The Effects of Buprenorphine on Self-administration of Cocaine and Heroin Speedball Combinations and Heroin Alone by Rhesus Monkeys 1

Transcription

1 /98/ $03.00/0 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 285, No. 2 Copyright 1998 by The American Society for Pharmacology and Experimental Therapeutics Printed in U.S.A. JPET 285: , 1998 The Effects of Buprenorphine on Self-administration of Cocaine and Heroin Speedball Combinations and Heroin Alone by Rhesus Monkeys 1 NANCY K. MELLO and S. STEVENS NEGUS Alcohol and Drug Abuse Research Center, McLean Hospital-Harvard Medical School, Belmont, Massachusetts Accepted for publication January 7, 1998 This paper is available online at ABSTRACT Concurrent abuse of cocaine and opioids is frequently observed clinically, and we have developed a model of speedball self-administration involving the simultaneous injection of cocaine and heroin combinations in rhesus monkeys (Mello et al. (1995) J Pharmacol Exp Ther 274:1325). In the present study, we evaluated the effects of buprenorphine ( mg/ kg/day i.v.) and saline on speedball combinations of cocaine [0.001, 0.01 or 0.10 mg/kg/inj] and heroin [ mg/ kg/inj]. We also examined the effects of buprenorphine (0.075 and mg/kg/day i.v.) on self-administration of heroin alone ( mg/kg/inj). Drug and food (1-g banana pellets) self-administration were maintained on a second-order FR4 (VR16:S) schedule in four 1-hr sessions each day. Each buprenorphine or saline control treatment was evaluated for 10 consecutive days, and monkeys returned to base-line performance between each treatment condition. Buprenorphine ( mg/kg/day) selectively reduced self-administration of speedball combinations of low-dose cocaine (0.001 mg/kg/ inj) and heroin (0.001 or mg/kg/inj) (P.05.01), and buprenorphine (0.237 mg/kg/day) shifted dose-effect curves for The abuse of cocaine in combination with opioids is a common form of polydrug abuse that often involves the simultaneous intravenous administration of heroin and cocaine known as the speedball. The medical consequences of speedball abuse include an enhanced risk for lethal drug overdose (NIDA, 1996), as well as exposure to disorders common to all intravenous drug abuse [e.g., compromised immune function, and vulnerability to viral infections (AIDS, hepatitis) and bacterial infections (endocarditis, pulmonary infections, abscesses)] (Kreek, 1987, 1991; O Brien, 1996; Pillai et al., 1991; Schoenbaum et al., 1989). The prevalence Received for publication September 2, This research was supported in part by grants K05-DA 00101, P50-DA and RO1-DA from the National Institute on Drug Abuse, National Institutes of Health. Preliminary data were presented to the College on Problems of Drug Dependence in 1995 and 1997 and to the American College of Neuropsychopharmacology in speedball combinations of cocaine (0.001 mg/kg/inj) and heroin ( mg/kg/inj) downward (P.05.01) and approximately 1 log unit to the right. Buprenorphine treatment was less effective in decreasing responding maintained by speedball combinations of heroin and 0.01 and 0.10 mg/kg/inj cocaine. Buprenorphine treatment (0.075 and mg/kg/day) also shifted the heroin dose-effect curve downward (P ) and to the right. Both speedball and heroin self-administration were associated with dose-dependent decreases in food-maintained responding during saline control treatment. However, food-maintained responding was often higher than control levels during buprenorphine treatment (P ), which suggests that buprenorphine antagonized the rate-decreasing effects of speedballs and of heroin. Buprenorphine s selective reduction of speedball and heroin self-administration is consistent with clinical treatment trials in opioid abusers and polydrug abusers. Thus, these primate models of speedball and heroin self-administration should be useful for preclinical evaluation of novel drug abuse treatment medications. of speedball abuse has been difficult to quantify, but in one recent study, 63% of 926 polydrug abusers reported speedball use within a 6-month period (Schützet al., 1994). Cocaine use by methadone-maintained patients is reported by many treatment programs (Condelli et al., 1991; Gastfriend et al., 1993; Kosten et al., 1989b; Schottenfeld et al., 1993). Pharmacotherapies for opioid abuse are less effective in reducing cocaine abuse, and as yet, no consistently effective pharmacotherapy for cocaine abuse has been identified (Mendelson and Mello, 1996). Thus dual dependence on cocaine and opioids presents a special challenge for pharmacological treatment (Mendelson and Mello, 1996). The development of new pharmacological treatments can be facilitated by the availability of animal models of drug abuse for evaluating medication efficacy (Mello and Negus, 1996). Preclinical models of speedball self-administration have been developed only re- ABBREVIATIONS: ANOVA, analysis of variance; FR, fixed ratio; VR, variable ratio; DSM-III-R, Diagnostic and Statistical Manual of Mental Disorders (3rd ed-revised); inj, injection. 444

2 1998 Buprenorphine and Speedball Abuse 445 cently, and this report describes the effects of one potential treatment medication, buprenorphine, on the self-administration of cocaine and heroin combinations by rhesus monkeys. Despite emerging evidence of the clinical importance of speedball abuse (NIDA, 1996), there have been relatively few studies of the behavioral effects of cocaine and opioid combinations. Anecdotal reports by polydrug abusers suggest that cocaine and heroin combinations may produce a more pleasurable effect than either drug alone and/or may attenuate each other s aversive effects (Brecher, 1972; Kosten et al., 1986; Rosen and Kosten, 1991; Tutton and Crayton, 1993). Controlled clinical laboratory studies also indicate that methadone maintenance may enhance the subjective effects of cocaine under some conditions (Foltin et al., 1995; Preston et al., 1996). When polydrug abusers were given speedballs consisting of cocaine and hydromorphone combinations, they often identified cocaine alone as a speedball (Walsh et al., 1996). Although speedballs produced slightly greater subjective effects on some measures than cocaine or hydromorphone alone, these differences were not statistically significant (Walsh et al., 1996). However, when experienced polydrug abusers were given cocaine and morphine (speedball) combinations, they reported a distinct profile of subjective effects that was different from the