- Email: [email protected]
Social stress by repeated defeat: effects on social behaviour and emotionality
ELSEVIER
Behavioural Brain Research 62 (1994) 187-193
BEHAVIOURAL BRAIN RESEARCH
Research Report
Social stress by repeated defeat: effects on social behaviour and emotionality M.E. Albonetti*, F. Farabollini lstituto di Fisiologia Umana, Universita' degli Studi. Via del Later#to 8, 53100 Siena, ltah' Received 19 October 1993; revised 8 March 1994; accepted 8 March 1994
Abstract
The consequences of repeated defeat stress on social and non-social behaviours were assessed in male rats 24 h after the last defeat. Aggressive, defensive, introductory and affiliative items &both experimental animals and their opponents were recorded in a social behaviour test, while emotionality, exploration and general motor activity were scored in the Emergency, Hole-Board, and Elevated Plus-Maze tests. In addition to a dramatic loss of body weight, a selective inhibition of aggression was observed in the stressed experimental subjects, paralleled by decreased defence in their opponents. In the stressed animals, no change was found in other social and non-social behaviours; in particular, defence and emotionality were unaffected. This shows that, under our experimental conditions, the inhibition of aggression, which has often been reported to parallel an increase in defence after social and non-social aversire stimulation, was not dependent on a concomitant activation of a prevailing defensive motivational system, sustained by increased emotionality and fear. As the same result, namely a selective inhibition of aggression with no effect on defence, was obtained after exposure to a non-social stressor (restraint), the hypothesis is advanced that the threshold for stress-induced behavioural changes is lower for aggression than for any other behavioural and motivational system, including that leading to defence. The inhibition of aggression would therefore be a direct response to stress and not a by-product of the activation of a fear-based defensive system.
Key words. Aggression; Defence; Emotionality; Social stress
1. Introduction In male rodents, repeated defeat by aggressive same-sex conspecifics is reported to cause long-lasting inhibition of aggression [39,41], increased defence and submission [56,62] with a shift from active defence (such as retreat) to passive defence (e.g. freezing [40,45]), decreased interest in conspecifics [39,68], a decline in motor activity [39,40], and weight loss [15,68]. The behavioural consequences of repeated defeat are the same as those of uncontrollable, as opposed to controllable, physical stress, such as electric shock [44,75]. This supports the view that defeat acts as a stressor [77] which is uncontrollable on account of the unpredictable occurrence of conspecific aggression, and the failure to inhibit it [76]. The inability to control the stressor is thought to increase emotionality [4,40,47] and to induce or facilitate the occurrence of fear [60,75 ], which is defined as "the emotional state presumed * Corresponding author. Fax: (39) (577) 287205. 0166-4328/94,'$7.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0166-4328(94)00045-H
to be normally induced by exposure to potentially dangerous objects, individuals or situations" [20]. In the literature on stress-based animal models of depression, the term fear appears to be used as a synonym of anxieO' [2,37,55,70]. It has been suggested that fear activates defensive and submissive motivation(s) and inhibits other motivational systems, including that leading to offence [ 5,13,19,52]. Despite the supposed involvement of anxiety in stress-induced modifications of agonistic behaviour, the possible relation between stress, anxiety (as measured by standard behavioural tests) and alterations of agonism is not yet well established. In the present paper, male rats were investigated for the effects of stress, due to repeated defeat, on: (1) social behaviour, with both agonistic and non-agonistic items being considered for an exhaustive description which takes into account the integrated nature of behaviour [20]; and (2) different aspects of emotionality, as indicated by parameters of non-social behaviour validated and widely used as measures of anxiety [8,38]. This allowed us to
t88
M.E. Alhonetti and F, Farabollini Behavioural Brain Researvh 62 1994 187-19.~
explore the possible relationship between stress-induced modifications of social behaviour and anxiety, Behavioural tests were carried out 20 to 24 h after the last defeat; this allowed us to observe behavioural modifications which could reasonably be attributed to the whole stressing treatment rather than mainly to the last session, and to focus on persistent effects of stress.
2. Materials and methods
observation room. In order to avoid acoustic and pheromonal communication (see [3 ]) between stressed and control experimental males, the stressing procedure and handling never occurred contemporaneously. Twenty h after the last stressing or handling session, the Experimental subjects underwent a series of 3 non-social behaviour tcsts providing measures of anxiety, that is. the Emergency test [721 . the Hole-Board test [291 and the Elevated PlusMaze test [46]. Four h later, thc~ underwent the Social behaviour test.
2.1. Animals
2.3. Non-social behaviour tests
Male Wistar rats were used, supplied by S. Morini (S. Polo d'Enza, RE, Italy). When 84-86 days old (mean body weight 350.0 g, 6.37 S.E.), Experimental animals were individually housed with a neutral companion, namely a 3-month-old female (mean body weight 294.8 g, 5.12 S.E.) ovariectomized 3 weeks previously [69,25]. Opponents were housed as unisexual groups of 3 or 4. For both Experimental subjects and Opponents, perspex home-cages (36 x 20 x 18h cm) were used. All the animals were kept under a 12:12 h reversed light:dark cycle (light on at 8.00 pm), with food and water ad libitum. 2.2. Procedure
The Experimental animals were randomly assigned to the Stress (n = 10) or the Control (n = 9) group and allowed to familiarize with their neutral companions for 2 weeks. Subjects belonging to the Stress group were then introduced into the home-cage of an unfamiliar aggressive conspecifie (Resident) once a day for 7 consecutive days. Residents were 16-month-old Wistar males (mean b.wt. 538.1 g, 7.82 S.E.), sexually experienced and trained as winners in diadic social encounters, which had been living in isolation since they were 8 months old [4]. Each Stress group male was introduced into a Resident's homecage, left there until defeat occurred or up to a maximum time of 15 min, then returned to its own home-cage. The criterion for defeat was the Stress group male either assuming the extreme submissive posture, On-the-Back [ 31 ], 3 times in succession, maintaining it for at least 10 sec each time, with no introductory or aggressive behaviours directed to the Resident between consecutive On-the-Back bouts, or freezing for at least 90 s. Each Stress group subject was confronted with a different Resident every day. Confrontations took place in an observation room under dim red light; daily time of occurrence varied randomly between 9.30 am and 5.00 p.m,. Subjects belonging to the Control group were handled gently for less than 1 rain once a day for 7 consecutive days [9]. Handling occurred randomly between 9.30 am and 5.00 p.m. in the
Emergenc,.v test. Each Experimental animal was intro-
duced into a cubic black perspex box (10 cm sides) with one lateral wall missing, placed in a Hole-Board. Latency to emerge from the box (LAT. secl was measured. Max~mum time allowed was 3 mm. Hole-board test. After the Emergency box was removed. locomotor activity (Loc; number of squares crossed/, exploration (Head-Dipping, HD. frequency) and defecation (number of fecal boluses. BOLl were recorded during a 5-min period. The Inner Outer ratio ( I N / O U T : locomotion m inner squares locomotion in outer squares × 100) was also computed. Elevated Plus-Maze test. The number of entries and time spent in the open and closed arms were separately recorded for a 5-min period. Entries into open arms as a percentage of total entries into both open and closed arms (OP. N°,;, I and time spent in open arms as a percentage of total time spent in the two types of arms (OP, T'?,~,) were then computed. Direct measures of anxiety were considered to be LAT [ 10,11.17.22.46,54,63,73] and BOL [ 1.1t,21,32,33]. Inverse measures included OP. N".,, and OP, T°~, [34,42, 46.49.53.64.65] as well as I N / O U T [8,6112.4. Social behaviour rest
Each Experimental animal was introduced into a clean observation cage together with a naive unfamiliar conspecific (Opponent) of the s a m e sex. age, body weight and pre-experiment social and environmental experience [9]. Each Opponent was used only once. In order to reduce as much as possible aversive aspects of confrontation with environmental novelty [45 ], observation cages were similar m every respect to home-cages The encounters lasted 20 min. They were video-taped and the social behaviour of both the Experimental subject and the Opponent was then recorded on a PC with the aid of specialized software. The following behavioural parameters, based on behaviour lists by Grant and Mackintosh [31] and Fernandez-Espejo and Mir [26], were considered:
M.E. Albonetti and F. Farabollini/ Behavioural Brain Research 62 (1994) 187-193 -
Introductory behaviour (IN, frequency and duration, sec): includes Attending, Approaching, Investigating and A n o - G e n i t a l Sniffing;
-
Allo-Grooming (GR, freq. and dur.);
-
Aggressive A l l o - G r o o m i n g ( G G , freq. and dur.);
-
Threat (TH, freq.): includes Lateral Threat and Upright Offence, considered to be low-intensity aggressive behaviours;
- Attack (AK, freq.); -
Bite (BI, freq.);
- O n - T o p (OT, freq. and dur.). This high-intensity aggressive item signals the winning of an agonistic confrontation and is considered as the full d o m i n a n c e posture
-
in consolidated d o m i n a n c e / s u b m i s s i o n diadic
relationships [5]; Boxing (BO, freq.). This agonistic item often has ambivalent (i.e. offensive/defensive) significance [5,16];
-
-
low-intensity Defence (LDE, freq.): includes Defensive Sideways Posture, Upright Defence and Retreat; On-the-Back (OB, freq. and dur.). This high-intensity
189
Table 1 Non-social bchaviour - - Experimentalsubjects: means with S.E. of the parameters recorded 20 h after the repeated-defeat treatment (Stress group) or repeated handling (Control group) Behavioural parameters
Stressed n = I0
Control n=9
Emergency test (3 min) EAT
158.80 (37.22)
165.88 (49.76)
Hole-Board test (5 rain) Loc IN/OUT HD f BOL
1(/3.40 (15.4(I) 58.11 (27.63) 5.30 (1.12) 4.90 (0.73)
85.66 (15.58) 42.70 (9.45) 4.66 (1.08) 5.11 (0.91)
Elevated Plus-Maze test (5 min) OP, N"o 0.32 (0.1)8) OP, T" o 10.08 (4.92)
0.34 (0.07) 11).56(2.75)
f= frequency (number of acts/session) See text for parameter labels.
