Cannabinoid receptor agonists enhance the antinociceptive effects of μ-opioid receptor agonists, which suggests that combinations of these drugs might enhance therapeutic effectiveness (e.g., analgesia). However, it is not clear whether combinations of these drugs also enhance abuse or dependence liability. This experiment examined whether combinations of cannabinoids and opioids that enhance antinociception also increase abuse-related effects by studying the effects of the cannabinoid receptor agonists 2-[(1R,2R,5R)-5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]-5-(2-methyloctan-2-yl)phenol (CP 55,940) and (R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone mesylate (WIN 55,212) on the antinociceptive, discriminative stimulus, and positive reinforcing effects of μ-opioid receptor agonists in rhesus monkeys. In one group of monkeys (n = 3), morphine (0.1–5.6 mg/kg s.c.), CP 55,940 (0.0032–0.032 mg/kg s.c.), and WIN 55,212 (0.1–1.0 mg/kg s.c.) dose-dependently increased tail withdrawal latency from 50°C water, and pretreatment with small, otherwise ineffective, doses of CP 55,940 and WIN 55,212 shifted the morphine dose-effect curve to the left. In monkeys (n = 3) discriminating 3.2 mg/kg morphine, CP 55,940 (0.01–0.032 mg/kg s.c.) and WIN 55,212 (0.1–1.78 mg/kg s.c.) attenuated the discriminative stimulus effects of morphine, shifting the dose-effect curve to the right. In monkeys (n = 4) self-administering heroin (0.32–32.0 µg/kg/infusion i.v.), CP 55,940 (0.001–0.032 mg/kg s.c.), and WIN 55,212 (0.1–1.0 mg/kg s.c.) shifted the heroin dose-effect curve rightward and downward. Cannabinoid receptor agonists CP 55,940 and WIN 55,212 enhanced the antinociceptive effects but not the discriminative stimulus or positive reinforcing effects of μ-opioid receptor agonists in rhesus monkeys, supporting the view that combining cannabinoid and opioid receptor agonists might result in enhanced treatment effectiveness for pain without similarly enhancing abuse and dependence liability.
Pain remains a significant clinical problem (Gaskin and Richard, 2012), and μ-opioid receptor agonists (e.g., morphine and hydrocodone) continue to be the most effective treatment of moderate to severe pain. However, the use of opioids to treat pain is limited by unwanted effects (e.g., constipation, respiratory depression, and nausea), and the use of opioids for extended periods can result in the development of tolerance and physical dependence (Gutstein and Akil, 2005; Ballantyne and Shin, 2008; Manchikanti and Singh, 2008). One strategy for enhancing therapeutic effectiveness of a drug is to administer it in combination with another drug that has a different mechanism of action and that shares a therapeutic effect.
μ-Opioid receptor agonists and cannabinoid receptor agonists [e.g., Δ9-tetrahydrocannabinol (Δ9-THC)] share many behavioral and pharmacological effects, including antinociception (see Walker and Hohmann, 2005, for a review), and cannabinoid receptor agonists are increasingly used for treatment of pain (Hosking and Zajicek, 2008). Opioid receptor agonists have been combined with cannabinoid receptor agonists to treat some types of pain (e.g., dronabinol) (Narang et al., 2008). However, the effectiveness of cannabinoids alone can be modest, and the doses required to achieve therapeutic effects can have unwanted effects, thus limiting their clinical use (Kraft, 2012). Despite these limitations, combining opioids and cannabinoids to treat pain continues to be a promising therapeutic approach. Drug combinations allow for the possibility that smaller doses of individual drugs can be combined to maintain or improve the therapeutic effects while reducing the likelihood of encountering the adverse effects associated with larger doses of either drug administered alone (Smith, 2008).
Preclinical research supports combining μ-opioid receptor agonists and cannabinoid receptor agonists to treat pain. For example, in rodents, Δ9-THC has robust antinociceptive effects (Welch and Stevens, 1992; Smith et al., 1998), and acute administration of Δ9-THC enhances the effectiveness of μ-opioid receptor agonists (e.g., morphine, codeine, or fentanyl) (Pertwee, 2001; Cichewicz, 2004; Welch, 2009). Similarly, morphine enhances Δ9-THC–induced antinociception in mice (Reche et al., 1996; Smith et al., 1998). Moreover, in rhesus monkeys, Δ9-THC and other cannabinoid receptor agonists [e.g., (R)-(+)-[2,3-Dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone mesylate (WIN 55,212)] have antinociceptive effects when administered alone (Vivian et al., 1998; Ko and Woods, 1999; Manning et al., 2001), and acute administration of Δ9-THC enhances the antinociceptive effects of morphine (Li et al., 2008).