effects of either drug alone (Foltin and Fischman, 1992). Preclinical studies of drug discrimination also have shown that cocaine and mu opioids may modulate each other s stimulus properties, and cocaine and heroin share discriminative stimulus effects under some conditions. For example, heroin substituted for cocaine in some rhesus monkeys, and its cocaine-like effects were antagonized by quadazocine, a mu opioid antagonist, but not by flupenthixol, a dopamine antagonist (Mello et al., 1995). These data suggested that heroin s cocaine-like effects were mediated by mu opioid receptors (Mello et al., 1995). We subsequently extended these findings to examine the extent to which a series of mu agonists shared discriminative stimulus properties with cocaine (Negus et al., 1998a). We found that mu agonists with a rapid onset of action (heroin, alfentanil) substituted for cocaine in more monkeys than slow-onset mu agonists (morphine, fentanyl) (Negus et al., 1998a). Morphine and fentanyl also did not consistently alter the discriminative stimulus effects of cocaine when administered as pretreatments to cocaine (Negus et al., 1998a). In contrast, in squirrel monkeys, mu opioid agonists did not substitute for cocaine, but pretreatment with mu opioids potentiated cocaine s discriminative stimulus effects (Spealman and Bergman, 1992, 1994). Another approach to the behavioral characterization of interactions between cocaine and opioids is to examine the reinforcing properties of a speedball combination of cocaine and heroin. We developed a model of speedball self-administration in the rhesus monkey and studied the reinforcing effects of nine speedball combinations of cocaine (0.001, 0.01 and 0.10 mg/kg/inj) and heroin (0.0001, and 0.01 mg/ kg/inj) (Mello et al., 1995). Each cocaine and heroin combination was evaluated for 10 days on a second-order schedule and compared with self-administration of cocaine alone and heroin alone (Mello et al., 1995). Intermediate doses of cocaine alone and heroin alone maintained equivalent high levels of drug self-administration, and combinations of cocaine and heroin usually maintained levels of drug self-administration similar to those maintained by either cocaine or heroin alone (Mello et al., 1995). Dose-dependent decreases in food-maintained responding occurred during cocaine, heroin and speedball self-administration, but speedball self-administration was not associated with any other overt toxic effects during the period of observation (Mello et al., 1995). In rats, cocaine ( mg/inj) and heroin (5.4 or 18 g/inj) combinations maintained responding under a FR10 schedule of reinforcement (Hemby et al., 1996). As in rhesus monkeys, self-administration of cocaine and heroin combinations was very similar to self-administration of either cocaine alone or heroin alone (Hemby et al., 1996). A recent study in rhesus monkeys also reported that cocaine alone and cocaine and heroin combinations maintained equivalent maximum numbers of injections under a progressive ratio schedule (Rowlett and Woolverton, 1997). When low doses of heroin ( g/kg/inj) that did not maintain drug self-administration alone were available in combination with cocaine, there was evidence for a leftward shift in the cocaine dose-effect curve (Rowlett and Woolverton, 1997). These demonstrations that speedballs reliably maintain self-administration over a wide dose range without toxic complications (Hemby et al., 1996; Mello et al., 1995; Rowlett and Woolverton, 1997) provide new preclinical models for evaluating drug abuse treatment medications. This report is the first study of the effects of buprenorphine, an opioid mixed agonist-antagonist, on self-administration of cocaine and heroin (speedball) combinations in rhesus monkeys. Buprenorphine has been previously reported to decrease the self-administration of opioids [heroin and hydromorphone] (Mello et al., 1983); [morphine] (Harrigan and Downs, 1981); [alfentanil] (Winger et al., 1992); and cocaine in rhesus monkeys (Lukas et al., 1995; Mello et al., 1989, 1990, 1992, 1993a, b) while producing minimal and transient effects on food self-administration. These basic findings of the effects of buprenorphine on cocaine self-administration have been replicated in several laboratories (Carroll et al., 1992; Carroll and Lac, 1992; Winger et al., 1992) (see Mello and Mendelson, 1995, for review). It is well established that buprenorphine reduces heroin self-administration in inpatient studies (Mello and Mendelson, 1980; Mello et al., 1982) and opioid abuse in outpatient clinical trials (Johnson et al., 1992; Strain et al., 1994) (see Bickel and Amass, 1995; Fudala and Johnson, 1995; Mello et al., 1993c, for review). Buprenorphine currently is being evaluated by the Food and Drug Administration as a treatment for opioid abuse as an alternative to methadone maintenance (Segal and Schuster, 1995). Buprenorphine also may be useful for the treatment of polydrug abuse, because it reduces both opioid and cocaine use in outpatient studies of persons dependent on both cocaine and opioids (Gastfriend et al., 1993; Kosten et al., 1989a, 1989b; Mello et al., 1993c; Schottenfeld et al., 1993). In controlled clinical laboratory studies in polydrug abusers, buprenorphine (4 mg sublingually) decreased the number of choices of high doses of cocaine (16 and 32 mg/70 kg) over tokens that could be exchanged for cigarettes, a variety of preferred foods and movie and music access (Foltin and Fischman, 1994). However, buprenorphine increased visual analog scale ratings of cocaine high and decreased reports of bad drug effect in comparison with cocaine alone (Foltin and Fischman, 1994). Subsequently, the effects of buprenor-