defensive item signals losing and is considered as the full submissive posture [5]. In the Experimental animals only, Self-Grooming (SG,
Control groups in the anxiety-related non-social behaviours recorded in the Emergency, the Hole-Board or the
dur.) was also recorded. The following scores were com-
Plus-Maze tests (Table 1).
puted:
With respect to social behaviour (Table 2), differences between the two groups were significant only for the parameters of aggression. While the high-intensity items
- total Aggression ( T A G , freq.; includes G G , T H and OT); - total Defence (TDE, freq.; includes L D and OB); - total A G freq. minus total D E freq. ( T A G - T D E ) ; - O T dur. minus OB dur. (OT-OB); -
Experimental's total D E divided by O p p o n e n t ' s total AG (TDE/Opponent's TAG).
2.5. Statistical analys& Experimental subjects. The Stress and Control groups were compared with the two-tailed Student's t-test for i n d e p e n d e n t samples [ 59] for each parameter of non-social and social behaviour, as well as for body weight increase during the 7-day experimental period. Pearson's correlation coefficient [59] between each measure of anxiety and each parameter of social behaviour was then computed separately for the two groups. Opponents. For each parameter of social behaviour, Student's t-test was used to compare the O p p o n e n t s confronted with the Control group subjects (Opponents-tocontrol) with the O p p o n e n t s confronted with the Stress group subjects (Opponents-to-stressed).
3 .
Table 2 Social behaviour - - Experimentalsubjects: mcans with S.E. of the parameters directed to the Opponents during the 20-mintest, 24 h after the repeated-defeat treatment (Stress group) or repeated handling (Control group) Behavioural parameters
Stress group n = 10
Control group n=9
IN f IN d GR f GR d GG f GG d TH OT f OT d TAG f LDE OB f OB d TDE f (TAG-TDE) f (OT-OB) d TDE/Opponent's TAG SG d
85.50 (5.61) 83.40 (9.02) 3.81 (1.07) 3.47 (1.25) 4.90 (1.531 20.43 (10.331 2.17 (0.75) 0.60 (0.22) 3.72 (1.(181 7.811(1.621 15.70 (5.11) 2.70 (2.38) 24.12 (9.16) 18.41) (7.08) - 10.60 (3.07) - 22.50 (19.881 2.23 (0.89) 92.16 (14.841
96.33 (8.50) 96.68 (12.54) 2.01 (0.58) 1.17 (0.36) 4.55 (1.17) 21.94 (9.59) 10.66 (4.17)** 6.33 (2.87)* 49.11 (15.94)* 21.55 (3.45)** 19.77 (8.07) 1.88 (1.1l) 8.48 (5.14) 21.66 (9.08) 0.51 (2.15)* 40.80 (20.33)* 1.75 (0.54) 59.88 (9.0(/)
R e s u l t s
Experimental subjects. N o significant difference was found between subjects belonging to the Stress and the
f= frequency (number of acts/session): d = duration (seconds). * Student's t, P< = 0.05; ** Student's t, P< = 0.01. See text for parameter labels.
190
M . E . A[bonetti and F. barahollini
Behavioural Brain Research 62 1994
Attack and Bite were never observed, both low-intensity Threat (TH; t = 3.21, nl = 10, n2 = 9, P = 0.005) and highintensity On-Top (OT; freq., t=2.10, P = 0 . 0 5 ; dur., t= 2.38, P=0.02), as well as total aggression (TAG, t = 2,91, P = 0.01), were significantly lower in stressed than in control animals. Aggression was markedly less frequent than defence in the stressed subjects, but such a difference was not found in the controls. Thus the balance of total aggression and total defence, as measured by (TAG-TDE), was significantly different between the Stress and Control groups (t= 2.31, P=0.03). Similarly, stressed animals spent much less time in the extreme aggressive posture, On-Top, than in the extreme defensive posture, On-Back, whereas the reverse was true for control subjects, and this difference was again significant (t = 2.22, P = 0.04). No significant correlation was found between any parameter of social behaviour and any measure of emotionality (OP, N°/~,; OP, T%I,; LAT; I N / O U T ; BOL). After the 7-day experimental period, stressed animals had lost weight, the mean decrease being 2.20 g (4.65 S.E.), whereas control animals gained an average of 20.77 g (10.41 SE). This difference was significant (t = 2.08,
?=0.05). Opponents. Opponents spent significantly more time investigating the stressed animals than the controls (t = 2.08, P = 0.05; Table 3), but directed to both groups similar amounts of aggression. However, striking differences emerged in defence; both low-intensity (LDE, t= 2.43, P = 0 . 0 2 ) and high-intensity (On-the-Back, OB; freq., t = 2.15, P = 0.04; dur., t = 2.33, P = 0,03)defensive param-
Table 3 Social behaviour - - Opponents: means with S.E. of the parameters directed to the Experimental animals during the 20-min test
Behavioural parameters
Opponentsto-stressed n = 10
IN f IN d GR f GR d GG f GG d TH OT f OT d TAG f LDE OB f OB d TDE f (TAG-TDE) f (OT-OB) d
79.80 (12.55) 92.35 (11.83) 4.12 (1.13) 3.30(1.31) 4.43 (1.51) 14.97 (6.37) 3.90 (1.32) 2.70 (2.38) 25.94 (10.83) 11.00 (3.08) 4.t(I (1.36) 0.75 (0.28) 7.30 (3.89) 4.80 (1.60) 6.20 (4.06) 18.64 (9.24)
* Student's t, 0 . 0 5 < P < 0 . 1 : olhers as in Table 2.
Opponentsto-control n=9 64.33 (t3.28) 45.31 (10.47)* 4.87 (0.96) 4.56(1.08) 4.90 (1.38) 16.18 (7.63) 5.00 (1.46) 1.77 (1.02) 8.96 (5.63) 11.66 (2.64) 16.67 (5.23)* 6.36 (2.72)* 52.98 ( 12.22)* 23.00 (7.69)* - 11.33 (8.63)' - 4 4 . 0 2 (14.86)*
1~ ~- 19/
eters were lower in the Opponent~ confronted with stressed subjects (Opponents-to-stressed j than in those confronted with control animals (Opponents-to-control). The Opponents-to-stressed were indeed more aggresstve than defensive, as suggested by the positive values of both {TAGTDE) and (OT-OB), whereas the reverse was true for Opponents-to-control. which had negative values for the same scores. The difference belween the two groups of Opponents was significant ~br (OT-OBI 11=-2.21. P = 0.041, and marginally. ,~) t\,~r ( I ' A G - T D E ) (t1.7t, P = 0.07 }.
4. Discussion In male rats. repeated defeat, expected to act as a sociat stressor [77] as indirectly confirmed b y the average decrease of body weight, caused a selective and specific inhibition of aggression in subsequent encounters with unfamiliar naive conspecifics (Opponents). Under our experimental conditions, no additional effect was found on other social beh aviours, including defence, nor on anxietyrelated non-social behaviours. Consistent with the decreased offence shown by the stressed animals, reduced defence was observed in their Opponents. Our findings are partially at variance with the pertinent literature on rats and other rodents, which reports repeated defeat to: (1) inhibit not only offence [41], but also other social behaviours such as the introductory items [40,68]; (2) inhibit non-social behaviours, such as exploranon and locomotor activity [40]; and (3) increase defence [62.67,68]. Behaviourat inhibition is indeed considered to be a consequence of the activation of defence. which is presumed to overcome any other motivational system E75]. It is however interesting that Vogel et al. [71 ], in a review of behavioural alterations typical of human depression which can be sinmlated in animals through exposure to uncontrollable stressors, mentioned "irritability" and "hostility". which may be considered as items related to aggression rather than defence. In the present work. the absence of any increase of defence suggests that the inhibition of offence was not due to the concomitant activation of a prevailing defensive motivational system. Tha( offence and defence may not alwa3 s act as strictly linked and interdependent behavioural systems [9] is consistent with the long-accepted view that offensive and defensive behaviours have different remote and proximate causation [51] and arc served by different neural substrates [5,18,23]. This belief appears to be supported by recent pharmacological investigations which characterized a number of drugs (such as fluprazine/acting selectively on aggresmon with no effect on defence [30,48,52].
