- Email: [email protected]
Naltrexone-reversible pain suppression in the isolated attacking mouse
BEHAVIORAL AND NEURAL BIOLOGY 50, 354--360
(1988)
BRIEF REPORT Naltrexone-Reversible Pain Suppression in the Isolated Attacking Mouse BERT SIEGFRIED AND HANS-RUDOLF FRISCHKNECHT 1
Institute of Pharmacology, University of Zurich, Gloriastrasse 32, CH-8006 Zurich, Switzerland Fighting pairs of isolated DBA/2 mice showed a significant increase in tailflick response latencies independent of whether opponents were losing or winning the combat. The effect lasted less than 10 rain in both animals. Elevated pain thresholds were also found in isolates that attacked a nonaggressive conspecific, and were prevented by naltrexone (0.2 mg/kg), while a larger dose (1.0 mg/kg) inhibited the attack behavior. A small increase in pain threshold was observed after exposure of isolates to the test box alone, while isolation per se had no effect on baseline tail-flick latencies. The data demonstrate that endogenous pain suppressing systems are activated during attack and suggest that this opioidmediated antinociception is a correlate of the isolation syndrome, reflecting enhanced arousal of the attacking animal. © 1988Academic Press, Inc.
Among several stressful procedures that activate pain suppressive mechanisms, murine aggressive encounters are of particular interest as they represent a biologically meaningful experimental setting. Typically, in this social stress model an aggressive resident mouse attacks an unexperienced intruder. The attacked and bitten intruder reacts with an opioid antagonist-sensitive analgesia, the amplitude and duration of which depends on the impact on the stressor (Miczek, Thompson, & Shuster, 1982; Rodgers & Randall, 1987; Siegfried, Frischknecht, Riggio, & Waser 1987). Less attention has been paid to pain threshold changes of attacking animals. Two of three studies reported unaltered levels of nociception in attacking mice (Miczek et al., 1982; Teskey & Kavaliers, 1987), while the other found a hyperalgesic effect (Rodgers & Hendrie, 1983). The present experiments assessed possible nociceptive changes in attacking isolated DBA/2 mice, since social deprivation causes well described changes in behavioral reactivity, including aggression, in this particular l Reprint requests should be addressed to B. Siegfried. 354
0163-1047/88 $3.00 Copyright © 1988 by Academic Press, Inc. All rights of reproduction in any form reserved.
NALTREXONE-REVERSIBLE PAIN SUPPRESSION
355
strain (Frischknecht, Siegfried, & Waser, 1985; Rodgers & Randall, 1987; Siegfried, Alleva, Oliverio, & Puglisi-Allegra, 1981). Male DBA/2J/Zur (DBA) mice (Institut ft~r Zuchthygiene, Universit~it Z~rich, Switzerland) were used. Upon arrival, at an age of 6-7 weeks, they were housed either in groups of 12 animals per cage (transparent Macrolon cages, 42 x 26 x 17 cm) or in individual cages (24 x 14 x 13 cm). They had free access to food and water and were kept on a natural light-dark cycle. After 6 weeks, the aggressiveness of isolated DBA mice was assessed by introducing for 10 min into their home cage a group-housed DBA mouse of the same age, twice daily on 2 successive days. Only animals consistently attacking the intruder within the first 2 min were used for the experiments (70-90% of all animals). Analgesia was tested by using a procedure and a tail-flick apparatus described recently (Siegfried et al., 1987). All experiments were run between 1 and 5 PM.
