Abstract
The cannabinoid-1 receptor (CB1R) has been implicated in the control of energy balance. To explore the pharmacological utility of CB1R inhibition for the treatment of obesity, we evaluated the efficacy of N-[(1S,2S)-3-(4-chlorophenyl)-2-(3-cyanophenyl)-1-methylpropyl]-2-methyl-2-{[5-(trifluoromethyl)pyridin-2-yl]oxy}propanamide (MK-0364) and determined the relationship between efficacy and brain CB1R occupancy in rodents. MK-0364 was shown to be a highly potent CB1R inverse agonist that inhibited the binding and functional activity of various agonists with a binding Ki of 0.13 nM for the human CB1R in vitro. MK-0364 dose-dependently inhibited food intake and weight gain, with an acute minimum effective dose of 1 mg/kg in diet-induced obese (DIO) rats. CB1R mechanism-based effect was demonstrated for MK-0364 by its lack of efficacy in CB1R-deficient mice. Chronic treatment of DIO rats with MK-0364 dose-dependently led to significant weight loss with a minimum effective dose of 0.3 mg/kg (p.o.), or a plasma Cmax of 87 nM. Weight loss was accompanied by the loss of fat mass. Partial occupancy (30–40%) of brain CB1R by MK-0364 was sufficient to reduce body weight. The magnitude of weight loss was correlated with brain CB1R occupancy. The partial receptor occupancy requirement for efficacy was also consistent with the reduced food intake of the heterozygous mice carrying one disrupted allele of CB1R gene compared with the wild-type mice. These studies demonstrated that MK-0364 is a highly potent and selective CB1R inverse agonist and that it is orally active in rodent models of obesity.
Cannabinoid-1 receptor (CB1R) is a G protein-coupled receptor predominantly expressed in the nervous system, and it has been identified as the brain receptor for exogenous molecules such as tetrahydrocannabinol (Berry and Mechoulam, 2002; Howlett et al., 2002; Pertwee, 2005; Thakur et al., 2005). Endogenous agonists (endocannabinoids) include anandamide and 2-arachidonyl glycerol, which have been found in the brains of vertebrates. These endocannabinoids are released from membrane precursors upon stimulation and play an important role in modulating intracellular cAMP level, ion channel activity, and synaptic transmission (Schlicker and Kathmann, 2001; Wilson and Nicoll, 2002).
In the species that have been studied thus far, CB1R is expressed in several brain regions, including hypothalamus, amygdala, hippocampus, basal ganglia, cortex, and cerebellum (Breivogel and Childers, 1998). In both amygdala and hippocampus, electrophysiological studies have revealed that CB1R agonists inhibit the release of GABA and glutamate, acting primarily at presynaptic CB1R (Schlicker and Kathmann, 2001; Wilson and Nicoll, 2002). Conversely, CB1R inverse agonists have been shown to increase synaptic transmission (Auclair et al., 2000; Wallmichrath and Szabo, 2002; Melis et al., 2004). Taken together, these data suggest that modulation of neurotransmitter release in the brain underlies the molecular and cellular basis of the in vivo pharmacological effect of CB1R ligands.
The anatomical localization of CB1R in the hypothalamus, hippocampus, amygdala, and cortex suggests that CB1R and endocannabinoids may modulate food intake and energy expenditure (Berthoud, 2002; Seeley and Woods, 2003; Schwartz and Porte, 2005; Swanson, 2005). Marijuana use is associated with increased appetite in humans (Greenberg et al., 1976). Experimental animal studies indicated that the active ingredient in marijuana, Δ9-tetrahydrocannabinol, is responsible for the appetite-enhancing effect through activation of CB1R (Trojniar and Wise, 1991). Cannabinoid agonists have also been shown to increase the reward value of food (Gardner and Vorel, 1998). Conversely, CB1R inverse agonist has been shown to inhibit the intake of food in lean rodents, diet-induced obese (DIO) rodents, marmosets, and humans (Arnone et al., 1997; Gardner and Vorel, 1998; Simiad et al., 1998; Freedland et al., 2000; Di Marzo et al., 2001; Blundell et al., 2006). In addition to the hypophagic effect, inverse agonists were shown to increase energy expenditure (Liu et al., 2005). Upon chronic administration of CB1R inverse agonists, weight loss has been observed in rodents and humans (Colombo et al., 1998; Van Gaal et al., 2005). The involvement of CB1R in energy balance is further supported by the observation that the endocannabinoid level seems to be modulated by leptin, an adipocyte-derived hormone. Acute treatment of rats with 125 μg of leptin led to a significant reduction of endocannabinoids (anandamide and 2-arachidonyl glycerol) in the hypothalamus (Di Marzo et al., 2001).
