The dietary polyphenols trans-resveratrol [5-[(1E)-2-(4-hydroxyphenyl)ethenyl]-1,3-benzenediol; found in red wine] and curcumin [1,7-bis(4-hydroxy-3-methoxyphenyl)-1E,6E-heptadiene-3,5-dione] (found in curry powders) exert anti-inflammatory and antioxidant effects via poorly defined mechanisms. It is interesting that cannabinoids, derived from the marijuana plant (Cannabis sativa), produce similar protective effects via CB1 and CB2 receptors. We examined whether trans-resveratrol, curcumin, and ASC-J9 [1,7-bis(3,4-dimethoxyphenyl)-5-hydroxy-1E,4E,6E-heptatriene-3-one] (a curcumin analog) act as ligands at cannabinoid receptors. All three bind to human (h) CB1 and mouse CB1 receptors with nanomolar affinities, displaying only micromolar affinities for hCB2 receptors. Characteristic of inverse agonists, the polyphenols inhibit basal G-protein activity in membranes prepared from Chinese hamster ovary (CHO)-hCB1 cells or mouse brain that is reversed by a neutral CB1 antagonist. Furthermore, they competitively antagonize G-protein activation produced by a CB1 agonist. In intact CHO-hCB1 cells, the polyphenols act as neutral antagonists, producing no effect when tested alone, whereas competitively antagonizing CB1 agonist mediated inhibition of adenylyl cyclase activity. Confirming their neutral antagonist profile in cells, the polyphenols similarly attenuate stimulation of adenylyl cyclase activity produced by a CB1 inverse agonist. In mice, the polyphenols dose-dependently reverse acute hypothermia produced by a CB1 agonist. Upon repeated administration, the polyphenols also reduce body weight in mice similar to that produced by a CB1 antagonist/inverse agonist. Finally, trans-resveratrol and curcumin share common structural motifs with other known cannabinoid receptor ligands. Collectively, we suggest that trans-resveratrol and curcumin act as antagonists/inverse agonists at CB1 receptors at dietary relevant concentrations. Therefore, these polyphenols and their derivatives might be developed as novel, nontoxic CB1 therapeutics for obesity and/or drug dependence.
Dietary polyphenols, such as resveratrol [5-[(1E)-2-(4-hydroxyphenyl)ethenyl]-1,3-benzenediol] (found in red wine) and curcumin [1,7-bis(4-hydroxy-3-methoxyphenyl)-1E,6E-heptadiene-3,5-dione] (found in curry powders), have been used safely for centuries as traditional medicines. As a consequence, increasing scientific investigation suggests that they may prove useful as therapeutics for a broad range of conditions (Scalbert et al., 2005), from inflammatory diseases (Rahman et al., 2006) to cancer (Hadi et al., 2007). The protective effects of resveratrol and curcumin seem to be related to their antioxidant (Fraga, 2007) and anti-inflammatory (Surh et al., 2005) properties. Although the specific mechanisms responsible for these beneficial effects remain unclear, the beneficial effects in vitro generally require relatively high concentrations (>1 μM) and are thought to involve multiple receptor- and nonreceptor-mediated processes (Stevenson and Hurst, 2007).
Recently, it has been reported that resveratrol and other polyphenols bind with high affinity to a distinct, yet unidentified, plasma membrane bound receptor that occurs in high density throughout the brain (Han et al., 2006). Cannabinoid receptors seem to share many characteristics with this newly discovered, uncharacterized resveratrol receptor. Originally isolated from the marijuana plant (Cannabis sativa), both synthetic and naturally occurring cannabinoids, such as Δ9-tetrahydrocannabinol, produce their effects by acting at two G-protein-coupled receptors, CB1 (Matsuda et al., 1990) and CB2 (Munro et al., 1993). CB1 receptors are expressed in high abundance throughout the central nervous system, whereas CB2 receptors are expressed predominantly in immune cells and non-neuronal tissues. Cannabinoids acting at both receptors produce antioxidant (Hampson et al., 1998) and anti-inflammatory (Klein, 2005) effects, similar to that reported for resveratrol and curcumin. Therefore, the current studies were conducted to determine whether two important dietary polyphenols, resveratrol and curcumin, and an analog of curcumin (ASC-J9) act as ligands at cannabinoid receptors. It is important that our study identifies the human CB1 cannabinoid receptor as a high-affinity target for all three polyphenols: resveratrol (Ki = 45 nM), curcumin (Ki = 6 nM), and ASC-J9 (Ki = 64 nM, an analog of curcumin). Furthermore, all polyphenols examined seem to act as CB1 antagonists/inverse agonists and share common structural motifs with other known cannabinoid receptor ligands. It is important that these results indicate that CB1 receptors are one of the highest affinity targets identified to date for resveratrol and curcumin and may have significant implications for future development of novel, nontoxic CB1 ligands.
