Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The prevalence of illicit cocaine abuse has created significant public health problems in the United States over the last two decades (Chilcoat and Johanson, 1998). There have been continuing efforts to better understand the pharmacological mechanisms that underlie this abuse (Leshner, 1997), which may contribute to the identification of new leads for the discovery of medical treatments for cocaine abuse (Johnson and Vocci, 1993). Because cocaine indirectly stimulates dopaminergic receptors by blocking reuptake of dopamine (Heikkila et al., 1979), and because dopaminergic mechanisms have been implicated as mediating cocaine abuse (Wise, 1984; Kuhar et al., 1991), studies have focused on the respective roles of subtypes of dopamine receptors in mediating the pharmacological effects of cocaine.

Several studies have examined the antagonism of cocaine by D1-type dopamine receptor antagonists. For example, Woolverton and Virus (1989) found that the dopamine D1 receptor antagonist SCH 23390 decreased rates of responding maintained by cocaine in rhesus monkeys. However, these decreases in cocaine-maintained behavior were obtained only at doses of the antagonist that also decreased rates of responding maintained by food reinforcement. The similarity of the potency of the antagonist in decreasing rates of responding maintained by either reinforcer suggests that this effect was not a specific action on the reinforcing effects of cocaine. Bergman et al. (1990) found that the D1 antagonist SCH 39166 as well as the D2 antagonist eticlopride shifted the dose-effect curve for cocaine to the right in squirrel monkeys (Saimiri sciureus) whose responding was maintained by cocaine. Although there were no comparisons to the effects of these antagonists on responding maintained by another reinforcer, the nature of the interaction, an antagonist-dose-dependent shift to the right in the cocaine dose-effect curve, suggested pharmacological specificity. Caine and Koob (1994) found that the dopamine D1 antagonists SCH 23390 and SCH 39166 but not A-69024 decreased rates of responding maintained by cocaine in rats at certain doses that did not alter rates of responding maintained by food reinforcement. Interestingly, the D1 antagonist A-69024 decreased rates of responding maintained by food reinforcement at doses that did not affect behavior maintained by cocaine. The specificity of the effect of the D1 antagonists in this study needs to be carefully considered because of the difference in effects obtained with A-69024.

The influential role of dopamine D1 receptors in the effects of cocaine was dramatically illustrated in studies of genetically altered mice in which the D1 receptor was targeted for removal (Xu et al., 1994). In these D1 knockout mice, cocaine was ineffective as a stimulant of locomotor activity, whereas the wild-type control mice exhibited a dose-related increase in locomotor activity. Furthermore, electrophysiological studies of cells in the nucleus accumbens showed a reduction in the inhibitory effects of cocaine on action potentials in knockout compared with wild-type subjects. These results clearly indicate a significant role of D1 dopamine receptors in the locomotor stimulant effects of cocaine and in the central effects within the nucleus accumbens presumed to mediate those effects. Nonetheless, the respective role of D1 dopamine receptors in other behavioral effects of cocaine is yet to be fully determined.

Although incomplete, there is more information on the effects of pure D1 antagonists than partial D1 agonists on the effects of cocaine (for review, see Winger, 1998). In one study, the effects of the partial agonist (±)-SKF 38393 were examined on cocaine- and food-maintained responding in squirrel monkeys. Decreases in responding maintained by cocaine under a 30-response fixed-ratio (FR) schedule were obtained at doses of (±)-SKF 38393 that had minimal effects on responding maintained under the same schedule by food presentation (Katz and Witkin, 1992). In addition, at an appropriate dose, (±)-SKF 38393 shifted the cocaine dose-effect curve ~3-fold to the right. Similarly, Spealman et al. (1997) showed that the R-enantiomer of SKF 38393 shifted the dose-effect function for the discriminative-stimulus effects of cocaine ~3-fold to the right, as did the partial agonist SKF 75670 and the antagonist SCH 39166.

In preliminary studies of the benzazepine antagonist SCH 39166, we found a maximal 3-fold shift to the right in the discriminative-stimulus effects of cocaine, with higher doses of the antagonist ineffective in producing greater shifts to the right in the dose-effect curve. This small shift to the right in the cocaine dose-effect curve, coupled with a clearly defined limit to the degree to which the discriminative stimulus effect could be shifted, stood in marked contrast to the types of interactions obtained with other compounds, such as opioid agonists and antagonists (Bertalmio and Woods 1987), and in the dramatic elimination of locomotor stimulant effects of cocaine in D1 knockout mice (Xu et al., 1994). Thus, we initiated this study to quantitatively characterize the interaction between the indirect agonist, cocaine, and dopamine D1 receptor blockade. The effects of the antagonists on the cocaine dose-effect curves for discriminative-stimulus effects and effects on locomotor activity were quantified in terms of the degree of rightward shift, and the change in maximal efficacy of cocaine. A greater rightward shift or decrease in efficacy of cocaine is indicative of a greater degree of dopamine D1 receptor involvement in mediating that effect of cocaine. To ensure that the findings were not idiosyncratic to a particular antagonist, several D1 ligands were used, including the benzazepine antagonists SCH 23390 and SCH 39166, and the structurally distinct tetrahydroisoquinoline A-69024. In addition the partial agonists SKF 77434 and SKF 38393 were studied; these compounds each produce ~50% of maximal stimulation of cyclase (O'Boyle et al., 1989; Andersen and Jansen, 1990).

