIs, half
number of physiological signs of stress (Fritzsche et al.,           were intact females and their cycles were monitored so that
2000; Gattermann and Weinandy, 1996). In Experiment 2,               they were not used on the day of estrus. The remainder of
we examined whether conditioned defeat could be induced              the female NAIs were ovariectomized. We observed each
in female hamsters following exposure to social defeat, and          cage of group-housed intruders before testing and marked
if so, whether the behavioral change was persistent.                 the dominant hamster so that it is not used as a NAI. All
    Finally, the typical sexual dimorphism in adrenal gland          animals were housed in polycarbonate cages (20 ⫻ 40 ⫻ 20
size and steroid production observed in laboratory rodents is        cm) with wire mesh tops and corn cob bedding, and food,
reversed in Syrian hamsters. Thus, male hamsters have                water, and cotton nesting materials were available ad libi-
larger adrenal glands and greater steroid production than do         tum. All procedures and protocols were approved by the
females, and this hypertrophic action appears to be depen-           Georgia State University Institutional Animal Care and Use
dent on testosterone in male hamsters in much the same way           Committee, and all methods were in accordance with the
K.L. Huhman et al. / Hormones and Behavior 44 (2003) 293–299                              295

standards outlined in the National Institutes of Health Guide         were subjected to social defeat as described above, whereas
for the Care and Use of Laboratory Animals.                           the other half (novel cage controls) were placed into the
                                                                      empty cage of a resident aggressor during training. All 20
Experiment 1a                                                         hamsters were tested the following day with a nonaggres-
                                                                      sive intruder and behavior was videotaped and scored as
Defeat training                                                       described above.
   On the day of training, 14 male hamsters were trans-
ported from the colony room to the behavioral testing room            Experiment 2
where they were placed into the cage of a resident aggressor
for four, 5-min defeat training trials. Training began at                Thirty-one adult female, Syrian hamsters were used in
0900 h with 2-h intervals between each 5-min training                 Experiment 2, and their estrous cycles were monitored be-
session. The resident aggressors were larger animals that             fore and during the experiment using visual assessment of
were known to be aggressive, and in all cases these animals           vaginal swabs. Some of the females (n ⫽ 18) were trained
became dominant over the subjects.                                    for conditioned defeat (training occurred only on Diestrus 1
                                                                      or Diestrus 2) in the same manner as in Experiment 1,
Behavioral testing                                                    whereas the rest of the females (n ⫽ 13) were left undis-
   Behavioral testing began 1 day following defeat training,          turbed and served as home cage controls. Behavioral testing
and all testing was completed during the first 2 h of the dark        for defeated and control (nondefeated) animals began 1 day
phase of the LD cycle to minimize circadian variation in              following the training procedure and was conducted in the
agonistic behavior (Landau, 1975). A resident/intruder pair-          same manner as in Experiment 1 except that no animal was
ing was used for testing consisting of placing a group-               tested on the day of sexual receptivity. All animals were
housed, nonaggressive intruder into the home cage of each             then retested at 4-day intervals or until they displayed ag-
subject for 5 min. Testing was then repeated every 3– 6 days          gression toward a nonaggressive intruder. Subjects were
for a period of 33 days or until an experimental male                 paired with either an intact, nonaggressive intruder that was
became aggressive toward the NAI placed in its cage. The              not in estrus or with an intruder that was ovariectomized.
NAIs were randomly paired with the experimental animals               Again, these intruders were group-housed and were smaller
so that animals were generally paired with an unfamiliar              than the experimental animals, and they exhibited no ag-
conspecific on each test. The behavior of the subjects was            gressive behavior toward the home cage animals. All ses-
scored during testing by two trained observers using a                sions were recorded on VHS tape and the behavior of
detailed behavioral inventory as described elsewhere (Huh-            subjects was analyzed as in Experiment 1.
