Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Drug self-administration procedures have been useful for investigating the neurobiological and pharmacological mechanisms underlying the reinforcing effects of cocaine, and such information may lead to the development of new and effective medications for cocaine abuse and dependence (Mello and Negus, 1996; Mendelson and Mello, 1996). Cocaine nonselectively blocks the reuptake of the monoamine neurotransmitters dopamine, norepinephrine, and serotonin (Taylor and Ho, 1978); however, the reinforcing effects of cocaine have been most compellingly related to its action as an indirect dopamine agonist (Wise and Bozarth 1987; Koob, 1992). In this regard, the potency with which many monoamine reuptake inhibitors serve as i.v. reinforcers is highly correlated with their binding to the dopamine transporter (Ritz et al., 1987; Bergman et al., 1989). Moreover, in contrast to selective dopamine reuptake inhibitors, selective norepinephrine and serotonin reuptake inhibitors have not been shown to produce reinforcing effects in drug self-administration procedures.

Studies with direct dopamine receptor agonists offer another pharmacological method for elucidating the neurobiological mechanisms underlying the abuse-related effects of cocaine. In contrast to cocaine or other dopamine reuptake inhibitors, which increase the activation of dopamine receptors nonselectively, direct dopamine agonists can be used to selectively activate D_1-like (D₁, D₅) or D_2-like (D₂, D₃, D₄) receptors which, in turn, have different anatomical and functional characteristics. For example, D_1-like agonists modulate the expression of dynorphin and Substance P within striatonigral neurons, whereas D_2-like agonists regulate enkephalin expression within striatopallidal neurons (Gerfen et al., 1990). In addition, D₁ receptors are positively coupled with adenylyl cyclase through stimulatory G-proteins, whereas D₂ receptors are negatively coupled to adenylyl cyclase through inhibitory G-proteins (Schwartz et al., 1992). Moreover, adaptations in these different systems occur in response to chronic cocaine administration (Terwilliger et al., 1991; Nestler and Aghajanian, 1997). Thus, studies with selective D_1-ike and D_2-like agonists may provide one way to clarify the roles of distinct neural pathways and molecular mechanisms involved in cocaine addiction.

Although it is generally believed that both D_1-like and D_2-like receptor mechanisms may be involved in the reinforcing and other behavioral effects of cocaine, recent studies suggest that D_1-like and D_2-like agonists have differing profiles of cocaine-related actions. On the one hand, both D_1-like and D_2-like agonists have been reported to maintain i.v. self-administration in animals previously trained to self-administer cocaine (Woolverton et al., 1984; Self and Stein, 1992; Caine and Koob, 1993, 1995; Weed and Woolverton, 1995; Nader and Mach, 1996), and this is consistent with findings that both D_1-like and D_2-like receptor blockers seem to antagonize cocaine self-administration behavior (Bergman et al., 1990; Caine and Koob, 1994). On the other hand, D_1-like agonists may not maintain responding as robustly or under as wide a range of conditions as was reported for cocaine and D_2-like agonists (Grech et al., 1996). In addition, in both rats and monkeys, D_1-like agonists may attenuate the ability of cocaine to reinstate responding after extinction of cocaine self-administration whereas D_2-like agonists appear to enhance this effect of cocaine (Self et al., 1996; Barrett-Larimore and Spealman, 1997). These latter results agree with previous findings that D_2-like agonists produced leftward shifts in the dose-effect function for cocaine self-administration (Caine and Koob, 1995; Caine et al., 1997). The effects of D_1-like agonists on cocaine self-administration are unknown, but, based on results from the studies described above, D_1-like agonists may differ qualitatively from D_2-like agonists in the manner by which they modify the abuse-related effects of cocaine.