M.E. Albonetti and F. Farabollini / Behavioural Brain Research 62 (1994) 187-193
Recently, Siegfried at al. [58], working on mice, proposed the hypothesis that nociceptive stimulation, induced by being attacked and bitten in a social conflict, may be responsible for the activation of a defensive motivational system. In the present experiment, activation of defensive motivation might have been prevented by a lack of nociceptive stimulation of sufficient intensity. As a matter of fact, high-intensity aggressive items which might possibly have caused physical damage (e.g. Attack, Bite) were never observed, presumably due to the use of the low-aggression Wistar strain [62]. However, this absence of high-intensity aggression did not prevent the occurrence of defeat both during the stress-induction phase and at testing. In the present experiment, repeated defeat failed to increase not only defence, but also anxiety. Social interaction, considered to be negatively correlated with anxiety [28,38], was unaffected by the social stress: the stressed rats did appear to be interested in the unfamiliar Opponents, and approached, explored and groomed them as much as did the control animals. In addition, no change was found in behavioural parameters widely assumed to be standard measures of anxiety (percentage entries and time spent in the open arms of an Elevated Plus-Maze; latency to exit from a box; defecation and internal vs. external locomotion in the Hole-Board; see Methods for references). Even Self-Grooming, which may be a substitutive activity indicating embarrassment and uncertainty and which has been viewed as a hint of the animal's emotionality [25,53], was similar in the stressed and control rats. Increased anxiety in the Plus-Maze has been found immediately after a single defeat in male rats [36] and mice [53]. Taken together, these and our results suggest either that defeat does increase anxiety but that this is a shortterm effect no longer appreciable 24 h afterwards, or that anxiety, increased by a single defeat, returns to basal levels when defeat is repeated. Aside, attention has been recently called on the possibility that different anxiety parameters, recorded under different testing conditions, measure different aspects, or kinds, of anxiety [27]. Thus, the possibility cannot be ruled out that anxiety tests we have carried out might have been inadequate to reveal anxiety supposedly induced by social stress. The hypothesis that the stressed rats were not especially anxious since their low aggression did not provoke retaliation, thus reducing the amount of aggression by the Opponents (see [57]), is disproved by the finding that the Opponents directed similar amounts of aggression to both the stressed and control Experimental animals. Therefore, in the stressed rats the reduction of aggression given, namely that directed to the Opponents, did not appear to be a strategic tool effective in reducing the aggression received. In other words, it did not act as a signal [76] able to modify the Opponents'behaviour. From an adaptive
191
point of view, decreased aggression reduced energy expenditure but did not prevent the possibility of being injured during agonistic encounters. More importantly, in species like the rat which are characterized by a hierarchical social organization, with aggression playing a crucial role in gaining high rank [6,7,12,62], inhibited aggression condemns an animal to a low ranking, which is detrimental to individual and inclusive fitness [ 14]. Interestingly, the Opponents investigated the stressed rats for a significantly longer time than the control animals; this was presumably aimed at the detection of olfactory cues revealing the condition of stress [24,43,66]. In a previous study [9], we obtained selective inhibition of offence, but no effect on defence, 24 h after both single and repeated exposure to a mild physical stressor, restraint [35]. Offence thus appears to have a greater sensitivity to aversive stimulation and a lower threshold for stress-induced changes than other behavioural systems, whereas stronger stressors are presumably required to activate and maintain defence and produce a persistent increase in anxiety. Interestingly, the nature of the stressor, either social or physical, does not seem to affect the behavioural responses to it. Differences in the ability ofstressors to induce behavioural modifications seem to be due to their quantitative rather than qualitative attributes--in other words, they seem to be a matter of how much rather than how (see [40]).
Acknowledgements We wish to thank Dr. P. Christie for revision of the English.
References [1] Abel, E.L., Behavior and corticosteroid response of maudsley reactive and nonreactive rats in the open field and forced swimming test, Physiol. Behav., 50 (1991) 151-153. [2] Abel, E i . , Altman, H.J. and Commissaris, R.L., Maudsley reactive and nonreactive rats in the forced swim test: comparison in fresh water and soiled water, Physiol. Behav., 52 (1992) 1117-1119. [3] Abel, E.L. and Bilitzke, P.J., A possible alarm substance in the forced swimming test, Physiol. Behav., 48 (1990) 233-239. [4] Adams, D.B., The relation &scent-marking, olfactory investigation and specific postures in the isolation induced fighting of rats, Behaviour, 56 (1976) 286-297. [5] Adams, D.B., Motivational systems ofagonistic behavior in muroid rodents: a comparative review and neural model, Aggress. Behav., 6 (1980) 295-346. [6] Adams, D.B., Boice, R., Development of dominance in domestic rats in laboratory and seminatural environments, Behav. Proc_ 19 (1989) 127-142. [7] Albert, D.J., Dyson, E.M., Walsh, M.L. and Wong, R., Defensive aggression and testosterone-dependent intermale social aggression are each elicited by food competition, Ph~wiol. Behav., 43 (1988) 21-28.
192
M.E. Albonetti and F Farabollilzi, Behavhmral Brain Research 62 19941 tb¢7-!9;
[8] Albonetti, M.E. and Faraboltini, F., Behavioural responses to single and repeated restraint in male and female rats, Behav. Proc.. 28 (1992) 97-110. [9] Albonetti, M.E. and Faraboltini, F., Efl?cts of single and repeated restraint on the social behavior of male rats, Physiol. Behav.. 53 11993) 937-942. [10] Anderson, E.E., The sex hormones and emotional behavior: IlL The effect of castration upon timidity in male and female rats, J Genet. Psychol., 56 (1940) 169-174. [ 11 ] Archer, J., Rodent sex differences in emotional and related behavior, Behav. Biol., 14 (1975) 451-479. [12] Benus, R.F., Bohus, B., Koolhaas, J.M. and van Oortmerssen, G.A., Behavioural strategies of aggressive and non-aggressive male mice in active shock avoidance, Behav. Proc., 20 (1989) 1-12. [13] Blanchard D.C. and Btanchard, R.J., Affect and aggression: an animal model applied to human behavior, Adv. St. Aggress'.. 1 (1984) 1-62. [14] Blanchard D.C. and Blanchard, R.J., Behavioral correlates of chronic dominance-subordination relationships of male rats in a seminatural situation, Neurosci. Biobehav. Rev., 14 (1990) 455-462. [15] Blanchard D.C., Cholvanich, P., Blanchard, R.J., Clow, D.W., Hammer Jr., R.P., Rowlett, J.K. and Bardo, M.T., Serotonin, but not dopamine, metabolites are increased in selected brain regions of subordinate male rats in a colony environment, Brain Res., 568 (1991) 61-66. [16] Blanchard, R.J,, Flannelly K. and Blanchard D.C., Life-span studies of dominance and aggression in established colonies of laboratory rats, Physiol. Behav., 43 (1988) 1-7. [17] Blanchard, R.J., Kelley, M J . and Blanchard D.C., Defensive reactions and exploratory behavior, J. Comp. Physiol. P.tvchoL, 87 (1974) 1129-1133. [18] Blanchard, R.J., Kleinschmidt, C.K., Flannelly K. and Blanchard D.C., Fear and aggression in the rat, Aggress. Behav.. 10 (1984) 309-315. [19] Bolles, R.C., Species-specific defense reactions and avoidance learning, Psychol. Rev., 77 (1970) 32-48. [20] Brain, P.F., Kusumorini N. and Benton. D., "Anxiety" in laboratory rodents: a brief review of some recent behavioural development, Behav. Proc., 25 (1991) 71-80. [21] Broadhurst, P.L., Determinants of cmotionality in the rat: 1. situational factors, Br. J. Psychol., 48 (1957) 1-12. [22] Chopin, P. and Briley, M., Animal models of anxiety: the effect of compounds that modify 5-HT neurotrasmission, Trends Pharmacol. Sci., 8 (1987) 383-388. [23] Depaulis, A. and Vergnes, M., Relationship between mousekilling and conspecific aggression in the male rat, Aggress. Behav., 9 (1983) 259-268. [24] Dua, J.K. and Dobson, M.J., Role of olfactory cues in acquisition and extinction of avoidance, J. Exp. Psvchol., 103 (1974) 461-465. [25] Farabollini, F., Hole, D.R. and Wilson, C.A., Behavioral effects in adulthood of serotonin depletion by p-chlorophenylalanine given neonatal!y to male rats, bit. J. Neurosci., 41 (1988) 187-199. [26] Fernandez-Espejo, E. and Mir, D., Ethological analysis of the male rat's socioagonistic behavior in a resident-intruder paradigm, Aggress. Behav., 16 (1990) 41-55. [27] File, S.E., Behavioural detection of anxiolytic action. In J.M. EItiott, D.J. Heal and C.A. Marsden (Eds.), ExperimentalApproaches to Anxiety and Depression, Wiley and Sons, 1993, pp. 25-44. [281 File, S.E. and Hide, J.R., Can social interaction be used to measure anxiety? Br. J. Pharmacol., 62 (1978) 19-24. [29] File, S.E., Wardill, A.G., The reliability of the holeboard apparatus, Psychopharmacology, 44 (1975) 47-51. [30] Flannelly, K.J., Muraoka, M.Y., Blanchard, D.C. and Blanchard, R.J., Specific anti-aggressive effects of fluprazine hydrochloride, Psychopharmacology (Berl.), 87 (1985)86-89. [31] Grant, E.C. and Mackintosh, J.H., A comparison of the social postures of some common laboratory rodents, Behaviour, 21 (1963) 246-259.