man et al., 1990; Jasnow et al., 1999). Interobserver reli-
ability was ⬎90%. The following behaviors were observed               Experiment 3
and recorded as total duration in seconds over the 5-min
testing period:                                                          Twenty female hamsters were matched by weight and
                                                                      divided into three groups that were randomly designated as
   1. Nonsocial: locomotor/exploratory, self-groom, nest-             residents (n ⫽ 7), intruders (n ⫽ 7), or home cage controls
      ing, feeding, sleeping.                                         (n ⫽ 6). On the day of the experiment, each intruder was
   2. Social: attend, approach, investigate, sniff, touching          placed in the home cage of a weight-matched resident and
      nose.                                                           the animals were allowed to interact for 15 min. Again, the
   3. Submissive/defensive: upright/side defense, tail lift,          encounters were videotaped for verification of the domi-
      tooth chatter, flee, full submissive posture.                   nance status that was assigned by two observers during
   4. Aggressive: upright/side offense, chase, bite, attack.          behavioral testing. Immediately following each pairing, the
Subjects were judged as displaying conditioned defeat if the          animals were brought one at a time to an adjacent room
only agonistic behaviors that they produced during testing            where they were rapidly decapitated, and trunk blood was
with the nonaggressive intruder were submissive (tail lift,           collected in vials containing 50 ␮l of heparin. The order in
flee, tooth chatter, full submissive posture) or defensive            which blood was collected (i.e., whether the dominant or
behaviors (upright and side defense). If a subject displayed          subordinate hamster was sampled first) was alternated and a
any aggressive behavior (upright or side offense, chase,              control animal was sacrificed between each pair. Blood
attack or bite) toward the nonaggressive intruder, then they          collection for each pair of hamsters occurred within 5 min
were no longer considered to be exhibiting conditioned                of the end of the encounter. Blood was kept on ice until it
defeat.                                                               was spun in a refrigerated centrifuge, and plasma was stored
                                                                      at ⫺20°C. The assay for plasma adrenocorticotropin
Experiment 1b                                                         (ACTH)-like immunoreactivity was performed by the En-
                                                                      docrine Core Laboratory, Yerkes Regional Primate Re-
   In a related experiment, 20 male hamsters were trans-              search Center of Emory University, using a commercially
ported to the behavioral testing room. Half of the animals            prepared kit (antibody-coated tubes) for human ACTH (Di-
296                                  K.L. Huhman et al. / Hormones and Behavior 44 (2003) 293–299

agnostic Systems Laboratories, Webster, TX). Intraassay
variability was less than 10% and all samples were run in
the same assay.

Statistical analysis

   The data in Experiments 1 and 2 violated the homoge-
neity of variance assumption (Levene’s Test for Equality of
Variances); therefore, only nonparametric statistical tests
were used in these experiments. In Experiment 2, there was
no difference in the behavior of the defeated or nondefeated
females based on whether an OVX or non-OVX intruder
was used during testing (P ⬎ 0.05); therefore, these data
were collapsed for further analysis. The behavioral data
(durations in seconds) of defeated and control animals were
analyzed using a Mann-Whitney U (two-tailed) with results
reported as Z scores. The proportion of defeated and non-             Fig. 1. Percentage of defeated males (N ⫽ 14) that exhibited conditioned
                                                                      defeat (submissive and defensive behaviors and no aggression) during
defeated males (Exp. 1b) and females (Exp. 2) exhibiting
                                                                      subsequent generalization tests when paired with a nonaggressive intruder.