The present study was designed to further evaluate differing roles of D_1-like and D_2-like receptors in mediating the abuse-related effects of cocaine by systematically comparing the effects of D_1-like and D_2-like agonists in rats trained to self-administer cocaine. First, D_1-like or D_2-like agonists were substituted for cocaine in rats with extensive cocaine self-administration experience to rigorously evaluate and compare the reinforcing effects of these agents. Second, modification of the reinforcing effects of cocaine by D_1-like and D_2-like agonists was studied by assessing the effects of pretreatments on full dose-effect functions for cocaine self-administration. In general, the results of this study suggest that D_2-like agonists, but not D_1-like agonists, maintain self-administration and enhance the effects of self-administered cocaine in rats with extensive cocaine self-administration training. These results are consistent with those from parallel studies in which the cocaine-discriminative stimulus generalized to D_2-like agonists, but not D_1-like agonists, in rats with extensive cocaine discrimination training (Caine et al., unpublished observations).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals

Cocaine self-administration studies were conducted in 32 male Wistar rats (Charles River, Wilmington, MA). The rats weighed approximately 350 g at the start of the study and were maintained in the range of 400 to 600 g with once-daily feedings of standard rat chow (Rat Diet 5012; PMI Feeds, Inc., St. Louis, MO). Bacon-flavored biscuits (Bio-Serve, Frenchtown, NJ) were also provided once or twice weekly, primarily for enrichment. Each rat was housed individually with free access to water in a temperature- and humidity-controlled facility that was maintained on a 12-h light/dark cycle (lights on from 7:00 AM to 7:00 PM). Behavioral testing generally began toward the end of the light cycle (4:00 PM to 6:00 PM). Animal maintenance and research were conducted in accordance with the guidelines provided by the National Institutes of Health Committee on Laboratory Animal Resources, and all protocols were approved by the Institutional Animal Care and Use Committee. The health of the rats was periodically monitored by consulting veterinarians.

Apparatus

Drug self-administration studies were conducted in experimental chambers (21 cm × 29.5 cm × 24.5 cm) placed within sound-attenuating cubicles equipped with a house light and an exhaust fan. Each chamber contained three response levers. On one wall of the chamber, two levers were located on either side of a pellet receptacle, 3 cm above the grid floor and 1.5 cm from the side walls. During preoperative training, responding on either of these two levers was maintained by 45-mg food pellets (A/I Rodent Pellets; P. J. Noyes Co., Lancaster, NH) that were delivered by a pellet dispenser (ENV-203) to the food receptacle located between the two levers and 2 cm above the floor. A third lever, which was used for drug-maintained responding, was located on the center of the opposite wall, 3 cm above the floor. A single white stimulus light located above each lever was used to indicate that responding had scheduled consequences.

A single-channel fluid swivel (Lomir Biomedical, Malone, NY) was mounted on a balance arm above the chamber and attached to a spring lead with an inner Tygon tubing. The other end of the swivel was connected to an infusion pump that permitted automated delivery of i.v. drug injections for specified durations (see below). The infusion pump (PHM-100; 3.3 rpm motor) and all other components of the operant chambers and associated hardware were supplied by MED Associates Inc. (Georgia, VT). Scheduling of experimental events and data collection were accomplished with a DOS-based microcomputer system that was equipped with programs written in MED Associates MedState Notation.

Drug Self-Administration Procedures

Summary of Groups and Procedures. All rats were first allowed to respond in 2-h drug self-administration sessions (5-6 days per week) with a single unit dose of cocaine (0.32 mg per injection; approximately 1.0 mg/kg/injection). After stable self-administration behavior was established (see below), the rats were divided into two groups of 16 rats each. One group of rats was trained further, with a multiple-component drug self-administration procedure. In this procedure, the unit dose of cocaine was varied during the test session to obtain cocaine self-administration dose-effect functions. After stable cocaine dose-effect functions were established (see below), D_1-like and D_2-like dopamine agonists were either substituted for cocaine or administered as pretreatments to cocaine self-administration, until at least eight rats had been tested with each of the following dopamine agonists: SKF 82958, SKF 77434, quinelorane, and 7-hydroxy-dipropylaminotetralin (7-OH-DPAT). The benzazepine D_1-like agonist SKF 82958 was extensively studied, because this agent was previously shown to maintain responding, and additionally, to attenuate cocaine's reinstating effects after extinction of cocaine self-administration in both rats and monkeys (Weed and Woolverton, 1995; Grech et al., 1996; Self et al., 1996; Barrett-Larimore and Spealman, 1997). SKF 77434, a relatively low-efficacy D₁ dopamine agonist, was reported to support drug self-administration in rats, but not in primates (Self and Stein, 1992; Weed and Woolverton, 1995). For comparison with D_1-like agonists, the D_2-like agonists quinelorane and 7-OH-DPAT were chosen, because these drugs maintained responding and/or enhanced the effects of self-administered cocaine in both rats and monkeys (Caine and Koob, 1995; Parsons et al., 1996; Nader and Mach, 1996; Caine et al., 1999b).

