There is now a large body of evidence supporting a role for glutamate receptor mechanisms in the behavioral effects of cocaine related to its abuse. Glutamate acts through both ligand-gated ionotropic receptors and G protein-coupled metabotropic receptors (mGluRs). Research on the role of glutamate mechanisms in cocaine abuse has focused extensively on the ionotropic glutamate receptors (Cornish and Kalivas, 2000; Di Ciano and Everitt, 2001; Tzschentke and Schmidt, 2003), and drugs acting at this class of receptors have been shown to modulate various behavioral effects of cocaine, including sensitization, self-administration, and reinstatement of drug seeking in animals (Li et al., 1997; Pierce et al., 1997; Kalivas, 2004).

Recent evidence suggests that mGluRs also play a significant role in the abuse-related effects of cocaine (Kenny and Markou, 2004; Lee et al., 2005; Weiss, 2005). The mGluRs constitute a family of eight receptor subtypes that are currently classified into three groups (groups I-III) based on sequence homology, pharmacology, and signal transduction mechanisms (Conn and Pin, 1997). Group II mGluRs (made up of the mGluR2 and three subtypes) are negatively coupled to adenylate cyclase (Conn and Pin, 1997; Schoepp, 2001) and are expressed at moderate-to-high levels in brain regions implicated in drug abuse (Xi et al., 2002a; Kenny and Markou, 2004). Repeated administration of cocaine has been shown to alter the function of these receptors as well as their regulation by cystine/glutamate exchange in the nucleus accumbens (Xi et al., 2002b; Moran et al., 2005), supporting a possible link between group II mGluR activity and the effects of cocaine.

Studies in rodents also suggest a role for group II mGluRs in the behavioral effects of cocaine and other stimulants. The selective group II mGluR agonist LY379268 (Monn et al., 1999), for example, blocked amphetamine-induced locomotion and rearing as well as the expression of sensitization in rats (Cartmell et al., 2000; Kim and Vezina, 2002). LY379268 also attenuated the reinstatement of cocaine seeking induced by a cocaine-associated stimulus in this species, although it was relatively ineffective in antagonizing the reinforcing effects of cocaine during self-administration (Baptista et al., 2004). Finally, Morishima et al. (2005) showed that knockout mice lacking the mGluR2 subtype exhibited enhanced cocaine sensitization and conditioned place preference compared with wild-type controls, effects essentially opposite to those induced by pharmacological stimulation of group II mGluRs. Although these findings implicate group II mGluRs in processes underlying cocaine's abuse-related effects, the role of these receptors in the expression of cocaine's reinforcing and stimulus effects is still largely unresolved.

The purpose of the present study was to investigate the interactions between the group II mGluR agonist LY379268 and cocaine in squirrel monkeys whose behavior was maintained under a second order schedule of i.v. cocaine self-administration either with or without presentations of a cocaine-paired stimulus, extinguished and then reinstated by priming injections of cocaine with or without a cocaine-paired stimulus, and controlled by the discriminative stimulus effects of cocaine. Antagonism studies were conducted with the group II mGluR antagonist LY341495 (Kingston et al., 1998) to determine its ability to block the cocaine-modulating effects of LY379268. Finally, quantitative observational studies were conducted to evaluate the effects of LY379268 and LY341495 on a range of species-typical behaviors, motor coordination, and muscle resistance. The results show that LY379268 reduced cocaine self-administration and cocaine-induced reinstatement of drug seeking in a dose-dependent, LY341495-reversible manner in the presence and absence of a cocaine-paired stimulus. However, LY379268 neither altered the discriminative stimulus effects of cocaine nor produced marked changes in species-typical behaviors at doses that attenuate cocaine's reinforcing and priming effects.

Materials and Methods

Subjects. A total of 14 adult male and female squirrel monkeys (Saimiri sciureus) weighing 0.7 to 1.1 kg were studied in daily experimental sessions. When animals were not being studied, they were housed in a climate-controlled vivarium with unlimited access to water. Animals used in cocaine self-administration, reinstatement, and observation studies had unlimited access to food (Teklad monkey diet; Harlan Teklad, Madison, WI) except during experimental sessions. Animals used in cocaine discrimination experiments were maintained at approximately 90% of their free-feeding body weights by adjusting their access to food in the home cages. All monkeys were maintained in accordance with the guidelines of the Committee on Animals of Harvard Medical School and the Guide for Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, Commission on Life Sciences, National Research Council, National Academy Press, Washington, DC, 1996). Research protocols were approved by the Harvard Medical School Institutional Animal Care and Use Committee.

Surgical Procedures. In cocaine self-administration and reinstatement studies, monkeys were implanted with i.v. catheters (polyvinyl chloride, i.d., 0.38 mm; o.d., 0.76 mm) using surgical procedures described by Platt et al. (2005). In brief, under isoflurane anesthesia and aseptic conditions, one end of the catheter was passed through the femoral or jugular vein to the right atrium. The distal end of the catheter was then passed s.c. and exited in the midscapular region. Catheters were flushed daily with 0.9% saline solution and sealed with stainless steel obturators when not in use. Monkeys wore nylon mesh jackets (Lomir Biomedical, Toronto, Canada) at all times to protect the catheter.