A first experiment compared nociceptive thresholds in eight pairs of losing and winning animals. After assessment of the preexposure tailflick latency (TFL), two aggressive isolated DBA mice were simultaneously placed into a transparent test box (24 x 24 x 30 cm). Two observers, each of them directing his attention to one animal, recorded the following parameters: latency to fight, latency to lose (defined as the first unambigously displayed upright submissive posture concomitant with vocalization), and numbers of bites received. Immediately after the criterion of losing was reached by an animal, both loser and winner were removed from the test box and were individually placed into a holding cage (26 x 20 x 14 cm). Postexposure TFLs were measured 1 and 10 min later. Between the two measurements the animals were housed in the holding cage. As control (repeated tail-flick measurement) another group of isolated animals (n = 12) was exposed to two tail-flick tests, spaced by a 6-min interval. Between the trials these animals stayed in their home cages. In a second experiment two groups of aggressive isolated DBA mice were exposed for 5 min to the test box either alone (n = 11) or together with a behaviorally unexperienced group-housed DBA mouse of the same age (n = 21). The number of attack bites directed to the group-housed DBA mouse was counted. As in the first experiment, TFLs of isolates were taken before and 1 min after exposure. In addition, preexposure TFLs were measured in group-housed DBA mice (n = 21) in order to assess whether isolation per se would alter basal nociceptive thresholds. A group of isolated DBA mice that was exposed to two tail-flick tests, spaced by a 6-rain interval, served as control (n = 11). A third experiment investigated the effect of naltrexone on pain thresholds in the attacking DBA mouse. Aggressive isolated mice were injected ip either with physiological saline solution (n = 7) or with naltrexone (0.2 mg/kg, n = 9; 1.0 mg/kg, n = 5, expressed as the base) in a volume
356
SIEGFRIED AND FRISCHKNECHT
1 min after losing
o
--
10 min after losing
5
_s3
~2
rt]l Control
Loser
Winner
Loser
I Winner
F]~. I. Time course of fighting-induced pain suppression (expressed as the difference between post- minus preexposure tail-flick latency; median, interquartile range) in losing and winning isolated DBA/2 mice. Preexposure TFLs did not significantly differ between the groups (control, 1.21 (1.11-1.34) s; loser, 1.32 (1.13-1.67) s; winner, 1.19 (1.15-1.53) s; H(2) = 1.03, NS).
of 0.1 ml/10 g body weight. Preexposure TFLs were assessed 10 min later, followed by a 5-min habituation period in the test box. A grouphoused DBA mouse was then introduced and latency to attack and number of bites delivered by the isolated DBA mouse during the 5-min testing period were registered. After the end of exposure the isolated mouse was placed into the holding cage for 1 min before postexposure TFL was measured. Data were analyzed by nonparametric Kruskal-Wallis analysis of variance (ANOVA) and/or by nonparametric Wilcoxon two-sample tests. Significant within-group differences were verified by the Wilcoxon onesample test. The median latency to attack in pairs of isolated mice was 10 (7.520) s and the median latency to lose the fight was 300 (108-373) s. Nociceptive thresholds of losing and winning animals are shown in Fig. 1. ANOVA over the ATFLs of control, loser, and winner groups revealed a significant fighting effect 1 min after the criterion of losing was reached (H(2) = 14.6, p < .001) and individual group comparisons showed significantly higher ATFLs in loser and winner animals when compared to controls (p < .005). The effect of fighting was also reflected by the significant positive values of the ATFLs within the loser and winner groups 1 rain (p < .02), but not 10 min, after the criterion of losing was reached. No significant difference was evident between the ATFLs of loser and winner animals, despite the significantly higher number of bites received by losers when compared to winners (median number of bites: losers 7 (6-8) bites, winners 1.5 (1-2.5) bites, p < .001).
NALTREXONE-REVERSIBLE PAIN SUPPRESSION
357
5
o
,
3 2
2_ <~ t
I Control
Cage
9 Bites 20 Bites
FIG. 2. Changes in nociception (expressed as the difference between post- minus preexposure tail-flick latency; median, interquartile range) in isolated DBA/2 mice after repeated tail-flick measure (control), after a 5-rain exposure to the test box (cage), or after attacking a group-housed DBA mouse during 5 rain (9 bites, 20 bites). Preexposure TFLs did not significantly differ between the groups (control, 1.19 (1.12-1.39) s; cage, 1.30 (1.26-1.34) s; 9 bites, 1.26 (1.13-1.30) s; 20 bites, 1.32 (1.25-1.95) s; H(3) = 5.46, NS).
