- Email: [email protected]
Recruitment time of conditioned opioid analgesia
Physiology&Behavior, Vol. 39, pp. 135-140.Copyright©PergamonJournals Ltd., 1987. Printedin the U.S.A.
0031-9384/87$3.00 + .00
Recruitment Time of Conditioned Opioid Analgesia L O U I S D. M A T Z E L A N D R A L P H R. M I L L E R 1
Department o f Psychology, SUNY-Binghamton, Binghamton, N Y 13901 R e c e i v e d 23 S e p t e m b e r 1986 MATZEL, L. D. AND R. R. MILLER. Recruitment time of conditioned opioid analgesia. PHYSIOL BEHAV 39(1) 135--140, 1987.---Threeexperiments examined the development of conditioned analgesia in rats exposed to stimuli that had previously been paired with footshock. In Experiment 1, tailflick latencies increased if the tailflick test for analgesia was immediately preceded by 90 sec of exposure to a context in which unsignaled shock had previously been administered. This analgesia was blocked by the opiate antagonist naloxone administered prior to exposure to the context on the test day. Experiment 2 determined that 90 and 300 sec of exposure to the conditioning context immediately prior to testing evoked comparable analgesia as indexed by increased latencies to pawlick in response to thermal stimulation (hotplate). However, no analgesia was evident in animals exposed to the aversive context for 5 sec immediately prior to the hotplate test relative to animals not exposed to that context. In Experiment 3, a 5-sec exposure to the aversive context produced analgesia comparable to a 90-see exposure if an 85-sec delay intervened between the 5-sec exposure and the hotplate test. These results suggest that brief exposure to stimuli previously paired with shock can activate the endogenous opioid system, but the analgesic action of these opioids is delayed. Implications for the role of endogenous opioids in learning are discussed. Analgesia
Hyperalgesia
Endogenous opioids
Naloxone
A stimulus which has previously been paired with aversive stimulation such as footshock will in some instances come to elicit a temporary decrease in an animal's sensitivity to pain. Although this conditioned analgesic effect has been observed under a variety of conditions with both punctate cues and static environmental stimuli as the eliciting agent (e.g., [6, 8, 13, 14]), the specific circumstances under which a conditioned stimulus (CS) for shock will produce analgesia are still uncertain. Although the mechanism by which an initially neutral stimulus comes to elicit an analgesic response appears to conform to many of the principles of classical conditioning [4, 12-14], the procedures that reliably establish associations between a CS and a shock unconditioned stimulus (US) as indexed through traditional conditioned responses (CRs) are not universally successful in establishing analgesia as a CR. For instance, Davis and Henderson [2] and Lysle and Fowler [7] found that a punctate CS that previously had been paired with footshock elicits not analgesia, but rather, hyperalgesia, despite the fact that the footshock US upon which the CR is based itself produces a temporary decrease in responsiveness to painful stimulation [7]. One possible source of this discrepancy is the duration of the exposure to the CS prior to the test for analgesia. While in at least one instance hyperalgesia has been observed [2] in response to a CS of the same duration that was shown to produce analgesia elsewhere [14], instances where hyperalgesia as a CR has been reported have employed CSs of 30
~Requests for reprints should be addressed to Ralph R. Miller.
135
Rats
sec or less (e.g., [2,7]), whereas longer duration CSs more typically appear to elicit analgesia (e.g., [4,8]). The experiments reported here attempted first to establish that the opioid system was responsible for the decreased pain sensitivity in rats elicited by a 90-sec exposure to a context in which unsignaled shock previously had been administered. The analgesic response elicited by this same treatment has been shown elsewhere (Matzel and Miller, manuscript under review) to impair the animal's learning about a CS-shock association. In Experiment 2, animals were exposed to shock as in Experiment 1. Subsequently, the duration of the exposure to the context in which shock had been administered was varied (between 5 and 300 sec) and pain sensitivity was assessed immediately upon removal from that context. Having established in Experiment 2 that short duration (5 sec) exposure to the aversive CS was insufficient to evoke analgesia whereas longer duration exposure (90 or 300 sec) did elicit analgesia, Experiment 3 assessed analgesia after an 85-sec delay that followed a 5-sec exposure to the conditioning context. EXPERIMENT 1 METHOD
Subjects Eighteen male and 18 female, naive, Sprague-Dawleyderived rats bred in our colony served as subjects. Male and
136
MATZEL AND MILLER
o_ _J u_ .J
32
IO
~ 9
z9
£ uJ m
bz
b z
7
w
23
u
_1
$
26
6
~ ao +B/As
+ B/An
+A/As
+A/An
0
TREATMENT
5
90
TREATMENT
II 300
FIG. 1. Mean latency to tailflick after exposure to Context A in Experiment 1. Context A had a common history of unsignaled shock for Groups +A/As and +A/An. Groups +A/As and +B/As received a saline injection prior to exposure to Context A on the test day, whereas Groups +A/An and +B/An received an injection of naloxone. Brackets indicate standard errors.
FIG. 2. Mean latency to pawlick after exposure to Context A in Experiment 2. Context A had a common history of unsignaled shock for all groups. Groups 0, 5, 90, and 300 were exposed to Context A for 0, 5, 90, and 300 sec, respectively, immediately prior to the hotplate test. Brackets indicate standard errors.
female body weight ranges were 310--420 g and 225-305 g, respectively. All animals were water deprived, receiving 10-rain access to water a day approximately 1 hr following completion of any experimental session. This deprivation schedule was imposed gradually during the week preceding initiation of the study and was intended to maintain comparability to previous studies in our laboratory in which analgesia was reported with similar procedures (Matzei and Miller, manuscript under review). Purina Lab Chow was freely available in the home cages. Animals were housed individually in hanging wire-mesh cages in a colony maintained on a 16 hr/8 hr light/dark cycle. All manipulations were conducted near the middle of the light portion of this cycle. Subjects were assigned to treatment groups (ns=9) counterbalancing for sex, litter of origin, and body weight.
bulbs. Unlike Context A, these grids ran the length of the chamber. This apparatus was housed in a large room that was brightly illuminated by overhead fluorescent lights. The ambient sound level was 56 dB (C). These chambers were designated Context B. Previous work in our laboratory found no evidence of generalization between Contexts A and B [9]. Contexts A and B were located in separate rooms in the laboratory. The tailflick device consisted of a metal box that supported a 46× 16-cm (lxw) aluminum plate. Twenty cm from the end of the plate, a variable intensity nominal 150-W projector lamp was mounted under the plate with light from the bulb passing through a 3-mm aperture. A rat could be placed in one of six commercially manufactured Plexiglas restraining tubes (Fisher, 01-280-10) that rested on the surface of the aluminum plate such that the rat's tail extended over the aperture. When activated, the intensity of the lamp was permanently set such that the average latency of naive rats to tailflick was approximately 7 sec.