These interactions of the D1 ligands with cocaine were assessed on several behavioral effects. Locomotor activity in rodents was used as an indication of psychomotor stimulant actions of cocaine and because of the substantive effects of D1 receptor elimination on this behavior in knockout subjects. The discriminative-stimulus effects of cocaine were used as an indication of the subjective effects of cocaine, which likely play an important role in its abuse. The effects on response rates during the drug-discrimination procedure also were assessed to provide an assessment of the general behavior-disrupting effects of cocaine, with less specificity for cocaine abuse. Our results indicated differences in the degree of involvement of D1-like dopamine receptors in these behavioral effects of cocaine.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Stimulation of Locomotor Activity. Mice (male Swiss-Webster) were obtained and allowed to habituate to the animal facility for at least 1 week before use with a 12-h light/dark cycle (lights on 7:00 AM). Subjects weighed 30 to 35 g and were ~6 to 7 weeks of age at the time of the study. Locomotor activity was studied in Digiscan activity monitors (Omnitech Electronics Inc., Columbus, OH), which consisted of 40-cm³ clear acrylic chambers equipped with photoelectric detectors placed 2.56 cm apart along the walls of the chamber. One activity count was registered each instance in which the subject crossed a single photo beam. Mice were injected and placed immediately in the chamber with one subject per chamber. Activity was assessed for a 1-h session. Each dose or dose combination was studied in eight subjects, and no subject was used more than once.

Cocaine Discriminative-Stimulus Effects. Adult male squirrel monkeys weighing between 750 and 1000 g served as subjects and were housed under a 12-h light/dark cycle (lights on 7:00 AM). They were fed a daily ration of food (Purina Monkey Chow supplemented with Teklad Monkey Diet) (Ralston Purina, St. Louis, MO; Teklad Premier Laboratory Diets, Madison, WI) at least 30 min after testing that maintained their body weights at a relatively constant level throughout the course of the study. Water was always available in the individual home cages. Five or six subjects were studied for each drug or drug combination. With cocaine alone, however, eight subjects were studied because some subjects were replaced during the study, and the newly introduced subjects all received cocaine. When subjects were replaced, it was due to health reasons not related to the study. All monkeys had been studied previously under experimental procedures similar to those described below and had received injections of various drugs; however, drugs had not been administered for at least 1 week before the initiation of these studies.

Drugs and Injection Procedures. ()-Cocaine HCl (Sigma Chemical Co., St. Louis, MO) and the dopamine D1 receptor antagonists (+)-SCH 23390 HCl (Research Biochemicals, Inc., Natick, MA), SCH 39166 (Schering Corp., Bloomfield, NJ), and A-69024 (Abbott Laboratories, Abbott Park, IL, and National Institute on Drug Abuse, Rockville, MD) were dissolved in distilled water or water with mild acidification (0.16% tartaric acid) and heat. The partial D1 agonists (±)-SKF 38393 HCl (RBI), (+)-SKF 38393 HCl (Research Biochemicals, Inc.), and (±)-SKF 77434 HCl (Research Biochemicals, Inc.) were dissolved in distilled water. Doses for monkeys were injected i.m. (calf or thigh) in a volume of 1.0 ml/kg body weight or less. Doses for mice were injected i.p. in a volume of 1.0 ml/100 g b.wt. Control injections were similar volumes of respective vehicle solutions. For mice, two injections were given immediately before the subjects were placed in the activity monitors. For monkeys, drugs were injected immediately before experimental sessions (i.e., 5 min before testing was initiated). Both (±)- and (+)-SKF 38393 also were administered 30 min before testing was initiated. When two injections were administered to monkeys, each was injected in a different leg. Studies of the drugs alone in monkeys were conducted with either as single injections, as cocaine injections along with antagonist vehicle control injection, or as antagonist with cocaine-vehicle control injections. The studies with vehicle control injections did not produce effects different from those obtained with the single injections of the drugs alone, and therefore the data were combined for presentation. All doses are expressed as milligrams of the above drug formulations per kilogram of the body weight of the subject.

