Abstract
μ-Opioid agonists decrease cocaine self-administration in laboratory studies and cocaine use by many cocaine- and opioid-dependent polydrug abusers. To assess the role of μ-agonist efficacy as a determinant of these effects, this study evaluated cocaine- and food-maintained responding by rhesus monkeys (Macaca mulatta) during chronic treatment with saline or the high-efficacy μ-agonist fentanyl (0.001–0.01 mg/kg/h), the intermediate-efficacy μ-agonist morphine (0.032–0.32 mg/kg/h), or the low-efficacy μ-agonists nalbuphine (0.1–1.0 mg/kg/h) and butorphanol (0.0032–0.032 mg/kg/h). Responding was maintained by cocaine and food under a second order schedule of reinforcement during multiple daily sessions of cocaine and food availability. Saline and each opioid dose were administered continuously for 7 consecutive days during availability of each cocaine dose. All four μ-agonists produced dose-dependent and sustained decreases in cocaine self-administration across a range of cocaine doses (0.0032–0.1 mg/kg/injection). Nalbuphine and butorphanol produced the greatest decreases in cocaine self-administration and the smallest effects on food-maintained responding. Morphine and fentanyl produced smaller decreases in cocaine self-administration, and undesirable effects precluded evaluation of higher fentanyl and morphine doses. Decreases in cocaine self-administration produced by nalbuphine and butorphanol probably did not reflect a general blockade of cocaine's abuse-related effects, because nalbuphine and butorphanol did not block the discriminative stimulus effects of cocaine in monkeys trained to discriminate 0.4 mg/kg cocaine from saline in a food-reinforced drug discrimination procedure. These results suggest that low-efficacy μ-agonists may decrease cocaine self-administration to a greater degree and with fewer undesirable effects than high-efficacy μ-agonists.
Cocaine abuse continues to be a significant public health problem, and no consistently effective pharmacotherapy has been identified (Mendelson and Mello, 1996). One approach to the evaluation of candidate medications for cocaine abuse has been to examine medication effects on cocaine-maintained responding in drug self-administration procedures. We have proposed that a promising medication might be one that produced dose-dependent and sustained decreases in cocaine self-administration across a broad range of cocaine doses while producing minimal evidence of toxicity (Mello and Negus, 1996). The mixed action opioid buprenorphine appears to approach these criteria. In rhesus monkeys (Macaca mulatta), for example, chronic treatment with buprenorphine decreased cocaine self-administration for up to 120 days while producing smaller changes in responding maintained by food presentation and little other evidence of toxicity in the same monkeys (Mello et al., 1989; for review, see Mello and Mendelson, 1995). The ability of buprenorphine to decrease cocaine self-administration without major toxicity has been confirmed and extended in studies with rodents (Carroll and Lac, 1992), nonhuman primates (Carroll et al., 1992; Winger et al., 1992), and humans (Foltin and Fischman, 1994).
The basis for buprenorphine's effects on cocaine self-administration is not well understood. Buprenorphine is a mixed action opioid that binds with high affinity to the three major types of opioid receptors, the μ-, κ-, and δ-receptors (Rothman et al., 1995). Buprenorphine has intermediate efficacy at μ-receptors and acts as a partial μ-agonist, and it has very low efficacy at κ- and δ-receptors and acts primarily as a κ- and δ-antagonist (Cowan, 1995; Dykstra and Negus, 1995; S. S. Negus, unpublished observations). Current evidence suggests that the μ-agonist effects of buprenorphine are most important for its effects on cocaine self-administration. Buprenorphine-induced decreases in cocaine self-administration were blocked by the μ-selective antagonist naltrexone (Mello et al., 1993b), and cocaine self-administration was decreased by acute administration of other μ-agonists (Winger et al., 1992). It is also interesting to note that maintenance on both buprenorphine and the more selective μ-agonist methadone has often been reported to decrease cocaine use by polydrug abusers (Kosten et al., 1989; Gastfriend et al., 1993; Strain et al., 1994; Schottenfeld et al., 1997; Borg et al., 1999). In contrast, the κ- and δ-receptor-mediated effects of buprenorphine appear to be less important, because neither κ- nor δ-selective antagonists produced consistent decreases in cocaine self-administration in rhesus monkeys (Negus et al., 1995; Mello and Negus, 2000).
The intermediate efficacy of buprenorphine at μ-receptors may contribute to its ability to decrease cocaine self-administration without producing other severe adverse effects. Efficacy can be defined as the ability of a drug to activate transduction mechanisms associated with its receptor (Kenakin, 1993), and the influence of efficacy on μ-receptor-mediated behavioral and physiological effects has been extensively studied (Picker and Dykstra, 1989; Gerak et al., 1994;Butelman et al., 1995; Gatch et al., 1995; Negus and Mello, 1999; Cook et al., 2000). It is now well established that high-efficacy μ-agonists produce a broad range of effects (including antinociception, respiratory depression, discriminative stimulus effects, and changes in schedule-controlled responding) within a relatively narrow dose range. Lower efficacy agonists, in contrast, produce some agonist effects but not others, and there can be substantial differences in either the potency or the maximal effects of a lower efficacy agonist across different experimental endpoints.
