Abstract
Synthetic cannabinoids (SCs) are an emerging class of abused drugs that differ from each other and the phytocannabinoid ∆9-tetrahydrocannabinol (THC) in their safety and cannabinoid-1 receptor (CB1R) pharmacology. As efficacy represents a critical parameter to understanding drug action, the present study investigated this metric by assessing in vivo and in vitro actions of THC, two well-characterized SCs (WIN55,212-2 and CP55,940), and three abused SCs (JWH-073, CP47,497, and A-834,735-D) in CB1 (+/+), (+/−), and (−/−) mice. All drugs produced maximal cannabimimetic in vivo effects (catalepsy, hypothermia, antinociception) in CB1 (+/+) mice, but these actions were essentially eliminated in CB1 (−/−) mice, indicating a CB1R mechanism of action. CB1R efficacy was inferred by comparing potencies between CB1 (+/+) and (+/−) mice [+/+ ED50 /+/− ED50], the latter of which has a 50% reduction of CB1Rs (i.e., decreased receptor reserve). Notably, CB1 (+/−) mice displayed profound rightward and downward shifts in the antinociception and hypothermia dose-response curves of low-efficacy compared with high-efficacy cannabinoids. In vitro efficacy, quantified using agonist-stimulated [35S]GTPγS binding in spinal cord tissue, significantly correlated with the relative efficacies of antinociception (r = 0.87) and hypothermia (r = 0.94) in CB1 (+/−) mice relative to CB1 (+/+) mice. Conversely, drug potencies for cataleptic effects did not differ between these genotypes and did not correlate with the in vitro efficacy measure. These results suggest that evaluation of antinociception and hypothermia in CB1 transgenic mice offers a useful in vivo approach to determine CB1R selectivity and efficacy of emerging SCs, which shows strong congruence with in vitro efficacy.
Introduction
Synthetic cannabinoids (SCs), comprising diverse structures and largely unknown pharmacology (Kronstrand et al., 2013; Louis et al., 2014; Sobolevsky et al., 2015), have emerged as drugs of abuse and represent a significant public health concern (Law et al., 2015; Trecki et al., 2015). In contrast to Δ9-tetrahydrocannabinol (THC), the primary psychoactive constituent of cannabis, SCs have been linked to physiologic toxicity (Freeman et al., 2013; Takematsu et al., 2014), greater psychologic complications (Celofiga et al., 2014; Meijer et al., 2014; Schwartz et al., 2015), and death (Behonick et al., 2014; Gerostamoulos et al., 2015; Shanks et al., 2015; Westin et al., 2015). These clinical observations suggest that SCs pose a greater public safety threat than cannabis/THC. Efforts to predict and limit the clinical harm of SC toxicity would benefit from new methods to characterize the pharmacology of emerging compounds.
The pharmacologic mechanisms that underlie heightened risk for medical complications by SCs compared with cannabis/THC are unknown and may vary according to the particular SC under consideration. Two important dimensions of SC pharmacology are efficacy at the cannabinoid receptor 1 (CB1R) and potential non-CB1R sites of action. Similar to THC, SCs bind and activate G-protein–coupled CB1Rs, which play an important role in mediating pharmacologic effects produced by marijuana and CB1R agonists (Rinaldi-Carmona et al., 1994; Zimmer et al., 1999; Huestis et al., 2001). Based largely on results from in vitro assays of agonist-stimulated [35S]GTPγS binding (Burkey et al., 1997a,b; Selley et al., 1996; Sim et al., 1996), however, THC is a low-efficacy CB1R agonist, whereas many SCs function as high-efficacy CB1R agonists (Wiley et al., 2013, 2015). Accordingly, heightened CB1R efficacy may contribute to the heightened risk for adverse effects of SCs; however, existing in vivo assays used to assess cannabinoid effects have poor resolution or slow throughput for distinguishing CB1R agonist efficacy. For example, although THC and the SC WIN55,212-2, respectively, possess low and high CB1R efficacy in stimulating [35S]GTPγS binding (Selley et al., 1996; Breivogel et al., 1998; Griffin et al., 1998), these drugs produce comparable maximal effects in commonly used assays (catalepsy, hypothermia, and antinociception) to assess the in vivo pharmacology of cannabinoids in mice (Fan et al., 1994).