conditioned defeat (operationally defined as an animal that
displayed at least some submissive/defensive behavior and
no aggression) or exhibiting only territorial aggression was          aggressor for defeat training, and 10 males were placed in
compared using a chi-squared analysis. Although the data              the cage of a resident aggressor with the aggressor removed
were generated in separate experiments, we conducted com-             from the cage (no-defeat controls). Three of the males in the
parisons between males and females for proportion of ham-             defeat group actually vigorously attacked the resident ag-
sters exhibiting conditioned defeat (log-linear analysis re-          gressor, so these males were excluded from further testing
ported as G statistic) and for the durations of behaviors             because they were never defeated. Previous defeat signifi-
exhibited by males and females [Kruskal-Wallis followed               cantly influenced the likelihood that a male hamster would
by Dunn’s multiple comparison (reported as Q statistic)]. In          exhibit conditioned defeat [␹2(1) ⫽ 17.0, P ⬍ 0.01] or
Experiment 3, the hormonal values for subordinate, domi-              territorial aggression [␹2(1) ⫽ 13.4, P ⬍ 0.01] when tested
nant, and control hamsters were compared using a one-way              the next day with a nonaggressive intruder. All of the
ANOVA. Statistical significance for all analyses was as-              defeated males exhibited conditioned defeat when tested
cribed at P ⬍ 0.05.                                                   with a nonaggressive intruder, whereas 9 of 10 of the con-
                                                                      trol males attacked the nonaggressive intruders. One control
                                                                      animal displayed only social behavior toward the intruder.
Results                                                               The durations (in seconds) of the behaviors and the propor-
                                                                      tion of defeated and nondefeated males exhibiting condi-
    None of the animals in this study were bitten such that           tioned defeat are shown in Figs. 2a and 3a, respectively.
they bled during defeat training; thus, no animals were               Defeated males exhibited significantly more submissive/
removed from the study because of injury. In Experiment               defensive behavior (Z ⫽ ⫺3.842, P ⬍ 0.01) and signifi-
1a, 100% (14 of 14) of the defeated males displayed con-              cantly less aggressive behavior (Z ⫽ ⫺3.246, P ⬍ 0.01)
ditioned defeat for at least 10 days (through the fourth test)        than did nondefeated males. Social and nonsocial behavior
following conditioned defeat training. On the 5th testing             did not differ between groups (P ⬎ 0.05).
day (16 days following defeat), 93% (13 of 14) of the                     In Experiment 2, the Mann-Whitney U between previ-
animals continued to display conditioned defeat. One ani-             ously defeated and nondefeated (control) females on the
mal displayed aggressive behavior when the NAI was in-                first day of testing revealed that there were no significant
troduced into its home cage, and this animal was scored as            differences in the duration of aggressive and nonsocial be-
no longer displaying conditioned defeat. On the 6th testing           haviors (P ⬎ 0.05). Significant differences, however, were
day (21 days following defeat), 79% (11 of 14) of the                 found in the duration of social (Z ⫽ ⫺2.102, P ⬍ 0.05) and
hamsters continued to display conditioned defeat. On test-            submissive/defensive behaviors (Z ⫽ ⫺2.263, P ⬍ 0.05)
ing day 7 (26 days following defeat), 71% (10 of 14) of the           (see Fig. 2b) with defeated females exhibiting more sub-
animals continued to display conditioned defeat behaviors.            missive/defensive and less social behavior. Previous defeat
Fifty-seven percent of the animals continued to exhibit               significantly influenced the likelihood that a female hamster
conditioned defeat through the end of the experiment (33              would exhibit conditioned defeat [␹2(1) ⫽ 4.31, P ⬍ 0.05]
days following defeat). The results of Experiment 1a are              or territorial aggression [␹2(1) ⫽ 6.48, P ⬍ 0.05] when
summarized in Fig. 1.                                                 tested with a nonaggressive intruder during the first behav-
    In Experiment 1b, 10 males were placed with a resident            ioral test (see Fig. 3b). On the second behavioral test, all of
K.L. Huhman et al. / Hormones and Behavior 44 (2003) 293–299                                              297

                                                                               Kruskal-Wallis nonparametric ANOVA. The Kruskal-Wal-
                                                                               lis was significant for submissive/defensive behavior (H ⫽
                                                                               29.7, P ⬍ 0.01). The Dunn’s multiple comparisons test
                                                                               indicated that defeated males exhibited significantly more
                                                                               submissive/defensive behavior than did defeated females (Q
                                                                               ⫽ 3.76, P ⬍ 0.05).