Previous studies indicated that disruption of both copies of CB1R gene led to a lean phenotype in mice and resistance to DIO (Cota et al., 2003; Ravinet Trillou et al., 2004; Osei-Hyiaman et al., 2005). However, body weight phenotype of the heterozygous mice has not been reported. In addition, it has not been reported whether complete blockade or partial inhibition of CB1R is required for chronic weight loss efficacy, whereas variable occupancy has been reported for different compounds to achieve acute food intake inhibition (Need et al., 2006). In the present study, we evaluated the in vitro and in vivo pharmacological properties of a novel compound MK-0364 (Lin et al., 2006). MK-0364 is structurally distinct from previously reported inverse agonists such as rimonabant or SLV319 (Rinaldi-Carmona et al., 1994; Lange et al., 2004). MK-0364 was shown to be a highly potent and selective CB1R inverse agonist. It inhibited food intake acutely in both mice and rats, and it caused reduction of body weight and fat mass following 14 days of once-a-day oral administration. It was further demonstrated that partial brain CB1R occupancy was sufficient for efficacy, and weight loss efficacy was correlated with brain CB1R occupancy. The partial receptor occupancy requirement for efficacy recapitulated the reduced food intake phenotype of the heterozygous mice carrying one copy of disrupted CB1R gene, providing consistent pharmacological and genetic validation for the development of MK-0364 as a new antiobesity agent.
Materials and Methods
Materials and Animals. MK-0364, AM2233, and [125I]AM2233 (specific activity 2000 Ci/mmol) were synthesized at Merck. AM251 and CP 55,940 were purchased from Tocris Cookson Inc. (Ellisville, MO). AM251 is a close analog of SR141716 (Thakur et al., 2005). Dex-fenfluramine (d-Fen) was purchased from Sigma/RBI (Natick, MA). All animal testing protocols were reviewed and approved by the Merck Research Laboratories Institutional Animals Care and Use Committee (Rahway, NJ).
In Vitro Pharmacology. The binding assay was performed by incubating various concentrations of MK-0364 with 0.5 nM [3H]CP 55,940, 1.5 μg of recombinant human CB1R-CHO membranes (or 0.1 μg of human CB2R-CHO membranes) in 50 mM Tris-HCl, pH 7.4, 5 mM MgCl2, 2.5 mM EDTA, 0.5 mg/ml fatty acid-free bovine serum albumin (BSA), 1× proteinase inhibitor mix (P8340; Sigma-Aldrich, St. Louis, MO), and 1% dimethyl sulfoxide. After 1-h incubation at 37°C, the reaction was stopped by filtration, and bound radioligand was separated from free radioligand by washing the filter plate. Total specifically bound radiolabel was approximately 10% of the total added radiolabel. Inhibitory IC50 values were calculated through nonlinear curve fitting, from which Ki values were then calculated. The CB1R density (Bmax = 5 pmol/mg based on [3H]CP 55,940 binding) in the recombinant human CB1R-CHO membranes was close to that from rat brain membranes (3–5 pmol/mg).
The intrinsic activity was measured by incubating recombinant CB1R-CHO cells with various concentrations of MK-0364 in the presence of 10 μM forskolin, 200 μM phosphodiesterase inhibitor, and 500 nM 3-isobutyl-1-methylxanthine in the assay buffer (Earle's balanced salt solution supplemented with 5 mM MgCl2, 10 mM HEPES, pH 7.3, and 1 mg/ml BSA) at room temperature for 30 min. Cells were lysed by boiling, and intracellular cAMP level was determined using the cAMP SPA kit (GE Healthcare, Little Chalfont, Buckinghamshire, UK). Because CB1R is a Gi-coupled receptor, in the absence of forskolin, cannabinoid compounds do not change intracellular cAMP level. When adenylyl cyclase is activated by forskolin, activation of Gi by CB1R in the presence of agonist (such as CP 55,940) will lead to an inhibition of the forskolin-stimulated cAMP increase, and inverse agonist will lead to a further increase of the forskolin-stimulated cAMP increase. The maximal CP 55,940-mediated inhibition of forskolin-stimulated cAMP increase is defined as 100% agonist efficacy, and the intrinsic activity of all other compounds is relative to the efficacy of CP 55,940. Negative efficacy denotes inverse agonism.
For selectivity counterscreen, 170 radioligand binding or enzymatic assays were carried at MDS Pharma Services (King of Prussia, PA) as a contract service to Merck. A summary of each assay protocol and the reference for each assay are listed in the MDS Pharma Services catalog.
Body Temperature Measurement. Male Sprague-Dawley rats (139 –181 g body weight; 6 weeks old) from Charles River Laboratories, Inc. (Wilmington, MA) were used. Rats were fed ad libitum with pelleted standard rodent chow (Harlan Teklad Diet 7012, 5% fat; 3.75 kcal/g; Harlan Teklad, Madison, WI). Animals were maintained on a 12:12-h light/dark cycle (lights off at 4:00 PM). They were randomized into groups (six rats/group) for compound and vehicle dosing. Animals were first dosed with vehicle (10% Tween 80 in water i.v.) or the CB1R inverse agonist MK-0364 (3 mg/kg i.v.) 30 min before challenge with vehicle (1% Tween 80 in saline i.p.) or the cannabinoid receptor agonist CP 55,940 (1 mg/kg i.p.). Acute measurement of body temperature was accomplished using a microprobe thermocouple thermometer (model BAT 10 thermometer and RET-2 temperature probe; Physiotemp Instruments Inc., Clifton, NJ). Measurements were performed at t =–30, 0, 60, and 75 min following CP 55,940 administration. Studies were conducted during the mid-light cycle (from 9:00 AM to 1:00 PM).
Mouse Acute Food Intake and Weight Change. For the first study involving only wild-type mice, male C57BL/6N wild-type mice were used. MK-0364 was dissolved or dispersed (with sonication) as a fine homogeneous suspension in 0.225% methylcellulose/10% Tween 80 in water for subsequent oral dosing of mice. All mice were weighed, and vehicle (0.225% methylcellulose/10% Tween 80 in water) or MK-0364 (1 or 3 mg/kg) was administered by oral gavage to male mice approximately 30 min before the onset of the dark phase of the light cycle (n = 12 per group, age 23 weeks, mean body weight 34.14 ± 0.53 g). Mice were fed ad libitum in the dark phase after dosing. A preweighed aliquot of a highly palatable medium-high fat diet (25% kcal from sucrose, 32% kcal from fat, 4.41 kcal/g; D12266B; Research Diets, New Brunswick, NJ) was provided in the food hopper of the cage 5 min before the onset of the dark phase of the light cycle and weighed 2 and 18 h after the onset of the dark phase of the light cycle. In addition, all mice were weighed 18 h after the onset of the dark phase of the light cycle. The study was of crossover design, i.e., vehicle and 1-mk/kg groups were dosed first. After a 4-day washout, the previous vehicle group was dosed with 3 mg/kg MK-0364, and the previous 1-mg/kg group was dosed with vehicle.
A second study was conducted comparing drug effect in wild-type and CB1R-deficient mice that were obtained from A. Zimmer (University of Bonn) (Zimmer et al., 1999). These mice were backcrossed onto the C57BL/6J genetic background for 10 generations by A. Zimmer before homozygous CB1R-deficient mice being rederived at Taconic Farms (Germantown, NY) onto the C57BL/6N genetic background. The resulting heterozygous mice were intracrossed to yield mice of all three possible genotypes (wild type, heterozygous, and homozygous). The resulting wild-type and knockout mice were then used to set up time-matched homozygous and wild-type intracrosses. These intracrosses were used to generate age-matched wild-type and knockout mice. Mice were maintained at all times on an irradiated regular rodent chow diet (Harlan Teklad 7012) except during the time of assay. At the time of the study, average body weight was 27 g (S.E.M. = 0.1 g) for the wild-type group and 24 g (S.E.M. = 0.4 g) for the CB1R-deficient group.
Acute Studies in DIO Rats. For acute experiments, male Sprague-Dawley DIO rats from Charles River Laboratories, Inc. were raised from 4 weeks of age on a diet moderately high in fat (32% kcal) and high in sucrose (25% kcal) (D12266B; Research Diets). Animals were used at 12 weeks of age and were maintained on a 12:12-h light dark cycle. The rats were randomized into groups (n = 6 rats/group) for compound and vehicle dosing. Rats were weighed 17 h after dosing to determine effects on overnight body weight gain. MK-0364 was administered orally to DIO rats 1 h before the start of the dark cycle (3:00 PM) at 0.3, 1, and 3 mg/kg p.o. Vehicle was 10% Tween 80 in water, and dosing volume was 2 ml/kg. Powdered food was provided in food cups that were weighed continuously at 5-min intervals over 18 h, and the data were recorded using a computerized system.
Chronic Studies in DIO Rats. For the 14-day chronic experiment, male Sprague-Dawley DIO rats were obtained as described above. Animals were used at 15 weeks of age and were maintained on a 12:12-h light/dark cycle. Rats were conditioned to dosing for 4 days before baseline measurements, using an oral gavage of 10% Tween 80 in water. Thereafter, animals were dosed daily with vehicle (10% Tween 80), MK-0364 (0.3, 1, or 3 mg/kg), or d-Fen (3 mg/kg). Compounds or vehicle was administered at 1 h before the dark cycle for 14 days. Body composition was measured by dual energy X-ray densitometry (DEXAscan; Hologic Inc., Waltham, MA) 5 days before the study and at the end of the 14-day study. Daily endpoints included body weight and food intake (change on day 14 was 18-h change, whereas changes on days 1 to 13 were 24-h changes). Terminal endpoints (18 h after the last dose) included DEXAscan analysis, white adipose tissue weights, and trough brain and plasma MK-0364 levels. The epididymal fat pad was associated with the testes along the vas deferens. The mesenteric fat pad was abdominal from the base of stomach and along the intestine to the colon.
In a parallel pharmacokinetic study, a subset of DIO rats with indwelling femoral arterial catheters was given MK-0364 at 0.3 mg/kg p.o. once daily for 14 days (n = 23 cannulated rats). To determine circulating concentrations of MK-0364 on day 1 or 7, blood was drawn before dosing and then at 15 and 30 min, and 1, 2, 4, 6, and 8 h postdosing. Some animals were euthanized at 2-h postdose on day 14 to evaluate brain exposure to MK-0364.
Ex Vivo Receptor Occupancy Measurement. Ex vivo receptor occupancy was determined by dosing MK-0364 or vehicle to rats, followed by cutting brain sections for subsequent in vitro binding of [125I]AM2233, which is a nonselective cannabinoid agonist (Gifford et al., 2002). Occupancy of CB1R in drug-treated animals can be determined by comparison with vehicle-treated animals.
In the single-dose study, nonfasted lean SD rats (males; 350–400 g) were dosed with MK-0364 (0.3, 1, 3, 10, or 30 mg/kg p.o.) or vehicle (10% Tween 80 in water). Two hours later, rats were euthanized by CO2 asphyxiation. Plasma samples were taken by cardiac puncture, and brains were quickly dissected and frozen in isopentane at –40°C and stored at –80°C until use. The degree of brain receptor occupancy was determined by quantitative receptor autoradiography in vitro. Thin coronal brain sections (20 μm) encompassing the substantia nigra were cut on a cryostat at –17°C. Sections were thaw-mounted onto microslides and dried completely at room temperature. Receptor binding was initiated by incubating the brain sections in binding buffer (50 mM Tris-HCl, pH 7.3, 2 mM MgCl2, 1 mM CaCl2, 5 mM KCl, and 1.0% BSA) containing 0.5 nM [125I]AM2233 for 30 min at room temperature. Nonspecific binding was defined by similar incubation conditions in the presence of 100 nM MK-0364. Brain sections were washed three times for 4 min in ice-cold washing buffer [50 mM Tris-HCl, pH 7.3, 2 mM MgCl2, 1 mM CaCl2, 5 mM KCl, and 0.05% Triton X-100 (v/v)], and then they were rinsed briefly in ice-cold distilled water before drying at room temperature. Brain sections were exposed to X-ray film overnight. Autoradiographic images were captured and analyzed with MCID/M2 image analysis system (Imaging Research, St. Catharines, ON, Canada). Relative optical density was determined for vehicle- or compound-treated groups (n = 3 for vehicle or each dose of compound treatment). Specific binding was obtained by subtracting the nonspecific binding (average pixel intensity of the area above substantia nigra in the presence of 100 nM MK-0364) from the corresponding total binding (average pixel intensity of the entire coronal section at the level including the substantia nigra). The maximal specific binding was defined by specific binding in the vehicle-treated rat brains. Oral dosing of MK-0364 resulted in a dose-dependent reduction of [125I]AM2233 binding to CB1R, and percentage of receptor occupancy was derived by the reciprocal of percentage of maximal binding.
To demonstrate that there was no significant dissociation of the bound MK-0364 from the tissue sections during incubation of radioligand binding, an in vitro control study was performed before the occupancy assay. Two equivalent sets (A and B) of coronal brain sections from naive rats (seven slides each) encompassing the same regions as the occupancy assay were preincubated with rising concentrations of MK-0364 (0, 1, 10, 100, 1000, 5000, and 10,000 nM) in the binding buffer (see above) for 30 min at room temperature. After a brief rinse with the buffer, both sets were subjected to [125I]AM2233 binding following the same protocol as the occupancy assay except that set A contained the radioligand plus the same rising concentrations of MK-0364 as in the preincubation, whereas set B contained radioligand only. The absence of a significant rightward shift of competition binding curve in set B with respect to set A suggested a lack of significant dissociation of prebound MK-0463 from tissues under these conditions (data not shown).
In the 14-day chronic study, male DIO rats (543-g average body weight; 16% body fat; n = 8 per group) were dosed orally once a day for 14 days with vehicle (10% Tween 80 in water), MK-0364 (0.1, 0.3, 1, or 3 mg/kg), or AM251 (3 mg/kg). Rats were dosed 1 h before the beginning of the dark cycle. Food intake and body weight were recorded daily. Fourteen days later, half of each treatment group (n = 4) was euthanized at either 2 or 24 h after the last dose. Brains were collected for receptor occupancy studies. Blood was collected for compound level determination.
Results
In Vitro Pharmacological Characterization. The apparent binding affinity of MK-0364 for the recombinant human or rat CB1R or CB2R is summarized in Table 1. It binds to human or rat CB1R with an IC50 of 0.3 and 0.4 nM, respectively, corresponding to a Ki value of 0.13 and 0.27 nM, respectively. MK-0364 binds to the human or rat CB2R with an IC50 value of 290 and 470 nM, respectively, corresponding to a Ki value of 170 and 310 nM, respectively. The selectivity ratio of CB1R over CB2R is approximately 1000-fold. As shown in Fig. 1A, MK-0364 is an inverse agonist at CB1R, because it causes a further increase of forskolin-induced cAMP increase, an opposite effect compared with the effect of an agonist such as CP 55,940. AM251 is also an inverse agonist, consistent with previous reports (Pertwee, 2005). The competitive antagonist activity of MK-0364 was demonstrated by determining the dose-response curve of the synthetic agonist CP 55,940 in the presence of various concentrations of MK-0364. MK-0364 shifted the agonist dose response to the right, and the Kb value from Schild's plot for MK-0364 against CP 55,940 is 0.5 nM (Fig. 1, B and C). MK-0364 similarly shifted the dose-response curve of methanandamide (data not shown).
Off-target activity of MK-0364 was evaluated in 170 assays of receptors, enzymes, ion channels, and transporters. Other than activity at CB1R and CB2R, none of the 170 assays revealed more than 50% inhibition by MK-0364 at concentrations less than 1 μM. Only 11 targets were inhibited with IC50 values between 1 and 10 μM (Table 2). Thus, MK-0364 showed more than 1000-fold selectivity over all tested targets.
Reversal of Agonist-Induced Hypothermia by MK-0364 in Rats. Agonist-induced hypothermia is one of the typical cannabinoid agonist effects observed in rodents. MK-0364 was evaluated in rats to determine its effect in reversing agonist-induced hypothermia. The vehicle-treated group had an average basal body temperature of 37°C. The CB1/2R receptor agonist CP 55,940 elicited a robust 4°C temperature decrease over the course of the study (60 min, –4.2°C; 75 min, –4.0°C; P < 0.0001 versus vehicle). Prior treatment with MK-0364 blocked this hypothermic response completely at both time points examined (60 min, 0.4°C, P < 0.0001 versus CP 55,940; 75 min, 0.3°C, P < 0.0001 versus CP 55,940). When monitored over the initial 30-min period before CP 55,940 challenge, MK-0364 alone did not cause a temperature change compared with vehicle (Fig. 2).
Acute Efficacy of MK-0364 on Food Intake and Body Weight Gain in Wild-Type and CB1R-Deficient Mice. MK-0364 dose-dependently inhibited 2 h and overnight food intake as well as overnight gains in body weight in C57BL/6N mice (Fig. 3). At the 1- and 3-mg/kg doses (p.o.), MK-0364 significantly inhibited 2-h food intake (36 and 69% reductions, respectively; P < 0.05 and P < 0.00001, respectively) and overnight food intake (13 and 40% reductions, respectively; P < 0.05 and P < 0.00001, respectively) as well as overnight gains in body weight (48 and 165% reductions, respectively; P < 0.01 and P < 0.00001, respectively). Animal behavior was not significantly affected by the treatment of MK-0364 based on visual inspection.
Based upon the dose-response study described above, a second study was carried out in which ad libitum-fed male wild-type mice (n = 10) were compared with male CB1R-deficient mice (n = 9) in their response to MK-0364 at 3 mg/kg. MK-0364 significantly inhibited 2-h food intake (65% reduction; P < 0.00001), overnight food intake (28% reduction; P < 0.00001), and overnight gains in body weight (73% reduction; P < 0.00005) in wild-type mice. In contrast, MK-0364 had no significant effect on any of these parameters in CB1R-deficient mice (Fig. 4).
Acute Efficacy of MK-0364 on Food Intake and Body Weight Gain in Diet Induced Obese Rats. The average cumulative overnight food intake for vehicle treated DIO rats was 18.8 g and the average body weight gain was 9.5 g (Fig. 5). The profile and extent of food intake in animals receiving MK-0364 (0.3 mg/kg) were similar to that of vehicle-treated animals. MK-0364 (1 and 3 mg/kg p.o.) decreased cumulative food intake significantly from 2 h postdosing until the end of the study. Average cumulative food intake overnight in the MK-0364 (1 mg/kg p.o.) group was 12.2 g. At this dose, food intake was reduced by 61% at 3 h (P = 0.0003 versus vehicle), 60% at 6 h (P = 0.001 versus vehicle), and 35% at 17 h (P = 0.004 versus vehicle). Average cumulative food intake overnight for rats given MK-0364 (3 mg/kg p.o.) was 9.6 g, and at this dose food intake was reduced by 64.4% at 3 h (P < 0.0001 versus vehicle), 70% at 6 h (P < 0.0001 versus vehicle), and 49% at 17 h (P = 0.0004 versus vehicle). MK-0364 administration (1 and 3 mg/kg p.o.) also led to significant decreases in body weight of 3 g (P = 0.001 versus vehicle) and 10 g (P < 0.0001 versus vehicle), respectively (Fig. 5B).
Two-Week Efficacy of MK-0364 on Food Intake and Body Weight Gain in Diet-Induced Obese Rats. Chronic weight loss efficacy of MK-0364 was evaluated in DIO rats. At all of the doses tested, MK-0364 decreased food intake, body weight gain, and adiposity (Fig. 6). Average starting body weight was 530.0 ± 6.8 g. Average weight gain of the vehicle-treated group over 13 days was 15 ± 4 g. Final body weight changes on day 13 from starting weight in the 0.3, 1, and 3 mg/kg groups were –3 ± 6, –6 ± 4, and –19 ± 6g, respectively. The relative weight loss compared with vehicle was 4, 5, and 7% for the 0.3-, 1-, 3-mg/kg groups, respectively, at the end of the study. The final weight loss in the 1 mg/kg MK-0364 group was comparable with that of the 3 mg/kg d-Fen group, whereas MK-0364 at 3 mg/kg caused a bigger weight loss than d-Fen at 3 mg/kg. Mesenteric and epididymal fat pad weights decreased significantly (P < 0.05) in a dose-dependent manner with MK-0364 treatment (Fig. 6C). There was no significant effect on lean mass (data not shown).
MK-0364 treatment resulted in a dose-related decrease in food intake that achieved statistical significance for the 0.3-, 1-, and 3-mg/kg groups (Fig. 6A). The reductions in cumulative food intake in comparison with the vehicle group were 17.5 ± 4, 20.7 ± 1.8, and 21.4 ± 2.8% for the 0.3-, 1-, and 3-mg/kg groups, respectively. Significant decreases in food intake were evident on day 1 for the 1- and 3-mg/kg MK-0364 groups, whereas the 0.3-mg/kg group showed significant decreases as early as day 2. In all three groups, significant food intake suppression was observed in at least 7 of the 14 days of the study. During the second week of MK-0364 treatment, daily food intake inhibition was maintained between 10 and 25% inhibition. Administration of d-Fen markedly inhibited food intake on day 1 of administration, and the extent of daily food intake inhibition declined on subsequent days, with a 15 to 30% inhibition relative to vehicle for most of the study duration.
At the end of the 14-day efficacy study, the plasma levels (18 h after the last dose) of MK-0364 was determined to be 3, 10, and 16 nM, respectively at 0.3, 1, and 3 mg/kg. The levels of MK-0364 in brain tissue (18 h after the last dose) were determined to be 0.068, 0.051, and 0.1 nmol/g tissue, respectively. Plasma cholesterol levels measured at the end of the study in the drug-treated groups were decreased compared with the vehicle group (82 ± 3, 77 ± 6, 69 ± 3, and 72 ± 6 mg/dl in the vehicle, 0.3-, 1-, and 3-mg/kg groups, respectively). Only the cholesterol reduction in the 1 mg/kg MK-0364 group achieved statistical significance (P < 0.05).
In a 14-day parallel PK study at 0.3 mg/kg p.o. of MK-0364 in DIO rats, plasma Cmax was 87 nM after the first dose (Tmax = 2 h), and area under the curve(0–8h) was 332 nM · h. The brain tissue level of MK-0364 at 2 h after the last dose was 0.09 nmol/g brain tissue on day 14. There was no significant difference in the plasma level of MK-0364 after one or seven doses, indicating a lack of accumulation or reduction of drug level over repeated doses. It should be pointed out that even though the total drug level in brain tissue reached 0.09 nmol/g at 0.3 mg/kg, this drug level does not necessarily represent drug available for receptor binding, and it probably represents a significant amount of drug bound to brain tissue. Hence, direct CB1R occupancy was determined using an ex vivo occupancy procedure.
CB1R Occupancy after Single Dose in Rats. CB1R occupancy by MK-0364 was determined in lean SD rats following oral administration of MK-0364 (0.3, 1, 3, 10, or 30 mg/kg p.o.). As shown in Fig. 7, there was a dose-dependent increase of CB1R occupancy as the dose was increased, and 50% receptor occupancy was reached at approximately 100 nM MK-0364 in the plasma at 2 h postdosing. More than 90% receptor occupancy was reached at a total plasma level of 1 μM.
CB1R Occupancy after 2-Week Treatment in Diet-Induced Obese Rats. To further correlate the CB1R occupancy requirement with chronic weight loss efficacy, another 14-day chronic study was performed with MK-0364 (at 0.1, 0.3, 1, or 3 mg/kg p.o.) in which both CB1R occupancy and weight loss efficacy were measured. Two hours after the last dose of the 14-day treatment period, the CB1R occupancy and plasma compound level were determined, and 50% occupancy was reached at approximately 90 nM MK-0364 in the plasma (Fig. 8). This occupancy50 value was very similar to that after single dose treatment (Fig. 7), suggesting a lack of accumulation of MK-0364 in the brain with repeated dosing. In addition, the brain CB1R occupancy by MK-0364 declined slower than its clearance from plasma, so that part of the 24 h occupancy-plasma level curve was shifted to the left compared with the 2-h occupancy-plasma level curve (Fig. 8).
After 14 days of once-a-day treatment, reduced weight gain (0.1 mg/kg MK-0364) or weight loss (0.3–3 mg/kg MK-0364 or AM251 at 3 mg/kg) were observed (Fig. 9). The dose-dependent weight loss efficacy was correlated with occupancy at 2 h for MK-0364. Weight loss efficacy of AM251 was close to the weight loss efficacy of MK-0364 at similar occupancy.
Haploinsufficiency of the Heterozygous Mice with CB1R Gene Disruption. We have evaluated the phenotype of wild-type mice, heterozygous mice with one disrupted allele of the CB1R gene, and homozygous CB1R-deficient mice. As shown in Fig. 10, four parameters were measured, including 1) cumulative weight gain from the age of 4 to 22 weeks, 2) body composition at 21 to 23 weeks of age, 3) cumulative food intake from 4 to 22 weeks of age, and 4) cumulative feed efficiency from 4 to 22 weeks of age. The heterozygous mice exhibited a trend of intermediate phenotype in all four parameters between the wild-type mice and CB1R-deficient mice, although only two of the parameters (food intake and body composition) were statistically different from either wild-type mice or CB1R-deficient mice. In situ binding with [3H]CP 55,940 confirmed the absence of radioligand binding in the homozygous mouse brain and reduced radioligand binding in the heterozygous mouse brain (data not shown). CB1R density in mouse brain membranes (Bmax) based on [3H]CP 55,940 binding also confirmed that the receptor was present at high levels in wild-type mice, could not be detected in CB1R-deficient mice, and it was significantly reduced by 40 to 50% in heterozygous mice (data not shown). The subtle haploinsufficiency phenotype suggested a gene dosage effect of CB1R gene disruption, and the genetic evidence was consistent with the partial receptor occupancy requirement for weight loss efficacy of MK-0364.
Discussion
The current studies evaluated the in vitro and in vivo pharmacological properties of MK-0364. The novel CB1R inverse agonist MK-0364 is structurally distinct from previously published CB1R ligands (Lin et al., 2006). In binding assays, MK-0364 is highly potent and selective, with a binding Ki of 0.13 nM at human CB1R and a selectivity ratio of at least 1000-fold against CB2R and 168 other targets (Tables 1 and 2). In whole-cell functional assays, MK-0364 exhibits inverse agonist efficacy analogous to SR141716A (Bouaboula et al., 1997; Pertwee, 2005). When MK-0364 and an agonist were present together, MK-0364 competitively inhibited the agonist effect. These data demonstrated that MK-0364 is an inverse agonist when agonist is not present and that it can also inhibit the action of agonists when both are present. Further evidence supporting the inverse agonist interpretation of the data in Fig. 1A is the existence of neutral antagonist O-2050 for human CB1R (Martin et al., 2002).
Consistent with the in vitro pharmacological activity, MK-0364 inhibited cannabinoid agonist-induced hypothermia in rats, demonstrating that MK-0364 is brain penetrating and inhibits one of the well known in vivo effects mediated by CB1R agonists (Compton et al., 1992). Studies using regional injection of CB1R agonist have indicated that the hypothalamus is involved in the agonist-induced hypothermia (Fitton and Pertwee, 1982). Although CB1R agonists are known to induce hypothermia, inverse agonists such as SR141716A or MK-0364 do not cause hyperthermia (Rinaldi-Carmona et al., 1994; this study). On the other hand, agonists increase food intake (Williams and Kirkham, 2002), whereas inverse agonists inhibit food intake (Arnone et al., 1997; this study). These data suggest that either the CB1R involved in the hypothermic response is not constitutively active or that there is no endogenous agonist tone. In contrast, the opposite effects of agonists and inverse agonists on food intake suggest that the CB1R involved in feeding control is either constitutively active or activated by an endogenous agonist tone.
Furthermore, MK-0364 inhibited spontaneous food intake in both mice and rats, and the anorectic effect was observed only in wild-type mice and not in CB1R-deficient mice. Thus, the in vivo efficacy of MK-0364 is highly selective and mediated entirely by CB1R, consistent with the more than 1000-fold selectivity profile of MK-0364 as determined by in vitro assays.
With chronic treatment of MK-0364 in DIO rats, significant weight loss was accompanied by loss of fat mass. Mechanistically, at least a portion of body weight loss occurs as a consequence of a decrease in food ingested. Although the maximal food intake inhibition occurred after the 1st day of treatment, lower level of food intake inhibition was sustained for the entire 2 weeks. The minimum effective dose for body weight reduction in 14 days was 0.3 mg/kg (Cmax = 87 nM). Such a high in vivo potency of MK-0364 is consistent with the high binding affinity in vitro.
To further establish the relationship between efficacy and brain CB1R occupancy, we have developed an ex vivo receptor occupancy procedure to determine what level of CB1R occupancy is needed for efficacy. As shown in Fig. 7, 50% occupancy can be achieved at 2 h after a single dose at the plasma level of 100 nM in lean rats, and more than 90% occupancy can be achieved at the plasma level of 1000 nM. Because the minimum effective dose for chronic weight inhibition was 0.3 mg/kg with a Cmax of 87 nM in DIO rats, the occupancy data suggested that approximately 30 to 40% occupancy is achieved at the minimum effect dose. As shown in Fig. 8, 50% occupancy can be achieved at 2 h after the last dose at the plasma level of approximately 90 nM in DIO rats. The data in Figs. 7 and 8 thus indicated that the 2-h occupancy versus plasma level relationship in lean rats after single dose is very similar to that in DIO rats after 14 repeated doses. The data in Fig. 8 further suggested that the brain elimination of MK-0364 is slower than that of plasma elimination in rats, because the 24-h occupancy versus plasma level curve is shifted to the left compared with the 2-h occupancy versus plasma level curve. It should be pointed out that future in vivo occupancy measurement or an in vivo positron emission tomography occupancy study will be desirable to confirm the current conclusion based on ex vivo occupancy data.
The aforementioned conclusion that partial occupancy is sufficient for efficacy was further demonstrated in the DIO rat weight loss study where both occupancy and efficacy were measured. As shown in Fig. 9, chronic weight loss efficacy can be achieved with at least 30% CB1R occupancy by MK-0364 at 2 h postdose. Higher occupancy is correlated with higher efficacy. These data demonstrate that complete blockade of CB1R is not required for efficacy in rodents. The requirement of partial occupancy for efficacy is also consistent with the genetic evidence demonstrating a subtle haploinsufficiency phenotype of the heterozygous CB1R mice, especially food intake and body composition phenotypes (Fig. 10). This would contrast the efficacy of NPY Y5 receptor antagonist where more than 90% receptor occupancy is required for efficacy (Ishihara et al., 2006).
In summary, the novel CB1R inverse agonist MK-0364 was shown to be highly potent and selective both in vitro and in vivo. Chronic weight loss was accompanied by the loss of fat mass. Significant weight loss was observed at partial receptor occupancy, and weight loss efficacy was proportional to brain CB1R occupancy. These rodent data suggest that MK-0364 has the potential to be a new agent for the treatment of obesity.
Acknowledgments
We thank Dr. Andreas Zimmer for allowing us to use the CB1R-deficient mice in this study.
Footnotes
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.106.118737.
-
ABBREVIATIONS: CB1R, cannabinoid-1 receptor; CB2R, cannabinoid-2 receptor; DIO, diet-induced obesity; MK-0364, N-[(1S,2S)-3-(4-chlorophenyl)-2-(3-cyanophenyl)-1-methylpropyl]-2-methyl-2-{[5-(trifluoromethyl)pyridin-2-yl]oxy}propanamide; d-Fen, dexfenfluramine; CHO, Chinese hamster ovary; BSA, bovine serum albumin; CP 55,940, (1R,3R,4R)-3-[2-hydroxy-4-(1,1-dimethylheptyl)phenyl]-4-(3-hydroxypropyl)cyclohexan-1-ol; SR 141716A, N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboximide hydrochloride; SLV319, (4S)-(–)-3-(4-chlorophenyl)-N-methyl-N′-[(4-chlorophenyl)sulfonyl]-4-phenyl-4,5-dihydro-1H-pyrazole-1-carboxamidine; AM-2233, (2-iodophenyl)(1-{[(2R)-1-methylpiperidin-2-yl]methyl}-1H-indol-3-yl)methane; AM-251, 1-(2,4-dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-(1-piperidinyl)-1H-pyrazole-3-carboxamide; SR141716, 5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-N-1-piperidinyl-1H-pyrazole-3-carboxamide; O-2050, N-{6-[(6aR)-1-hydroxy-6,6,9-trimethyl-6a,7,10,10a-tetrahydro-6H-benzo[c]chromen-3-yl]hex-4-yn-1-yl}methansulfonamide.
-
↵ The online version of this article (available at http://jpet.aspetjournals.org) contains supplemental material.
-
↵1 Current affiliation: Merck Research Laboratories, Boston, Massachusetts.
- Received December 18, 2006.
- Accepted February 26, 2007.
- The American Society for Pharmacology and Experimental Therapeutics