- Email: [email protected]
Stress-induced social avoidance
Physiology & Behavior 77 (2002) 327 – 332
Stress-induced social avoidance: A new model of stress-induced anxiety? J. Haller*, N. Bakos Institute of Experimental Medicine, P.O. Box 67, 1450 Budapest, Hungary Received 25 February 2002; received in revised form 25 March 2002; accepted 23 July 2002
Abstract We have studied the long-term behavioral effects of a single stressor in male rats by using an approach/avoidance situation as the behavioral endpoint. A single exposure to social defeat or electric shocks was used as stressors. Behavioral testing was performed in a twocompartment cage divided by an opaque wall and connected by a short tunnel. The larger compartment contained an unfamiliar male rat that was separated from the rest of the compartment by a transparent, perforated Plexiglas wall. The subject was placed in the small compartment and allowed to explore the cage for 5 min. The test was performed on Days 1, 5, or 10 after stress application. Unstressed rats spent 90% of time in the large compartment that contained the unfamiliar male. Social defeat dramatically reduced the exploration of the large compartment, without time-related changes in this response. A mild electric shock had a similar effect that lasted more than 5 days but less than 10 days. The exploration of an empty cage was significantly less inhibited by stress than the exploration of a cage that contained the stimulus rat. The test could be applied repeatedly in the same rat, without major changes in the response. Chlordiazepoxide applied 1 h before behavioral testing abolished completely the stress-induced behavioral deficit. We suggest that the model can be used for studying the effects of various compounds on stress-induced anxiety. D 2002 Elsevier Science Inc. All rights reserved. Keywords: Stress; Anxiety; Anxiolytic; Model; Rat
1. Introduction Anxiety and the anxiolytic potential of various compounds are studied mostly in animals with basal anxiety levels. This in sharp contrast with the study of other disorders (e.g., depression, schizophrenia), where the respective condition is induced by behavioral or pharmacological techniques. Anxiety tests involving anxious animals did not become popular probably because (i) benzodiazepines, for which the available laboratory tests were developed, provide reliable results in animals with basal anxiety levels, and (ii) the widely used chronic stress paradigms that lead to anxiety are laborious. However, the advent of new classes of anxiolytics questions the applicability of old methods, mostly because anxiolytic agents other than benzodiazepines provide unreliable results with the usual anxiety tests [1,2]. Recently, a model
*
Corresponding author. Tel.: +36-1-210-9406; fax: +36-1-210-9951. E-mail address: [email protected] (J. Haller).
has been developed for the study of anxiolytics in animals that were made anxious by the anxiogenic compound mchlorophenyl-piperazine [3]. Despite the interesting results, one wonders what anxiety types are modeled by the application of this compound, the effects of which occur rapidly and are short-lived. More promising are models that make use of the anxiogenic potential of stress. Stress is among the aetiological factors of anxiety and human patients are frequently exposed to stressors during treatment. Moreover, anxiety increases the sensitivity towards everyday stressors. In addition, stress was shown to affect the efficacy of many compounds that affect anxiety (e.g., serotonergic and cholecystokininergic compounds) [4– 7]. Regarding the effects of stress on anxiety and the efficacy of anxiolytics, single stressors provide a good alternative to the time-consuming and laborious chronic stress procedures. It has been shown that strong single stressors have surprisingly long-term consequences in rats. A single exposure to electric shocks, social defeat, or a cat affects stress reactivity, neurochemical measures, and behavior for days and even weeks [8– 11]. Data suggest that the long-
0031-9384/02/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved. PII: S 0 0 3 1 - 9 3 8 4 ( 0 2 ) 0 0 8 6 0 - 0
328
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
term consequences of a single stressor include an increase in anxiety [8,12]. In the present work, we have studied the long-term effects of single stressors on social avoidance. We have also assessed the possibility of developing an anxiety model based on the long-term consequences of single stressors. We have applied electric shocks or social defeat to male rats and investigated the time course of the resulting behavioral deficit in a social approach/avoidance situation. The reasons for choosing this testing procedure were the following: (i) It has been shown that social avoidance is associated with several types of human anxiety, including social anxiety, panic, and generalized anxiety disorder (GAD) [13 – 15]. Noteworthy, social avoidance is common in GAD patients lacking phobic avoidant behavior and which are consequently diagnosed as having GAD and not social phobia [15]. (ii) The widely used anxiety tests (e.g., the plus-maze, light/dark, open-field, passive avoidance tests) make heavy use of approach/avoidance behavior towards aversive stimuli. It has been also shown that social avoidance is increased by anxiogenic drugs in both rats and monkeys [16,17]. In addition, a single social defeat resulted in a reduction of social contacts during a subsequent encounter with an unfamiliar, nonaggressive conspecific [18]. (iii) The approach/avoidance situation provides a simple testing procedure in a social context. There are a large number of nonsocial anxiety tests, but only one that investigates the social dimension of anxiety (the social interaction test [19]). This test provides reliable and valuable information on anxiety, but is rather laborious and requires special training. One objective of the present work was to find easily applicable methods for measuring the long-term, anxiety-related consequences of single stressors.