Materials and Methods
All drugs used in this study were obtained from Tocris Bioscience (Ellisville, MO). [3H]CP-55,950 (168 Ci/mmol) and [35S]GTPγS (1250 Ci/mmol) were purchased from PerkinElmer Life and Analytical Sciences (Waltham, MA). [3H]Adenine (26 Ci/mmol) was obtained from Vitrax (Placenia, CA). All other reagents were purchased from Thermo Fisher Scientific (Waltham, MA).
CHO-K1 cells stably expressing hCB1 receptors (CHO-hCB1) were a generous gift from Dr. Debra A. Kendall (University of Connecticut, Storrs, CT). Stably transfected CHO-hCB2 cells were generated in our laboratory (Shoemaker et al., 2005). Cells were incubated in a humidified atmosphere of 5% CO2 and 95% air at 37°C in Dulbecco's modified Eagle's medium with 10% fetal calf serum, 100 units/ml penicillin, 100 mg/ml streptomycin, and 2.5 mg/ml Geneticin (G418).
Brain tissue was collected from decapitated male and female B6SJL mice obtained from an in-house breeding colony. Whole brains were pooled before beginning homogenization. Pellets of frozen/thawed cells or freshly harvested brain tissue were resuspended in a homogenization buffer containing 50 mM HEPES, pH 7.4, 3 mM MgCl2, and 1 mM EGTA. Using a 40-ml Dounce glass homogenizer (Wheaton, Philadelphia PA), samples were subjected to 10 complete strokes and centrifuged at 18,000 rpm for 10 min at 4°C. After repeating the homogenization procedure twice more, the samples were resuspended in HEPES buffer (50 mM, pH 7.4) and subjected to 10 strokes utilizing a 7-ml glass homogenizer. Membranes were stored in aliquots of approximately 1 mg/ml at -80°C.
Competition Receptor Binding
Increasing concentrations of WIN-55,212-2 or different polyphenols were incubated with 0.1 nM (mouse brain or CHO-hCB2) or 0.5 nM (CHO-hCB1) [3H]CP-55,940 in a final volume of 1 ml of binding buffer as described previously (Shoemaker et al., 2005). Each binding assay contained 100 (mouse brain or CHO-hCB2) or 150 (CHO-hCB1) μg of membrane protein, and reactions were incubated for 90 min at room temperature with mild agitation. Nonspecific binding was defined as binding observed in the presence of 1 μM nonradioactive CP-55,940. Reactions were terminated by rapid vacuum filtration through Whatman GF/B glass fiber filters (Whatman, Clifton, NJ) followed by two washes with ice-cold binding buffer. Analysis of the binding data were performed using the nonlinear regression (Curve Fit) function of GraphPad Prism version 4.0b (GraphPad Software Inc., San Diego, CA) to determine the concentration of the drug that displaced 50% of [3H]CP-55,940 (IC50). A measure of affinity (Ki) was derived from the IC50 values utilizing the Cheng-Prusoff equation (Cheng and Prusoff, 1973).
[35S]GTPγS binding assays were performed with minor modifications as described previously (Shoemaker et al., 2005) in a buffer containing 20 mM HEPES, pH 7.4, 100 mM NaCl, 10 mM MgCl2, and 0.1% bovine serum albumin. Each binding reaction contained 100 (mouse brain or CHO-hCB2) or 150 μg (CHO-hCB1) of membrane protein, cannabinoid ligands, 0.1 nM [35S]GTPγS, and 10 μM GDP. Nonspecific binding was defined by 10 μM nonradioactive GTPγS. After incubation at 30°C for 2 h, the reaction was terminated by filtration, and bound radioactivity was determined by liquid scintillation counting.
Measurement of cAMP Levels in Intact Cells
The conversion of [3H]adenine-labeled ATP pools to cAMP was used as a functional measure of cannabinoid activity (Shoemaker et al., 2005). CHO-hCB1 cells were seeded into 24-well plates and cultured to confluence. Dulbecco's modified Eagle's medium containing 0.9% NaCl, 500 μM 3-isobutyl-1-methylxanthine, and 2 μCi/well [3H]adenine was added to the cells for 2 h at 37°C. The [3H]adenine mixture was removed, and the cannabinoids were added for 15 min in a Krebs-Ringer-HEPES buffer containing 500 μM 3-isobutyl-1-methlyxanthine and 10 μM forskolin. The reaction was terminated with 50 μl of 2.2 N HCl and [3H]cAMP separated by alumina column chromatography.