Although current behavioral and pharmacological evidence indicates that combining cannabinoid receptor agonists with μ-opioid receptor agonists might be an especially useful strategy to enhance antinociceptive effects, less is known about interactions between these drugs with regard to effects that are predictive of abuse. Opioid abuse continues to be a significant public health problem (Compton and Volkow, 2006). Cannabinoids (e.g., marijuana) are also abused and might confer physical dependence as a consequence of repeated use (Lichtman and Martin, 2005). Despite the potential for increased therapeutic effectiveness, the benefit of combining opioids and cannabinoids to treat pain could be undermined by the perceived risk of enhancing unwanted effects, especially if combinations increase abuse. Some studies indicate that Δ9-THC increases the reinforcing and discriminative stimulus effects of opioids in rodents (Norwood et al., 2003; Manzanedo et al., 2004; Solinas et al., 2005); however, in rhesus monkeys, Δ9-THC fails to enhance and can attenuate the discriminative stimulus effects of morphine and heroin (Li et al., 2008). Moreover, acute administration of Δ9-THC does not enhance and, at large doses, suppresses heroin self-administration (Li et al., 2012). Inconsistencies in the reported interactions between cannabinoids and opioids could be the consequence of numerous differences across studies related to species, drug, dose, or experimental history. Nevertheless, it appears that, in human and nonhuman primates, Δ9-THC enhances the antinociceptive effects of morphine across a range of doses that fail to enhance and possibly attenuate the discriminative stimulus and reinforcing effects of μ-opioid receptor agonists, suggesting that combining cannabinoids and opioids might not increase abuse liability.
The current study examined interactions between μ-opioid receptor agonists and cannabinoid receptor agonists by studying the effects of cannabinoid receptor agonists on the antinociceptive, discriminative stimulus, and reinforcing effects of μ-opioid receptor agonists in rhesus monkeys. These studies extend previous research (Li et al., 2008, 2012) by examining interactions of opioids with two cannabinoid receptor agonists, 2-[(1R,2R,5R)-5-hydroxy-2-(3-hydroxypropyl) cyclohexyl]-5-(2-methyloctan-2-yl)phenol (CP 55,940) and WIN 55,212, which have overlapping but not identical pharmacology (e.g., greater efficacy at cannabinoid receptors) (Breivogel and Childers, 2000), compared with Δ9-THC. Moreover, the current experiment investigated the mechanism(s) underlying such interactions by assessing the ability of the cannabinoid receptor antagonist rimonabant to attenuate the effects of CP 55,940 and WIN 55,212 on morphine discrimination and heroin self-administration.
Materials and Methods
Ten adult rhesus monkeys (Macaca mulatta) served as subjects. Body weight (range, 5–11 kg) was maintained by postsession feeding (High Protein Monkey Diet; Harlan Tekland, Madison, WI). Monkeys also received fresh fruit and peanuts daily; water was continuously available in the home cage. Subjects were housed individually in a colony room and under a 14/10-hour light/dark cycle (lights on at 6:00 am). Animals used in these studies were maintained in accordance with the Institutional Animal Care and Use Committee, The University of Texas Health Science Center at San Antonio, and the 2011 Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animals Resources on Life Sciences, National Research Council, National Academy of Sciences).
Monkeys in the antinociception experiment were tested while seated in primate chairs (Model R001-T; Primate Products, Miami, FL). Warm water baths were used to maintain water at the appropriate temperatures. Tails were dipped in plastic insulated mugs containing water, and tail withdrawal latencies were measured using a silent handheld stop watch. Monkeys participating in the morphine discrimination and heroin self-administration experiments were seated in primate chairs (Model R001; Primate Products) and positioned in sound-attenuating operant conditioning chambers containing two horizontally aligned response levers located approximately 32 cm apart. Located above each lever, were circular, translucent disks that could be trans-illuminated green or red. Extraneous sounds were masked by white noise and an exhaust fan. Experimental events were arranged and data were recorded by an interface (Med Associates, Inc., East Fairfield, VT) connected to a PC computer running Med-PC IV software (Med Associates, Inc.). During drug discrimination sessions, feet were placed in shoes containing brass electrodes to which a brief (250 millisecond, 3 mA) electric stimulus could be delivered from a remote current generator (H13-15; Coulbourn Instruments, Allentown, PA). During drug self-administration sessions, infusions were delivered intravenously by connecting a subcutaneous vascular access port (Access Technologies, Skokie, IL) to a 185-cm extension set (Abbott Laboratories, Stone Mountain, GA) via a 20-g Huber-point needle (Access Technologies). The opposing end of the extension set was connected to a 30-ml syringe that was mounted in a syringe driver (Razel Scientific Instruments, Inc., Stamford, CT) located outside the chamber.