larger resident rat. The aggressiveness of residents was checked previously and only residents showing high levels of attacks were used. The encounter lasted 30 min. The number of biting attacks delivered by residents was recorded. Electric shocks were delivered via the grid floor of a Plexiglas box (30 30 30 cm). An alternating current of 100 V and 8 mA was applied. Shock duration was 0.01 s; shocks were delivered in trains lasting 1 s with an intershock interval of 0.02 s. Two trains per minute were delivered. Shocks were delivered for 5 min, i.e., each rat received 10 shock trains of 1s. Animals returned to their home cage after stress exposure and were left undisturbed until testing. 2.3. The approach/avoidance test The apparatus consisted of two chambers connected by a tunnel (diameter: 10 cm, length: 5 cm) (Fig. 1). The dimensions of the small and large chambers were 40 20 20 and 40 40 20 cm, respectively. The tunnel was closed by a sliding door. A large male rat was enclosed in a subchamber of the large chamber (40 15 15 cm), which was separated from the rest of the chamber by a transparent perforated Plexiglas wall. The apparatus was lit by dim red light. The test started by placing the subject in the small chamber. After an acclimation period of 3 min, the sliding door was removed and the subject was allowed to explore the whole apparatus for 5 min. Behavior was video-recorded from above and later analyzed. The camera was placed 2.5 m above floor. Two variables were recorded: the number of visits made to the large compartment and the time spent in the large compartment.
2. Materials and methods 2.1. Animals Subjects were male Wistar rats weighing approximately 300 g (Charles River Laboratories, Hungary). They were kept in a 12:12-h reverted day –night schedule, with lights off at 1000 h. Rats were maintained under standard laboratory conditions (temperature: 21 ± 2 °C; humidity: 60% ± 10%). Standard laboratory food (Charles River) and tap water were available ad libitum. All animals were isolated 3 days before stress application and remained isolated throughout, because it has been shown that social isolation is a prerequisite of the long-term changes induced by single stressors [20]. In addition, the approach/avoidance behavior of socially housed animals was highly variable in preliminary experiments. 2.2. Stress procedures Social stress was induced by exposing subjects to inescapable attacks in the home cage (40 50 60 cm) of a
Fig. 1. The test apparatus for social avoidance: 1 = small chamber; 2 = tunnel with sliding door; 3 = large chamber; 4 = transparent, perforated Plexiglas wall; 5 = subchamber of the unfamiliar male. The subject is placed in the small chamber and, after a 3-min habituation, the sliding door is removed. The rat is allowed to explore the whole apparatus for 5 min.
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
329
2.4. Experimental design Both stress application and testing were performed in the early hours of the dark (active) period. Animals were individually housed 3 days before stress application and individual housing was maintained throughout. Six experiments were run. In Experiment 1, rats were socially stressed and tested 1, 5, or 10 days thereafter in the approach/avoidance paradigm. Control rats were left undisturbed in their home cage. Different rats were used for each time-point (n = 6 per group). In Experiment 2, the rats were socially stressed and left undisturbed for 10 days. On the 10th day, they were injected either with vehicle or chlordiazepoxide-hydrochloride (Sigma). The latter was dissolved in saline and administered intraperitoneally. The dose was 5 mg/kg. One hour after the injections, rats were tested in the approach/avoidance paradigm (n = 8 per group). In Experiments 3 and 4, rats were exposed to electric shocks and the behavioral test was run 1, 5, or 10 days after shock exposure. In Experiment 3, different rats were used for each time-point (n = 8 per group). In Experiment 4, rats were tested three times at 1, 5, and 10 days (n = 8 per group). In Experiment 5, the effects of chlordiazepoxide (5 mg/kg) was tested in animals stressed by electric shocks, 5 days after stress exposure (n = 8 per group). The lag of 5 days was chosen, because the previous experiment showed that the effects of electric shocks are shorter lasting than the effects of social stress. Experiment 6 assessed the role of the stimulus rat in the electric shock-induced avoidance of the large compartment. Rats were tested 1 day after shock exposure by exposing them to either an empty cage or a large compartment containing the stimulus rat (n = 8 per group). 2.5. Statistics In Experiments 1, 3, and 4, results were analyzed by twofactor ANOVA. Factor 1 was stress application, while Factor 2 was time. Data were analyzed by one-factor ANOVA in Experiments 2, 5, and 6.
3. Results Unstressed controls spent most of the time (around 90%) in the large compartment, by investigating the box containing the unfamiliar stimulus rat (Figs. 2 and 3). This behavior was very robust, with small individual variability. Results were almost identical in the five experiments performed. The number of visits performed was between four and six and also showed large consistency. Socially stressed animals received 5.72 ± 0.84 attacks during the stress period. The time spent in the large compartment was more than halved in stressed rats and this effect did not show time-related changes (Factor 1
Fig. 2. The effect of social stress on social avoidance. Stress was applied 1, 5, or 10 days before testing, as indicated. Chlordiazepoxide (5 mg/kg) was applied 1 h before behavioral testing. * Statistically significant differences compared with unstressed animals in post hoc comparisons.