[ 32 [ Gray, J.A.. Sex differences m the emotional behavior of laboratory rodents: comment. Br. 1. Ps~:ho[.. 70 (19791 3 5 - 3 6 [33] Hall. C S . . Emotional behavior )n the rat. I. Defecation and urinanon as measures of individual differences in c~notionalit~. 1 Comp. Psychol,, 18 11934) 385-403~ [34 t H andley, S.L. and Mithani. S.. Eit~2cts of atpha-adrenoceptor agonists and antagonists m a mazc
[40] Kudryavtseva, N.N.. Bakshtanovska)a, 1.'~. and Kor~akina, k 'A~,. Social model of depression in mice of C57BL.'6J strata. Pharmacol Biochem Behav.. 38 11991) 315-321i [41 ] Leshner. ~.I.. The role of hormones m the control of submissiveness. In P.F. Brain and D Benton reds.), Multidisciplinao" "lp!)rmwhes' to .4ggression Research. Elsevier North-Holland. Amsterdam. 1981. pp. 309-322. [42] Lister. R.G.. The use of a plus-maze to measure anx~et)in the mouse. Ps3"chopharmacology {Berl. ,;, 92 f 19871 180-185 [43] Macka~-Sim. A. and Laing, D.G,, Discrimination of odors from stressed rats by non-stressed rats. Physiol. Behav.. 24 (1980) 699704. [44] Maier. S,F.. Drugan. R , Gram J.~A .. 1tyson. R., Macl,ennan, A.J.. Mo3e. T.. Madden. J. and Barchas~ J.D., Learned helplessness. pare inhibition and the endogenou~ opmtes, In M.D. Zeiler and P. Harzem (Eds.). Advances b7 An#rra~ Behavior. Vol. 3. Wiley. New York. 1983. pp. 275-323. [45] Misslin R. and Cigrang. M.. Doc~ ncophobia neccssaril5 impl 5 fear or anxiety? Behav Proc , 12 (1986} 45-50. [46] Montgomery, K.C.. The relation between fiear induced by novelty stimulation and exploratory behaviour, ,L Comp. Physiol. Ptvchol., 48 11955) 254-260 [47] Musty. R,E.. Jordan. M.P. and Lcm)x. R.H.. Criterion for learned helplessness in the rat: a redefinition. Pharmacol, Bioehem. Behav. 36 11990t 739-744 [48] Olivier. B., Mos. J. and Hartog. J., Eltoprazine: a specific annaggressive drug ~serenic) in animal aggression models. Aggress. Behav.. 15 f1989) 9 5 - 1 1 2 [49] Pellm~,. 51.. Chopin. P.. File. S.I. and Brile~. M.. Validation of open:closed arm entries m an elevated plus maze as a measure ol anxiety in the rat. J. Neurosci Meth.. 14 11985] 149-t67. [50] Ptaznik. ~\.. Stefanski. R. and Kostowski. W.. Restraint stressinduced changes in saccharin weference: the effects of antidepresslvc treatment and diazepam. PhatvnacoL Biochem Behav.. 33 (1989) 755-759. [5t] Rasa. O.A.. Aggression: appetite or aversion? A ggress. Behav.. 2 1976J 213-222. [52] Rawleigh. J.M. and Kemble. E.D.. 'Iest-specific effects ot EG-7142 on isolation-induced aggression m mice. Pharmacol. Biochet~. Behav.. 42 ~1992) 317-321 [531 Rodgers. R.J. and Cole. J . ( . , Anxlet) enhancement m the murmc elevated plus maze b) immediate prior exposure to social stressors. PhvsioL Behav.. 53 (19931 383-3 ~
M.E. Albonetti and F. Farabollini / Behavioural Bra#t Research 62 (1994) 187-193 [54] Rodriguez Echandia, E.L., Gonzalez, A.S., Cabrera, R. and Fracchia, L.N., A further analysis of behavioral and endocrine effects of unpredictable chronic stress, Physiol. Behav., 43 (1988) 789-795. [55] Russel, P.A., Open field defecation in rats: relationships with body weight and basal defecation level, Br. J. Psvchol., 64 (1973) 109114. [56] Scholtens, J., van Haaren, F. and van de Poll, N.E., Effects of losing and testosterone upon subsequent behavior in male and female $3 (Tryon Maze Dull) rats, Aggress. Behav., 14 (1988) 371387. [57] Senar, J.C., Agonistic communication in social species: what is communicatcd? Beha~im~r, 112 (1990) 270-282. [58] Siegfried, B., Frischknecht H. and Nunes de Souza, R.L., An ethological model for the study of activation and interaction of pain, memory and defensive systems in the attacked mouse. Role of endogenous opioids, Neurosci. Biobehav. Rev., 14 (1990) 481-490. [59] Snedecor, G.W. and Cochran, W.G., Statistical Methods, Iowa State Univ. Press, 1980, 5(17 pp. [60] Steenbergen, H.L., Heinsbroek, R.P.W, Van Haaren, F. and Van de Poll, N.E., Sex-dependent effects of inescapable shock administration on behavior and subsequent escape performance in rats, Physiol. Behav., 45 (1989) 781-787. [61] Stefanski, R., Palejko, W., Kostowski, W. and Plaznik, A., The comparison of benzodiazepine derivatives and serotonergic agonists and antagonists in two animal models of anxiety, Neuropharmacoh~gy, 3 l (1992) 1251 - 1258. [62] Swanson, H.H., Sex differences in behavioral consequences of defeat in the rat arc not organized by testosterone during early development, Aggress. Behav., 16 (1990) 341-344. [63] Takahashi, L.K., Kalin, N.H., Vanden Burgt, J.A. and Sherman, J. E., Corticotropin-relcasing factor modulates defensive withdrawal and explorator~ behavior in rats, Behav. Neurosci., 103 (1989) 648654. [64] Taukulis, H.K. and Goggin, C.E., Diazepam-stress interactions in the rat: effects on autoanalgesia and a plus-maze model of anxiety, Behav. Neural Biol., 53 (1990)205-216. [65] Trcit, D., Menard, J. and Royan. C., Anxiogenic stimuli in the
[66] [67]
[68]
[69]
[70]
[71]
[72] [73]
[74]
[75]
[76] [77]
193
elevated plus-maze, Pharmacol. Biochem. Behav., 44 (1993)463469. Valenta, J.G. and Rigby, M.K., Discrimination of the odor of stressed rats, Science, 161 (1968) 599-601. van de Poll, N.E., de Jonge, F., van Oyen, H.G. and van Pelt, J., Aggressive behavior in rats: effects of winning and losing on subsequent aggressive interactions, Behav. Pro~., 7 (1982) 143-155. van de Poll, N.E., Smeets, J., van Oyen, H.G. and van der Zwan, S.M., Behavioral Consequences of agonistic experience in rats: sex differences and the effects of testosterone, .I. Comp. Physiol. Psychol., 96 (1982) 893-903. van de Poll, N.E., Swanson, H.H. and van Oyen, H.G., Gonadal hormones and sex differences in aggression in rats. In P.F. Brain, P.F. and D. Benton, D. (Eds.), The Biology ¢!fAg,~,ression, Sijthoff and Noordhoff. Amsterdam, 1981, pp. 244-252. van Dijken, H.H., Mos, J., van der Heyden, J.A. and Tilders, F.J., Characterization of stress-induced long-term behavioural changes in rats: evidence in favor of anxiety, Physiol. Behav., 52 (1992) 945-951. Vogel, G., Neill, D., Hagler, M. and Kors, D., A new animal model of endogenous depression: a summary of present findings, Neurosci. Biobehav. Rev., 14 (1990) 85-91. Wallace, R.J., Latency measures indicate new place neophobia in Rattus species, Behav. Proc., 17 (1988) 63-67. Welker, W.I., Escape, exploratory, and food-seeking responses of rats in novel situation, J. Comp. Physhd. Psychol., 52 (1959) 106111. Wiepkema, P.R., Koolhaas, J.M. and Olivier-Aardema, R., Adaptive aspects of neuronal elements in agonistic behavior, Prog. Brain Res., 53 (1980) 369-384. Williams, J.L. and Lierle, D.M., Effects of repeated defeat by a dominant conspecific on subsequent pain sensitivity, open field activity, and escape learning, Anita. Learn. Behav., 16 (1988) 477485. Wilson, E.O., Sociobiology: the New Synthesi.~, Harvard University Press, Cambridge, 1975, pp. 671. Zayan, R., The specificity of social stress, Behav. Proc., 25 (1991) 81-93.