Apparatus. In studies involving cocaine self-administration, cocaine-induced reinstatement, and cocaine discrimination, daily sessions were conducted in ventilated, sound-attenuated chambers (MED Associates, St. Albans, VT), which were provided with white noise to mask extraneous sounds. Within the chambers, monkeys sat in Plexiglas chairs (MED Associates) facing an aluminum panel equipped with either one or two response levers (depending on type of study) and red and white lights, which could be illuminated to serve as visual stimuli. Each press of the lever with a downward force of approximately 0.25 N was recorded as a response. In experiments involving catheterized subjects, catheters were connected to syringe pumps (MED Associates) located outside the chamber. Each operation of the pump lasted 1 s and delivered a volume of 0.18 ml into the catheter. In cocaine discrimination experiments, 190-mg sucrose pellets (Bioserve, Frenchtown, NJ) could be delivered to a receptacle in the front panel of the chair.

Observational studies were conducted in a ventilated, transparent Plexiglas arena (114 × 122 × 213 cm) located in a lighted room isolated from other animals. The arena was equipped with perches, suspended plastic chains, and a wood-chip substrate to permit a range of species-typical behaviors. A digital video camera and recorder were positioned approximately 1 m in front of the chamber and operated continuously during the session.

Cocaine Self-Administration. Six monkeys were trained to self-administer cocaine under a second order fixed-interval (FI), fixed-ratio (FR) schedule of i.v. cocaine injection using the procedures described by Lee et al. (2005). Each monkey had self-administered cocaine under this schedule for at least 8 months and had been tested with other drugs as pretreatments before being assigned to the present study. Under the second order schedule, a white light was illuminated at the start of the session, and the completion of a 10-response fixed-ratio (FR10) produced a 2-s change in illumination from white to red. This cycle was then repeated throughout a 10-min FI, and the first FR10 completed after the expiration of the FI 10-min produced the 2-s light change paired with an i.v. injection of cocaine. A 60-s timeout (TO) period, during which all lights were off and responses had no scheduled consequences, followed each cocaine injection. If the FR requirement was not completed within 8 min following the expiration of the FI (limited hold), the TO period was started automatically without an injection. Daily sessions were completed after five cycles of the second order schedule or a maximal session length of 90 min. In initial studies, the dose of self-administered cocaine was varied over a 3- to 10-fold range to determine the dose that maintained maximal rates of responding for individual monkeys (0.18 mg/kg/injection for five monkeys; 0.3 mg/kg/injection for the sixth monkey). These unit doses then were held constant for the remainder of the study.

The effects of a range of doses of LY379268 (0.1-1.0 mg/kg i.m.) on self-administration of cocaine were determined in five monkeys. Monkeys were pretreated with LY379268 or vehicle 30 min before each of five consecutive sessions. On the first day of the experiment, subjects received a vehicle pretreatment before the session (baseline session). On the 2nd day of the experiment, the monkeys received an injection of LY379268 before the session, and on the following 3 days, they received vehicle before the session. This testing protocol was used to capture the time course of effects of LY379268, which were sometimes apparent beyond the 1st test day in preliminary studies. Doses of LY379268 were tested in a different order in each subject, and tests with different doses were separated by at least 1 week without intervening drug pretreatments.

Additional studies were conducted in six monkeys to determine whether the effects of LY379268 on cocaine self-administration could be reversed by coadministration of the group II mGluR antagonist LY341495. These studies were conducted using the general protocol described above with maximally effective doses of LY379268 (0.56 and 1.0 mg/kg) combined with either 10.0 mg/kg LY341495 or vehicle. LY341495 or vehicle was administered i.m. 35 min prior to the session, and LY379268 or vehicle was administered i.m. 30 min before the session. Tests with different combinations of LY379268, LY341495, and vehicle were conducted in irregular order across monkeys and were separated by at least 1 week without intervening drug pretreatments. An additional test in four monkeys determined the effects of 10.0 mg/kg LY341495 on cocaine self-administration in the absence of LY379268. The dose of LY341495 was selected on the basis of a previous report demonstrating central activity and antagonism of the effects of LY379268 in rodents (Feinberg et al., 2005) and on preliminary observations that the dose was well tolerated in squirrel monkeys.

A final experiment in four monkeys investigated the effects of LY379268 (0.56 and 1.0 mg/kg) on cocaine self-administration under the second order schedule in which the cocaine-paired stimulus was omitted (cf. Goldberg et al., 1979). Testing under this schedule was conducted as described above after allowing baseline response rates to restabilize under the second order schedule of cocaine injection without the cocaine-paired stimulus (8-19 days for individual subjects).

