Introduction

Top Abstract Introduction Methods Results Discussion References

The reinforcing effects of cocaine and other psychomotor stimulant drugs have been related to increased dopaminergic transmission via blockade of the presynaptic DA transporter (Ritz et al., 1987). A functional effect of such a blockade is prolonged stimulation of DA receptors by endogenous DA. There are several subtypes of DA receptors and the roles of the various DA receptor subtypes in mediating these reinforcing effects are still controversial. The original classification scheme proposed two types of CNS DA receptors, D1 and D2, distinguished by their ability to stimulate or inhibit, respectively, cAMP production (Kebabian and Calne, 1979; Stoof and Kebabian, 1981). More recently, the number of proposed CNS DA receptors has increased, and they are generally divided into two groups, D1-like and D2-like (Demchyshyn et al., 1995; Sibley and Monsma, 1992; Sugamori et al., 1994; Tiberi et al., 1991). For convenience, the terms "D1" and "D2" will be used here to refer the receptors in the respective groups.

One approach to investigating the role of DA receptors in the reinforcing effects of cocaine, or other drugs, is to study the reinforcing effects of selective agonists. Both D1 (Grech et al., 1996; Self and Stein, 1992; Weed et al., 1993; Weed and Woolverton, 1995) and D2 (Grech et al., 1996; Woolverton et al., 1984; Yokel and Wise, 1978) receptor agonists have been shown to function as positive reinforcers in monkeys and rats. Some, but not all, D1 receptor agonists functioned as reinforcers in rats and monkeys. The reason(s) for this difference between agonists is unclear, although intrinsic efficacy at D1 receptors may be an important variable. Those agonists found to function as positive reinforcers have generally been reported to have high efficacy in in vitro studies of cAMP formation in rat brain tissue, whereas those that did not function as positive reinforcers had low efficacy (Grech et al., 1996; Weed and Woolverton, 1995). Therefore, the present experiments were designed to address the question of why some, but not all, D1 receptor agonists can function as reinforcers in rhesus monkeys.

The overall approach of the present study was to correlate the results of in vivo assays of reinforcing efficacy with in vitro efficacy in relevant pharmacological assays. A PR paradigm of self-administration was used to provide a quantitative evaluation of the reinforcing potency and efficacy of several phenyl-benzazepine D1 receptor agonists (Weinstock et al., 1985) and cocaine. A PR paradigm progressively increases the number of responses required before reinforcement until at some point, called the breaking point (BP), the ratio of responses to reinforcer is large enough that responding stops. PR paradigms have been used to measure the reinforcing effects ("reward strength") of a variety of reinforcers including food (Hodos, 1961), liquid (Hodos and Kalman, 1963), electrical brain stimulation (Hodos, 1965) and drugs (Griffiths et al., 1978; Hoffmeister, 1979), and have been used to compare the relative reinforcing efficacy of drugs (Griffiths et al., 1978; Richardson and Roberts, 1996; Woolverton, 1995). D1 agonist potency and efficacy were established in vitro by evaluating the ability of the compounds to stimulate the production of the second messenger cAMP in striatal tissue. Rhesus monkey tissue was used in this assay so that in vivo/in vitro comparisons could be made within a species. The cAMP signal transduction system has been studied extensively in the rodent brain; however, comparatively little research on this transduction system has been done with primate brain tissue. Therefore, for comparison with existing literature, in vitro studies were also conducted with rat brain tissue.

Mechanisms other than D1 receptor stimulation may also play a role in the reinforcing effects of these compounds. For example, several phenyl-benzazepines, both D1 receptor agonists and antagonists, have recently been reported to increase synaptic DA in the rat dorsal striatum through an interaction with the DA transporter (Tomiyama et al., 1995). Therefore, the potency and efficacy of the D1 receptor agonists to block the uptake of DA into synaptosomal preparations of rat striata were also evaluated.