3 446 Mello and Negus Vol. 285 phine pretreatment on the subjective effects of speedballs were evaluated in polydrug abusers (Foltin and Fischman, 1995). In contrast to the effects of buprenorphine on subjective responses to cocaine alone (Foltin and Fischman, 1994), buprenorphine decreased ratings of drug liking and quality as well as Profile of Mood States elation and positive mood scores after administration of speedball combinations of morphine (5 or 10 mg/70 kg) and high doses of cocaine (32 mg/70 kg) but not after low doses of cocaine (8 mg/70 kg) and morphine combinations (Foltin and Fischman, 1995). Foltin and Fischman (1995) concluded that buprenorphine may be more effective in reducing cocaine abuse in speedball users than in persons who use cocaine independently of opioids. One goal of the present study was to evaluate the effects of chronic treatment with buprenorphine on the self-administration of speedball combinations consisting of a range of doses of heroin and low, intermediate and high doses of cocaine. A second goal was to re-examine the effects of buprenorphine on heroin self-administration by rhesus monkeys using the same range of heroin doses as in the heroin and cocaine (speedball) combinations. Our previous study examined only one dose of heroin in combination with buprenorphine (Mello et al., 1983), and other investigators have studied the effects of buprenorphine on self-administration of morphine (Harrigan and Downs, 1981) and alfentanil (Winger et al., 1992). We also were concerned with assessing the adequacy of these drug self-administration models for evaluation of treatment medications. Because buprenorphine has been studied extensively in clinical trials (see Mello and Mendelson, 1995; Mello et al., 1993c, for review), the present study allowed us to compare preclinical and clinical findings. As we have discussed elsewhere, cross-validation of findings in animal models of pharmacotherapy evaluation with results from clinical trials will be important for establishing the validity and the predictive value of preclinical drug selfadministration models (see Mello and Negus, 1996, for review). We now report that buprenorphine reduced self-administration of cocaine and heroin (speedball) combinations at most doses studied; however, occasional increases in high dose speedball-maintained responding also were observed. Buprenorphine also reduced self-administration of heroin alone by rhesus monkeys. For the most part, the effects of buprenorphine were selective for speedball and heroin selfadministration. Food-maintained responding during buprenorphine treatment tended to be equivalent to or higher than food-maintained responding during saline control treatment. It appeared that buprenorphine antagonized the ratedecreasing effects of high doses of both speedball and heroin. Buprenorphine s effects on speedball and heroin self-administration by rhesus monkeys are concordant with findings from clinical treatment trials. These data suggest the potential usefulness of this speedball self-administration model for evaluation of new pharmacotherapies for drug abuse treatment. Methods Drug self-administration procedures were used to study the effects of chronic buprenorphine treatment on speedball combinations of cocaine and heroin in comparison with heroin alone. The present report describes three studies of the effects of buprenorphine on drug and food self-administration. Study 1 evaluated the effects of buprenorphine ( mg/kg/day) on three speedball combinations that maintained high levels of drug self-administration. Study 2 examined the effects of an intermediate dose of buprenorphine (0.237 mg/kg/day) on dose-effect curves for speedballs containing a single dose of cocaine (0.001, 0.01 or 0.10 mg/kg/inj) and three or four doses of heroin ( mg/kg/inj). In study 3, the effects of buprenorphine (0.075 and mg/kg/day) on heroin self-administration dose-effect curves ( mg/kg/inj) were determined. The effects of buprenorphine on cocaine self-administration in rhesus monkeys studied under similar conditions have been reported previously (Lukas et al., 1995; Mello et al., 1989, 1990, 1992, 1993a, b). Subjects Subjects were five rhesus monkeys (Macaca mulatta) (4 males, 1 female) weighing 8 to 12 kg. Three monkeys had a history of selfadministration of cocaine and cocaine-heroin combinations and were subjects in a previous study of speedball (cocaine and heroin) selfadministration (606.5, 13441, 12780) (Mello et al., 1995). One monkey had a history of cocaine self-administration (89B057), and the female monkey (R800) had a history of alcohol self-administration before the beginning of the present series of studies. At least three monkeys were studied in each treatment condition, but not all monkeys were studied at all doses. Monkeys were maintained at ad libitum weight and given multiple vitamins, fresh fruit and vegetables and Lab Diet Jumbo Monkey Biscuits (PMI Feeds Inc., St. Louis, MO) to supplement a banana pellet diet. Food supplements were given between 5:00 and 5:30 P.M. Water was continuously available. A 12-hr light-dark cycle was in effect (lights on 7 A.M. 7 P.M.), and the experimental chamber was dark during food and drug self-administration sessions. Animal maintenance and research were conducted in accordance with the guidelines provided by the Institute of Laboratory Animal Resources (ILAR-NRC, 1996). The facility is licensed by the U.S. Department of Agriculture, and protocols were approved by the Institutional Animal Care and Use Committee. The health of the monkeys was monitored periodically by consultant veterinarians trained in primate medicine. Operant food and drug acquisition procedures provided an opportunity for environmental manipulation and enrichment (Line, 1987; Line et al., 1989). Monkeys had visual, auditory and olfactory contact with other monkeys throughout the study. Surgical Procedures Double-lumen Silicone rubber catheters (internal diameter, inch; outside diameter, inch) were implanted surgically in the jugular or femoral vein and exited in the midscapular region. All surgical procedures were performed under aseptic conditions. Monkeys initially were sedated with ketamine (5 mg/kg s.c.), then a surgical level of anesthesia was induced with sodium thiopental (10 mg/kg i.v.). Atropine (0.05 mg/kg) was administered to reduce salivation. An endotracheal tube was inserted, and anesthesia was maintained with halothane (1% mixed with oxygen). After surgery, monkeys were given 200,000 units of Combiotic Dihydrostreptomycin and Penicillin G every other day for a total of five i.m. injections. Aspirin or acetaminophen ( mg/day p.o.) also was administered for 3 days. The intravenous catheter was protected by a tether system consisting of a custom-fitted nylon vest connected to a flexible stainless steel cable and fluid swivel (Lomir Biomedical, Inc., Malone, NY). This flexible tether system permits monkeys to move freely. Catheter patency was evaluated periodically by administration of either a short-acting barbiturate, methohexital sodium (3 mg/kg i.v.) or ketamine (5 mg/kg) through the catheter lumen. If muscle tone decreased within 10 sec after drug administration, the catheter was considered patent.