Accordingly, the present study examined the hypothesis that relative efficacy at the μ-receptor might be an important determinant of the magnitude and selectivity of μ-opioid agonist effects on cocaine self-administration. Specifically, we postulated that opioids with low-to-intermediate efficacy at μ-receptors would decrease cocaine self-administration more selectively than high-efficacy μ-agonists. To test this hypothesis, we selected four opioid agonist analgesics that differ in relative efficacy at the μ-receptor, and that produce behavioral effects mediated primarily by μ-opioid receptors in rhesus monkeys (Gerak et al., 1994; Butelman et al., 1995; Gatch et al., 1995;Emmerson et al., 1996). Specifically, fentanyl has relatively high efficacy, morphine has intermediate efficacy, and butorphanol and nalbuphine have low efficacy at μ-opioid receptors. It should be noted, however, that these drugs also differ in their relative selectivity for μ-receptors, and nalbuphine and butorphanol in particular bind with relatively low selectivity to both μ- and κ-opioid receptors and may produce some effects in monkeys that are mediated by κ-opioid receptors (Vivian et al., 1999). The effects of continuous treatment for 7 consecutive days with all four opioids were examined on behavior maintained under identical second order schedules by 1-g food pellets and by a range of cocaine doses. Parallel studies examined the effects nalbuphine and butorphanol on the discriminative stimulus effects of cocaine in rhesus monkeys. These studies were conducted to assess the degree to which selective decreases in cocaine self-administration produced by nalbuphine and butorphanol could be attributed to a selective blockade of the abuse-related effects of cocaine. For comparison, we have shown previously that fentanyl and morphine produced cocaine-like discriminative stimulus effects in some monkeys, and these high-efficacy μ-agonists also produced additive effects with cocaine when they were administered in combination with cocaine (Negus et al., 1998).
Materials and Methods
Subjects
A total of nine male and three female rhesus monkeys were subjects in drug self-administration studies, and seven male monkeys were subjects in drug discrimination studies. All monkeys had an experimental history involving the evaluation of dopaminergic and/or opioid compounds in assays of drug self-administration or drug discrimination. Monkeys weighed 5 to 12 kg and were maintained on a diet of multiple vitamins, fresh fruit, and Lab Diet Jumbo Monkey biscuits (PMI Feeds, Inc., St. Louis, MO). In addition, monkeys could receive 1-g banana flavored pellets (Precision Primate Pellets Formula L/I Banana Flavor; P.J. Noyes Company, Inc., Lancaster, NH) during daily operant sessions (see below). Water was continuously available. A 12-h light/dark cycle was in effect (lights on from 7:00 AM to 7:00 PM).
Animal maintenance and research were conducted in accordance with the guidelines provided by the National Institutes of Health Committee on Laboratory Animal Resources. The facility was licensed by the United States Department of Agriculture, and protocols were approved by the Institutional Animal Care and Use Committee. The health of the monkeys was periodically monitored by consulting veterinarians. Monkeys had visual, auditory, and olfactory contact with other monkeys throughout the study. Operant procedures and foraging toys provided opportunities for environmental manipulation and enrichment. Music or nature video tapes were also played daily in animal housing rooms to provide additional environmental enrichment.
Cocaine Self-Administration Procedures
Apparatus.
Each monkey was housed individually in a well ventilated stainless steel chamber (64 × 64 × 79 cm) equipped with a custom-designed operant panel (28 × 28 cm) mounted on the front wall. Three square translucent response keys (6.4 × 6.4 cm) were arranged 2.54 cm apart in a horizontal row 3.2 cm from the top of the operant panel. Each key could be transilluminated by red or green stimulus lights (Superbright LEDs; Fairchild Semiconductor, San Jose, CA). In addition, three circular translucent panels (1.9 cm in diameter) were located in a vertical column below the center response key and could be transilluminated by red or green stimulus lights. Each housing chamber was also equipped with a pellet dispenser (model G5210; Gerbrands, Arlington, MA) and two syringe pumps (model B5P-lE; Braintree Scientific, Braintree, MA; or model 980210; Harvard Apparatus, South Natick, MA), one for each lumen of the double-lumen catheter. Operation of the operant panels and data collection were accomplished with microprocessors and software purchased from MED Associates (Georgia, VT).
Surgical Procedures.
Double-lumen silicone rubber catheters (inside diameter 0.028 in.; outside diameter 0.088 in.; Saint Gobain Performance Plastics, Beaverton, MI) were implanted in the internal jugular or femoral vein and exited in the midscapular region. All surgical procedures were performed under aseptic conditions. Monkeys were initially sedated with ketamine (5–10 mg/kg), and anesthesia was induced with sodium thiopental (10 mg/kg i.v.). Atropine (0.05 mg/kg) was administered i.m. or s.c. to reduce salivation. After insertion of an endotracheal tube, anesthesia was maintained with isoflurane (1–2% in oxygen). After surgery, monkeys were given the antibiotic procaine penicillin G (300,000 units/kg/day i.m.) for 5 days. An analgesic dose of buprenorphine (0.032 mg/kg i.m.) was administered twice daily 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, Malone, NY). This flexible tether system permitted monkeys to move freely. Catheter patency was periodically evaluated by i.v. administration of ketamine (5 mg/kg) or the short-acting barbiturate methohexital (3 mg/kg) through the catheter lumen. The catheter was considered to be patent if i.v. administration of ketamine or methohexital produced a loss of muscle tone within 10 s.
Behavioral Procedures.
Procedures for the evaluation of cocaine- and food-maintained responding were identical to those used in our previous studies of the effects of κ-opioids on cocaine self-administration (Mello and Negus, 2000). The conditions of food and cocaine availability were associated with different colored stimulus lights projected on the center response key of the operant response panel. The two side keys were not transilluminated during the food and drug self-administration studies, and responding on these keys had no scheduled consequences. During sessions of food availability, the center key was transilluminated with a red stimulus light, whereas during sessions of cocaine availability, the center key was transilluminated with a green stimulus light. Four food sessions and four drug sessions were conducted during each experimental day. Food sessions began at 6:00 AM, 11:00 AM, 3:00 PM, and 7:00 PM, and drug sessions began at 7:00 AM, 12:00 PM, 4:00 PM, and 8:00 PM. At all other times, responding had no scheduled consequences. The experimental room was dark during all food and drug sessions. Each food and drug session lasted 1 h or until 25 food pellets or 20 injections had been delivered, whichever occurred first. Monkeys could earn a maximum of 100 food pellets per day and 80 injections per day.
Training Procedures.