[stress] = 30.05, P < .0001; Factor 2 [time] = 0.51, P < .6; Stress Time interaction = 0.06, P < .9) (Fig. 2). The number of visits was not changed (Factor 1 = 0.89, P < .3; Factor 2 = 1.11, P < .3; Stress Time interaction = 0.6, P < .6). In the second experiment, behavioral effects were tested 10 days after defeat. Social stress produced the effects observed earlier; moreover, the number of visits was also decreased in this experiment. A similar decrease was noticed in Experiment 1, but differences were not significant. Chlordiazepoxide applied 1 h before behavioral testing abolished stress-induced changes completely ( F = 110.69, P < .0001 for time spent in the large compartment, and F = 5.83, P < .01 for the number of visits made to the large compartment). Electric shocks decreased the duration of large compartment visits dramatically on Days 1 and 5 (Fig. 3). In contrast to social stress, the effects of electric shocks disappeared after 10 days (Factor 1 = 117.20, P < .0001; Factor 2 = 36.08, P < .0001; Stress Time interaction = 37.87, P < .0001). Similar results were obtained when the same animals were repeatedly tested on Days 1, 5, and 10. The only difference was that the effect was still significant at 10 days, i.e., the effect of stress was diminished but significant on the 10th day (Factor 1 = 294.6, P < .0001; Factor 2 = 17.47, P < .001; Stress Time interaction = 15.65, P < .001). The number of visits showed some time-related changes in this case, but the stress had no effect on this variable (Factor 1 = 1.33, P < .3;
330
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
Factor 2 = 5.92, P < .004; Stress Time interaction = 2.83, P < .07). Chlordiazepoxide effects were tested after 5 days and, again, the behavioral effects of stress were completely abolished by chlordiazepoxide ( F = 232.6, P < .0001 for time spent in the large compartment; F = 0.17, P < .8 for the number of visits). Electric shocks reduced the duration of the exploration of the large compartment, even when this did not contain a stimulus rat ( F = 38.17, P < .0001; Fig. 4). However, the exploration of the large compartment lasted significantly less, when it contained a stimulus rat ( F = 34.7, P < .0001). The frequency of large compartment visits did not show significant variations. Thus, the electric shock-
Fig. 4. The effect of the stimulus rat on the electric shock-induced avoidance of the large compartment. Stress was applied 1 day before testing. Data were expressed as percentage of change compared with control. Stressed animals showed a significant avoidance of the large compartment, even when this was empty. However, the avoidance was significantly stronger, when the stimulus rat was present.
induced avoidance of the large compartment showed a strong social component.
4. Discussion
Fig. 3. The effect of exposure to electric shocks on social avoidance. Stress was applied 1, 5, or 10 days before testing, as indicated. Upper panels = different animals used for each time-point; middle panels = animals tested on Day 1 and retested on Days 5 and 10; lower panels = the effect of chlordiazepoxide (5 mg/kg) applied 1 h before behavioral testing on Day 5. * Statistically significant differences compared with unstressed animals in post hoc comparisons.
In the approach/avoidance situation employed, naive rats spent approximately 90% of time in the compartment containing the stimulus rat. Socially stressed rats spent less than 50% of time in the compartment containing the stimulus rat, with a tendency of reducing also the number of visits. This effect did not change during the 10 days of the experiment. Electrically shocked rats showed a similar behavior, but the effect lasted less then 10 days. Repeating the test in the same animals did not lower the effects of electric shocks; moreover, a slight increase in the effect was noticed. The stress-induced avoidance of the large compartment showed a strong social component. Chlordiazepoxide abolished the consequences of both social stress and electric shocks. It has been shown earlier, that single stressors have longterm consequences in the rat. A single exposure to social defeat, electric shocks, or a cat, induced long-term neurochemical, endocrinological, and behavioral changes [8 –11]. Our results are in agreement with these data, since a single social defeat or electric shock induced a lasting behavioral effect in our rats. The single difference was the relatively fast ( < 10 days) recovery from the effects of electric shocks,
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
which is in contrast with earlier findings [11,12]. The fast recovery may be explained by the mild electric shocks applied, since our rats received fewer shocks and the duration of stress application was lower. One can assume that the effects of more frequent shocks administered for a longer period may have longer lasting effects. Noteworthy, the effects of stress were smaller but significant 10 days after shock exposure, when the rats were repeatedly tested. This suggests that testing (the contact with an unfamiliar male) prolonged the effects of stress. The behavior of controls was not changed by repeated testing. It has been suggested earlier that the long-term behavioral effects of single stressors are caused by anxiety, which is a consequence of a long-term increase in stress responsiveness [4,8,12]. This assumption is supported by the strong effect of chlordiazepoxide on the stress-induced social avoidance in our experiment. Traumatic events in humans lead to a long-term increase in anxiety and generate the symptoms of post-traumatic stress disorder [21]. Thus, the long-term effects of single stressors may model a human type of anxiety. Taken conjointly, our data suggest that stress-induced social avoidance may be used as a model of stress-induced anxiety and, prospectively may be used to assess the effects of anxiolytic agents. This assumption is supported by several lines of reasoning. (i) Stress is an important component of anxiety, either as an aetiological factor or as a factor concurrent with pharmacological treatment. (ii) Social avoidance is a common symptom in various anxiety disorders, including social anxiety, panic, and generalized anxiety [13 –15]. In addition, social avoidance is increased by anxiogenic drugs in animals [16,17]. Noteworthy, the avoidance of (nonsocial) aversive stimuli is the main variable measured in the current anxiety tests. (iii) The stress-induced social avoidance model allows the study of drug effects in animals with increased anxiety. This is in sharp contrast with the current practice, which involves the study of animals showing natural anxiety levels. It is important to emphasize that stress-induced social avoidance occurred also in a heterotypic situations, i.e., rats showed social avoidance even when the stressor was nonsocial (it was an electric shock). (iv) The model allows the study of the effects of anxiolytics at multiple levels: during stress exposure, chronically after stress exposure, and acutely before testing. It is important to emphasize that the test appears repeatable, i.e., it may allow the repeated testing of the same animals. This may lower sample size in chronic studies. (v) Testing is simple, rapid, and does not require special training. In conclusion, single stressors induce long-lasting social avoidance in rats. The effect appears reversible by anxiolytic agents, which suggest that stress-induced social avoidance may be used as model for the study of the interaction among stress, anxiety, and pharmacological anxiolysis.
