Effects of acute and prolonged opiate abstinence on extinction behaviour in rats

Effects of acute and prolonged opiate abstinence on extinction behaviour in rats

Kim G C Hellemans

Abstract We examined the role of withdrawal in relapse to drug-seeking and drug-taking by testing the effects of opiate abstinence on extinction behaviour in rats trained to self-administer heroin. Male Long-Evans rats responded for IV heroin under a heterogeneous chain (VI 120 s; FR 1) schedule in which “seeking” responses preceded a “taking” response which produced a drug infusion. Responding was then measured in extinction during acute (6, 12, and 24 hr) and prolonged (3, 6, 12, and 25 day) abstinence. Sucrose consumption and somatic withdrawal were assessed at each testing period. During acute abstinence, responses on the “drug-seeking” manipulandum increased at 24 hr, whereas responses on the “drug-taking” manipulandum increased at 6 hr. Both responses were elevated during the 12-day abstinence test. Sucrose consumption was reduced and somatic withdrawal scores were increased in opiate-experienced rats at each test period. Results suggest that heroin abstinence has different effects on drug-seeking and drug-taking and that these effects do not temporally coincide with somatic measures of opioid withdrawal.

Theories of drug addiction traditionally include drug abstinence as a critical feature of compulsive drug use. Different classes of drugs produce distinct withdrawal symptoms, but in all cases the aversive effects of drug abstinence are alleviated by the drug itself (Koob, Stinus, Le Moal, & Bloom, 1989). This suggests that drug abstinence may play a key role in maintaining drug self-administration, and may contribute to relapse after cessation of drug intake. Indeed, based on the premise that withdrawal acts as a negative reinforcer, many theories posit a major role for drug abstinence in the maintenance of, and relapse to, drug use (Gawin, 1989; Koob et al., 1989; Solomon & Corbit, 1974; Tiffany, 1990; Wilder, 1973). Nevertheless, there is surprisingly little evidence that drug abstinence plays a significant role in maintaining drug self-administration (Robinson & Berridge, 1993; Stewart & Vezina, 1988; Wise & Bozarth, 1987). For example, in both nonhuman and human subjects, there is no direct correlation between the severity of withdrawal symptoms and relapse tendencies following prolonged cocaine (Hughes, Higgins, & Bickel, 1994) or opiate (Shaham, Rajabi, & Stewart, 1996) use. Moreover, relapse to drug-taking in humans can occur even after overt signs of withdrawal have subsided (jaffe, 1990; Tiffany, 1990). To account for this latter phenomenon, it was suggested that conditioned cues associated with drug abstinence induce withdrawal and subsequent relapse following prolonged abstinence from morphine (Goldberg, Woods, & Schuster, 1969; Wilder & Pescor, 1970) or alcohol (Ludwig, Wilder, & Stark, 1974). Empirical evidence, however, does not appear to support this idea because the correlation between cue– induced craving and withdrawal symptoms, at least in cocaine addicts, is weak (Childress, McLellan, Ehrman, & O’Brien, 1988), and self-reported craving for cocaine (Jaffe, Cascella, Kumor, & Sherer, 1989) and opiates (Ehrman, Ternes, O’Brien, & McLellan, 1992; Koob et al., 1989; Meyer & Mirin, 1979) is often highest immediately after drug administration, when withdrawal symptoms are alleviated.

Direct tests of the role of withdrawal in opiate selfadministration have also produced inconclusive findings. Rats trained under a “free-choice” drinking procedure for morphine increase their drug consumption during early withdrawal (Wikler & Pescor, 1970), but consumption levels remain elevated up to 72 days after the last morphine administration, even after overt somatic signs of withdrawal have ceased (Wilder & Pescor, 1967, 1970). Moreover, several studies reported that precipitating withdrawal in opiate-dependent animals by an injection of the opiate antagonist, naloxone, increases morphine self-administration (Goldberg, Woods, & Schuster, 1971; Goldberg et al., 1969). The interpretation of these findings is problematic because behavioural responses were similar to those of animals self-administering low doses of opioid agonists. Thus, increased responding following an opioid antagonist injection may be due to a diminished effect of morphine, rather than an attempt to alleviate withdrawal symptoms. In fact, recent studies demonstrate that neither a high dose (5mg/kg) nor a low dose of naloxone (0.1 mg/kg) are effective in reinstating heroin-seeking following extinction (Shaham & Stewart, 1995; Shaham et al., 1996). In contrast, spontaneous withdrawal was found to reinstate heroin-seeking, but it cannot be concluded from this study whether this effect was due to the motivational effects of opioid withdrawal or from state-dependent mechanisms (see Shaham et al., 1996 for a discussion).

Thus, despite many years of research, the role of opiate abstinence in relapse to drug-taking behaviour in animal models has not been established. One difficulty in studying this relationship is that no single animal paradigm can model all of the behaviours associated with addiction. In particular, few tests dissociate drug-seeking and drug-taking behaviours despite the evidence that these may be mediated by distinct neural and psychological mechanisms (Robinson & Berridge, 1993; Tiffany, 1990). In humans, drug-seeking involves efforts to obtain the drug (e.g., contacting the dealer, stealing money for drugs, going to the store to purchase cigarettes), whereas drug-taking refers specifically to ingesting the drug (e.g., snorting cocaine, injecting heroin, drinking alcohol).

Drug-seeking and drug-taking can be measured in animal studies using a heterogeneous chain schedule in which performance of a seeking response provides access to the opportunity to perform a taking response (Olmstead, Parkinson, Miles, Everitt, & Dickinson, 2000). In the case of cocaine, the two responses are distinct because seeking and taking responses for a drug infusion are differentially affected by changes in dose and the introduction of a timeout period. Finally, when rats are responding for either cocaine or sucrose, extinction of the taking response reduces seeking responses in a later test (Johnston, Beninger, & Olmstead, 2001; Olmstead, Lafond, Everitt, & Dickinson, 2001). Seeking responses under a heterogeneous chain schedule, therefore, are maintained by the contingency between seeking and taking links in the chain.

Using a variation on the chain schedule described above, the current study examined whether the duration of heroin abstinence influences previously reinforced drug-seeking and drug-taking responses in extinction. Rats were trained to respond for intravenous heroin under a heterogeneous chain schedule in which responding in the drug-seeking link gave access to the opportunity to perform a different, drugtaking response. Responses on the drug-seeking and drug-taking manipulanda were measured in extinction, 6, 12, and 24 hr after heroin self-administration. In addition, rats were repeatedly tested under extinction conditions 3, 6, 12, and 25 days after exposure to heroin. In order to relate the rates of operant responding to measures of opioid withdrawal, somatic symptoms (Bhargava, 1994; Blasig, Herz, Reinhold, & Zieglgansberger, 1973), and sucrose intake (Lieblich, Yirmiya, & Liebeskin, 1991; Zellner, Dacanay, & Riley, 1984) were assessed at each test period.



The subjects were 16 (8 self-administration and 8 control) male hooded Long-Evans rats (Charles River, QUE) weighing between 300-350 g at the start of the experiment. The rats were housed singly and maintained on a reversed light-dark cycle throughout the duration of the experiment (lights on from 7:00 PM to 7:00 AM). Following recovery from surgery, rats were food restricted to 10 pellets (Purina rat chow) per day, maintaining body weight at 85-90% of the free-feeding weight. Rats had access to water at all times, except during behavioural testing. Animal care and all experiments were conducted in accordance with the guidelines provided by the Queen’s University Animal Care Committee and the Canadian Council on Animal Care.


Eight rats were surgically implanted with intravenous silastic catheters under anaesthesia. The catheters were made using guide cannulae (C313DC/CAC; Plastics One, Roanoke, VA, U.S.A.), silastic tubing (0.02xOD:0.037 and 0.012xOD:0.025; VWR, Missassauga, ON, Canada), and shrink tubing (3/64″ length; Electrosonic, Willowdale, ON, Canada). All catheters were sterilized in an autoclave prior to implantation. Rats were anaesthetized using a mixture of xylazine (20 mg/ml IP; Bayer Inc., Etobicoke, ON, Canada) and ketamine (100mg/ml IP; Rogar/STB Inc., London, ON, Canada). The catheters were implanted with the proximal end reaching the heart through the right jugular vein, and secured with silk sutures. After implantation, the distal end was passed subcutaneously to the top of the skull where it was then connected to a bent 22-gauge cannula (C312G-Sup; Plastics One, Roanoke, VA, U.S.A.) and mounted to the skull with jewelers’ screws and dental cement (ASH Temple Ltd.; Ottawa, ON, Canada). An analgesic, buprenorphine (0.3 mg/ml; Reckitt & Colman Pharmaceuticals Inc., Richmond, vA, U.S.A.; 0.2 ml/kg s.c.) and an antibiotic, trimethoprim (40mg/ml; Schering-Plough, Pointe-Claire, QUE, Canada; 0.3 ml s.c.), were administered 30 min post-operatively. The catheters were flushed daily with 0.2 ml physiological saline and once a week with 0.1 ml Heparin solution (Fisher Scientific, Fair Lawn, NJ, U.S.A.; 30 units/ml 0.9% sterile saline). The rats commenced training at a minimum of 7 days after surgery.


Training and testing took place in four operant chambers (26.5 x 22 x 20 cm), each housed in a sound– attenuating box (constructed in-house). Each chamber was fitted with two retractable levers 4 cm wide, equidistant from the outer walls, 13 cm apart and 9 cm from the Plexiglas floor. Each box was equipped with a Razel infusion pump (Razel Scientific Instruments Inc.; Stamford, CT, U.S.A.), which delivered intravenous infusions of heroin through a single channel liquid swivel with connector attachments (Lomir; Ile Perrot, QUE, Canada). A response wheel was fitted on the opposite side of the box, 9 cm from the Plexiglas flooring, and measured 8 cm in height. The operant chambers were fixed with a ceiling house light that served as the discriminative stimulus (SD). External noise was masked by ventilating fans, which were mounted on the side of the sound-attenuating boxes.


Heroin hydrochloride (Health Canada) was dissolved in sterile 0.9% saline to yield a dose of 0.12 mg/kg. Each 0.2-ml infusion was delivered over 7.28 s.

Behavioural Procedures


Overview. Rats were trained to self-administer heroin under a heterogeneous chain schedule in which responses on a drug-seeking manipulandum (the wheel) in the first link gave access to the second link of the chain. The second link was signaled by the presentation of the SD and the insertion of both levers. One response on the drug-taking lever delivered an infusion of heroin and turned off the SD. Responses on the inactive lever had no programmed consequences. The assignments of the left and right levers as drug-taking or inactive were counterbalanced across rats. Each training session lasted 3 hr, and occurred either from 811 AM or from 1-4 PM. These session times were selected in order to control for the phase of the light/dark cycle during testing. Session time was randomly assigned and counterbalanced across rats.