	Methods

Top Abstract Introduction Methods Results Discussion References

In Vivo: Self-administration

Animals and apparatus. Seven adult rhesus monkeys, Macaca mulatta (six male, one female), weighing between 5.0 and 12 kg, were used as subjects. One subject was experimentally and drug naive at the start of the experiment (9129). Two monkeys had histories of drug self-administration in this PR paradigm (8607, 9126, evaluating reinforcing effects of cocaine and procaine and/or evaluating temporal parameters of the PR paradigm; Woolverton, 1995; Rowlett et al., 1996). Four monkeys (8612, 8903, 11084, C53) had histories of self-administration of D1 agonists and cocaine under a fixed-ratio 10 (FR 10) self-administration paradigm (Weed et al., 1993; Weed and Woolverton, 1995). The drug reinforcer maintaining initial shaping of lever pressing for one monkey (C53) was SKF 81297 (0.1 and 0.03 mg/kg/injection, respectively; Weed and Woolverton, 1995). Monkeys were fed sufficient amounts of Teklad Monkey Diet (Harlan, Indianapolis, IN) to maintain stable body weights. Water was continuously available. In addition, monkeys were given a chewable multiple vitamin tablet 3 days a week and occasionally received fresh fruit.

Procedure. After adaptation to the cubicle and restraint system, each monkey was anesthetized with a combination of ketamine and isofluorane and a chronic indwelling venous catheter was surgically implanted (see Woolverton, 1995, for details). After recovery from surgery, daily sessions began at noon with illumination of the white stimulus lights above both levers. The experimentally naive subject was initially trained to lever-press by successive approximation under a FR 1 of cocaine delivery (base-line dose: 0.1 mg/kg/injection). During an injection the white lights were extinguished, the red lights were illuminated and responding was not counted. When the subjects demonstrated acquisition of cocaine-maintained lever-pressing under the FR 1, the response requirement was gradually raised until responding was reliable under a FR 30 schedule. At this point, a discrete-trials procedure was implemented. A total of 20 trials was available each day. The beginning of a trial was signaled by the illumination of the white stimulus lights located over both levers. Responses on the right lever led to delivery of a drug injection, whereas responses on the left lever were counted but had no other programmed consequences. Trials ended either by: 1) delivery of an injection or 2) the expiration of a fixed time period without obtaining an injection (LH). During an injection, the white lever lights were extinguished and the red lever lights were illuminated. At the end of the injection, all stimulus lights were extinguished and a TO began during which cocaine was not available. Monkeys with a history of responding under FR schedules began the discrete-trials procedure immediately. For naive monkeys, the length of the TO and LH were gradually increased to the terminal conditions of 30 and 15 min, respectively. Simultaneously, the response requirement was increased in small increments after the completion of four trials, which allowed for a maximum of five progressively greater ratios during a session. Response requirements were raised gradually, until animals were responding under terminal conditions. Daily sessions began at 12:00 P.M. each day and were terminated after the 20th trial or after the failure to complete the response requirement for drug delivery within the LH on two consecutive trials.

Data analysis. Number of injections, number of responses per session and BP (last completed FR) were recorded for each monkey. Data presented are individual and group means of the number of injections per session and running rate [responses emitted/(reinforcer latency  response latency)] over the last three stable sessions of availability of a dose. Dose-response functions for both measures were analyzed by repeated measures ANOVA by use of the within factors of drug, dose and replication upon the triplicate replications representing the last three stable days of drug availability for each dose of drug (SuperAnova, Abacus Concepts, Berkeley, CA). Contrast tests were used to compare the maxima between drugs. The single-point maxima for the D1 receptor agonists were compared with both of the means of the doses of cocaine that maintained maximum responding (0.17 and 0.3 mg/kg/injection). Because BP violates assumptions of homogeneity of variance (Depoortere et al., 1993; Rowlett et al., 1996; Woolverton, 1995), it was not analyzed statistically. For injections per session, ED₅₀ values were calculated on an individual basis by least-squares linear regression. Three or four doses, taken from the visibly linear portion of the dose-response function, were used in these calculations. If one dose engendered between 20 and 80% of the maximum effect observed in a monkey, then one dose with <20% effect and one dose with >80% effect were also used. If two doses engendered between 20 and 80% of maximum, then one dose with <20% effect and one dose with >80% effect were also used. If no doses engendered between 20 and 80% of maximum, then two doses with <20% effect and two doses with >80% effect were used. Mean ED₅₀ values and 95% CIs were calculated for the group.

In Vitro: D1 Agonist Efficacy

Animals. Both rats and monkeys were used to study D1 agonist efficacy. Male Sprague-Dawley rats (Harlan) lived in standard plastic rat cages and were maintained on a 12-h light/dark cycle (lights on at 6:00 A.M.). Water and food were available ad libitum. Rhesus monkeys [n = 12, 9 males (12278, 972, C53, 8805, 8217, 9165, 8236, 8711, 9001) and 3 females (11083, 7976, 8619)] lived in standard stainless steel primate cages and were maintained on a 16:8-h light/dark cycle (lights on at 6:00 A.M.). Water was available ad libitum, and they were fed sufficient monkey chow to maintain stable adult body weights.