4 1998 Buprenorphine and Speedball Abuse 447 Buprenorphine and Saline Administration Procedures The effects of daily treatment with saline or buprenorphine on speedball or heroin self-administration and food self-administration were studied. Each treatment condition was in effect for 10 days. At the end of each treatment condition, monkeys were returned to saline control treatment and the maintenance dose of cocaine for at least 2 days and until responding for cocaine and food returned to base-line levels. Cocaine (0.032 mg/kg/inj) was used as the maintenance drug to ensure high base-line rates of drug-maintained responding before each speedball or heroin substitution and treatment condition. Speedball combinations or heroin were substituted for the maintenance dose of cocaine in an irregular order. All buprenorphine doses were not studied in all monkeys. Either saline or buprenorphine was administered by slow infusion in a volume of 5 ml through one lumen of the double-lumen catheter from 9:30 to 10:20 each morning. These procedures were identical with those used in our previous studies of the effects of buprenorphine on cocaine self-administration (Lukas et al., 1995; Mello et al., 1989, 1990, 1992, 1993a, b). Saline delivery through the second lumen of the double-lumen catheter was automatically programmed as a pretreatment and to maintain catheter patency. For example, during saline control treatment conditions, saline was delivered from 9:30 to 10:20 A.M. every morning at a rate of 0.1 ml/min for a total of 5.0 ml. For the remaining 23 hr of each experimental day, 0.1 ml saline was delivered every 20 min for a total of 6.9 ml. Rationale for buprenorphine dose selection. We have reported previously that buprenorphine treatment at daily doses of 0.237, 0.32, 0.40 and 0.70 mg/kg/day (i.v.) significantly reduced cocaine self-administration under experimental conditions identical to those used in the present study (Lukas et al., 1995; Mello et al., 1989, 1990, 1992, 1993a, b). In study 1, we examined the effects of a range of buprenorphine doses ( mg/kg/day) on speedball selfadministration. These doses were lower or equivalent to the dose range we used in our previous studies of the effects of buprenorphine on cocaine self-administration. In study 2, we examined the lowest dose of buprenorphine (0.237 mg/kg/day) that significantly reduced cocaine self-administration in our earlier study (Mello et al., 1990). In study 3, the effects of the same dose of buprenorphine (0.237 mg/kg/day) and a lower dose (0.075 mg/kg/day) on heroin self-administration were examined. Behavioral Procedures Monkeys were housed individually in stainless steel chambers equipped with a custom-designed operant response panel, a banana pellet dispenser (Gerbrands model G5210, Arlington, MA) and two syringe pumps (model , Harvard Apparatus, Inc., South Natick, MA), one for each lumen of the double-lumen catheter. During food self-administration sessions, the response key on the operant panel was illuminated with a red light, and responding under an FR4 (VR16:S) schedule resulted in presentation of a 1-g banana pellet (P.J. Noyes Co., Lancaster, NH). During drug self-administration sessions, the response key was illuminated with a green light, and responding under an FR4 (VR16:S) schedule resulted in delivery of 0.1 ml of saline or a drug solution over 0.9 sec through one lumen of the double-lumen catheter. A 10-sec time-out followed delivery of each drug or saline injection or food pellet. The operation of the syringe pump was audible to the monkey. Schedules of reinforcement were programmed by custom-designed software and run on Apple IIGS microcomputers. Details of this apparatus have been described previously (Mello et al., 1990). Food and drug sessions each lasted for 1 hr or until 20 drug or saline injections or 25 food pellets had been delivered. Monkeys could earn a maximum of 80 injections per day and a maximum of 100 food pellets per day. Food sessions began at 6 A.M., 11 A.M., 3 P.M. and 7 P.M., and drug sessions began at 7 A.M., 12 noon, 4 P.M. and 8 P.M. At all other times, responding had no scheduled consequences. The experimental room was dark during all drug and food sessions. Drug Self-Administration Procedures All monkeys were trained to self-administer cocaine (0.01 or mg/kg/inj i.v.) and subsequently given access to speedball combinations of cocaine and heroin and to heroin alone. During speedball self-administration, cocaine and heroin were prepared in a single solution and delivered through one catheter lumen as in our previous study (Mello et al., 1995). Simultaneous administration of cocaine and heroin combinations was designed to simulate speedball selfadministration by humans. Dose-effect curve determinations. Stable food- and cocainemaintained responding under the terminal FR4 (VR16:S) schedule were defined by the following criteria: 1) 3 consecutive days during which the number of drug injections on each day differed by no more than 20% from the mean number of drug injections per day and there was no upward or downward trend; and 2) during the same 3 consecutive days, the mean number of both drug injections per day and food pellets per day was greater than 50. After cocaine-maintained responding was stable under the final schedule requirement [FR4 (VR16:S)], the effects of substituting saline and cocaine alone ( mg/kg/inj) for the maintenance dose of cocaine were determined. Each substitution remained in effect for at least 5 days and until responding had stabilized. Responding was considered to have stabilized when one of the following three criteria had been met: 3 consecutive days during which the number of drug injections delivered on each day differed by no more than 20% from the mean number of drug injections per day and there were no upward or downward trends, 3 consecutive days during which the number of injections per day on each day was 20 or less, or the substitution treatment had been