Mice. Animal use protocols employed in this study were approved by the University of Arkansas for Medical Sciences Institutional Animal Care and Use Committee and conducted in accordance with the United States Public Health Service policy on humane care and use of laboratory animals. Male and female B6SJL mice were obtained from an in house breeding colony.
Hypothermia Experiments. Body temperature of age- and weight-matched mice was measured by a digital thermometer (model 17025; Thermo Fisher Scientific) inserted ∼1 cm into the rectum. Body temperature was measured 1 h after a subcutaneous injection of CP-55,940, a time interval resulting in maximal hypothermia (data not shown). When testing CB1 antagonism, drugs were given 30 min before CP-55,940 injections by the intraperitoneal route. For all experiments, body temperature was measured before any injection, 30 min after antagonist or vehicle injection and 1 h after injection of CP-55,940. The injection vehicle used for these experiments contained 50% polyethylene glycol and 50% saline.
Body Weight Reduction Experiments. Age- and weight-matched mice were injected intraperitoneally with the indicated doses of test drugs twice daily for 3 days. Body weight (in grams) was recorded each morning before drug injection and finally at 9:00 AM on day 4 of the study, 12 h after the last drug dose. Animals were fed ad libitum during the 3-day experiment. The injection vehicle for these experiments contained 50% polyethylene glycol and 50% saline.
Curve-fitting and statistical analyses were conducted utilizing GraphPad Prism version 4.0b. Data obtained from three or more experimental groups were analyzed by a one-way ANOVA, followed by a Dunnett's post hoc comparison of individual groups. A nonpaired Student's t test was employed to statistically compare data obtained from two experimental groups.
trans-Resveratrol, Curcumin, and the Curcumin Analog ASC-J9 Selectively Bind with Nanomolar Affinities to Human CB1 Receptors Stably Expressed in CHO Cells. Homologous competition receptor binding with the CB1/CB2 agonist [3H]CP-55,940 showed that stably transfected CHO-hCB1 cells express a density of CB1 cannabinoid receptors of 0.26 ± 0.14 pmol/mg protein (n = 3; data not shown). Saturation binding studies with [3H]CP-55,940 demonstrated that CHO-hCB2 cells express a density of hCB2 receptors of 1.4 ± 0.24 pmol/mg protein (Shoemaker et al., 2005). [3H]CP-55,940 binds nonselectively to hCB1 and hCB2 receptors expressed in CHO cells, with a Kd of 1.0 ± 0.3 or 0.38 ± 0.06 nM for each, respectively. Competition binding (Fig. 1A) demonstrates that the nonselective cannabinoid agonist WIN-55,212-2 also has relatively equivalent nanomolar affinity (Ki) for hCB1 (7.7 ± 1.3 nM; n = 5) and hCB2 (5.8 ± 1.2 nM; n = 4) receptors (Fig. 1A; Table 1). In contrast, all three polyphenols examined selectively bind to hCB1, relative to hCB2 receptors (Fig. 1, B–D). Curcumin is 446-fold selective, binding to hCB1 with an affinity of 5.9 ± 2.1 nM (n = 6) and having a Ki for hCB2 of over 2 μM (Fig. 1B). ASC-J9, an analog of curcumin, although demonstrating a slightly lower affinity (64 ± 17 nM; n = 3), also binds to hCB1 with a 201-fold selectively over hCB2 (13 ± 1.3 μM; n = 4) (Fig. 1C). trans-Resveratrol is highly selective, binding to hCB1 with a Ki of 45 ± 17 nM (n = 3), while failing to significantly displace [3H]CP-55,940 from hCB2 at concentrations up to 100 μM (Fig. 1D). It is important that cis-resveratrol failed to displace [3H]CP-55,940 from hCB1 at concentrations up to 100 μM (data not shown). All polyphenols (100 μM) fail to reduce [3H]CP-55,940 binding in wild-type CHO cells (data not shown).
It is curious that approximately 10 to 15% of residual [3H]CP-55,940 binding was observed in both CHO-hCB1 and CHO-hCB2 homogenates for all ligands examined (including WIN-55,212-2), even when high concentrations of the nonradioactive drugs were employed for competition. It is possible that the residual biding was due, in part, to the use of nonradioactive CP-55,940 to define nonspecific binding. Under certain conditions, employing the same nonradioactive compound to define nonspecific binding of the radioactive compound may identify binding erroneously as specific, when in reality it is nonspecific but inhibitable. Therefore, experiments were performed to compare the maximal displacement of [3H]CP-55,940 produced by CP-55,940 (1 μM) and a second high-affinity nonselective cannabinoid agonist HU-210 (1 μM) (data not shown). Results from these experiments revealed that nonradioactive CP-55,940 and HU-210 produce near-identical maximal displacement of [3H]CP-55,940 in membrane homogenates prepared from mouse brain, CHO-hCB1, and CHO-hCB2 cells. This suggests that the residual [3H]CP-55,940 binding observed for all cannabinoid ligands tested was not due to the use of nonradioactive CP-55,940 to define nonspecific binding. Although the exact reason for the observed residual binding is unknown, it is possible that the highly hydrophobic properties of the ligands tested, relative to CP-55,940, might contribute these results.