Warm Water Tail Withdrawal.
One male (JO) and two female (SA and ME) monkeys were studied in the warm water tail withdrawal study. The lower portion (approximately 15 cm) of the shaved tail was inserted into a mug containing 40, 50, or 55°C water, and the time until the tail was completely removed from the mug was recorded. In the event that the monkey failed to remove the tail from the water within 20 seconds, the mug was removed and the tail withdrawal latency was recorded as 20 seconds (maximum effect). Each temperature was tested once every 15 minutes (once per cycle), and temperatures were tested in random order across consecutive tests and across monkeys. Sessions comprised 8 cycles and began with one saline cycle followed by cumulative doses of morphine across successive cycles; the dose of morphine was increased in 0.25-log unit increments across cycles. Test sessions in which a morphine dose-effect curve was determined ended after eight cycles or when the tail withdrawal latency reached 20 seconds with 50°C water, whichever occurred first. When tested in combination with morphine, WIN 55,212 was administered during the first cycle (substituted for saline), and CP 55,940 was administered 60 minutes before the first cycle.
Three female monkeys (CI, AM, and HA) were trained previously to discriminate morphine from saline during daily sessions; each of two to eight cycles comprised a 10-minute timeout period followed by a 5-minute response period. During the timeout period, all lights were off and responding had no programmed consequence. Injections occurred during the first minute of each timeout period. The beginning of the response period was signaled by illumination of both side key lights red, in the presence of which a brief electric stimulus was scheduled to be delivered every 45 seconds. Thirty consecutive responses on the correct lever (determined by the training or testing condition as described below) extinguished the red lights and suspended the schedule of electric stimulus presentation; after 30 seconds, the red key lights were illuminated and the electric stimulus presentation schedule was restarted. Responses on the incorrect lever reset the response requirement on the correct lever and, otherwise, had no programmed consequence. Response periods lasted either until 5 minutes elapsed or until four electric stimuli were delivered, whichever occurred first.
During training sessions, levers were designated as correct on the basis of the injection given during the first minute of the cycle. For two monkeys (CI and HA), the left lever was designated as correct after saline injections; conversely, the right lever was designated correct after injections of 3.2 mg/kg s.c. morphine (i.e., training dose). For monkey AM, the contingencies were reversed. Individual monkeys were trained under the same lever designations for the entire experiment. During saline training sessions, saline or sham (a capped needle was pressed firmly against the back) injections were administered before two to six cycles. During morphine training sessions, the training dose of morphine was administered at the beginning of one cycle that was preceded by zero to five saline or sham training cycles. The number of saline or sham cycles preceding the morphine training cycle varied quasi-randomly across sessions.
All monkeys in this study had previously met the initial testing criteria, which were five consecutive or six of seven sessions in which at least 80% of total responses occurred on the correct (injection-appropriate) lever and fewer than 30 responses occurred on the incorrect lever before the first reinforcer of the cycle. Thereafter, monkeys were tested when the training criteria were met for two consecutive sessions consisting of one drug training session and one saline training session. Test sessions were identical to training sessions except that completion of the ratio requirement on either lever was reinforced.
Morphine dose-effect curves were determined under test conditions with use of a cumulative dosing method. Saline was administered during the first cycle of the session; thereafter, increasing doses of morphine (in 0.25-log unit increments) were administered across successive cycles until at least 80% of total responses in a cycle occurred on the drug-associated lever, four electric stimuli were delivered, or eight cycles were completed, whichever occurred first. When tested in combination with morphine, CP 55,940 was administered subcutaneously 60 minutes before determination of a morphine dose-effect curve, and WIN 55,212 was administered during the first cycle of the session. After the effects of CP 55,940 and WIN 55,212 were studied alone and in combination with morphine, selected doses of each cannabinoid receptor agonist were studied in combination with morphine (dose-effect curve determination) and the cannabinoid receptor antagonist rimonabant (1.0 mg/kg). Rimonabant was always administered 45 minutes before the session. Control dose-effect curves (morphine alone) were determined immediately before and after a series of tests with both CP 55,940 and WIN 55,212.