in effect for 10 days. After preliminary determination of cocaine dose-effect curves, dose-effect curves for heroin alone ( mg/kg/inj) and combinations of cocaine and heroin were determined. Saline and each test dose of heroin or a cocaine/heroin mixture were substituted for the maintenance dose of cocaine for 10 consecutive days. The uniform use of 10-day substitution conditions in this phase of the study was designed to facilitate comparisons with subsequent 10-day periods of buprenorphine treatment. At the conclusion of each substitution, the maintenance dose of cocaine was reinstated for at least 4 days and until responding for cocaine and food returned to base-line levels. Test conditions were presented in an irregular order. Group data for some doses of cocaine, heroin and speedball during saline treatment conditions were reported previously (Mello et al., 1995), but additional determinations were conducted for the present study and two additional monkeys were studied. Heroin and cocaine (speedball) dose combinations. The effects of buprenorphine on the self-administration of cocaine/heroin combinations or of heroin alone were examined in a series of three studies. In study 1, four monkeys (606.5, 12780, and 89B057) were given access to three different speedball combinations during chronic treatment with either saline or buprenorphine ( mg/kg/day). The three different speedball combinations used in these studies were designed to vary the proportion of cocaine and heroin in the speedball mixture as follows. (1) Low dose cocaine/medium dose heroin: This speedball combination was composed of a low dose of cocaine that maintained low levels of self-administration (0.001 mg/ kg/inj) in combination with a medium dose of heroin selected from the peak of the heroin dose-effect curve (0.001 mg/kg/inj in monkeys 606.5, and 13441; mg/kg/inj in monkey 89B057). (2) Medium dose cocaine/medium dose heroin: This speedball combination was composed of a medium dose of cocaine selected from the peak of the cocaine dose-effect curve (0.01 mg/kg/inj) in combination with a medium dose of heroin from the peak of the heroin dose-effect curve (0.001 or mg/kg/inj). (3) High dose cocaine/medium dose heroin: This speedball combination was composed of a high dose of cocaine on the descending limb of the cocaine dose-effect curve (0.10 mg/kg/inj) in combination with a medium dose of heroin from the peak of the heroin dose-effect curve (0.001 or mg/kg/inj).

5 448 Mello and Negus Vol. 285 Speedball doses and buprenorphine doses were administered in an irregular order. As noted earlier, each treatment condition was separated by at least 2 days of saline control treatment until responding for drugs and food returned to base-line levels. In study 2, the effects of an intermediate dose of buprenorphine (0.237 mg/kg/day) were examined on self-administration of speedball combinations consisting of various doses of heroin ( mg/kg/inj) in combination with low, medium and high doses of cocaine ( mg/kg/inj). For the purposes of experimental design, analysis and presentation, these speedball dose-effect curves are displayed graphically as a function of heroin dose. Thus, one speedball dose-effect curve consisted of a low dose of cocaine (0.001 mg/ kg/inj) in combination with each of four heroin doses. The second speedball dose-effect curve consisted of an intermediate dose of cocaine (0.01 mg/kg/inj) in combination with three or four of the same doses of heroin. The third speedball dose-effect curve consisted of a high dose of cocaine (0.1 mg/kg/inj) in combination with three or four doses of heroin. The heroin mg/kg/inj cocaine speedball studies were conducted in monkeys 606.5, and The heroin 0.01 mg/kg/inj cocaine studies were conducted in monkeys 606.5, 89B057 and R800. The heroin 0.1 mg/kg/inj cocaine studies were conducted in monkeys 606.5, 89B057 and In study 3, the effects of two doses of buprenorphine (0.075 and mg/kg/day) were examined on the complete dose-effect curves for heroin alone ( mg/kg/inj) in monkeys 606.5, R800 and 89B057. Drugs Cocaine HCl, heroin (3,6-diacetylmorphine HCl) and buprenorphine HCl were obtained in crystalline form from the National Institute on Drug Abuse, National Institutes of Health. The purity of cocaine and heroin was certified by Research Triangle Institute, Research Triangle Park, NC, to be greater than 98%. All drugs were dissolved in sterile saline or sterile water (buprenorphine), filtersterilized with a 0.22 micron Millipore filter and stored in sterile, pyrogen-free vials. Data Analysis The dependent variables were the total number of injections per day and the total number of food pellets per day acquired during four drug and four food sessions. In study 1, statistical analyses were based on the mean ( S.E.M.) of the number of injections and food pellets per day delivered during the last 3 days of a 10-day period of buprenorphine or saline control treatment. Changes in drug- and food-maintained responding during buprenorphine treatment were compared statistically with the saline treatment base-line with an ANOVA for repeated measures. The statistical significance of all ANOVAs was evaluated with Huynh-Feldt Epsilon factors for degrees of freedom adjustment of within-group means (Super ANOVA Software Manual, Abacus Concepts, Inc., Berkeley, CA, 1989). If the ANOVA showed a significant main effect, Dunnett s multiple comparisons procedure was used to determine which points differed significantly. Results of studies 2 and 3 are shown as speedball or heroin dose-effect curves based on the mean ( S.E.M.) of the number of injections and food pellets per day delivered during the last 3 days of each 10-day treatment condition. Differences between corresponding speedball or heroin dose points during saline control treatment and buprenorphine treatment also were compared where possible with a two-factor ANOVA for repeated measures and Huynh- Feldt Epsilon factors for degrees of freedom adjustment of between group means. In addition, the mean numbers of injections and food pellets delivered each day during a 10-day buprenorphine and a 10-day saline treatment condition are shown for one speedball heroin and cocaine dose combination and for heroin alone. Probability values of P.05 are reported as statistically significant. Results Cocaine, Heroin and Speedball Dose-Effect Curves Dose-effect curves were determined for self-administration of cocaine alone and for heroin alone to provide a basis for comparison with dose-effect curves for cocaine and heroin (speedball) combinations. The dose-effect curve for self-administration of cocaine alone ( mg/kg/inj) was an inverted-u shape similar to that usually observed in this laboratory (Lukas et al., 1995; Mello et al., 1995; Negus et al., 1996, 1997). As shown in figure 1, the lowest dose of cocaine alone (0.001 mg/kg/inj) maintained an average of 21 injections per day, whereas at the intermediate dose (0.01 mg/kg/ inj) and high dose (0.10 mg/kg/inj) of cocaine alone, cocaine self-administration averaged 72 and 58 injections per day, respectively. Dose-effect curves for heroin alone ( mg/kg), described later in study 3, were also similar to those reported in our previous study (Mello et al., 1995). Speedball self-administration dose-effect curves were determined under saline control and buprenorphine treatment conditions, and these data are described below in study 2. Study 1: Effects of Buprenorphine on Self-Administration of Selected Heroin and Cocaine Combinations The effects of three doses of buprenorphine ( mg/kg/day i.v.) on three speedball cocaine and heroin combinations were studied to establish the behaviorally active dose range of buprenorphine. The lowest dose of buprenorphine studied ( mg/kg/day) had minimal effects on a speedball combination consisting of a low dose of cocaine (0.001 mg/kg/inj) and a medium dose of heroin (0.001 or mg/kg/inj) (data not shown). Food-maintained responding was also unaffected by low-dose buprenorphine treatment, and this dose of buprenorphine was not evaluated subsequently. As shown in figure 2, higher buprenorphine doses (0.075 and 0.75 mg/kg/day) each significantly decreased responding maintained by a speedball combination consisting of the low dose of cocaine and medium dose of heroin (P.05). At speedball combinations consisting of a medium dose of cocaine (0.01 mg/kg/inj) and a medium dose of heroin, speedball-maintained responding was lower during buprenorphine treatment (0.075 and 0.75 mg/kg/day) than during saline control treatment, but these differences did not Fig. 1. Cocaine dose-effect curves. Effects of manipulating the unit dose of cocaine on the number of injections per day (left ordinate) and food pellets per day (right ordinate) in monkeys trained to respond for cocaine (closed circles) and food (open squares). Abscissa, Saline, the number of injections per day and food pellets per day delivered when saline was substituted for cocaine; and the unit dose of cocaine available for selfadministration. Each unit dose of cocaine was available for at least 5 consecutive days, and each point shows the average ( S.E.M.) data from the last 3 days in each of five monkeys.

6 1998 Buprenorphine and Speedball Abuse 449 Fig. 2. Effects of buprenorphine on speedball combinations of cocaine and heroin. The number of drug injections per day or food pellets per day (x S.E.) is shown on the left and right ordinates, respectively. Speedball combinations consisted of a low (0.001 mg/kg/inj), medium (0.01 mg/kg/ inj) or high (0.10 mg/kg/inj) dose of cocaine in combination with a medium (0.001 or mg/kg/inj) dose of heroin as shown on the abscissa. Saline treatment is shown as a closed circle. Buprenorphine treatments (0.075 and 0.75 mg/kg/day) are shown as shaded squares and shaded triangles, respectively. Data presented are the average ( S.E.M.) of the last 3 days of 10 days of treatment with saline or buprenorphine. Each point is the average of data from three or four monkeys. Statistically significant differences between corresponding speedball doses during saline control treatment and buprenorphine treatment are indicated as follows: *0.075 mg/kg/day buprenorphine different from saline, P.05; 0.75 mg/kg/day buprenorphine different from saline, P.05. achieve statistical significance. At the highest dose of cocaine (0.10 mg/kg/inj) in combination with the medium dose of heroin, buprenorphine (0.075 and 0.75 mg/kg/day) reduced speedball self-administration significantly compared with saline control treatment (P.05). Food-maintained responding did not differ significantly during buprenorphine treatment and saline treatment (fig. 2). Fig. 3. Effects of buprenorphine (0.237 mg/kg/day) on speedball doseeffect curves for individual monkeys. Dose-effect curves for a low dose of cocaine (0.001 mg/kg/inj) in combination with one of four doses of heroin ( mg/kg/inj) are shown for individual monkeys (left panel). The unit dose of heroin in combination with cocaine is shown on the abscissa. Points above S show data from saline treatment sessions when saline was the solution available for self-administration (open circles). Self-administration of each cocaine-heroin speedball combination during saline treatment (closed circles) and during buprenorphine treatment (0.237 mg/kg/day) (shaded squares) are shown on the left ordinate as injections per day. Each data point is the average of the last 3 days of 10 days of self-administration of each cocaine and heroin combination for each monkey (x S.E.M.). Food-maintained responding during saline self-administration (open circles), self-administration of cocaine and heroin combinations during saline treatment (closed circles) and during buprenorphine treatment (0.237 mg/kg/day) (shaded squares) is shown in the right panel. The number of banana pellets per day earned during each condition are shown on the right ordinate. Statistically significant differences between corresponding