Procedure. Monkeys were sacrificed by exsanguination during deep pentobarbital anesthesia. Brains were rapidly removed immediately after sacrifice, and the caudate nucleus and putamen were dissected from coronal sections according to the atlas of Snider and Lee (1961). Time between sacrifice and homogenization of membranes ranged from 1 to 6 h during which time the brain or homogenates were kept on ice. Approximately equal proportions of caudate and putamen were homogenized by hand with a Teflon/glass homogenizer until just homogeneous (about 10 strokes) in 25 volumes of ice-cold 10 mM imidazole HCl and 2 mM EGTA (pH 7.4). Homogenates were centrifuged at 27,000 × g for 15 min at 4°C. The resulting pellet was resuspended in fresh buffer and recentrifuged. The final pellet was homogenized in 30 to 40 volumes of fresh buffer with the addition of 10% glycerol to stabilize frozen samples. Homogenates which were frozen for assay at another time were fast-frozen in liquid nitrogen or on solid CO₂ and stored at 80°C. Storage times ranged from a minimum of 2 months to approximately 2 years. Most experiments with rhesus tissue were performed by use of frozen tissue homogenates.

8

3

Data analysis. Mean background radioactivity was subtracted from that of each sample tube to control for [³H]cAMP, which was not adsorbed by the charcoal mixture. A single-site binding isotherm was fit to the corrected [³H]cAMP standard curve, and this function was used to interpolate cAMP concentrations for the remainder of the samples (Prism, Graph Pad, San Diego, CA). Basal and DA-stimulated cAMP levels were converted into picomoles per milligram of protein per minute. Agonist-stimulated cAMP levels were converted to percentage of the level of 100 µM DA and pooled across assays. Outliers in the data were excluded based on the following criteria: 1) values which fell outside of the standard curve were automatically excluded by the curve-fitting program during the interpolation process; 2) values which fell outside two standard deviations of the pooled mean were excluded.

In Vitro: DA Uptake

Animals. The subjects were male Sprague-Dawley rats weighing between 200 and 224 g (Harlan). They lived in standard plastic rat cages and were maintained on a 12-h light/dark cycle (lights on at 6:00 A.M.). Water and food were available ad libitum.

Procedure. Rat striata were homogenized with 16 strokes of a Teflon pestle homogenizer (clearance, 0.003 inches; Kontes Co, Vineland, NJ) in 20 ml of an ice-cold solution of 0.32 M sucrose (Aldrich, Milwaukee, WI) and 5 mM NaHCO₃ (Fisher Scientific, Pittsburgh, PA), pH 7.4. The homogenate was centrifuged at 2,000 × g for 10 min at 4°C. The supernatant was centrifuged at 20,000 × g for 15 min at 4°C. The resulting pellet was resuspended in 1.2 ml of Krebs-HEPES buffer (KRH) consisting of 125 mM NaCl, 5 mM KCl, 1.25 mM CaCl₂, 1.5 mM KH₂PO₄, 0.1 mM ethylenediaminetetraacetic acid (all Fisher Scientific), 10 mM D-glucose (Aldrich), 1.5 mM MgSO₄, 25 mM HEPES, 0.1 mM pargyline and 0.1 mM ascorbic acid, pH 7.4, saturated with 95% O₂/5% CO₂.

Data analysis. Results are expressed as picomoles of DA uptake per minute per milligram of protein ± S.E.M. Inhibition values (IC₅₀; K_i) were obtained by a nonlinear iterative curve-fitting procedure (Prism, Graph Pad, San Diego, CA). ANOVA followed by Duncan's New Multiple Range test determined significant differences between log K_i values for the compounds tested.

Chemicals

Cocaine HCl and SKF 81297 [6-chloro-PB: 6-chloro-7,8-dihydroxy-1-phenyl-2,3,4,5-tetrahydro-(1H)-3-benzazepine HCl and HBr] were provided by the National Institute on Drug Abuse; Rockville, MD. SKF 81297 HBr, R(+)-6-BrAPB HBr (6-bromo-N-allyl-7,8-dihydroxy-1-phenyl-2,3,4,5-tetrahydro-[1H]-3-benzazepine) and SKF 82958 HBr (6-chloro-APB: 6-chloro-N-allyl-7,8-dihydroxy-1-phenyl-2,3,4,5-tetrahydro-[1H]-3-benzazepine) were purchased from Research Biochemicals Incorporated (Natick, MA). DA HCl was purchased from Sigma. For self-administration, cocaine was dissolved in 0.9% saline and the D1 agonists were dissolved in a vehicle of 2% ethanol and 98% saline. Injection volumes ranged from 0.8 to 1.33 ml/injection, and concentrations of drugs were adjusted accordingly to give the reported milligrams per kilogram per injection dose. Drug solutions administered intravenously were sterilized with a 0.22-micron filter unit (Millipore, Bedford, MA). Doses are for the salts of the drugs (HBr for SKF 81297).