Behavioural Brain Research 62 (1994) 187-193
BEHAVIOURAL BRAIN RESEARCH
Research Report
Social stress by repeated defeat: effects on social behaviour and emotionality M.E. Albonetti*, F. Farabollini lstituto di Fisiologia Umana, Universita' degli Studi. Via del Later#to 8, 53100 Siena, ltah' Received 19 October 1993; revised 8 March 1994; accepted 8 March 1994
Abstract
The consequences of repeated defeat stress on social and non-social behaviours were assessed in male rats 24 h after the last defeat. Aggressive, defensive, introductory and affiliative items &both experimental animals and their opponents were recorded in a social behaviour test, while emotionality, exploration and general motor activity were scored in the Emergency, Hole-Board, and Elevated Plus-Maze tests. In addition to a dramatic loss of body weight, a selective inhibition of aggression was observed in the stressed experimental subjects, paralleled by decreased defence in their opponents. In the stressed animals, no change was found in other social and non-social behaviours; in particular, defence and emotionality were unaffected. This shows that, under our experimental conditions, the inhibition of aggression, which has often been reported to parallel an increase in defence after social and non-social aversire stimulation, was not dependent on a concomitant activation of a prevailing defensive motivational system, sustained by increased emotionality and fear. As the same result, namely a selective inhibition of aggression with no effect on defence, was obtained after exposure to a non-social stressor (restraint), the hypothesis is advanced that the threshold for stress-induced behavioural changes is lower for aggression than for any other behavioural and motivational system, including that leading to defence. The inhibition of aggression would therefore be a direct response to stress and not a by-product of the activation of a fear-based defensive system.
Key words. Aggression; Defence; Emotionality; Social stress
1. Introduction In male rodents, repeated defeat by aggressive same-sex conspecifics is reported to cause long-lasting inhibition of aggression [39,41], increased defence and submission [56,62] with a shift from active defence (such as retreat) to passive defence (e.g. freezing [40,45]), decreased interest in conspecifics [39,68], a decline in motor activity [39,40], and weight loss [15,68]. The behavioural consequences of repeated defeat are the same as those of uncontrollable, as opposed to controllable, physical stress, such as electric shock [44,75]. This supports the view that defeat acts as a stressor [77] which is uncontrollable on account of the unpredictable occurrence of conspecific aggression, and the failure to inhibit it [76]. The inability to control the stressor is thought to increase emotionality [4,40,47] and to induce or facilitate the occurrence of fear [60,75 ], which is defined as "the emotional state presumed * Corresponding author. Fax: (39) (577) 287205. 0166-4328/94,'$7.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0166-4328(94)00045-H
to be normally induced by exposure to potentially dangerous objects, individuals or situations" [20]. In the literature on stress-based animal models of depression, the term fear appears to be used as a synonym of anxieO' [2,37,55,70]. It has been suggested that fear activates defensive and submissive motivation(s) and inhibits other motivational systems, including that leading to offence [ 5,13,19,52]. Despite the supposed involvement of anxiety in stress-induced modifications of agonistic behaviour, the possible relation between stress, anxiety (as measured by standard behavioural tests) and alterations of agonism is not yet well established. In the present paper, male rats were investigated for the effects of stress, due to repeated defeat, on: (1) social behaviour, with both agonistic and non-agonistic items being considered for an exhaustive description which takes into account the integrated nature of behaviour [20]; and (2) different aspects of emotionality, as indicated by parameters of non-social behaviour validated and widely used as measures of anxiety [8,38]. This allowed us to
t88
M.E. Alhonetti and F, Farabollini Behavioural Brain Researvh 62 1994 187-19.~
explore the possible relationship between stress-induced modifications of social behaviour and anxiety, Behavioural tests were carried out 20 to 24 h after the last defeat; this allowed us to observe behavioural modifications which could reasonably be attributed to the whole stressing treatment rather than mainly to the last session, and to focus on persistent effects of stress.
2. Materials and methods
observation room. In order to avoid acoustic and pheromonal communication (see [3 ]) between stressed and control experimental males, the stressing procedure and handling never occurred contemporaneously. Twenty h after the last stressing or handling session, the Experimental subjects underwent a series of 3 non-social behaviour tcsts providing measures of anxiety, that is. the Emergency test [721 . the Hole-Board test [291 and the Elevated PlusMaze test [46]. Four h later, thc~ underwent the Social behaviour test.
2.1. Animals
2.3. Non-social behaviour tests
Male Wistar rats were used, supplied by S. Morini (S. Polo d'Enza, RE, Italy). When 84-86 days old (mean body weight 350.0 g, 6.37 S.E.), Experimental animals were individually housed with a neutral companion, namely a 3-month-old female (mean body weight 294.8 g, 5.12 S.E.) ovariectomized 3 weeks previously [69,25]. Opponents were housed as unisexual groups of 3 or 4. For both Experimental subjects and Opponents, perspex home-cages (36 x 20 x 18h cm) were used. All the animals were kept under a 12:12 h reversed light:dark cycle (light on at 8.00 pm), with food and water ad libitum. 2.2. Procedure
The Experimental animals were randomly assigned to the Stress (n = 10) or the Control (n = 9) group and allowed to familiarize with their neutral companions for 2 weeks. Subjects belonging to the Stress group were then introduced into the home-cage of an unfamiliar aggressive conspecifie (Resident) once a day for 7 consecutive days. Residents were 16-month-old Wistar males (mean b.wt. 538.1 g, 7.82 S.E.), sexually experienced and trained as winners in diadic social encounters, which had been living in isolation since they were 8 months old [4]. Each Stress group male was introduced into a Resident's homecage, left there until defeat occurred or up to a maximum time of 15 min, then returned to its own home-cage. The criterion for defeat was the Stress group male either assuming the extreme submissive posture, On-the-Back [ 31 ], 3 times in succession, maintaining it for at least 10 sec each time, with no introductory or aggressive behaviours directed to the Resident between consecutive On-the-Back bouts, or freezing for at least 90 s. Each Stress group subject was confronted with a different Resident every day. Confrontations took place in an observation room under dim red light; daily time of occurrence varied randomly between 9.30 am and 5.00 p.m,. Subjects belonging to the Control group were handled gently for less than 1 rain once a day for 7 consecutive days [9]. Handling occurred randomly between 9.30 am and 5.00 p.m. in the
Emergenc,.v test. Each Experimental animal was intro-
duced into a cubic black perspex box (10 cm sides) with one lateral wall missing, placed in a Hole-Board. Latency to emerge from the box (LAT. secl was measured. Max~mum time allowed was 3 mm. Hole-board test. After the Emergency box was removed. locomotor activity (Loc; number of squares crossed/, exploration (Head-Dipping, HD. frequency) and defecation (number of fecal boluses. BOLl were recorded during a 5-min period. The Inner Outer ratio ( I N / O U T : locomotion m inner squares locomotion in outer squares × 100) was also computed. Elevated Plus-Maze test. The number of entries and time spent in the open and closed arms were separately recorded for a 5-min period. Entries into open arms as a percentage of total entries into both open and closed arms (OP. N°,;, I and time spent in open arms as a percentage of total time spent in the two types of arms (OP, T'?,~,) were then computed. Direct measures of anxiety were considered to be LAT [ 10,11.17.22.46,54,63,73] and BOL [ 1.1t,21,32,33]. Inverse measures included OP. N".,, and OP, T°~, [34,42, 46.49.53.64.65] as well as I N / O U T [8,6112.4. Social behaviour rest
Each Experimental animal was introduced into a clean observation cage together with a naive unfamiliar conspecific (Opponent) of the s a m e sex. age, body weight and pre-experiment social and environmental experience [9]. Each Opponent was used only once. In order to reduce as much as possible aversive aspects of confrontation with environmental novelty [45 ], observation cages were similar m every respect to home-cages The encounters lasted 20 min. They were video-taped and the social behaviour of both the Experimental subject and the Opponent was then recorded on a PC with the aid of specialized software. The following behavioural parameters, based on behaviour lists by Grant and Mackintosh [31] and Fernandez-Espejo and Mir [26], were considered:
M.E. Albonetti and F. Farabollini/ Behavioural Brain Research 62 (1994) 187-193 -
Introductory behaviour (IN, frequency and duration, sec): includes Attending, Approaching, Investigating and A n o - G e n i t a l Sniffing;
-
Allo-Grooming (GR, freq. and dur.);
-
Aggressive A l l o - G r o o m i n g ( G G , freq. and dur.);
-
Threat (TH, freq.): includes Lateral Threat and Upright Offence, considered to be low-intensity aggressive behaviours;
- Attack (AK, freq.); -
Bite (BI, freq.);
- O n - T o p (OT, freq. and dur.). This high-intensity aggressive item signals the winning of an agonistic confrontation and is considered as the full d o m i n a n c e posture
-
in consolidated d o m i n a n c e / s u b m i s s i o n diadic
relationships [5]; Boxing (BO, freq.). This agonistic item often has ambivalent (i.e. offensive/defensive) significance [5,16];
-
-
low-intensity Defence (LDE, freq.): includes Defensive Sideways Posture, Upright Defence and Retreat; On-the-Back (OB, freq. and dur.). This high-intensity
189
Table 1 Non-social bchaviour - - Experimentalsubjects: means with S.E. of the parameters recorded 20 h after the repeated-defeat treatment (Stress group) or repeated handling (Control group) Behavioural parameters
Stressed n = I0
Control n=9
Emergency test (3 min) EAT
158.80 (37.22)
165.88 (49.76)
Hole-Board test (5 rain) Loc IN/OUT HD f BOL
1(/3.40 (15.4(I) 58.11 (27.63) 5.30 (1.12) 4.90 (0.73)
85.66 (15.58) 42.70 (9.45) 4.66 (1.08) 5.11 (0.91)
Elevated Plus-Maze test (5 min) OP, N"o 0.32 (0.1)8) OP, T" o 10.08 (4.92)
0.34 (0.07) 11).56(2.75)
f= frequency (number of acts/session) See text for parameter labels.