Reinstatement of Cocaine Seeking. Five of the six monkeys, who served in the cocaine self-administration studies, also were used to investigate the effects of LY379268 on the reinstatement of extinguished cocaine seeking. Of these five monkeys, three were tested first in the reinstatement studies described below, and two were tested first in cocaine self-administration studies described in the previous section. The procedures used to investigate reinstatement of cocaine-seeking behavior were similar to those described by Lee et al. (2005). In brief, after a period of stable responding under the second order schedule of cocaine self-administration, cocaine-seeking behavior was extinguished by substituting vehicle for cocaine injections and omitting presentations of the cocaine-paired stimulus. Extinction sessions were conducted daily until response rates declined to 10% or less of the rate maintained by cocaine self-administration (4-15 sessions depending on the monkey), at which time reinstatement testing was started. Reinstatement test sessions used the same procedures as cocaine self-administration sessions except that only vehicle was available for self-administration. Presentations of the cocaine-paired stimulus also were restored in initial tests because earlier studies showed that reinstatement of cocaine-seeking behavior was greatest when cocaine priming was accompanied by presentations of the cocaine-paired stimulus (Spealman et al., 2004).

The effects of a range of doses of LY379268 (0.1-1.0 mg/kg) and vehicle on cocaine-induced reinstatement of drug seeking were determined in five monkeys. In these experiments, monkeys received an i.m. pretreatment with LY379268 30 min before the test session followed by an i.v. injection of cocaine (1.0 mg/kg) immediately before the session. Doses of LY379268 were tested in a different order in each subject, and tests with different doses were separated by at least 1 week.

After an intervening period during which cocaine self-administration was re-established and then extinguished again, additional studies were conducted in four monkeys to determine whether the effects of LY379268 on cocaine-induced reinstatement of drug seeking could be antagonized by LY341495. These studies were conducted using the general procedure described above with LY341495 or its vehicle injected i.m. 35 min before the test session and LY379268 or its vehicle injected i.m. 30 min before the session; priming with cocaine (1.0 mg/kg i.v.) was administered immediately before the session. Tests with different combinations of LY379268, LY341495, and vehicles were separated by at least five sessions without drug pretreatments or cocaine priming.

After another intervening period during which cocaine self-administration was re-established and subsequently extinguished, a final study investigated the effects of LY379268 (0.56 and 1.0 mg/kg) on cocaine-induced reinstatement of drug seeking in the absence of the cocaine-paired stimulus. These tests were conducted in four monkeys using procedures identical to those described above except that the white stimulus light remained on at all times except during the TO (i.e., no stimulus change upon completion of the FR10).

Cocaine Discrimination. Four monkeys were trained to discriminate cocaine from saline using procedures similar to those described by Spealman et al. (1991). In brief, each monkey was trained to respond differentially on the left and right levers, depending on whether cocaine (0.3 mg/kg i.m.) or vehicle was injected. During each training session, 10 consecutive responses on one lever (left for two monkeys, right for two monkeys) produced a pellet of food if cocaine was injected, whereas 10 consecutive responses on the other lever produced a pellet of food if vehicle was injected. Each response on the inappropriate lever (e.g., the vehicle lever when cocaine was injected) reset the FR10 requirement. Delivery of each food pellet was followed by a 10-s TO, during which the chamber was dark and responses had no scheduled consequences.

Daily training sessions consisted of a variable number of components (n = 1-4) of the basic procedure described above. The number of components per session was varied irregularly from day-to-day with the restriction that each number of components occurred equally often within a block of 20 sessions. Each component ended after 10 food pellets had been delivered or after 5 min had elapsed, whichever occurred first. An extended (10-min) TO, during which injections of cocaine or vehicle could be administered, preceded each component. During training sessions, vehicle was injected at the midpoint of the 10-min TO that preceded the first n - 1 components, and cocaine was injected during the TO that preceded the nth component of the session. Periodically, saline was injected during TO periods preceding all four components of a session to prevent an invariant association between cocaine injection and the last component.

Drug testing began once monkeys made ≥90% of responses on the injection appropriate lever for a least five consecutive training sessions. Thereafter, test sessions were conducted at approximately weekly intervals with at least 2 training days preceding each test session and at least 2 days off following each test session. Test sessions were conducted only if ≥90% of responses were made on the injection-appropriate lever during the three most recent training sessions. Test sessions consisted of four components, each preceded by a 10-min TO period. During each component of a test session, completion of 10 consecutive responses on either lever resulted in delivery of a food pellet. Dose-response functions for cocaine and LY379268 were determined in four monkeys using the cumulative-dosing procedure described by Spealman et al. (1991). In brief, incremental doses of drug (0.25-0.5 log unit increments) were injected during the 10-min timeout period that preceded sequential components of a test session, permitting a four-point cumulative dose-response function to be determined in a single session.