Two male (PE and NA) and two female (BE and MA) monkeys participated in the self-administration study. Monkeys were initially sedated with 10 mg/kg s.c. ketamine, and anesthesia was maintained by isoflurane throughout surgery. After being placed in either the jugular or femoral vein, the polyurethane catheter (5-Fr; SIMS Deltec Inc., St. Paul, MN) was tunneled subcutaneously to the midscapular region where it was attached to a subcutaneous access port. During daily sessions, monkeys lever-pressed under a fixed-ratio 30 schedule for intravenous infusions of either heroin or saline. Sessions comprised four 25-minute cycles. Each cycle began with a 5-minute timeout period, during which all lights were extinguished and responding had no programmed consequence, followed by a 20-minute response period. Each response period began with a priming infusion of the unit dose of heroin that was available for self-administration in that component. During the response period, a green light located above the active lever (right lever for MA and BE and left lever for PE and NA) was illuminated, and 30 responses on the active lever turned the lever light red for 2 seconds, activated the syringe pump, and initiated a 180-second timeout, during which all lights were extinguished. Responses on the active lever during timeouts and on the inactive lever at any time were recorded but had no programmed consequence. A maximum of seven infusions could be obtained per cycle.
Heroin dose-effect curves were determined within single sessions; multiple unit doses of heroin (0.32–32.0 µg/kg/infusion) were available within a single session, and a different unit dose was available in each cycle. Doses were presented in ascending order, and the range of doses of heroin was adjusted for individual monkeys to include doses comprising the ascending limb of the dose-effect curve; the unit dose of heroin was altered across cycles by varying the infusion duration, which ranged from 0.6 to 55 seconds. After a range of doses was identified for each monkey, that range remained constant for a particular monkey for the remainder of the experiment. Heroin dose-effect curves were considered to be stable when the mean number of infusions received in each of the four cycles did not vary by more than ±1.5 infusions, compared with the number of infusions obtained in the corresponding cycle for three consecutive sessions. After stable heroin dose-effect curves were established, the effects of CP 55,940 were determined first by administering the drug subcutaneously 60 minutes before the beginning of a heroin self-administration session. Tests occurred no more frequently than once every four sessions and only when responding was stable (see above). The effects of WIN 55,212 were determined in the same fashion; however, WIN 55,212 was administered subcutaneously immediately before the beginning of the self-administration session. After the effects of CP 55,940 and WIN 55,212 on heroin self-administration were studied, selected doses of the cannabinoid agonist were studied in combination with rimonabant (1.0 mg/kg) before a heroin self-administration session. Rimonabant was administered 45 minutes before the session.
For the antinociception study, tail withdrawal latencies were expressed as a percentage of maximal possible effect (MPE) according to the following formula: % MPE = [(test latency – control latency)/(20 seconds – control latency)] × 100. Control latencies were determined in the absence of drug. The mean (± S.E.M.) MPE was calculated for each agonist dose administered alone and in combination and plotted as a function of dose of morphine. For the discrimination studies, the dependent measures were the distribution of responses between the two levers and response rate. For each cycle, the total number of responses on the drug lever was divided by the total number of responses on either lever and multiplied by 100 to yield the percentage of drug-lever responding. Response rate was calculated by dividing the total number of responses by the duration of the response period excluding timeouts. The mean (± S.E.M.) percentage drug-lever responding and response rate were calculated for each dose of morphine alone and in combination with test compounds and plotted as a function of dose of morphine. Control dose-effect curves for morphine antinociception and discrimination represent the mean of two determinations. Tests with CP 55,940, WIN 55,212, and rimonabant were determined once. For antinociception and discrimination studies, potency was estimated by calculating the dose required to produce the ED50 for individual monkeys by fitting straight lines to dose-effect data. Only the linear portion of the curve was fit; this portion of the curve ranged from the largest dose that produced less than 20% to the smallest dose that produced more than 80% of the maximum effect. For self-administration studies, the mean (± S.E.M.) number of infusions received per component was plotted as a function of unit dose of heroin. Control heroin dose-effect curves represent the mean (± S.E.M.) number of infusions obtained for the three consecutive sessions immediately before the first tests during which responding was deemed stable. All curve fits and analyses were conducted using GraphPad Prism (version 5.0; GraphPad Software, Inc., San Diego, CA).