                                                                                   ANOVA revealed a significant effect [F(2,15) ⫽ 3.87, P
                                                                               ⬍ 0.05] of group on plasma ACTH-like immunoreactivity
                                                                               with subordinate hamsters displaying higher plasma ACTH
                                                                               levels than did home cage controls. Dominant females and
                                                                               controls exhibited similar ACTH levels. Mean (⫾ SEM)
                                                                               plasma ACTH-like immunoreactivity in controls was 167 ⫾
                                                                               28 pg/ml. Dominant females had a mean ACTH level of 172
                                                                               ⫾ 22 pg/ml and controls exhibited a mean of 249 ⫾ 20
                                                                               pg/ml.

Fig. 2. Average duration (mean ⫾ SEM in seconds) of submissive/defen-
sive (sub/def), nonsocial, social, and aggressive behavior of defeated (dark
bars, N ⫽ 7) and nondefeated (hatched bars, N ⫽ 10) males (a) and
defeated (dark bars, N ⫽ 18) and nondefeated (hatched bars, N ⫽ 13)
females (b) during a subsequent generalization test with a nonaggressive
intruder. Nonparametric statistics revealed that defeated males and females
exhibited significantly more submissive/defensive behaviors (*P ⬍ 0.01)
than did their nondefeated counterparts. Defeated males displayed signif-
icantly less aggression (**P ⬍ 0.01) than did nondefeated males and
defeated females displayed significantly less social behavior (***P ⬍ 0.05)
compared with nondefeated females. Defeated males exhibited signifi-
cantly more submissive/defensive behaviors than did defeated females (P
⬍ 0.05).

the defeated and nondefeated animals exhibited aggression
toward the nonaggressive intruders; thus, there were no
females that displayed long-lasting conditioned defeat.
    The proportion of males and females that exhibited con-
ditioned defeat, aggression, or social behavior (Fig. 3) was
compared using log-linear analysis. There was a sex differ-                    Fig. 3. Percentage of defeated (dark bars) and nondefeated (blank bars)
ence only in the proportion of hamsters exhibiting condi-                      males (a) and females (b) that exhibited either conditioned defeat, only
tioned defeat [G(2) ⫽ 10.3, P ⬍ 0.01] with more males than                     social behavior and no submission or aggression, or aggressive behavior.
                                                                               Prior defeat significantly altered the proportions of males and females that
females showing conditioned defeat. The durations of sub-                      exhibited conditioned defeat and aggression. Significant differences among
missive/defensive, aggressive, and social behavior emitted                     groups are indicated by brackets and an asterisk. Group N values are shown
by males and females (Fig. 2) were compared using a                            above each bar.
298                                  K.L. Huhman et al. / Hormones and Behavior 44 (2003) 293–299

Discussion                                                            defeat occurs. The hamsters experience the initial social
                                                                      defeat in the home cage of a larger, more aggressive oppo-
    Conditioned defeat occurs in males of a number of ro-             nent. The defeated animals are then tested in their own
dent species including rats, mice, and hamsters (Corrigan             home cages against a smaller, group-housed intruder that
and Flannelly, 1979; Frishknecht et al., 1982; Potegal et al.,        exhibits no aggression. Thus, the males appear to generalize
1993). We classify a hamster as exhibiting conditioned                considerably from the training to the testing situation. One
defeat if that individual produces only submissive and de-            might expect that the repeated exposure to a smaller, non-
fensive agonistic behaviors and does not emit any aggres-             aggressive opponent would facilitate extinction of the con-
sive behaviors. By this definition, numerous male hamsters            ditioned defeat response and reinstatement of normal terri-
in the present study continued to display conditioned defeat          torial aggression. The present study does not support this
in response to a nonaggressive intruder for a period of at            possibility and instead leaves open the possibility that the
least 33 days following the single initial defeat experience.         repeated exposure might actually be prolonging the re-
This response was emitted despite the fact that these males           sponse. While we have not tested this possibility using the
were never attacked after the initial training day. This study        precise social defeat methodology presented here, we have
also indicates that the response to social defeat in hamsters         examined this possibility following a more severe defeat
is sexually dimorphic. Only 28% of defeated female ham-               experience (Potegal et al., 1993). In that study, conditioned
sters exhibited conditioned defeat, and the duration of sub-          defeat diminished at the same rate in hamsters that were
missive/defensive behavior produced by these females was              tested repeatedly and hamsters that were tested only once
considerably less than that observed in males. The present            following a delay, indicating that the subsequent exposures
data also do not support the hypothesis that female hamsters          led to neither habituation nor sensitization of conditioned
fail to show conditioned defeat because they do not produce           defeat.