For in vitro studies, the supplier of all chemicals was Sigma. DA HCl was dissolved in buffer. D1 agonists were dissolved at 1 mM in 25% ethanol and ultrapure water and then diluted in buffer. Drugs were mixed fresh before each assay.

	Results

Top Abstract Introduction Methods Results Discussion References

In Vivo: Self-administration

Cocaine and the three D1 receptor agonists maintained responding under the PR schedule (fig. 1). For all drugs, the number of injections per session increased with dose of drug to a maximum (top panel). The mean maximum number of injections per session were: cocaine, 16.0 injections/session (n = 7); SKF 82958, 15.2 injections/session (n = 6); R(+)-6-BrAPB, 14.3 injections/session (n = 6); SKF 81297, 13.1 injections/session (n = 6). The dose-response function for each D1 agonist was significantly different from the dose-response function of cocaine (for each, P < .05), and the dose-response function of SKF 81297 differed from that of each other compound (for each, P < .05). The maxima for SKF 81297, SKF 82958 and R(+)-6-BrAPB were not significantly different, nor were the maxima for cocaine, R(+)-6-BrAPB and SKF 82958 significantly different. The maxima for SKF 81297 and cocaine were significantly different (P < .05). Cocaine and each D1 agonist maintained the same group-average maximum BP of 960 responses per injection. Whereas the cocaine dose-response function was asymptotic, the dose-response functions for the D1 agonists were biphasic. On a milligram per kilogram basis, cocaine, SKF 82958 and R(+)-6-BrAPB were equipotent, whereas SKF 81297 was 3- to 10-fold less potent than the others. On a molar basis, the ED₅₀ values were: R(+)-6-BrAPB 42.1 (95% CI, 31.1-53.2) nmol/kg/injection; SKF 82958, 61.7 (95% CI, 49.9-73.5) nmol/kg/injection; SKF 81297, 247.9 (95% CI ,143.1-352.6) nmol/kg/injection. A dose-dependent decrease in running rate was maintained by each D1 agonist (P < .01), whereas the running rate maintained by cocaine increased with dose (P < .01; fig.1, bottom panel).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Self-administration of D1 agonists and cocaine under a PR schedule. Each point represents the mean of seven monkeys for cocaine and six monkeys for the D1 receptor agonists. Vertical bars represent the S.E.M. (Top Panel) Group means of PR performance maintained by cocaine and D1 receptor agonists. x-axis, dose of drug in milligrams per kilogram per injection; left y-axis, number of injections per session; right y-axis, break point (BP) associated with the number of injections per session. An asterisk indicates statistical significance at the P < .05 level between the maxima of SKF 81297 and cocaine. (Bottom Panel) Group means for running rate of responding. y-axis represents responses per second.

Considered individually, the dose-response functions of cocaine increased to an asymptote in most monkeys (n = 5: 8612, C53, 9129, 11084, 8903; data not shown). In one monkey (9126) responding only increased with dose, and in another responding was biphasic (8607). Two monkeys occasionally reached the procedural maximum of 20 injections/session for cocaine (8903, 0.3 mg/kg/injection, 1 of 3 days; 9126, 0.56 mg/kg/injection, 2 of 3 days). In all other cases responding maintained by cocaine and D1 receptor agonists was not constrained by the limitation of 20 possible trials. Similar to cocaine, the dose-response functions of the D1 receptor agonists exhibited the three shapes, at least in one monkey per agonist. However, in contrast to cocaine, the functions had biphasic shapes in three monkeys for each drug. The number of injections per session maintained by high doses of SKF 81297, SKF 82958, R(+)-6-Br APB and cocaine seemed to decrease when the TO was increased to 1 h. For 0.1 mg/kg/injection cocaine (n = 2), injections per sessions averaged 10.2 at the 30-min TO and 8.1 at the 60-min TO. Similarly, mean injections per session of SKF 81297 decreased from 15.3 to 11.3 at 0.3 mg/kg/injection (n = 1) and from 7.7 to 6.0 at 0.56 mg/kg/injection (n = 1). The comparable values were 10.3 to 9.3 for 0.3 mg/kg/injection SKF 82958 (n = 1), 16.7 to 8 for 0.3 mg/kg/injection R(+)-6-BrAPB (n = 1).