trans-Resveratrol, Curcumin, and ASC-J9 Act as Antagonists/Inverse Agonists at Human CB1 Receptors in Membrane Preparations of CHO-hCB1 Cells. To determine the intrinsic activity concerning G-protein function, the ability of the three polyphenols to modulate [35S]GTPγS binding in CHO-hCB1 membranes was examined (Fig. 2). Characteristic of agonists, the nonselective full CB1/CB2 agonist WIN-55,212-2 produces a concentration-dependent increase of approximately 90% in the binding of [35S]GTPγSto CHO-hCB1 membranes, with an ED50 of 31 ± 6.4 nM (Fig. 2A; n = 3). In marked contrast, when tested alone, all polyphenols produce a concentration-dependent decrease of [35S]GTPγS binding to CHO-hCB1 membranes. All polyphenols (100 μM) fail to produce any change in [35S]GTPγS binding to membranes prepared from wild-type CHO cells (data not shown). This suggests that the polyphenols act as inverse agonists, suppressing G-protein activation produced by constitutively active hCB1 receptors. However, the potency (e.g., IC50) of the polyphenols required to observe inverse agonism is relatively low (curcumin, 1.3 ± 0.3 μM, n = 3; ASC-J9, 56 ± 22 μM, n = 3; trans-resveratrol, 47 ± 17 μM, n = 3) compared with their high nanomolar affinity for hCB1 receptors (Fig. 1). Consistent with an antagonist/inverse agonist profile, coincubation with a fixed concentration of each polyphenol that produced minimal reduction of [35S]GTPγS binding alone resulted in a significant reduction in the potency of the agonist WIN-55,212-2 to activate G-proteins (Fig. 2B). Both curcumin and trans-resveratrol produced a significant (p < 0.05), 3-fold shift to the right in the concentration-effect curve of WIN-55,212-2 (+curcumin, 120 ± 3.5 nM, n = 3; +trans-resveratrol, 130 ± 25 nM, n = 3) (Fig. 2B). It is interesting that coincubation with ASC-J9 resulted in a much greater, 63-fold reduction in the potency of WIN-55,212-2 to activate G-proteins in CHO-hCB1 membranes (2500 ± 640 nM, n = 3).
trans-Resveratrol, Curcumin, and ASC-J9 Act as Neutral Antagonists at Human CB1 Receptors in Intact CHO-hCB1 Cells. Cannabinoid receptors activate the Gi/Go class of G-proteins, modulating the activity of the effector adenylyl cyclase. As a consequence, cannabinoid agonists reduce, neutral antagonists do not alter, and inverse agonists increase the levels of intracellular cAMP (Fig. 3). WIN-55,212-2 produced a concentration-dependent reduction of over 60% in intracellular cAMP levels in intact CHO-hCB1 cells, with an ED50 of 12.0 ± 4.6 nM (n = 8; Fig. 3, A and B). In marked contrast, exposure of CHO-CB1 cells to all three polyphenols with concentrations as high as 10 μM did not alter intracellular levels of cAMP (Fig. 3A). Therefore, all polyphenols tested act as neutral antagonists, rather than inverse agonists, in intact CHO-hCB1 cells. Indicative of competitive antagonism, coincubation with a fixed concentration of each of the polyphenols with the agonist WIN-55,212-2 resulted in a significant (p < 0.05) 7- to 10-fold parallel shift to the right in the concentration-effect curve (WIN + curcumin, 130 ± 65 nM, n = 5; WIN + ASC-J9, 90 ± 25 nM, n = 4; WIN + trans-resveratrol, 100 ± 25 nM, n = 4) (Fig. 3B). Characteristic of neutral antagonists, all polyphenols examined attenuated not only the inhibitory effects of the agonist WIN-55,212-2 (Fig. 3C) but also the stimulatory action of the inverse agonist AM-251 (Fig. 3D). Lastly, neither WIN-55,212-2 nor any of the polyphenols tested altered intracellular cAMP levels in wild-type CHO cells not transfected with hCB1 (data shown). It is interesting that the inability of trans-resveratrol to alter cAMP levels in CHO cells suggests that these cells respond differently than MCF-7 breast cancer cells, in which resveratrol has been shown to directly stimulate adenylyl cyclase activity (El-Mowafy and Alkhalaf, 2003).