The present study used a novel strategy for using CB1 (+/+), (+/−), and (−/−) mice to facilitate efficient and precise in vivo evaluation of SC selectivity for, and efficacy at, CB1Rs. Specifically, the effects of six cannabinoids and two noncannabinoid comparison drugs were examined in all three mouse genotypes using well-established in vivo assays sensitive to cannabimimetic activity (i.e., catalepsy, hypothermia, antinociception). We made two predictions. First, we predicted that drug selectivity at CB1Rs for antinociception, hypothermia, and catalepsy will be revealed by comparing effects in CB1 (+/+) and (−/−) mice, such that selective CB1R agonists will produce effects in CB1 (+/+) mice, but not in CB1 (−/−) mice. Second, we predicted that differences in drug efficacy at CB1Rs will be revealed by comparing dose-effect curves between CB1 (+/+) and (+/−) mice. CB1R density is 50% lower in CB1 (+/−) than in (+/+) mice (Selley et al., 2001) and decreases in receptor density produce greater rightward or downward shifts in dose-effect curves for low- than for high-efficacy agonists (Comer et al., 1992). Accordingly, we predicted that dose-effect curves will be shifted further rightward or downward for low- than high-efficacy cannabinoids in CB1 (+/−) relative to (+/+) mice. The drugs selected for study included THC, the well-characterized SCs CP55,940 and WIN55,212-2, two SCs associated with abuse CP47,497, JWH-073 (Atwood et al., 2011), and the heat degradant of the SC A-834,735 (Frost et al., 2010). Published (Griffin et al., 1998; Grim et al., 2016), and preliminary data suggest that these compounds display a range of CB1R efficacies as assessed in vitro by maximal stimulation of [35S]GTPγS binding, such that low (e.g., THC), moderate (e.g., JWH-073, CP47,497), and high (e.g., CP55,940, WIN55,212-2, A-834,735D) degrees of CB1R efficacy are represented. The noncannabinoids morphine and chlorpromazine, which are active in subsets of these assays (Wiley and Martin, 2003), served as comparison drugs. Additionally, [3H]SR141716A binding was conducted to confirm a 50% decrease in CB1R density in (+/−) mice. Finally, the cannabinoids were evaluated for their efficacy to stimulate [35S]GTPγS binding in transgenic mouse tissue to provide an in vitro correlate of in vivo efficacy measures. The in vitro assays used cerebellar and spinal cord tissues, which represent central nervous system (CNS) regions expressing high and low CB1R density, respectively. Correlations between in vivo and in vitro measures of efficacy were used to identify optimal strategies for in vivo assessment of drug efficacy at CB1Rs.
Materials and Methods
Subjects
Male and female CB1 (+/+), (+/−), and (−/−) mice (Zimmer et al., 1999) derived from CB1 (+/−) breeding pairs backcrossed at least 15 generations on a C57BL/6J background served as subjects. Mice had ad libitum access to food and water and were maintained on a 12-hour light/dark cycle. A total of 224 mice between 8 and 36 weeks of age were used for all experiments, which consisted of 192 mice in the in vivo experiments with 24 mice per drug, with n = 7–10 for each genotype, such that at least seven CB1R (+/+), seven (+/−), and seven (−/−) mice. Thirty-two mice for the in vitro studies in which cerebella and spinal cords were dissected from 12 (+/+), 12 (+/−), and 8 (−/−) mice. In all cases, at least three male and three female mice were included in each genotype, for each experiment. The experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Virginia Commonwealth University Institutional Animal Care and Use Committee.
Drugs
Studies used THC and the following five SCs reported to have higher efficacy than THC at CB1Rs (in order from purported highest to lowest efficacy): A-834,735 degradant (A-834,735D), WIN55,212-2 and CP55,940, JWH-073, and CP47,497 (Breivogel et al., 1998; Griffin et al., 1998; Auwärter et al., 2009; Atwood et al., 2011; Thomas and Wiley, 2014; Grim et al., 2016).The μ opioid receptor agonist morphine and dopamine receptor antagonist chlorpromazine were tested as controls that were expected to produce in vivo effects independent of CB1R density. A-834,735D, WIN55,212-2, CP47,497, JWH-073, and chlorpromazine were obtained from Cayman Chemical (Ann Arbor, MI), and morphine, THC, and CP55,940 were generously supplied by the National Institute on Drug Abuse Drug Supply Program. Each drug was administered in a vehicle consisting of ethanol, Emulphor EL-620, and 0.9% saline in a ratio of 1:1:18. For binding assays, THC-Certified Reference Material and CP55,940 (both in methanol) were acquired from Cayman, and [3H]SR141716A was purchased from Perkin-Elmer (Waltham, MA).