In Vitro: D1 Agonist Efficacy

There were no differences in the amount of cAMP produced (either basal or DA-stimulated) which could be attributed to the experimental history of the monkeys (fig. 2). Therefore, individual replications across all monkeys were converted to percentage of stimulation produced by 100 µM DA for each monkey, and these normalized replications were pooled across monkeys. EC₅₀ values were determined by fitting a sigmoidal curve to the pooled data. Dopamine stimulated the production of cAMP in both rat and rhesus striata to nearly twice that of the basal levels (fig. 3). Group means for rats were 1.99 times basal and for monkeys were 1.96 times basal. Based on 95% CIs which did not overlap, both basal and DA-stimulated production of cAMP in rat striatal homogenates was significantly higher than that seen with rhesus striatal homogenates whether the assays were performed in fresh or frozen homogenates. Freezing homogenates before assay had no effect on either basal or DA-stimulated cAMP production in rat tissue, but significantly decreased both in the rhesus homogenates. No effect of storage time in the 80°C freezer was apparent upon either basal or 100 µM DA-stimulated levels of cAMP production.

View larger version (41K): [in this window] [in a new window]   Fig. 2.   Effects of drug history on basal (B) and 100 µM DA-stimulated cAMP production. y-axis represents rate of cAMP production in picomoles per milligram of protein per minute. Bars represent individual monkeys. Error bars represent 95% CIs. Representative monkeys are grouped according to drug history into the following groups: control (11083, 12278), more than 2 months drug-free before sacrifice (8711, 8619, 9001, 7976) and less than 2 months drug-free before sacrifice (8805, 972, C53).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Basal (B) and DA-stimulated production of cAMP in striatal membrane homogenates. The results of two representative experiments are presented for both species. Homogenates were prepared immediately after sacrifice. Data marked "Fresh" are from experiments performed immediately after homogenate preparation. Data marked "Frozen" are from experiments performed with aliquots from the same tissue pools stored for at least 2 months at 80°C. (Top Panel) absolute rates of basal and 100 µM DA-stimulated cAMP production. (Bottom Panel) the same data as above normalized to percentage of basal stimulation. Error bars represent 95% CIs.

In both rhesus and rat striatal homogenates, D1 agonists stimulated the production of cAMP in a concentration-dependent manner (fig. 4). Although mean values were highest for SKF 81297 (96.3% and 101.2% of DA in rhesus and rat, respectively; table 1), apparent differences between SKF 81297, SKF 82958 and R(+)-6-BrAPB did not achieve statistical significance. SKF 38393 and SKF 77434 were significantly less efficacious than these three compounds (P < .05). Any stimulation of cAMP by S()-6-BrAPB was minimal (maximum rhesus: 15.9%; rat: 11.6%) and unrelated to drug concentration. SKF 38393 stimulated the production of cAMP to 32.5% of DA in rhesus striata but to 71.7% of DA in rat striata. The difference in efficacy of SKF 38393 between rhesus and rat striata was the only significant difference between efficacies in the two species seen in this study (P < .05). As judged by overlap of the 95% CIs of EC₅₀ estimates, none of the compounds displayed any potency differences either within or between species (table 1).

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Stimulation of cAMP production in rhesus and rat striata. x-axis, concentration of drugs in log molar units; y-axis, cAMP production as percentage of 100 µM DA. (Top Panel) rhesus striata (n = 6-8). (Bottom Panel) rat striata (n = 4-6). Error bars represent 95% CIs.

In Vitro: DA Uptake

The concentration-effect curves for all drugs to inhibit DA uptake into striatal synaptosomes were best fit by a one-site model. Cocaine and GBR 12909 fully blocked the uptake of DA into synaptosomes from rat striatal tissue, with GBR 12909 being significantly more potent (about 50-fold) than cocaine (fig. 5; table 2). Additionally, all D1 agonists fully blocked the uptake of DA. All the D1 agonists tested were significantly less potent (ranging from approximately 8-fold for R(+)-6-BrAPB to 90-fold for SKF 77434) than cocaine in this assay. Significant differences in potency were found among the D1 agonists. The enantiomers of 6-BrAPB were equipotent and significantly more potent than SKF 77434, SKF 82985 and SKF 38393. Furthermore, SKF 81297 was significantly more potent than SKF 77434. 