Similar to Human CB1 Receptors, trans-Resveratrol, Curcumin, and ASC-J9 Bind with Nanomolar Affinity to and Act as Antagonists/Inverse Agonists at Mouse CB1 Receptors in Membrane Preparations of Whole-Brain Tissue. Before conducting in vivo studies in mice, in vitro studies were conducted to determine the affinity and intrinsic activity of the polyphenols at mouse CB1 receptors (Fig. 4). Homologous competition receptor binding with [3H]CP-55,940 showed that mouse brain membranes contain a density of mCB1 cannabinoid receptors of 0.59 ± 0.14 pmol/mg protein, to which CP-55,940 binds with an affinity (Kd) of 2.6 ± 0.55 nM (n = 3, data not shown). The affinity of WIN-55,212-2 for brain mCB1 receptors (3.4 ± 1.5 nM, n = 5; Fig. 4A; Table 1) is similar to that observed in CHO-hCB1 membranes (Fig. 1). Furthermore, all three polyphenols bind to mCB1 receptors with high nanomolar affinity and with the same rank order of potency as observed for hCB1 (Table 1). Curcumin binds mCB1 with the highest affinity (73 ± 24 nM, n = 7), followed by similar, but lower, Ki values for trans-resveratrol (270 ± 160 nM, n = 6) and ASC-J9 (190 ± 110 nM, n = 8). WIN-55,212-2 activates G-proteins in mouse brain membranes with an ED50 of 76 ± 38 nM (n = 3; Fig. 4B). Similar to that observed for hCB1, all polyphenols produce concentration-dependent inhibition of [35S]GTPγS binding to mouse brain membranes (Fig. 4B). However, curcumin and ASC-J9 act as inverse agonists at lower concentrations (IC50 = 660 ± 370 nM, n = 3; 360 ± 86 nM, n = 3, respectively) than trans-resveratrol (IC50 = 6.7 ± 2.5 μM, n = 4). This suggests that at mCB1 receptors in the brain, trans-resveratrol acts a pure neutral antagonist, whereas curcumin and ASC-J9 are inverse agonists. Consistent with these predictions, the neutral antagonist trans-resveratrol significantly attenuated G-protein activation by WIN-55,212-2 (Fig. 4C, left), whereas the decrease in [35S]GTPγS binding produced by the putative inverse agonists curcumin and ASC-J9 was reversed by coincubation with the neutral CB1 antagonist O-2050 (Fig. 4C, center and right).
In Mice, Acute Administration of trans-Resveratrol, Curcumin, and ASC-J9 Antagonizes Hypothermia Produced by a CB1 Agonist. Cannabinoid agonists produce a classic tetrad of effects in mice (hypothermia, analgesia, catalepsy, and reduced locomotor activity), mediated by activation of CB1 receptors (Smith et al., 1994). To determine whether the polyphenols act as antagonists/inverse agonists at mCB1 receptors in vivo (as predicted by in vitro assays), the ability of each compound to antagonize hypothermia produced by the cannabinoid agonist CP-55,940 was examined (Fig. 5, A and B). One hour after subcutaneous injections, CP-55,940 produces a dose-related decrease in body temperature in mice of over 6°C with an ED50 of 0.25 mg/kg (0.23–0.28, 95% confidence interval; n = 5–22 mice per dose) (Fig. 5B). When coadministered with a 0.2 mg/kg dose of CP-55,940, employed to produce approximately a half-maximal effect (-3.1 ± 0.82°C, n = 22), all three polyphenols tested produce a significant dose-dependent reversal of CP-55,940-induced hypothermia with a rank order of potency (IC50) for reversal of trans-resveratrol (14 mg/kg) > curcumin (52 mg/kg) > ASC-J9 (88 mg/kg) (Fig. 5A). Curcumin is most efficacious, resulting in a complete antagonism of hypothermia. Indicative of competitive antagonism, coadministration with a fixed dose of trans-resveratrol (5 mg/kg) results in a significant (p < 0.05), parallel rightward shift in the hypothermia dose-effect curve for CP-55,940 to 0.51 mg/kg (0.48–0.55, 95% confidence interval; n = 5 mice per dose). CP-55,940-induced hypothermia is completely blocked by the CB1 antagonist/inverse agonist AM-251 (10 mg/kg, data not shown). In addition, when administered alone, polyphenol concentrations that produce maximal antagonism of hypothermia induced by CP-55,940 (trans-resveratrol, 50 mg/kg; curcumin, 500 mg/kg; ASC-J9, 200 mg/kg) produce no hypothermia (data not shown).