Measurement of Effects and Statistical Analysis. Locomotor activity in mice was assessed for a 1-h period with counts collected during each successive 10-min epoch; counts during the first and last three epochs were cumulated for separate analyses of the first and last 30 min of the 1-h test session. Counts averaged across subjects were used to construct dose-effect curves that were further analyzed (see below). In general, the interactions between cocaine and the D1 dopaminergic ligands obtained in the first and second 30-min periods were comparable and, although both are shown in the graphs, only the analysis of data from the first 30 min is described in detail.

D1 Receptor Binding. Brains from male Sprague-Dawley rats weighing 200 to 225 g (Taconic Farms Inc., Germantown, NY) were removed and striatum was dissected and rapidly frozen. Membranes were prepared by homogenizing tissues in 20 volumes (w/v) of 50 mM Tris, pH 7.4 at 25°C (buffer A), with a Brinkmann Polytron (Brinkmann Instruments, Inc., Westbury, NY) (setting 6 for 20 s), and centrifuged at 50,000g for 10 min at 4°C. The resulting pellet was resuspended in buffer and centrifuged again at the same parameters. The supernatant was discarded, and the final pellet was resuspended in cold Tris-HCl containing 120 mM NaCl, 5 mM CaCl₂, and 1 mM MgCl₂, pH 7.4 at 25°C (buffer B) to a concentration of 3.3 mg/ml (original wet weight). Ligand-binding experiments were conducted in assay tubes containing 1.0 ml of buffer B for 30 min at 37°C. Each tube contained 0.3 nM [³H]SCH 23390 (New England Nuclear, Boston, MA), 1 µM mianserin (Research Biochemicals Inc.), and 2.5 mg of striatal tissue. Nonspecific binding was determined with 1 µM fluphenazine. Incubations were terminated by rapid filtration through Whatman GF/B filters (Whatman Specialty Products, Inc., Fairfield, NJ) with a Brandel cell harvester (Brandel Laboratories, Gaithersburg, MD). The filters were washed twice with 5 ml of cold buffer B and transferred to scintillation vials. Beckman Ready Safe (3.0 ml) was added, and the vials were counted the next day with a Beckman 6000 liquid scintillation counter (Beckman Instruments, Fullerton, CA). Triplicate samples were used in each assay and assays were replicated at least three times. Data were analyzed with GraphPad Prism software (San Diego, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Direct Effects of Cocaine and Dopamine D1 Receptor Ligands. Cocaine produced a dose-related stimulation of locomotor activity in Swiss-Webster mice. Maximal stimulation to 700 counts/min was produced by a dose of 59 µmol/kg, with both higher and lower doses producing less stimulation (Fig. 1). The other compounds produced only dose-related decreases in locomotor activity across the range of active doses in the first 30 min following injection (Fig. 1, top panel). The dopamine D1 antagonist SCH 23390 was ~5-fold more potent than SCH 39166 in producing decreases in locomotor activity (Fig. 1, top panel, squares and upward pointing triangles, respectively), followed by A-69024 (open circles), exhibiting an ~34-fold lower potency than SCH 23390. The other compounds were appreciably less potent than these three compounds (Table 1, column A). Similar results were obtained in the second 30 min after injection; however, the partial agonist SKF 77434 exhibited a significant stimulation in locomotor activity at the 30-µmol/kg (10-mg/kg) dose.

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Effects of cocaine, SCH 23390, SCH 39166, A-69024, SKF 77434, (±)-SKF 38393, and (+)-SKF 38393 on locomotor activity in Swiss-Webster mice. Top panel, locomotor activity counts during the first 30 min of a 1-h observation period as a function of dose administered. Bottom panel, locomotor activity counts during the second 30 min of a 1-h observation period as a function of dose administered. Ordinates, average locomotor activity counts (in counts/min); abscissae, dose (mg/kg) of drug, log scale. Each point is the average of eight subjects, with vertical bars representing ±1 S.E. Points above C represent the effects of vehicle injections.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Effects of cocaine, SCH 23390, SCH 39166, A-69024, SKF 77434, (±)-SKF 38393, and (+)-SKF 38393 in squirrel monkeys discriminating 0.3 mg/kg cocaine from saline injections under a FR 30-response schedule of food presentation. Top panel, effects on the distribution of responses on the two response keys expressed as a percentage of responses on the cocaine-appropriate key. Each point is the average of at least three of six or eight (cocaine) subjects, with data excluded for subjects failing to respond at a rate of 0.02 response/s. If more than half of the subjects tested at a given dose failed to meet the response-rate requirement, no mean value was calculated for percentage of drug-appropriate responding at that dose. Ordinates, average percentage of responses occurring on the cocaine-appropriate key; abscissae, dose (mg/kg) of drug, log scale. Bottom panel, effects on rates of responding under the FR schedule. Each point is the average of all of the subjects studied. Ordinates, average response rates expressed as a percentage of control response rates; abscissae, dose (mg/kg) of drug, log scale. Control rates of responding (from saline training sessions), under the FR schedule averaged 3.62 ± 0.80 (mean ± S.E.) responses/s. The average number of these sessions for the various subjects tested was six. Rates of responding during saline test sessions averaged 3.40 ± 0.81 responses/s, with 0.51 ± 0.39% of the responses on the cocaine-appropriate response key.