View larger version (32K): [in this window] [in a new window]   Fig. 5.   Blockade of DA uptake in rat striatal synaptosomes. The y-axis represents percentage of specific [3H]DA uptake. The x-axis represents concentration of drugs in log molar units. Each point is the mean of four rats and error bars represent the S.E.M.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Consistent with previous reports of reinforcing effects of D1 agonists (Grech et al., 1996; Self and Stein, 1992; Weed et al., 1993; Weed and Woolverton, 1995) SKF 81297, SKF 82958 and R(+)-6-BrAPB all functioned as reinforcers under a PR schedule. This finding extends the conditions under which D1 agonists have been found to function as positive reinforcers. SKF 82958 and R(+)-6-BrAPB were equipotent, whereas SKF 81297 was significantly less potent than the other two. Further, all appeared to have relatively high reinforcing efficacy. SKF 82958 and R(+)-6-BrAPB maintained PR performance that, at maximum, was similar to that maintained by cocaine, a highly efficacious reinforcer. Only PR performance maintained by SKF 81297 and cocaine differed significantly, with the maximum number of injections per session maintained by SKF 81297 being significantly lower than that of cocaine (13.1 vs. 16.0 injections/session, P < .05), consistent with lower reinforcing efficacy. On the other hand, when the D1 agonists were compared with each other, the maximum number of injections per session and BP maintained by SKF 81297, SKF 82958 and R(+)-6-BrAPB were not different, which suggested similar reinforcing efficacies. The difference between SKF 81297 and cocaine is an important positive control which demonstrates that this paradigm can resolve differences in the number of injections per session maintained by drugs (see also Woolverton, 1995). Thus, although the results are somewhat ambiguous, it is arguable that SKF 81297 had somewhat lower reinforcing efficacy than the other compounds. Grech et al. (1996) reported that SKF 81297 did not maintain self-administration as consistently as did SKF 82958, a finding which also is consistent with a somewhat lower reinforcing efficacy for SKF 81297.

In most cases, the dose-response function for cocaine increased with dose to an asymptote, which implied that the reinforcing effect reached a maximum that was maintained as dose was increased further (see also Woolverton, 1995). In contrast, the dose-response functions of the D1 agonists were biphasic. Biphasic dose-response functions are widely seen in self-administration paradigms, and have been reported in PR self-administration paradigms as well (Griffiths et al., 1978). One interpretation of this result is that the reinforcing efficacy of D1 receptor agonists decreased at higher doses. Alternatively, it may be that high doses of D1 receptor agonists have direct effects on responding that extend beyond the 30-min TO used in the present procedure and that increase as drug levels and response requirement accumulate during the session. The reductions in running rate seen with each D1 receptor agonist suggest a direct effect on rate. In fact, little is known about the pharmacokinetics of these D1 agonists in rhesus monkeys. However, the pattern of self-administration under an FR10 schedule maintained by SKF 81297, SKF 82958 and R(+)-6-BrAPB was one of steady high-rate responding, similar to the pattern seen with short-acting drugs such as cocaine (Weed and Woolverton, 1995). To examine the hypothesis of direct effects on responding more directly, three monkeys were tested with the TO increased to 1 h. The increased duration of the TO allowed more time for drug clearance between injections that should have reduced drug accumulation. The number of injections per session maintained by high doses of D1 receptor agonists was not higher under the 1-h TOs than under the half-hour TOs. This result suggests that drug accumulation cannot wholly explain the biphasic dose-response functions. It is possible that a 1-h TO was not sufficient for these effects to dissipate. Alternatively, there may be other effects (e.g., punishing effects) of higher doses of D1 receptor agonists that suppress responding.