331
Acknowledgements This study was supported by OTKA Grant No. T 032345, ETT Grant No. 286/2001, and OMFB Biotechnology Grant No. BIO-00142/2001.
References [1] Rodgers RJ. Animal models of ‘anxiety’: where next? Behav Pharmacol 1997;8:477 – 96. [2] Griebel G. Is there a future for neuropeptide receptor ligands in the treatment of anxiety disorders? Pharmacol Ther 1999;82:1 – 61. [3] Bilkei-Gorzo A, Gyertyan I, Levay G. mCPP-induced anxiety in the light – dark box in rats—a new method for screening anxiolytic activity. Psychopharmacology 1998;136:291 – 8. [4] Adamec RE, Shallow T, Budgell J. Blockade of CCK(B) but not CCK(A) receptors before and after the stress of predator exposure prevents lasting increases in anxiety-like behavior: implications for anxiety associated with posttraumatic stress disorder. Behav Neurosci 1997;111:435 – 49. [5] Haller J, Leveleki C, Halasz J, Baranyi J, Makara GB. The effect of glucocorticoids on the anxiolytic efficacy of buspirone. Psychopharmacology 2001;157:388 – 94. [6] Koks S, Mannisto PT, Bourin M, Shlik J, Vasar V, Vasar E. Cholecystokinin-induced anxiety in rats: relevance of pre-experimental stress and seasonal variations. J Psychiatry Neurosci 2000;25: 33 – 42. [7] McNaughton N, Panickar KS, Logan B. The pituitary – adrenal axis and the different behavioral effects of buspirone and chlordiazepoxide. Pharmacol Biochem Behav 1996;54:51 – 6. [8] Adamec R, Kent P, Anisman H, Shallow T, Merali Z. Neural plasticity, neuropeptides and anxiety in animals—implications for understanding and treating affective disorder following traumatic stress in humans. Neurosci Biobehav Rev 1998;23:301 – 18. [9] Koolhaas JM, Meerlo P, De Boer SF, Strubbe JH, Bohus B. The temporal dynamics of the stress response. Neurosci Biobehav Rev 1997;21:775 – 82. [10] Meerlo P, De Boer SF, Koolhaas JM, Daan S, Van den Hoofdakker RH. Changes in daily rhythms of body temperature and activity after a single social defeat in rats. Physiol Behav 1996;59:735 – 9. [11] van Dijken HH, de Goeij DC, Sutanto W, Mos J, de Kloet ER, Tilders FJ. Short inescapable stress produces long-lasting changes in the brain – pituitary – adrenal axis of adult male rats. Neuroendocrinology 1993;58:57 – 64. [12] van Dijken HH, Mos J, van der Heyden JA, Tilders FJ. Characterization of stress-induced long-term behavioural changes in rats: evidence in favor of anxiety. Physiol Behav 1992;52:945 – 51. [13] Johnson HD, LaVoie JC, Spenceri MC, Mahoney-Wernli MA. Peer conflict avoidance: associations with loneliness, social anxiety and social avoidance. Psychol Rep 2001;88:227 – 35. [14] Whittal ML, Goetsch VL. The impact of panic expectancy and social demand on agoraphobic avoidance. Behav Res Ther 1997;35: 813 – 21. [15] Rickels K, Rynn M. Overview and clinical presentation of generalized anxiety disorder. Psychiatr Clin North Am 2001;24:1 – 17. [16] Cutler MG. Comparison of the effects of yohimbine and clonidine on the behaviour of female mice during social encounters in an ‘‘approach-avoidance’’ situation. Neuropharmacology 1993;32: 411 – 7. [17] Maestripieri D, Martel FL, Nevison CM, Simpson MJ, Keverne EB. Anxiety in rhesus monkey infants in relation to interactions with their mother and other social companions. Dev Psychobiol 1991;24: 571 – 81. [18] Meerlo P, Overkamp GJ, Daan S, Van Den Hoofdakker RH, Koolhaas
332
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
JM. Changes in behaviour and body weight following a single or double social defeat in rats. Stress 1996;1:21 – 32. [19] File SE, Hyde JR. Can social interaction be used to measure anxiety? Br J Pharmacol 1978;62:19 – 24. [20] Ruis MA, te Brake JH, Buwalda B, De Boer SF, Meerlo P, Korte SM,
et al. Housing familiar male wildtype rats together reduces the longterm adverse behavioural and physiological effects of social defeat. Psychoneuroendocrinology 1999;24:285 – 300. [21] Silverman JJ. Post-traumatic stress disorder. Adv Psychosom Med 1986;16:115 – 40.