Drug interaction experiments subsequently were conducted for cocaine after pretreatment with LY379268 or vehicle. LY379268 (0.3 or 1.0 mg/kg i.m.) or vehicle was administered 30 min before the test session, and cumulative doses of cocaine were administrated during the session as described above. Each drug combination was tested in four monkeys. An additional study using three of these monkeys was conducted by administering vehicle or 1.0 mg/kg LY379268 24 h before the cocaine test session as described above.

Observation Studies. Observation studies were conducted with four monkeys using procedures similar to those described by Lee et al. (2005). In brief, after habituation to the observation arena and handling and injection procedures, 30-min observation sessions were conducted daily, during which the animal's behavior was recorded continuously. Scoring of video recordings was conducted by two observers, who were trained in the use of the behavioral scoring system described by Platt et al. (2000) but were not informed about the drugs under investigation. Before beginning the study, each observer underwent at least 20 h of training until they reached an interobserver reliability criterion of ≥90% based on percentage agreement scores. The behavioral scoring system included ten categories (Table 1), which were scored by recording the presence or absence of each behavior in 15-s intervals during three 5-min observation periods across the session (0-5, 12-17, and 24-29 min). Frequency scores were calculated from these data as the number of 15-s intervals in which a particular behavior was observed.

In addition, during the 6th, 18th, and 30th min of each session, the monkey was removed from the observational arena by a trained handler and evaluated for ataxia [defined as the inability to balance on and/or grasp a stainless steel pole (56.0 cm in length; 1.0 cm in diameter) held in a horizontal plane] and muscle resistance (defined as resistance to hind limb extension). For ataxia, a score of 0 indicated that the monkey was able to balance normally on the pole, a score of 1 indicated inability to balance effectively, and a score of 2 indicated that the monkey could neither balance on nor grasp the pole. For muscle resistance, a score of 0 indicated no change in resistance to hind-limb extension, a score of +1 indicated increased resistance to extension and/or clinging to the grid floor, and a score of -1 indicated decreased resistance to extension and/or flaccidity. Drug test sessions were conducted once per week, with saline control sessions on intervening days. LY379268 (0.1-1.0 mg/kg) and LY341495 (10.0 mg/kg) were administered i.m. 30 min prior to placing the subject in the observation arena.

Data Analysis. In studies involving cocaine self-administration and reinstatement of drug seeking, response rates and number of injections were computed for individual subjects in each component of the experimental. Data for individual subjects (n = 4-6/experiment) were then averaged over the entire session. In drug self-administration experiments, data during the first component of the session (i.e., responses made before the first self-administration of cocaine) also were analyzed separately to provide a measure of reinforced drug seeking in the absence of the direct pharmacological effects of cocaine (cf. Spealman and Goldberg, 1978; Everitt and Robbins, 2000). In studies involving cocaine discrimination, the percentage of responses on the cocaine lever and the response rate (regardless of lever) were calculated for individual subjects in each component of a session. Individual data were then averaged across subjects (n = 3 or 4/experiment). In observation studies, individual scores for each behavior were averaged across the three 5-min observation periods of the session because no significant differences among the three periods were observed for any measure (determined by repeated measures ANOVA). Scores for individual subjects were then averaged for the group (n = 4). Data from all experiments were analyzed with one- or two-way repeated measures ANOVA and with Bonferroni's Student's t tests for planned comparisons (α level, p < 0.05) using the SigmaStat 3.1 statistical software package (Systat Software, Inc., Richmond, CA).

Drugs. Cocaine hydrochloride (Sigma-Aldrich, St. Louis, MO) and LY379268 (Eli Lilly and Co., Indianapolis, IN) were dissolved in 0.9% saline solution and sterile distilled water, respectively. LY341495 (Tocris Bioscience, Ellisville, MO) was dissolved in sterile distilled water containing 0.1 N NaOH, and the pH was adjusted by adding small amounts of 0.1 N HCl.

Results

Cocaine Self-Administration. Self-administered cocaine maintained consistently high baseline response rates under the second order schedule of i.v. drug injection throughout the study. Averaged across all control sessions that preceded drug pretreatment tests, response rates for individual subjects ranged from 0.72 to 1.30 responses/s, with a mean (±S.E.M.) of 0.89 ± 0.1 responses/s for the group of six monkeys. All monkeys self-administered the maximal possible injections per session (five) under control conditions.

Pretreatment with LY379268 30 min before the session resulted in dose-dependent decreases in the rate of responding (Fig. 1, left), and the number of injections per session (data not shown) was maintained by self-administered cocaine. Repeated measures ANOVA showed a significant main effect of LY379268 pretreatment on response rate (F_4,16 = 11.8, p < 0.001) and number of injections per session (F_4,16 = 22.3, p < 0.001). Subsequent pairwise comparisons revealed significant reductions in response rate after pretreatment with 0.3 to 1.0 mg/kg LY379268 and a significant reduction in the number of injections per session after pretreatment with 1.0 mg/kg LY379268 (p < 0.05, Bonferroni's t test). Emesis was observed in the majority of subjects following pretreatment with the highest dose of LY379268.