After initial shaping of key pressing by using food reinforcement, responding was maintained on a variable ratio (VR) schedule that was gradually increased from a VR1 to a VR16. Under this schedule, the response requirement for individual ratios varied randomly across a range of ±25%, and under the terminal VR16 schedule, the response requirement for individual ratios varied from 12 to 20. Completion of the VR response requirement resulted in the delivery of a single food pellet. In addition, completion of the VR response requirement resulted in 1) the illumination of a red stimulus light for 1 s below the center response key, and 2) the initiation of a 10-s time-out period, during which the stimulus light illuminating the center response key was turned off and responding had no scheduled consequences. After monkeys received at least 50 food pellets per day for three consecutive days under the VR16 schedule, behavior was maintained on a second order schedule that consisted of two components, a VR and fixed ratio (FR). Under this schedule, completion of each VR schedule component resulted in the illumination of the red stimulus for 1 s below the center response key and initiation of the 10-s time-out, and this is designated as [VR:S]. However, a food pellet was delivered only after completion of a fixed number (or fixed ratio) of VR components. The FR component was gradually increased to 2 or 4, so that monkeys had to complete two or four VR components for each food pellet. The terminal second order schedule response requirements were designated as FR2 [VR 16:S] or FR4 [VR 16:S]. The final response requirement averaged 32 (range 26–39) or 64 (range 53–78) responses.
Once monkeys received at least 50 food pellets per day for at least three consecutive days under the terminal second order schedule, the intravenous double-lumen catheter was implanted as described above. After recovery from surgery for at least 1 week, key pressing maintained by drug reinforcement (0.032 mg/kg/inj i.v. cocaine injections) was shaped under a series of increasing variable ratios identical to those used during training of food-maintained responding. Completion of the response requirement resulted in the delivery of 0.1 ml of cocaine solution over 1 s through one lumen of the double lumen catheter. In addition, completion of the response requirement during drug sessions resulted in the illumination of a green stimulus light for 1 s below the center response key and the initiation of a 10-s time-out. For each monkey, the final second order schedule response requirement was identical for food- and drug-maintained responding (FR2 [VR16:S] or FR4 [VR16:S]).
Monkeys were trained until they met the following criteria for stable food and cocaine self-administration under the terminal schedule: 1) three consecutive days during which the number of drug injections/day differed by no more than 20% from the mean number of drug injections/day during those 3 days and there was no upward or downward trend; and 2) during the same three consecutive days, the mean number of both drug injections per day and food pellets per day was greater than 50.
Testing Procedures.
The effects of repeated treatments with saline and the high-efficacy μ-agonist fentanyl, the intermediate-efficacy μ-agonist morphine, and the low-efficacy μ-agonists nalbuphine and butorphanol were studied. Saline and each dose of each opioid agonist were examined for 7 consecutive test days. Saline and opioids were administered by i.v. injection through the second lumen of the double-lumen catheter. Injections were delivered every 20 min from 10:30 AM each day until 9:30 AM the next morning for a total of 3 inj/h and 69 inj/day. No injections were delivered between 9:30 AM and 10:30 AM, and during this period, monkeys received their morning ration of food, and their health status was evaluated by the technical staff. At the conclusion of each 7-day test period with an opioid agonist, the maintenance dose of cocaine (0.032 mg/kg/inj) and saline control treatment were reinstated for a period of at least 4 days and until the number of reinforcers per day maintained by cocaine and food returned to baseline levels. This interval between successive treatments was designed to reduce the possibility of carryover effects from one treatment condition to the next.
Studies with each μ-agonist were conducted in two phases. The first phase of these experiments compared the effects of treatment with saline, fentanyl (0.001–0.018 mg/kg/h), morphine (0.032–0.32 mg/kg/h), nalbuphine (0.1–1.0 mg/kg/h), or butorphanol (0.0032–0.032 mg/kg/h) on food- and cocaine-maintained responding in groups of four monkeys during availability of 0.01 mg/kg/inj cocaine. These doses of μ-agonists were based on preliminary dose-ranging studies and on the relative potencies of fentanyl, morphine, nalbuphine, and butorphanol in producing other behavioral effects in rhesus monkeys (e.g., antinociception; Gatch et al., 1995). A unit dose of 0.01 mg/kg/injection cocaine was used for the initial studies because it was the lowest dose to reliably maintain high rates of cocaine self-administration in all monkeys, and because previous studies have shown that behavior maintained by this unit dose of cocaine is especially sensitive to the effects of pretreatment compounds (Mello and Negus, 2000). Data for 7 days of treatment with the highest dose of 0.018 mg/kg/h fentanyl were obtained in only two monkeys, and as a result these data were not included in statistical analysis. This dose of fentanyl was not studied for the full 7 days in the other two monkeys due to concerns about the toxic effects of high-dose fentanyl treatment (see Results). Opioids and opioid doses were tested in an irregular order across monkeys.
The second phase of these experiments examined the effects of treatment with saline or fentanyl (0.0032 and 0.01 mg/kg/h), morphine (0.1 and 0.32 mg/kg/h), nalbuphine (0.1 and 0.32 mg/kg/h), or butorphanol (0.01 and 0.032 mg/kg/h) on responding maintained by a broader range of cocaine doses (0.0032–0.1 mg/kg/inj) in groups of three monkeys. Evaluation of saline and each dose of opioid during availability of each dose of cocaine was conducted during a block of 7 consecutive test days as described above. Treatment conditions were studied in an irregular order across monkeys. One monkey exhausted its catheter sites before studies with 0.0032 mg/kg/h fentanyl were completed, so data for these combinations were obtained in only two monkeys.
Data Analysis.
The total numbers of injections or food pellets delivered per day were determined. Data for the first series of experiments were evaluated using a two-factor analysis of variance, with opioid dose and treatment day as the two factors. For the second series of experiments, data are expressed as the mean (±S.E.M.) numbers of injections per day and pellets per day for the 7 days of each test condition. Data were evaluated using a two-factor analysis of variance with cocaine dose and opioid dose as the two factors. A significant analysis of variance was followed by individual means comparison using Duncan's post hoc test. The criterion for significance was set at p < 0.05.
Cocaine Discrimination Procedures
Apparatus.