In Mice, Repeated Administration of trans-Resveratrol and Curcumin Produces Dose-Dependent Reduction in Body Weight, Similar to That Produced by the CB1 Antagonist/Inverse Agonist AM-251. CB1 antagonists/inverse agonists produce reductions in food intake and body weight in mice (Pavon et al., 2008). Because in vitro assays suggest that all polyphenols tested act as antagonists/inverse agonists at mCB1, the ability of trans-resveratrol and curcumin to reduce body weight in mice was examined (Fig. 5C). As anticipated, the CB1 antagonist/inverse agonist AM-251 (10 mg/kg) administered twice daily for 3 days results in a significant (p < 0.01) weight loss of 2.8 ± 0.47 g (Fig. 5C, left; n = 6). Likewise, repeated administration of curcumin produces a dose-related weight loss, equivalent to that produced by AM-251 (Fig. 5C, center; n = 5). Although slightly higher doses are required, trans-resveratrol also results in significant (p < 0.05), dose-dependent weight loss (Fig. 5C, right, n = 5).
In Silico Comparison of the Structures of trans-Resveratrol and Curcumin with Known Cannabinoids Reveals Common Structural Motifs. Molecular modeling studies employing CAChe molecular modeling software (Fujitsu America, Inc., Sunnyvale, CA) with structure minimizations performed with a PM5 wave function in water reveals that the favored conformation of trans-resveratrol (Fig. 6A, in red) is similar to that of a series of novel synthetic resorcinol-derived cannabinoids (Wiley et al., 2002), as graphically illustrated by comparison with the resorcinol O-1422 (Fig. 6A, in green). When the resorcinol rings of both molecules are overlaid, the similarities are striking. Although the cyclohexyl group of O-1422 is not present in trans-resveratrol, the dimethylheptyl side chain (also present in many other cannabinoids) of O-1422 is similar in length to the trans-double bond and phenol ring of resveratrol.
In addition, a subsequent overlay of the CB1-selective ligand rimonabant (Fig. 6B, in blue), trans-resveratrol (Fig. 6B, in red), and curcumin (Fig. 6B, in purple) reveals several areas of similarity that closely match a three-dimensional pharmacophore model of CB1-selective ligands recently proposed by Wang et al. (2008). For example, an aromatic region (A) and a hydrophobic region (B), which are located in aromatic rings containing electron-withdrawing groups, are present in all three molecules. Furthermore, the amide carbonyl (of rimonabant), the carbonyl (of curcumin), and the phenol (of trans-resveratrol) all contain electron-donating oxygens (hydrogen bond acceptors) and are all located in the middle region; hence, they are designated as electron-donating region C.
Based on inferences drawn from a model proposed by Song et al. (1999), it might be predicted that aromatic rings contained in region A of these ligands probably interact with some combination of CB1 receptor residues F3.25(189), W5.43(279), F5.42(278), and Y5.39(275). Moreover, it is also probable that the hydrophobic region B of these compounds might interact with CB1 receptor residue F3.36(200). In any case, it is clear that trans-resveratrol and curcumin share several common structural motifs with known cannabinoid ligands, and these motifs probably contribute to their ability to bind with high affinity to CB1 receptors.
The most significant finding of this study is the identification of human CB1 cannabinoid receptors as a high-affinity target for three distinct polyphenols; trans-resveratrol (Ki = 45 nM), curcumin (Ki = 6 nM), and ASC-J9 (Ki = 64 nM, an analog of curcumin). All polyphenols examined seem to act as CB1 antagonists/inverse agonists, at dietary-relevant concentrations, in both in vitro and in vivo assays. Furthermore, in silico comparison of the structures of trans-resveratrol and curcumin with known cannabinoids reveals common structural motifs. Coupled with their proven safety, these studies indicate that trans-resveratrol, curcumin, and/or their derivatives might be developed as novel, nontoxic CB1 therapeutics for use in obesity, diabetes, drug dependence, and additional disease states in which CB1 antagonists have shown efficacy.
Polyphenols, including trans-resveratrol and curcumin, are known to produce many biological effects by acting on multiple targets (Stevenson and Hurst, 2007). trans-Resveratrol and curcumin are very efficacious antioxidant (Fraga, 2007) and anti-inflammatory (Surh et al., 2005) agents; however, their in vitro effects require relatively high concentrations (>1 μM) and are thought to involve multiple receptor- and nonreceptor-mediated processes. Therefore, the specific molecular mechanisms responsible for these effects remain unclear. This study identifies CB1 receptors as one of the highest affinity targets for trans-resveratrol and curcumin reported to date. For example, although trans-resveratrol inhibits the activity of quinone reductase 2, with a dissociation constant of 35 to 50 nM (Buryanovskyy et al., 2004), much higher concentrations are required to stimulate adenylyl cyclase (800 nM) (El-Mowafy and Alkhalaf, 2003) or inhibit the activity of Iκβ kinase (1 μM) (Kundu et al., 2006) and lipoxygenase (3.7 μM) (Jang et al., 1997). Likewise, curcumin inhibits the activity of glycogen synthase kinase-2β with an IC50 of 63 nM (Bustanji et al., 2008); however, significantly greater concentrations are required to reduce the aggregation of β-amyloid (800 nM) (Yang et al., 2005) or inhibit glutathione transferases (0.04–5 μM) (Hayeshi et al., 2007). Therefore, compared with the affinity for most other identified targets, it is likely that CB1 receptors clearly play an important role in the molecular mechanism of action for trans-resveratrol and curcumin, requiring relatively low, physiologically attainable concentrations to produce near-full CB1 receptor occupancy.
The present findings are additionally important because they identify a specific, high-affinity, receptor-mediated mechanism that probably contributes to many of the reported beneficial effects of these and other structurally related polyphenols in a variety of disease states. For example, both CB1 antagonists/inverse agonists and polyphenols (including trans-resveratrol and curcumin) are efficacious anti-inflammatory agents (Rahman et al., 2006; Muccioli, 2007) and seem to be promising therapeutics for use in cardiovascular disease, cancer, stroke, and diabetes (Scalbert et al., 2005). In addition, curcumin has been used for centuries in the traditional Indian Ayurveda system of medicine to reduce the hallucinatory effects of many psychotropic drugs, including hashish, a potent form of cannabis (Tilak et al., 2004). However, the most direct evidence supporting our observations that certain polyphenols may produce actions through CB1 receptors is provided by the recent report that trans-resveratrol and several other polyphenols bind to a specific, yet unidentified, binding site in rat brain (Han et al., 2006). Similar to CB1 receptors, these binding sites are localized to plasma membranes, expressed in high density, and widely distributed throughout the brain. It is most interesting that [3H]trans-resveratrol binds to these unidentified sites, with an affinity (Kd) of 220 nM, very similar to its affinity (Ki) for mCB1 receptors of 270 nM reported in this study. It is certainly possible that [3H]trans-resveratrol might also bind to the orphan receptor GPR55 or to other noncannabinoid G-protein-coupled receptors, such as dopamine receptors, to which cannabinoid receptor ligands also bind.
It is interesting that all three polyphenols were shown to possess both neutral antagonist and inverse agonist properties, depending on the assay or tissue/cell homogenate examined. These data suggest that the polyphenols tested might act as protean agonists at CB1 receptors, similar to that recently described for the CB2 ligand AM-1241 (Yao et al., 2006). A protean agonist is a compound that changes its apparent intrinsic activity to exhibit agonist, antagonist, or inverse agonist activity at the same receptor, depending on the specific assay systems employed for detection. Alternatively, a more simple explanation for the current observations might be due to differences between assay conditions used for the GTPγS binding assay (employing membrane homogenates and relatively high concentrations of guanine nucleotides), relative to that employed for the cAMP assay (employing whole cells).
trans-Resveratrol and curcumin, like most polyphenols, are extensively and rapidly metabolized by glucuronidation and sulfation in the liver and other tissues (Singh et al., 2008). This predicts that relatively poor bioavailability, particularly in the central nervous system, might preclude observation of significant antagonism of effects mediated by central CB1 receptors in mice as reported here. However, even with such unfavorable pharmacokinetic properties, peak serum concentrations in mice of approximately 1 to 2 μM parent drug after a single, acute intraperitoneal injection of moderate doses (∼20–100 mg/kg) of either trans-resveratrol (Asensi et al., 2002) or curcumin (Pan et al., 1999) have been reported. In addition, curcumin can accumulate to concentrations as high as 1 to 2 μM in the brains of mice fed a relatively low dose of 2 mg/kg/day over a period of 3 to 4 months (Begum et al., 2008). Very low doses of trans-resveratrol protect against neuronal damage after cerebral ischemia, providing evidence that this polyphenol is also able to cross the blood-brain barrier in sufficient concentrations to provide neuroprotection (Wang et al., 2003). Lastly, in humans, consumption of a single oral 7.5 μg/kg dose of dietary trans-resveratrol contained in red wine results in a serum concentration of ∼26 nM and a 25 mg/70 kg body weight oral dose of pure trans-resveratrol results in a serum concentration of ∼37 nM (for review, see Baur and Sinclair, 2006). Although no chronic consumption studies in humans have been conducted, it might be predicted that serum levels of trans-resveratrol occurring in daily red wine drinkers might be even higher than those observed after a single exposure. Based on their high nanomolar affinities for CB1 receptors reported here, if such micromolar (or even high nanomolar) concentrations of trans-resveratrol, curcumin, or ASC-J9 are attained in the brain, near-full receptor occupancy would be predicted. Alternatively, it is also certainly possible that a metabolite of trans-resveratrol and/or curcumin might bind with high (or superior) affinity to CB1 receptors to mediate the in vivo effects reported here. In any case, because of the potential therapeutic promise of these drugs in a number of disease states, several methods to improve their systemic bioavailability, including the development of liposomal and nanoparticle preparations, are actively being pursued (Anand et al., 2007). Based on the present findings, future development of polyphenol-based CB1 ligands should include similar studies to improve systemic bioavailability.