In Experiment 2, isolated mice that attacked a nonaggressive grouphoused conspecific were split, according to their aggressive behavior, into two groups delivering a median number of 9 (5-11) bites (n = 10) or 20 (14-21) bites (n = 1l) within the 5-min testing period. Nociceptive thresholds of the various groups are illustrated in Fig. 2. ANOVA revealed a significant exposure effect (H(3) = 19.3, p < .001) and individual group comparisons showed a significant higher ATFL, when compared to controls, in animals exposed to the test cage (p < .05) or in animals delivering a low or high number of attack bites (p < .005). Antinociception in the attacking animal did not depend on the number of bites directed to grouphoused mice, but was significantly more pronounced in attacking animals than in animals exposed to the test cage only (9 bites vs cage, p < .1; 20 bites vs cage, p < .01). The median preexposure TFL of group-housed mice (1.20 (1.11-1.35) s) was in the same range as the preexposure values of the isolated animals mentioned in the legend to Fig. 2. In Experiment 3, injection of 1.0 mg/kg naltrexone completely prevented attacks in isolated mice. Animals treated with a lower dose of 0.2 mg/kg naltrexone or with saline showed an identical median attack latency of 5 (5-25) s, and the number of bites directed to the intruder was not significantly different between them (NaC1, 5 (4-11) bites; naltrexone, 0.2 mg/kg, 8 (5-13) bites). Figure 3 depicts the effect of naltrexone on increase in TFL after social confrontation. Animals injected with 0.2 mg/kg naltrexone showed a significantly smaller ATFL (p < .02) when compared to the NaC1 group. In fact, 0.2 mg/kg naltrexone completely
358
SIEGFRIED AND FRISCHKNECHT
g
1.0
:-2 0.5 _L
~, I I NoCI
'
I
Nix0.2
Ntx 1.0
FIG. 3. Social encounter-induced changes in nociception (expressed as the difference between post- minus preexposure tail-flick latency; median, interquartile range) in isolated DBA/2 mice after ip injection of NaC1 or naltrexone (0.2, 1.0 mg/kg). Preexposure TFLs did not significantly differ between the groups (NaC1, 1.53 (1.36-1.68) s; 0.2 mg/kg naltrexone, 1.30 (1.10-1.47) s; 1.0 mg/kg naltrexone, 1.49 (1.31-1.67) s; H(2) = 3.10, NS).
prevented elevation of pain threshold in the attacking animal as indicated by the nonsignificant increase in TFL in this group. The results demonstrate that naltrexone-reversible pain inhibitory mechanisms are activated in the isolated attacking DBA/2 mouse. This supports a previous study that reported a naloxone-sensitive increase in nociceptive thresholds in rats after mouse-killing behavior (Kromer & Dum, 1982), but contrasts with investigations on attacking mice. Miczek et al. (1982) found no tail-flick response changes in CWF residents that attacked B6AF1 mice. Similarly, Teskey and Kavaliers (1987) reported no alterations of hot-plate response latencies in isolated fighting CF-1 mice, while Rodgers and Hendrie (1983) showed a naloxone-sensitive decrease of paw lick latencies in isolated BKW residents attacking a group-housed conspecific. The interpretation of the hyperalgesic effect in the latter study, however, is complicated by the significant isolationinduced increase of baseline hot-plate response latencies. In the present study, no differences in baseline tail-flick response latencies were evident between isolated and group-housed mice. The discrepancy found in the effect of attack on modulation of endogenous pain inhibitory systems favors the view that antinociception in the aggressor is not a general phenomenon but more likely depends on the different strains and/or housing conditions of mice used in this social stress model. Isolated mice develop a complex syndrome, one characteristic of which is the high excitability of animals toward sensory stimuli. In fact, isolated DBA/2 mice, when compared to group-housed mice, showed a higher reactivity upon tactile body stimulation and an increased excitability, as defined by the EEG desynchronization elicited by tactile stimuli under urethane anesthesia (Siegfried et al., 1981). Moreover, the failure of postattack footshock administration to inhibit aggressive behavior of isolated DBA/2 mice 24 h later may be ascribed to an overexcitation, interfering with optimal memory processing (Frischknecht et al., 1985).