Because preliminary observations indicated that the effects of LY379268 could persist beyond the 1st day of administration, subjects were tested for an additional three sessions with vehicle administered 30 min before each session (Fig. 1, right). Repeated measures ANOVA revealed a significant effect of time (F_3,36 = 12.1, p < 0.001), dose (F_3,36 = 5.0, p < 0.05), and dose × time interaction (F_9,36 = 3.6, p < 0.01) on response rate as well as a significant effect of time (F_3,36 = 19.4, p < 0.001), dose (F_3,36 = 18.4, p < 0.001), and their interaction (F_9,36 = 17.8, p < 0.001) on the number of injections per session. Pair-wise comparisons showed significant decreases in response rate (Fig. 1, right) but not on the number of injections per session (data not shown) on the 2nd day following administration of 0.56 and 1.0 mg/kg LY379268. Response rates and number of injections per session were at or near control levels on days 3 and 4 following administration of all doses of LY379268. Emesis was not observed in any subject on test days 2 to 4.

Combined treatment with the group II mGluR antagonist LY341495 (10.0 mg/kg, 35 min before the test session) and LY379268 (0.56 and 1.0 mg/kg, 30 min before the test session) resulted in a significant attenuation of the inhibitory effects of LY379268 on both response rate (Fig. 2) and number of cocaine injections per session (data not shown). Two-way repeated measures ANOVA revealed significant main effects of LY379268 (F_2,6 = 25.0, p < 0.001) and LY341495 (F_1,6 = 7.2, p < 0.05) on response rate and a significant main effect of LY379268 on the number of injections per session (F_2,6 = 5.3, p < 0.05). Pairwise comparisons showed significant attenuation of the effects of 0.56 and 1.0 mg/kg LY379268 by 10.0 mg/kg LY341495 on response rate (p < 0.05, Bonferroni's t test) and a significant attenuation of the effects of 1.0 mg/kg LY379269 by 10.0 mg/kg LY341495 on the number of injections per session (from 2.3 ± 0.4 injections/session after LY379268 to 5.0 ± 0.0 injections/session after the combined drugs; p < 0.05, Bonferroni's t test). Pretreatment with LY341495 in the absence of LY379268 did not significantly alter the response rate or the number of cocaine injections per session compared with vehicle pretreatment. No significant effects of LY341495 alone or in combination with LY379268 were observed on test days 2 to 4 (data not shown).

Because an earlier study in rats suggested that LY379268 may preferentially attenuate drug seeking controlled by a cocaine-associated stimulus, we compared the effect of LY379268 on self-administration of cocaine under the second order schedule in which the cocaine-paired stimulus was either presented upon completion of each FR 10 or was omitted entirely. Under baseline conditions, omission of the cocaine-paired stimulus resulted in a significant reduction in response rate (Fig. 3, left pair of bars; p < 0.05, Bonferroni's t test) but not in the number of cocaine injections per session. Repeated measures ANOVA using dose of LY379268 and presence versus absence of the cocaine-paired stimulus as factors showed a significant main effect of LY379268 on response rate (F_2,6 = 13.7, p < 0.01) and an effect of LY379268 that approached significance on the number of injections per session (F_2,6 = 4.5, p = 0.06). Pairwise comparisons revealed that response rate was reduced significantly following pretreatment with both doses of LY379268 when the cocaine-paired stimulus was present and with 1.0 mg/kg LY379268 when the stimulus was absent (p < 0.05, Bonferroni's t test). The number of cocaine injections per session was reduced significantly following pretreatment with 1.0 mg/kg LY379268 under both stimulus conditions (to 2.5 ± 1.1 injections/session with the stimulus present and 1.3 ± 1.2 injections/session with the stimulus absent; p < 0.05, Bonferroni's t test).

Because there was a significant effect of omitting the cocaine-paired stimulus on response rate under baseline conditions, we reanalyzed the results as a percentage of the baseline response rate using data from the entire session as well as data collected during the first component of the session (i.e., before the first injection of self-administered cocaine). The latter analysis provided a measure of reinforced cocaine seeking that was unaffected by the direct pharmacological effects of recently administered cocaine (cf. Everitt and Robbins, 2000). Repeated measures ANOVA using data during the first component of the session (Fig. 4, left) showed a significant main effect of LY379268 (F_2,6 = 26.9, p < 0.001) but not presence versus absence of the cocaine-paired stimulus. Subsequent pairwise comparisons (Bonferroni's t test) revealed that, following pretreatment with 0.56 mg/kg LY379268, there was a significant reduction in response rate when the stimulus was absent (p < 0.05) and a near-significant reduction when the stimulus was present (p = 0.08). Following pretreatment with 1.0 mg/kg, there was a significant reduction in response rate under each stimulus condition (p < 0.05). A similar analysis using data from the entire session (Fig. 4, right) also showed a significant main effect of LY379268 (F_2,6 = 10.3, p < 0.01), but not of the stimulus, with a significant reduction in response rate after pretreatment with 0.56 mg/kg LY379268 when the stimulus was present as well as a significant reduction in response rate after pretreatment with 1.0 mg/kg LY379268 under both stimulus conditions (p < 0.05, Bonferroni's t test).