Preoperative Training. Lever responding was initially shaped by using food reinforcement. During daily 1-h sessions of food availability, the stimulus lights were illuminated above the two levers located on either side of the pellet receptacle, and rats were trained to press either of these two levers under a fixed ratio (FR) 1 schedule of food reinforcement. Food training continued until rats earned 100 food pellets in a single session. Rats were then implanted with an i.v. catheter for drug self-administration studies.

Surgery. Rats were anesthetized with an isofluorane/oxygen vapor mixture and prepared with chronic indwelling i.v. catheters as previously described (Caine et al., 1993), with minor modifications (Emmett-Oglesby et al., 1993). Each catheter consisted of a 13-cm length of Silastic tubing fitted to a 22-gauge guide cannula that was bent at a right angle. The guide cannula was encased in dental cement anchored with a 0.5-inch-diameter circular nylon mesh. The tubing was passed s.c. from the animal's back to the right external jugular vein. All animals were allowed to recover for approximately 7 days before they were given access to i.v. cocaine. After surgery, a prophylactic dose of ticarcillin (20 mg/kg, i.v.) was delivered for 5 days to forestall infection. Thereafter, catheters were flushed daily with sterile physiological saline containing heparin (3 USP U/0.1 ml). If blood could not be withdrawn through the catheter, then catheter patency was tested by administering a solution containing 25 mg/ml ketamine and 1.25 mg/ml midazolam (0.05-0.1 ml, i.v.). Animals with patent catheters exhibited prominent signs of sedation within 3 s of i.v. injection. Animals with faulty catheters were prepared with a new catheter as described above, unless both external jugular veins were exhausted.

Initial Cocaine Self-Administration Training. Daily 2-h sessions began 1 week after surgery. Rats were placed in the experimental chambers and received a noncontingent automated infusion to fill the catheter (17 µl) and to deliver one unit dose of cocaine (0.32 mg). The unit dose was delivered in 56 µl over ca. 3.2 s. Thereafter, a cue light above the drug self-administration lever was illuminated, and responding on that lever was maintained by i.v. cocaine injections under a FR 1 timeout (TO) 20-s schedule of reinforcement. Under this schedule, completion of the response requirement produced an i.v. infusion of the unit dose of cocaine and initiated a 20-s TO period during which the cue light was turned off and responses had no scheduled consequences. In subsequent training sessions, the response requirement was gradually increased to a FR 5. Each rat was trained until cocaine self-administration behavior stabilized, with stability defined as three consecutive sessions during which there was less than 10% variation in the total number of cocaine reinforcers earned per session.

Multiple-Component Cocaine Self-Administration Training. A procedure for rapid assessment of responding maintained by different doses of cocaine was adapted from a previous study (Caine and Koob, 1995). These multiple-component sessions consisted of three or four 20-min components separated by 2-min TO periods. Dose-effect functions were determined by increasing the volume of cocaine injections in successive components so that 0-, 17-, 56-, or 178-µl injections were delivered in approximately 0, 1, 3.2, and 10 s, respectively. Drug solutions consisted of 0.56 mg/ml cocaine, 1.78 mg/ml cocaine, or 5.6 mg/ml cocaine, yielding unit doses of 0, 0.01, 0.032, 0.10, 0.32, and 1.0 mg total cocaine per injection (approximately 0.032-3.2 mg/kg/injection). During any one session, different doses of cocaine were always presented in an ascending order, and an injection of the cocaine dose available during each component was noncontingently administered at the beginning of the component. Substitution or pretreatment testing began after drug self-administration behavior stabilized, with stability defined as three consecutive multiple-component sessions during which the dose that maintained peak responding remained stable within a half log-unit range.