Stress-induced social avoidance: A new model of stress-induced anxiety? J. Haller*, N. Bakos Institute of Experimental Medicine, P.O. Box 67, 1450 Budapest, Hungary Received 25 February 2002; received in revised form 25 March 2002; accepted 23 July 2002
Abstract We have studied the long-term behavioral effects of a single stressor in male rats by using an approach/avoidance situation as the behavioral endpoint. A single exposure to social defeat or electric shocks was used as stressors. Behavioral testing was performed in a twocompartment cage divided by an opaque wall and connected by a short tunnel. The larger compartment contained an unfamiliar male rat that was separated from the rest of the compartment by a transparent, perforated Plexiglas wall. The subject was placed in the small compartment and allowed to explore the cage for 5 min. The test was performed on Days 1, 5, or 10 after stress application. Unstressed rats spent 90% of time in the large compartment that contained the unfamiliar male. Social defeat dramatically reduced the exploration of the large compartment, without time-related changes in this response. A mild electric shock had a similar effect that lasted more than 5 days but less than 10 days. The exploration of an empty cage was significantly less inhibited by stress than the exploration of a cage that contained the stimulus rat. The test could be applied repeatedly in the same rat, without major changes in the response. Chlordiazepoxide applied 1 h before behavioral testing abolished completely the stress-induced behavioral deficit. We suggest that the model can be used for studying the effects of various compounds on stress-induced anxiety. D 2002 Elsevier Science Inc. All rights reserved. Keywords: Stress; Anxiety; Anxiolytic; Model; Rat
1. Introduction Anxiety and the anxiolytic potential of various compounds are studied mostly in animals with basal anxiety levels. This in sharp contrast with the study of other disorders (e.g., depression, schizophrenia), where the respective condition is induced by behavioral or pharmacological techniques. Anxiety tests involving anxious animals did not become popular probably because (i) benzodiazepines, for which the available laboratory tests were developed, provide reliable results in animals with basal anxiety levels, and (ii) the widely used chronic stress paradigms that lead to anxiety are laborious. However, the advent of new classes of anxiolytics questions the applicability of old methods, mostly because anxiolytic agents other than benzodiazepines provide unreliable results with the usual anxiety tests [1,2]. Recently, a model
*
Corresponding author. Tel.: +36-1-210-9406; fax: +36-1-210-9951. E-mail address: [email protected] (J. Haller).
has been developed for the study of anxiolytics in animals that were made anxious by the anxiogenic compound mchlorophenyl-piperazine [3]. Despite the interesting results, one wonders what anxiety types are modeled by the application of this compound, the effects of which occur rapidly and are short-lived. More promising are models that make use of the anxiogenic potential of stress. Stress is among the aetiological factors of anxiety and human patients are frequently exposed to stressors during treatment. Moreover, anxiety increases the sensitivity towards everyday stressors. In addition, stress was shown to affect the efficacy of many compounds that affect anxiety (e.g., serotonergic and cholecystokininergic compounds) [4– 7]. Regarding the effects of stress on anxiety and the efficacy of anxiolytics, single stressors provide a good alternative to the time-consuming and laborious chronic stress procedures. It has been shown that strong single stressors have surprisingly long-term consequences in rats. A single exposure to electric shocks, social defeat, or a cat affects stress reactivity, neurochemical measures, and behavior for days and even weeks [8– 11]. Data suggest that the long-
0031-9384/02/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved. PII: S 0 0 3 1 - 9 3 8 4 ( 0 2 ) 0 0 8 6 0 - 0
328
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
term consequences of a single stressor include an increase in anxiety [8,12]. In the present work, we have studied the long-term effects of single stressors on social avoidance. We have also assessed the possibility of developing an anxiety model based on the long-term consequences of single stressors. We have applied electric shocks or social defeat to male rats and investigated the time course of the resulting behavioral deficit in a social approach/avoidance situation. The reasons for choosing this testing procedure were the following: (i) It has been shown that social avoidance is associated with several types of human anxiety, including social anxiety, panic, and generalized anxiety disorder (GAD) [13 – 15]. Noteworthy, social avoidance is common in GAD patients lacking phobic avoidant behavior and which are consequently diagnosed as having GAD and not social phobia [15]. (ii) The widely used anxiety tests (e.g., the plus-maze, light/dark, open-field, passive avoidance tests) make heavy use of approach/avoidance behavior towards aversive stimuli. It has been also shown that social avoidance is increased by anxiogenic drugs in both rats and monkeys [16,17]. In addition, a single social defeat resulted in a reduction of social contacts during a subsequent encounter with an unfamiliar, nonaggressive conspecific [18]. (iii) The approach/avoidance situation provides a simple testing procedure in a social context. There are a large number of nonsocial anxiety tests, but only one that investigates the social dimension of anxiety (the social interaction test [19]). This test provides reliable and valuable information on anxiety, but is rather laborious and requires special training. One objective of the present work was to find easily applicable methods for measuring the long-term, anxiety-related consequences of single stressors.
larger resident rat. The aggressiveness of residents was checked previously and only residents showing high levels of attacks were used. The encounter lasted 30 min. The number of biting attacks delivered by residents was recorded. Electric shocks were delivered via the grid floor of a Plexiglas box (30 30 30 cm). An alternating current of 100 V and 8 mA was applied. Shock duration was 0.01 s; shocks were delivered in trains lasting 1 s with an intershock interval of 0.02 s. Two trains per minute were delivered. Shocks were delivered for 5 min, i.e., each rat received 10 shock trains of 1s. Animals returned to their home cage after stress exposure and were left undisturbed until testing. 2.3. The approach/avoidance test The apparatus consisted of two chambers connected by a tunnel (diameter: 10 cm, length: 5 cm) (Fig. 1). The dimensions of the small and large chambers were 40 20 20 and 40 40 20 cm, respectively. The tunnel was closed by a sliding door. A large male rat was enclosed in a subchamber of the large chamber (40 15 15 cm), which was separated from the rest of the chamber by a transparent perforated Plexiglas wall. The apparatus was lit by dim red light. The test started by placing the subject in the small chamber. After an acclimation period of 3 min, the sliding door was removed and the subject was allowed to explore the whole apparatus for 5 min. Behavior was video-recorded from above and later analyzed. The camera was placed 2.5 m above floor. Two variables were recorded: the number of visits made to the large compartment and the time spent in the large compartment.
2. Materials and methods 2.1. Animals Subjects were male Wistar rats weighing approximately 300 g (Charles River Laboratories, Hungary). They were kept in a 12:12-h reverted day –night schedule, with lights off at 1000 h. Rats were maintained under standard laboratory conditions (temperature: 21 ± 2 °C; humidity: 60% ± 10%). Standard laboratory food (Charles River) and tap water were available ad libitum. All animals were isolated 3 days before stress application and remained isolated throughout, because it has been shown that social isolation is a prerequisite of the long-term changes induced by single stressors [20]. In addition, the approach/avoidance behavior of socially housed animals was highly variable in preliminary experiments. 2.2. Stress procedures Social stress was induced by exposing subjects to inescapable attacks in the home cage (40 50 60 cm) of a
Fig. 1. The test apparatus for social avoidance: 1 = small chamber; 2 = tunnel with sliding door; 3 = large chamber; 4 = transparent, perforated Plexiglas wall; 5 = subchamber of the unfamiliar male. The subject is placed in the small chamber and, after a 3-min habituation, the sliding door is removed. The rat is allowed to explore the whole apparatus for 5 min.
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
329
2.4. Experimental design Both stress application and testing were performed in the early hours of the dark (active) period. Animals were individually housed 3 days before stress application and individual housing was maintained throughout. Six experiments were run. In Experiment 1, rats were socially stressed and tested 1, 5, or 10 days thereafter in the approach/avoidance paradigm. Control rats were left undisturbed in their home cage. Different rats were used for each time-point (n = 6 per group). In Experiment 2, the rats were socially stressed and left undisturbed for 10 days. On the 10th day, they were injected either with vehicle or chlordiazepoxide-hydrochloride (Sigma). The latter was dissolved in saline and administered intraperitoneally. The dose was 5 mg/kg. One hour after the injections, rats were tested in the approach/avoidance paradigm (n = 8 per group). In Experiments 3 and 4, rats were exposed to electric shocks and the behavioral test was run 1, 5, or 10 days after shock exposure. In Experiment 3, different rats were used for each time-point (n = 8 per group). In Experiment 4, rats were tested three times at 1, 5, and 10 days (n = 8 per group). In Experiment 5, the effects of chlordiazepoxide (5 mg/kg) was tested in animals stressed by electric shocks, 5 days after stress exposure (n = 8 per group). The lag of 5 days was chosen, because the previous experiment showed that the effects of electric shocks are shorter lasting than the effects of social stress. Experiment 6 assessed the role of the stimulus rat in the electric shock-induced avoidance of the large compartment. Rats were tested 1 day after shock exposure by exposing them to either an empty cage or a large compartment containing the stimulus rat (n = 8 per group). 2.5. Statistics In Experiments 1, 3, and 4, results were analyzed by twofactor ANOVA. Factor 1 was stress application, while Factor 2 was time. Data were analyzed by one-factor ANOVA in Experiments 2, 5, and 6.
3. Results Unstressed controls spent most of the time (around 90%) in the large compartment, by investigating the box containing the unfamiliar stimulus rat (Figs. 2 and 3). This behavior was very robust, with small individual variability. Results were almost identical in the five experiments performed. The number of visits performed was between four and six and also showed large consistency. Socially stressed animals received 5.72 ± 0.84 attacks during the stress period. The time spent in the large compartment was more than halved in stressed rats and this effect did not show time-related changes (Factor 1
Fig. 2. The effect of social stress on social avoidance. Stress was applied 1, 5, or 10 days before testing, as indicated. Chlordiazepoxide (5 mg/kg) was applied 1 h before behavioral testing. * Statistically significant differences compared with unstressed animals in post hoc comparisons.
[stress] = 30.05, P < .0001; Factor 2 [time] = 0.51, P < .6; Stress Time interaction = 0.06, P < .9) (Fig. 2). The number of visits was not changed (Factor 1 = 0.89, P < .3; Factor 2 = 1.11, P < .3; Stress Time interaction = 0.6, P < .6). In the second experiment, behavioral effects were tested 10 days after defeat. Social stress produced the effects observed earlier; moreover, the number of visits was also decreased in this experiment. A similar decrease was noticed in Experiment 1, but differences were not significant. Chlordiazepoxide applied 1 h before behavioral testing abolished stress-induced changes completely ( F = 110.69, P < .0001 for time spent in the large compartment, and F = 5.83, P < .01 for the number of visits made to the large compartment). Electric shocks decreased the duration of large compartment visits dramatically on Days 1 and 5 (Fig. 3). In contrast to social stress, the effects of electric shocks disappeared after 10 days (Factor 1 = 117.20, P < .0001; Factor 2 = 36.08, P < .0001; Stress Time interaction = 37.87, P < .0001). Similar results were obtained when the same animals were repeatedly tested on Days 1, 5, and 10. The only difference was that the effect was still significant at 10 days, i.e., the effect of stress was diminished but significant on the 10th day (Factor 1 = 294.6, P < .0001; Factor 2 = 17.47, P < .001; Stress Time interaction = 15.65, P < .001). The number of visits showed some time-related changes in this case, but the stress had no effect on this variable (Factor 1 = 1.33, P < .3;
330
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
Factor 2 = 5.92, P < .004; Stress Time interaction = 2.83, P < .07). Chlordiazepoxide effects were tested after 5 days and, again, the behavioral effects of stress were completely abolished by chlordiazepoxide ( F = 232.6, P < .0001 for time spent in the large compartment; F = 0.17, P < .8 for the number of visits). Electric shocks reduced the duration of the exploration of the large compartment, even when this did not contain a stimulus rat ( F = 38.17, P < .0001; Fig. 4). However, the exploration of the large compartment lasted significantly less, when it contained a stimulus rat ( F = 34.7, P < .0001). The frequency of large compartment visits did not show significant variations. Thus, the electric shock-
Fig. 4. The effect of the stimulus rat on the electric shock-induced avoidance of the large compartment. Stress was applied 1 day before testing. Data were expressed as percentage of change compared with control. Stressed animals showed a significant avoidance of the large compartment, even when this was empty. However, the avoidance was significantly stronger, when the stimulus rat was present.
induced avoidance of the large compartment showed a strong social component.
4. Discussion
Fig. 3. The effect of exposure to electric shocks on social avoidance. Stress was applied 1, 5, or 10 days before testing, as indicated. Upper panels = different animals used for each time-point; middle panels = animals tested on Day 1 and retested on Days 5 and 10; lower panels = the effect of chlordiazepoxide (5 mg/kg) applied 1 h before behavioral testing on Day 5. * Statistically significant differences compared with unstressed animals in post hoc comparisons.
In the approach/avoidance situation employed, naive rats spent approximately 90% of time in the compartment containing the stimulus rat. Socially stressed rats spent less than 50% of time in the compartment containing the stimulus rat, with a tendency of reducing also the number of visits. This effect did not change during the 10 days of the experiment. Electrically shocked rats showed a similar behavior, but the effect lasted less then 10 days. Repeating the test in the same animals did not lower the effects of electric shocks; moreover, a slight increase in the effect was noticed. The stress-induced avoidance of the large compartment showed a strong social component. Chlordiazepoxide abolished the consequences of both social stress and electric shocks. It has been shown earlier, that single stressors have longterm consequences in the rat. A single exposure to social defeat, electric shocks, or a cat, induced long-term neurochemical, endocrinological, and behavioral changes [8 –11]. Our results are in agreement with these data, since a single social defeat or electric shock induced a lasting behavioral effect in our rats. The single difference was the relatively fast ( < 10 days) recovery from the effects of electric shocks,
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
which is in contrast with earlier findings [11,12]. The fast recovery may be explained by the mild electric shocks applied, since our rats received fewer shocks and the duration of stress application was lower. One can assume that the effects of more frequent shocks administered for a longer period may have longer lasting effects. Noteworthy, the effects of stress were smaller but significant 10 days after shock exposure, when the rats were repeatedly tested. This suggests that testing (the contact with an unfamiliar male) prolonged the effects of stress. The behavior of controls was not changed by repeated testing. It has been suggested earlier that the long-term behavioral effects of single stressors are caused by anxiety, which is a consequence of a long-term increase in stress responsiveness [4,8,12]. This assumption is supported by the strong effect of chlordiazepoxide on the stress-induced social avoidance in our experiment. Traumatic events in humans lead to a long-term increase in anxiety and generate the symptoms of post-traumatic stress disorder [21]. Thus, the long-term effects of single stressors may model a human type of anxiety. Taken conjointly, our data suggest that stress-induced social avoidance may be used as a model of stress-induced anxiety and, prospectively may be used to assess the effects of anxiolytic agents. This assumption is supported by several lines of reasoning. (i) Stress is an important component of anxiety, either as an aetiological factor or as a factor concurrent with pharmacological treatment. (ii) Social avoidance is a common symptom in various anxiety disorders, including social anxiety, panic, and generalized anxiety [13 –15]. In addition, social avoidance is increased by anxiogenic drugs in animals [16,17]. Noteworthy, the avoidance of (nonsocial) aversive stimuli is the main variable measured in the current anxiety tests. (iii) The stress-induced social avoidance model allows the study of drug effects in animals with increased anxiety. This is in sharp contrast with the current practice, which involves the study of animals showing natural anxiety levels. It is important to emphasize that stress-induced social avoidance occurred also in a heterotypic situations, i.e., rats showed social avoidance even when the stressor was nonsocial (it was an electric shock). (iv) The model allows the study of the effects of anxiolytics at multiple levels: during stress exposure, chronically after stress exposure, and acutely before testing. It is important to emphasize that the test appears repeatable, i.e., it may allow the repeated testing of the same animals. This may lower sample size in chronic studies. (v) Testing is simple, rapid, and does not require special training. In conclusion, single stressors induce long-lasting social avoidance in rats. The effect appears reversible by anxiolytic agents, which suggest that stress-induced social avoidance may be used as model for the study of the interaction among stress, anxiety, and pharmacological anxiolysis.
331
Acknowledgements This study was supported by OTKA Grant No. T 032345, ETT Grant No. 286/2001, and OMFB Biotechnology Grant No. BIO-00142/2001.
References [1] Rodgers RJ. Animal models of ‘anxiety’: where next? Behav Pharmacol 1997;8:477 – 96. [2] Griebel G. Is there a future for neuropeptide receptor ligands in the treatment of anxiety disorders? Pharmacol Ther 1999;82:1 – 61. [3] Bilkei-Gorzo A, Gyertyan I, Levay G. mCPP-induced anxiety in the light – dark box in rats—a new method for screening anxiolytic activity. Psychopharmacology 1998;136:291 – 8. [4] Adamec RE, Shallow T, Budgell J. Blockade of CCK(B) but not CCK(A) receptors before and after the stress of predator exposure prevents lasting increases in anxiety-like behavior: implications for anxiety associated with posttraumatic stress disorder. Behav Neurosci 1997;111:435 – 49. [5] Haller J, Leveleki C, Halasz J, Baranyi J, Makara GB. The effect of glucocorticoids on the anxiolytic efficacy of buspirone. Psychopharmacology 2001;157:388 – 94. [6] Koks S, Mannisto PT, Bourin M, Shlik J, Vasar V, Vasar E. Cholecystokinin-induced anxiety in rats: relevance of pre-experimental stress and seasonal variations. J Psychiatry Neurosci 2000;25: 33 – 42. [7] McNaughton N, Panickar KS, Logan B. The pituitary – adrenal axis and the different behavioral effects of buspirone and chlordiazepoxide. Pharmacol Biochem Behav 1996;54:51 – 6. [8] Adamec R, Kent P, Anisman H, Shallow T, Merali Z. Neural plasticity, neuropeptides and anxiety in animals—implications for understanding and treating affective disorder following traumatic stress in humans. Neurosci Biobehav Rev 1998;23:301 – 18. [9] Koolhaas JM, Meerlo P, De Boer SF, Strubbe JH, Bohus B. The temporal dynamics of the stress response. Neurosci Biobehav Rev 1997;21:775 – 82. [10] Meerlo P, De Boer SF, Koolhaas JM, Daan S, Van den Hoofdakker RH. Changes in daily rhythms of body temperature and activity after a single social defeat in rats. Physiol Behav 1996;59:735 – 9. [11] van Dijken HH, de Goeij DC, Sutanto W, Mos J, de Kloet ER, Tilders FJ. Short inescapable stress produces long-lasting changes in the brain – pituitary – adrenal axis of adult male rats. Neuroendocrinology 1993;58:57 – 64. [12] van Dijken HH, Mos J, van der Heyden JA, Tilders FJ. Characterization of stress-induced long-term behavioural changes in rats: evidence in favor of anxiety. Physiol Behav 1992;52:945 – 51. [13] Johnson HD, LaVoie JC, Spenceri MC, Mahoney-Wernli MA. Peer conflict avoidance: associations with loneliness, social anxiety and social avoidance. Psychol Rep 2001;88:227 – 35. [14] Whittal ML, Goetsch VL. The impact of panic expectancy and social demand on agoraphobic avoidance. Behav Res Ther 1997;35: 813 – 21. [15] Rickels K, Rynn M. Overview and clinical presentation of generalized anxiety disorder. Psychiatr Clin North Am 2001;24:1 – 17. [16] Cutler MG. Comparison of the effects of yohimbine and clonidine on the behaviour of female mice during social encounters in an ‘‘approach-avoidance’’ situation. Neuropharmacology 1993;32: 411 – 7. [17] Maestripieri D, Martel FL, Nevison CM, Simpson MJ, Keverne EB. Anxiety in rhesus monkey infants in relation to interactions with their mother and other social companions. Dev Psychobiol 1991;24: 571 – 81. [18] Meerlo P, Overkamp GJ, Daan S, Van Den Hoofdakker RH, Koolhaas
332
J. Haller, N. Bakos / Physiology & Behavior 77 (2002) 327–332
JM. Changes in behaviour and body weight following a single or double social defeat in rats. Stress 1996;1:21 – 32. [19] File SE, Hyde JR. Can social interaction be used to measure anxiety? Br J Pharmacol 1978;62:19 – 24. [20] Ruis MA, te Brake JH, Buwalda B, De Boer SF, Meerlo P, Korte SM,
et al. Housing familiar male wildtype rats together reduces the longterm adverse behavioural and physiological effects of social defeat. Psychoneuroendocrinology 1999;24:285 – 300. [21] Silverman JJ. Post-traumatic stress disorder. Adv Psychosom Med 1986;16:115 – 40.