Each monkey was housed individually in a well ventilated, stainless steel chamber (56 × 71 × 69 cm) equipped with an operant panel identical to the one described above. In addition, the response panel supported an externally mounted pellet dispenser (model G5310; Gerbrands) that delivered 1-g fruit-flavored food pellets (P.J. Noyes Company, Inc.) to a food receptacle mounted on the cage beneath the response panel. Operation of the response panels and data collection were accomplished with IBM-compatible computers and interface systems (MED Associates) located in a separate room.
Training Procedures.
Drug discrimination procedures were identical to those used in our previous studies of the effects of opioids on cocaine discrimination (Negus et al., 1998). Discrimination sessions consisted of multiple cycles and were conducted 5 days/week. Each cycle consisted of a 15-min time-out period followed by a 5-min response period. During the time-out, all stimulus lights were off and responding had no scheduled consequences. During the response period, the right and left response keys were transilluminated red or green, and monkeys could receive up to 10 food pellets by responding under an FR30 schedule of food presentation. For three of the seven monkeys, the left key was illuminated green and the right key was illuminated red. For the other four monkeys, the colors of the response keys were reversed. The center key was not illuminated at any time, and responding on the center key had no scheduled consequences. If all available food pellets were delivered before the end of the 5-min response period, the stimulus lights transilluminating the response keys were turned off, and responding had no scheduled consequences for the remainder of that response period.
On training days, monkeys were given an i.m. injection of either saline or 0.40 mg/kg cocaine 5 min after the beginning of each time-out period (i.e., 10 min before the response period). After administration of saline, responding on only the green key (the saline-appropriate key) produced food, whereas after administration of 0.40 mg/kg cocaine, only responding on the red key (the drug-appropriate key) produced food. Responses on the inappropriate key reset the FR requirement. Daily sessions consisted of one to five cycles, and if the training dose of cocaine was administered, it was administered only during the last cycle.
During the response period of each cycle, three dependent variables were determined: 1) Percentage of injection-appropriate responding before delivery of the first reinforcer [(injection-appropriate responses emitted before first reinforcer ÷ total responses emitted before first reinforcer) × 100]; 2) percentage of injection-appropriate responding for the entire response period [(injection-appropriate responses emitted during response period ÷ total responses emitted during response period) × 100]; and 3) response rate (total responses emitted during response period ÷ total time stimulus lights were illuminated).
Monkeys were considered to have acquired cocaine discrimination when the following three criteria were met for seven of eight consecutive training sessions: 1) the percentage of injection-appropriate responding before delivery of the first reinforcer was greater than or equal to 80% for all cycles; 2) the percentage of injection-appropriate responding for the entire cycle was greater than or equal to 90% for all cycles; and 3) at least one pellet was earned during all training cycles.
Testing Procedures.
Once monkeys met criterion levels of cocaine discrimination, testing began. Test sessions were identical to training sessions except that responding on either key produced food, and cocaine, nalbuphine, or butorphanol was administered as described below. Two series of experiments were conducted to characterize the effects of nalbuphine and butorphanol administered alone or as pretreatments to cocaine. The effects of fentanyl and morphine on cocaine discrimination in this procedure have been described previously (Negus et al., 1998).
In the first series of experiments, cumulative doses of nalbuphine (0.0032–3.2 mg/kg) or butorphanol (0.00032–0.032 mg/kg) were administered at the beginning of each time-out period, instead of either saline or the training dose of cocaine. Dose-effect curves were determined twice in each monkey, and each dose increased the total dose by one-half log unit. Both opioids were tested up to doses that eliminated responding in most monkeys. In the second series of experiments, the effects of opioid pretreatment on cocaine discrimination were determined. Cumulative cocaine dose-effect curves were determined either without pretreatment or after pretreatment with nalbuphine (0.032–0.32 mg/kg) or butorphanol (0.0032–0.032 mg/kg). Nalbuphine and butorphanol pretreatments were administered 15 min before the first cumulative dose of cocaine. Control cocaine dose-effect curves were determined twice in each monkey, and each opioid pretreatment dose was studied once in each monkey.
Test sessions were conducted only if the three criteria listed above under “Training Procedures” were met during the training day immediately preceding the test day. If responding did not meet criterion levels of discrimination performance then training was continued until criterion levels of performance were obtained for at least two consecutive days. In general, testing was conducted on Tuesdays and Fridays, and training sessions were conducted on Mondays, Wednesdays, and Thursdays.
Data Analysis.
Graphs for percentage of cocaine-appropriate responding and response rates were plotted as a function of the cumulative dose of test compound (log scale). The percentage of cocaine-appropriate responding for a given cycle was calculated and reported only if the monkey emitted enough responses to earn at least one food pellet (i.e., 30 response, equivalent to a response rate of 0.1 responses/s). In group graphs, data for percentage of cocaine-appropriate responding are plotted only if at least four of the seven monkeys met the response rate criterion.
For substitution tests, nalbuphine and butorphanol were each tested twice in each monkey, and data from multiple determinations were averaged. A test drug was considered to substitute for cocaine if it produced ≥90% cocaine-appropriate responding. Pretreatment studies to examine the effects of each dose of nalbuphine and butorphanol in combination with cocaine were conducted once in each monkey. For each test, an ED50 value for cocaine discrimination was defined as the dose of cocaine that produced 50% cocaine-appropriate responding. ED50 values were calculated by interpolation when only two data points were available (one below and one above 50% cocaine-appropriate responding) or by linear regression when at least three data points were available on the linear portion of the dose-effect curve. Individual ED50 values were averaged to yield a mean ED50 value (±95% confidence limits). In some pretreatment tests, all data points for cocaine-appropriate responding fell above 50%. For the purposes of analysis, ED50 values in these cases were estimated by assuming that the next lowest half log dose would have produced 0% cocaine-appropriate responding. In other tests, the pretreatment drug decreased responding in combination with most or all doses of cocaine, and in these cases, an ED50 value was not calculated.