The literature on cAMP production in brain tissue of non-human primates is limited. Basal levels of cAMP production in monkeys in the present experiment were generally similar to those found in previous studies using tissue from non-human primates (Izenwasser and Katz, 1993; Pifl et al., 1991; Vermeulen, et al., 1994). In addition, DA and the D1 receptor agonists stimulated the production of cAMP in membrane homogenates from rhesus striata. In the present study, the D1 receptor agonists did not differ significantly in potency from DA or from each other, but differences in efficacy of the compounds in stimulating cAMP production were evident. SKF 81297, SKF 82958 and R(+)-6-BrAPB stimulated cAMP production to at least 59% of DA levels, SKF 38393 and SKF 77434 had low efficacy and there was no evidence for D1 agonist efficacy of S()-6-BrAPB. Previous studies with cAMP production have also found SKF 38393 to be a partial D1 agonist in rhesus monkey tissue (Pifl et al., 1991; Vermeulen et al., 1994). On the other hand, Vermeulen et al. (1994) found that SKF 81297 had low efficacy in astrocyte cultures from rhesus monkey brain. Similar to the present results, Izenwasser and Katz (1993) reported that SKF 82958 was approximately equipotent and equiefficacious with DA in squirrel monkey tissue. On the other hand, SKF 81297 and R(+)-6-BrAPB generally lacked efficacy in that study. Thus, there is agreement as to the partial efficacy of SKF 38393 in non-human primates, but some ambiguity as to the D1 agonist potency and efficacy of other compounds. The fact that assay conditions were virtually identical in this and the Izenwasser and Katz (1993) study raises the possibility of differences between species of non-human primates, e.g., structural differences in the receptors or G-proteins which leads to differences in signal transduction. Unfortunately, the present data cannot address these mechanisms. Clearly, additional information from non-human primates is needed before firm conclusions can be drawn.

The findings of the present study were generally consistent with the literature on cAMP production in rat tissue (Kebabian and Calne, 1979; Miller et al., 1974). For most compounds, potency and efficacy were similar to previous reports. The efficacy of SKF 38393 in rat homogenates in the present study (72% of DA) was higher than some previously reported values (59%, Izenwasser and Katz, 1993; 39%, Pifl et al., 1991); however, other studies have reported similar or higher values for SKF 38393 (69%, Arnt et al., 1992; 75-80%, Weinstock et al., 1985). The stereoselectivity of BrAPB is consistent with previous reports on the stimulation of the production of cAMP by the enantiomers of SKF 38393 (Kaiser et al., 1982) or D1 receptor binding, which show that the R(+)-enantiomers of several of the phenyl-benzazepine compounds tested here have higher affinities for the D1 receptor than the S()-enantiomers (Neumeyer et al., 1992). As noted previously (Pifl et al., 1991), there was a difference between species in the efficacy of SKF 38393, with the production of cAMP being higher in the rat than in the monkey. In the present study, relative potency and efficacy of the other D1 agonists were similar in rats and rhesus monkeys. This is in contrast to the findings of Izenwasser and Katz (1993), who reported differences between rat and squirrel monkey tissue for cAMP production by several D1 receptor agonists (SKF 38393, SKF 81297 and R(+)-6-BrAPB).

In a previous studies (Weed and Woolverton, 1995; Woolverton et al., 1984), high-efficacy D1 agonists (SKF 81297, R(+)-6-BrAPB, SKF 82958) functioned as positive reinforcers in monkeys responding under an FR 10 schedule of reinforcement, whereas low efficacy agonists (SKF 38393, SKF 77434, S()-6-BrAPB) did not. The goals of the present study were to extend those studies and directly correlate in vitro receptor potency and efficacy with in vivo reinforcing potency and efficacy. R(+)-6-BrAPB and SKF 82958 were equipotent both in vitro and in vivo. SKF 81297 was less potent in vivo than predicted by the in vitro data. Although this difference in potency is inconsistent with the D1 receptor mechanisms hypothesis, pharmacokinetic factors could have contributed to the apparent discrepancy. Indeed, Pfeiffer et al. (1982) found that the presence of an N-allyl group enhanced the lipid solubility of benzazepines. SKF 81297 is the only compound in the present group of compounds without the N-allyl substituent. Regarding efficacy, the data are consistent with a qualitative relationship between D1 receptor efficacy and reinforcing efficacy. When grouped together, the D1 receptor agonists which functioned as reinforcers produced higher maximal stimulation of cAMP production than the compounds which did not function as reinforcers. The mean maximum stimulation produced by SKF 81297, SKF 82958 and R(+)-6-BrAPB (reinforcers) in rhesus tissue was significantly higher than the mean maximum stimulation produced by SKF 38393, SKF 77434, and S()-6-BrAPB (nonreinforcers) at the .05 level. This result is consistent with previous findings (Weed and Woolverton, 1995) and with the notion of a "threshold" effect (Ruffolo, 1982), i.e., that a certain level of intrinsic efficacy is required for an effect. That is, it would appear that a minimum D1 receptor efficacy is necessary for a compound to function as a positive reinforcer. In evaluating this conclusion, possible limitations on in vivo measurement should be considered. It may be that the present procedure (or any PR, for that matter) imposes some ceiling on responding, other than the maximum reinforcing effect, that influenced measurement. Regardless, the central hypothesis that D1 receptor efficacy plays a role in the reinforcing effects of these compounds in monkeys is supported.