Phase 1. During this phase, the wheel was removed from the chambers. Initially, both levers were inserted at the start of each session, and the SD (house light) was illuminated during the entire session. These conditions remained until the rats demonstrated a stable response pattern (10-15 responses on the drug-taking lever and 0-2 responses on the inactive lever) over 3 days. In subsequent sessions, the levers were retracted and the SD was turned off after each drug infusion. In addition, a latency to turn on the SD after each drug infusion was introduced. This latency was increased from 1 to 2 to 3 min once responding had stabilized at each time point. The final latency was set at 3 min because pilot studies revealed that rats wait about 3 min after an infusion before they begin to respond on the drug-seeking manipulandum.

Phase 2. During this phase, the wheel was introduced into the chamber. Wheel responses (a measure of drug-seeking) were maintained under a variable interval (VO schedule. The first response (defined as one 189 pull on the wheel) meeting the VI contingency terminated the first link of the chain, turned on the SD, and led to the insertion of the levers into the chamber. Subsequently, a response on the drug-taking lever produced a heroin infusion, retracted both levers, and turned off the house light. At the end of the heroin infusion, the next cycle of the chain schedule started. The VI contingency of the drug-seeking link was increased across sessions from 2, 5, 15, 30, 60, to 120 s. The last three sessions under the VI 120-s schedule provided the baseline measure for responses during extinction testing.


The effects of acute and prolonged heroin abstinence were assessed using three different measures: I) operant responding in extinction, II) sucrose consumption, and III) somatic withdrawal symptoms. For acute abstinence testing, each rat was assessed under each measure at 6, 12, and 24 hr after their last self-administration session. The order of tests was random and counterbalanced across subjects. For example, if one rat underwent a 6-hr test for sucrose intake, a 12-hr test for operant responding, and a 24-hr test for somatic withdrawal, the next two consecutive days would involve baseline training in the operant chamber prior to undergoing the next set of tests. For prolonged abstinence testing, each subject was tested in all three behavioural paradigms at 3, 6, 12, and 25 days after the last self-administration session. There were no baseline sessions between these tests. The order of tests was random and counterbalanced across subjects. Each heroin-experienced subject had a matched drug-naive control animal that did not undergo any surgery or selfadministration training. These rats were tested in each paradigm at the same time that their matched experimental rat underwent the acute abstinence tests. The 3-, 6-, 12-, and 25-day abstinence tests were conducted at the same point in the circadian cycle as the 24-hr acute abstinence tests. Therefore, the 24-hr sucrose intake and somatic withdrawal tests in controls were used as comparisons for these measures in heroin– experienced rats during prolonged abstinence testing. Control rats were tested in the operant paradigm 3, 6, 12, and 25 days after the last acute abstinence test.

Operant responding in extinction. The effects of acute and prolonged heroin abstinence on responses on the drug-seeking manipulandum, the drug-taking lever, and the inactive lever were determined under extinction conditions. During testing, the wheel and the two levers were in the chamber throughout the 3-hr sessions and the SD was presented for 60 s on a variable time (VT) 3-min schedule. This design provides information on how rats respond during abstinence when drug-taking behaviour is no longer contingent on drug– seeking behaviour. That is, we measured drug-seeking and drug-taking independently and thus determined how each response was influenced by the presentation of a stimulus previously associated with availability of the drug. In order to minimize the effect of extinction on operant responding during the acute abstinence tests (6, 12, and 24 hr), rats received two baseline sessions under the VI 120-sec/FR-1 chain schedule between each test session.

Sucrose intake. Rats decrease their consumption of palatable solutions for up to two weeks after cessation of morphine injections (Lieblich et al., 1991; Zellner et al., 1984). Thus, consumption of sucrose at different periods of abstinence was also assessed. Sucrose dissolved in tap water (10% w/v) was provided in 50-ml drinking bottles in the home cage, and the volume consumed over a 1-hr period was measured.

Somatic symptoms. Somatic symptoms of withdrawal were used to quantify the severity of heroin withdrawal. The rats were placed in Plexiglas boxes and their behaviour was observed and recorded with a video camera for 1 hr (scoring was done by an observer who was blind to the treatment condition). Four categories of withdrawal behaviours were scored: wet dog shakes; paw tremors; mouth movements; and ear wipes (adapted from MacRae & Siegel, 1997).


Training. Drug-seeking responses were converted to response rates (responses/min) when the SD was on and off. As mentioned, the SD turned on when the VI 120-s drug-seeking schedule was completed and turned off when the drug infusion started. Drug-seeking responses when the stimulus was on were calculated as the number of responses made from the onset of the SD to the first drug-taking response (coinciding with the SD turning off). Because there is a period of selfimposed abstinence that follows each infusion and varies with drug dose (Olmstead et al., 2000; Wilson et al., 1971; Wise et al., 1995), the drug-seeking response rate when the SD was off was calculated as responding during the period from the first drug-seeking response in each cycle to the completion of the schedule (coinciding with the SD turning on). Drug-seeking response rates were then analyzed using a two-way, repeated measures ANOVA, with stimulus condition (on vs. off) and baseline session (1-7 days) as factors. Drug-taking responses and inactive lever responses were evaluated using a two-way, repeated measures ANOVA, with lever (drug-taking vs. inactive) and baseline session (1-7 days) as the two factors.

Testing. Data from the acute (6, 12, and 24 hr) and prolonged (3, 6, 12, and 25 day) abstinence tests were analyzed separately because no baseline training sessions were given between the prolonged abstinence tests.

Operant extinction tests. The rate of responding on the drug-seeking manipulandum during extinction tests was analyzed using a two-way, repeated measures AVOVA, with acute (6, 12, and 24 hr) or prolonged (3, 6, 12, 25 day) abstinence period and SD (on vs. off) as the factors. The rate of responding on the previously drugtaking lever was analyzed using three-way, repeated measures ANOVA, with acute (6, 12, or 24 hr) or prolonged (3, 5, 12, or 25 days) abstinence period, stimulus condition (on vs. off), and lever (drug-taking vs. inactive levers) as the factors. Significant ANOVAs (p

Sucrose intake. Data for the sucrose intake tests were analyzed using a two-way, repeated measures ANOVA, with acute (6, 12, and 24 hr) or prolonged (3, 6, 12, and 25 days) abstinence period as the factor. Data from the control rats at the 24-hr test were compared to the data from heroin-experienced rats at each of the prolonged abstinence tests. Group (heroin-experienced vs. drug-naive) served as the between-groups factor. A significant ANOVA was followed by individual means comparison using Tukey’s HSD post-hoc tests. The criterion for significance was set atp

Somatic withdrawal symptoms. Scores for wet dog shakes, paw tremors, mouth movements, and ear wipes were summed for each rat to yield a total withdrawal score. Data from the control rat at the 24-hr test were compared to the data from heroin-experienced rats at each of the prolonged abstinence tests. These data were analyzed using two-way, repeated measures ANOVA, with acute (6, 12, and 24 hr) or prolonged (3, 6, 12, and 25 days) abstinence period as the repeated measures factor. The group (heroin-experienced vs. drug-naive) served as the between-groups factor.



Rats took 6-7 days to get through the first phase of training, averaging 14.69 +/- 1.98 infusions per 3-hr session (data not shown). Rats responded significantly more on the drug-taking lever (14.5 +/- 1.37) than on the inactive lever (3.5 +/- 0.56), suggesting that they were able to discriminate between levers prior to starting the chain schedule, t(14) = 7.39, p

+/- 1.25 infusions per 3-hr session (data not shown). Thus, rats had approximately 18 consecutive selfadministration sessions (one per day) prior to the first acute abstinence test. They had two additional selfadministration sessions between the first and second, and the second and third acute abstinence tests (approximately 22 sessions in total). Prolonged abstinence testing occurred at days 3, 6, 12, and 25 after the last self-administration session. On the intervening days, rats were left in their home cages. Thus, the whole experiment took approximately 47 days.

The rats maintained stable response rates on the drug-seeking, F(3,21) = 0.16, p > 0.05, and the drug-taking, F(3,21) = 1.26, p > 0.05, manipulanda during the baseline sessions conducted prior to and between the acute withdrawal extinction tests (see Figure 1). In addition, drug-seeking response rates were higher when the SD was off than when it was on, F(1,7) = 10.02, p


Acute abstinence. Figure 2A shows response rates on the drug-seeking manipulandum during the acute abstinence tests. Statistical analyses of responses on the drug-seeking manipulandum during the acute abstinence period (6, 12, and 24 hr) yielded a significant main effect of Period, F(2,14) = 4.32, p 0.05) and Period x Stimulus, F(2,14) = 2.15, p > 0.05, indicate that there was no difference in responding in the presence and absence of the SD.

Responding on the previously drug-taking lever and on the inactive lever in acute heroin abstinence are shown in Figures 2C and 2E. Statistical analyses performed on responses during the acute abstinence periods (6, 12, and 24 hr) yielded significant effects of Stimulus, F(1,7) = 28.25, p 0.05, Lever, F(1,7) = 2.79, p > 0.05, Period x Stimulus, F(2,14) = 2.14, p > 0.05, and Stimulus x Lever, F(1,7) = 1.47, p > 0.05, interactions were not significant.

Prolonged abstinence. Statistical analyses performed on the rate of responding on the drug-seeking manipulandum during prolonged abstinence (see Figure 2B) yielded a significant main effect of Period, x(3,21) = 3.29, p 0.05). The effects of Stimulus, 1-07) = 1.12, p > 0.05, and Period x Stimulus, F(3,21) = 0.41, p > 0.05, were not significant. Again, these results imply that there was no difference in responding in the presence and absence of the SD.

Responding on the previously drug-taking lever and on the inactive lever (see Figures 2D and 2F) followed a similar pattern to that observed for responding on the drug-seeking manipulandum. The 3-way ANOVA yielded significant effects of Period, F(3,18) = 4.07, p 0.05. These analyses indicate that responding did not differ significantly between the drug-taking and inactive levers; however, rats responded significantly more on both levers in the presence of the SD. Post-hoc tests revealed that difference was also due to the increase in responding at the 12-day test as compared to the 3-, 6and 25-day tests (p 0.05, Period x Lever, F(3,18) = 2.5, p > 0.05, Lever x Stimulus, F(1,6) = 1.54, p > 0.05, and Period x Lever x Stimulus, F(3,18) = 0.26, p > 0.05, interactions were not significant.


Acute abstinence. Sucrose intake at the different periods of heroin abstinence is shown in Table 1. Heroin– experienced rats consumed significantly less sucrose than drug-naive animals during each of the acute abstinence tests, as evidenced by the significant group effect, F(1,14) = 41.61,p

Prolonged abstinence. When compared to control rats tested at 24 hr after the last self-administration session, drug-experienced rats consumed significantly less sucrose during each prolonged abstinence test, F0,13) = 20.24, p


Data from one session for one rat was lost due to improper recording.

Acute abstinence. Table 1 shows the sum of all somatic withdrawal signs for heroin-experienced and drug– naive rats during acute heroin abstinence. Heroin– experienced rats showed significantly more withdrawal signs than drug-naive controls at every abstinence period, F1,12) = 22.07,p 0.05), and Period x Group, F(2,24) = 1.2, p > 0.05), were not significant.

Prolonged abstinence. Results for the sum of all somatic withdrawal signs in prolonged heroin abstinence are also shown in Table 1. Heroin-experienced rats showed significantly more withdrawal signs than drug– naive controls at each abstinence period, F0,13) = 21.60, p 0.05, and Period x Group, q3,39) = 1.56,p > 0.05, were not significant.


The present experiment suggests that previously reinforced drug-seeking and drug-taking behaviours are influenced by both acute and prolonged periods of heroin abstinence. In acute abstinence (6, 12, and 24 hr), response rates on both the drug-seeking and drugtaking manipulanda vary with period of abstinence. Whereas rates of responding on the lever previously associated with drug-taking and on the inactive lever were significantly elevated at the 6-hr test, response rates on the drug-seeking manipulandum were significantly higher at the 24-hr test. In prolonged abstinence (3, 6, 12, and 25 days), responding on both the drug– seeking and drug-taking manipulanda was initially low at 3 and 6 days, but was elevated at 12 days after the last self-administration session. Somatic measures of opiate withdrawal and sucrose intake tests suggest that heroin-experienced rats were experiencing the aversive effects of opiate abstinence up to 12 days after the last self-administration session.

Data from this and other studies (Johnston et al., 2001; Olmstead et al., 2000; Olmstead et al., 2001) demonstrate that measures of drug-seeking and drugtaking under a heterogeneous chain schedule are distinct. During baseline training, rats learned that access to the drug-taking lever was contingent upon making the drug-seeking response (a wheel turn) in the absence of the SD. In addition, rats learned to discriminate between the drug-taking and the inactive levers. Finally, overall rates of responding over the 3-hr sessions did not appear to be influenced by previous experience with the manipulanda in extinction, as both drug-seeking and drug-taking remained stable over the seven baseline sessions even though extinction tests (i.e., acute abstinence) occurred after baseline sessions 3, 5, and 7.

The suppression of responding on the drug-seeking manipulandum 6 hr after the last drug self-administration session suggests that early abstinence states may lead to an initial decrease in motivation to gain access to the drug. Conversely, if the opportunity to selfadminister the drug is readily available (i.e., signalled by the SD), the rat will engage in behaviours previously associated with drug-taking. Dissociation of the two behaviours is not unusual; indeed, some theorists argue that drug-seeking (or “urges”) may not be necessary for drug use (Tiffany, 1990). In contrast to the 6-hr test, rats were willing to work harder for access to the drug (as measured by responses on the drug-seeking manipulandum) at later stages of acute abstinence (e.g., 24 hr). These results suggest that periods of short-term abstinence may differentially affect behaviours distal and proximal to the delivery of the reinforcer.

Because rats were trained daily at specific times, circadian cues likely contributed to the discriminative and conditioned reinforcing properties of the operant sessions. These conditions may mimic human situations in which users ingest drugs at certain (habitual) times of the day. No doubt, the higher response rates at the 24-hr abstinence test could be due, in part, to the conditioned circadian cues. This test, therefore, provides insight into how addicts may behave under these conditions. Similarly, an inherent part of the 12-hr abstinence test is that it was conducted during the opposite circadian cycle to that of the training sessions. Although we can not rule out the possibility entirely, it is unlikely that the increase in responding at 24-hr postdrug is clue to the fact that this test was conducted during the active portion of the rats’ cycle because there was no difference in response rates during the 6hr abstinence tests that were conducted during the light and dark phases of the circadian cycle. Moreover, if changes in activity levels were responsible for the increased responding at 24 hr, responses on the inactive lever would likely increase during this test. Thus, if rats were trained during the light, rather than the dark phase, we believe that the cues would simply be “shifted” accordingly (much like individuals who work night shifts), and that we would still see the same pattern of results. Nonetheless, it is possible that circadian factors not associated with the training sessions may have contributed to the differences in responding during the 12- and 24-hr abstinence tests.

The significant peak in responding on both drug– seeking and drug-taking manipulanda at the 12-day prolonged abstinence test was unexpected, particularly because the 12-day test was the fourth successive extinction test these rats received. It is not likely that this change in responding is artefactual (i.e., due to previous drug exposure) because responding for nondrug reinforcers such as food or water generally decreases across extinction sessions, and successive extinction tests for drugs of abuse typically produce a decline in responding, both within and across sessions (Grimm & See, 2000). Although our results did not follow this characteristic pattern, they are in line with previous results showing that drug priming and stress induced by intermittent footshock reliably reinstate heroin- (Erb, Shaham, & Stewart, 1996), cocaine(Shaham & Stewart, 1995), and nicotine-seeking (Shaham, Adamson, Grocki, & Corrigall, 1997) behaviours after prolonged (4-6 weeks) extinction training and drug abstinence. It was recently found that animals trained to respond for IV heroin reinstate heroin– seeking after exposure to footshock stress up to 11 weeks after their last drug self-administration session (Shalev, Morales, Hope, Yap, & Shaham, 2001), and that both extinction behaviour and cue-induced reinstatement progressively increases over a 2-month withdrawal period (Grimm, Hope, Wise, & Shaham, 2001). Despite these consistencies, the time-dependent changes in extinction behaviour in the present experiment should be interpreted with caution because rats were repeatedly tested under extinction at 3, 6, 12, and 25 days of abstinence. Nonetheless, our data do suggest that the motivation to seek heroin peaks after prolonged rather than acute abstinence periods.

The observed changes in drug-seeking and drug-taking across different abstinence periods is not surprising because the neurochemical and behavioural state of the rat following short-term (hours to a few days) abstinence is very different from the state induced by longer (weeks to months) drug-free periods (Olmstead et al., 2000; Wise & Bozarth, 1987). For example, dopaminergic D2 receptor activation reinstates nonreinforced responding for both heroin and cocaine in acute abstinence states (Self, Belluzi, Kossuth, & Stein, 1996; Wise, Newton, Leeb, & Burnett, 1990), but only reinstates cocaine-seeking following long-term extinction of IV drug self-administration (De Vries, Schoffelmeer, Binnekade, & Vanderschuren, 1999).

Interestingly, despite a clear bias for responding on the drug-seeking manipulandum in the absence of the SD during baseline training (see Figure 1A), we found no difference in responding on the drug-seeking manipulandum in the presence and absence of the SD during extinction testing (see Figure 2A). The differential responding during baseline sessions was not surprising because the light signalled the availability of the drug. If the SD had acquired stimulus-activating properties during these sessions, it may have increased responding during extinction tests, thus masking stimulus control over drug-seeking. This is further substantiated by the fact that responding on the drug-taking manipulandum was dramatically increased in the presence of the SD during the same extinction sessions. In fact, responses on both the drug-taking and inactive levers increased, suggesting that the SD had a general activating effect on responding. Rats that had no prior experience in the self-administration paradigm did not show increased responding in the presence of the SD (data not shown), confirming that the effect is due to previous associations between the SD and the drug infusion. This interpretation is consistent with evidence from studies examining the secondary reinforcing properties of drug-paired stimuli. For example, clinical data show that drug-paired stimuli can produce increased craving in human addicts in a drug-free state (Childress et al., 1988; Ehrman et al., 1992; Ludwig et al., 1974; O’Brien, Childress, McLellan, & Ehrman, 1992), an effect that is both robust and long lasting. Similarly, studies with laboratory animals have shown that presentation of drug-paired stimuli can lead to reinstatement of drug-seeking and drug-taking (Ettenberg, Petit, & Bloom, 1996; Goldberg, Morse, & Goldberg, 1976; Gracy et al., 2000; Markou, Weiss, Gold, Caine, Schulteis, & Koob, 1993; McFarland & Ettenberg, 1997; Meil & See, 1996), even up to 11 weeks after the last reinforced session (Shalev et al., 2001).

Conventional measures of opiate withdrawal confirmed that rats were experiencing the aversive effects of opiate abstinence during extinction testing. Our findings also corroborate previous reports (Lieblich et al., 1991) that withdrawn rats consume less of a sweet solution than do opiate-naive control subjects. Opiate– experienced rats continue to show decrements in intake up to 12 days after their last drug self-administration session. We have considered the possibility that a conditioned taste aversion to sucrose may have contributed to this persistent decrement in intake over the prolonged abstinence tests. This interpretation is unlikely because rats with limited opiate experience do not differ from controls in saccharin intake when tested 8 days after initial withdrawal testing (Lieblich et al., 1991). Our findings from the somatic withdrawal tests were similar to those of the sucrose intake tests in that withdrawal symptoms were elevated in opiate-experienced rats at every period of assessment. Over the course of training, and prior to any extinction testing, self-administering rats received approximately 18 days of heroin experience, taking an average of 10-12 infusions per 3-hr session. Early studies demonstrate that rats administered morphine for a 6-week period show a significant increase in frequency of wet dog shakes up to 23 days later (Wilder & Pescor, 1967, 1970); this behaviour persists up to 180 days after morphine exposure (Martin, Wilder, Eades, & Pescor, 1963). It should be pointed out, however, that the lack of decline in somatic symptoms over time in our study is unusual. The increased frequency of somatic symptoms during testing may be due to quantifying these behaviours via observation of video recording, although our numbers are similar to those reported by Bardin, Kim, and Seigel (2000). As well, sucrose intake, which is not a subjective measure, changed very little over prolonged abstinence tests. Most importantly, if somatic symptoms had declined over prolonged abstinence testing, the dissociation between operant measures and somatic withdrawal symptoms would have been even greater than currently reported at the 12-day test.

In conclusion, our findings from the operant extinction tests suggest that the time course of withdrawal may involve distinct somatic and motivational aspects. These findings are in agreement with previous studies in which the motivation aspects of opioid withdrawal were measured in a conditioned place preference paradigm (Mucha, 2001). For example, rats administered methylnaloxonium, an opiate antagonist, show a place aversion when the drug is injected into a variety of brain sites, but only some of these sites are capable of inducing the classic overt signs of opiate withdrawal (Stinus, Le Moal, & Koob, 1990). Furthermore, the neuroleptic alpha-flupenthixol blocks the aversive motivational effects of naloxone-precipitated withdrawal, but does not attenuate naloxone-precipitated somatic withdrawal signs (Bechara, Nader, & van der Kooy, 1995). In conjunction with results from our study, these findings imply that our operant measures of drug-seeking and drug-taking are probably more closely related to motivational states of withdrawal, which are dissociated from observable somatic withdrawal symptoms.


Bardin, L., Kim, J. A., & Siegel, S. (2000). The role of formalin-induced pain in morphine tolerance, withdrawal, and reward. Experimental and Clinical Psychopharmacology, 8, 61-67.

Bechara, A., Nader, K., & van der Kooy, D. (1995). Neurobiology of withdrawal motivation: Evidence for two separate aversive effects produced in morphine– naive versus morphine-dependent rats by both naloxone and spontaneous withdrawal. Behavioral Neuroscience, 109, 91-105.

Bhargava, H. N. (1994). Diversity of agents that modify opioid tolerance, physical dependence, abstinence syndrome and self-administration behavior. Pharmacological Reviews, 46, 293-324.

Blasig, J., Herz, A., Reinhold, K., & Zieglgansberger, S. (1973). Development of physical dependence on morphine in respect to time and dosage and quantification of the precipitated withdrawal syndrome in rats. Psychopharmacologia, 33, 19-38.

Childress, A. R., McLellan, A. T., Ehrman, R., & O’Brien, C. P. (1988). Classically conditioned responses in opioid and cocaine dependence: A role in relapse? NIDA Research Monograph, 84, 25-43.

De Vries, T., Schoffelmeer, A. N. M., Binnekade, R., & Vanderschuren, L. J. M. J. (1999). Dopaminergic mechanisms mediating the incentive to seek cocaine and heroin following long-term withdrawal of IV drug self-administration. Psychopharmacology, 143, 254-260.

Ehrman, R., Ternes, J., O’Brien, C. P., & McLellan, A. T. (1992). Conditioned tolerance in human opiate addicts. Psychopharmacology, 108, 218-224.

Erb, S., Shaham, Y., Sr Stewart, J. (1996). Stress reinstates cocaine-seeking behavior after prolonged extinction and a drug-free period. Psychopharmacology, 128, 408-412.

Ettenberg, A., MacConnell, L. A., & Geist, T. D. (1996). Effects of haloperidol in a response-reinstatement model of heroin relapse. Psychopharmacology, 124, 205-210.

Gawin, F. H. (1989). Cocaine addiction: Psychology and neurophysiology. Science, 251, 1580-1586.

Goldberg, S. R., Morse, W. H., & Goldberg, M. (1976). Behavior maintained under a second-order schedule by intramuscular injection of morphine or cocaine in rhesus monkeys. Journal of Pharmacology and Experimental Therapeutics, 199, 278-286.

Goldberg, S. R., Woods, J. H., & Schuster, C. R. (1969). Morphine: Conditioned increases in self-administration in

rhesus monkeys. Science, 166, 1306-1307.

Goldberg, S. R., Woods, J. H., & Schuster, C. R. (1971). Nalorphine-induced changes in morphine self-administration in rhesus monkeys. Journal of Pharmacology and Experimental Therapeutics, 176, 464-471.

Gracy, K. N., Dankiewicz, L. A., Weiss, F., & Koob, G. F (2000). Heroin-specific stimuli reinstate operant heroin– seeking behavior in rats after prolonged extinction. Pharmacology, Biochemistry and Behavior, 65, 489-494.

Grimm, J. W., Hope, B. T., Wise, R. A., & Shaham, Y. (2001). Incubation of cocaine craving after withdrawal. Nature, 412, 141-142.

Grimm, J. W., & See, R. E. (2000). Dissociation of primary and secondary reward-relevant limbic nuclei in an animal model of relapse. Neuropsychopharmacology, 22, 473-479.

Hughes, J. R., Higgins, S. T., & Bickel, W. K. (1994). Nicotine withdrawal versus other drug withdrawal syndromes: Similarities and dissimilarities. Addiction, 89, 1461-1470.

Jaffe, J. H. (1990). Drug addiction and drug abuse. In A. G. Gilman, T. W. Rall, A. S. Nies, & P. Taylor (Eds.), Goodman and Gilman’s the pharmacological basis of therapeutics (p. 522). New York: Pergamon Press.

Jaffe, J. H., Cascella, N. G., Kumor, K. M., & Sherer, M. A. (1989). Cocaine-induced cocaine craving. Psychopharmacology, 97, 59-64.

Johnston, L. D., Beninger, R. J., & Olmstead, M. C. (2001). Pimozide, like extinction, devalues stimuli associated with sucrose-taking. Pharmacology, Biochemistry and Behavior, 68, 583-590.

Koob, G. F., Stinus, L., Le Moal, M., & Bloom, F. E. (1989). Opponent process theory of motivation: Neurobiological evidence from studies of opiate dependence. Neuroscience and Biobehavioral Reviews, 13, 135-140.

Lieblich, L, Yirmiya, R., & Liebeskind, J. C. (1991). Intake of and preference for sweet solutions are attenuated in morphine-withdrawn rats. Behavioral Neuroscience, 105, 965-970.

Ludwig, A. M., Wilder, A., & Stark, L. H. (1974). The first drink. Archives of General Psychiatry, 30, 539-547. MacRae, J., & Siegel, S. (1997). The role of self-administra

tion in morphine withdrawal in rats. Psychobiology, 25, 77-82.

Markou, A., Weiss, F., Gold, L. H., Caine, S. B., Schulteis, G., & Koob, G. F. (1993). Animal models of drug craving. Psychopharmacology, 112,163-182.

Martin, M. R., Wilder, A., Eades, C. G., & Pescor, F. T. (1963). Tolerance to and physical dependence on morphine in rats. Psychopharmacologia, 4, 247-260.

McFarland, K., & Ettenberg, A. (1997). Reinstatement of drug-seeking behavior produced by heroin-predictive environmental stimuli. Psychopharmacology, 131, 89-92.