In Vivo Assays
To assess in vivo cannabimimetic activity, cumulative dose-response curves of each test compound in producing catalepsy, hypothermia, and antinociception were established, as previously described (Falenski et al., 2010). The bar test was used to assess catalepsy. In this assay, the mouse’s forepaws were placed on a metal bar 4.5 cm above the workspace, and immobility time (with the exception of movement related to respiration) was measured for a 60-second period. If the mouse removed its forepaws from the bar, they were placed back on the bar with a maximum of four occasions. In the event the mouse removed its forepaws from the bar for a fifth time, the test ended and the immobility time scored up to that time point was recorded. To assess antinociception, the distal 1 cm of the tail was immersed in a 52°C water bath, and the latency to remove the tail from the water was recorded. Rectal temperature was assessed by inserting a thermocouple probe (Physitemp Instruments, Clifton, NJ) 2 cm into the rectum. Completion of the three tests required approximately 10 minutes for a group of six mice. Thus, each group of mice was injected every 40 minutes with increasing doses of drug and tested 30 minutes after each injection. An injection volume of 10 μl/g was used for each dose, with the exception of doses in which the drug concentration exceeded solubility or suspension. In these cases, the final injection volumes were increased as follows: 16.6 μl/g for WIN55,212-2, 14.4 μl/g for CP55,940, 16.6 μl/g for JWH-073, and 13 μl/g for THC. The injection volumes to achieve the two highest doses of CP47,497 (56 mg/kg and 100 mg/kg) were 13 μl/g and 22 μl/g, respectively. A-834,735D did not require additional volume for any of the injections.
Binding Assays
Membrane Preparation.
Male and female CB1 (+/+), (+/−), and (−/−) mice were euthanized via rapid decapitation. Cerebella were collected and hemisected, and whole spinal cords were removed as previously described (Nguyen et al., 2012). Tissue was stored at −80°C until assay. At the start of the binding assay, samples were placed in 5 ml of cold membrane buffer (50 mM Tris-HCl, 3 mM MgCl2, 1 mM EGTA, pH 7.4) and homogenized. The homogenates were centrifuged at 50,000g for 10 minutes, the supernatant was removed, and the samples were resuspended in assay buffer (100 mM NaCl, 3 mM MgCl2, 0.2 mM EGTA, 50 mM Tris-HCl, pH 7.4). For agonist-stimulated [35S]GTPγS experiments, membranes (8 µg/ml of protein) were incubated with adenosine deaminase (4 mU/ml) in assay buffer at 30°C for 12 minutes before addition to assay tubes.
[3H]SR141716A Binding.
Using established methods (Selley et al., 2001), cerebellum and spinal cord samples were diluted with assay buffer to 10 μg/ml and 15 μg/ml, respectively. Membrane homogenates were then incubated with [3H]SR141716A (0.03–10 nM) in the absence and presence of a saturating concentration of unlabeled SR141716A (5 μM) to assess specific and nonspecific binding. The assay was incubated until equilibrium was attained (90 minutes) at 30°C, and then the reaction was terminated by rapid filtration under vacuum through Whatman GF/B glass fiber filters presoaked in Tris buffer containing 0.5% bovine serum albumin (BSA), followed by three washes with ice-cold Tris buffer. Bound radioactivity was measured via liquid scintillation spectrophotometry at 45% efficiency after a 9-hour delay.
Agonist-Stimulated [35S]GTPγS Binding.
Concentration-effect curves were generated by incubating varying concentrations of cannabinoid agonist with 30 μM guanosine diphosphate (GDP), 0.1 nM [35S]GTPγS, and 0.1% bovine serum albumin in duplicate and incubated at 30°C for 2 hours (Nguyen et al., 2012). Basal binding was determined in the absence of agonist, and nonspecific binding was determined in the presence of 20 μM unlabeled GTPγS. The reaction was terminated by vacuum filtration through grade GF/B glass fiber filters, followed by two washes with cold Tris buffer (50 mM, pH 7.4). Bound radioactivity was assessed via liquid scintillation spectrophotometry at 95% efficiency after overnight extraction in scintillation fluid (Research Products International, Mount Prospect, IL). For agonist-stimulated [35S]GTPγS binding, nonspecific binding was subtracted from each drug curve, and data were expressed as percent stimulation ((stimulated-basal)/basal) × 100). As a small magnitude of stimulation was detected in CB1 (−/−) tissue in certain instances (e.g., WIN55,212-2) (Breivogel et al., 2001; Monory et al., 2002), the stimulation from (−/−) tissue was subtracted from CB1 (+/+) and (+/−) curves to provide an accurate representation of CB1R-mediated agonist-stimulated binding.