Apparatus Twelve chambers housed in controlled-environment shells were used. These chambers were clear Plexiglas boxes measuring 22.75 x 8.25 × 13.00 cm (1 x w x h) with electrifiable stainless steel grid floors (0.48-cm-diameter rods spaced 1.91 cm center-to-center). The grids were electrically linked in series by NE-2 neon bulbs. Constant-current footshock was provided by a high-voltage AC source in series with a 1-Mohm resistor and the neon bulbs. A water-filled lick tube was located on one short wall of each chamber, left-right centered and 3.8 cm above the grid floor. Although lick rates were not monitored in the present studies, these lick tubes were present to maintain comparability with other work in our laboratory [10]. Each environmental shell was illuminated by a nominal 7-W (at 120 VAC) white light bulb driven at 57 VAC. Background noise, produced by a ventilation fan, was 70 dB (C) (re. SPL). These chambers were designated Context A. Twelve additional chambers, highly dissimilar from the 12 Context A chambers, were used for administration of USs outside of Context A. They consisted of a 5 0 x l 6 . 5 x 5 0 - c m ( l x w x h ) stall with opaque green Plexiglas walls and floors composed of 0.64-cm diameter, electrifiable grids (spaced 1.59 cm center-to-center) interconnected by NE-2H neon
Procedure On Days 1-4, each subject was placed for 20 min in each of the two contexts. Order of exposure was A-B on Days 1 and 3, and B-A on Days 2 and 4. These initial days were intended to acclimate the animals to the chambers so that novelty of the context by itself would not evoke an analgesic response. Approximately 60 min following these sessions, all animals were placed in one of the Plexiglas restraining tubes for 10 min. The tube rested in place on the tailflick apparatus. The projector bulb was not activated during these adaptation sessions. On Days 5-10, all animals received six days of unsignaled shock presentations. Groups + A / A n and + A / A s received this treatment in Context A (therefore designated by +A), while the remaining animals, Group +B/An and +B/As received identical treatment in Context B (designated +B). During each of these 60-min sessions, all animals received six unsignaled 0.80-mA, 60-Hz, 500-msec footshocks. The footshocks were pseudorandomly scheduled throughout each session with an inter-US interval (ISI) range of 4-12
R E C R U I T M E N T TIME
137 administered, and this analgesia was opioid in nature, as indicated by the lack of conditioned analgesia in animals that received the opiate antagonist naloxone prior to exposure to the context in which they previously had been shocked (Group +A/An). Group + A / A n did not differ from either Group +B/As or Group +B/An for which Context A did not have a history of shock, and Groups +B/As and +B/An did not differ from each other, F s < 1.
~: 26 ¢,.).
1
D..
~ 23
i --
20
z w -J
EXPERIMENT 2 ~7
0
5
9O
5/85
TREATMENT
FIG. 3. Mean latency to pawlick after exposure to Context A in Experiment 3. Context A had a common history of unsignaled shock for all groups. Groups 0, 5, and 90 were exposed to Context A for 0, 5, and 90 sec, respectively, immediately prior to tlae hotplate test. Group 5/85 was exposed to Context A for 5 sec, but the hotplate test was delayed for 85 sec. Brackets indicate standard errors.
min, and a mean ISI of 8 min. On the third and fifth day of this shock exposure, the first shock in the session was scheduled within 5 sec of the start of the session to minimize the possibility o f animals'learning that Context A was safe for the first few minutes of each session. On Day 11, all animals received an intraperitoneal injection of either 1 cc/kg isotonic saline (Groups + A / A s and +B/As, designated by s) or a 10 mg/kg dose of naloxone hydrochloride (Groups + A/An and + B/An, designated by n) dissolved 10 mg/cc saline. Twenty min post-injection, each animal was placed in Context A for 90 sec. Following removal from Context A, animals were individually placed in their restraining tube and were placed on the tailflick apparatus, after which the projector lamp was activated and the latency to withdraw the tail from the beam of light was recorded. Five latencies were recorded for each rat. The first measure was taken approximately midway between the base and tip of the tail; subsequent measures were taken at l-cm increments toward the tip. A trial was terminated i f a tailflick did not occur within 15 sec. In practice, this latency was reached on less than 10% of the trials and was most frequently reached in Group +A/As. RESULTS
AND
DISCUSSION
A single latency was assigned to each subject by averaging its five tailflick latencies. Two subjects were discarded due to illness and two were discarded because they failed to acclimate to the restraining tube. Figure 1 depicts the mean tailflick latencies for each group. A two-factor A N O V A indicated a significant effect of the type of injection administered prior to exposure to Context A on the test day, F(1,28)=9.30, p<0.01, a significant effect of the context in which the animals were shocked, F(1,28)=7.95,p<0.01, and a significant interaction of these two factors, F(1,28)=4.65, p<0.05. Planned comparisons determined that Group + A / A s , which received a saline injection prior to exposure to the context in which it was previously shocked, had longer tailflick latencies than any of the remaining groups, Fs(1,28)~>12.57, ps<0.01. This suggests that analgesia was a conditioned response to the context in which shock was
Consistent with previous work on conditioned analgesia [1, 4, 8], decreased pain sensitivity was observed in Experiment 1 in animals exposed to a context in which unsignaled shock had previously been administered. Moreover, this analgesia was opioid in nature, as it was attenuated by the opiate antagonist naloxone. Experiment 2 attempted to determine if varying durations of exposure to the context paired with shock would influence the magnitude of the analgesia observed. In this study, all animals received unsignaled shock in Context A in a manner identical to Groups +A/An and + A / A s in the previous study. The test for analgesia consisted of placing the animals in Context A for 0, 5, 90, or 300 sec followed immediately by a hotplate test for analgesia. Since the analgesia observed in Experiment 1 was determined to be associative, Experiment 2 included no animals for which Context A was associatively neutral; however, the group that was not exposed to Context A prior to the hotplate test (0-sec exposure) was placed in Context B (which was associatively neutral) for 300 sec prior to testing in order to provide handling comparable to that of the remaining groups. Since this study was intended to examine the temporal onset of analgesia, the hotplate test rather than the tailflick test was employed because the hotplate allowed more immediate testing after removal from Context A, i.e., no time was lost placing the animals in the restraining tubes or using multiple tests. Prior work in our laboratory has found greater sensitivity to analgesia with a single test using the hotplate than with a single test using the tailflick apparatus. METHOD
Subjects Twenty-four male and 24 female rats of the same general description as those used in Experiment 1 served as subjects.