Meil, M. W., & See, R. E. (1996). Conditioned cue recovery

of responding following prolonged withdrawal from selfadministered cocaine in rats: An animal model of relapse. Behavioural Pharmacology, 7, 754-76392.

Meyer, R. E., & Mirin, S. M. (1979). The heroin stimulus. implications for the theory of addiction. New York: Plenum Medical Book Co.

Mucha, R. F. (2001). Preferences for tastes paired with a nicotine antagonists in rats chronically treated with nicotine. Pharmacology, Biochemistry and Behavior, 56, 175-179.

O’Brien, C. P., Childress, A. R., McLellan, A. T., & Ehrman, R. (1992). A learning model of addiction. In C. P. O’Brien & J. Jaffe (Eds.), Advances in understanding the addictive states (pp. 157-177). New York: Raven Press.

Olmstead, M. C., Lafond, M. V., Everitt, B. J., & Dickinson, A. (2001). Cocaine-seeking by rats is a goal directed action. Behavioral Neuroscience, 115, 247-291.

Olmstead, M. C., Parkinson, J. A., Miles, F., Everitt, B. J., & Dickinson, A. (2000). Cocaine-seeking by rats: Regulation, reinforcement and activation. Psychopharmacology, 152, 123-131.

Robinson, T. E., & Berridge, K. C. (1993). The neural basis of drug craving: An incentive-sensitization theory of addiction. Brain Research Reviews, 18, 247-291.

Self, D. W., Beluzzi, J. D., Kossuth, S., & Stein, L. (1996). Self-administration of the DI agonist SKF 82958 is mediated by D1, not D2, receptors. Psychopharmacology, 123, 303-306.

Shaham, Y., Adamson, L. K., Grocki, S., & Corrigall, W. A. (1997). Reinstatement and spontaneous recovery of nicotine seeking in rats. Psychopharmacology, 130, 396403.

Shaham, Y., Rajabi, H., & Stewart, J. (1996). Relapse to heroin-seeking in rats under opioid maintenance: The effects of stress, heroin priming, and withdrawal. Journal of Neuroscience, 16, 1957-1963.

Shaham, Y., & Stewart, J. (1995). Stress reinstates heroin– seeking in drug-free animals: An effect mimicking heroin, not withdrawal. Psychopharmacology, 119, 334-341.

Shalev, U., Morales, M., Hope, B., Yap, J., & Shaham, Y. (2001). Time dependent changes in extinction behavior and stress-induced reinstatement of drug seeking following withdrawal from heroin in rats. Psychopharmacology, 156, 98-107.

Solomon, R. L., & Corbit, J. D. (1974). An opponent– process theory of motivation: I. Temporal dynamics of affect. Psychological Review, 81, 119-145.

Stewart, J., & Vezina, P. (1988). A comparison of the effects of intra-accumbens injections of amphetamine and morphine on reinstatement of heroin intravenous self-administration behavior. Brain Research, 457, 287-294.

Stinus, L., Le Moal, M., & Koob, G. F. (1990). Nucleus accumbens and amygdala are possible substrates for the aversive stimulus effects of opiate withdrawal.

Neuroscience, 37, 767-773.

Tiffany, S. T. (1990). A cognitive model of drug urges and drug-use behavior: Role of automatic and nonautomatic processes. Psychological Review, 97, 147-168.

Wilder, A. (1973). Dynamics of drug dependence. Archives of General Psychology, 28, 611-616.

Wilder, A., & Pescor, F. T. (1967). Classical conditioning of a morphine abstinence phenomenon, reinforcement of opioid-drinking behavior and ‘relapse’ in — addicted rats. Psychopharmacologia, 10, 255-284.

Wilder, A., & Pescor, F. T. (1970). Persistence of ‘relapsetendencies’ of rats previously made physically dependent on morphine. Psychopharmacologia, 16, 375-384.

Wilson, M. C., Hitomi, M., & Schuster, C. R. (1971). Psychmotor stimulant self-administration as a function of dosage per injection in the rhesus monkey. Psychopharmacologia, 22, 271-281.

Wise, R. A., & Bozarth, M. A. (1987). A psychomotor stimulant theory of addiction. Psychological Review, 94, 469492.

Wise, R. A., Murray, A., & Bozarth, M. A. (1990). Bromocriptine self-administration and bromocriptine– reinstatement of cocaine-trained and heroin-trained lever pressing in rats. Psychopharmacology, 100, 355-360.

Wise, R. A., Newton, P., Leeb, K., Burnette, B., Pocock, D., & Justice, J.B. (1995). Fluctuations in nucleus accumbens dopamine concentration during intravenous cocaine self-administration in rats. Psychopharmacology, 120, 10-20.

Zellner, D. A., Dacanay, R. J., & Riley, A. L. (1984). Opiate withdrawal: The results of conditioning or physiological mechanisms. Pharmacology, Biochemistry and Behavior 20, 175-180.

KIM G. C. HELLEMANS, Queen’s University YAVIN SHAHAM, National Institute on Drug Abuse MARY C. OLMSTEAD, Queen’s University

Send correspondence to Mary C. Olmstead, Department of Psychology, Queen’s University, Kingston, Ontario, K7L 3N6. (Tel: 613-533-6208; Fax: 613-533-2499; E-mail: olmstead@psyc.queensu.ca).

Copyright Canadian Psychological Association Dec 2002

Provided by ProQuest Information and Learning Company. All rights Reserved