Reinstatement of Cocaine Seeking. During extinction sessions in which saline was substituted for cocaine and the cocaine-paired stimulus was omitted, response rates and the number of injections per session declined for all subjects, with group means of 0.03 (±0.02) responses/s and 1.3 (±0.6) injections/session. Priming with cocaine (1.0 mg/kg) and restoration of the cocaine-paired stimulus induced robust reinstatement of drug-seeking to a response rate (0.81 ± 0.12 responses/s) and number of injections (4.8 ± 0.2 injections/session) that approached the levels maintained by active cocaine self-administration (see above).

Pretreatment with LY379268 produced dose-dependent reductions in the rate of responding (Fig. 5) and number of injections per session (data not shown) engendered by cocaine priming. Repeated measures ANOVA showed a significant main effect of LY379268 on both response rate (F_4,16 = 12.6, p < 0.01) and the number of injections per session (F_4,16 = 4.2, p < 0.05) engendered by cocaine priming. Subsequent pairwise comparisons revealed significant reductions in cocaine-induced responding following pretreatment with 0.3 to 1.0 mg/kg LY379268 and a significant reduction in the number of injections per session following pretreatment with 1.0 mg/kg LY379268 (from 4.8 ± 0.2 to 2.4 ± 0.6 injections/session; p < 0.05, Bonferroni's t test). Emesis was observed in most subjects following pretreatment with 1.0 mg/kg LY379268.

Combined pretreatment with LY341495 (10.0 mg/kg) and LY379268 (0.56 and 1.0 mg/kg) before priming with 1.0 mg/kg cocaine resulted in a significant reversal of the effects of LY379268 on cocaine-induced drug seeking (Fig. 6). Repeated measures ANOVA revealed a significant main effect of LY379268 on both the rate of responding (F_2,6 = 19.6, p < 0.01) and the number of injections per session (F_2,6 = 10.3, p < 0.05). Subsequent comparisons showed that LY341495 significantly reversed the effects of 0.56 and 1.0 mg/kg LY379268 on cocaine-induced reinstatement of responding as well as the effects of 1.0 mg/kg LY379268 on the number of injections per session (from 1.5 ± 0.8 to 4.5 ± 0.3 injections/session; p < 0.05, Bonferroni's t test). Pretreatment with LY341495 in the absence of LY379268 did not have a significant effect on either the response rate or the number of injections per session engendered by cocaine priming.

A final experiment compared the effects of LY379268 on cocaine-induced reinstatement of drug seeking when the cocaine-paired stimulus was omitted during the test session (Fig. 7). Omission of the cocaine-paired stimulus during reinstatement testing with 1.0 mg/kg cocaine in the absence of LY379268 resulted in a nonsignificant reduction in the response rate induced by cocaine priming and no change in the number of injections per session. In reinstatement tests following pretreatment with LY379268, repeated measures ANOVA revealed a significant main effect of LY379268 on both response rate (F_2,6 = 13.6, p < 0.001) and the number of injections per session (F_2,6 = 84.5, p < 0.001) engendered by cocaine priming. Pairwise comparisons showed that 0.56 and 1.0 mg/kg LY379268 significantly reduced the rate of responding induced by cocaine priming in both the presence and absence of the cocaine-paired stimulus (p < 0.05, Bonferroni's t test). Likewise, the number of injections engendered by cocaine priming was reduced significantly by 1.0 mg/kg LY379268 in the presence (1.0 ± 0.0 injections/session) and absence (0.8 ± 0.5 injections/session) of the cocaine-paired stimulus (p < 0.05, Bonferroni's t test).

Cocaine Discrimination. Cocaine maintained consistent stimulus control over behavior throughout the study. During training sessions that preceded drug test sessions, individual monkeys made an average of ≥94% responses on the cocaine lever after injection of the training dose of cocaine (0.3 mg/kg) and ≥96% responses on the vehicle lever after injection of vehicle. Under test conditions, cocaine engendered dose-related increases in the percentage of responses on the cocaine lever, reaching a maximum of nearly 100% cocaine-lever responses after cumulative doses ≥0.3 mg/kg (Fig. 8, top). Repeated measures ANOVA showed that pretreatment with LY379268 (0.3 and 1.0 mg/kg) 30 min before the test session did not systematically alter the dose-response function for cocaine's discriminative stimulus effects. There was, however, a main effect of LY379268 on the rate of responding (F_3,18 = 13.9, p < 0.01), with modest although significant reductions in response rate observed following pretreatment with 1.0 mg/kg LY379268 compared with pretreatment with vehicle (Fig. 8, bottom; p < 0.05, Bonferroni's t test). When the pretreatment time for LY379268 was increased to 24 h, 1.0 mg/kg LY379268 had no systematic effect on the dose-response function for either the discriminative stimulus effects of cocaine or the rate of responding (data not shown). Finally, when tested in the absence of cocaine, LY379268 did not engender significant responding on the cocaine lever regardless of dose. The maximal percentage of cocaine-lever responding (3 ± 3%) was observed after a cumulative dose of 1.0 mg/kg LY379268, which also induced emesis in two of four monkeys.

