Transition from Moderate to Excessive Drug Intake: Change in Hedonic Set Point

See allHide authors and affiliations

Science  09 Oct 1998:
Vol. 282, Issue 5387, pp. 298-300
DOI: 10.1126/science.282.5387.298


Differential access to cocaine self-administration produced two patterns of drug intake in rats. With 1 hour of access per session, drug intake remained low and stable. In contrast, with 6 hours of access, drug intake gradually escalated over days. After escalation, drug consumption was characterized by an increased early drug loading and an upward shift in the cocaine dose-response function, suggesting an increase in hedonic set point. After 1 month of abstinence, escalation of cocaine intake was reinstated to a higher level than before. These findings may provide an animal model for studying the development of excessive drug intake and the basis of addiction.

A critical problem in drug addiction research is to understand the differences between controlled and uncontrolled drug use, the latter being an essential feature of drug addiction (1, 2). These two patterns of drug use may be observed simultaneously in different individuals or they may represent successive stages in the same individuals. The transition from drug use to addiction often involves a gradual process of escalated drug intake, whereby an individual's consumption becomes exaggerated with chronic exposure to a drug (2). Because escalation of drug use defines a common feature of drug addiction, the study of the factors that govern its development may help to explain the transition from drug use to drug addiction.

In animal models of drug self-administration, availability plays a role in determining the pattern of drug intake, as suggested by different studies with different drug access conditions (3). Numerous studies have restricted drug access to a few hours per day and produced a regular and stable pattern of consumption. In contrast, others have shown that, with continuous access to the drug, different patterns of drug intake develop, including the binge-like patterns of psychomotor stimulant use observed in both animals and humans (3,4). Unknown, however, are the mechanisms responsible for escalation of drug intake that may set the stage for addiction. Here, we directly assessed the effect of the history of drug use on cocaine-taking behavior and have begun to characterize the factors responsible for the development of escalation of cocaine use and its reversal and relapse.

The duration of access to cocaine (250 μg per infusion) dramatically influenced intake (5). With short access (ShA), cocaine intake remained stable over time, whereas with long access (LgA), cocaine intake gradually escalated from 71 to 110 infusions (Fig. 1A). Escalation of cocaine use appeared as early as the first hour of access to the drug (Fig. 1B), which rules out an acute within-session tolerance (6). Furthermore, increased cocaine use did not result from an increase in general activity (7). Within the first hour, the temporal course of cocaine intake was unchanged in ShA rats with repeated testing (Fig. 1C); in LgA rats with repeated testing, cocaine intake increased dramatically the first 10 min and then dropped and stabilized at a larger amount (Fig. 1D). This increased drug-loading behavior may reflect an acquired “need state” for a higher level of cocaine intoxication (1).

Figure 1

Effect of drug availability on cocaine intake (mean ± SEM). (A) In LgA rats (n= 6) but not in ShA rats (n = 7), total cocaine intake started to increase significantly from the fifth session (P< 0.05; sessions 5 to 12 compared with session 1) and continued to increase thereafter (P < 0.05; session 5 compared with sessions 10 to 12). (B) During the first hour, LgA rats took progressively more infusions than ShA rats (P < 0.05; sessions 8, 10, 11, and 12). (C and D) In LgA rats (D) but not in ShA rats (C), cocaine intake measured at 10, 50, and 60 min increased (P < 0.05; first session compared with last session). As shown in the inset in (D), this change (last minus first session) was greater the first 10 min (P < 0.05; compared with the first 10 min). *, P < 0.05 (Student'st test after appropriate one-way and two-way analysis of variance).

To establish the generality of these observations and to further determine the nature of the changes underlying increased cocaine use, escalated intake was produced in a different experimental condition (8). Again, changing the duration of drug availability induced two patterns of cocaine intake (Fig. 2, A and B). Escalation of total cocaine intake showed no sign of stabilization even after 22 sessions of self-administration (Fig. 2A). Moreover, the temporal pattern of cocaine intake in LgA rats peaked in the first 10 min and then dropped to stabilize at an amount higher than that in ShA rats. Therefore, escalated cocaine use is a robust phenomenon that is observable under different experimental conditions.

Figure 2

Reproduction of escalated cocaine use in another setting. (A) In LgA rats (n= 12) but not in ShA rats (n = 12), mean total cocaine intake (± SEM) started to increase significantly from session 5 (P < 0.05; sessions 5 to 22 compared with session 1) and continued to increase thereafter (P < 0.05; session 5 compared with sessions 8 to 10, 12, 13, and 17 to 22). (B) During the first hour, LgA rats self-administered more infusions than ShA rats during sessions 5 to 8, 11, 12, 14, 15, and 17 to 22 (P < 0.05). (C) Mean infusion (± SEM) per cocaine dose tested. LgA rats took significantly more infusions than ShA rats at doses of 31.25, 62.5, 125, and 250 μg per infusion (P < 0.05). (D) After escalation, LgA rats took more cocaine than ShA rats regardless of the dose (P < 0.05). *, P < 0.05 (Student's t test after appropriate one-way and two-way analysis of variance).