defensive item signals losing and is considered as the full submissive posture [5]. In the Experimental animals only, Self-Grooming (SG,
Control groups in the anxiety-related non-social behaviours recorded in the Emergency, the Hole-Board or the
dur.) was also recorded. The following scores were com-
Plus-Maze tests (Table 1).
puted:
With respect to social behaviour (Table 2), differences between the two groups were significant only for the parameters of aggression. While the high-intensity items
- total Aggression ( T A G , freq.; includes G G , T H and OT); - total Defence (TDE, freq.; includes L D and OB); - total A G freq. minus total D E freq. ( T A G - T D E ) ; - O T dur. minus OB dur. (OT-OB); -
Experimental's total D E divided by O p p o n e n t ' s total AG (TDE/Opponent's TAG).
2.5. Statistical analys& Experimental subjects. The Stress and Control groups were compared with the two-tailed Student's t-test for i n d e p e n d e n t samples [ 59] for each parameter of non-social and social behaviour, as well as for body weight increase during the 7-day experimental period. Pearson's correlation coefficient [59] between each measure of anxiety and each parameter of social behaviour was then computed separately for the two groups. Opponents. For each parameter of social behaviour, Student's t-test was used to compare the O p p o n e n t s confronted with the Control group subjects (Opponents-tocontrol) with the O p p o n e n t s confronted with the Stress group subjects (Opponents-to-stressed).
3 .
Table 2 Social behaviour - - Experimentalsubjects: mcans with S.E. of the parameters directed to the Opponents during the 20-mintest, 24 h after the repeated-defeat treatment (Stress group) or repeated handling (Control group) Behavioural parameters
Stress group n = 10
Control group n=9
IN f IN d GR f GR d GG f GG d TH OT f OT d TAG f LDE OB f OB d TDE f (TAG-TDE) f (OT-OB) d TDE/Opponent's TAG SG d
85.50 (5.61) 83.40 (9.02) 3.81 (1.07) 3.47 (1.25) 4.90 (1.531 20.43 (10.331 2.17 (0.75) 0.60 (0.22) 3.72 (1.(181 7.811(1.621 15.70 (5.11) 2.70 (2.38) 24.12 (9.16) 18.41) (7.08) - 10.60 (3.07) - 22.50 (19.881 2.23 (0.89) 92.16 (14.841
96.33 (8.50) 96.68 (12.54) 2.01 (0.58) 1.17 (0.36) 4.55 (1.17) 21.94 (9.59) 10.66 (4.17)** 6.33 (2.87)* 49.11 (15.94)* 21.55 (3.45)** 19.77 (8.07) 1.88 (1.1l) 8.48 (5.14) 21.66 (9.08) 0.51 (2.15)* 40.80 (20.33)* 1.75 (0.54) 59.88 (9.0(/)
R e s u l t s
Experimental subjects. N o significant difference was found between subjects belonging to the Stress and the
f= frequency (number of acts/session): d = duration (seconds). * Student's t, P< = 0.05; ** Student's t, P< = 0.01. See text for parameter labels.
190
M . E . A[bonetti and F. barahollini
Behavioural Brain Research 62 1994
Attack and Bite were never observed, both low-intensity Threat (TH; t = 3.21, nl = 10, n2 = 9, P = 0.005) and highintensity On-Top (OT; freq., t=2.10, P = 0 . 0 5 ; dur., t= 2.38, P=0.02), as well as total aggression (TAG, t = 2,91, P = 0.01), were significantly lower in stressed than in control animals. Aggression was markedly less frequent than defence in the stressed subjects, but such a difference was not found in the controls. Thus the balance of total aggression and total defence, as measured by (TAG-TDE), was significantly different between the Stress and Control groups (t= 2.31, P=0.03). Similarly, stressed animals spent much less time in the extreme aggressive posture, On-Top, than in the extreme defensive posture, On-Back, whereas the reverse was true for control subjects, and this difference was again significant (t = 2.22, P = 0.04). No significant correlation was found between any parameter of social behaviour and any measure of emotionality (OP, N°/~,; OP, T%I,; LAT; I N / O U T ; BOL). After the 7-day experimental period, stressed animals had lost weight, the mean decrease being 2.20 g (4.65 S.E.), whereas control animals gained an average of 20.77 g (10.41 SE). This difference was significant (t = 2.08,
?=0.05). Opponents. Opponents spent significantly more time investigating the stressed animals than the controls (t = 2.08, P = 0.05; Table 3), but directed to both groups similar amounts of aggression. However, striking differences emerged in defence; both low-intensity (LDE, t= 2.43, P = 0 . 0 2 ) and high-intensity (On-the-Back, OB; freq., t = 2.15, P = 0.04; dur., t = 2.33, P = 0,03)defensive param-
Table 3 Social behaviour - - Opponents: means with S.E. of the parameters directed to the Experimental animals during the 20-min test
Behavioural parameters
Opponentsto-stressed n = 10
IN f IN d GR f GR d GG f GG d TH OT f OT d TAG f LDE OB f OB d TDE f (TAG-TDE) f (OT-OB) d
79.80 (12.55) 92.35 (11.83) 4.12 (1.13) 3.30(1.31) 4.43 (1.51) 14.97 (6.37) 3.90 (1.32) 2.70 (2.38) 25.94 (10.83) 11.00 (3.08) 4.t(I (1.36) 0.75 (0.28) 7.30 (3.89) 4.80 (1.60) 6.20 (4.06) 18.64 (9.24)
* Student's t, 0 . 0 5 < P < 0 . 1 : olhers as in Table 2.
Opponentsto-control n=9 64.33 (t3.28) 45.31 (10.47)* 4.87 (0.96) 4.56(1.08) 4.90 (1.38) 16.18 (7.63) 5.00 (1.46) 1.77 (1.02) 8.96 (5.63) 11.66 (2.64) 16.67 (5.23)* 6.36 (2.72)* 52.98 ( 12.22)* 23.00 (7.69)* - 11.33 (8.63)' - 4 4 . 0 2 (14.86)*
1~ ~- 19/
eters were lower in the Opponent~ confronted with stressed subjects (Opponents-to-stressed j than in those confronted with control animals (Opponents-to-control). The Opponents-to-stressed were indeed more aggresstve than defensive, as suggested by the positive values of both {TAGTDE) and (OT-OB), whereas the reverse was true for Opponents-to-control. which had negative values for the same scores. The difference belween the two groups of Opponents was significant ~br (OT-OBI 11=-2.21. P = 0.041, and marginally. ,~) t\,~r ( I ' A G - T D E ) (t1.7t, P = 0.07 }.
4. Discussion In male rats. repeated defeat, expected to act as a sociat stressor [77] as indirectly confirmed b y the average decrease of body weight, caused a selective and specific inhibition of aggression in subsequent encounters with unfamiliar naive conspecifics (Opponents). Under our experimental conditions, no additional effect was found on other social beh aviours, including defence, nor on anxietyrelated non-social behaviours. Consistent with the decreased offence shown by the stressed animals, reduced defence was observed in their Opponents. Our findings are partially at variance with the pertinent literature on rats and other rodents, which reports repeated defeat to: (1) inhibit not only offence [41], but also other social behaviours such as the introductory items [40,68]; (2) inhibit non-social behaviours, such as exploranon and locomotor activity [40]; and (3) increase defence [62.67,68]. Behaviourat inhibition is indeed considered to be a consequence of the activation of defence. which is presumed to overcome any other motivational system E75]. It is however interesting that Vogel et al. [71 ], in a review of behavioural alterations typical of human depression which can be sinmlated in animals through exposure to uncontrollable stressors, mentioned "irritability" and "hostility". which may be considered as items related to aggression rather than defence. In the present work. the absence of any increase of defence suggests that the inhibition of offence was not due to the concomitant activation of a prevailing defensive motivational system. Tha( offence and defence may not alwa3 s act as strictly linked and interdependent behavioural systems [9] is consistent with the long-accepted view that offensive and defensive behaviours have different remote and proximate causation [51] and arc served by different neural substrates [5,18,23]. This belief appears to be supported by recent pharmacological investigations which characterized a number of drugs (such as fluprazine/acting selectively on aggresmon with no effect on defence [30,48,52].