Multiple-Component Substitution Tests. D_1-like and D_2-like agonists were substituted for cocaine, to assess their effects in the multiple-component self-administration procedure. Drugs and doses were selected on the basis of previous studies (Self and Stein, 1992; Caine and Koob, 1993), and included SKF 82958 (0.001-0.032 mg/injection), SKF 77434 (0.001-0.1 mg/injection), quinelorane (0.001-0.1 mg/injection), and 7-OH-DPAT (0.01-0.32 mg/injection). In each rat, test drugs were evaluated by studying overlapping ranges of doses across three or four sessions. Each drug was initially substituted for cocaine in two consecutive sessions. If levels of responding maintained by any dose of the substitution drug were observed to be above levels observed during saline substitution, then the substitution drug was studied in the third consecutive session. However, if during the first two substitution sessions with a drug, no dose of the substitution drug maintained responding above levels observed during saline substitution, then multiple-component cocaine self-administration sessions were resumed, and stable cocaine-maintained behavior was re-established before a third session with the substitution drug. For subjects in which self-administration behavior occurred during the third drug substitution test only, the experiment was continued for an additional (fourth) drug substitution test.

Pretreatment Tests. The doses of cocaine available for self-administration during multiple-component sessions were determined by pseudorandom design such that sessions began with unit doses of 0, 0.01, 0.032, or 0.1 mg cocaine. Occasionally, 2-h single-component sessions of cocaine self-administration (see below) or multiple-component sessions of saline availability (0, 17, 56, 178 µl) were interspersed with multiple-component sessions of cocaine self-administration. Pretreatments were administered i.p. (1 ml/kg) immediately before the test session, and included the following drugs and doses: SKF 82958 (0.1, 0.32 mg/kg); SKF 77434 (0.32, 1.0 mg/kg); quinelorane (0.0032, 0.01 mg/kg); 7-OH-DPAT (0.32, 1.0 mg/kg). These doses were selected on the basis of preliminary results showing that lower doses were ineffective, whereas higher doses eliminated responding completely. For each rat, each dose of a pretreatment drug was tested in at least two sessions with overlapping cocaine dose ranges. Preliminary studies indicated that the duration of action of the behavioral effects of D_1-like and D_2-like agonists were at least 60 and 90 min, respectively. Therefore, in the case of D_2-like agonist pretreatments, a component of no cocaine availability (cocaine-associated cue light alone) and three doses of cocaine were tested in each session (88 min). In the case of D_1-like agonist pretreatments, either a component of no cocaine availability and two doses of cocaine, or three doses of cocaine were tested in each session (66 min). Each pretreatment test was preceded by a baseline session (i.e., a session during which the dose of cocaine that maintained maximal responding deviated by  0.5 log unit from the most recent session in which no pretreatment was administered).

Data Analysis. Data from multiple-component test sessions were expressed as the total number of drug injections obtained in each 20-min component. For each rat, multiple determinations were averaged. For comparison with data from substitution tests or pretreatment tests, baseline values were averaged from three recent multiple-component self-administration sessions that spanned the entire cocaine dose range, and in which no pretreatments were administered. Reinforcing effects of a drug during substitution test sessions were analyzed with a one-way within-subjects ANOVA using drug dose as the factor. Significant main effects were followed by pair-wise comparisons of responding maintained by saline, with those maintained by each drug dose. Pretreatment effects on the cocaine self-administration dose-effect function were analyzed with a two-way within-subjects ANOVA by using pretreatment dose and cocaine dose as factors. Significant main effects were followed by pair-wise comparisons of each cocaine dose under baseline conditions, with the same cocaine dose under each pretreatment condition. Pretreatment effects on the latency to the first response or the mean interinfusion interval during single-component self-administration sessions were analyzed by a one-way within-subjects ANOVA using pretreatment dose as the factor. Significant main effects were followed by pair-wise comparisons of each drug pretreatment dose with saline pretreatment. The criterion for significance was p < .05 for all analyses (main effects followed by Duncan pair-wise comparisons).