In addition, ED25 values for rate suppression were defined as the dose of cocaine that decreased response rates to 25% of saline control values. Control values, in turn, were defined as the average response rate during saline training cycles on the training day preceding each test day. ED25 values were used to analyze response rate data, because high doses of nalbuphine or butorphanol alone often decreased response rates to less than 50% of control. ED25 values were calculated by interpolation when only two data points were available (one below and one above 25% control response rate) or by linear regression when at least three data points were available on the linear portion of the dose-effect curve. Individual ED25 values were averaged to yield a mean ED25 value (±95% confidence limits).
Because drug doses were incremented on a logarithmic scale, ED50 and ED25 values were converted to their log values for calculation of mean and confidence limits and for statistical analysis. Mean ED50and ED25 values and confidence limits were converted back to their linear values for presentation in Table1. For pretreatment tests, the pretreatment drug was considered to significantly alter the cocaine ED50 or ED25 values if the 95% confidence limits for cocaine alone did not overlap with the 95% confidence limits for cocaine after pretreatment.
Drugs
Cocaine HCl, fentanyl HCl, and morphine sulfate was obtained from the National Institute on Drug Abuse (Bethesda, MD) and were dissolved in sterile water. Nalbuphine HCl was purchased from Sigma/RBI (Natick, MA) and was also dissolved in sterile water. Butorphanol tartrate was used as the commercially available Torbutrol solution (Fort Dodge Animal Health, Overland Park, KS). All drug solutions were filter-sterilized using a 0.22-μm Millipore filter and stored in pyrogen-free vials. Doses were calculated using the salt forms of the drugs given above.
Results
Dose Dependence of μ-Agonist Effects on Cocaine- and Food-Maintained Responding.
Figure 1shows the mean effects of treatment with saline and the opioid agonists fentanyl, morphine, nalbuphine, and butorphanol on responding maintained by 0.01 mg/kg/inj cocaine and food pellets. During saline treatment, monkeys responded for approximately 70 to 80 cocaine injections per day and for 80 to 100 food pellets per day. All four opioid agonists produced a dose-dependent and significant decrease in cocaine self-administration. However, the selectivity of these effects depended on the efficacy of the μ-agonist.
Treatment with the high-efficacy μ-agonist fentanyl (0.01 mg/kg/h) produced a significant decrease in cocaine self-administration to 38.8 ± 10.5 inj/day. This dose of fentanyl did not significantly alter food-maintained responding (p = 0.39), although responding for food was decreased in some monkeys. A higher dose of fentanyl (0.018 mg/kg/h) was evaluated in three monkeys (data not shown). In two of these monkeys, 0.018 mg/kg/h fentanyl was evaluated for the entire 7 days, and this dose of fentanyl further decreased cocaine self-administration while having little affect on food-maintained responding. However, in the third monkey, 0.018 mg/kg/h fentanyl nearly eliminated all responding and produced profound sedation. As a result, the treatment was terminated in this monkey, and 0.018 mg/kg/h fentanyl was not examined in the fourth monkey.
Treatment with the intermediate-efficacy μ-agonist morphine decreased cocaine self-administration to 42.6 ± 8.2 and 24.6 ± 5.0 inj/day at doses of 0.1 and 0.32 mg/kg/h, respectively. However, the higher dose of morphine also significantly decreased levels of food-maintained responding to 48.5 ± 20.3 pellets/day.
All doses of the low efficacy μ-agonist nalbuphine decreased cocaine self-administration, and the highest dose of 1.0 mg/kg/h nalbuphine decreased cocaine self-administration to 17.7 ± 8.3 inj/day. Moreover, none of the nalbuphine doses significantly altered food-maintained responding. Thus, nalbuphine produced robust and selective decreases in the self-administration of 0.01 mg/kg/inj cocaine.
Butorphanol also produced robust and selective decreases in cocaine self-administration. The highest dose of 0.032 mg/kg/h butorphanol decreased cocaine self-administration to 7.5 ± 1.7 inj/day but did not produce a significant change in food-maintained responding.
Time Course of μ-Agonist Effects on Cocaine- and Food-Maintained Responding.
Figure 2 shows mean data for each of the 7 days of treatment with saline and with the highest dose of each of the μ-agonists. Monkeys responded for most of the food pellets and 0.01 mg/kg/inj cocaine injections throughout the 7 days of saline treatment. Cocaine self-administration was decreased during treatment with each of the μ-agonists. In general, peak decreases in cocaine self-administration were apparent by the 2nd day, and these decreases were sustained throughout the 7-day treatment period. The higher efficacy μ-agonists fentanyl and morphine also tended to decrease food-maintained responding, but the lower efficacy μ-agonists nalbuphine and butorphanol produced selective effects on cocaine self-administration with little or no change in food-maintained responding throughout the 7-day treatment period.
Figure 2 also shows recovery of cocaine- and food-maintained responding after the termination of treatment with the highest doses of each μ-agonist. During these post-treatment periods, baseline conditions of 0.032 mg/kg/inj cocaine availability and saline treatment were reinstated for at least 5 days. In general, cocaine self-administration recovered rapidly within 1 or 2 days when μ-agonist treatment was terminated. In contrast, rates of food-maintained responding initially decreased to low levels after treatment with morphine and nalbuphine before recovering toward baseline levels. Smaller decreases in food-maintained responding were observed after treatment with 0.01 mg/kg/h fentanyl; however, as noted above, two monkeys were treated for 7 days with a higher dose of 0.018 mg/kg/h fentanyl. After this treatment, food-maintained responding decreased in both monkeys, and in one monkey, food-maintained responding was almost eliminated on the 2nd day after treatment. This monkey also vomited on the 2nd day after fentanyl treatment. In comparison with the other μ-agonists, termination of high-dose butorphanol treatment produced relatively small changes in rates of food-maintained responding.
μ-Agonist Effects on Cocaine Self-Administration Dose-Effect Curve.