Apparatus The apparatus was identical to that of Experiment 1 except that a hotplate replaced the tailflick apparatus for the analgesia test. The hotplate consisted of a 22 x 22-cm copper plate on top of a constant temperature water bath. The surface was maintained at 50°C. The surface of the plate was enclosed in clear Plexiglas walls with a clear Plexiglas ceiling 9.5 cm above the copper plate that prevented rearing by the animals.
Procedure Initial acclimation to Contexts A and B was conducted on Days 1-4 as in the previous study. Subsequently, all subjects received six days of unsignaled shock exposure in Context A. Shock intensity and scheduling was identical to that of Experiment 1. On Day 11 (the day following the last shock session), the hotplate test for analgesia was conducted.
138
MATZEL AND MILLER
Group 5, Group 90, and Group 300 (ns= 12) were placed in Context A for 5, 90, and 300 sec, respectively. Group 0 (n= 12) was placed in Context B (where no shocks had been administered) for 300 sec. Upon removal from either context, each animal was immediately placed on the hotplate and the latency to lick a hind paw was recorded. This study was conducted in two balanced replications. R E S U L T S A N D DISCUSSION
Mean pawlick latencies are depicted in Fig. 2. An A N O V A with treatment and replications as factors was conducted. Neither replication nor the treatment x replication interaction proved significant, ps>0.10. The effect of treatment was reliable, F(3,40)=3.66, p<0.02. Planned comparisons indicated no difference between Groups 0 and 5, F=0.14, or between Groups 90 and 300, F=0.70. Group 0 differed both from Group 90, F(1,40)=8.17, p<0.01, and from Group 300, F(1,40)=4.08, p<0.05. Similarly, Group 5 differed from Group 90, F(1,40)=6.15, p<0.02. The tendency toward a difference between Groups 5 and 300 was not statistically reliable, F(1,40)=2.70, 0.10
to Context A for 5 sec, one group was exposed for 90 sec, and a third group was not exposed. These animals then received an immediate hotplate test for analgesia. An additional group was exposed to Context A for 5 sec and upon removal was placed in a holding cage for 85 sec prior to the hotplate test, thereby providing a 90-sec interval between onset of exposure to Context A and the hotplate test. In order to equate handling and exposure to the holding cage, the former three groups were placed in the holding cage for 85 sec immediately prior to test day exposure to Context A. METHOD
Subjects Thirty male and 30 female rats of the same general description as those of Experiments 1 and 2 were used.
Apparatus The apparatus was identical to that of Experiment 2 except for the addition of six holding cages. These cages were identical to the animal's wire mesh home cages and were located in a separate room. In order to decrease pawlick tatencies and thereby possibly increase sensitivity to hyperalgesia, the surface temperature of the hotplate was increased from 50° to 52°C (see the Results and Discussion Section of Experiment 2).
Procedure To maintain comparability to the previous experiments, all animals were adapted to Contexts A and B on Days 1-4 in the same manner as in those studies; however, Context B was not used again in this study after Day 4. In addition, approximately 60 min following any session on Days 1-4, each animal was placed in one of the six wire holding cages for 10 min in order to adapt the animals to these cages. During one hour sessions on Days 5-10, all animals received six shock presentations a day in the same manner as in the previous studies. On Day 11 the analgesia test was conducted. Animals in Groups 0, 5, and 90 (ns=14, 14, 18, respectively) were first placed in their individual holding cages for 85 sec, Group 0 was then placed directly on the hotplate, while Groups 5 and 90 were first placed in Context A for 5 or 90 sec, respectively. Animals in Group 5/85 (n= 14) were placed in Context A for 5 sec followed by an 85-sec delay during which they were kept in their holding cages. Upon removal from the holding cages, these animals were transferred to the hotplate. Latency to lick a hind paw was recorded for all animals. R E S U L T S AND DISCUSSION
One animal in Group 5 died during the course of the study. The mean latency to lick a hind paw for each group is illustrated in Fig. 3. A one-factor A N O V A found a significant effect of treatment, F(3,55)=4.99, p<0.01. Planned comparisons found no difference between Groups 0 and 5 or between Groups 90 and 5/85, Fs(1,55)
R E C R U I T M E N T TIME
139
exposed to the same context for 90 sec were analgesic. Additionally, these results were extended by our demonstrating that a 5-see exposure was sufficient to elicit the analgesic response provided the test for analgesia was delayed for 85 sec following exposure to the excitatory context. This result argues against the possibility that the endogenous opioid system is activated at the average time of onset of the first shock because the system was apparently activated in Group 5/85 by 5 sec of exposure to the stimuli associated with shock, although the analgesic response was not immediately observable. Moreover, the present observation is contrary to the view that the magnitude of the analgesic response is proportional to the duration of exposure to the excitatory context on the day of testing. GENERAL DISCUSSION The current experiments identified time between onset of an aversive CS and a test for analgesia as a crucial factor in determining whether conditioned analgesia will be manifest (see Watkins and Mayer [15] for a review of other factors that influence the development of analgesia). Experiment 1, consistent with previous work on conditioned analgesia [1, 4, 8, 12-14], found that pain sensitivity decreased after exposure to contextual stimuli previously paired with the occurrence of shock, and that this conditioned analgesia was attenuated by the opiate antagonist naloxone. Experiment 2 determined that while long exposure (90 or 300 sec) to these contextual stimuli elicited the analgesic response, brief exposure (5 sec) did not. Experiment 3 found that the duration of exposure per se did not contribute to the manifestation of analgesia in behavior, but rather, the delay between the initial exposure to the aversive context on the test day and the test for analgesia was critical, i.e., a 5-sec exposure to the stimuli paired with shock was as effective in eliciting analgesia as a 90-sec exposure, provided than an 85-sec delay intervened between the 5-sec exposure and the test of pain sensitivity. Thus, the inferior conditioned responding seen with long intervals as opposed to short intervals between CS onset and US onset, at least in the aversive case, may be due in part to conditioned analgesia. Several reports have suggested that a stimulus previously paired with shock may elicit hyperalgesia, i.e., an increase in pain sensitivity [2,7]. This effect was not observed in the present studies, although when the baseline pawlick latency was reduced in Experiment 3 (by increasing the surface temperature of the hotplate), the nonsignificant difference between Groups 0 and 5 was in the direction of hyperalgesia in Group 5. Thus, the present data argue neither for nor against the possibility that short exposure to an aversive CS results in hyperalgesia. Fanselow [3] has suggested that the observed preference