Morphine sulfate, heroin hydrochloride, and rimonabant were provided by the National Institute on Drug Abuse (Research Technology Branch, Rockville, MD). CP 55,940 was purchased from Sigma-Aldrich (St. Louis, MO), and WIN 55,212 was purchased from Tocris (Ellisville, MO). Morphine was dissolved in sterile water, and heroin was dissolved in 0.9% saline. Rimonabant, CP 55,940, and WIN 55,212 were dissolved in a 1:1:18 mixture of absolute ethanol, Emulphor-620 (Rhone-Poulenc Inc., Princeton, NJ) and 0.9% saline. All doses are expressed as the salt, and all drugs except heroin were administered subcutaneously in the back in a volume of 0.2–0.8 ml. Heroin was administered intravenously during self-administration studies.
Under control conditions, the mean (± S.E.M.) tail withdrawal latencies for 40, 50, and 55°C water were 20 ± 0, 2.52 ± 0.28, and 1.12 ± 0.01 seconds, respectively. For all monkeys, doses of CP 55,940 up to and including 0.01 mg/kg and doses of WIN 55,212 up to and including 0.32 mg/kg failed to increase tail withdrawal latencies above control levels in 50°C water; ED50 values were 0.018 (SA), 0.011 (ME), and 0.021 (JO) mg/kg for CP 55,940 and 0.35 (SA), 0.37 (ME), and 0.52 (JO) mg/kg for WIN 55,212 (Fig. 1). Morphine dose-dependently increased tail withdrawal latency in 50°C water (ED50 of 0.51, 0.70, was 2.57 for monkeys SA, ME, JO, respectively) (Fig. 1). Pretreatment with 0.01 mg/kg CP 55,940 (Fig. 1), a dose that was ineffective in all monkeys when administered alone, reduced the ED50 for morphine to 0.35 (SA), 0.14 (ME), and 0.19 (JO) mg/kg, resulting in a 1.6- (SA), 5.2- (ME), and 13.4-fold (JO) leftward shift in the morphine dose-effect curve. For monkey JO, pretreatment with 0.0178 mg/kg CP 55,940 (Fig. 1) did not increase tail withdrawal latency when administered alone and reduced the ED50 for morphine to 0.11 mg/kg, resulting in a 23.1-fold leftward shift. Likewise, pretreatment with 0.32 mg/kg WIN 55,212 (Fig. 1), a dose that was ineffective when administered alone, reduced the ED50 for morphine to 0.14 (SA), 0.11 (ME), and 0.47 (JO) mg/kg, resulting in a 3.9- (SA), 7.3- (ME), and 5.4-fold (JO) leftward shift in the morphine dose-effect curve, respectively. In one monkey (ME), 0.1 mg/kg WIN 55,212 also shifted the morphine dose-effect curve 6.6-fold to the left.
In monkeys trained to discriminate 3.2 mg/kg morphine, saline occasioned less than 1% responding on the drug lever (Fig. 2). Morphine dose-dependently increased drug-lever responding (Fig. 2), with 3.2 (CI), 0.56 (HA), and 1.0 (AM) mg/kg occasioning greater than 80% responding on the drug lever. The ED50 values for morphine control dose-effect curves were 1.72 (CI), 0.46 (HA), and 0.69 (AM) mg/kg. CP 55,940 alone did not occasion drug-lever responding in any monkey (Fig. 2) and dose-dependently shifted the morphine dose-effect curve rightward up to 6.6- (AM) and 5.7-fold (HA) and, in some cases, flattened the curve (CI; larger doses were not studied). Often, rightward shifts in the discrimination dose-effect curves occurred in the absence of substantial decreases in response rate (Table 1), with the exception on 0.032 mg/kg in monkey AM. When administered alone, rimonabant (1.0 mg/kg) had no effect on the morphine dose-effect curve for any monkey (data not shown) and attenuated the effects of CP 55,940, resulting in a leftward shift in the discrimination dose-effect curve back to near control. Similar to CP 55,940, WIN 55,212 did not occasion drug-lever responding in any monkey (Fig. 2) and dose-dependently shifted the morphine dose-effect curve rightward up to 2.5-fold (AM) and downward (HA and CI). Shifts in the discrimination dose-effect curves occurred in the absence of substantial decreases in response rate (Table 2). Rimonabant (1.0 mg/kg) attenuated the effects of WIN 55,212 on the morphine discrimination dose-effect curves in two monkeys (0.32 mg/kg in HA and 1.78 mg/kg in AM), indicated by leftward shifts in the dose-effect curves, but was ineffective in preventing the effects of WIN 55,212 in monkey CI.