Data Analyses
The dose-effect curve of each drug on each in vivo endpoint was analyzed by two-way mixed design analysis of variance with dose (within subject) and genotype (between subjects) as the two factors. The Holm-Sidak post hoc test was used to assess dose-dependent changes within genotype as well as differences in drug effect among genotypes at each dose. In addition, ED50 values and 95% confidence limits for drug effects on behavioral measures were determined via linear regression (Colquhoun, 1971), and ED50 values were considered to differ if 95% confidence limits did not overlap. The ED50 value was defined as the dose to produce immobility for 30 seconds in the bar test for catalepsy, a 4°C loss in body temperature for hypothermia, and 50% of the maximum possible effect in the tail withdrawal test for antinociception. To assess changes in pharmacologic effects produced by reducing CB1R density [i.e., in CB1 (+/+) versus CB1 (+/−) mice] dose ratios were calculated using the equation (CB1 (+/+) ED50/CB1 (+/−) ED50) for each agonist on each in vivo measure. This dose ratio served as an in vivo measure of CB1R agonist efficacy.
For [3H]SR141716A radioligand binding assays, saturation binding (Bmax) and affinity (KD) were determined by nonlinear regression saturation analysis in GraphPad Prism 6.0 and was expressed as mean values ± standard error of the mean. Maximal stimulation (Emax) and EC50 values for agonist-stimulated [35S]GTPγS binding were determined via nonlinear regression using GraphPad Prism 6.0 software. Significant differences between (+/+) and (+/−) Emax values were determined by Student’s t test for each drug. Differences in Emax across (+/+) samples in cerebellum and spinal cords were analyzed using one-way analysis of variance followed by a Tukey post hoc test.
Pearson correlations were calculated between dose ratios [CB1 (+/+) ED50 / CB1 (+/−) ED50] from in vivo (catalepsy, hypothermia, and antinociception) measures and Emax ratios (CB1 (+/+) Emax/CB1 (+/−) Emax) from agonist-stimulated [35S]GTPγS binding assays in cerebellar and spinal cord tissue.
Results
In Vivo Effects of Cannabinoids in CB1R (+/+), (+/−), and (−/−) Mice.
The cumulative dose-response evaluations of the six tested cannabinoids in CB1R transgenic mice revealed stratification for antinociceptive efficacy. A-834,735D (Fig. 1A), WIN55,212-2 (Fig. 1B), CP55,940 (Fig. 1C), JWH-073 (Fig. 1D), CP47,497 (Fig. 1E), and THC (Fig. 1F) produced significant dose-dependent antinociceptive effects in (+/+) mice, showed varying degrees of rightward shifts as well as decreased Emax values in (+/−) mice, and generally lacked activity in (−/−) mice (see Supplemental Table 1 for statistics and Supplemental Table 2 for baseline conditions). Dose ratios for each of the cannabinoids revealed dramatic decreases in potency in (+ /-) mice compared with (+/+) mice (Table 1). The maximum %MPE values for JWH-073, CP47,497, and THC were lower in (+/−) mice than in (+/+) mice. As the magnitude of effects of CP47,497 and THC did not surpass 50% and 25% MPE, respectively, neither the antinociceptive ED50 values in (+/−) mice nor the dose-ratios (CB1 (+/+)/CB1 (+/−) mice) were calculated. The diminished antinociceptive effects of all drugs in the (+/−) mice were revealed by parametric statistics (see Supplemental Table 1 and post hoc comparisons in Fig. 1). No consistent sex differences were observed for antinociception or the other measures (Supplemental Table 3). A single high dose of each drug produced antinociception from 0.5 to 6 hours in (+/+) mice, and (+/−) mice displayed decreased effects and shorter durations of action, although the actions of 100 mg/kg THC were greatly attenuated in both genotypes (Supplemental Figs. 1–6, Panel A for each figure).