displayed by rats for signaled over unsignaled shock results in part from the signal's ability to evoke an analgesic response which serves to decrease the effective magnitude of the US. The results reported here suggest that this would be likely only when there is a long interval between signal onset and shock onset. Otherwise, the signal for s h o c k would serve to activate the analgesic response only after the shock had already occurred. In fact, aside from the evidence that the signal for shock may in some instances result in hyperalgesia as measured by standard tests for pain sensitivity [2,7], psychophysical evidence has been reported that suggests that a brief, discrete signal that precedes a shock may serve to increase the perceived intensity of the shock [11]. While the signal for shock initiates the action of the analgesic system very soon after exposure, the modulation of pain perception apparently does not occur until somewhat later (Experiment 3). Such a lag in the analgesic system may be adaptive in that a signal for danger might better elicit a response that eliminates the danger [5,11] before eliciting an analgesic response that reduces the perceived intensity of stimulation that accompanies the danger. Analgesia has previously been shown to interfere with learning escape responses to shock [9,16], and CS-shock associations (Matzel and Miller, manuscript under review). Having escaped from potentially damaging stimuli, the analgesic response then might reduce the perceived intensity of any lingering pain such that normal functioning is less apt to be impaired through distraction. Such a possibility is equally applicable to conditioned and unconditioned analgesic responses. In each case, it would be of little adaptive value to attenuate pain as soon as it is inflicted because such a response might undermine the functional value of pain receptors. The fact that analgesia is often observed in response to stimuli that signal impending shock may be a function of the relatively long stimulus durations commonly employed in the conditioning laboratory relative to the typical duration of stimuli which precede danger in the natural environment. Since hyperalgesia seems in some instances [2, 7, 11] to be the immediate response to signals of short duration which precede shock, it may be that the more common functional outcome of classical conditioning in the aversive case is to increase the perceived intensity of impending danger.
ACKNOWLEDGEMENTS Support for this research was provided by NIMH Grant 3381 and NSF Grant BNS 86-00755. Thanks are extended to J. T. Cannon for his helpful technical suggestions regarding this research and S. Hallam and J. Navarro for their comments on an early draft of the manuscript.
REFERENCES
1. Chance, W. T., G. M. Krynock and J. A. Rosecrans. Antinociception following lesion-induced hyperemotionality and conditioned fear. Pain 4: 243-252, 1978. 2. Davis, H. D. and R. W. Henderson. Effects of conditioned fear on responsiveness to pain: Long-term retention and reversibility by naloxone. Behav Neurosci 99: 277-289, 1985. 3. Fanselow, M. S. Naloxone attenuates rat's preference for signaled shock. Physiol Psychol 7: 70-74, 1979.
4. Fanselow, M. S. Shock-induced analgesia on the formalin test: Effects of shock severity, naloxone, hypophysectomy, and associative variables. Behav Neurosci 98: 79-95, 1984. 5. Fanselow, M. S. and R. A. Sigmundi. Species-specific danger signals, endogenous opioiod analgesia, and defensive behavior. J Exp Psychol [Anita Behav] 12: 301-309, 1986.
140 6. Hayes, R. L., G. J. Bennet, P. G. Newlon and D. J. Mayer. Behavioral and physiological studies of non-narcotic analgesia in the rat elicited by certain environmental stimuli. Brain Res 155: 69-90, 1978. 7. Lysle, D. T. and H. Fowler. Pain sensitivity changes produced by first- and second-order fear excitors. Paper presented at the 57th annual meeting of the Eastern Psychological Association, 1986. 8. MacLennan, A. J., R. L. Jackson and S. F. Maier. Conditioned analgesia in the rat. Bull Psychon Soc 15: 387-390, 1980. 9. Maier, S. F., S. Davies, J. W. Grau, R. L. Jackson, D. H. Morrison, T. Moye, J. Madden, IV and J. D. Barchas. Opiate antagonists and long-term analgesic reaction induced by inescapable shock in rats. J Comp Physiol Psychol 94:1172-1183, 1980. 10. Matzel, L. D., A. M. Brown and R. R. Miller. Associative effects of US preexposure: Retarded conditioned responding mediated by an excitatory training context. J Exp Psychol [Anita Behav], in press.
MATZEL AND MILLER 11. Miller, R. R., C. Greco, M. Vigorito and N. A. Marlin. Signaled tailshock is perceived as similar to a stronger unsignaled tailshock: Implications for a functional analysis of classical conditioning. J Exp Psychol [Anim Behav] 9: 105-131, 1983. 12. Ross, R. T. Blocking and unblocking of conditioned analgesia. Learn Motiv 16: 173-189, 1985. 13. Ross, R. T. Pavlovian second-order conditioned analgesia..I Exp Psychol [Anita Behav] 12: 32-39, 1986. 14. Ross, R. T. and A. Randich. Unconditioned stress-induced analgesia following exposure to brief footshock. J Exp Psychol [Anim Behav] 10: 127-137, 1985. 15. Watkins, L. R. and D. J. Mayer. Organization of endogenous opiate and nonopiate pain control systems. Science 216:11851192, 1982. 16. Whitehouse, W. G., J. Walker, D. L. Margules and P. J. Bersh. Opiate antagonists overcome the learned helplessness effect but impair competent escape performance. Physiol Behav 3 0 : 7 3 1 734, 1983.