NALTREXONE-REVERSIBLE PAIN SUPPRESSION
359
It seems therefore likely that in the present experiments the antinociception observed after exposure to the test box alone, after moderate or vigorous attack, as well as after fighting reflects the high excitability described in isolated DBA/2 mice. This is especially appearing in regard to the possibility that activation of endogenous pain inhibitory systems may be taken as a sensitive index of the level of arousal experienced in a given situation (Tazi, Dantzer, & Le Moal, 1987). Our data suggest participation of endogenous opioids in this overexcitation since correlates of arousal at the physiological (pain suppression), as well as at the behavioral (isolationinduced aggression), level were inhibited by low and high doses of naltrexone, respectively. The high arousability may, speculatively, be ascribed to the fact that social encounter represents a stressful experience to longterm isolated animals and thus naltrexone, at a low dose, antagonized the stress-induced increase in pain threshold, and, at a high dose, prevented social contact from being perceived as a stressful event. The arousal interpretation may explain the fact that antinociception, similar in amplitude and time course, was found in losing and winning animals, despite different numbers of bites received and despite qualitatively different behaviors displayed by the opponents at the end of the aggressive encounter: the losing animal engaged in a defensive upright posture accompanied by vocalization, while the winner showed attack behavior and biting with the absence of vocalization. Increased excitability may have masked the reported relationship between certain defeat postures and the extent of analgesia (Frischknecht & Siegfried, 1988; Rodgers & Randall, 1987). Nevertheless, antinociception in the attacking animal may not only be a correlate of unspecific arousal devoid of any adaptive value, but may optimize goal-directed fighting by suppressing pain-oriented behavior. Further comparison of pain threshold changes in attacking dominant and isolated animals of the same strain will decide whether pain inhibition in the attacking mouse is the expression of an aberrant mechanism or whether it represents a biologically meaningful one. REFERENCES Frischknecht, H. R., & Siegfried, B. (1988). Emergence and development of stress-induced analgesia and concomitant behavioral changes in mice exposed to social conflict. Physiology and Behavior, 44, in press. Frischknecht, H. R., Siegfried, B., & Waser, P. G. (1985). Postaggression footshock inhibits aggressive behavior in dominant but not in isolated mice. Behavioral and Neural Biology, 44, 132-138. Kromer, W., & Dum, J. E. (1980). Mouse-kiliing in rats induces a naloxone-blockable increase in nociceptive threshold. European Journal of Pharmacology, 63, 195-198. Miczek, K. A., Thompson, M. L., & Shuster, L. (1982). Opioid-like analgesia in defeated mice. Science, 215, 1520-1522. Rodgers, R. J., & Hendrie, C. A. (1983). Social conflict activates status-dependent analgesic and hyeralgesic mechanisms in male mice: Effects of naloxone on nociception and behavior. Physiology and Behavior, 30, 775-780.
360
SIEGFRIED AND FRISCHKNECHT
Rodgers, R. J., & Randall, J. I. (1987). Defensive analgesia in rats and mice. The Psychological Record, 37, 335-347. Siegfried, B., Alleva, E., Oliverio, A., & Puglisi-Allegra, S. (1981). Effects of isolation on activity, reactivity, excitability and aggressive behavior in two inbred strains of mice. Behavioural Brain Research, 2, 211-218. Siegfried, B., Frischknecht, H. R., Riggio, G., & Waser, P. G. (1987). Long-term analgesic reaction in attacked mice. Behavioral Neuroscience, 101, 797-805. Tazi, A., Dantzer, R., & Le Moal, M. (1987). Prediction and control of food rewards modulate endogenous pain inhibitory systems. Behavioural Brain Research, 23, 197204. Teskey, G. C., & Kavaliers, M. (1987). Aggression, defeat and opioid activation in mice: Influences of social factors, size and territory. Behavioural Brain Research, 23, 7784.