a HPA axis response to social defeat.                                     Exposure to social defeat does not appear to elicit major
    This study provides a clear demonstration of the ability          changes in the future agonistic behavior of defeated female
of social defeat to alter agonistic behavior in male hamsters.        hamsters. We did find that the submissive/defensive and
All defeated males subsequently failed to exhibit territorial         social behavior of defeated females was significantly differ-
aggression when tested in their home cages with a nonag-              ent from that of nondefeated females. This difference, how-
gressive intruder, whereas all nondefeated males attacked             ever, was only observable during the initial generalization
the nonaggressive intruders. Of possible concern is the fact          test, and the behavioral change was neither profound nor
that we excluded three hamsters from the defeat group                 long-lasting. The duration of submissive/defensive behavior
because they were not defeated during initial training but            produced by defeated females was significantly greater than
instead attacked the resident aggressor. This occurrence was          that displayed by nondefeated females, but this difference
unprecedented. Our hamsters are group-housed before they              could hardly be considered striking. By the definition dis-
arrive in our laboratory, and thus they have acquired con-            cussed earlier, only 5 of the 18 defeated females exhibited
siderable social experience and already possess either ag-            conditioned defeat and only 1 of these females produced a
gressive or submissive tendencies. We do not prescreen or             high frequency of submissive/defensive behaviors (157/300
select our subjects in any way. Nevertheless, our resident            s). By contrast, males that were similarly defeated all ex-
aggressors generally attack and defeat all of our experimen-          hibited conditioned defeat and 5 of 7 of these males dis-
tal subjects, and all of these defeated males then exhibit            played more than 135 s of submissive/defensive behavior.
conditioned defeat. Very occasionally, we do have an ex-              Furthermore, we found conditioned defeat to persist for at
perimental animal that defeats a resident aggressor, and              least four behavioral tests in all defeated males whereas all
these animals are excluded from further testing. Based on             defeated females exhibited normal territorial aggression by
our considerable experience with this model (Jasnow et al.,           the second behavioral test. A sex difference in responsive-
1999), we do not feel that the results of this experiment are         ness to social stress and defeat has also been reported in rats
biased by the unusual occurrence of aggressive hamsters in            (Haller et al., 1998a) with defeat being more stressful in
the defeat group because the duration of submissive/defen-            males. Rats are social animals and females are not particu-
sive behaviors observed was similar to that which we have             larly aggressive in the absence of pups, so it is not clear if
observed in other studies (Jasnow and Huhman, 2001;                   this finding can be generalized to solitary-living hamsters.
Whitten et al., 2001; Jasnow et al., 1999). Finally, although             Our results also have interesting implications for the
the precise average duration of submissive behavior dis-              debate on the role of pituitary-adrenocortical hormones in
played by the defeated groups is somewhat variable among              agonistic behavior (Leshner, 1983; Albers et al., 2002;
studies (as should be expected with complex social behav-             Haller et al., 1998b). Briefly, there are data suggesting that
ior), the inhibitory effect of social defeat on subsequent            ACTH and glucocorticoids increase submissiveness 24 and
territorial aggression in male hamsters is robust, reproduc-          48 h after social defeat in mice (Roche and Leshner, 1979).
ible, and comparable among studies.                                   Leshner (1975, 1983) has maintained that the stress hor-
    It is interesting to note that conditioned defeat is ob-          mone response to defeat provides not only a mechanism
served in a different context from that in which the initial          whereby an animal can adapt to the stressor but that these
K.L. Huhman et al. / Hormones and Behavior 44 (2003) 293–299                                                299

hormones (particularly ACTH) also feed back to influence                      Haller, J., Fuchs, E., Halasz, J., Makara, G.B., 1998a. Defeat is a major
an animal’s current and future agonistic behavior. The                            stressor in males while social instability is stressful mainly in females:
                                                                                  towards the development of a social stress model in female rats. Brain
present data do not support this role for ACTH in our model.                      Res. Bull. 50, 33–39.