0031-9384/87$3.00 + .00
Recruitment Time of Conditioned Opioid Analgesia L O U I S D. M A T Z E L A N D R A L P H R. M I L L E R 1
Department o f Psychology, SUNY-Binghamton, Binghamton, N Y 13901 R e c e i v e d 23 S e p t e m b e r 1986 MATZEL, L. D. AND R. R. MILLER. Recruitment time of conditioned opioid analgesia. PHYSIOL BEHAV 39(1) 135--140, 1987.---Threeexperiments examined the development of conditioned analgesia in rats exposed to stimuli that had previously been paired with footshock. In Experiment 1, tailflick latencies increased if the tailflick test for analgesia was immediately preceded by 90 sec of exposure to a context in which unsignaled shock had previously been administered. This analgesia was blocked by the opiate antagonist naloxone administered prior to exposure to the context on the test day. Experiment 2 determined that 90 and 300 sec of exposure to the conditioning context immediately prior to testing evoked comparable analgesia as indexed by increased latencies to pawlick in response to thermal stimulation (hotplate). However, no analgesia was evident in animals exposed to the aversive context for 5 sec immediately prior to the hotplate test relative to animals not exposed to that context. In Experiment 3, a 5-sec exposure to the aversive context produced analgesia comparable to a 90-see exposure if an 85-sec delay intervened between the 5-sec exposure and the hotplate test. These results suggest that brief exposure to stimuli previously paired with shock can activate the endogenous opioid system, but the analgesic action of these opioids is delayed. Implications for the role of endogenous opioids in learning are discussed. Analgesia
Hyperalgesia
Endogenous opioids
Naloxone
A stimulus which has previously been paired with aversive stimulation such as footshock will in some instances come to elicit a temporary decrease in an animal's sensitivity to pain. Although this conditioned analgesic effect has been observed under a variety of conditions with both punctate cues and static environmental stimuli as the eliciting agent (e.g., [6, 8, 13, 14]), the specific circumstances under which a conditioned stimulus (CS) for shock will produce analgesia are still uncertain. Although the mechanism by which an initially neutral stimulus comes to elicit an analgesic response appears to conform to many of the principles of classical conditioning [4, 12-14], the procedures that reliably establish associations between a CS and a shock unconditioned stimulus (US) as indexed through traditional conditioned responses (CRs) are not universally successful in establishing analgesia as a CR. For instance, Davis and Henderson [2] and Lysle and Fowler [7] found that a punctate CS that previously had been paired with footshock elicits not analgesia, but rather, hyperalgesia, despite the fact that the footshock US upon which the CR is based itself produces a temporary decrease in responsiveness to painful stimulation [7]. One possible source of this discrepancy is the duration of the exposure to the CS prior to the test for analgesia. While in at least one instance hyperalgesia has been observed [2] in response to a CS of the same duration that was shown to produce analgesia elsewhere [14], instances where hyperalgesia as a CR has been reported have employed CSs of 30
~Requests for reprints should be addressed to Ralph R. Miller.
135
Rats
sec or less (e.g., [2,7]), whereas longer duration CSs more typically appear to elicit analgesia (e.g., [4,8]). The experiments reported here attempted first to establish that the opioid system was responsible for the decreased pain sensitivity in rats elicited by a 90-sec exposure to a context in which unsignaled shock previously had been administered. The analgesic response elicited by this same treatment has been shown elsewhere (Matzel and Miller, manuscript under review) to impair the animal's learning about a CS-shock association. In Experiment 2, animals were exposed to shock as in Experiment 1. Subsequently, the duration of the exposure to the context in which shock had been administered was varied (between 5 and 300 sec) and pain sensitivity was assessed immediately upon removal from that context. Having established in Experiment 2 that short duration (5 sec) exposure to the aversive CS was insufficient to evoke analgesia whereas longer duration exposure (90 or 300 sec) did elicit analgesia, Experiment 3 assessed analgesia after an 85-sec delay that followed a 5-sec exposure to the conditioning context. EXPERIMENT 1 METHOD
Subjects Eighteen male and 18 female, naive, Sprague-Dawleyderived rats bred in our colony served as subjects. Male and
136
MATZEL AND MILLER
o_ _J u_ .J
32
IO
~ 9
z9
£ uJ m
bz
b z
7
w
23
u
_1
$
26
6
~ ao +B/As
+ B/An
+A/As
+A/An
0
TREATMENT
5
90
TREATMENT
II 300
FIG. 1. Mean latency to tailflick after exposure to Context A in Experiment 1. Context A had a common history of unsignaled shock for Groups +A/As and +A/An. Groups +A/As and +B/As received a saline injection prior to exposure to Context A on the test day, whereas Groups +A/An and +B/An received an injection of naloxone. Brackets indicate standard errors.
FIG. 2. Mean latency to pawlick after exposure to Context A in Experiment 2. Context A had a common history of unsignaled shock for all groups. Groups 0, 5, 90, and 300 were exposed to Context A for 0, 5, 90, and 300 sec, respectively, immediately prior to the hotplate test. Brackets indicate standard errors.
female body weight ranges were 310--420 g and 225-305 g, respectively. All animals were water deprived, receiving 10-rain access to water a day approximately 1 hr following completion of any experimental session. This deprivation schedule was imposed gradually during the week preceding initiation of the study and was intended to maintain comparability to previous studies in our laboratory in which analgesia was reported with similar procedures (Matzei and Miller, manuscript under review). Purina Lab Chow was freely available in the home cages. Animals were housed individually in hanging wire-mesh cages in a colony maintained on a 16 hr/8 hr light/dark cycle. All manipulations were conducted near the middle of the light portion of this cycle. Subjects were assigned to treatment groups (ns=9) counterbalancing for sex, litter of origin, and body weight.
bulbs. Unlike Context A, these grids ran the length of the chamber. This apparatus was housed in a large room that was brightly illuminated by overhead fluorescent lights. The ambient sound level was 56 dB (C). These chambers were designated Context B. Previous work in our laboratory found no evidence of generalization between Contexts A and B [9]. Contexts A and B were located in separate rooms in the laboratory. The tailflick device consisted of a metal box that supported a 46× 16-cm (lxw) aluminum plate. Twenty cm from the end of the plate, a variable intensity nominal 150-W projector lamp was mounted under the plate with light from the bulb passing through a 3-mm aperture. A rat could be placed in one of six commercially manufactured Plexiglas restraining tubes (Fisher, 01-280-10) that rested on the surface of the aluminum plate such that the rat's tail extended over the aperture. When activated, the intensity of the lamp was permanently set such that the average latency of naive rats to tailflick was approximately 7 sec.