In all monkeys self-administering heroin, the number of infusions obtained per cycle increased as the unit dose of heroin increased (Fig. 3). When saline was available, the mean (± S.E.M.) number of infusions obtained for all four cycles of the session was 1.9 ± 0.4 (NA), 1.4 ± 0.3 (PE), 0.8 ± 0.2 (BE), and 0.8 ± 0 (MA) and mean response rates were 0.06 ± 0.02 (NA), 0.05 ± 0.01 (PE), 0.03 ±0.01 (BE), and 0.03 ± 0 (MA) responses per second (data not shown). The largest number of infusions was obtained and the highest response rates were maintained at a unit dose of 3.2 µg/kg/infusion for two monkeys (PE and MA) and a unit dose of 10 µg/kg/infusion for two other monkeys (NA and BE), with the highest mean number of infusions ranging from 4.7 ± 0.3 (MA) to 6.0 ± 0 (NA and PE) infusions and response rates ranging from 0.4 ± 0.1 (MA) to 4.4 ± 0.1 (NA) responses per second.
CP 55,940 dose-dependently decreased heroin self-administration in all four monkeys (Fig. 3), shifting heroin dose-effect curves rightward in two monkeys (NA and PE) and downward in two monkeys (BE and MA). Monkey NA was most sensitive to the effects of CP 55,940; 0.0032 mg/kg (upright triangles) reduced the number of infusions obtained per cycle at the three largest unit doses (3.2, 10, and 32 µg/kg/infusion). For monkeys PE and MA, 0.01 mg/kg CP 55,940 reduced the number of infusions obtained per cycle at intermediate unit doses of heroin (3.2 and 10 µg/kg/infusion). For monkey BE, 0.01 mg/kg CP 55,940 had no effect on the number of infusions received; however, 0.032 mg/kg reduced the number of infusions obtained per cycle at each of the three highest unit doses (3.2, 10, and 32 µg/kg/infusion). That same dose of CP 55,940 increased the number of infusions obtained at the lowest unit dose of heroin (1.0 µg/kg/infusion) from a mean of 0.7 to 3.0 infusions. Pretreatment with 1.0 mg/kg rimonabant alone did not impact heroin self-administration in any monkey (data not shown) and attenuated the effects of CP 55,940 when administered in combination with 0.0032 (NE), 0.01 (PE and MA), and 0.032 (BE) mg/kg, in each case, resulting in heroin dose-effect curves that were similar to those obtained before treatment.
WIN 55,212 also dose-dependently reduced heroin self-administration in all four monkeys, resulting in substantial rightward (NA and PE) and downward (BE and MA) shifts in the heroin dose-effect curves (Fig. 3). Doses of 0.32 (NA and ME) and 1.0 (PE and BE) mg/kg generally reduced the number of infusions received of all unit doses of heroin and, in no case, increased the number of infusions received. For all four monkeys, 1.0 mg/kg rimonabant attenuated the effects of WIN 55,212 on heroin self-administration when administered in combination with effective doses of WIN 55,212 (0.32 mg/kg for NE and MA and 1.0 mg/kg for PE and BE), resulting in each case, in heroin dose-effect curves similar to those obtained before treatment.
Pain remains a significant public health problem, and despite their widespread use for treating pain, opioids are limited by unwanted effects, ineffectiveness in some patients, and the development of physical dependence after repeated administration. Consequently, there is an unmet need for more effective pain treatments. Cannabinoid and opioid systems interact at cellular, neurochemical, and behavioral levels, and some interactions are likely to be mediated by common pathways (Vigano et al., 2005; Bushlin et al., 2010) that might be important for developing novel therapeutics that target both systems. Cannabinoid receptor agonists have robust antinociceptive effects in nonhumans (reviewed in Pertwee, 2001; Welch, 2009), but their effectiveness in humans is unclear (Kraft, 2012). One strategy for enhancing therapeutic effectiveness is to combine cannabinoids with other drugs. In several species, Δ9-THC enhances the antinociceptive effects of opioids (Welch and Stevens, 1992; Finn et al., 2004; Li et al., 2008). Moreover, supplementing ongoing opioid pain treatments with cannabinoids (e.g., dronabinol) can increase treatment effectiveness (Narang et al., 2008). However, because opioids (e.g., prescription analgesics) and cannabinoids (e.g., marijuana) are abused individually, there is some concern that combining opioids and cannabinoids might increase abuse. The current study examined interactions between μ-opioid receptor agonists and cannabinoid receptor agonists by studying the effects of cannabinoid receptor agonists CP 55,940 and WIN 55,212 on the antinociceptive, discriminative stimulus, and reinforcing effects of μ-opioid receptor agonists morphine and heroin in rhesus monkeys.