Interactions of Cocaine with Dopamine D1 Receptor Antagonists. At the lowest dose studied (0.01 mg/kg), SCH 23390 only marginally changed the potency of cocaine for stimulation of locomotor activity (Fig. 3, left panel, triangles). The ED₅₀ value for cocaine in the presence of SCH 23390 was not appreciably different from that for cocaine alone (Table 2, column A). However, the relative potency analysis indicated an ~1.5-fold shift to the right in the dose effect curve (Table 2, column B), which although small, was significant (95% CL of that value excluded the value 1.0). The maximal stimulation produced by cocaine was decreased by this dose of SCH 23390, as reflected by a significant effect of preparations in the relative potency analysis. This dose of SCH 23390 was inactive when administered alone (Fig. 3, left panel). Higher doses of SCH 23390 (0.03 mg/kg, 0.1 mg/kg) produced further dose-related decreases in the maximal effect of cocaine; these doses, respectively, produced marginal or substantial decreases in activity when administered alone (Fig. 3). Only the highest dose of SCH 23390 appreciably altered the ED₅₀ value for cocaine relative to cocaine alone (Table 2, column A), although this and the intermediate dose significantly decreased the potency of cocaine as indicated by the relative potency analysis (Table 2, column B).

View larger version (41K): [in this window] [in a new window]   Fig. 3.   Antagonism of the effects of cocaine on locomotor activity in Swiss-Webster mice by SCH 23390, SCH 39166, and A-69024. Top panel, locomotor activity counts during the first 30 min of a 1-h observation period as a function of cocaine dose. Bottom panel, locomotor activity counts during the second 30 min of a 1-h observation period as a function of cocaine dose. Ordinates, average locomotor activity counts (counts/min); abscissae, dose (mg/kg) of drug, log scale. Each point is the average of eight subjects, with vertical bars representing ±1 S.E. Points above C represent effects of the antagonists with vehicle injections (, , , , ), or two vehicle injections ().

View larger version (42K): [in this window] [in a new window]   Fig. 4.   Antagonism of the discriminative effects of cocaine by SCH 23390, SCH 39166, and A-69024 in squirrel monkeys trained to discriminate 0.3 mg/kg cocaine from saline injections. Top panel, effects on the distribution of responses on the cocaine-appropriate key. Each point for the combined effects of the drugs is the average of at least three of the six subjects studied, with data excluded for subjects failing to respond at a rate of 0.02 response/s. If three or more of the subjects tested at a given dose failed to meet the response-rate requirement, no mean value was calculated for percentage of drug-appropriate responding at that dose. Ordinates, average percentage of responses occurring on the cocaine-appropriate key; abscissae, dose (mg/kg) of drug, log scale. Bottom panel, effects on rates of responding under the FR schedule. Each point for the combined effects of the drugs is the average of all of the subjects studied. Ordinates, average response rates expressed as a percentage of control response rates; abscissae, dose (mg/kg) of drug, log scale. The average number of control saline training sessions for the various subjects tested was six. All other details are as in Fig. 2.

Interactions of Cocaine with Dopamine D1 Receptor Partial Agonists. Each of the doses of SKF 77434 examined decreased the maximal effects of cocaine (Fig. 5, left panel). The two lowest doses studied (3.0 and 10.0 mg/kg) had comparable effects, and only at these doses were some significant stimulant effects of cocaine preserved. There was a trend toward an increase in cocaine ED₅₀ values at these doses of SKF 77434 (Table 3, column A), and the relative potency analysis indicated that both 3.0 and 10.0 mg/kg SKF 77434 significantly shifted the cocaine dose-effect curve to the right (Table 3, column B). These changes in the effects of cocaine were obtained with doses that were inactive when administered alone (Fig. 5, left panel, disconnected points at left).

View larger version (39K): [in this window] [in a new window]   Fig. 5.   Interactions among D1 dopamine partial agonists and cocaine on locomotor activity in Swiss-Webster mice. Top panel, locomotor activity counts during the first 30 min of a 1-h observation period as a function of cocaine dose. Bottom panel, locomotor activity counts during the second 30 min of a 1-h observation period as a function of cocaine dose. Ordinates, average locomotor activity counts (counts/min); abscissae, dose (mg/kg) of drug, log scale. Each point is the average of eight subjects, with vertical bars representing ±1 S.E. Points above C represent effects of the antagonists with vehicle injections (, , , ,), or two vehicle injections ().