Apparatus Twelve chambers housed in controlled-environment shells were used. These chambers were clear Plexiglas boxes measuring 22.75 x 8.25 × 13.00 cm (1 x w x h) with electrifiable stainless steel grid floors (0.48-cm-diameter rods spaced 1.91 cm center-to-center). The grids were electrically linked in series by NE-2 neon bulbs. Constant-current footshock was provided by a high-voltage AC source in series with a 1-Mohm resistor and the neon bulbs. A water-filled lick tube was located on one short wall of each chamber, left-right centered and 3.8 cm above the grid floor. Although lick rates were not monitored in the present studies, these lick tubes were present to maintain comparability with other work in our laboratory [10]. Each environmental shell was illuminated by a nominal 7-W (at 120 VAC) white light bulb driven at 57 VAC. Background noise, produced by a ventilation fan, was 70 dB (C) (re. SPL). These chambers were designated Context A. Twelve additional chambers, highly dissimilar from the 12 Context A chambers, were used for administration of USs outside of Context A. They consisted of a 5 0 x l 6 . 5 x 5 0 - c m ( l x w x h ) stall with opaque green Plexiglas walls and floors composed of 0.64-cm diameter, electrifiable grids (spaced 1.59 cm center-to-center) interconnected by NE-2H neon
Procedure On Days 1-4, each subject was placed for 20 min in each of the two contexts. Order of exposure was A-B on Days 1 and 3, and B-A on Days 2 and 4. These initial days were intended to acclimate the animals to the chambers so that novelty of the context by itself would not evoke an analgesic response. Approximately 60 min following these sessions, all animals were placed in one of the Plexiglas restraining tubes for 10 min. The tube rested in place on the tailflick apparatus. The projector bulb was not activated during these adaptation sessions. On Days 5-10, all animals received six days of unsignaled shock presentations. Groups + A / A n and + A / A s received this treatment in Context A (therefore designated by +A), while the remaining animals, Group +B/An and +B/As received identical treatment in Context B (designated +B). During each of these 60-min sessions, all animals received six unsignaled 0.80-mA, 60-Hz, 500-msec footshocks. The footshocks were pseudorandomly scheduled throughout each session with an inter-US interval (ISI) range of 4-12
R E C R U I T M E N T TIME
137 administered, and this analgesia was opioid in nature, as indicated by the lack of conditioned analgesia in animals that received the opiate antagonist naloxone prior to exposure to the context in which they previously had been shocked (Group +A/An). Group + A / A n did not differ from either Group +B/As or Group +B/An for which Context A did not have a history of shock, and Groups +B/As and +B/An did not differ from each other, F s < 1.
~: 26 ¢,.).
1
D..
~ 23
i --
20
z w -J
EXPERIMENT 2 ~7
0
5
9O
5/85
TREATMENT
FIG. 3. Mean latency to pawlick after exposure to Context A in Experiment 3. Context A had a common history of unsignaled shock for all groups. Groups 0, 5, and 90 were exposed to Context A for 0, 5, and 90 sec, respectively, immediately prior to tlae hotplate test. Group 5/85 was exposed to Context A for 5 sec, but the hotplate test was delayed for 85 sec. Brackets indicate standard errors.
min, and a mean ISI of 8 min. On the third and fifth day of this shock exposure, the first shock in the session was scheduled within 5 sec of the start of the session to minimize the possibility o f animals'learning that Context A was safe for the first few minutes of each session. On Day 11, all animals received an intraperitoneal injection of either 1 cc/kg isotonic saline (Groups + A / A s and +B/As, designated by s) or a 10 mg/kg dose of naloxone hydrochloride (Groups + A/An and + B/An, designated by n) dissolved 10 mg/cc saline. Twenty min post-injection, each animal was placed in Context A for 90 sec. Following removal from Context A, animals were individually placed in their restraining tube and were placed on the tailflick apparatus, after which the projector lamp was activated and the latency to withdraw the tail from the beam of light was recorded. Five latencies were recorded for each rat. The first measure was taken approximately midway between the base and tip of the tail; subsequent measures were taken at l-cm increments toward the tip. A trial was terminated i f a tailflick did not occur within 15 sec. In practice, this latency was reached on less than 10% of the trials and was most frequently reached in Group +A/As. RESULTS
AND
DISCUSSION
A single latency was assigned to each subject by averaging its five tailflick latencies. Two subjects were discarded due to illness and two were discarded because they failed to acclimate to the restraining tube. Figure 1 depicts the mean tailflick latencies for each group. A two-factor A N O V A indicated a significant effect of the type of injection administered prior to exposure to Context A on the test day, F(1,28)=9.30, p<0.01, a significant effect of the context in which the animals were shocked, F(1,28)=7.95,p<0.01, and a significant interaction of these two factors, F(1,28)=4.65, p<0.05. Planned comparisons determined that Group + A / A s , which received a saline injection prior to exposure to the context in which it was previously shocked, had longer tailflick latencies than any of the remaining groups, Fs(1,28)~>12.57, ps<0.01. This suggests that analgesia was a conditioned response to the context in which shock was
Consistent with previous work on conditioned analgesia [1, 4, 8], decreased pain sensitivity was observed in Experiment 1 in animals exposed to a context in which unsignaled shock had previously been administered. Moreover, this analgesia was opioid in nature, as it was attenuated by the opiate antagonist naloxone. Experiment 2 attempted to determine if varying durations of exposure to the context paired with shock would influence the magnitude of the analgesia observed. In this study, all animals received unsignaled shock in Context A in a manner identical to Groups +A/An and + A / A s in the previous study. The test for analgesia consisted of placing the animals in Context A for 0, 5, 90, or 300 sec followed immediately by a hotplate test for analgesia. Since the analgesia observed in Experiment 1 was determined to be associative, Experiment 2 included no animals for which Context A was associatively neutral; however, the group that was not exposed to Context A prior to the hotplate test (0-sec exposure) was placed in Context B (which was associatively neutral) for 300 sec prior to testing in order to provide handling comparable to that of the remaining groups. Since this study was intended to examine the temporal onset of analgesia, the hotplate test rather than the tailflick test was employed because the hotplate allowed more immediate testing after removal from Context A, i.e., no time was lost placing the animals in the restraining tubes or using multiple tests. Prior work in our laboratory has found greater sensitivity to analgesia with a single test using the hotplate than with a single test using the tailflick apparatus. METHOD
Subjects Twenty-four male and 24 female rats of the same general description as those used in Experiment 1 served as subjects.