The relationship between reinforcing efficacy and efficacy to stimulate the production of cAMP in rats cannot be evaluated at this time because only two D1 receptor agonists have been tested in rats. Both a high- (SKF 82958) and a low-efficacy (SKF 77434) D1-receptor agonist functioned as reinforcers in the Self and Stein (1992) study. From this study, one can determine that the behavioral effects of low-efficacy D1 receptor agonists may differ between rats and rhesus monkeys in that SKF 77434 did not function as a reinforcer in rhesus monkeys (Weed and Woolverton, 1995). However, from these data one cannot infer whether rats simply have a lower threshold for D1 receptor efficacy in the reinforcing effects of D1 receptor agonists than do rhesus monkeys, or whether there is a different relationship altogether. Clearly, further study in this area is needed.

Like cocaine and GBR 12909, the phenyl-benzazepines were also fully efficacious in blocking the uptake of DA. DA transporter blockade increases synaptic concentrations of DA and has been associated with reinforcing effects (Ritz et al., 1987). Recently, Tomiyama et al. (1995) demonstrated that intrastriatally administered phenyl-benzazepine compounds transiently increased the efflux of DA into striatal dialysate. Although it is possible that inhibition of DA transport played a role in the reinforcing effects of these phenyl-benzazepines, several points argue against this mechanism. Compounds that have been found not to have reinforcing effects (SKF 38393, SKF 77434, S()-6-BrAPB; Weed and Woolverton, 1995) were fully efficacious as DA uptake inhibitors. The R(+)- and S()-enantiomers of the phenyl-benzazepines were equipotent in inhibiting DA transport in contrast to the stereoselectivity observed for the reinforcing effects. Further, SKF 82958 and R(+)-6-BrAPB were equipotent with cocaine as reinforcers, but were 1 to 2 orders of magnitude less potent than cocaine as DA uptake blockers. Therefore, the ability to of the phenyl-benzazepines to interact with the DA transporter does not appear to be related to activity at the D1 receptor, which suggests that the interaction with the D1 receptor pharmacophore and the DA transporter pharmacophore rely on different elements of the phenyl-benzazepine structure.

Two other issues should be noted in evaluating the results of the present experiment: the use of frozen tissue and the use of tissue from monkeys with a history of drug exposure. It is unlikely that the use of frozen tissue altered the conclusions of the present study. Freezing striatal homogenates produced a decrease in both basal and 100 µM DA-stimulated levels of cAMP production. However, 100 µM DA produced a similar percentage increase in cAMP production, relative to basal levels, in either fresh or frozen striatal homogenates. Furthermore, frozen homogenates remained viable after 2 years of storage at 80°C. The use of frozen homogenates is important to the study of cAMP production in non-human primate tissue because it can greatly increase the amount of data that can be collected in each monkey. It also seems unlikely that drug history influenced the present results. No differences were apparent in either basal or 100 µM DA-stimulated levels of cAMP between monkeys that had been drug-free for at least 2 months and those that had self-administered dopaminergic drugs within 2 months of sacrifice. Although homogenates from one control monkey (12278) had relatively low basal levels, 100 µM DA-stimulated cAMP production homogenates from the other control monkey (11083) had relatively high levels. Izenwasser and Katz (1993) used one drug-naive and three drug-experienced squirrel monkeys in their experiment. They reported that the basal levels of stimulation in the drug-naive monkey were only one-third that of the three drug-experienced monkeys. The effect of drug history would be better elucidated in a study which was designed to evaluate drug history specifically.

In summary, the reinforcing effects of D1 receptor agonists in rhesus monkeys can be predicted by their ability to stimulate the production of cAMP, but not by their ability to inhibit DA transporter function. The current data are consistent with the interpretation that a certain level of efficacy to stimulate the production of cAMP is necessary for a D1 receptor agonist to function as a reinforcer in rhesus monkeys, but once this threshold is reached, the D1 receptor agonist will function as a high-efficacy reinforcer. An additional consideration is that self-administration of a drug by animals is predictive of abuse liability in humans. The fact that D1 agonists can maintain self-administration in several non-human species under a variety of conditions raises the possibility that high-efficacy D1 agonists may have abuse liability in humans.