View larger version (41K): [in this window] [in a new window]   Fig. 6.   Interactions among D1 dopamine partial agonists and cocaine in squirrel monkeys trained to discriminate 0.3 mg/kg cocaine from saline injections. Top panel, effects on the distribution of responses on the cocaine-appropriate key. Each point is the average of at least three of the five or six subjects studied, with data excluded for subjects failing to respond at a rate of 0.02 response/s. If three or more of the subjects tested at a given dose failed to meet the response rate requirement, no mean value was calculated for percentage of drug-appropriate responding at that dose. Ordinates, average percentage of responses occurring on the cocaine-appropriate key; abscissae, dose (mg/kg) of drug, log scale. Bottom panel, effects on rates of responding under the FR schedule. Each point for the combined effects of the drugs is the average of all of the subjects studied. Ordinates, average response rates expressed as a percentage of control response rates; abscissae, dose (mg/kg) of drug, log scale. The average number of control saline training sessions for the various subjects tested was five. All other details are as in Fig. 2.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, the effects of several compounds acting at dopamine D1 receptors were examined alone and in combination with cocaine. The effects examined were the cocainelike discriminative-stimulus effects and the concurrently assessed effects on rates of lever pressing, as well as the effects on locomotor activity. Locomotor activity in rodents is used as an indication of psychomotor stimulant actions, and the discriminative-stimulus effect of cocaine is used as an indication of the subjective effects of these drugs, which likely play an important role in cocaine abuse. The effects on response rates during the drug-discrimination procedure provide an index of the disruption in behavior produced by cocaine, which is not an effect that is specifically related to cocaine abuse.

Cocaine and the D1 partial agonist SKF 77434 produced increases in locomotor activity, but only cocaine produced substantial stimulation. Results similar to those of our study have been reported by others. For example, stimulation of locomotor activity in rats was reported for SKF 77434 but not for (±)-SKF 38393 (Murray and Waddington, 1989; Meyer and Shults, 1993; although see Tirelli and Terry, 1993). Similarly, decreases in locomotor activity with the remaining compounds, or schedule-controlled behavior with all of the dopaminergic D1 agents, have been reported (Bergman et al., 1991, 1995, 1996; Criswell et al., 1992; Katz et al., 1995).

The ED₅₀ values for producing each of the various effects examined in the present study and their apparent affinities computed from the antagonism experiments were related to their binding affinities determined from displacement of [³H]SCH 23390. The benzazepine antagonists SCH 23390 and SCH 39166 were most potent in decreasing locomotor activity in mice and in decreasing response rates in monkeys discriminating cocaine injections. These two compounds were equally potent in decreasing response rates and had the highest affinity in the present study of displacement of [³H]SCH 23390. Our affinity values are consistent with those previously reported in rats (Chipkin et al., 1988) and primates (Madras et al., 1988; Bergman et al., 1991), and with previous assessments of potencies in vivo (Bergman et al., 1991, 1995; Criswell et al., 1992). The relatively lower potency of A-69024 in decreasing locomotor behavior or response rates is consistent with its lower affinity (present study; Kerkman et al., 1989; Kassiou et al., 1995). The ~30-fold difference in potency between SKF 77434 and SCH 23390 in decreasing response rates is consistent with its ~12-fold lower affinity for D1 dopamine receptors (Table 4; Andersen et al., 1985; Andersen and Jansen, 1990). However, in decreasing locomotor activity, the potency of SKF 77434 was disproportionately lower than SCH 23390.

The partial agonist SKF 77434 stimulated locomotor activity at lower doses than those that decreased locomotor activity (albeit during the second 30 min after injection). If its potency as a locomotor stimulant is considered along with the potencies of the other compounds, the agreement among in vivo potencies and affinities for D1 receptor binding among all of the compounds is relatively high. This agreement is consistent with an interpretation that the stimulation of activity produced by the partial agonist SKF 77434 is mediated by agonist actions at D1 dopamine receptors, whereas the decreases in locomotor activity produced by the remaining ligands is the result of antagonist effects at D1 dopamine receptors. The differences in activity of SKF 77434 and the other D1 ligands might be explained on the basis of the greater intrinsic efficacy of SKF 77434 compared with the antagonists; however, remaining without explanation is the lack of stimulant effects of (±)- or (+)-SKF 38393. The racemic form of this compound has been shown to maximally stimulate locomotor activity in C57BL mice between 50 and 90 min after injection (Tirelli and Terry, 1993). Thus, it is possible that the stimulant effects of either form of SKF 38393 would have been obtained with our Swiss-Webster mice, although at an even later time point. Alternatively, SKF 38393 may have decreased locomotor activity by a mechanism other than D1 dopamine receptor activation (cf. Terry and Katz, 1992).

In monkeys trained to discriminate cocaine from saline injections, cocaine produced dose-related increases in responding on the cocaine-appropriate lever, with virtually exclusive responding on that lever at the training dose. None of the D1 dopaminergic agents fully substituted for cocaine, although with SCH 39166, SCH 23390, and SKF 77434 there was a substitution that was significantly greater than saline levels. Similar results have been reported in rodents and primates (Kleven et al., 1990; Callahan et al., 1991; Witkin et al., 1991; Terry et al., 1994; Spealman et al., 1997). The partial substitution for cocaine by the compounds with some agonist activity has been interpreted as consistent with the notion that activation of D1 dopamine receptors contributes to, but does not fully reproduce, the discriminative effects of cocaine.

The partial substitution of D1 dopamine receptor antagonists is not currently understood. Several studies have found paradoxical agonist actions of dopamine D1 receptor antagonists. For example, Starr and Starr (1986) first reported substantial grooming behavior induced by relatively low doses of SCH 23390, below those that antagonized the effects of SKF 38393 (see also Gorell et al., 1986; Wachtel et al., 1992). Similarly, Wachtel and White (1995) found a D1 agonist-like effect of both SCH 23390 and SCH 39166 on glutamate-stimulated firing in neurons of the nucleus accumbens of rats. These effects also were seen at low doses, and furthermore, appeared to depend on whether a D2 agonist treatment had occurred in the past. Wachtel and White (1995) ruled out artifacts of iontophoretic drug administration and local anesthetic effects, in part by examining the enantiomer of SCH 23390. The actions of SCH 23390 on 5-hydroxytryptamine (5-HT) systems (Bischoff et al., 1986) appeared unlikely to account for these paradoxical agonistlike effects because the effect also was obtained with the more D1-selective antagonist SCH 39166. Interestingly, the agonist actions in the previous studies were generally obtained at low doses relative to those having antagonist effects, whereas in our study, the substitution for cocaine of the D1 antagonists occurred at doses marginally higher than those having antagonist effects. The previous results showing antagonism at doses lower than those at which agonist effects are obtained indicates that the "paradoxical agonist" effects of these antagonists are not likely due to some, albeit limited, intrinsic efficacy of the antagonists. Our results showing that substitution for cocaine does not occur with the partial agonists (±)- and (+)-SKF 38393, and that greater substitution was obtained with the antagonist SCH 39166, further support that these paradoxical agonist actions are not due to limited intrinsic efficacy of these compounds at D1 dopamine receptors.

The D1 antagonists A-69024, SCH 23390, and SCH 39166 and the partial agonists SKF 77434, (±)-SKF 38393, and (+)-SKF 38393 each produced a rightward shift in the cocaine dose-effect curve for stimulation of locomotor activity (ascending limb of the dose-effect curve). The antagonism of the effects of cocaine on locomotor activity by the D1 antagonist SCH 23390 has been reported (Cabib et al., 1991); however, characterization of the cocaine dose-effect curve as a result of that antagonism has not been previously reported. One interesting result of our study is that the antagonists not only shifted the cocaine dose-effect curve to the right but also decreased maximal efficacy for stimulation of locomotor activity, such that the antagonism was not fully surmountable.

The shift to the right in the dose-effect curve concomitant with a decrease in maximal effect is consistent with what is expected with the antagonism of an indirectly acting agonist by a competitive antagonist (Kenakin, 1993). In addition, the characteristics of the antagonism reported herein are consistent with a differential antagonism of the ascending and descending portions of the cocaine dose-effect curve. If these two limbs of the dose-effect curve represent differently mediated and offsetting effects, then a preferential shift to the right in the ascending limb of the curve would result in a decrease in maximal effect. Finally, the antagonist-induced decreases in maximal effects of cocaine could represent a functional antagonism resulting from the opposing effects of cocaine (stimulation of activity) and the D1 ligands (decreases in activity) observed when the drugs are administered alone. However, the decreases in maximal effect of cocaine often occurred at doses of the antagonists and partial agonists that alone had no substantial effects on locomotor activity, as well as at doses that alone decreased activity. Therefore, the decrease in maximal cocaine effect in the presence of the D1 ligands is at least not uniformly due to a functional or physiological antagonism. Thus, the shift to the right in the cocaine dose-effect curve concomitant with a decrease in maximal effect in the studies of locomotor activity appears to be either characteristic of antagonism of the effects of an indirect-acting agonist, or differential antagonism of the differently mediated limbs of the bitonic dose-effect curve.

One major difference in the antagonism of the behavioral effects examined herein was that in the cocaine discrimination, there were generally no changes in the efficacy of cocaine in the presence of increasing doses of the D1 antagonists. At all of the doses of the D1 ligands, the antagonism was surmounted by appropriate doses of cocaine, a result that contrasts with that obtained for the effects of cocaine on locomotor activity. If the antagonism of the effects of cocaine on locomotor activity is characteristic of indirect agonist-antagonist interactions (as described above), the question remains as to why the antagonism of the discriminative-stimulus effects was fully surmountable. These differences in antagonism may be related to the differences in species used to assess discriminative effects and locomotor stimulation (however, see below), or to differences in exposure to cocaine as the discrimination procedure necessitates repeated administration of the training drug. Alternatively, it is possible that there are differences in dopamine receptor occupancy necessary for the two effects. Kenakin (1993) notes that even with interactions between indirect-acting agonists and competitive antagonists, when small amounts of pharmacological stimulus are necessary for an effect, the amount of parallel shift in dose-effect curves will be extended before there are decreases in maximal effect. In the opioid system for example, there is compelling evidence that discriminative-stimulus effects are "sensitive" based on their appearance at doses lower than those producing other opioid effects (Woods et al., 1988). It follows from these observations that the discriminative effects of opioids would require a small receptor occupancy for the expression of a full effect compared with other effects. Our results suggest a similar relationship for the discriminative-stimulus effects of cocaine.

Several studies have demonstrated an antagonism of the discriminative-stimulus effects of cocaine by D1 dopamine antagonists. For example, Kleven et al. (1990) showed an ~3-fold shift to the right in the discriminative effects of cocaine in rhesus monkeys by SCH 23390. Spealman et al. (1991) showed similar degrees of antagonism by SCH 39166. In our study, the antagonists dose-dependently shifted the cocaine dose-effect curve for effects on locomotor activity, often reaching doses that completely prevented stimulation at any dose of cocaine studied. In contrast, in the cocaine discrimination studies the antagonists exhibited a maximal degree to which the cocaine dose-effect curve could be shifted (typically 3-fold), with higher doses ineffective in producing greater shifts to the right in the discriminative-stimulus effects of cocaine. Interestingly, among the published studies of antagonism of cocaine by D1 antagonists, ~3-fold shifts to the right in the dose-effect curve are generally the rule (Kleven et al., 1990; Spealman, 1990; Bergman et al., 1990; Spealman et al., 1991; Vanover et al., 1991; Spealman et al., 1997).

A relatively greater antagonism of the discriminative-stimulus effects compared with the response-rate-decreasing effects of cocaine could explain the ~3-fold limit to the antagonism of the discriminative-stimulus effects if pronounced decreases in response rates precluded an assessment of further antagonism of the discriminative effects. However, even at doses that decreased response rates, there was usually a rate of responding sufficient to adequately assess the discriminative-stimulus effects.

The differences in the sensitivity of the discriminative-stimulus effects and the effects on response rates to antagonism by the D1 antagonists suggest differing mechanisms underlying these effects of cocaine. The degree of D1 receptor involvement in the effects of cocaine on response rate appears to be substantially less than that for the discriminative-stimulus effects of cocaine. Similarly, the degree of D1 receptor involvement in the effects of cocaine on locomotor activity appears to be greater than that for the other effects. This latter conclusion, however, is tempered by the species differences in assessing these effects. However, a study on D1 dopamine receptor knockout mice (Miner et al., 1995) lends support to the conclusions of differential D1 receptor involvement in different behavioral effects of cocaine. In that study, conditioned place preference induced by cocaine was retained in the D1 knockout mice. As mentioned in the Introduction, Xu et al. (1994) found D1 knockout mice to be insensitive to the locomotor stimulant effects of cocaine. These results along with our findings clearly indicate a significant role of D1 dopamine receptors in the locomotor stimulant effects of cocaine. Moreover, these findings together suggest a more complex and possibly more limited involvement of D1 dopamine receptors in the effects mediating the covert stimulus effects underlying discriminative effects and possibly conditioned place preference.

The maximal shift in the dose-effect curve for the discriminative-stimulus effects of cocaine is not entirely surprising when considered in light of the multiplicity of the actions of cocaine. Cocaine not only inhibits the uptake of dopamine but also that of serotonin and norepinephrine. With its indirect dopaminergic agonist actions, each of the dopamine receptor subtypes should be stimulated. Selective antagonism of any one of these dopaminergic sites should remove part, but not all of the interoceptive stimulus effects of cocaine. How these various actions of cocaine contribute to its discriminative-stimulus effects is not clear. However, the current study suggests that there is a limit to the potential blockade of the discriminative effects of cocaine through antagonist actions at dopaminergic D1 receptors. It is possible that the limited antagonism of the discriminative effects of cocaine by selective antagonists occurs because only the D1 component of the stimulus is eliminated. As such, the discriminative-stimulus effects of cocaine may resemble those of a compound stimulus, with individual components combining in some manner to produce the discriminative effect. The elimination of any one component may not be sufficient to preclude a discriminative effect produced by the remaining components (Stolerman et al., 1991). The results of substitution studies with direct agonists, however, are not consistent with this interpretation. For example, many of the selective D2-like and D1-like agonists do not fully substitute for cocaine when administered alone, regardless of dose (Callahan et al., 1991; Witkin et al., 1991). However, full substitution has been on occasion reported following administration of D2-like agonists (Barrett and Appel, 1989; Ukai et al., 1993). Thus, there may be differences in the relative contributions of D1-like and D2-like dopamine receptors to the discriminative stimulus complex produced by cocaine.

For the discriminative effects of cocaine, there were differences in the interactions with cocaine of antagonists SCH 23390 and SCH 39166, and the partial agonists SKF 77434, (±)-SKF 38393, and (+)-SKF 38393. The antagonists each produced 3-fold shifts, whereas the partial agonists produced less of a shift, if at all. The present results with (+)-SKF 38393 somewhat differ from a previous finding; in that study (Spealman et al., 1997), the dose-effect curve for cocaine was shifted ~3-fold to the right by (+)-SKF 38393. In that study, the effects of cocaine were determined with a cumulative-dosing procedure, whereas in the present study single doses were administered. Although there have been reports that the cumulative-dosing procedure can interject effects specific to its use (Schindler et al., 1990; Walker and Branch, 1998), it is not clear that those effects were involved, or how they would have contributed to the differences between our study and the one by Spealman et al. (1997).

Interestingly, SKF 77434 is similar to SKF 38393 in terms of its stimulation of adenylyl cyclase activity in rodent tissue (O'Boyle et al., 1989), and SKF 77434 also antagonized the effects of cocaine in a manner like that of the D1 antagonists. Because SKF 77434 is also a partial agonist, it is clear that in the present study the antagonism of cocaine could be obtained with drugs having some intrinsic efficacy. The lack of a robust antagonism of cocaine by SKF 38393 may have been due to its actions mediated by other systems. For example, Zarrindast et al. (1991) found behavioral effects of SKF 38393 to be antagonized by the 5-HT antagonist metergolide, rather than by the dopamine D1 antagonist SCH 23390. Those results suggest that the actions of this drug that are mediated by 5-HT receptors (Biscoff et al., 1986) are more potent than its actions at D1 receptors. These 5-HT-mediated effects could have interfered with its ability to antagonize the effects of cocaine via a D1 partial agonist action.

The shifts in the cocaine dose-effect curves for discriminative-stimulus effects produced by the D1 antagonists in the present study were relatively modest, ~3-fold shifts. Greater shifts were obtained in the effects of cocaine on locomotor activity. These findings indicate differences in the involvement of D1 dopamine receptors in mediating these effects of cocaine. It is currently not clear which of these effects, if either, will be predictive of the therapeutic effects of the compounds in treating cocaine abuse. The discriminative-stimulus effects of cocaine are thought to be indicative of subjective effects that are presumably involved in the abuse of cocaine. If so, the current 3-fold shifts in the cocaine dose-effect curve may be, at least on first blush, disappointing. However, a 3-fold shift in the cocaine dose-effect curve may be adequate for a therapeutic agent. Surmounting even a minimal 3-fold antagonism of the effects of cocaine may very well be beyond the economic means of at least a subset of cocaine abusers.

In summary, several dopaminergic agents were examined alone and in combination with cocaine to better understand the role of D1 dopamine receptors in certain behavioral effects of cocaine. There was a clear limit to the degree to which the D1 ligands could shift the dose-effect curve for the effects of cocaine on response rates, with a greater shift in its discriminative-stimulus effects. There also were defined limits to the shifts in cocaine discriminative-stimulus effects, with a maximal 3-fold shift in the dose-effect curve. Furthermore, there was evidence that the activity of D1 dopaminergic receptors may be redundant with other mechanisms mediating the discriminative effects of cocaine. The locomotor-stimulant effects of cocaine were shifted to the right to an extent greater than that for the other behavioral effects examined and the maximal effects of cocaine were decreased by the antagonists. The antagonism was progressively increased with increasing doses of the D1 antagonists up to doses that produced an insurmountable antagonism. These results indicate a contribution of D1-receptor agonist actions in each of these effects of cocaine, however, with appreciable differences in the respective roles of D1 dopamine receptors in mediating each of these behavioral effects of cocaine.