Activation of peripheral CB1 receptors is effective at suppressing inflammation that leads to chronic pain states (Gutierrez et al., 2007). However, the potential use of current CB1 agonists for this application is severely limited by concurrent stimulation of central CB1 receptors, resulting in unacceptable psychotropic side effects. Furthermore, the CB1 antagonist/inverse agonist rimonabant is very effective for management of obesity (Pavon et al., 2008). However, several adverse effects, presumed to be mediated via blockade of central CB1 receptors, resulted in the recent discontinuance of all ongoing clinical trials of rimonabant in Europe (Jones, 2008), thus virtually assuring a lack of future Food and Drug Administration approval for use in the United States. Several studies indicate that the metabolic benefits of CB1 antagonists/inverse agonists in obese animals is due to action at peripheral, but not central, CB1 receptors (Pavon et al., 2008). Results from the present study demonstrating that repeated administration of curcumin or trans-resveratrol produces a dose-dependent reduction in body weight provide additional evidence for this observation. It is interesting that, although not attributed to action at CB1 receptors, others also report that trans-resveratrol reduces body weight in Zucker obese rats (Lekli et al., 2008). Therefore, polyphenol-derived, peripherally restricted CB1 agonists or antagonists might be developed as a novel class of nontoxic cannabinoids. The observation that high doses of either trans-resveratrol (Espín et al., 2007) or curcumin (Chainani-Wu, 2003) seem to be well tolerated and produce a very limited number of adverse side effects in humans provides further support for this hypothesis.
Last, as an additional advantage, it is likely that polyphenol-derived CB1 ligands could be developed that posses multiple therapeutic actions because of their pleiotropic action at several distinct targets simultaneously, in addition to their action at CB1 receptors. Such novel CB1 antagonists/inverse agonists might be particularly useful for the treatment of several disease states. For example, current CB1 antagonists/inverse agonists seem to be very efficacious for the management of obesity (Pavon et al., 2008). Antioxidants also reduce many adverse consequences associated with obesity (Vincent et al., 2007). As such, novel polyphenol-derived CB1 antagonists, because of combined CB1 antagonism and anticipated antioxidant properties (Fraga, 2007), might provide additive or even synergistic improvement of obesity symptoms.
We thank Dr. Debra A. Kendall (University of Connecticut, Storrs, CT) for the generous gift of the CHO-hCB1 cells used in the study.
This work was supported by the Amyotrophic Lateral Sclerosis Association [Grant 1311].
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: ASC-J9, 1,7-bis(3,4-dimethoxyphenyl)-5-hydroxy-1E,4E,6E-heptatriene-3-one; [3H]CP-55,950, (-)-cis-3-[2-hydroxy-4-(1,1-dimethylheptyl)phenyl]-trans-4-(3-hydroxypropyl) cyclohexanol; [35S]GTPγS, guanosine 5′-O-(3-[35S]thio)triphosphate; CHO, Chinese hamster ovary; h, human; WIN-55,212-2, (R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone mesylate; CP-55,940, (1R,3R,4R)-3-[2-hydroxy-4-(1,1-dimethylheptyl)phenyl]-4-(3-hydroxypropyl)cyclohexan-1-ol; ANOVA, analysis of variance; HU-210, (-)-11-hydroxy-δ(8)-tetrahydrocannabinol-dimethylheptyl; AM-251, N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2, 4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide; m, mouse; O-2050, (6aR,10aR)-3-(1-methanesulfonylamino-4-hexyn-6-yl)-6a,7,10,10a-tetrahydro-6,6,9-trimethyl-6H-dibenzo[b,d]pyran; rimonabant, 5-(p-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-N-piperidinopyrazole-3-carboxamide hydrochloride; AM1241, (R,S)-3-(2-iodo-5-nitrobenzoyl)-1-(1-methyl-2-piperidinylmethyl)-1H-indole.
- Received January 29, 2009.
- Accepted April 8, 2009.
- The American Society for Pharmacology and Experimental Therapeutics