Morphine, CP 55,940, and WIN 55,212 dose-dependently increased tail withdrawal latency from 50°C water (Dykstra and Woods, 1986; Vivian et al., 1998). Pretreatment with doses of CP 55,940 and WIN 55,212 that did not increase tail withdrawal latency when tested alone shifted the morphine dose-effect curve leftward, consistent with effects in rodents (Welch and Stevens, 1992; Smith et al., 1998; Miller et al., 2012) and rhesus monkeys (Li et al., 2008). Because CP 55,940 and WIN 55,212 increased the antinociceptive effects of morphine, it might be expected that they would similarly enhance other (e.g., discriminative stimulus and positive reinforcing) effects of μ-opioid receptor agonists. To the contrary, CP 55,940 and WIN 55,212 shifted the morphine discriminative stimulus and heroin self-administration dose-effect curves rightward and downward. The mechanisms underlying the differential effects of CP 55,940 and WIN 55,212 in these studies (i.e., enhancing some but not other effects of opioids) remain unclear. Differences in interactions are not likely to be attributable to pharmacokinetics, because the routes of administration and pretreatment times for morphine and cannabinoid receptor agonists were consistent across studies. Moreover, it is clear from the consistent rank order potency of agonists and from quantitatively consistent antagonism studies that these behavioral effects of morphine and heroin are mediated by μ-opioid receptors (Bowen et al., 2002; Gerak et al., 1994).
Opioid receptor agonists and cannabinoid receptor agonists individually can have robust and pharmacologically distinct discriminative stimulus effects. μ-Opioid receptor agonists (e.g., morphine and hydromorphone) fail to occasion drug-appropriate responding in rhesus monkeys (Wiley et al., 1995a; McMahon, 2006; Li et al., 2008) or humans (Lile et al., 2009) discriminating Δ9-THC. Similarly, Δ9-THC does not occasion drug-appropriate responding in rhesus monkeys discriminating morphine (Li et al., 2008). Consistent with these findings, neither CP 55,940 nor WIN 55,212 occasioned drug-appropriate responding in monkeys discriminating morphine. However, CP 55,940 and WIN 55,212 dose-dependently shifted the morphine discriminative stimulus dose-effect curve rightward and did so at doses that generally did not impact rates of responding. Attenuation of the discriminative stimulus effects of morphine could result from pharmacological antagonism or some other mechanism (e.g., physiologic antagonism). Because CP 55,940 and WIN 55,212 did not attenuate the antinociceptive effects of morphine and that μ receptors mediate antinociceptive and discriminative stimulus effects of morphine, pharmacological antagonism seems to be an unlikely mechanism. Instead, CP 55,940 and WIN 55,212 might reduce the discriminability of morphine, which necessitates an increase in the dose of morphine required for drug-appropriate responding. Masking is a phenomenon whereby one stimulus interferes with the ability of the subject to detect another stimulus; CP 55,940 and WIN 55,212 might mask the discriminative stimulus effects of morphine (Li et al., 2008). Consistent with this view, the discriminative stimulus effects of morphine can be attenuated by nonopioid drugs (e.g., amphetamine) (Gauvin and Young, 1989), and Δ9-THC attenuates the discriminative stimulus effects of noncannabinoid drugs (e.g., phencyclidine) (Doty et al., 1994). Thus, CP 55,940 and WIN 55,212 might attenuate the discriminative stimulus effects of morphine through perceptual (i.e., nonpharmacologic) mechanisms. Whether this interaction between opioids and cannabinoids is attributable to masking and whether such attenuation of drug effects has implications for drug abuse and drug abuse treatment (i.e., reduction in subjective effects) remain to be determined.
Cannabinoid receptor agonists have positive reinforcing effects under some conditions (Tanda et al., 2000) and can enhance the positive reinforcing effects (self-administration) of heroin in rats (e.g., Δ9-THC and WIN 55,212) (Solinas et al., 2005). In the current study, neither CP 55,940 nor WIN 55,212 enhanced heroin self-administration; instead, both agonists decreased the number of heroin infusions received, shifting the heroin dose-effect curve rightward and downward. It is unclear whether reductions in heroin self-administration are attributable specifically to attenuation of the positive reinforcing effects, because decreased self-administration could be the result of several factors. For example, as in the discrimination study, CP 55,940 and WIN 55,212 might attenuate the discriminative stimulus effects of heroin that impact self-administration. Alternatively, decreases in heroin self-administration could be attributable to nonselective suppression of responding. Doses of CP 55,940 (0.0032–0.032 mg/kg) and WIN 55,212 (0.32–1.0 mg/kg) that decreased heroin self-administration did not markedly reduce rates of responding maintained by shock avoidance (current discrimination study), whereas those doses decreased rates of responding maintained by other reinforcers (e.g., food) in rhesus monkeys (McMahon, 2011).