The left panels of Figs. 3(fentanyl and morphine) and 4 (nalbuphine and butorphanol) show cocaine self-administration dose-effect curves during saline treatment and during treatment with the μ-agonists. Data for responding maintained by food are shown in the right panels of Figs. 3 and 4. During saline treatment, monkeys responded at low levels for saline injections (points over Sal). Cocaine (0.001–0.1 mg/kg/inj) dose dependently maintained drug self-administration responding. The cocaine dose-effect curve had an inverted U-shape, and peak levels of cocaine self-administration occurred at unit doses of 0.01 to 0.032 mg/kg/inj. Monkeys responded at high, stable levels for food during availability of saline and cocaine doses up to 0.032 mg/kg/inj. Rates of food-maintained responding decreased to varying degrees across monkeys during availability of the high dose of 0.1 mg/kg/inj cocaine.
Fentanyl, morphine, nalbuphine and butorphanol all produced dose-dependent downward shifts in the cocaine self-administration dose-effect curves. For the highest dose of fentanyl (0.01 mg/kg/h), this tendency to decrease cocaine self-administration was significant only during availability of 0.032 mg/kg/inj cocaine. Fentanyl also tended to decrease rates of food-maintained responding, and this effect was significant during availability of 0.0032 mg/kg/inj cocaine. Higher doses of fentanyl were not tested in cocaine dose-effect studies due to the toxic effect of high-dose fentanyl treatment described above.
The low dose of morphine (0.1 mg/kg/h) significantly decreased responding maintained by low doses of 0.0032 and 0.01 mg/kg/inj cocaine but not by a higher unit dose of 0.032 mg/kg/inj cocaine. This dose of morphine did not alter food-maintained responding. A higher dose of 0.32 mg/kg/h morphine decreased cocaine self-administration across a broader range of cocaine doses, but 0.32 mg/kg/h morphine also significantly decreased food-maintained responding. Treatment with 0.32 mg/kg/h morphine did not significantly alter cocaine- or food-maintained responding during availability of a high dose of 0.1 mg/kg/inj cocaine, but this test was conducted in only two monkeys, and mean levels of cocaine- and food-maintained responding were decreased in both monkeys.
The low doses of nalbuphine (0.1 mg/kg/h) and butorphanol (0.01 mg/kg/h) decreased cocaine self-administration without altering food-maintained responding across all cocaine doses tested (0.0032–0.032 mg/kg/inj). A higher dose of nalbuphine (0.32 mg/kg/h) and butorphanol (0.032 mg/kg/h) produced greater decreases in cocaine self-administration across this dose range and also significantly decreased responding for 0.1 mg/kg/inj cocaine. Food-maintained responding was significantly decreased only during availability of the high dose of 0.1 mg/kg/inj cocaine.
Effects of Nalbuphine and Butorphanol on Cocaine Discrimination.
The effects of nalbuphine and butorphanol on cocaine discrimination were evaluated for comparison with their effects on cocaine self-administration. The effects of fentanyl and morphine on cocaine discrimination in this procedure have been described previously (Negus et al., 1998).
Figure 5 (left) shows that nalbuphine administered alone produced a dose-dependent partial substitution for cocaine that plateaued at 50% cocaine-appropriate responding. This partial substitution in the mean data reflected complete substitution of nalbuphine for cocaine in three of the seven monkeys and partial substitution (30–40%) in two other monkeys. In the final two monkeys, nalbuphine produced primarily saline-appropriate responding up to doses that eliminated responding. Figure 5 (right) shows the effects of nalbuphine pretreatments to cocaine. Discrimination ED50 values and rate suppression ED25 values for cocaine administered alone or after nalbuphine pretreatment are shown in Table 1. Nalbuphine produced dose-dependent leftward shifts in the dose-effect curves for cocaine discrimination and cocaine rate suppression. All doses of nalbuphine significantly decreased the cocaine discrimination ED50 value, and the highest dose of 0.32 mg/kg nalbuphine decreased the cocaine rate suppression ED25 value. Nalbuphine tended to produce the greatest leftward shifts in the cocaine dose-effect curve in those monkeys in which nalbuphine alone produced the highest levels of cocaine-appropriate responding. Nalbuphine pretreatments did not attenuate the discriminative stimulus effects of cocaine in any monkeys (data not shown).
Figure 6 (left) shows that butorphanol also produced a dose-dependent partial substitution for cocaine that peaked at 47% cocaine-appropriate responding. Butorphanol substituted completely for cocaine in two of the seven monkeys, and partial substitution (40–75%) was observed in two other monkeys. With one exception, butorphanol and nalbuphine produced high levels of cocaine-appropriate responding in the same monkeys. In one monkey, however, nalbuphine substituted completely for cocaine, but butorphanol produced primarily saline-appropriate responding. Figure 6 (right) shows the effects of butorphanol pretreatments to cocaine. Discrimination ED50 values and rate suppression ED25 values for cocaine administered alone or after butorphanol pretreatment are shown in Table 1. Butorphanol produced dose-dependent leftward shifts in the cocaine discrimination and rate-suppression dose-effect curves, and the highest dose of 0.032 mg/kg butorphanol significantly decreased the cocaine discrimination ED50 value and rate suppression ED25 value. As with nalbuphine, butorphanol tended to produce the greatest leftward shifts in the cocaine dose-effect curve in those monkeys in which nalbuphine alone produced the highest levels of cocaine-appropriate responding. Butorphanol pretreatments did not attenuate the discriminative stimulus effects of cocaine in any monkeys (data not shown).
Discussion
Effects of μ-Agonists on Cocaine Self-Administration.