Apparatus The apparatus was identical to that of Experiment 1 except that a hotplate replaced the tailflick apparatus for the analgesia test. The hotplate consisted of a 22 x 22-cm copper plate on top of a constant temperature water bath. The surface was maintained at 50°C. The surface of the plate was enclosed in clear Plexiglas walls with a clear Plexiglas ceiling 9.5 cm above the copper plate that prevented rearing by the animals.
Procedure Initial acclimation to Contexts A and B was conducted on Days 1-4 as in the previous study. Subsequently, all subjects received six days of unsignaled shock exposure in Context A. Shock intensity and scheduling was identical to that of Experiment 1. On Day 11 (the day following the last shock session), the hotplate test for analgesia was conducted.
138
MATZEL AND MILLER
Group 5, Group 90, and Group 300 (ns= 12) were placed in Context A for 5, 90, and 300 sec, respectively. Group 0 (n= 12) was placed in Context B (where no shocks had been administered) for 300 sec. Upon removal from either context, each animal was immediately placed on the hotplate and the latency to lick a hind paw was recorded. This study was conducted in two balanced replications. R E S U L T S A N D DISCUSSION
Mean pawlick latencies are depicted in Fig. 2. An A N O V A with treatment and replications as factors was conducted. Neither replication nor the treatment x replication interaction proved significant, ps>0.10. The effect of treatment was reliable, F(3,40)=3.66, p<0.02. Planned comparisons indicated no difference between Groups 0 and 5, F=0.14, or between Groups 90 and 300, F=0.70. Group 0 differed both from Group 90, F(1,40)=8.17, p<0.01, and from Group 300, F(1,40)=4.08, p<0.05. Similarly, Group 5 differed from Group 90, F(1,40)=6.15, p<0.02. The tendency toward a difference between Groups 5 and 300 was not statistically reliable, F(1,40)=2.70, 0.10
to Context A for 5 sec, one group was exposed for 90 sec, and a third group was not exposed. These animals then received an immediate hotplate test for analgesia. An additional group was exposed to Context A for 5 sec and upon removal was placed in a holding cage for 85 sec prior to the hotplate test, thereby providing a 90-sec interval between onset of exposure to Context A and the hotplate test. In order to equate handling and exposure to the holding cage, the former three groups were placed in the holding cage for 85 sec immediately prior to test day exposure to Context A. METHOD
Subjects Thirty male and 30 female rats of the same general description as those of Experiments 1 and 2 were used.
Apparatus The apparatus was identical to that of Experiment 2 except for the addition of six holding cages. These cages were identical to the animal's wire mesh home cages and were located in a separate room. In order to decrease pawlick tatencies and thereby possibly increase sensitivity to hyperalgesia, the surface temperature of the hotplate was increased from 50° to 52°C (see the Results and Discussion Section of Experiment 2).
Procedure To maintain comparability to the previous experiments, all animals were adapted to Contexts A and B on Days 1-4 in the same manner as in those studies; however, Context B was not used again in this study after Day 4. In addition, approximately 60 min following any session on Days 1-4, each animal was placed in one of the six wire holding cages for 10 min in order to adapt the animals to these cages. During one hour sessions on Days 5-10, all animals received six shock presentations a day in the same manner as in the previous studies. On Day 11 the analgesia test was conducted. Animals in Groups 0, 5, and 90 (ns=14, 14, 18, respectively) were first placed in their individual holding cages for 85 sec, Group 0 was then placed directly on the hotplate, while Groups 5 and 90 were first placed in Context A for 5 or 90 sec, respectively. Animals in Group 5/85 (n= 14) were placed in Context A for 5 sec followed by an 85-sec delay during which they were kept in their holding cages. Upon removal from the holding cages, these animals were transferred to the hotplate. Latency to lick a hind paw was recorded for all animals. R E S U L T S AND DISCUSSION
One animal in Group 5 died during the course of the study. The mean latency to lick a hind paw for each group is illustrated in Fig. 3. A one-factor A N O V A found a significant effect of treatment, F(3,55)=4.99, p<0.01. Planned comparisons found no difference between Groups 0 and 5 or between Groups 90 and 5/85, Fs(1,55)
R E C R U I T M E N T TIME
139
exposed to the same context for 90 sec were analgesic. Additionally, these results were extended by our demonstrating that a 5-see exposure was sufficient to elicit the analgesic response provided the test for analgesia was delayed for 85 sec following exposure to the excitatory context. This result argues against the possibility that the endogenous opioid system is activated at the average time of onset of the first shock because the system was apparently activated in Group 5/85 by 5 sec of exposure to the stimuli associated with shock, although the analgesic response was not immediately observable. Moreover, the present observation is contrary to the view that the magnitude of the analgesic response is proportional to the duration of exposure to the excitatory context on the day of testing. GENERAL DISCUSSION The current experiments identified time between onset of an aversive CS and a test for analgesia as a crucial factor in determining whether conditioned analgesia will be manifest (see Watkins and Mayer [15] for a review of other factors that influence the development of analgesia). Experiment 1, consistent with previous work on conditioned analgesia [1, 4, 8, 12-14], found that pain sensitivity decreased after exposure to contextual stimuli previously paired with the occurrence of shock, and that this conditioned analgesia was attenuated by the opiate antagonist naloxone. Experiment 2 determined that while long exposure (90 or 300 sec) to these contextual stimuli elicited the analgesic response, brief exposure (5 sec) did not. Experiment 3 found that the duration of exposure per se did not contribute to the manifestation of analgesia in behavior, but rather, the delay between the initial exposure to the aversive context on the test day and the test for analgesia was critical, i.e., a 5-sec exposure to the stimuli paired with shock was as effective in eliciting analgesia as a 90-sec exposure, provided than an 85-sec delay intervened between the 5-sec exposure and the test of pain sensitivity. Thus, the inferior conditioned responding seen with long intervals as opposed to short intervals between CS onset and US onset, at least in the aversive case, may be due in part to conditioned analgesia. Several reports have suggested that a stimulus previously paired with shock may elicit hyperalgesia, i.e., an increase in pain sensitivity [2,7]. This effect was not observed in the present studies, although when the baseline pawlick latency was reduced in Experiment 3 (by increasing the surface temperature of the hotplate), the nonsignificant difference between Groups 0 and 5 was in the direction of hyperalgesia in Group 5. Thus, the present data argue neither for nor against the possibility that short exposure to an aversive CS results in hyperalgesia. Fanselow [3] has suggested that the observed preference