Drugs. Cocaine was provided by the National Institute on Drug Abuse, National Institutes of Health. All other drugs (SKF 82958, SKF 77434, R-6-Br-APB, SKF 81297, quinelorane, 7-OH-DPAT) were obtained from Research Biochemicals International (Natick, MA). All drugs were dissolved in physiological saline, and, with the exception of cocaine, the saline was warmed to assure dissolution. All doses refer to the weights of the respective salts (cocaine, HCl; quinelorane, 2 HCl; all other drugs, HBr).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Cocaine, D_1-like Agonists, or D_2-like Agonists Alone (Fig. 1). Cocaine dose-dependently and reliably maintained responding when the unit dose of cocaine was varied in 20-min components of multiple-component test sessions. Across a 100-fold range of unit doses (0.01-1.0 mg/injection), cocaine produced a stable, inverted U-shaped dose-effect function that is characteristic of drug self-administration data (Fig. 1, left panel). Intermediate doses of cocaine (0.032 and 0.1 mg/injection) maintained the highest levels of cocaine self-administration (10.5 ± 4.4 and 12.2 ± 1.2 injections/20-min component, respectively), which were significantly different from levels produced by saline substitution (p < .01). In comparison with intermediate doses of cocaine, lower (0.01 mg/injection) and higher doses (0.32 and 1.0 mg/injection) maintained lower levels of self-administration that were not statistically different from those observed during saline substitution. However, self-administration of high doses of cocaine differed qualitatively from responding maintained by saline, in that high doses of cocaine produced response patterns characterized by regular interinfusion intervals throughout the entire response period (for more detail, see Fig. 4, below).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Responding maintained by i.v. injections of cocaine, the D1-like agonists SKF 82958 and SKF 77434, or the D2-like agonists quinelorane and 7-OH-DPAT. Abscissae, unit dose in mg/injection (log scale). The point above "S" in the left panel shows responding maintained by saline. Ordinates, total number of injections earned during a 20-min period of drug availability. Values shown are group means (± S.E.), calculated from three to four overlapping dose-effect determinations in each of eight rats. Asterisks indicate significant differences from responding maintained by saline by Duncan's pair-wise comparison after a significant main effect of dose (*p < .05, **p < .01).

Effects of Pretreatment with D_1-like or D_2-like Agonists on Cocaine Self-Administration (Fig. 2). Pretreatment with the D_1-like agonists SKF 82958 and SKF 77434 dose-dependently decreased self-administration of cocaine doses that maintained maximal self-administration behavior, resulting in a downward shift in the cocaine dose-effect function (Fig. 2, left panels). The lower doses of SKF 82958 (0.1 mg/kg) or SKF 77434 (0.32 mg/kg) almost completely eliminated responding maintained by 0.032 mg/injection cocaine, whereas the next higher dose of each D_1-like agonist significantly decreased responding maintained by both 0.032 and 0.1 mg/injection cocaine. Higher doses of SKF 82958 (1.0 mg/kg) and SKF 77434 (3.2 mg/kg) produced more severe disruptions of behavior and were studied only in experiments with self-administration of a single dose of cocaine (see below).

View larger version (38K): [in this window] [in a new window]   Fig. 2.   Self-administration of cocaine alone or after pretreatments with the D1-like agonists SKF 82958 and SKF 77434, or the D2-like agonists quinelorane and 7-OH-DPAT. Abscissae, unit dose of cocaine in mg/injection (log scale). Points above "0" on the abscissae show responding maintained only by the cocaine-associated cue light. Ordinates, total number of injections earned during a 20-min period of drug availability. Values shown are group means (± S.E.) calculated from two overlapping dose-effect determinations in each of eight rats. Open circles represent cocaine alone (i.e., no pretreatment). Asterisks indicate significant differences from responding after no pretreatment by Duncan's pair-wise comparison after a significant main effect of dose (*p < .05, **p < .01).

Effects of Pretreatment with D_1-like or D_2-like Agonists on Patterns of Self-Administration Maintained by a Single Dose of Cocaine (Figs. 3 and 4). Both D_1-like and D_2-like agonists decreased self-administration of a dose of cocaine on the descending limb of the cocaine dose-effect function (0.32 mg/injection). When this dose of cocaine was available under control conditions (saline pretreatment), rats self-administered an average of 24.5 ± 0.9 injections in 2 h, and drug intake was distributed evenly throughout the session. Pretreatment with R-6-Br-APB, SKF 77434, SKF 82958, SKF 81297, or quinelorane dose-dependently decreased the number of cocaine injections that were self-administered (Fig. 3). The D_2-like agonist quinelorane was approximately 2 log units more potent in decreasing cocaine self-administration than were the D_1-like agonists, whereas potency did not differ significantly among the latter drugs (Table 1).