M.E. Albonetti and F. Farabollini / Behavioural Brain Research 62 (1994) 187-193
Recently, Siegfried at al. [58], working on mice, proposed the hypothesis that nociceptive stimulation, induced by being attacked and bitten in a social conflict, may be responsible for the activation of a defensive motivational system. In the present experiment, activation of defensive motivation might have been prevented by a lack of nociceptive stimulation of sufficient intensity. As a matter of fact, high-intensity aggressive items which might possibly have caused physical damage (e.g. Attack, Bite) were never observed, presumably due to the use of the low-aggression Wistar strain [62]. However, this absence of high-intensity aggression did not prevent the occurrence of defeat both during the stress-induction phase and at testing. In the present experiment, repeated defeat failed to increase not only defence, but also anxiety. Social interaction, considered to be negatively correlated with anxiety [28,38], was unaffected by the social stress: the stressed rats did appear to be interested in the unfamiliar Opponents, and approached, explored and groomed them as much as did the control animals. In addition, no change was found in behavioural parameters widely assumed to be standard measures of anxiety (percentage entries and time spent in the open arms of an Elevated Plus-Maze; latency to exit from a box; defecation and internal vs. external locomotion in the Hole-Board; see Methods for references). Even Self-Grooming, which may be a substitutive activity indicating embarrassment and uncertainty and which has been viewed as a hint of the animal's emotionality [25,53], was similar in the stressed and control rats. Increased anxiety in the Plus-Maze has been found immediately after a single defeat in male rats [36] and mice [53]. Taken together, these and our results suggest either that defeat does increase anxiety but that this is a shortterm effect no longer appreciable 24 h afterwards, or that anxiety, increased by a single defeat, returns to basal levels when defeat is repeated. Aside, attention has been recently called on the possibility that different anxiety parameters, recorded under different testing conditions, measure different aspects, or kinds, of anxiety [27]. Thus, the possibility cannot be ruled out that anxiety tests we have carried out might have been inadequate to reveal anxiety supposedly induced by social stress. The hypothesis that the stressed rats were not especially anxious since their low aggression did not provoke retaliation, thus reducing the amount of aggression by the Opponents (see [57]), is disproved by the finding that the Opponents directed similar amounts of aggression to both the stressed and control Experimental animals. Therefore, in the stressed rats the reduction of aggression given, namely that directed to the Opponents, did not appear to be a strategic tool effective in reducing the aggression received. In other words, it did not act as a signal [76] able to modify the Opponents'behaviour. From an adaptive
191
point of view, decreased aggression reduced energy expenditure but did not prevent the possibility of being injured during agonistic encounters. More importantly, in species like the rat which are characterized by a hierarchical social organization, with aggression playing a crucial role in gaining high rank [6,7,12,62], inhibited aggression condemns an animal to a low ranking, which is detrimental to individual and inclusive fitness [ 14]. Interestingly, the Opponents investigated the stressed rats for a significantly longer time than the control animals; this was presumably aimed at the detection of olfactory cues revealing the condition of stress [24,43,66]. In a previous study [9], we obtained selective inhibition of offence, but no effect on defence, 24 h after both single and repeated exposure to a mild physical stressor, restraint [35]. Offence thus appears to have a greater sensitivity to aversive stimulation and a lower threshold for stress-induced changes than other behavioural systems, whereas stronger stressors are presumably required to activate and maintain defence and produce a persistent increase in anxiety. Interestingly, the nature of the stressor, either social or physical, does not seem to affect the behavioural responses to it. Differences in the ability ofstressors to induce behavioural modifications seem to be due to their quantitative rather than qualitative attributes--in other words, they seem to be a matter of how much rather than how (see [40]).
Acknowledgements We wish to thank Dr. P. Christie for revision of the English.
References [1] Abel, E.L., Behavior and corticosteroid response of maudsley reactive and nonreactive rats in the open field and forced swimming test, Physiol. Behav., 50 (1991) 151-153. [2] Abel, E i . , Altman, H.J. and Commissaris, R.L., Maudsley reactive and nonreactive rats in the forced swim test: comparison in fresh water and soiled water, Physiol. Behav., 52 (1992) 1117-1119. [3] Abel, E.L. and Bilitzke, P.J., A possible alarm substance in the forced swimming test, Physiol. Behav., 48 (1990) 233-239. [4] Adams, D.B., The relation &scent-marking, olfactory investigation and specific postures in the isolation induced fighting of rats, Behaviour, 56 (1976) 286-297. [5] Adams, D.B., Motivational systems ofagonistic behavior in muroid rodents: a comparative review and neural model, Aggress. Behav., 6 (1980) 295-346. [6] Adams, D.B., Boice, R., Development of dominance in domestic rats in laboratory and seminatural environments, Behav. Proc_ 19 (1989) 127-142. [7] Albert, D.J., Dyson, E.M., Walsh, M.L. and Wong, R., Defensive aggression and testosterone-dependent intermale social aggression are each elicited by food competition, Ph~wiol. Behav., 43 (1988) 21-28.
192
M.E. Albonetti and F Farabollilzi, Behavhmral Brain Research 62 19941 tb¢7-!9;
[8] Albonetti, M.E. and Faraboltini, F., Behavioural responses to single and repeated restraint in male and female rats, Behav. Proc.. 28 (1992) 97-110. [9] Albonetti, M.E. and Faraboltini, F., Efl?cts of single and repeated restraint on the social behavior of male rats, Physiol. Behav.. 53 11993) 937-942. [10] Anderson, E.E., The sex hormones and emotional behavior: IlL The effect of castration upon timidity in male and female rats, J Genet. Psychol., 56 (1940) 169-174. [ 11 ] Archer, J., Rodent sex differences in emotional and related behavior, Behav. Biol., 14 (1975) 451-479. [12] Benus, R.F., Bohus, B., Koolhaas, J.M. and van Oortmerssen, G.A., Behavioural strategies of aggressive and non-aggressive male mice in active shock avoidance, Behav. Proc., 20 (1989) 1-12. [13] Blanchard D.C. and Btanchard, R.J., Affect and aggression: an animal model applied to human behavior, Adv. St. Aggress'.. 1 (1984) 1-62. [14] Blanchard D.C. and Blanchard, R.J., Behavioral correlates of chronic dominance-subordination relationships of male rats in a seminatural situation, Neurosci. Biobehav. Rev., 14 (1990) 455-462. [15] Blanchard D.C., Cholvanich, P., Blanchard, R.J., Clow, D.W., Hammer Jr., R.P., Rowlett, J.K. and Bardo, M.T., Serotonin, but not dopamine, metabolites are increased in selected brain regions of subordinate male rats in a colony environment, Brain Res., 568 (1991) 61-66. [16] Blanchard, R.J,, Flannelly K. and Blanchard D.C., Life-span studies of dominance and aggression in established colonies of laboratory rats, Physiol. Behav., 43 (1988) 1-7. [17] Blanchard, R.J., Kelley, M J . and Blanchard D.C., Defensive reactions and exploratory behavior, J. Comp. Physiol. P.tvchoL, 87 (1974) 1129-1133. [18] Blanchard, R.J., Kleinschmidt, C.K., Flannelly K. and Blanchard D.C., Fear and aggression in the rat, Aggress. Behav.. 10 (1984) 309-315. [19] Bolles, R.C., Species-specific defense reactions and avoidance learning, Psychol. Rev., 77 (1970) 32-48. [20] Brain, P.F., Kusumorini N. and Benton. D., "Anxiety" in laboratory rodents: a brief review of some recent behavioural development, Behav. Proc., 25 (1991) 71-80. [21] Broadhurst, P.L., Determinants of cmotionality in the rat: 1. situational factors, Br. J. Psychol., 48 (1957) 1-12. [22] Chopin, P. and Briley, M., Animal models of anxiety: the effect of compounds that modify 5-HT neurotrasmission, Trends Pharmacol. Sci., 8 (1987) 383-388. [23] Depaulis, A. and Vergnes, M., Relationship between mousekilling and conspecific aggression in the male rat, Aggress. Behav., 9 (1983) 259-268. [24] Dua, J.K. and Dobson, M.J., Role of olfactory cues in acquisition and extinction of avoidance, J. Exp. Psvchol., 103 (1974) 461-465. [25] Farabollini, F., Hole, D.R. and Wilson, C.A., Behavioral effects in adulthood of serotonin depletion by p-chlorophenylalanine given neonatal!y to male rats, bit. J. Neurosci., 41 (1988) 187-199. [26] Fernandez-Espejo, E. and Mir, D., Ethological analysis of the male rat's socioagonistic behavior in a resident-intruder paradigm, Aggress. Behav., 16 (1990) 41-55. [27] File, S.E., Behavioural detection of anxiolytic action. In J.M. EItiott, D.J. Heal and C.A. Marsden (Eds.), ExperimentalApproaches to Anxiety and Depression, Wiley and Sons, 1993, pp. 25-44. [281 File, S.E. and Hide, J.R., Can social interaction be used to measure anxiety? Br. J. Pharmacol., 62 (1978) 19-24. [29] File, S.E., Wardill, A.G., The reliability of the holeboard apparatus, Psychopharmacology, 44 (1975) 47-51. [30] Flannelly, K.J., Muraoka, M.Y., Blanchard, D.C. and Blanchard, R.J., Specific anti-aggressive effects of fluprazine hydrochloride, Psychopharmacology (Berl.), 87 (1985)86-89. [31] Grant, E.C. and Mackintosh, J.H., A comparison of the social postures of some common laboratory rodents, Behaviour, 21 (1963) 246-259.