ACTH was significantly elevated following social defeat in                    Haller, J., Halasz, J., Makara, F., Kruk, M., 1998b. Acute effects of
female hamsters, yet the majority of defeated females sub-                        glucocorticoids: behavioral and pharmacological perspectives. Neuro-
sequently exhibit normal territorial aggression. In male                          sci. Biobehav. Rev. 23, 337–344.
hamsters, we have shown that animals still display condi-                     Huhman, K.L., Bunnell, B.N., Mougey, E.H., Meyerhoff, J.L., 1990. Ef-
                                                                                  fects of social conflict on POMC-derived peptides and glucocorticoids
tioned defeat even if the pituitary-adrenocortical response is
                                                                                  in male golden hamsters. Physiol. Behav. 47, 949 –956.
completely blocked (Jasnow et al., 1999). Thus, in hamsters                   Huhman, K.L., Hebert, M.A., Meyerhoff, J.L., Bunnell, B.N., 1991a.
our data do not appear to support a role for ACTH (and                            Plasma cyclic AMP increases in hamsters following exposure to a
presumably glucocorticoids) in the acquisition or mainte-                         graded footshock stressor. Psychoneuroendocrinology 16, 559 –563.
nance of conditioned defeat.                                                  Huhman, K.L., Moore, T.O., Ferris, C.F., Mougey, E.H., Meyerhoff, J.L.,
                                                                                  1991b. Acute and repeated exposure to social conflict in male golden
                                                                                  hamsters: increases in plasma POMC-peptides and cortisol and de-
                                                                                  creases in plasma testosterone. Horm. Behav. 25, 206 –216.
Acknowledgments                                                               Huhman, K.L., Moore, T.O., Mougey, E.H., Meyerhoff, J.L., 1992. Hor-
                                                                                  monal responses to fighting in hamsters: separation of physical and
    This work was supported by NIH MH62044 to K.L.H.                              psychological causes. Physiol. Behav. 51, 1083–1086.
and MH12907 to A.M.J. This material is based upon work                        Jasnow, A.M., Banks, M.C., Owens, E.C., Huhman, K.L., 1999. Differen-
                                                                                  tial effects of two corticotropin-releasing factor antagonists on condi-
supported, in part, by the STC program of the NSF under
                                                                                  tioned defeat in male Syrian hamsters (Mesocricetus auratus). Brain
agreement #IBN-9876754. The authors thank Susie Lackey                            Res. 846, 122–128.
at the Endocrine Core Laboratory, Yerkes Regional Primate                     Jasnow, A.M., Drazen, D.L., Huhman, K.L., Nelson, R.J., Demas, G.E.,
Research Center of Emory University, for her assistance                           2001. Acute and chronic social defeat suppresses humoral immunity of
with the hormonal assay. We also thank David Marshall,                            male Syrian hamsters (Mesocricetus auratus). Horm. Behav. 40, 428 –
Charlie Smeeten, and Patricia Hicks for their expert animal                       433.
                                                                              Jasnow, A.M., Huhman, K.L., 2001. Activation of GABA(A) receptors in
care.                                                                             the amygdala blocks the acquisition and expression of conditioned
                                                                                  defeat in Syrian hamsters. Brain Res. 920, 142–150.
                                                                              Landau, I.T., 1975. Light-dark rhythms in aggressive behavior of the male
References                                                                        golden hamster. Physiol. Behav. 14, 767–774.