speedball doses during saline control treatment and buprenorphine treatment are indicated by asterisks (*P.05; **P.01; ***P.001). Study 2: Effects of Buprenorphine (0.237 mg/kg/day) on Speedball Self-Administration Dose-Effect Curves Figures 3, 4 and 5 show the effects of saline and buprenorphine treatment (0.237 mg/kg/inj) on the dose-effect curves for self-administration of speedballs consisting of a low dose of cocaine (0.001 mg/kg/inj), an intermediate dose of cocaine (0.01 mg/kg/inj) or a high dose of cocaine (0.10 mg/kg/inj) in combination with heroin ( mg/kg/inj). Data are shown for the last 3 days of each 10-day treatment period for individual monkeys. At speedball combinations consisting of low (0.001 mg/kg/inj) and high (0.10 mg/kg/inj) doses of cocaine and heroin, the speedball dose-effect curve was a typical inverted-u shape (figs. 3 and 5). At speedball combinations consisting of an intermediate dose of cocaine (0.10 mg/ kg/inj) and heroin, the lowest doses studied maintained relatively high levels of responding, but responding decreased as the heroin dose increased (fig. 4). Similarly, there was a speedball dose-dependent decrease in food-maintained responding at all speedball combinations, with two exceptions. Food-maintained responding was stable across all speedball combinations in monkey 89B057 (fig. 4) and increased at the highest speedball dose studied in monkey (fig. 5). At the low dose cocaine and heroin combination, buprenorphine shifted the ascending limb of the speedball dose-response curve downward and approximately one log unit to the right in all three monkeys (fig. 3). Speedball-maintained responding at doses of mg/kg/inj cocaine in combination with mg/kg/inj heroin was decreased significantly during buprenorphine treatment (P.05). These effects of buprenorphine were selective for speedball self-administration, because food-maintained responding was not decreased in comparison with the saline control treatment condition. Food-maintained responding decreased at the highest doses of speedball self-administration during saline control treatment. However, during buprenorphine treatment, food-maintained responding was higher at the highest speedball dose than during saline-control treatment in all monkeys (P ). When an intermediate dose of cocaine (0.01 mg/kg/inj) was combined with heroin, buprenorphine also shifted the speedball dose-effect curve to the right (fig. 4). Some speedball doses maintained higher rates of drug self-administration during buprenorphine treatment than during saline treatment conditions, but these differences were statistically significant only in monkey R800 (P.05) (fig. 4, row 2). During saline control treatment, food-maintained responding decreased at the highest speedball doses in monkeys and R800. During buprenorphine treatment, levels of food-maintained responding were lower than during saline-control treatment, but these differences were not statistically significant except in monkey at one dose (P.01). In monkey R800, at the highest speedball dose, food-maintained

7 450 Mello and Negus Vol. 285 Fig. 4. Effects of buprenorphine on speedball dose-effect curves for individual monkeys. Dose-effect curves for a medium dose of cocaine (0.01 mg/kg/inj) in combination with one of three or four doses of heroin ( mg/kg/inj) are shown for individual monkeys (left panel). All other details are the same as described in the legend for figure 3. Fig. 5. Effects of buprenorphine on speedball dose-effect curves for individual monkeys. Dose-effect curves for a high dose of cocaine (0.10 mg/kg/inj) in combination with one of three or four doses of heroin ( mg/kg/inj) are shown for individual monkeys (left panel). All other details are the same as described in the legend for figure 3. responding was significantly higher during buprenorphine treatment than during saline control treatment (P.05). Figure 5 shows the effects of saline and buprenorphine (0.237 mg/kg/day) treatment on individual dose-effect curves for speedballs consisting of a high dose of cocaine (0.10 mg/kg/inj) in combination with heroin. Buprenorphine treatment produced a nonsignificant down shift in the speedball dose-effect curve self-administration in monkey (P.0599) and a right shift in monkey 89B057. Buprenorphine was associated with higher speedballmaintained responding at high speedball doses but these differences were statistically significant in only one instance (monkey 12780; P.05). Food self-administration also was decreased at some speedball doses, but these effects were not consistently speedball dose-dependent. Food-maintained responding was sometimes higher during buprenorphine treatment than during saline treatment at some speedball doses, but these differences were not statistically significant (fig. 5). Figure 6 shows the effects of daily buprenorphine administration (0.237 mg/kg/day) on cocaine and heroin combinations in individual monkeys during 10 days of observation. Each heroin/cocaine dose combination includes a medium dose of heroin (0.001 or mg/kg/inj) in combination with a low (0.001 mg/kg/inj), medium (0.01 mg/ kg/inj) or high (0.10 mg/kg/inj) dose of cocaine. These are the same speedball combinations examined during treatment with lower (0.075 mg/kg/day) and higher (0.75 mg/ kg/day) doses of buprenorphine and shown earlier in figure 2. At the lowest dose of cocaine (0.001 mg/kg/inj) in combination with a medium dose of heroin, buprenorphine (0.237 mg/kg/day) reduced speedball self-administration within 1 or 2 days in two of the three monkeys, and this effect persisted across the 10 days of treatment (fig. 6, left panel). Buprenorphine did not reduce speedball self-administration until after 5 days of treatment in monkey The effects of buprenorphine on food-maintained responding were minimal and transient under these conditions. At a medium dose of cocaine (0.01 mg/kg/inj) in combination with heroin, buprenorphine had only transient effects on speedball