[ 32 [ Gray, J.A.. Sex differences m the emotional behavior of laboratory rodents: comment. Br. 1. Ps~:ho[.. 70 (19791 3 5 - 3 6 [33] Hall. C S . . Emotional behavior )n the rat. I. Defecation and urinanon as measures of individual differences in c~notionalit~. 1 Comp. Psychol,, 18 11934) 385-403~ [34 t H andley, S.L. and Mithani. S.. Eit~2cts of atpha-adrenoceptor agonists and antagonists m a mazc
[40] Kudryavtseva, N.N.. Bakshtanovska)a, 1.'~. and Kor~akina, k 'A~,. Social model of depression in mice of C57BL.'6J strata. Pharmacol Biochem Behav.. 38 11991) 315-321i [41 ] Leshner. ~.I.. The role of hormones m the control of submissiveness. In P.F. Brain and D Benton reds.), Multidisciplinao" "lp!)rmwhes' to .4ggression Research. Elsevier North-Holland. Amsterdam. 1981. pp. 309-322. [42] Lister. R.G.. The use of a plus-maze to measure anx~et)in the mouse. Ps3"chopharmacology {Berl. ,;, 92 f 19871 180-185 [43] Macka~-Sim. A. and Laing, D.G,, Discrimination of odors from stressed rats by non-stressed rats. Physiol. Behav.. 24 (1980) 699704. [44] Maier. S,F.. Drugan. R , Gram J.~A .. 1tyson. R., Macl,ennan, A.J.. Mo3e. T.. Madden. J. and Barchas~ J.D., Learned helplessness. pare inhibition and the endogenou~ opmtes, In M.D. Zeiler and P. Harzem (Eds.). Advances b7 An#rra~ Behavior. Vol. 3. Wiley. New York. 1983. pp. 275-323. [45] Misslin R. and Cigrang. M.. Doc~ ncophobia neccssaril5 impl 5 fear or anxiety? Behav Proc , 12 (1986} 45-50. [46] Montgomery, K.C.. The relation between fiear induced by novelty stimulation and exploratory behaviour, ,L Comp. Physiol. Ptvchol., 48 11955) 254-260 [47] Musty. R,E.. Jordan. M.P. and Lcm)x. R.H.. Criterion for learned helplessness in the rat: a redefinition. Pharmacol, Bioehem. Behav. 36 11990t 739-744 [48] Olivier. B., Mos. J. and Hartog. J., Eltoprazine: a specific annaggressive drug ~serenic) in animal aggression models. Aggress. Behav.. 15 f1989) 9 5 - 1 1 2 [49] Pellm~,. 51.. Chopin. P.. File. S.I. and Brile~. M.. Validation of open:closed arm entries m an elevated plus maze as a measure ol anxiety in the rat. J. Neurosci Meth.. 14 11985] 149-t67. [50] Ptaznik. ~\.. Stefanski. R. and Kostowski. W.. Restraint stressinduced changes in saccharin weference: the effects of antidepresslvc treatment and diazepam. PhatvnacoL Biochem Behav.. 33 (1989) 755-759. [5t] Rasa. O.A.. Aggression: appetite or aversion? A ggress. Behav.. 2 1976J 213-222. [52] Rawleigh. J.M. and Kemble. E.D.. 'Iest-specific effects ot EG-7142 on isolation-induced aggression m mice. Pharmacol. Biochet~. Behav.. 42 ~1992) 317-321 [531 Rodgers. R.J. and Cole. J . ( . , Anxlet) enhancement m the murmc elevated plus maze b) immediate prior exposure to social stressors. PhvsioL Behav.. 53 (19931 383-3 ~
M.E. Albonetti and F. Farabollini / Behavioural Bra#t Research 62 (1994) 187-193 [54] Rodriguez Echandia, E.L., Gonzalez, A.S., Cabrera, R. and Fracchia, L.N., A further analysis of behavioral and endocrine effects of unpredictable chronic stress, Physiol. Behav., 43 (1988) 789-795. [55] Russel, P.A., Open field defecation in rats: relationships with body weight and basal defecation level, Br. J. Psvchol., 64 (1973) 109114. [56] Scholtens, J., van Haaren, F. and van de Poll, N.E., Effects of losing and testosterone upon subsequent behavior in male and female $3 (Tryon Maze Dull) rats, Aggress. Behav., 14 (1988) 371387. [57] Senar, J.C., Agonistic communication in social species: what is communicatcd? Beha~im~r, 112 (1990) 270-282. [58] Siegfried, B., Frischknecht H. and Nunes de Souza, R.L., An ethological model for the study of activation and interaction of pain, memory and defensive systems in the attacked mouse. Role of endogenous opioids, Neurosci. Biobehav. Rev., 14 (1990) 481-490. [59] Snedecor, G.W. and Cochran, W.G., Statistical Methods, Iowa State Univ. Press, 1980, 5(17 pp. [60] Steenbergen, H.L., Heinsbroek, R.P.W, Van Haaren, F. and Van de Poll, N.E., Sex-dependent effects of inescapable shock administration on behavior and subsequent escape performance in rats, Physiol. Behav., 45 (1989) 781-787. [61] Stefanski, R., Palejko, W., Kostowski, W. and Plaznik, A., The comparison of benzodiazepine derivatives and serotonergic agonists and antagonists in two animal models of anxiety, Neuropharmacoh~gy, 3 l (1992) 1251 - 1258. [62] Swanson, H.H., Sex differences in behavioral consequences of defeat in the rat arc not organized by testosterone during early development, Aggress. Behav., 16 (1990) 341-344. [63] Takahashi, L.K., Kalin, N.H., Vanden Burgt, J.A. and Sherman, J. E., Corticotropin-relcasing factor modulates defensive withdrawal and explorator~ behavior in rats, Behav. Neurosci., 103 (1989) 648654. [64] Taukulis, H.K. and Goggin, C.E., Diazepam-stress interactions in the rat: effects on autoanalgesia and a plus-maze model of anxiety, Behav. Neural Biol., 53 (1990)205-216. [65] Trcit, D., Menard, J. and Royan. C., Anxiogenic stimuli in the
[66] [67]
[68]
[69]
[70]
[71]
[72] [73]
[74]
[75]
[76] [77]
193
elevated plus-maze, Pharmacol. Biochem. Behav., 44 (1993)463469. Valenta, J.G. and Rigby, M.K., Discrimination of the odor of stressed rats, Science, 161 (1968) 599-601. van de Poll, N.E., de Jonge, F., van Oyen, H.G. and van Pelt, J., Aggressive behavior in rats: effects of winning and losing on subsequent aggressive interactions, Behav. Pro~., 7 (1982) 143-155. van de Poll, N.E., Smeets, J., van Oyen, H.G. and van der Zwan, S.M., Behavioral Consequences of agonistic experience in rats: sex differences and the effects of testosterone, .I. Comp. Physiol. Psychol., 96 (1982) 893-903. van de Poll, N.E., Swanson, H.H. and van Oyen, H.G., Gonadal hormones and sex differences in aggression in rats. In P.F. Brain, P.F. and D. Benton, D. (Eds.), The Biology ¢!fAg,~,ression, Sijthoff and Noordhoff. Amsterdam, 1981, pp. 244-252. van Dijken, H.H., Mos, J., van der Heyden, J.A. and Tilders, F.J., Characterization of stress-induced long-term behavioural changes in rats: evidence in favor of anxiety, Physiol. Behav., 52 (1992) 945-951. Vogel, G., Neill, D., Hagler, M. and Kors, D., A new animal model of endogenous depression: a summary of present findings, Neurosci. Biobehav. Rev., 14 (1990) 85-91. Wallace, R.J., Latency measures indicate new place neophobia in Rattus species, Behav. Proc., 17 (1988) 63-67. Welker, W.I., Escape, exploratory, and food-seeking responses of rats in novel situation, J. Comp. Physhd. Psychol., 52 (1959) 106111. Wiepkema, P.R., Koolhaas, J.M. and Olivier-Aardema, R., Adaptive aspects of neuronal elements in agonistic behavior, Prog. Brain Res., 53 (1980) 369-384. Williams, J.L. and Lierle, D.M., Effects of repeated defeat by a dominant conspecific on subsequent pain sensitivity, open field activity, and escape learning, Anita. Learn. Behav., 16 (1988) 477485. Wilson, E.O., Sociobiology: the New Synthesi.~, Harvard University Press, Cambridge, 1975, pp. 671. Zayan, R., The specificity of social stress, Behav. Proc., 25 (1991) 81-93.