The high-efficacy μ-agonist fentanyl, the intermediate-efficacy μ-agonist morphine, and the low-efficacy μ-agonists nalbuphine and butorphanol all produced dose-dependent decreases in cocaine self-administration across a broad range of cocaine doses. These findings agree with previous reports that μ-agonists decreased cocaine self-administration by humans (Foltin and Fischman, 1994), rhesus monkeys (Mello et al., 1989; Carroll et al., 1992; Winger et al., 1992; Mello and Mendelson, 1995) and rodents (Carroll and Lac, 1992). The role of μ-agonist efficacy as a determinant of these effects has not been extensively examined. In a study that used procedures similar to those used herein, the intermediate-efficacy μ-agonist buprenorphine dose-dependently decreased cocaine self-administration, whereas the μ-antagonist naltrexone had little effect (Mello et al., 1990). Moreover, buprenorphine-induced decreases in cocaine self-administration were blocked by naltrexone (Mello et al., 1993b). These results suggested that some μ-agonist activity was required to produce robust decreases in cocaine self-administration. Similarly, Winger et al. (1992) examined the acute effects of a series of μ-opioids on complete cocaine self-administration dose-effect curves in rhesus monkeys. The μ-agonists nalbuphine, buprenorphine and heroin all dose-dependently decreased cocaine self-administration and produced downward shifts in the cocaine self-administration dose-effect curves, whereas the μ-selective opioid antagonist quadazocine had little effect on cocaine self-administration. Taken together with the results of the present study, these findings suggest that cocaine self-administration is decreased by μ-agonists with a wide range of efficacies at μ-opioid receptors.
All four μ-agonists tested in the present study produced sustained decreases in cocaine self-administration across the entire 7-day treatment period. Previous studies in rhesus monkeys found that chronic treatment with buprenorphine (Mello et al., 1992) and morphine (Winger and Woods, 2001) continued to decrease cocaine self-administration for up to 120 days of treatment. These results contrast with the transient decreases in cocaine self-administration produced in rhesus monkeys by treatment with dopamine receptor antagonists (Mello and Negus, 1996).
Selectivity of μ-Agonist Effects on Cocaine Self-Administration.
At least one dose of each μ-agonist produced a selective decrease in 0.01 mg/kg/inj cocaine self-administration without significantly altering behavior maintained by food presentation. This finding agrees with previous reports that μ-agonist effects on rates of operant responding may vary as a function of the reinforcer used to maintain responding (Katz and Goldberg, 1986; Mello et al., 1989). The low-efficacy μ-agonists nalbuphine and butorphanol produced the greatest decreases in cocaine self-administration across the broadest range of μ-agonist doses while producing the smallest effects on food-maintained responding. In contrast, the higher efficacy μ-agonists morphine and fentanyl selectively decreased cocaine self-administration across a relatively narrow range of doses, and higher doses produced pronounced undesirable effects, including sedation and significant decreases in food-maintained responding. These undesirable effects limited the range of high-efficacy μ-agonist doses that could be tested. These results support our hypothesis that opioids with lower efficacy at μ-receptor decrease cocaine self-administration more selectively than higher efficacy μ-agonists.
The selectivity of μ-agonist effects on cocaine self-administration has also been examined with varying results in other studies (Mello et al., 1989, 1992, 1993a; Carroll et al., 1992; Carroll and Lac, 1992;Winger et al., 1992; for review, see Mello and Mendelson, 1995). For example, in agreement with the present findings, buprenorphine produced relatively selective decreases in cocaine self-administration while producing smaller or more transient effects on food-maintained responding (Mello et al., 1989, 1992, 1993b; Mello and Mendelson, 1995). In contrast, a previous study of the effects of nalbuphine and butorphanol on cocaine self-administration reported that these low-efficacy μ-agonists produced nonselective decreases in cocaine self-administration (Mello et al., 1993a). Although that previous study and the present study were similar in most respects, the discrepancy between the present results and these previous findings can be attributed at least in part to two important procedural differences. First, the previous study administered nalbuphine (0.1 to 3 mg/kg/day) and butorphanol (0.01 to 0.3 mg/kg/day) during a 50-min infusion once each morning (Mello et al., 1993a). Both nalbuphine and butorphanol have relatively short durations of action, and the effects of both drugs dissipated over the course of each experimental day. In contrast, the present study administered μ-agonists by repeated injections once every 20 min throughout the day. This more continuous delivery of nalbuphine and butorphanol may have contributed to more selective effects on cocaine self-administration.
A second difference between the two studies is that they used different unit doses of cocaine. The previous study used relatively high cocaine unit doses of 0.05 to 0.1 mg/kg/injection cocaine (Mello et al., 1993a), whereas the present study used a wide range of cocaine unit doses (0.0032–0.1 mg/kg/injection) that spanned both the ascending and descending limbs of the cocaine dose-effect curve. μ-Agonists were most potent and most selective in decreasing self-administration of low cocaine doses, and this is consistent with the more general finding that behavior maintained by low-magnitude reinforcers is more vulnerable to disruption by pharmacological treatments than behavior maintained by higher magnitude reinforcers (for further discussion of this point, see Mello and Negus, 1996).
Acute treatment with μ-agonists also produced nonselective decreases in cocaine self-administration in comparison with effects on alfentanil self-administration in rhesus monkeys (Winger et al., 1992). However, the relatively high-efficacy μ-agonist heroin decreased cocaine and opioid self-administration at a dose (0.1 mg/kg) only 3- to 10-fold lower than toxic doses (Winger et al., 1992). In contrast, the low-efficacy μ-agonist nalbuphine decreased drug self-administration at doses of 0.1 to 1.0 mg/kg, and doses up to 56 mg/kg can be safely administered to monkeys (Gerak et al., 1994). Taken together, these results suggest that low-efficacy μ-agonists may decrease cocaine self-administration with a wider margin of safety than high-efficacy μ-agonists.
Effects of Withdrawal from μ-Agonist Treatment.
After termination of chronic μ-agonist treatment, abstinence signs characteristic of μ-agonist withdrawal (Woods and Gmerek, 1985) were rarely observed (e.g., vomiting was observed in only one monkey after termination of treatment with the highest dose of fentanyl). In addition, cocaine self-administration recovered rapidly, and this finding is consistent with previous studies (Mello and Mendelson, 1995;Winger and Woods, 2001). However, termination of high-dose μ-agonist treatment often produced a transient decline in food-maintained responding, and decreases in food-maintained responding after termination of chronic opioid administration have been interpreted as a sensitive withdrawal sign indicative of physical dependence (Holtzman and Villarreal, 1973). These results suggest that μ-agonist doses that decreased cocaine self-administration in the present study may have produced modest levels of physical dependence.