displayed by rats for signaled over unsignaled shock results in part from the signal's ability to evoke an analgesic response which serves to decrease the effective magnitude of the US. The results reported here suggest that this would be likely only when there is a long interval between signal onset and shock onset. Otherwise, the signal for s h o c k would serve to activate the analgesic response only after the shock had already occurred. In fact, aside from the evidence that the signal for shock may in some instances result in hyperalgesia as measured by standard tests for pain sensitivity [2,7], psychophysical evidence has been reported that suggests that a brief, discrete signal that precedes a shock may serve to increase the perceived intensity of the shock [11]. While the signal for shock initiates the action of the analgesic system very soon after exposure, the modulation of pain perception apparently does not occur until somewhat later (Experiment 3). Such a lag in the analgesic system may be adaptive in that a signal for danger might better elicit a response that eliminates the danger [5,11] before eliciting an analgesic response that reduces the perceived intensity of stimulation that accompanies the danger. Analgesia has previously been shown to interfere with learning escape responses to shock [9,16], and CS-shock associations (Matzel and Miller, manuscript under review). Having escaped from potentially damaging stimuli, the analgesic response then might reduce the perceived intensity of any lingering pain such that normal functioning is less apt to be impaired through distraction. Such a possibility is equally applicable to conditioned and unconditioned analgesic responses. In each case, it would be of little adaptive value to attenuate pain as soon as it is inflicted because such a response might undermine the functional value of pain receptors. The fact that analgesia is often observed in response to stimuli that signal impending shock may be a function of the relatively long stimulus durations commonly employed in the conditioning laboratory relative to the typical duration of stimuli which precede danger in the natural environment. Since hyperalgesia seems in some instances [2, 7, 11] to be the immediate response to signals of short duration which precede shock, it may be that the more common functional outcome of classical conditioning in the aversive case is to increase the perceived intensity of impending danger.
ACKNOWLEDGEMENTS Support for this research was provided by NIMH Grant 3381 and NSF Grant BNS 86-00755. Thanks are extended to J. T. Cannon for his helpful technical suggestions regarding this research and S. Hallam and J. Navarro for their comments on an early draft of the manuscript.
REFERENCES
1. Chance, W. T., G. M. Krynock and J. A. Rosecrans. Antinociception following lesion-induced hyperemotionality and conditioned fear. Pain 4: 243-252, 1978. 2. Davis, H. D. and R. W. Henderson. Effects of conditioned fear on responsiveness to pain: Long-term retention and reversibility by naloxone. Behav Neurosci 99: 277-289, 1985. 3. Fanselow, M. S. Naloxone attenuates rat's preference for signaled shock. Physiol Psychol 7: 70-74, 1979.
4. Fanselow, M. S. Shock-induced analgesia on the formalin test: Effects of shock severity, naloxone, hypophysectomy, and associative variables. Behav Neurosci 98: 79-95, 1984. 5. Fanselow, M. S. and R. A. Sigmundi. Species-specific danger signals, endogenous opioiod analgesia, and defensive behavior. J Exp Psychol [Anita Behav] 12: 301-309, 1986.
140 6. Hayes, R. L., G. J. Bennet, P. G. Newlon and D. J. Mayer. Behavioral and physiological studies of non-narcotic analgesia in the rat elicited by certain environmental stimuli. Brain Res 155: 69-90, 1978. 7. Lysle, D. T. and H. Fowler. Pain sensitivity changes produced by first- and second-order fear excitors. Paper presented at the 57th annual meeting of the Eastern Psychological Association, 1986. 8. MacLennan, A. J., R. L. Jackson and S. F. Maier. Conditioned analgesia in the rat. Bull Psychon Soc 15: 387-390, 1980. 9. Maier, S. F., S. Davies, J. W. Grau, R. L. Jackson, D. H. Morrison, T. Moye, J. Madden, IV and J. D. Barchas. Opiate antagonists and long-term analgesic reaction induced by inescapable shock in rats. J Comp Physiol Psychol 94:1172-1183, 1980. 10. Matzel, L. D., A. M. Brown and R. R. Miller. Associative effects of US preexposure: Retarded conditioned responding mediated by an excitatory training context. J Exp Psychol [Anita Behav], in press.
MATZEL AND MILLER 11. Miller, R. R., C. Greco, M. Vigorito and N. A. Marlin. Signaled tailshock is perceived as similar to a stronger unsignaled tailshock: Implications for a functional analysis of classical conditioning. J Exp Psychol [Anim Behav] 9: 105-131, 1983. 12. Ross, R. T. Blocking and unblocking of conditioned analgesia. Learn Motiv 16: 173-189, 1985. 13. Ross, R. T. Pavlovian second-order conditioned analgesia..I Exp Psychol [Anita Behav] 12: 32-39, 1986. 14. Ross, R. T. and A. Randich. Unconditioned stress-induced analgesia following exposure to brief footshock. J Exp Psychol [Anim Behav] 10: 127-137, 1985. 15. Watkins, L. R. and D. J. Mayer. Organization of endogenous opiate and nonopiate pain control systems. Science 216:11851192, 1982. 16. Whitehouse, W. G., J. Walker, D. L. Margules and P. J. Bersh. Opiate antagonists overcome the learned helplessness effect but impair competent escape performance. Physiol Behav 3 0 : 7 3 1 734, 1983.