                                                                              Lerwill, C.J., Makings, P., 1971. The agonistic behavior of the golden
Albers, H.E., Huhman, K.L., Meisel, R.L., 2002. Hormonal basis of social          hamster Mesocricetus auratus (Waterhouse). Anim. Behav. 19, 714 –
    conflict and communication, in: Pfaff, D.W., Arnold, A.P., Etgen,             721.
    A.M., Fahrbach, S.E., Rubin, R.T. (Eds.), Hormones, Brain and Be-         Leshner, A.I., 1975. Theoretical review: a model of hormones and agonis-
    havior, Vol. 1, Academic Press, San Diego, pp. 393– 433.                      tic behavior. Physiol. Behav. 15, 225–235.
Blanchard, D.C., Spencer, R.L., Weiss, S.M., Blanchard, R.J., McEwen,         Leshner, A.L., 1983. The hormonal responses to competition and their
    B., Sakai, R.R., 1995. Visible burrow system as a model of chronic            behavioral significance, in: Svare, B.B. (Ed.), Hormones and Aggres-
    social stress: behavioral and neuroendocrine correlates. Psychoneu-           sive Behavior, Plenum Press, New York, pp. 393– 404.
    roendocrinology 20, 117–134.                                              Marques, D.M., Valenstein, E.S., 1977. Individual differences in aggres-
Bohus, B., Versteeg, C.A., De Jong, W., Cransberg, K., Kooy, J.G., 1983.          siveness of female hamsters: response to intact and castrated males and
    Neurohypophysial hormones and central cardiovascular control. Prog.           to females. Anim. Behav. 25, 131–139.
    Brain Res. 60, 445– 457.                                                  Nowack, R.M., Paradiso, J.L., 1983. Walker’s Mammals of the World.
Brain, P.F., 1972. Effects of isolation/grouping on endocrine function and        Johns Hopkins University Press, Baltimore, MD.
    fighting behavior in male and female golden hamsters (Mesocricetus        Payne, A.P., Swanson, H.H., 1970. Agonistic behaviour between pairs of
    auratus Waterhouse). Behav. Biol. 7, 349 –357.                                hamsters of the same and opposite sex in a neutral observation area.
Corrigan, J.G., Flannelly, K.J., 1979. Ultrasonic vocalizations of defeated       Behaviour 36, 260 –269.
    male rats. J. Comp. Physiol. Psychol. 93, 105–115.                        Potegal, M., Huhman, K., Moore, T., Meyerhoff, J., 1993. Conditioned
Frishknecht, J.-R., Seigfreid, B., Waser, P.G., 1982. Learning of submis-         defeat in the Syrian golden hamster (Mesocricetus auratus). Behav.
    sive behavior in mice: a new model. Behav. Processes 7, 235–245.              Neural Biol. 60, 93–102.
Fritzsche, P., Riek, M., Gattermann, R., 2000. Effects of social stress on    Roche, K.E., Leshner, A.I., 1979. ACTH and vasopressin treatments im-
    behavior and corpus luteum in female golden hamsters (Mesocricetus            mediately after a defeat increase future submissiveness in male mice.
    auratus). Physiol Behav. 68, 625– 630.                                        Science 204, 1343–1344.
Gaskin, J., Kitay, J., 1970. Adrenocortical function in the hamster: sex      Swanson, H.H., 1967. Effects of pre- and post-pubertal gonadectomy on
    differences and effects of gonadal hormones. Endocrinology 87, 779 –          sex differences in growth, adrenal and pituitary weights of hamsters. J.
    786.                                                                          Endocrinol. 39, 555–564.
Gattermann, R., Weinandy, R., 1996. Time of day and stress response to        Whitten, R.D., Jasnow, A.M., Albers, H.E., Martin-Schild, S., Zadina, J.E.,
    different stressors in experimental animals. Part I: golden hamster           Huhman, K.L., 2001. The effects of endomorphin-1 on conditioned
    (Mesocricetus auratus Waterhouse, 1839). J. Exp. Anim Sci. 38, 66 –           defeat in Syrian hamsters (Mesocricetus auratus). Brain Res. 914,
    76.                                                                           74 – 80.
You can also read