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Effects of pretreatments with saline, the D1-like agonists SKF 82958, SKF 77434, SKF 81297, and R-Br-APB, or the D2-like agonist quinelorane, on the self-administration of 0.32 mg/injection of cocaine. Abscissa, dose of pretreatment compound in mg/kg (log scale). The point above "S" shows data obtained after saline pretreatment. Ordinate, total number of cocaine injections delivered during a single-component 2-h session. Values shown are group means (± S.E.) calculated from single determinations in either eight (quinelorane, SKF 82958) or six rats. See Table 1 for description of A50 values.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Comparison of the effects of pretreatments with the D1-like agonist SKF 82958 or the D2-like agonist quinelorane on patterns of responding maintained by 0.32 mg/injection cocaine. Abscissae (top panels), dose of pretreatment drug in mg/kg (log scale). Points above "S" show data obtained after saline pretreatment. Ordinate (top left panel), latency from the start of the session to the first lever press in minutes. Ordinate (top right panel), average interval that separated cocaine infusions in minutes. Values shown are group means (± S.E.), calculated from single determinations in each of the same eight rats. Asterisks indicate significant differences from responding after saline pretreatment by Duncan's pair-wise comparison after a significant main effect of dose (*p < .05, **p < .01). The bottom panel shows event records from self-administration sessions of a representative rat to illustrate the qualitative differences in the effects of the D1-like and D2-like agonist pretreatments on cocaine self-administration. Horizontal lines denote time, and each downward tick represents the delivery of one cocaine injection. The type of pretreatment is printed above each event record.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Self-Administration of D_2-like Agonists. In the present study, direct dopamine agonists with high affinity for D_2-like receptors maintained i.v. drug self-administration in cocaine-trained subjects in a manner consistent with previous findings in primate and rodent species. For example, a broad range of dopamine D_2-like agonists differing in potency and selectivity for D_2-like receptors have been shown to reliably maintain self-administration behavior in cocaine-experienced monkeys (Woolverton et al., 1984; Grech et al., 1996; Nader and Mach, 1996). Similarly, in rats with cocaine self-administration experience, the selective D_2-like agonists quinpirole and 7-OH-DPAT have been shown to maintain responding with performance described by inverted U-shaped dose-effect functions characteristic for drug self-administration behavior (Caine and Koob, 1993, 1995). The present results extend the latter observations to include the selective D_2-like agonist quinelorane.

Self-Administration of D_1-like Agonists. Although the reinforcing effects of D_2-like agonists in the present experiments are consistent with previous findings, results with D_1-like agonists in this study differ markedly from those in earlier reports. For example, Self and Stein (1992) previously reported that both the D_1-like partial agonist SKF 77434 and the full agonist SKF 82958 maintained i.v. self-administration in rats. Moreover, several D_1-like agonists (SKF 82958, R-6-Br-APB, SKF 81297) were found to maintain responding under selected conditions in monkeys (Weed and Woolverton, 1995; Grech et al., 1996). However, in the present experiments, SKF 82958 and SKF 77434 clearly did not maintain i.v. self-administration in cocaine-experienced rats. The reasons for these qualitatively different findings are not immediately obvious, but probably do not include dose selection, inasmuch as the doses tested here included all of the doses previously reported to be effective in rats (Self and Stein, 1992).

Modification of Cocaine Self-Administration by Pretreatment with D_2-like Agonists. Consistent with previous findings, pretreatment with D_2-like agonists shifted the dose-effect function for cocaine self-administration to the left (Caine and Koob, 1995; Caine et al., 1997, 1999). The effects of D_2-like agonists on cocaine self-administration depended on the pretreatment dose as well as the unit dose of cocaine. A low dose of quinelorane slightly suppressed responding maintained by a low cocaine-dose, an effect that may be due to a preferential action at D_2-like autoreceptors to reduce postsynaptic dopaminergic tone (Eilam and Szechtman, 1989; Depoortere et al., 1996). In contrast, higher doses of quinelorane or 7-OH-DPAT, which presumably increase postsynaptic D_2-like dopaminergic receptor activation, increased responding maintained by low cocaine doses and dose-dependently increased interinjection intervals during self-administration of a higher dose of cocaine (0.32 mg/injection). This profile of effects is characteristic for D_2-like agonists in rats (Caine and Koob, 1993, 1995; Parsons et al., 1996; Caine et al., 1997), and similar effects of quinelorane and 7-OH-DPAT have recently been observed in rhesus monkeys (Caine et al., 1999). Collectively, these findings suggests that pretreatment with D_2-like agonists may augment the reinforcing effects of cocaine.