The failure of CP 55,940 and WIN 55,212 to enhance the reinforcing effects of heroin might be a consequence of the manner in which drug combinations were studied. Contingency of drug administration can impact the behavioral effects of some drugs (Lecca et al., 2007), and it is possible that allowing monkeys to self-administer CP 55,940 and WIN 55,212 each in combination with heroin would have enhanced the reinforcing effects of heroin. However, in a previous study (Li et al., 2012) in which rhesus monkeys could self-administer Δ9-THC in combination with heroin, the combinations resulted in an overall decrease in responding and in the amount of drug self-administered with no evidence of enhancement. Although CP 55,940 and WIN 55,212 were administered response-independently in the current study, on the basis of previous work, it appears to be unlikely that these cannabinoid receptor agonists would enhance the reinforcing effects of heroin in rhesus monkeys under other conditions.
Rimonabant (1.0 mg/kg) completely blocked the effects of CP 55,940 and WIN 55,212 on the discriminative stimulus effects of morphine and on the reinforcing effects of heroin, confirming that these effects of CP 55,940 and WIN 55,212 are mediated by CB1 receptors (Wiley et al., 1995b; McMahon, 2006). When administered alone, the same dose of rimonabant had no effect on the discriminative stimulus effects of morphine or heroin self-administration. In rats, rimonabant decreases heroin self-administration that is maintained under a progressive ratio schedule (Solinas et al., 2003); conversely, the μ-opioid receptor antagonist naltrexone attenuates the discriminative stimulus effects of Δ9-THC in rats (Solinas and Goldberg, 2005). Taken together, these studies demonstrate that, in rodents, there is a significant interaction between cannabinoid and opioid systems with regard to the discriminative stimulus and reinforcing effects of cannabinoid receptor and μ-opioid receptor agonists (reviewed in Solinas et al., 2007). However, the nature of interaction between cannabinoid and opioid systems in human and nonhuman primates is less clear. Naltrexone has inconsistent effects on the subjective and physiologic effects of Δ9-THC in humans (Greenwald and Stitzer, 2000; Haney et al., 2003; Haney, 2007). Moreover, Δ9-THC attenuates the discriminative stimulus effects of morphine and heroin (Li et al., 2008) and reduces heroin self-administration (Li et al., 2012) in rhesus monkeys.
In summary, acute administration of the cannabinoid receptor agonists CP 55,940 and WIN 55,212 enhanced the antinociceptive but not the discriminative stimulus or positive reinforcing effects of the μ-opioid receptor agonists morphine and heroin. Taken together with previous work (Li et al., 2008, 2012), these data indicate that combinations of cannabinoids and opioids might enhance therapeutic effectiveness of opioids without similarly enhancing their abuse liability; the current study also demonstrates that these interactions are likely to be mediated through CB1 receptors. Moreover, in the current and in previous studies, cannabinoid receptor agonists robustly attenuated the discriminative stimulus and reinforcing effects of opioids. This suggests that administration of cannabinoids might attenuate the subjective effects of opioids and thereby reduce effects related to abuse liability but not the therapeutic effects (e.g., antinociception). It is important to determine whether the nature of these drug interactions are the same or different during the course of long-term treatment.
The authors thank Hillary Halderman, Jeff Pressley, Sara Schirmer, and Monica Shroyer for excellent technical assistance.
Participated in research design: Maguire, Yang, France.
Conducted experiments: Maguire, Yang.
Performed data analysis: Maguire.
Wrote or contributed to the writing of the manuscript: Maguire, France.
- Received February 15, 2013.
- Accepted March 26, 2013.
This work was supported by the National Institutes of Health National Institute on Drug Abuse [Grants R01DA005018, T32DA031115, and K05DA17918].
- CP 55,940
- 2-[(1R,2R,5R)-5-hydroxy-2-(3-hydroxypropyl) cyclohexyl]-5-(2-methyloctan-2-yl)phenol
- maximal possible effect
- WIN 55,212
- (R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone mesylate
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics