Chronic Δ⁹-tetrahydrocannabinol treatment in rhesus monkeys: differential tolerance and cross-tolerance among cannabinoids

Abstract

BACKGROUND AND PURPOSE

The extent to which behavioural effects vary as a function of CB₁ receptor agonist efficacy is not clear. These studies tested the hypothesis that cannabinoid tolerance and cross-tolerance depend upon the CB₁ agonist efficacy of drugs to which tolerance/cross-tolerance develops.

EXPERIMENTAL APPROACH

Sensitivity to cannabinoids, including the cannabinoid antagonist rimonabant, low efficacy agonist Δ⁹-tetrahydrocannabinol (Δ⁹-THC), and high efficacy agonists CP 55940 and WIN 55212-2, was determined before and after chronic Δ⁹-THC treatment in rhesus monkeys. Two measures of behavioural effect were assessed: effects of drugs to decrease fixed ratio responding for food presentation and stimulus-shock termination and discriminative stimulus effects in monkeys discriminating Δ⁹-THC (0.1 mg·kg⁻¹, i.v.).

KEY RESULTS

Δ⁹-THC decreased responding for both food presentation and stimulus-shock termination; these effects were antagonized by the CB₁ antagonist rimonabant. Chronic Δ⁹-THC (1 mg·kg⁻¹ per 12 h, s.c.) resulted in tolerance to the rate-decreasing effects of Δ⁹-THC and cross-tolerance to CP 55940 and WIN 55212-2; however, cross-tolerance was less than tolerance. Chronic Δ⁹-THC increased sensitivity to rimonabant without changing sensitivity to the non-cannabinoids midazolam and ketamine. In monkeys discriminating Δ⁹-THC (0.1 mg·kg⁻¹, i.v.), both CP 55940 and WIN 55212-2 produced high levels of drug-lever responding. Chronic Δ⁹-THC (1 mg·kg⁻¹ per day, s.c.) decreased sensitivity to Δ⁹-THC without producing cross-tolerance to CP 55940 or WIN 55212-2.

CONCLUSIONS AND IMPLICATIONS

In Δ⁹-THC-treated monkeys, the magnitude of tolerance and cross-tolerance to other CB₁ receptor agonists varied inversely with agonist efficacy, suggesting that CB₁ agonist efficacy is an important determinant of behavioural effects.

Introduction

Pharmacological efficacy has been defined as the relationship between the magnitude of physiological response produced by a drug or ligand at a certain amount of receptor occupancy (Ariens, 1954; Stephenson, 1956; Kenakin, 2002). Agonists acting at a common receptor subtype can vary in efficacy; for example, an agonist producing the maximum response at full or less-than-full receptor occupancy has higher efficacy than agonists producing less than the maximum at full receptor occupancy. When a low efficacy agonist is able to produce the same maximum response as that obtained with a high efficacy agonist, the low efficacy agonist occupies a greater number of receptors than a higher efficacy agonist at the smallest concentrations or doses producing the maximum response. Agonists not only can differ from each other in efficacy, but the same agonist can vary in its potency and maximum response obtained, depending on receptor density and the efficiency of receptor coupling to second messenger pathways (e.g. G-protein-coupled receptors) which, in turn, can vary among different organ systems or responses (Kenakin, 1981). Therefore, when receptor density and coupling efficiency are high, an agonist can be relatively more potent and produce a higher level of response than when receptor density and coupling efficiency are low.

For some drug classes (e.g. µ opioid agonists), in vivo effects critically depend on agonist efficacy (Bergman et al., 2000 for review), with increasing levels of µ agonist efficacy associated with increasing levels of effect (e.g. analgesia and abuse liability). However, for other drugs (e.g. cannabinoids), relationships between agonist efficacy and behavioural effects are less clear. CB₁ receptors are coupled to inhibitory G-proteins and CB₁ agonists typically inhibit adenylyl cyclase (Howlett, 2004). G-protein activation assessed with [³⁵S]-guanosine 5′-3′thio-triphosphate binding has been used to demonstrate marked differences in CB₁ receptor agonist efficacy among ligands; for example, rank order efficacy was WIN 55212-2 > CP 55940 > Δ⁹-tetrahydrocannabinol (Δ⁹-THC; Breivogel and Childers, 2000). However, despite having low efficacy, Δ⁹-THC often produces the same maximum effect in vivo as that obtained with high efficacy cannabinoid agonists. For example, Δ⁹-THC, CP 55940 and WIN 55212-2 produced similar maximum effects in drug discrimination procedures (e.g. McMahon, 2006) and assays concurrently assessing hypothermia, antinociception, locomotor activity and immobility (Fan et al., 1994). The ability of Δ⁹-THC to achieve the same level of response as high efficacy agonists across a broad range of conditions likely reflects the large number of CB₁ receptors (i.e. spare receptors) in the nervous system (Gifford et al., 1999) and high coupling efficiency to G-proteins.

One strategy for assessing relative efficacy among agonists is to decrease receptor function through chronic agonist treatment, and studies with µ opioids provide a compelling example of this approach. Chronic morphine treatment decreases sensitivity to morphine (i.e. tolerance) and other µ agonists (i.e. cross-tolerance) by decreasing downstream signalling through µ receptors (He et al., 2002; Martini and Whistler, 2007). In morphine-treated animals, the magnitude of decrease in sensitivity to some behavioural effects varies with agonist efficacy such that greater loss of sensitivity occurs for low efficacy as compared with high efficacy agonists (Walker et al., 1997), as would be predicted by receptor theory. Chronic Δ⁹-THC treatment can decrease CB₁ receptor number and function (Sim et al., 1996; Breivogel et al., 1999) and can produce tolerance and cross-tolerance (Fan et al., 1994); however, relationships between tolerance/cross-tolerance and CB₁ agonist efficacy have not been firmly established. Such a relationship would suggest that the potency and maximum effect of various cannabinoid agonists, including endogenous ligands such as anandamide, vary significantly as a function of different levels of marijuana smoking or other cannabinoid use (oral Δ⁹-THC) and corresponding CB₁ receptor desensitization.

In the current study, rhesus monkeys received chronic Δ⁹-THC treatment. Changes in sensitivity to cannabinoids were assessed with two measures of behavioural effect: decreases in the rate of fixed ratio responding and discriminative stimulus effects. One group of monkeys responded under fixed ratio schedules of food presentation and stimulus-shock termination (SST) and sensitivity to the effects of cannabinoids (rimonabant, Δ⁹-THC, CP 55940 and WIN 55212-2) and non-cannabinoids (midazolam and ketamine) to decrease rate of responding was assessed before, during and after discontinuation of chronic Δ⁹-THC treatment (1 mg·kg⁻¹ per 12 h, s.c. for 71 days). Another group of monkeys discriminated Δ⁹-THC (0.1 mg·kg⁻¹, i.v.) and sensitivity to the discriminative stimulus effects of Δ⁹-THC and to the Δ⁹-THC-like discriminative stimulus effects of WIN 55212-2 and CP 55940 was assessed before and after daily Δ⁹-THC (1 mg·kg⁻¹, s.c.) treatment; discrimination training was suspended during daily Δ⁹-THC treatment. Overall, these studies tested the hypothesis that the magnitude of tolerance and cross-tolerance to cannabinoid agonists in Δ⁹-THC-treated animals varies inversely as a function of cannabinoid agonist efficacy.

Methods

Animals

One group of adult rhesus monkeys (Macaca mulatta), consisting of two females and four males, responded for food presentation and SST. Another group of adult rhesus monkeys, consisting of one female and four males, discriminated Δ⁹-THC. Monkeys were housed individually on a 14 h light/10 h dark schedule, were maintained at 95% free-feeding weight (range 4–9 kg) with a diet consisting of primate chow (High Protein Monkey Diet, Harlan Teklad, Madison, WI, USA), fresh fruit and peanuts, and were provided with water in the home cage. Monkeys responding for food presentation and SST were pharmacologically and behaviourally naïve prior to this study; monkeys discriminating Δ⁹-THC were trained previously and had received cannabinoids and non-cannabinoids (e.g. McMahon, 2006, 2009). All monkeys were maintained in accordance with the Institutional Animal Care and Use Committee, The University of Texas Health Science Center at San Antonio, and with the ‘Guidelines for the Care and Use of Mammals in Neuroscience and Behavioral Research’ (National Research Council, 1996).

Catheter insertion

Monkeys were anaesthetized with ketamine (10 mg·kg⁻¹, i.m.) and isoflurane (1.5–3.0% inhalation). Using sterile techniques, chronic indwelling catheters (heparin-coated polyurethane, outer diameter = 1.68 mm, inner diameter = 1.02 mm; Instech Solomon, Plymouth Meeting, PA) were inserted into a femoral or subclavian vein. Suture silk (coated vicryl, Ethicon Inc., Somerville, NJ, USA) was used to anchor the catheter to the vessel and to ligate the section of the vessel proximal to the catheter insertion. The distal end of the catheter was attached to a vascular access port (Mida-cbas-c50, Instech Solomon) positioned s.c. at the mid-scapular region of the back.

Apparatus

Monkeys were seated in chairs (Model R001, Primate Products, Miami, FL, USA) and were placed in ventilated, sound-proof chambers equipped with levers and lights. Feet were placed in shoes containing brass electrodes to which a brief electric stimulus (3 mA, 250 ms) could be delivered from an a/c generator. An interface (MedAssociates, St. Albans, VT, USA) connected the chambers to a computer, which controlled and recorded experimental events with Medical-PC software.

Schedules of food presentation and SST: acute effects of Δ⁹-THC and rimonabant

In the presence of a green light, two female and two male monkeys responded under a schedule of continuous reinforcement to obtain 300-mg banana-flavoured food pellets (BioServ, Frenchtown, NJ, USA); the response requirement was systematically increased to an FR5 over days. Thereafter, sessions were divided into multiple cycles, each beginning with a 15-min timeout; the chamber was dark and responses had no programmed consequence. The timeout was followed by a 5-min response period; the green light was illuminated and a food pellet could be obtained upon completion of the response requirement up to delivery of 10 food pellets in a cycle. If 10 food pellets were obtained in less than 5 min, then the green light was extinguished and the remainder of the 5-min period was a timeout.

Sessions consisted of two to six cycles and were conducted 7 days per week. Vehicle was administered i.v. in the first minute of the first cycle, followed by vehicle or sham (dull pressure applied to the skin overlying the vascular access port) non-systematically in the first minute of subsequent cycles. Sessions were conducted until the response rate for individual monkeys stabilized, defined as five consecutive days with response rate within ±20% of the mean rate for those days. Thereafter, tests were conducted by administering vehicle or a dose (0.32–3.2 mg·kg⁻¹, i.v.) of rimonabant, 15 min before sessions in which vehicle was administered in the first cycle, followed by Δ⁹-THC i.v. in cumulative doses increasing by 0.25 or 0.5 log unit per cycle. The dose-effect function included ineffective doses (i.e. doses not modifying response rate) up to doses that markedly decreased response rate (i.e. doses resulting in delivery of no food pellets in a cycle). The order of testing with different doses of rimonabant was non-systematic, and two control dose-effect tests with Δ⁹-THC were determined, one before and the other after tests with rimonabant in combination with Δ⁹-THC were completed. Test sessions were conducted twice weekly so long as response rate in the preceding training session was within ±20% of the mean rate for the five previous vehicle-sessions; otherwise, testing was postponed until this criterion was satisfied.

Upon completion of tests under the schedule of food presentation, monkeys were trained to respond for SST under a schedule of continuous reinforcement. A red light was illuminated and an electric stimulus was scheduled for delivery every 40 s until completion of the response requirement or until four electric stimuli were delivered, whichever occurred first. Completion of the response requirement extinguished the light and prevented delivery of the electric stimulus for 30 s; thereafter, the light was illuminated and the schedule was repeated. The response requirement was systematically increased to an FR5 and sessions were divided into multiple cycles consisting of a 15-min timeout followed by a 5-min period of SST. Tests under the schedule of SST ended when four electric stimuli were delivered in a cycle. The criterion for testing and order of Δ⁹-THC and rimonabant administration in monkeys responding for SST were identical to that described for food-maintained responding; test sessions ended when electric stimuli were delivered.

Multiple schedule of food presentation and SST: sensitivity of drugs before and during chronic treatment with Δ⁹-THC

Chronic Δ⁹-THC was studied in one female monkey that participated in the experiment described above and two other naïve male monkeys that were trained to respond for food presentation and SST as described above. In order to study food presentation and SST simultaneously during chronic treatment, monkeys responded for the reinforcers under a multiple FR5/FR5 schedule. Sessions consisted of two to six, 20-min cycles, each comprising a 12-min timeout period during which the chamber was dark and responses had no programmed consequence. This was followed by an 8-min period consisting of food presentation followed by SST. Monkeys responded for food under an FR5 schedule as described above; the schedule of food presentation ended after 2 min or the delivery of 10 food pellets, in which case the remainder of the 2-min food component was a timeout. The food component was followed by a 1-min timeout and then a 5-min period of SST with parameters identical to that described above. Vehicle i.v. or sham was administered at the beginning of cycles. Each cycle ended after 8 min of schedule presentations (including the 2-min food component, 1-min timeout and 5-min SST component).

Training was conducted until stable rates of responding were established during both components, defined as five consecutive days with response rates for both components within ±20% of the mean rate for those days. Thereafter, monkeys received cumulative i.v. doses of Δ⁹-THC, CP 55940, WIN 55212-2, rimonabant, midazolam and ketamine. Δ⁹-THC was tested first followed by the other drugs in non-systematic order. Doses were increased by 0.25 or 0.5 log unit each cycle and the largest dose resulted in delivery of electric stimuli. Monkeys were then treated daily with Δ⁹-THC (1 mg·kg⁻¹ per 12 h, s.c. at 0600 h and 1800 h) for 71 days; sessions were conducted daily at noon. During chronic Δ⁹-THC treatment, monkeys received vehicle or sham during sessions until all three monkeys satisfied the criterion above. Beginning on day 24 of treatment, monkeys were tested with cumulative i.v. doses of Δ⁹-THC, CP 55940, WIN 55212-2, rimonabant, midazolam and ketamine; order of testing was non-systematic. The effects of Δ⁹-THC, CP 55940, WIN 55212-2 and rimonabant were determined twice on separate days; midazolam and ketamine were studied once. After Δ⁹-THC was administered at 0600 h on day 71 of treatment, Δ⁹-THC treatment was discontinued and sessions were conducted by administering vehicle or sham. Twenty-eight days after the discontinuation of chronic Δ⁹-THC treatment, cumulative doses of Δ⁹-THC and rimonabant were administered in separate test sessions. Tests with drugs before, during and after discontinuation of chronic Δ⁹-THC treatment were separated by at least two vehicle-sessions and were only conducted when responding for both food presentation and SST in the immediately preceding vehicle-session was ±20% of the mean rate for the five previous vehicle-sessions; otherwise, testing was postponed until the criterion was satisfied.

Drug discrimination procedure

Five monkeys discriminated Δ⁹-THC (0.1 mg·kg⁻¹, i.v.) from vehicle (1 part absolute ethanol, 1 part Emulphor-620 and 18 parts saline) while responding under a fixed ratio 5 (FR5) schedule of SST in a multiple-cycle procedure. Each cycle began with a 15-min timeout; responses during the timeout had no programmed consequence. The timeout was followed by a 5-min schedule of SST, which was signalled by illumination of red lights (one positioned above each lever). Five consecutive responses on the correct lever extinguished the red lights, prevented delivery of an electric stimulus and initiated a 30-s timeout. Otherwise, an electric stimulus was delivered every 40 s. Responding on the incorrect lever reset the response requirement on the correct lever. Determination of correct levers varied among monkeys (i.e. left lever associated with drug; right lever associated with vehicle) and remained the same for that monkey for the duration of the study.

Training sessions were conducted by administering Δ⁹-THC (0.1 mg·kg⁻¹) or vehicle within the first minute of a cycle, followed by vehicle or sham within the first minute of subsequent cycles. Δ⁹-THC training consisted of two cycles and was preceded by zero to four vehicle-training cycles; vehicle training consisted of two to six cycles. Completion of the FR on the correct lever was required for a reinforcer during each training cycle. Monkeys had previously satisfied the criteria for testing, that is, at least 80% of the total responses occurred on the correct lever and fewer than five responses occurred on the incorrect lever before completion of the FR on the correct lever for all cycles during five consecutive or six of seven training sessions. During test sessions, five consecutive responses on either lever postponed the shock schedule. Δ⁹-THC, CP 55940 and WIN 55212-2 were studied by administering vehicle in the first cycle, followed by cumulative i.v. doses increasing by 0.5 log unit in subsequent cycles. The dose-effect function included ineffective doses (i.e. doses producing responses predominantly on the vehicle lever) up to doses that produced greater than 80% of responses on the Δ⁹-THC lever.

The day after a Δ⁹-THC dose-effect determination, vehicle test sessions were conducted on consecutive days for 1, 3 or 6 days by administering vehicle in the first cycle followed by sham in a second cycle; monkeys received Δ⁹-THC (1 mg·kg⁻¹, s.c.) after each of these sessions. On the day after 1, 3 or 6 days of Δ⁹-THC treatment, a test session was conducted by administering vehicle in the first cycle followed by cumulative doses of Δ⁹-THC in subsequent cycles. After the 6-day treatment regimen, cumulative Δ⁹-THC dose-effect tests were conducted every other day until sensitivity was not different from the initial dose-effect findings, as determined for individual monkeys; vehicle-test sessions consisting of two cycles were conducted on intervening days. After sensitivity to Δ⁹-THC was determined before and after once daily Δ⁹-THC (1 mg·kg⁻¹) treatment for 1, 3, or 6 days, sensitivity to CP 55940 and WIN 55212-2 was determined in non-systematic order among monkeys before and after 3 days of Δ⁹-THC treatment, in the same manner as that described for Δ⁹-THC. Following tests with CP 55940 and WIN 55212-2, sensitivity to Δ⁹-THC was determined a second time before and after 3 days of Δ⁹-THC treatment. Successive episodes of daily Δ⁹-THC treatment were separated by at least 30 days. Before the initial dose-effect tests prior to Δ⁹-THC treatment, performance for consecutive training sessions, including both vehicle and drug training sessions, satisfied the test criteria. The type of training sessions preceding these test sessions varied non-systematically.

Drugs

Drug/molecular target nomenclature conforms to the British Journal of Pharmacology Guide to Receptors and Channels (Alexander et al., 2009). Rimonabant, Δ⁹-THC (100 mg·mL⁻¹ in absolute ethanol; The Research Technology Branch, National Institute on Drug Abuse, Rockville, MD, USA), CP 55940 ((1R,3R,4R)-3-[2-hydroxy-4-(1,1-dimethylheptyl) phenyl]-4-(3-hydroxypropyl)cyclohexan-1-ol; Tocris, Ellisville, MO, USA) and WIN 55212-2 ((R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo-[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone mesylate; Sigma, St. Louis, MO, USA) were dissolved in 1 part absolute ethanol, 1 part Emulphor-620 (Rhodia Inc., Cranbury, NJ, USA) and 18 parts physiological saline. Midazolam hydrochloride (5 mg·mL⁻¹ in physiological saline; Bedford Laboratories, Bedford, OH, USA) and ketamine hydrochloride (100 mg·mL⁻¹ in physiological saline; Bioniche, Athens, GA, USA) were diluted in physiological saline as needed. All drugs were administered i.v., and dose was expressed as the weight of the forms listed above.

Data analyses

Control rate of responding (responses·s⁻¹) was calculated as the average rate for all cycles in the five non-drug (i.e. vehicle) sessions immediately preceding a drug session, excluding vehicle sessions conducted the day after a drug session. The control response rate for data collected during chronic Δ⁹-THC treatment was calculated as the average rate of the five non-drug (i.e. vehicle) sessions immediately preceding Δ⁹-THC treatment. Response rate data were expressed as a percentage of the control response rate and data were analysed separately for each reinforcer (food presentation and SST). Discrimination data were expressed as a percentage of responses on the Δ⁹-THC lever out of total responses on both the Δ⁹-THC and vehicle levers. When response rate was decreased to less than 50% of the control or responding on the Δ⁹-THC lever was increased to greater than 50%, potency was calculated by simultaneously fitting straight lines to individual dose-effect data by means of GraphPad Prism version 5.0 for Windows (San Diego, CA, USA) with linear regression. Straight lines were fitted to the linear portion of dose-effect curves. The slopes of dose-effect curves were compared with an F-ratio test using GraphPad; if the slopes were not significantly different; then, a common, best-fitting slope was used for further analyses (Kenakin, 1997). Doses corresponding to the 50% level of the effect (ED₅₀ values), potency ratios and their 95% confidence limits were calculated with parallel line analyses of data from individual subjects (Tallarida, 2000). The potencies were considered significantly different when the 95% confidence limits of the potency ratio did not include 1.

Rimonabant alone (i.e. without Δ⁹-THC treatment) did not decrease responding to below 50% of control in all monkeys (see Results); differences in potency to modify rate of responding were examined by determining whether slopes of rimonabant dose–response curves were significantly different from 0. Rimonabant tended to decrease response rate when administered in single bolus doses before cumulative doses of Δ⁹-THC; therefore, antagonism of the rate-decreasing effects of Δ⁹-THC by rimonabant was analysed by expressing data as a percentage of the response rate following each respective dose of rimonabant alone.

Results

The effects of Δ⁹-THC to decrease response rate: antagonism by rimonabant

For 10 non-drug sessions, average (SEM) rates of lever-pressing for food presentation in four monkeys were 0.90 (0.05), 1.18 (0.06), 1.34 (0.10) and 1.58 (0.07) responses·s⁻¹. For the schedule of SST, average (SEM) rates of responding were 0.84 (0.06), 0.56 (0.05), 0.94 (0.06) and 1.84 (0.14) responses·s⁻¹ for the same monkeys respectively. Response rate did not significantly differ as a function of reinforcer (P > 0.05).

When administered as a single bolus dose, rimonabant (3.2 mg·kg⁻¹) decreased responding to 54% and 70% of control for food presentation and SST, respectively [Figure 1, top, diamonds above vehicle (VEH)]; the slope of the rimonabant dose–response curve was significantly different from 0 when responding was maintained by SST (P < 0.05) but not food presentation (P > 0.05). Δ⁹-THC produced a marked, dose-dependent decrease in responding for food and SST (Figure 1, circles, left and right, respectively). The ED₅₀ values determined from the two control Δ⁹-THC dose–response curves (i.e. one determined before and the other after tests with rimonabant in combination with Δ⁹-THC) were not significantly different from each other; these data were averaged for graphic presentation (Figure 1) and further analyses. The slopes of the Δ⁹-THC dose–response curves alone and in combination with various doses of rimonabant were not significantly different. The ED₅₀ values (95% confidence limits) for Δ⁹-THC to decrease responding for food presentation and SST were 0.17 and 0.058 mg·kg⁻¹ respectively (Table 1). Rimonabant dose-dependently antagonized the effects of Δ⁹-THC to decrease responding. For food presentation, rimonabant (3.2 mg·kg⁻¹) significantly decreased the potency of Δ⁹-THC 14-fold; smaller doses of rimonabant did not produce significant antagonism (Table 1, Food). For SST, doses of 1 and 3.2 mg·kg⁻¹ of rimonabant significantly decreased the potency of Δ⁹-THC 6.7- and 14-fold, respectively, whereas the smallest dose (0.32 mg·kg⁻¹) of rimonabant did not produce significant antagonism (Table 1, SST).

Figure 1.

Effects of rimonabant and Δ⁹-tetrahydrocannabinol (Δ⁹-THC), alone and in combination, on responding for food presentation (left) and stimulus-shock termination (SST; right). Abscissae: vehicle (VEH) or i.v. dose of Δ⁹-THC in mg·kg⁻¹ body weight. Ordinates: mean (±SEM) rate of responding expressed as a percentage of control. Points above VEH represent the average response rate after vehicle or a dose (0.32-3.2 mg·kg⁻¹, i.v.) of rimonabant alone. Rimonabant (0.32-3.2 mg·kg⁻¹) was administered prior to cumulative doses of Δ⁹-THC.

Table 1.

ED₅₀ values and 95% confidence limits for Δ⁹-tetrahydrocannabinol (Δ⁹-THC) alone and in combination with various doses (0.32–3.2 mg·kg⁻¹) of rimonabant in monkeys responding for food presentation (Food) and stimulus-shock termination (SST)

	ED50 in mg·kg−1 (95% confidence limits)	Potency ratio† (95% confidence limits)
Food
Δ9-THC alone	0.17 (0.055–0.54)
+0.32 Rimonabant	0.22 (0.097–0.48)	1.3 (0.4–4.0)
+1 Rimonabant	0.60 (0.19–1.9)	3.5 (1.0–13)
+3.2 Rimonabant	2.3* (0.60–8.8)	14 (3.6–50)
SST
Δ9-THC alone	0.058 (0.024–0.14)
+0.32 Rimonabant	0.13 (0.058–0.38)	2.2 (0.6–8.6)
+1 Rimonabant	0.39* (0.06–2.6)	6.7 (1.1–41)
+3.2 Rimonabant	0.79* (0.31–2.0)	14 (4.0–46)

^†Potency ratios and 95% confidence limits are equal to the ED₅₀ values of Δ⁹-THC in combination with a dose of rimonabant divided by the ED₅₀ value of Δ⁹-THC alone.

^*Significantly different from Δ⁹-THC alone.

The effects of drugs on response rate before and during chronic Δ⁹-THC treatment

In three monkeys responding for food presentation and SST under the multiple schedule, average (SEM) rates of lever-pressing for 10 non-drug sessions conducted before chronic Δ⁹-THC treatment were 0.46 (0.02), 0.51 (0.04) and 0.64 (0.03) responses·s⁻¹ for food presentation and 1.75 (0.12), 2.35 (0.28) and 1.70 (0.09) responses·s⁻¹ for SST for each respective monkey. Response rate was significantly lower for food presentation as compared with SST under the multiple schedule (P < 0.05).

Before and during chronic Δ⁹-THC treatment, Δ⁹-THC, CP 55940 and WIN 55212-2 markedly decreased responding for food presentation and SST (Figure 2). Analysis of dose–response curves separately for each drug and reinforcer showed that the slopes did not significantly differ as a function of Δ⁹-THC treatment (P > 0.05). Before chronic Δ⁹-THC treatment, the ED₅₀ values of Δ⁹-THC were 0.094 and 0.047 mg·kg⁻¹ in the food and SST components respectively (Table 2). The ED₅₀ values of CP 55940 were 0.0085 and 0.0095 mg·kg⁻¹ in the food and SST components respectively (Table 2); CP 55940 was 11- and 5-fold more potent than Δ⁹-THC in the respective schedule components. WIN 55212-2 was slightly less potent than Δ⁹-THC, that is, the ED₅₀ values of WIN 55212-2 were 0.24 and 0.21 mg·kg⁻¹ in the food and SST components respectively.

Figure 2.

Effect of Δ⁹-tetrahydrocannabinol (Δ⁹-THC) (top), CP 55940 (middle) and WIN 55212-2 (bottom) on responding for food presentation (left) and stimulus-shock termination (SST; right) before and during chronic Δ⁹-THC (1 mg·kg⁻¹ per 12 h, s.c.). Abscissae: vehicle (VEH) or dose administered i.v. in mg·kg⁻¹ body weight. Ordinates: mean (±SEM) rate of responding expressed as a percentage of control. Squares (top) show the effects Δ⁹-THC determined after discontinuation of Δ⁹-THC (1 mg·kg⁻¹ per 12 h, s.c.) treatment.

Table 2.

ED₅₀ values and 95% confidence limits of drugs determined before and during daily Δ⁹-tetrahydrocannabinol (Δ⁹-THC) treatment (1 mg·kg⁻¹ per 12 h, s.c.) in monkeys responding for food presentation (Food) and stimulus-shock termination (SST)

	ED50 in mg·kg−1 (95% confidence limits)	Potency ratio† (95% confidence limits)
Δ9-THC food
Before	0.094 (0.013–0.65)
During	2.2* (0.27–17)	23 (1.8–290)
After††	0.31 (0.028–3.5)	3.3 (0.2–49)
Δ9-THC SST
Before	0.047 (0.016–0.14)
During	7.6* (1.2–46)	160 (27–930)
After††	0.24* (0.066–0.90)	5.1 (1.2–22)
CP 55940 food
Before	0.0085 (0.0024–0.029)
During	0.082* (0.019–0.34)	9.6 (1.9–49)
CP 55940 SST
Before	0.0095 (0.0052–0.017)
During	0.085* (0.047–0.15)	8.9 (4.2–19)
WIN 55212-2 food
Before	0.24 (0.037–1.6)
During	1.2 (0.32–4.4)	4.9 (0.8–32)
WIN 55212-2 SST
Before	0.21 (0.047–0.93)
During	2.4* (0.85–6.7)	11 (2.4–54)
Rimonabant food
Before	3.2
During	0.87 (0.17–4.5)	0.27^
Rimonabant SST
Before	3.2
During	0.50 (0.21–1.2)	0.16^
Midazolam food
Before	0.17 (0.05–0.59)
During	0.24 (0.08–0.73)	1.4 (0.4–4.9)
Midazolam SST
Before	0.057 (0.029–0.12)
During	0.073 (0.04–0.13)	1.3 (0.6–2.9)
Ketamine food
Before	1.2 (0.81–1.7)
During	0.61 (0.33–1.1)	0.51 (0.2–1.1)
Ketamine SST
Before	1.2 (0.59–2.6)
During	0.84 (0.45–1.6)	0.70 (0.3–1.7)

^†Potency ratios and 95% confidence limits are equal to the ED₅₀ values determined during or after discontinuation of daily Δ⁹-THC treatment divided by the ED₅₀ value determined before daily Δ⁹-THC treatment.

^††ED₅₀ values and 95% confidence limits for Δ⁹-THC determined after discontinuation of Δ⁹-THC treatment also are shown.

^*Significantly different from Before.

^{^}ED₅₀ value Before equal to largest dose studied; 95% confidence limits could not be calculated.

Upon chronic Δ⁹-THC (1 mg·kg⁻¹ per 12 h, s.c.) treatment, responding in both schedule components was decreased as compared with before treatment (Figure 3, compare Days 1–24, during chronic Δ⁹-THC to before). Beginning on day 24 of chronic Δ⁹-THC treatment, sensitivity to Δ⁹-THC and other drugs was assessed. Sensitivity to the rate-decreasing effects of Δ⁹-THC was decreased during chronic Δ⁹-THC treatment (Figure 2, top, compare diamonds to circles). Before chronic treatment, 0.32 mg·kg⁻¹ of Δ⁹-THC decreased responding in the food and SST components to 27% and 12% of control respectively. During chronic Δ⁹-THC treatment, larger doses (3.2 and 10 mg·kg⁻¹) were required to produce comparable decreases in responding. Chronic Δ⁹-THC treatment significantly increased the ED₅₀ value of Δ⁹-THC by 23- and 160-fold in the food and SST components respectively (Table 2). Sensitivity to the rate-decreasing effects of CP 55940 also was decreased during chronic Δ⁹-THC treatment (Figure 2, middle); the ED₅₀ values of CP 55940 were increased 9.6- and 8.9-fold in the food and SST components respectively (Table 2). During chronic Δ⁹-THC treatment, the potency of WIN 55212-2 to decrease responding in the SST component decreased significantly by 11-fold (Table 2). Although the potency of WIN 55212-2 to decrease responding for food was less during chronic Δ⁹-THC as compared with before treatment, the difference was not statistically significant. Overall, during chronic Δ⁹-THC treatment, sensitivity to Δ⁹-THC decreased more than sensitivity to CP 55490 and WIN 55212-2 (Figure 2, compare top panel to middle and bottom panels).

Figure 3.

Vehicle performance before, during and after discontinuation of chronic Δ⁹-tetrahydrocannabinol (Δ⁹-THC) treatment in monkeys responding for food presentation (top) and stimulus-shock termination (SST; bottom). Abscissae: consecutive sessions before, during and after discontinuation of chronic Δ⁹-THC (1 mg·kg⁻¹ per 12 h, s.c.) treatment. Ordinates: mean (±SEM) rate of responding expressed as a percentage of control.

Rimonabant, when administered up to a cumulative dose of 3.2 mg·kg⁻¹ before chronic Δ⁹-THC treatment, decreased responding in the food and SST components to a maximum of 68% and 85% of control respectively (Figure 4; circles). The effects of rimonabant before chronic Δ⁹-THC did not significantly vary as a function of dose, that is, the slope of the dose–response curve was not significantly different from 0 (P > 0.05). During chronic Δ⁹-THC treatment, rimonabant (1 mg·kg⁻¹) decreased responding in the food and SST components to 39% and 35%, respectively; the dose–response curve for rimonabant during chronic Δ⁹-THC treatment was significantly different from 0 in the food component (F_1,13= 4.82; P < 0.05) and approached significance in the SST component (F_1,13= 3.20; P= 0.09). The ED₅₀ values for rimonabant during chronic Δ⁹-THC were 0.87 and 0.50 mg·kg⁻¹ for the food and SST components respectively (Table 2). Before chronic Δ⁹-THC treatment, a conservative estimate of the ED₅₀ was the largest dose (3.2 mg·kg⁻¹) tested. Although sensitivity to rimonabant was increased 3.7- and 6.4-fold for the respective schedule components, the data were not amenable to statistical analysis (i.e. 95% confidence limits of the potency ratios could not be calculated).

Figure 4.

Effects of rimonabant on responding for food presentation (left) and stimulus-shock termination (SST; right) before, during and after discontinuation of chronic Δ⁹-tetrahydrocannabinol (Δ⁹-THC); (1 mg·kg⁻¹ per 12 h, s.c.). Abscissae: vehicle (VEH) or dose administered i.v. in mg·kg⁻¹ body weight. Ordinates: mean (±S.E.M.) rate of responding expressed as a percentage of control.

Before chronic Δ⁹-THC treatment, the non-cannabinoids midazolam and ketamine dose-dependently decreased responding (Figure 5, circles). In the food and SST components, the ED₅₀ values for midazolam were 0.17 and 0.057 mg·kg⁻¹, respectively; the ED₅₀ values for ketamine were 1.2 mg·kg⁻¹ in both schedule components (Table 2). In contrast to the cannabinoids, sensitivity to the effects of midazolam and ketamine were not significantly modified by chronic Δ⁹-THC treatment; the ED₅₀ values did not differ by more than 2-fold before as compared with during chronic Δ⁹-THC treatment (Table 2).

Figure 5.

Effects of midazolam (top) and ketamine (bottom) on responding for food presentation (left) and stimulus-shock termination (SST) before and during chronic Δ⁹-tetrahydrocannabinol (Δ⁹-THC) (1 mg·kg⁻¹ per 12 h, s.c.). Abscissae: vehicle (VEH) or dose administered i.v. in mg·kg⁻¹ body weight. Ordinates: mean (±SEM) rate of responding expressed as a percentage of control.

When chronic Δ⁹-THC (1 mg·kg⁻¹ per 12 h, s.c.) treatment was discontinued, there was a trend for food-maintained responding to decrease and then increase over days (Figure 3, top), whereas SST-maintained responding was elevated during the first session conducted during the discontinuation period and, thereafter, not substantially modified relative to control (Figure 3, bottom). In the food component, sensitivity to Δ⁹-THC determined after discontinuation of chronic Δ⁹-THC treatment was no longer significantly different from sensitivity determined before chronic Δ⁹-THC treatment (Figure 2, top left, compare squares and circles; Table 2). In contrast, in the SST component, sensitivity to Δ⁹-THC remained significantly decreased (5.1-fold) after discontinuation of chronic treatment as compared with before chronic treatment, albeit the ED₅₀ value of Δ⁹-THC determined after discontinuation of Δ⁹-THC treatment was markedly (32-fold) less than the ED₅₀ value of Δ⁹-THC determined during chronic treatment (Figure 2, top right, compare square and diamonds; Table 2). After discontinuation of Δ⁹-THC treatment, sensitivity to the rate-decreasing effects of rimonabant was less than that determined during chronic Δ⁹-THC treatment (Figure 4); up to a dose of 3.2 mg·kg⁻¹, the slope of the dose–response functions in each schedule component were not significantly different from 0 (P > 0.05).

Discriminative stimulus effects of drugs before and immediately after Δ⁹-THC treatment

In monkeys discriminating Δ⁹-THC, the training drug dose-dependently increased the percentage of responses on the Δ⁹-THC lever, that is, at the training dose (0.1 mg·kg⁻¹, i.v.), percentage responding on the Δ⁹-THC lever was 100% (Figure 6, top, circles). In contrast, vehicle produced 0% of the responses on the Δ⁹-THC lever (Figure 6, top, circles above VEH). During the test conducted 1 day after a single injection of 1 mg·kg⁻¹ of Δ⁹-THC s.c., % Δ⁹-THC lever responding was 25% when monkeys received vehicle during that test (Figure 6, top left, diamond above VEH). In subsequent cycles, Δ⁹-THC dose-dependently increased Δ⁹-THC appropriate responding (Figure 6, top left, diamonds). The slopes and ED₅₀ values (Table 3) of the Δ⁹-THC dose–response curves determined before and after 1 day of Δ⁹-THC treatment (1 mg·kg⁻¹, s.c.) were not significantly different (P > 0.05). In contrast, both 3 and 6 days of treatment with Δ⁹-THC (1 mg·kg⁻¹, s.c.) produced a significant decrease in the potency of Δ⁹-THC to produce discriminative stimulus effects (Figure 6, top, three rightmost panels). Sensitivity to Δ⁹-THC decreased comparably (2.7- to 3.0-fold; Table 3) for 6 days of treatment and both determinations of the 3-day Δ⁹-THC treatment, that is, one occurring before and the other after studies with CP 55490 and WIN 55212-2. Sensitivity to Δ⁹-THC was assessed periodically after the 6-day treatment without intervening training sessions; the ED₅₀ value was no longer significantly different from control 7 days after treatment was discontinued (Figure 6, top right, compare triangle and circles; Table 3). For 10 non-drug sessions, average (SEM) rates of lever-pressing in five monkeys were 0.84 (0.07), 0.85 (0.04), 0.90 (0.03), 0.92 (0.10) and 1.63 (0.03) responses·s⁻¹. Up to a dose of 0.32 mg·kg⁻¹, Δ⁹-THC did not significantly modify response rate during any of the tests conducted before and after 1, 3 or 6 days of Δ⁹-THC treatment (P > 0.05).

Figure 6.

Discriminative stimulus effects of Δ⁹-tetrahydrocannabinol (Δ⁹-THC) in rhesus monkeys before and after 1, 3 or 6 days of Δ⁹-THC (1 mg·kg⁻¹ per day, s.c.). Abscissae: vehicle (VEH) or dose administered i.v. of Δ⁹-THC in mg·kg⁻¹ body weight. Ordinates: mean (± SEM) percentage of responding on the Δ⁹-THC lever (top) and rate of responding expressed as a percentage of control (bottom). The effects of the 3-day treatment were determined twice (First and Second; see Methods for details). The squares and triangles (rightmost panel) show sensitivity to Δ⁹-THC determined 3 and 7 days after discontinuation of the 6-day treatment regimen respectively.

Table 3.

ED₅₀ values and 95% confidence limits determined before and 1 day after Δ⁹-tetrahydrocannabinol (Δ⁹-THC) treatment (1 mg·kg⁻¹ per day, s.c. for up to 6 days) in monkeys discriminating Δ⁹-THC

	ED50 in mg·kg−1 (95% confidence limits)	Potency ratio† (95% confidence limits)
Δ9-THC
1 day Δ9-THC
Before	0.041 (0.013–0.13)
1 day after	0.030 (0.0097–0.093)	0.7 (0.2–2.9)
3 days Δ9-THC (First)
Before	0.046 (0.021–0.097)
1 day after	0.13* (0.068–0.24)	2.8 (1.3–6.4)
3 days Δ9-THC (Second)
Before	0.021 (0.012–0.037)
1 day after	0.057* (0.030–0.11)	2.7 (1.3–5.7)
6 days Δ9-THC
Before	0.037 (0.022–0.069)
1 day after	0.11* (0.080–0.16)	3.0 (1.6–5.4)
7 days after	0.042 (0.025–0.077)	1.1 (0.5–2.7)
CP 55940
3 days Δ9-THC
Before	0.0045 (0.0020–0.010)
1 day after	0.0047 (0.0025–0.0090)	1.0 (0.4–2.4)
WIN 55212-2
3 days Δ9-THC
Before	0.12 (0.032–0.44)
1 day after	0.083 (0.029–0.24)	0.7 (0.2–2.7)

^†Potency ratios and 95% confidence limits are equal to the ED₅₀ values determined 1 day after Δ⁹-THC treatment divided by the ED₅₀ value determined before daily Δ⁹-THC treatment.

^*Significantly different from before.

CP 55940 and WIN 55212-2 both dose-dependently increased responding on the Δ⁹-THC lever (Figure 7, top left and right, respectively, circles); the ED₅₀ values were 0.0045 and 0.12 mg·kg⁻¹ respectively (Table 3). When the dose–response curves were re-determined after 3 days of Δ⁹-THC (1 mg·kg⁻¹) treatment (Figure 7, top, diamonds), the potencies of CP 55940 and WIN 55212-2 were not significantly different from their respective controls (P > 0.05; Table 3). There was no significant relationship between dose (up to 0.032 mg·kg⁻¹) of CP 55940 and rate of responding (Figure 7, bottom left). WIN 55212-2, up to a dose of 0.32 mg·kg⁻¹, tended to decrease response rate, that is, the slope of the WIN 55212-2 dose–response curve determined before Δ⁹-THC treatment was significantly different from 0 (P < 0.05; Figure 7, bottom right, circles).

Figure 7.

Discriminative stimulus effects of CP 55940 (left) and WIN 55212-2 (right) in rhesus monkeys before and after 3 days of Δ⁹-tetrahydrocannabinol (Δ⁹-THC); (1 mg·kg⁻¹ per day, s.c.). Abscissae: vehicle (VEH) or dose administered i.v. in mg·kg⁻¹ body weight. Ordinates: mean (± SEM) percentage of responding on the Δ⁹-THC lever (top) and rate of responding expressed as a percentage of control (bottom).

Discussion

Chronic Δ⁹-THC treatment resulted in tolerance when measuring the effects of Δ⁹-THC to decrease rates of operant responding and to produce discriminative stimulus effects; the development of tolerance was time-dependent and reversible (i.e. tolerance diminished after treatment was discontinued). Tolerance to Δ⁹-THC was accompanied by a smaller loss of sensitivity to the rate-decreasing effects of CP 55940 and WIN 55212-2 (i.e. cross-tolerance), whereas sensitivity to the rate-decreasing effects of the cannabinoid antagonist rimonabant increased. Chronic Δ⁹-THC treatment had pharmacologically selective effects inasmuch as there was no change in sensitivity to the rate-decreasing effects of non-cannabinoids (midazolam and ketamine). The potency of drugs to decrease responding and the pattern of tolerance/cross-tolerance did not vary with the type of reinforcer (food presentation versus SST). For discriminative stimulus effects, tolerance to Δ⁹-THC developed without cross-tolerance to CP 55940 and WIN 55212-2. The inverse relationship between magnitude of tolerance/cross-tolerance in the current study and cannabinoid agonist efficacy of the drugs to which tolerance/cross-tolerance developed, as established elsewhere in vitro (Breivogel and Childers, 2000), suggests that CB₁ agonist efficacy is an important determinant of behavioural effects.

According to receptor theory (Kenakin, 1997), low efficacy agonists occupy more receptors than high efficacy agonists at equal levels of effect. When CB₁ receptor number and function are decreased (e.g. by chronic Δ⁹-THC treatment; Sim et al., 1996; Breivogel et al., 1999), fractional occupancy and dose need to increase in order to retain the effect (i.e. tolerance and cross-tolerance develop). Because low efficacy agonists occupy a relatively large number of receptors, receptor down-regulation produces a disproportionate change and greater increase in the fractional occupancy needed to retain effect for a low as compared with a high efficacy agonist which, in turn, can result in greater tolerance/cross-tolerance to the low efficacy agonist as compared with the high efficacy agonist. With sufficient receptor down-regulation, the maximum effect can also decrease for the low efficacy agonist. Chronic Δ⁹-THC treatment produced a greater decrease in the potency of the low efficacy agonist Δ⁹-THC than the higher efficacy agonists CP 55940 and WIN 55212-2. These results strongly suggest that, for some behavioural effects, Δ⁹-THC needs to occupy a greater number of cannabinoid receptors than CP 55940 and WIN 55212-2 for all three to achieve the same level of effect. However, the maximum effect was unchanged, suggesting the presence of unoccupied or spare CB₁ receptors.

An alternative explanation for these data is that Δ⁹-THC, CP 55940 and WIN 55212-2 vary in selectivity for different receptor subtypes including CB₁, CB₂ and perhaps other receptors (Palmer et al., 2002). However, in rhesus monkeys discriminating 0.1 mg·kg⁻¹ of Δ⁹-THC i.v. (i.e. the discrimination assay used in the current study), the results of a previous study strongly suggested that Δ⁹-THC, CP 55940 and WIN 55212-2 produced their discriminative stimulus effects through a common receptor type (McMahon, 2006). In that study, Schild analysis of rimonabant in combination with Δ⁹-THC, CP 55940 and WIN 55212-2 was consistent with simple, competitive and reversible antagonism; moreover, the potency (i.e. apparent affinity or pA₂) of rimonabant was similar in the presence of all three agonists. Rimonabant is more selective for CB₁ than CB₂ receptors (Rinaldi-Carmona et al., 1994); moreover, a selective CB₂ agonist did not substitute for the Δ⁹-THC discriminative stimulus, and the Δ⁹-THC discriminative stimulus was not blocked by a selective CB₂ antagonist (McMahon, 2006). Collectively, these results suggest that differences in the magnitude of tolerance and cross-tolerance to discriminative stimulus effects is not due to differences in receptor-subtype selectivity, but rather to differences in agonist efficacy at a common site (i.e. CB₁ receptors).

The effects of drugs to decrease fixed ratio responding have been used to assess tolerance and cross-tolerance to a variety of drugs including γ-aminobutyric acid_A modulators (McMahon and France, 2002) and opioids (Brandt and France, 2000). However, relatively large doses of drug are often required to decrease operant responding and the mechanisms underlying rate-decreasing effects sometimes differ from the mechanisms responsible for other behavioural effects (e.g. discriminative stimulus effects) of the same drug. For example, low efficacy µ agonists (e.g. nalbuphine) shared discriminative stimulus effects with other µ opioids and those effects were blocked by the µ opioid antagonist naltrexone; in contrast, the rate-decreasing effects of nalbuphine were not blocked by naltrexone (Walker and Young, 1993; Walker et al., 1994). However, in the current study, rimonabant antagonized the rate-decreasing effects of Δ⁹-THC, implicating a CB₁ receptor mechanism. Antagonism occurred at doses of rimonabant that also decreased responding, consistent with previous studies (Järbe et al., 2003; De Vry and Jentzsch, 2004). The marked tolerance to Δ⁹-THC (23- and 160-fold for food presentation and SST, respectively) and the lesser cross-tolerance to CP 55940 (10- and 9-fold, respectively) and WIN 55212-2 (5- and 11-fold, respectively) are consistent with a rank order efficacy in vivo of WIN 55212-2 ≥ CP 55940 > Δ⁹-THC, similar to what has been reported at CB₁ receptors in vitro (Breivogel and Childers, 2000). However, the current study cannot reject the possibility that differential tolerance/cross-tolerance is due, in part, to differential binding to multiple receptor subtypes that mediate rate-decreasing effects. Moreover, drug metabolism and disposition were not assessed in the current study to determine whether chronic treatment selectively increases metabolism of Δ⁹-THC. However, previous studies suggested that Δ⁹-THC tolerance was not accompanied by significant changes in metabolism in birds (Dewey et al., 1973; McMillan et al., 1973), rats (Siemens and Kalant, 1974), dogs (Martin et al., 1976) and humans (Hunt and Jones, 1980; but see Lemberger et al., 1971).

Tolerance to Δ⁹-THC has been previously demonstrated for many effects, including discriminative stimulus effects and effects on rates of operant responding, rectal temperature, locomotor activity, nociception, ingestion and learning (McMillan et al., 1970; Miczek, 1979; Wiley et al., 1993; Fan et al., 1994; Delatte et al., 2002; McKinney et al., 2008). When treatment was constant, tolerance to Δ⁹-THC did not develop equally for all effects (e.g. Miczek, 1979; Fan et al., 1994), perhaps reflecting differences in cannabinoid receptor number and G-protein-coupling efficiency in brain regions that differentially mediated the various effects of Δ⁹-THC. In the current study, tolerance to Δ⁹-THC for producing rate-decreasing and discriminative stimulus effects was generated with different treatment parameters, thereby precluding assessment of whether tolerance develops equally for the two behavioural measures.

Chronic agonist treatment with opioids (Goldberg and Schuster, 1967) and benzodiazepines (McMahon and France, 2002) can increase the sensitivity of a pharmacologically-related antagonist to decrease rates of operant responding. This can reflect agonist dependence inasmuch as the effects of the antagonist are due to withdrawal. Δ⁹-THC treatment increased the sensitivity of monkeys to the rate-decreasing effects of rimonabant, consistent with rimonabant inducing Δ⁹-THC withdrawal. However, in the absence of Δ⁹-THC treatment, rimonabant tended to decrease responding at a dose (3.2 mg·kg⁻¹) 0.5–1 log unit larger than the smallest doses decreasing response rate during chronic Δ⁹-THC treatment. The mechanism(s) responsible for the effects of rimonabant alone have not been established. Δ⁹-THC treatment (e.g. 1 mg·kg⁻¹ per 12 h, s.c.) was shown in another study to increase sensitivity to other effects of rimonabant (e.g. head shaking) in rhesus monkeys (e.g. Stewart and McMahon, 2010). Such enhancement might reflect CB₁ receptor inverse agonism (Bouaboula et al., 1997) in as much as chronic treatment with other types of agonist can increase sensitivity to inverse agonists at the same receptor (e.g. benzodiazepine receptor ligands; Sannerud et al., 1991).

In summary, in Δ⁹-THC-treated monkeys, there was a predictable inverse relationship between the magnitude of tolerance and cross-tolerance and the efficacy of the agonists to which tolerance and cross-tolerance developed, suggesting that cannabinoid agonist efficacy is an important determinant of behavioural effects. These results have significant implications for marijuana and other cannabinoid use inasmuch as the effectiveness of low efficacy agonists will be impacted to a greater degree than high efficacy agonists. Overall, these studies underscore the broad utility of using receptor theory to predict the effects of G-protein-coupled receptor agonists in animals, and suggest that the potential psychotherapeutic and other behavioural effects of cannabinoids depend on the magnitude to which they stimulate CB₁ receptors and downstream effector mechanisms.

Glossary

Abbreviations

CB

cannabinoid

CP 55940

(1R,3R,4R)-3-[2-hydroxy-4-(1,1-dimethylheptyl) phenyl]-4-(3-hydroxypropyl)cyclohexan-1-ol

Δ⁹-THC

Δ⁹-tetrahydrocannabinol

FR

fixed ratio

WIN 55212-2

(R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo-[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone mesylate

Conflicts of interest

None.

Supporting Information

Supporting Information: Teaching Materials; Figs 1–7 as PowerPoint slide.