A-834,735D (Fig. 2A), WIN55,212-2 (Fig. 2B), CP55,940 (Fig. 2C), JWH-073 (Fig. 2D), CP47,497 (Fig. 2E), and THC (Fig. 2F) produced hypothermic effects in the same rank order as antinociception. Each compound produced dose-dependent hypothermia in (+/+) and (+/−) mice but did not substantially affect rectal temperatures in (−/−) mice, with the exception of high doses of THC (300 and 560 mg/kg). These doses of THC significantly lowered rectal temperature in (+/−) and (−/−) mice but were substantially less in magnitude than found in (+/+) mice (see Supplemental Table 1 for statistics and Supplemental Table 2 for baseline values). Each drug was more potent in (+/+) mice than in (+/−) mice (Table 1). The decreased magnitude of hypothermia in (+/−) mice, which did not exceed the 4°C drop required for calculation, precluded the calculation of the ED50 values as well as the dose ratios (CB1 (+/+)/ CB1 (+/−) mice); however, the diminished hypothermic effects in the (+/−) mice compared with (+/+) mice were revealed by parametric statistics (see Supplemental Table 1) with appropriate post hoc analyses (Fig. 2). The time-course studies showed that each cannabinoid reduced body temperature from 0.5 to 6 hours (Supplemental Figs. 1-6, Panel B for each figure).
In contrast to antinociception and hypothermia, the catalepsy dose-response relationship of each drug was minimally affected by a 50% reduction in CB1R. Each cannabinoid yielded comparable dose-response relationships between (+/+) and (+/−) mice but was inactive in (−/−) mice for A-834,735D (Fig. 3A), WIN55,212-2 (Fig. 3B), CP55,940 (Fig. 3C), JWH-073 (Fig. 3D), CP47,497 (Fig. 3E), and THC (Fig. 3F). The statistical analyses and baseline conditions are shown in Supplemental Table 1 and 2, respectively. The time course of each drug in producing catalepsy is shown in Supplemental Figs. 1–6 (Panel C for each figure).
In the final in vivo experiment, we examined dose-response relationships of two noncannabinoid drugs, morphine and chlorpromazine, as well as repeated vehicle injections in CB1R transgenic mice. Altering CB1R number did not affect the pharmacologic effects of either drug in a systematic manner. The dose-response relationships of morphine for antinociception (Fig. 4A) and hypothermia (Fig. 4B) were not significantly affected by genotype (Supplemental Table 1). Morphine produced a significant statistical interaction for immobility in the bar test, which was due to a relatively small increase of immobility in CB1 (−/−) mice compared with the other genotypes (Fig. 4C). Chlorpromazine did not significantly stimulate antinociception in any of genotypes (Fig. 4D), and its hypothermic effects were equipotent in (+/+), (+/−), and (−/−) mice (Fig. 4E; Supplemental Table 1). The cataleptic effects of chlorpromazine differed statistically among genotypes (Fig. 4F; Supplemental Table 1), but no differences in potency were found as determined by linear regression. As previously reported (Falenski et al., 2010), repeated vehicle injections were generally without effect in each genotype; although CB1 (−/−) mice displayed increased immobility after the fifth injection (Supplemental Fig. 7).
[3H]SR141716A and Agonist-Stimulated [35S]GTPγS Binding in CB1R (+/+), (+/−), and (−/−) Mice.
[3H]SR141716A binding was used to determine the density of CB1Rs in cerebellum and spinal cord (Table 2). CB1R levels in the cerebellum were approximately 4-fold higher than levels measured in the spinal cord. Accordingly, cerebellum was considered a high CB1R expression region, and spinal cord was defined as a low CB1R expression region. Consistent with previous reports (Selley et al., 2001), the level of CB1Rs in (+/−) mice was approximately half the number of receptors in (+/+) mice for both regions.