Observed Behavior. LY379268 had no significant effects on the majority of recorded behaviors over the range of doses that significantly attenuated cocaine self-administration and cocaine-induced reinstatement of drug seeking. LY379268 (0.1-1.0 mg/kg) neither increased nor decreased locomotion, object exploration, foraging, self-grooming, scratching, or vocalization. It also did not induce a significant change in resting posture, static posture, balance, or muscle resistance. However, it did have a significant effect on visual scanning (F_3,9 = 8.0, p < 0.01), with a maximal increase of 26% observed following administration of 1.0 mg/kg (p < 0.05, Bonferroni's t test; data not shown). LY341495 (10.0 mg/kg) had no significant effect on any measure.

Discussion

In the present study, the group II mGluR agonist LY379268 attenuated cocaine self-administration by squirrel monkeys under a second order schedule of i.v. drug injection, resulting in dose- and time-dependent reductions in the rate of responding and the number of cocaine injections per session. The comparatively long duration of action of LY379268 in our study (up to 24 h after administration of high doses) is consistent with pharmacokinetic observations that significant concentrations of LY379268 can be detected in brain tissue 25 h after systemic administration in rodents (Bond et al., 2000).

Antagonism experiments showed that the effects of LY379268 were significantly reversed by the group II mGluR antagonist LY341495, suggesting that LY379268-induced attenuation of cocaine self-administration was mediated, at least in part, via activation of group II mGluRs. High doses of LY341495 have been reported to block mGluR8 (a subtype of group III mGluRs) in addition to the group II mGluRs (Schoepp, 2001), and it is conceivable that reversal of the effects of LY379268 by LY341495 in our study involved LY341495-induced antagonism at the mGluR8 subtype. Although this possibility cannot be ruled out, a direct interaction between LY379268 and LY341495 at mGluR8 seems unlikely in light of the low affinity that LY379268 has at this receptor (Monn et al., 1999; Schoepp, 2001).

Our finding that LY379268 attenuated self-administration of cocaine in monkeys seems to compliment the report by Morishima et al. (2005) that transgenic mice lacking the mGluR2 subtype show enhanced cocaine conditioned place preference and cocaine-induced sensitization (i.e., ostensibly opposite effects of pharmacological activation versus genetic inactivation of relevant mGluRs). Our results, however, are seemingly at odds with those of Baptista et al. (2004), who found that LY379268 was largely ineffective in reducing cocaine self-administration in rats. The apparently different results in our study and the study by Baptista and colleagues could be due to a number of factors but may reflect differences in the levels of exposure to self-administered cocaine prior to testing with LY379268. In a follow-up study by Baptista et al. (2005), for example, LY379268 dose-dependently attenuated cocaine self-administration in rats given comparatively long periods of access to cocaine each day, whereas only the highest dose of LY379268 affected cocaine self-administration in rats given shorter periods of access to cocaine each day. The similar results observed in rats with extended daily access to cocaine and in monkeys with extended histories (≥8 months in our study) of daily cocaine self-administration suggest that neuroadaptations resulting from exposure to sufficiently large total amounts of cocaine may augment the ability of LY379268 to attenuate cocaine self-administration.

Group II mGluRs act primarily as inhibitory autoreceptors and have been shown to regulate the release of glutamate and dopamine in the nucleus accumbens and other brain regions implicated in cocaine addiction (Conn and Pin, 1997; Hu et al., 1999; Cartmell and Schoepp, 2000; Xi et al., 2002a). Moreover, repeated exposure to cocaine induces significant reductions in Group II mGluR function, cystine/glutamate exchange, and basal extracellular levels of glutamate in the nucleus accumbens (Pierce et al., 1996; Xi et al., 2002b; Moran et al., 2005), along with an enhanced ability of cocaine and other stimulants to increase extracellular glutamate and dopamine in this region (Pierce et al., 1996; Kim et al., 2005). These interrelated functional changes would be expected to affect the ability of LY379268 to modulate cocaine-induced neurotransmitter release and cocaine's reinforcing effects in subjects with significant histories of drug self-administration. Consistent with this possibility, Kim et al. (2005) reported that LY379268 blocked the enhanced self-administration of amphetamine as well as the enhanced overflow of glutamate and dopamine in the nucleus accumbens as a result of repeated prior exposure to amphetamine.