(1988)
BRIEF REPORT Naltrexone-Reversible Pain Suppression in the Isolated Attacking Mouse BERT SIEGFRIED AND HANS-RUDOLF FRISCHKNECHT 1
Institute of Pharmacology, University of Zurich, Gloriastrasse 32, CH-8006 Zurich, Switzerland Fighting pairs of isolated DBA/2 mice showed a significant increase in tailflick response latencies independent of whether opponents were losing or winning the combat. The effect lasted less than 10 rain in both animals. Elevated pain thresholds were also found in isolates that attacked a nonaggressive conspecific, and were prevented by naltrexone (0.2 mg/kg), while a larger dose (1.0 mg/kg) inhibited the attack behavior. A small increase in pain threshold was observed after exposure of isolates to the test box alone, while isolation per se had no effect on baseline tail-flick latencies. The data demonstrate that endogenous pain suppressing systems are activated during attack and suggest that this opioidmediated antinociception is a correlate of the isolation syndrome, reflecting enhanced arousal of the attacking animal. © 1988Academic Press, Inc.
Among several stressful procedures that activate pain suppressive mechanisms, murine aggressive encounters are of particular interest as they represent a biologically meaningful experimental setting. Typically, in this social stress model an aggressive resident mouse attacks an unexperienced intruder. The attacked and bitten intruder reacts with an opioid antagonist-sensitive analgesia, the amplitude and duration of which depends on the impact on the stressor (Miczek, Thompson, & Shuster, 1982; Rodgers & Randall, 1987; Siegfried, Frischknecht, Riggio, & Waser 1987). Less attention has been paid to pain threshold changes of attacking animals. Two of three studies reported unaltered levels of nociception in attacking mice (Miczek et al., 1982; Teskey & Kavaliers, 1987), while the other found a hyperalgesic effect (Rodgers & Hendrie, 1983). The present experiments assessed possible nociceptive changes in attacking isolated DBA/2 mice, since social deprivation causes well described changes in behavioral reactivity, including aggression, in this particular l Reprint requests should be addressed to B. Siegfried. 354
0163-1047/88 $3.00 Copyright © 1988 by Academic Press, Inc. All rights of reproduction in any form reserved.
NALTREXONE-REVERSIBLE PAIN SUPPRESSION
355
strain (Frischknecht, Siegfried, & Waser, 1985; Rodgers & Randall, 1987; Siegfried, Alleva, Oliverio, & Puglisi-Allegra, 1981). Male DBA/2J/Zur (DBA) mice (Institut ft~r Zuchthygiene, Universit~it Z~rich, Switzerland) were used. Upon arrival, at an age of 6-7 weeks, they were housed either in groups of 12 animals per cage (transparent Macrolon cages, 42 x 26 x 17 cm) or in individual cages (24 x 14 x 13 cm). They had free access to food and water and were kept on a natural light-dark cycle. After 6 weeks, the aggressiveness of isolated DBA mice was assessed by introducing for 10 min into their home cage a group-housed DBA mouse of the same age, twice daily on 2 successive days. Only animals consistently attacking the intruder within the first 2 min were used for the experiments (70-90% of all animals). Analgesia was tested by using a procedure and a tail-flick apparatus described recently (Siegfried et al., 1987). All experiments were run between 1 and 5 PM.