Modification of Cocaine Self-Administration by Pretreatment with D_1-like Agonists. In contrast to the effects of D_2-like agonists on cocaine self-administration, the D_1-like agonists SKF 77434 and SKF 82958 decreased responding maintained by a broad range of doses of cocaine, and produced downward shifts in the cocaine dose-effect function. Notably, the D_1-like agonists SKF 77434 and SKF 82958 have been reported to differ in agonist efficacy, yet had similar effects on cocaine self-administration in the present study. These findings are consistent with previous results indicating that efficacy-based differences in behavioral effects of D_1-like agonists may be a less relevant factor for the modification of cocaine's effects in rats than in monkeys. In monkeys, for example, low-efficacy D_1-like agonists including SKF 38393 and SKF 75670, such as the D_1-like receptor blocker SCH 39166, have been shown to surmountably antagonize the behavioral effects of self-administered cocaine, resulting in rightward shifts in dose-effect functions for cocaine self-administration (Katz and Witkin, 1992; Bergman and Rosenzweig-Lipson, 1992). In rats, however, the low-efficacy agonist SKF 77434 produces behavioral effects that are qualitatively similar to those produced by high-efficacy agonists including SKF 82958 (see Waddington et al., 1995, for review). In view of the similar agonist effects of SKF 77434 and SKF 82958 in rats, the present results suggest that reductions in cocaine self-administration produced by D_1-like agonists, in contrast to those produced by D₁ receptor blockers, may not be surmountable by increased doses of cocaine (Bergman et al., 1990; Caine and Koob, 1994).

Implications for Neurobiological Mechanisms Underlying Cocaine's Reinforcing Effects. The different effects of D_1-like and D_2-like agonists in assays of cocaine self-administration and cocaine discrimination may support the results of previous studies aimed at identifying the neuroanatomical and molecular bases of the abuse-related effects of cocaine. For example, D_2-like agonists, but not D_1-like agonists, regulate gene expression within striatopallidal neurons (Gerfen et al., 1990), and a critical role for the ventral pallidum in cocaine self-administration has been proposed (Hubner and Koob, 1990; Robledo and Koob, 1993). In addition, D_2-like agonists inhibit adenylyl cyclase through activation of inhibitory G proteins, whereas D_1-like receptors are positively linked to adenylyl cyclase through stimulatory G proteins. Inhibitory G proteins have been proposed as a common signal-transduction mechanism for the reinforcing effects of abused drugs, and inactivation of inhibitory G proteins in the nucleus accumbens attenuates some behavioral effects of self-administered cocaine (Self et al., 1994; Self and Nestler, 1995). The present results and those from a companion study of cocaine discrimination (Caine et al., unpublished observations) indicate that D_2-like agonists produce prominent reinforcing effects and cocaine-like discriminative stimulus effects in cocaine-trained animals. Consequently, these findings support previous results and suggest that D₂ receptor-expressing pathways (e.g., striatopallidal neurons) and D₂ receptor-linked molecular events (e.g., inhibitory G proteins) may be especially important in mediating the abuse-related effects of cocaine.

Summary. In the present study, D_1-like and D_2-like dopamine agonists produced different behavioral effects in rats trained under a FR schedule of cocaine-maintained responding. When substituted for cocaine, D_2-like agonists dose-dependently and reliably maintained i.v. self-administration, whereas D_1-like agonists failed to consistently maintain responding. When administered as pretreatments, D_2-like agonists increased self-administration of low doses of cocaine and also appeared to enhance the behavioral effects of higher doses of cocaine. In contrast, D_1-like agonists only decreased self-administration across a broad range of doses of cocaine. The present results are consistent with results of a companion study in rats trained to discriminate cocaine from saline (Caine et al., unpublished observations), in which D_2-like agonists readily substituted for and enhanced the effects of cocaine under a wider range of conditions than did D_1-like agonists. Collectively, these findings suggest that the behavioral effects of cocaine overlap with those of D_2-like agonists to a greater extent than with those of D_1-like agonists.