self-administration in one monkey (R800), and had little or no effect on speedball-maintained responding in two monkeys (606.5 and 89B057) (fig. 6, middle panel). Food-maintained responding was decreased more than speedball-maintained responding at several time points. The effects of buprenorphine treatment on speedball combinations consisting of the highest dose of cocaine (0.10 mg/kg/inj) and heroin during 10 days of observation are shown in the right panel of figure 6. Buprenorphine treatment resulted in a sustained and selective decrease in speedball self-administration in one monkey (606.5). Similar effects were observed during the first 3 days of treatment in the other two monkeys, but these effects were not sustained throughout the 10 days of observation. Food-maintained responding was variable and no sustained increases or decreases were observed. Study 3: Effects of Buprenorphine on Self-Administration of Heroin Alone The dose-effect curve for the self-administration of heroin alone ( mg/kg/inj) had an inverted-u shape, and peak rates of drug self-administration were maintained by unit doses of to mg/kg/inj heroin during saline control treatment (fig. 7, left panel). Figure 7 shows the effects of treatment with buprenorphine (0.075 and mg/kg/day) on the heroin self-administration dose-effect curves. Both doses of buprenorphine shifted the heroin dose-effect curve downward (P ) and to the right. This effect was greatest during treatment with the higher dose of buprenorphine at heroin doses of mg/kg/inj in monkeys and R800 (P ) and at heroin doses of 0.01 mg/kg/inj in monkeys and 89B057 (P ). Heroin self-administration was associated with a dose-dependent decrease in food-maintained responding during saline control treatment (fig. 7,

8 1998 Buprenorphine and Speedball Abuse 451 Fig. 6. Daily effects of buprenorphine (0.237 mg/kg/day) on the self-administration of speedball combinations of cocaine (0.001, 0.01 or 0.10 mg/kg/inj) and heroin (0.001 or mg/kg/inj) and food in individual monkeys. Abscissae: Consecutive days of treatment with buprenorphine (0.237 mg/kg/day). Left ordinates: number of speedball injections per day (filled circles). Right ordinates: number of food pellets per day (open circles). Each panel shows the results of a single 10-day period of buprenorphine treatment in one monkey. Points above Sal in each panel show the mean number of injections/day and food pellets/day during saline treatment baseline control conditions [x S.E.M.]. right panel). Food-maintained responding also decreased at high doses of heroin self-administration during buprenorphine treatment. However, monkeys and R800 sometimes acquired more food pellets during buprenorphine treatment than during saline control treatment, but these differences were significant at only one heroin dose in one monkey (P.05). Figure 8 shows the effects of 10 days of saline treatment (x S.E.) and the effects of each dose of buprenorphine on daily heroin and food self-administration during 10 consecutive days. Each dose of heroin was near the peak of the dose-effect curve during saline control treatment. It is apparent that both doses of buprenorphine produced rapid and sustained decreases in heroin self-administration with minimal effects on food self-administration. These data are consistent with the interpretation that buprenorphine selectively reduced heroin self-administration in these monkeys. Time Course of the Effect of Buprenorphine on Speedball and Heroin Self-Administration The patterns of speedball and heroin-maintained responding across the four daily sessions were examined. Buprenorphine-induced decreases in heroin or speedball self-administration were distributed evenly across all four daily sessions. Decreases in drug-maintained responding were not restricted to the sessions that occurred in closest temporal proximity to buprenorphine administration. These data are consistent with buprenorphine s long duration of action (more than 24 hr) previously reported in clinical (Jasinski et al., 1978; Mello and Mendelson, 1980, Mello et al., 1982) and preclinical investigations (Mello et al., 1990, 1993a). Discussion Effects of Buprenorphine on Speedball Self-Administration This is the first evaluation of the effects of buprenorphine, a mixed opioid agonist-antagonist, on speedball self-administration by rhesus monkeys. Buprenorphine was selected for study because it reduced cocaine self-administration and opioid self-administration in both clinical and preclinical studies (see reviews by Bickel and Amass, 1995; Fudala and Johnson, 1995; Mello and Mendelson, 1995; Mello et al., 1993c). In the present study, the effects of buprenorphine on speedball self-administration depended both on the dose of buprenorphine and on the unit doses of heroin and cocaine in the speedball mixture. When speedballs consisted of a low dose of cocaine combined with heroin, buprenorphine (0.237 mg/kg/ day) consistently shifted the speedball dose-effect curve downward and to the right (fig. 3). Both lower (0.075 mg/kg/ day) and higher (0.75 mg/kg/day) doses of buprenorphine also significantly decreased responding maintained by a speedball combination consisting of a low dose of cocaine and a medium dose of heroin (fig. 2). However, when higher doses of cocaine were combined with heroin, the effects of buprenorphine on speedball-maintained responding were more variable. Buprenorphine (0.237 mg/kg/day) tended to shift the speedball dose-effect curve to the right and/or downward in five of six determinations (figs. 4 and 5). Similarly, lower and higher doses of buprenorphine significantly decreased responding maintained by a medium dose of heroin combined with a high dose of cocaine, but not by a medium dose of cocaine (fig. 2). Thus the effects of buprenorphine on speedball self-administration were related inversely to the cocaine dose-effect curve shown in figure 1. Buprenorphine was least effective in reducing self-administration of speedball combi-