Introduction
In drug addicts, craving1, 2 or automatic3 responses resulting from the association between environmental stimuli and the subjective effects of the drug are widely implicated as long-lasting risk factors for relapse. It is unknown, however, whether drug-related conditioning also contributes to drug use at early stages of the addictive cycle. Therefore, we investigated whether environmental stimuli paired with a single cocaine self-administration experience acquire sufficient strength to elicit cocaine-seeking behavior when subsequently presented without the drug.
Rats previously trained to press a lever for food pellets were given response-contingent access to cocaine for 2 h. This session was initiated by onset of white noise (70 dB) that served as a contextual stimulus associated with drug availability (S+). Each response at the right-hand lever resulted in an intravenous (i.v.) infusion of cocaine hydrochloride (0.25 mg in 0.1 ml saline solution), followed by a 20-s period (signaled by a white cue light) during which the lever remained inactive to prevent overdose. At the end of the session, rats were removed from the operant conditioning chamber for 1 h. Subsequently, a 2-h conditioning session was conducted during which only saline (non-reward) was available, signaled by illumination of the chamber's house light (S-). Responses at the right-hand lever resulted in an infusion of saline (0.1 ml) signaled by a 20-s intermittent tone. Responses at the left-hand lever were recorded but had no programmed consequence. Rats responded with significantly greater frequency during the 2 h of cocaine availability than during the saline/non-reward session (Fig. 1). Some rats (
50%, n = 4) showed an initial burst of responding, possibly reflecting extinction responses linked to the previous food-shaping procedure. Regardless of the presence or absence of an initial extinction-like response pattern, however, the majority of rats (85%) developed sustained responding with the emergence of increasingly regular inter-response intervals characteristic of i.v. cocaine-reinforced behavior (Fig. 2). Twenty-four hours later, responding was extinguished in daily 1-h sessions during which levers were available, but i.v. solutions and the corresponding stimuli were withheld. Rats reached the extinction criterion defined as five or fewer responses in each of three consecutive sessions within 11.3 
2.3 d (mean s.e.m.).
Figure 1: Reinstatement of cocaine-seeking after a single drug experience.
All procedures were conducted in strict adherence to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. (a) Self-administration: responses during access to intravenous cocaine (COC) or saline (SAL) in the presence of distinct stimuli paired with availability of cocaine (S+) or saline (S-). ##Significantly different from SAL/S- (ANOVA: F1,6 = 14.08; P < 0.01). Reinstatement: mean number of responses across the last 3 d of extinction (EXT) and responses during subsequent exposure to the stimuli previously paired with cocaine (S+) versus saline (S-) availability. **Significantly different from EXT (P < 0.01), Newman Keuls post hoc tests following ANOVA (immediate: F2,6 = 18.40, P < 0.01; 3 months: F2,6 = 22.10, P < 0.01; 6 months: F2,5 = 16.83, P < 0.01; 9 months: F2,5 = 13.20,P < 0.01). S+ effects at 12 months approached statistical significance (F2,4 = 4.25; P < 0.055). (b) Responding at the inactive lever was negligible.
Full size image (24 KB)Figure 2: Response patterns during access to cocaine and saline.
Lever-press responses are shown as tick marks. Rats 1, 2 and 3 were representative of 85% of the animals (n = 6) and showed sustained responding throughout the cocaine session (COC/S+) with the beginning development of regular inter-response intervals characteristic of i.v. cocaine-self-administration. Rat 4 showed an initial response burst during cocaine availability. During the saline session (SAL/S-), scattered response bursts characteristic of extinction were observed.