Each cannabinoid agonist was then assessed using [35S]GTPγS binding in membrane homogenates from the cerebellum and spinal cord of CB1 (+/+), (+/−), and (−/−) mice. As previously reported (Breivogel et al., 2001; Monory et al., 2002), WIN55,212-2 significantly stimulated [35S]GTPγS binding in cerebellar homogenates from CB1 (−/−) mice (Supplemental Fig. 8, Supplemental Table 4). None of the other ligands stimulated G-protein activity in cerebellar homogenates from CB1 (−/−) mice. To eliminate non-CB1R–mediated stimulation from (+/+) and (+/−) binding dose-effect curves, the magnitude of stimulation in the CB1 (−/−) tissue was subtracted from the total stimulation in tissue from the (+/+) and (+/−) mice (Fig. 5). The Emax values of A-834,735D [T (6) = 2.54, P < 0.05], WIN55,212-2 [T (6) = 5.48, P < 0.01], CP55,940 [T (6) = 3.47, P < 0.05], JWH-073 [T (6) = 21.0, P < 0.0001], CP47,497 [T (6) = 6.84, P < 0.001], and THC [T (6) = 4.44, P < 0.01] were significantly reduced in cerebellar membranes prepared from CB1 (+/−) compared with (+/+) mice (Table 3, top). Moreover, the THC Emax value in cerebellum from (+/+) mice was significantly lower than the Emax value of each of the other ligands [F (5,18) = 23.14, P < 0.0001]. In (+/−) cerebellum, the Emax of JWH-073 differed significantly from that of A-834,735D, WIN55,212-2, and CP55,940 [F (5,18) = 15.26, P < 0.0001]. Likewise, Emax values from spinal cord membranes were significantly lower in CB1 (+/−) mice than in (+/+) mice for A-834,735D [T (6) = 5.82, P < 0.01], WIN55,212-2 [T (6) = 4.28, P < 0.01], CP55,940 [T (6) = 2.60, P < 0.05], and CP47,497 [T (6) = 5.33, P < 0.01] (Table 3, middle; Fig. 6). Although the Emax values of JWH-073 (P = 0.052) and THC (P = 0.11) in spinal cord from CB1 (+/−) and (+/+) mice did not significantly differ, THC did not stimulate [35S]GTPγS binding above basal levels in spinal cords from CB1 (+/−) mice. Statistical comparison of Emax values from (+/−) spinal cord tissue revealed significant differences between JWH-073 and both WIN55,212-2 and A-834,735D, whereas CP47,497 differed only from WIN55,212-2 [F (4, 13) = 6.00, P < 0.01]. Finally, THC [F (5, 18) = 15.26, P < 0.0001] produced significantly less stimulation than each other ligand in the spinal cords of (+/+) mice.
CB1R Agonist Efficacy: Relationship between In Vivo and In Vitro Measures.
The relationships between in vivo dose ratios [CB1 (+/−) ED50/CB1 (+/+) ED50) for each of the three dependent variables and in vitro (+/+) Emax values from [35S]GTPγS binding assays (CB1 (+/−) Emax/CB1 (+/+) Emax) in cerebellum and spinal cord membranes are depicted in (Fig. 7). CB1 (+/−) and (+/+) binding correlated between cerebellum and spinal cord. The in vivo dose ratios for antinociception (r = 0.87) and hypothermia (r = 0.94), but not catalepsy, significantly correlated with [35S]GTPγS binding in spinal cord. No significant correlation was found between any of the in vivo measures and [35S]GTPγS binding in cerebellum. The structures for all agonists used in these studies can be seen in Supplemental Fig. 9.
Discussion
Receptor theory predicts that in vivo pharmacologic efficacy differences will be most apparent for pharmacologic effects mediated by CNS regions containing low receptor reserve and least evident for effects mediated by high receptor reserve (Rang, 2006). The results presented here apply this theory to CB1R-mediated pharmacologic effects. Through the use of: 1) a cumulative dose-response procedure and 2) CB1 transgenic mice that express varying densities of CB1Rs, we established a straightforward in vivo method to discern relative differences in agonist selectivity for, and efficacy at, CB1Rs. Specifically, we tested the impact of reducing total CB1R density on in vivo potency and efficacy of THC, five SCs, and two noncannabinoids, as well as the relationship between a metric of in vivo efficacy and in vitro functional activity.
Of the three in vivo dependent measures, antinociception was the most sensitive to a 50% reduction in CB1R density. CB1 (+/−) mice showed substantial reductions in antinociceptive potency and “apparent” efficacy to each cannabinoid compared with (+/+) mice. A-834,735D, WIN55,212-2, and CP55,940 produced maximal antinociceptive effects in (+/+) and (+/−) mice, but CB1 (+/−) mice displayed 2-fold decreases in potency to A-834,735D, and WIN55,212-2 and ∼3-fold decrease in potency to CP55,940. THC was the least potent cannabinoid in producing antinociception in (+/+) mice and produced less than 25% MPE in CB1 (+/−) mice. Similarly, JWH-073 and CP47,497, within the dose ranges tested, produced significantly reduced magnitudes of maximal antinociceptive effects in (+/−) mice compared with (+/+) mice. None of the tested cannabinoids produced significant antinociceptive effects in CB1 (−/−) mice. In contrast, morphine produced full antinociceptive effects irrespective of genotype, demonstrating selectivity of this in vivo model. This pattern of findings suggests relatively low CB1R reserve for cannabinoid-induced antinociception, which is consistent with the observation that CB1R density is relatively low in CNS areas (e.g., periaqueductal gray, dorsal horn of the spinal cord) (Herkenham et al., 1990, 1991; Matsuda et al., 1990) purported to mediate antinociception. Moreover, comparison of the dose-response relationship for each cannabinoid ligand between (+/+) and (+/−) mice revealed that THC, CP47,497, and JWH-073 behaved as low-efficacy CB1R agonists, CP55,940 had moderate efficacy, and A-834,735D and WIN55,212-2 were the highest efficacy compounds. This pattern of findings is in agreement with in vitro [35S]GTPγS binding experiments.
The dose-response curve for each agonist in producing hypothermia showed a rightward shift in CB1 (+/−) mice compared with (+/+) mice. Interestingly, the Emax of THC-induced hypothermia was profoundly reduced in CB1 (+/−) mice. The observation that the magnitude of hypothermia produced by 300 and 560 mg/kg THC did not statistically differ between (+/−) and (−/−) mice suggests CB1R-independent effects. In contrast, none of the other cannabinoids, at the doses assessed, produced hypothermia in CB1 (−/−) mice. Morphine and chlorpromazine elicited dose-dependent hypothermia irrespective of genotype, supporting the utility of using CB1R transgenic mice to infer selectivity. These findings, taken together with the low level of CB1R expression in the preoptic area of the hypothalamus (Herkenham et al., 1990), a region believed to mediate cannabinoid-induced hypothermia (Rawls et al., 2002), suggest relatively low CB1R reserve for this effect.
In contrast to the antinociceptive and hypothermic measures, the dose-response relationships of the cataleptic effects of each cannabinoid tested did not statistically differ between (+/+) and (+/−) mice, suggesting a relatively high CB1R reserve in brain regions that mediate catalepsy. Chlorpromazine produced catalepsy in all three genotypes, again indicating the utility of the assay to discern CB1R selectivity. The minimal rightward shift in the dose-response relationship of cannabinoids in CB1 (+/−) mice is consistent with high levels of CB1R expression in brain areas mediating catalepsy, including the dorsal striatum (∼3 or 4 pmol/mg), cerebellum (4–6 pmol/mg), and globus pallidus (≥6 pmol/mg), which represent among the highest levels in brain (Selley et al., 2001). Work from Dhawan et al. (2006) also suggests that cannabinoid-induced catalepsy requires low CB1R occupancy. Accordingly, a 50% reduction in CB1R expression appears insufficient to decrease potency and Emax catalepsy values of low efficacy CB1R agonists. Although these findings confirm that CB1Rs mediate cannabinoid-induced catalepsy, this dependent measure does not distinguish CB1R agonist efficacy and may suggest a high receptor reserve for this effect; however, the observations that cannabinoid agonists were not more potent in producing catalepsy compared with antinociception or hypothermia in CB1 (+/+) mice are inconsistent with the predicted results based upon receptor reserve. A limitation of determining potency in the in vivo dependent measures is the placement of arbitrary maximums that may not reflect true maximum effects.
Data from agonist-stimulated [35S]GTPγS binding experiments in spinal cord generally corroborated the a priori selection of CB1R agonists, which varied from high to low efficacy (A-834,735D ≥ WIN55,212-2 > CP55,940 > JWH-073 ≥ CP47,497 > THC), when relative Emax differences of (+/+) and (+/−) mice were taken into account. Significant correlations detected between Emax ratios from [35S]GTPγS binding in spinal cord and the in vivo hypothermia and antinociception measures were consistent with hypothesis that these effects were mediated by low CB1R reserve. Conversely, the absence of a correlation between catalepsy and receptor-mediated G-protein activation is consistent with the idea of high CB1R reserve. In the present study, we selected to investigate agonist-stimulated [35S]GTPγS binding in cerebellum and spinal cord because these CNS regions are subserved respectively by high and low CB1R reserve. Thus, it will be important in future studies to examine CB1Rs in other brain regions that subserve relevant pharmacologic effects of cannabinoids in models of learning and memory, pain, reward, drug dependence, feeding, and stress responses. Future studies may focus on investigating where novel, abused SCs fall along the efficacy continuum. Although not currently available, irreversible CB1R antagonists would offer great utility to investigate the consequences of pharmacologically reducing CB1R density on relevant endpoints. This approach has been implemented successfully for the μ opioid receptor (Walker et al., 1998; Pawar et al., 2007; Madia et al., 2009).
In conclusion, the present study establishes a straightforward in vivo approach, based on pharmacologic principles, to provide valuable insight into the pharmacology of THC and emerging abused SCs. Here, we used CB1 (+/+), (+/−), and (−/−) mice to assess in vivo efficacy, potency, and selectivity of six cannabinoids and two noncannabinoids by assessing their pharmacologic effects in established assays sensitive to CB1R agonists. In particular, the high correlations between efficacy in stimulating [35S]GTPγS binding in spinal cord and producing antinociception and hypothermia in CB1 (+/−) mice suggests that these endpoints represent useful predictors of in vivo efficacy. Specifically, we found using in vivo and in vitro assays that THC, CP47,497, and JWH-073 acted as low efficacy CB1R agonists, whereas A-834,735D, WIN55,212-2, and CP55,940 behaved as high efficacy CB1R agonists. More generally, our results support the idea that low CB1R reserve mediates cannabinoid-induced antinociception and hypothermia, whereas CNS regions with high CB1R reserve may subserve cannabinoid-induced catalepsy. Thus, the use of CB1R transgenic mice to assess the antinociceptive and hypothermic effects of CB1R agonists in vivo along with agonist-stimulated [35S]GTPγS binding in CB1R mouse spinal cord tissue in vitro possesses utility in determining the efficacy of emerging abused SCs. This model may also be extended to assess cannabinergic effects of novel therapeutic CB1R agonists and to determine whether efficacy represents an important determinant of the severe health complications associated with SC abuse compared with THC/cannabis. In sum, the results of the present study suggest that low receptor reserve conditions reveal stratification of CB1R ligands by efficacy and further bridge the gap of knowledge between in vivo and in vitro pharmacologic actions of CB1R agonists.
Acknowledgments
The authors thank Brittany Mason, Mohammed Mustafa, Pamela Weller, and Jolene Windle for assistance with breeding most of the mice used for these studies and the National Institute on Drug Abuse for the contribution of drugs used for these studies.
Authorship Contributions
Participated in research design: Grim, Morales, Gonek, Wiley, Thomas, Sim-Selley, Selley, Negus, Lichtman.
Conducted experiments: Grim, Morales, Gonek, Sim-Selley.
Contributed new reagents or analytic tools: Endres, Sim-Selley, Selley.
Performed data analysis: Grim, Morales, Gonek, Selley, Negus, Lichtman.
Wrote or contributed to the writing of the manuscript: Grim, Morales, Wiley, Thomas, Sim-Selley, Selley, Negus, Lichtman.
Footnotes
- Received February 26, 2016.
- Accepted July 21, 2016.
This research was supported by the National Institutes of Health [Grants T32DA007027, R01DA032933, R01DA03672, R01DA030404, and P30DA033934].
↵This article has supplemental material available at jpet.aspetjournals.org.
Abbreviations
- A-834,735 degradant
- 3,3,4-trimethyl-1-(1-((tetrahydro-2H-pyran-4-yl)methyl)-1H-indol-3-yl)pent-4-en-1-one
- CB1R
- cannabinoid 1 receptor
- CNS
- central nervous system
- CP47,497
- rel-5-(1,1-dimethylheptyl)-2-[(1R,3S)-3-hydroxycyclohexyl]-phenol
- CP55,9440
- 5-(1,1-dimethylheptyl)-2-[(1R,2R,5R)-5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]-phenol
- EGTA
- ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetra-acetic acid
- JWH-073
- (1-butyl-1H-indol-3-yl)-1-naphthalenyl-methanone
- chlorpromazine
- SC
- synthetic cannabinoid
- THC
- Δ9-tetrahydrocannabinol
- WIN55,212-2
- (R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-napthalenylmethanone
- Copyright © 2016 by The American Society for Pharmacology and Experimental Therapeutics