In the simple reinforcement schedule used here, the cocaine dose-response curve has a negative slope; above a certain threshold dose, an increase in the unit dose produces a proportional decrease in self-infusions (9). This phenomenon suggests that the animals regulate their intoxication around some endogenous reference or “hedonic set point” (1, 9). It is possible that escalated cocaine use results from elevation of a hedonic set point, a hypothesis consistent with the dramatic drug-loading behavior described previously (Fig. 1D). This hypothesis also predicts that the entire dose-response curve should be shifted upward after escalation. Although drug tolerance (or even sensitization) could account for some aspects of increased use, pharmacological tolerance or sensitization usually reflects a change in drug sensitivity that translates into a horizontal shift of the dose-response curve either to the right (tolerance) or to the left (sensitization) (10, 11).

To test these predictions, we examined a wide range of cocaine doses (12). Decreasing the dose produced an increase in cocaine self-infusions (Fig. 2C), a phenomenon that is true for both ShA and LgA rats. This compensatory process ceased abruptly when the dose was too low to maintain responding (that is, when responding for this dose was similar to responding for vehicle, indicated by the broken line inFig. 2C). Importantly, the minimum dose able to maintain cocaine-taking behavior (31.25 μg per infusion) did not differ between groups, suggesting that drug sensitivity was unchanged (Fig. 2C). In fact, escalated cocaine use produced an upward shift without a horizontal shift in the dose-response curve. At all doses tested above the threshold, LgA rats self-administered more cocaine than ShA rats (Fig. 2D). This effect did not result from an increase in general activity, because both groups responded similarly for vehicle (12). These findings suggest that escalated cocaine use results from a change in the hedonic set point for cocaine instead of from a simple change in sensitivity to the drug (1). This change induced by increasing drug availability suggests that an even longer access to the drug could lead to a further increase in set point.

To assess the persistence of escalated cocaine use, drug access was interrupted for 35 days and then restored (13). In ShA rats, abstinence had no effect on cocaine intake (Fig. 3). In contrast, in LgA rats, cocaine intake dropped to the quantity observed in ShA rats after abstinence, indicating a complete recovery from prior increased cocaine use. However, escalated cocaine intake was reinstated by reexposing rats to the LgA regimen. In the same rats, reinstatement of escalated cocaine intake was more dramatic than when it was originally established (compare Fig. 3, left and right). This suggests that, although abstinence promotes a short-lived return to previous levels of use, a history of drug escalation facilitates the relapse. The return of cocaine use to baseline during abstinence is reminiscent of a reversal of tolerance; however, the more dramatic escalation observed during reinstatement is more consistent with an allostatic dysregulation in hedonic set point (1).

Figure 3

Relapse to escalated cocaine use after abstinence. During the initiation phase (left), escalated cocaine use in LgA rats (n = 5) was significant after session 19 (P < 0.05). During reinstatement (right), escalated cocaine intake in LgA rats was fully reinstated after only six reinstatement sessions (P < 0.05). The session-by-session fluctuation observed in ShA rats (n = 6) during reinstatement was not significant (simple main effect). *, P < 0.05 (Student'st test after appropriate one-way and two-way analysis of variance).

This study demonstrates the existence of two patterns of cocaine-taking behavior. Rats allowed short access to cocaine maintained a stable and low intake for many weeks, and this demonstrates that self-administering a drug repeatedly does not lead obligatorily to a process of escalated use analogous to that described in human drug addicts. This observation shows the limitations of equating drug-taking behavior per se with drug addiction and of analyzing drug addiction only in terms of the positive-reinforcing effects of a drug (1). Indeed, positive-reinforcement analyses explain how behavior is initiated and maintained by a drug but do not explain why an individual self-administers a particular amount of drug and why such an amount may or may not change over time.

In this context, the escalation in drug intake observed in rats that had longer access to cocaine (6 hours) may offer a better animal model of the transition of drug use to drug addiction. Escalated cocaine use produced an increase in drug loading and an upward shift in the cocaine dose-effect function, suggesting an increase in hedonic set point as a potential mechanism for increased drug use. Animals, and perhaps human addicts, may take more cocaine after escalated intake not because they are simply tolerant to its rewarding effects but because they are trying to reach and then to maintain a higher state of intoxication. Certainly, tolerance, as classically defined, may contribute to this allostatic dysregulation; however, the current hypothesis is that an increase in hedonic set point would characterize the transition to the addicted state (1).

Finally, the temporal course of escalated drug use suggests that the “switch mechanism” of the transition to drug addiction may be gradual rather than abrupt. This is consistent with the slow accumulation of some drug-responsive gene products in the brain described recently during chronic exposure to cocaine (14). As shown by the dramatic reinstatement of escalated cocaine use after abstinence, part of these modifications are long lasting. Analysis of the brain changes responsible for the escalation in self-administration of cocaine observed in this animal model may help to explain the chronic, relapsing nature of drug addiction.

  • * To whom correspondence should be addressed. E-mail: aserge{at}


Navigate This Article