For the conditioned reinstatement test, stimuli previously paired with cocaine availability (S+) or non-reward (S-) were reintroduced. On all testing occasions, cocaine-seeking behavior was determined in counterbalanced order on the first and second day after the last extinction session. Rats resumed responding at the previously active lever when exposed to the S+; the S- had no effect (Fig. 1a; 'immediate'). The animals then were returned to the vivarium and retested at 3-month intervals for 1 year. Each test was preceded by a set of extinction sessions to reestablish responding consistent with the extinction criterion (requiring 9.7 2.3, 7.3 1.8, 6.0 1.8 and 5.2 1.0 sessions for the 3, 6, 9 and 12-month tests, respectively). Reinstatement of responding under the S+ (but not the S-) condition remained significantly different from extinction until 9 months and approached statistical significance (P < 0.055) at 12 months (Fig. 1a). Responding at the inactive lever was negligible (Fig. 1b).
To test whether the persistence of the motivating effects of the cocaine S+ is specific to drug-related conditioning or extends to conventional reinforcers, we prepared a second group of eight rats. The same general procedures were used, with the exception that sweetened condensed milk—a highly potent non-drug reinforcer4—was substituted for cocaine. As in the cocaine group, responding during the single session of reward availability was significantly greater than during the non-reward session, and the extinction criterion (
5 responses/h) was reached within 5.5 0.6 sessions (Fig. 3). Subsequent exposure to the milk-related S+ produced only a trend (P < 0.1) toward increased responding compared to extinction, and the S+ failed to elicit recovery of responding in the second test conducted 3 months later (4.8 0.5 extinction sessions at criterion; Fig. 3).
Figure 3: Conditioned reinstatement after one-time access to a highly palatable conventional reinforcer, sweetened condensed milk (SCM).
(a) Self-administration: responses during availability and non-availability (non-reward) of SCM in the presence of contextual stimuli paired with SCM-reward (S+) versus non-reward (S-). ##Significantly different from non-reward/S- (F1,7 = 36.67; P < 0.01). Reinstatement: average responses across the last three extinction (EXT) sessions and responses during subsequent exposure to the SCM S+ and S-. The SCM S+ failed to elicit significant recovery of responding on both the immediate (F2,7 = 3.53; nonsignificant, n.s.) and 3 months 'abstinence' (F2,5 = 2.95; n.s.) testing occasions. (b) Responses at the inactive lever. ##Significantly different from SCM/S- (F1,7 = 22.52; P < 0.01).
Full size image (21 KB)Ample evidence indicates that stimuli paired with drug use acquire the ability to elicit craving and facilitate relapse1, 2, 5, 6, 7. However, such effects are thought to require frequent reinforcement of the association between a substance of abuse and specific environmental cues. To date, long-lasting conditioned reinstatement by drug cues in animals has been documented only after multiple conditioning sessions5, 7. Challenging this presumption, the data reported here show that the motivating effects of a stimulus paired with only a single cocaine self-administration experience are robust and long-lasting.
A second observation is that drug-related environmental stimuli exert more powerful and long-lasting control over behavior than stimuli associated with potent conventional reinforcers. The mechanisms underlying the differential efficacy of the drug- versus non-drug-related stimuli are not known. Nonetheless, this finding suggests that learning and associative processes responsible for drug-related conditioning are distinct from those responsible for conditioning with conventional reinforcers. It is possible that neural plasticity induced by cocaine (but not conventional reinforcers) strengthens stimulus-reward associations and facilitates memory consolidation or storage of drug-related stimuli8, 9. In rats, a single in vivo exposure to cocaine can initiate neurophysiological events closely resembling long-term potentiation, and this increase in synaptic strength lasts for several days10. Consistent with a role for drug-induced neural plasticity in the motivating effects of drug cues, a single cocaine treatment is sufficient to induce conditioned behavioral sensitization10, 11. Similarly, a single treatment with amphetamine or morphine induces long-lasting locomotor sensitization associated with hyper-responsiveness of meso-corticolimbic dopamine and acetylcholine transmission12, 13.
Finally, our results suggest that conditioned responses to drug-related stimuli not only facilitate relapse, but could also contribute to the transition from initial drug use to addiction. Clearly, a host of biological and other factors such as social influence and drug availability are involved in this process. However, drug-related learning may have an important role in perpetuating desire for the drug once drug use has begun.