A first experiment compared nociceptive thresholds in eight pairs of losing and winning animals. After assessment of the preexposure tailflick latency (TFL), two aggressive isolated DBA mice were simultaneously placed into a transparent test box (24 x 24 x 30 cm). Two observers, each of them directing his attention to one animal, recorded the following parameters: latency to fight, latency to lose (defined as the first unambigously displayed upright submissive posture concomitant with vocalization), and numbers of bites received. Immediately after the criterion of losing was reached by an animal, both loser and winner were removed from the test box and were individually placed into a holding cage (26 x 20 x 14 cm). Postexposure TFLs were measured 1 and 10 min later. Between the two measurements the animals were housed in the holding cage. As control (repeated tail-flick measurement) another group of isolated animals (n = 12) was exposed to two tail-flick tests, spaced by a 6-min interval. Between the trials these animals stayed in their home cages. In a second experiment two groups of aggressive isolated DBA mice were exposed for 5 min to the test box either alone (n = 11) or together with a behaviorally unexperienced group-housed DBA mouse of the same age (n = 21). The number of attack bites directed to the group-housed DBA mouse was counted. As in the first experiment, TFLs of isolates were taken before and 1 min after exposure. In addition, preexposure TFLs were measured in group-housed DBA mice (n = 21) in order to assess whether isolation per se would alter basal nociceptive thresholds. A group of isolated DBA mice that was exposed to two tail-flick tests, spaced by a 6-rain interval, served as control (n = 11). A third experiment investigated the effect of naltrexone on pain thresholds in the attacking DBA mouse. Aggressive isolated mice were injected ip either with physiological saline solution (n = 7) or with naltrexone (0.2 mg/kg, n = 9; 1.0 mg/kg, n = 5, expressed as the base) in a volume
356
SIEGFRIED AND FRISCHKNECHT
1 min after losing
o
--
10 min after losing
5
_s3
~2
rt]l Control
Loser
Winner
Loser
I Winner
F]~. I. Time course of fighting-induced pain suppression (expressed as the difference between post- minus preexposure tail-flick latency; median, interquartile range) in losing and winning isolated DBA/2 mice. Preexposure TFLs did not significantly differ between the groups (control, 1.21 (1.11-1.34) s; loser, 1.32 (1.13-1.67) s; winner, 1.19 (1.15-1.53) s; H(2) = 1.03, NS).
of 0.1 ml/10 g body weight. Preexposure TFLs were assessed 10 min later, followed by a 5-min habituation period in the test box. A grouphoused DBA mouse was then introduced and latency to attack and number of bites delivered by the isolated DBA mouse during the 5-min testing period were registered. After the end of exposure the isolated mouse was placed into the holding cage for 1 min before postexposure TFL was measured. Data were analyzed by nonparametric Kruskal-Wallis analysis of variance (ANOVA) and/or by nonparametric Wilcoxon two-sample tests. Significant within-group differences were verified by the Wilcoxon onesample test. The median latency to attack in pairs of isolated mice was 10 (7.520) s and the median latency to lose the fight was 300 (108-373) s. Nociceptive thresholds of losing and winning animals are shown in Fig. 1. ANOVA over the ATFLs of control, loser, and winner groups revealed a significant fighting effect 1 min after the criterion of losing was reached (H(2) = 14.6, p < .001) and individual group comparisons showed significantly higher ATFLs in loser and winner animals when compared to controls (p < .005). The effect of fighting was also reflected by the significant positive values of the ATFLs within the loser and winner groups 1 rain (p < .02), but not 10 min, after the criterion of losing was reached. No significant difference was evident between the ATFLs of loser and winner animals, despite the significantly higher number of bites received by losers when compared to winners (median number of bites: losers 7 (6-8) bites, winners 1.5 (1-2.5) bites, p < .001).
NALTREXONE-REVERSIBLE PAIN SUPPRESSION
357
5
o
,
3 2
2_ <~ t
I Control
Cage
9 Bites 20 Bites
FIG. 2. Changes in nociception (expressed as the difference between post- minus preexposure tail-flick latency; median, interquartile range) in isolated DBA/2 mice after repeated tail-flick measure (control), after a 5-rain exposure to the test box (cage), or after attacking a group-housed DBA mouse during 5 rain (9 bites, 20 bites). Preexposure TFLs did not significantly differ between the groups (control, 1.19 (1.12-1.39) s; cage, 1.30 (1.26-1.34) s; 9 bites, 1.26 (1.13-1.30) s; 20 bites, 1.32 (1.25-1.95) s; H(3) = 5.46, NS).
In Experiment 2, isolated mice that attacked a nonaggressive grouphoused conspecific were split, according to their aggressive behavior, into two groups delivering a median number of 9 (5-11) bites (n = 10) or 20 (14-21) bites (n = 1l) within the 5-min testing period. Nociceptive thresholds of the various groups are illustrated in Fig. 2. ANOVA revealed a significant exposure effect (H(3) = 19.3, p < .001) and individual group comparisons showed a significant higher ATFL, when compared to controls, in animals exposed to the test cage (p < .05) or in animals delivering a low or high number of attack bites (p < .005). Antinociception in the attacking animal did not depend on the number of bites directed to grouphoused mice, but was significantly more pronounced in attacking animals than in animals exposed to the test cage only (9 bites vs cage, p < .1; 20 bites vs cage, p < .01). The median preexposure TFL of group-housed mice (1.20 (1.11-1.35) s) was in the same range as the preexposure values of the isolated animals mentioned in the legend to Fig. 2. In Experiment 3, injection of 1.0 mg/kg naltrexone completely prevented attacks in isolated mice. Animals treated with a lower dose of 0.2 mg/kg naltrexone or with saline showed an identical median attack latency of 5 (5-25) s, and the number of bites directed to the intruder was not significantly different between them (NaC1, 5 (4-11) bites; naltrexone, 0.2 mg/kg, 8 (5-13) bites). Figure 3 depicts the effect of naltrexone on increase in TFL after social confrontation. Animals injected with 0.2 mg/kg naltrexone showed a significantly smaller ATFL (p < .02) when compared to the NaC1 group. In fact, 0.2 mg/kg naltrexone completely
358
SIEGFRIED AND FRISCHKNECHT
g
1.0
:-2 0.5 _L
~, I I NoCI
'
I
Nix0.2
Ntx 1.0
FIG. 3. Social encounter-induced changes in nociception (expressed as the difference between post- minus preexposure tail-flick latency; median, interquartile range) in isolated DBA/2 mice after ip injection of NaC1 or naltrexone (0.2, 1.0 mg/kg). Preexposure TFLs did not significantly differ between the groups (NaC1, 1.53 (1.36-1.68) s; 0.2 mg/kg naltrexone, 1.30 (1.10-1.47) s; 1.0 mg/kg naltrexone, 1.49 (1.31-1.67) s; H(2) = 3.10, NS).
prevented elevation of pain threshold in the attacking animal as indicated by the nonsignificant increase in TFL in this group. The results demonstrate that naltrexone-reversible pain inhibitory mechanisms are activated in the isolated attacking DBA/2 mouse. This supports a previous study that reported a naloxone-sensitive increase in nociceptive thresholds in rats after mouse-killing behavior (Kromer & Dum, 1982), but contrasts with investigations on attacking mice. Miczek et al. (1982) found no tail-flick response changes in CWF residents that attacked B6AF1 mice. Similarly, Teskey and Kavaliers (1987) reported no alterations of hot-plate response latencies in isolated fighting CF-1 mice, while Rodgers and Hendrie (1983) showed a naloxone-sensitive decrease of paw lick latencies in isolated BKW residents attacking a group-housed conspecific. The interpretation of the hyperalgesic effect in the latter study, however, is complicated by the significant isolationinduced increase of baseline hot-plate response latencies. In the present study, no differences in baseline tail-flick response latencies were evident between isolated and group-housed mice. The discrepancy found in the effect of attack on modulation of endogenous pain inhibitory systems favors the view that antinociception in the aggressor is not a general phenomenon but more likely depends on the different strains and/or housing conditions of mice used in this social stress model. Isolated mice develop a complex syndrome, one characteristic of which is the high excitability of animals toward sensory stimuli. In fact, isolated DBA/2 mice, when compared to group-housed mice, showed a higher reactivity upon tactile body stimulation and an increased excitability, as defined by the EEG desynchronization elicited by tactile stimuli under urethane anesthesia (Siegfried et al., 1981). Moreover, the failure of postattack footshock administration to inhibit aggressive behavior of isolated DBA/2 mice 24 h later may be ascribed to an overexcitation, interfering with optimal memory processing (Frischknecht et al., 1985).
NALTREXONE-REVERSIBLE PAIN SUPPRESSION
359
It seems therefore likely that in the present experiments the antinociception observed after exposure to the test box alone, after moderate or vigorous attack, as well as after fighting reflects the high excitability described in isolated DBA/2 mice. This is especially appearing in regard to the possibility that activation of endogenous pain inhibitory systems may be taken as a sensitive index of the level of arousal experienced in a given situation (Tazi, Dantzer, & Le Moal, 1987). Our data suggest participation of endogenous opioids in this overexcitation since correlates of arousal at the physiological (pain suppression), as well as at the behavioral (isolationinduced aggression), level were inhibited by low and high doses of naltrexone, respectively. The high arousability may, speculatively, be ascribed to the fact that social encounter represents a stressful experience to longterm isolated animals and thus naltrexone, at a low dose, antagonized the stress-induced increase in pain threshold, and, at a high dose, prevented social contact from being perceived as a stressful event. The arousal interpretation may explain the fact that antinociception, similar in amplitude and time course, was found in losing and winning animals, despite different numbers of bites received and despite qualitatively different behaviors displayed by the opponents at the end of the aggressive encounter: the losing animal engaged in a defensive upright posture accompanied by vocalization, while the winner showed attack behavior and biting with the absence of vocalization. Increased excitability may have masked the reported relationship between certain defeat postures and the extent of analgesia (Frischknecht & Siegfried, 1988; Rodgers & Randall, 1987). Nevertheless, antinociception in the attacking animal may not only be a correlate of unspecific arousal devoid of any adaptive value, but may optimize goal-directed fighting by suppressing pain-oriented behavior. Further comparison of pain threshold changes in attacking dominant and isolated animals of the same strain will decide whether pain inhibition in the attacking mouse is the expression of an aberrant mechanism or whether it represents a biologically meaningful one. REFERENCES Frischknecht, H. R., & Siegfried, B. (1988). Emergence and development of stress-induced analgesia and concomitant behavioral changes in mice exposed to social conflict. Physiology and Behavior, 44, in press. Frischknecht, H. R., Siegfried, B., & Waser, P. G. (1985). Postaggression footshock inhibits aggressive behavior in dominant but not in isolated mice. Behavioral and Neural Biology, 44, 132-138. Kromer, W., & Dum, J. E. (1980). Mouse-kiliing in rats induces a naloxone-blockable increase in nociceptive threshold. European Journal of Pharmacology, 63, 195-198. Miczek, K. A., Thompson, M. L., & Shuster, L. (1982). Opioid-like analgesia in defeated mice. Science, 215, 1520-1522. Rodgers, R. J., & Hendrie, C. A. (1983). Social conflict activates status-dependent analgesic and hyeralgesic mechanisms in male mice: Effects of naloxone on nociception and behavior. Physiology and Behavior, 30, 775-780.
360
SIEGFRIED AND FRISCHKNECHT
Rodgers, R. J., & Randall, J. I. (1987). Defensive analgesia in rats and mice. The Psychological Record, 37, 335-347. Siegfried, B., Alleva, E., Oliverio, A., & Puglisi-Allegra, S. (1981). Effects of isolation on activity, reactivity, excitability and aggressive behavior in two inbred strains of mice. Behavioural Brain Research, 2, 211-218. Siegfried, B., Frischknecht, H. R., Riggio, G., & Waser, P. G. (1987). Long-term analgesic reaction in attacked mice. Behavioral Neuroscience, 101, 797-805. Tazi, A., Dantzer, R., & Le Moal, M. (1987). Prediction and control of food rewards modulate endogenous pain inhibitory systems. Behavioural Brain Research, 23, 197204. Teskey, G. C., & Kavaliers, M. (1987). Aggression, defeat and opioid activation in mice: Influences of social factors, size and territory. Behavioural Brain Research, 23, 7784.