In addition to reducing cocaine self-administration in our study, LY379268 dose-dependently attenuated the reinstatement of drug seeking induced by cocaine priming. Attenuation of the priming effects of cocaine was observed over the same range of doses of LY379268 that reduced cocaine self-administration, and these effects were significantly reversed by LY341495. The similar effects of LY379268 and its interaction with LY341495 in these two sets of experiments suggest that attenuation of the reinforcing and priming effects of cocaine by LY379268 is mediated via similar LY341495-sensitive mechanisms.

Previous studies in rodents have shown that LY379268 can attenuate the reinstatement of drug seeking in animals with histories of cocaine, heroin, or alcohol self-administration (Baptista et al., 2004; Backstrom and Hyytia, 2005; Bossert et al., 2005). These studies, which focused primarily on reinstatement of drug seeking induced by discrete or contextual stimuli associated with previous drug self-administration, have prompted speculation that LY379268 may preferentially affect drug seeking induced by conditioned environmental stimuli. Although this conclusion may be valid under some conditions (e.g., limited exposure to drug self-administration prior to testing), our findings suggest that LY379268 is effective in blocking reinstated drug seeking induced by cocaine priming alone or priming combined with presentations of a cocaine-paired stimulus. Recently, Baptista et al. (2005) also reported that LY379268 was effective in blocking cocaine priming-induced reinstatement of drug seeking in rats with histories of long (but not short) periods of daily access to self-administered cocaine. Collectively, these findings suggest that significant exposure to cocaine may induce functional changes in group II mGluRs that augment the effectiveness of LY379268 to block cocaine-induced reinstatement of drug seeking even in the absence of a cocaine-paired stimulus.

Under baseline conditions without LY379268 pretreatment, inclusion of a cocaine-paired stimulus during cocaine priming and cocaine self-administration engendered higher rates of responding than those observed in the corresponding baseline conditions when the stimulus was omitted. The most striking example of this was in the case of cocaine self-administration where the average response rate maintained by cocaine combined with presentations of the cocaine-paired stimulus was nearly twice as high as the response rate maintained by cocaine in the absence of the stimulus (see Fig. 3, left pair of bars). These results confirm earlier findings that omission of a cocaine-paired stimulus resulted in markedly reduced rates of responding under second order schedules of i.v. drug injection similar to the one used in the present study (Goldberg et al., 1979, 1981). The ability of a cocaine-paired stimulus to augment response rates under the second order schedule of drug injection appears to reflect at least in part the conditioned reinforcing effects of the stimulus because substituting a visual stimulus that was never paired with cocaine injections maintained low response rates compared with response rates maintained by the cocaine-paired stimulus (Goldberg et al., 1979). These observations in conjunction with the present results suggest that LY379268 does not preferentially attenuate cocaine seeking controlled by conditioned reinforcement, at least in the context of the second order schedule. This conclusion is further supported by the finding that LY379268 effectively reduced drug seeking in the first component of the self-administration session regardless of the presence or absence of the cocaine-paired stimulus. Had differences in drug effects due to conditioned reinforcement emerged, they would have been detected least ambiguously in this component, where drug seeking was not under the influence of an immediately preceding self-injection of cocaine (cf. Everitt and Robbins, 2000).

LY379268 has been found to reduce locomotor activity, rapid eye movement sleep, and fast (10-50 Hz) electroencephalographic wave recordings in rats, leading to speculation that the drug acts as a functional depressant of central arousal systems (Feinberg et al., 2002, 2005). Observational studies in squirrel monkeys, however, revealed no significant effect of LY379268 on locomotor activity, resting posture, or other behaviors that might reflect reduced arousal. The single significant effect of LY379268 in theses studies was an increase in visual scanning (also termed vigilance checking), which is not correlated with depressed arousal in squirrel monkeys (Winslow et al., 1989). Although the highest dose of LY379268 (1.0 mg/kg) reduced food-reinforced operant behavior in the cocaine discrimination study, possibly reflecting its emetic effects, our results suggest that the ability of LY379268 to attenuate cocaine self-administration and cocaine-induced reinstatement of drug seeking was not simply the result of a generalized reduction in arousal or operant behavior. These observations notwithstanding, the near-complete antagonism of cocaine self-administration and cocaine-induced reinstatement of drug seeking, which was observed after pretreatment with 1.0 mg/kg LY379268, may well reflect the drug's emetic side effects in additional to functional antagonism.

In contrast to the results obtained in the drug self-administration and reinstatement experiments, LY379268 did not systematically alter the discriminative stimulus effects of cocaine. These findings are noteworthy because they suggest a potentially unique ability of LY379268 to attenuate cocaine-maintained behaviors at doses that do not alter cocaine's effects as a discriminative stimulus. Although this conclusion should be viewed cautiously because of the different protocols under which cocaine was administered in the different studies, LY379268 might be expected to blunt cocaine's reinforcing and priming effects without altering its perceptibility. This pharmacological profile might be advantageous in the development of candidate medications to treat cocaine abuse and relapse. Although emetic and other potential side effects of LY379268 would be expected to limit its acceptability as a therapeutic candidate, group II mGluR agonists with fewer untoward effects might prove to be viable alternatives.