Previous studies reported some evidence to suggest that maximal levels of μ-opioid-induced physical dependence may be related to agonist efficacy at μ-receptors. For example, chronic administration of morphine (12 mg/kg/day), nalbuphine (128 mg/kg/day) or butorphanol 24 mg/kg/day) for 31 to 38 days produced physical dependence as defined by the emergence of overt abstinence signs after drug withdrawal in rhesus monkeys, and withdrawal from morphine produced greater and more severe abstinence signs than withdrawal from either nalbuphine or butorphanol (Woods and Gmerek, 1985). Similarly, in the present study, the greatest abstinence signs were observed after withdrawal from the high-efficacy μ-agonist fentanyl, whereas the weakest abstinence signs were examined after withdrawal from butorphanol. However, the present study was not designed to provide a comprehensive comparison of the ability of different opioids to produce physical dependence, and overall, abstinence signs were relatively mild after withdrawal from all the μ-agonists tested. The more important point is that μ-agonists decreased cocaine self-administration at doses that did not produce evidence of severe physical dependence and that were below doses that have generally been used to study physical dependence in other studies (Woods and Gmerek, 1985).
Potential Role of κ-Receptor-Mediated Effects for Nalbuphine and Butorphanol.
Nalbuphine and butorphanol produce agonist effects mediated primarily by μ-opioid receptors in rhesus monkeys (Gerak et al., 1994; Butelman et al., 1995). However, these compounds have only modest selectivity for μ- versus κ-receptors and may produce κ-receptor-mediated effects under some conditions. In vitro studies indicated that both compounds have relatively low efficacy at κ-receptors, and behavioral studies in rhesus monkeys found that irreversible blockade of μ-receptors unmasked weak κ-agonist effects of butorphanol but not of nalbuphine (Vivian et al., 1999). We reported previously that low-efficacy κ-ligands had little effect on cocaine self-administration (Mello and Negus, 2000), and as a result, it is unlikely that the low-efficacy actions of nalbuphine and butorphanol at κ-receptors contributed to their effects in the present study. However, intermediate- to high-efficacy κ-agonists do decrease cocaine self-administration in rhesus monkeys, and mixed action μ/κ-agonists produce greater decreases in cocaine self-administration with fewer undesirable effects than highly selective κ-agonists (Mello and Negus, 2000). Thus, we cannot exclude a role for the κ-receptor-mediated effects of nalbuphine and butorphanol in decreasing cocaine self-administration.
μ-Agonist Effects on Cocaine Discrimination.
The low-efficacy μ-agonists nalbuphine and butorphanol substituted for cocaine in some monkeys and produced dose-dependent rate decreasing effects in all monkeys. When nalbuphine and butorphanol were administered as pretreatments to cocaine, the discriminative stimulus and rate-decreasing effects of the opioids and cocaine were approximately additive. Consequently, nalbuphine and butorphanol produced leftward shifts in the cocaine discrimination and rate suppression dose-effect curves. These results agree with a previous study from our laboratory, which found that the intermediate- to high-efficacy μ-agonists morphine and fentanyl also mimicked the discriminative stimulus effects of cocaine in some rhesus monkeys and produced approximately additive effects when administered in combination with cocaine (Negus et al., 1998). These findings in rhesus monkeys also agree with studies in humans, which found that both acute and chronic treatment with μ-agonists increased abuse-related subjective effects of cocaine (Foltin and Fischman, 1992, 1994; Foltin et al., 1995; Preston et al., 1996), even though cocaine self-administration was reduced (Foltin and Fischman, 1994). One implication of these findings is that μ-agonists may enhance some effects of cocaine while decreasing cocaine self-administration. A second implication of these findings is that decreases in cocaine self-administration produced by low- and high-efficacy μ-agonists do not result from a general blockade of the abuse-related behavioral effects of cocaine.
Clinical Implications.
μ-Agonist-induced decreases in cocaine self-administration are concordant with the clinical finding that μ-agonists often decrease cocaine use by cocaine-dependent polydrug abusers (Kosten et al., 1989; Gastfriend et al., 1993; Strain et al., 1994; Schottenfeld et al., 1997; Borg et al., 1999). The extent to which these clinically observed decreases in cocaine use reflect a selective pharmacological effect of the opioid treatment medication on cocaine's reinforcing effects is unclear. Certainly, a number of other factors related to treatment, such as exposure to counseling, may have contributed to these effects. It is also important that cocaine abuse persists in many patients on opioid maintenance (Mello and Mendelson, 1995), and some patients may even initiate cocaine use while on treatment (Condelli et al., 1991; Borg et al., 1999). Finally, it should be noted that even for the low-efficacy μ-agonists, high doses were required to decrease cocaine self-administration relative to doses that are used clinically to produce analgesia (e.g., the lowest effective dose of butorphanol in the present study, 0.01 mg/kg/h, is equivalent to about 16 mg/day butorphanol in a 70-kg human). Nonetheless, the finding that μ-agonists decrease both cocaine self-administration in preclinical studies and cocaine use in many polydrug abusers suggests that opioid maintenance drugs may deserve careful consideration as candidate pharmacotherapies for cocaine abuse.
Acknowledgments
We thank Peter Fivel, Bradford Fischer, Ashton Koo, and David Linsenmayer for excellent technical assistance and Beth Moseley, D.V.M., and Kate Banks, D.V.M., for veterinary assistance.
Footnotes
-
This work was supported by Grants R01-DA02519, P50-DA04059, and K05-DA00101 from National Institute on Drug Abuse, National Institutes of Health.
- Abbreviations:
- VR
- variable ratio
- FR
- fixed ratio
- inj
- injection
- Received September 17, 2001.
- Accepted November 30, 2001.
- The American Society for Pharmacology and Experimental Therapeutics