Molecular targets for cannabidiol and its synthetic analogues: e€ect on vanilloid VR1 receptors and on the cellular uptake and enzymatic hydrolysis of anandamide

1 (7)-Cannabidiol (CBD) is a non-psychotropic component of Cannabis with possible therapeutic use as an anti-in¯ammatory drug. Little is known on the possible molecular targets of this compound. We investigated whether CBD and some of its derivatives interact with vanilloid receptor type 1 (VR1), the receptor for capsaicin, or with proteins that inactivate the endogenous cannabinoid, anandamide (AEA).
2 CBD and its enantiomer, (+)-CBD, together with seven analogues, obtained by exchanging the C-7 methyl group of CBD with a hydroxy-methyl or a carboxyl function and/or the C-5′ pentyl group with a di-methyl-heptyl (DMH) group, were tested on: (a) VR1-mediated increase in cytosolic Ca2+ concentrations in cells over-expressing human VR1; (b) [14C]-AEA uptake by RBL-2H3 cells, which is facilitated by a selective membrane transporter; and (c) [14C]-AEA hydrolysis by rat brain membranes, which is catalysed by the fatty acid amide hydrolase. 3 Both CBD and (+)-CBD, but not the other analogues, stimulated VR1 with EC50=3.2 ± 3.5 mM, and with a maximal e.ect similar in e.cacy to that of capsaicin, i.e. 67 ±70% of the e.ect obtained with ionomycin (4 mM). CBD (10 mM) desensitized VR1 to the action of capsaicin. The e.ects of maximal doses of the two compounds were not additive. 4 (+)-5′-DMH-CBD and (+)-7-hydroxy-5′-DMH-CBD inhibited [14C]-AEA uptake (IC50=10.0 and 7.0 mM); the (7)-enantiomers were slightly less active (IC50=14.0 and 12.5 mM). CBD and (+)- CBD were also active (IC50=22.0 and 17.0 mM). 5 CBD (IC50=27.5 mM), (+)-CBD (IC50=63.5 mM) and (7)-7-hydroxy-CBD (IC50=34 mM), but not the other analogues (IC504100 mM), weakly inhibited [14C]-AEA hydrolysis. 6 Only the (+)-isomers exhibited high a.nity for CB1 and/or CB2 cannabinoid receptors. 7 These ®ndings suggest that VR1 receptors, or increased levels of endogenous AEA, might mediate some of the pharmacological e.ects of CBD and its analogues. In view of the facile high yield synthesis, and the weak a.nity for CB1 and CB2 receptors, (7)-5′-DMH-CBD represents a valuable candidate for further investigation as inhibitor of AEA uptake and a possible new therapeutic agent. British Journal of Pharmacology (2001) 134, 845 ± 852

Keywords: Cannabinoid; endocannabinoid; vanilloid; receptors; FAAH; anandamide transporter
Abbreviations: AEA, arachidonoylethanolamide; AMT, anandamide membrane transporter; CBD, (7)-cannabidiol; [Ca2+]i, cytosolic calcium concentration; DMH, di-methyl-heptyl; FAAH, fatty acid amide hydrolase; HEK, human hembryonic kidney; HEK-hVR1, HEK cells transfected

Molecular targets for cannabidiol and its synthetic analogues: e€ect on vanilloid VR1 receptors and on the cellular uptake and enzymatic hydrolysis of anandamide

1Tiziana Bisogno, 3LumõÂr HanusÆ, 2Luciano De Petrocellis, 3Susanna Tchilibon, 3Datta E. Ponde, 1Ines Brandi, 2Aniello Schiano Moriello, 4John B. Davis, 3Raphael Mechoulam & *,1Vincenzo Di Marzo

1Endocannabinoid Research Group, Istituto per la Chimica di Molecole di Interesse Biologico, Consiglio Nazionale delle Ricerche, Via Campi Flegrei 34, Comprensorio Olivetti, Fabbricato 70, 80078 Pozzuoli (Napoli), Italy; 2Endocannabinoid Research Group, Istituto di Cibernetica, Consiglio Nazionale delle Ricerche, Via Campi Flegrei 34, Comprensorio Olivetti, Fabbricato 70, 80078 Pozzuoli (Napoli), Italy; 3Department of Medicinal Chemistry and Natural Products, Faculty of Medicine, Hebrew University, Jerusalem 91120, Israel and 4Neurology-CEDD, GlaxoSmithKline, New Frontiers Science Park, Third Avenue, Harlow, Essex CM19 5AW

Introduction

Among the bioactive constituents of Cannabis sativa, (7)- cannabidiol (CBD, Figure 1) is one of those with the highest potential for therapeutic use (Mechoulam, 1999). Although the pharmacological properties of the other major Cannabis component, (7)-D9-tetrahydrocannabinol (THC), have been more thoroughly investigated (Mechoulam, 1999; Pertwee, 1999, for reviews), THC, unlike CBD, exhibits potent psychotropic e.ects, which have complicated the full assessment of its therapeutic potential. Little is known of the molecular mechanism(s) of action of CBD, which, unlike THC, has very little a.nity for either cannabinoid receptor subtypes identi®ed so far, the CB1 and CB2 receptors (Pertwee, 1997, for review). Recent studies,  together with the earlier ®nding of the anti anxiety (Guimaraes et al., 1994), neuro-protective and anti-con- vulsive activity of CBD and some of its analogues (Consroe et al., 1981; Martin et al., 1987), indicate that CBD may also exert cyto-protective e.ects by inhibiting the release of in¯ammatory cytokines from blood cells (Srivastava et al., 1998; Malfait et al., 2000), thus producing an anti- in¯ammatory action, for example against rheumatoid arthritis (Malfait et al., 2000). These e.ects of CBD may be due to its anti-oxidant properties (Hampson et al., 1998), to its direct interaction with cytochrome p450- enzymes (Bornheim & Correia, 1989) and other enzymes of the `arachidonate cascade’ (Burstein et al., 1985), or to an action at a speci®c receptor. Recent studies have investigated whether CBD interacts with proteins of the `endocannabinoid signalling system’ other than the CB1/CB2 receptors. These proteins are: (i) fatty acid amide hydrolase (FAAH) (Cravatt et al., 1996), the intracellular enzyme catalysing the hydrolysis of the endogenous cannabinoid ligand, anandamide (arachidonoylethanolamide, AEA) (Ueda et al., 2000, for review); and (ii) the `anandamide membrane transporter’ (AMT) (Di Marzo et al., 1994), which facilitates the transport of AEA across the cell membrane and, subsequently, its intracellular degradation (Hillard & Jarrahian, 2000, for review). It was found that CBD inhibits both AEA hydrolysis by FAAH-containing membrane preparations (Watanabe et al., 1996), and AEA uptake by RBL± 2H3 cells via the AMT (Rakhshan et al., 2000). Although these e.ects were observed at high mM concentrations, these ®ndings raised the possibility that some of the pharmacological actions of CBD might be due to inhibition of AEA degradation, with subsequent enhancement of the endogenous levels of this mediator, for which neuroprotective (Hansen et al., 1998) and anti- in¯ammatory (Di Marzo et al., 2000a) properties have been previously suggested.

Many pharmacological activities of CBD have been established only in vivo, hence some of them may be due to CBD metabolites. The metabolism of CBD is well estab- lished. The primary step is hydroxylation on C-7, leading to (7)-7-hydroxy-CBD, followed by further oxidation to (7)-7- carboxy-CBD (Agurell et al., 1986). Although the metabolism of the dimethyl-heptyl homologue of CBD and of the (+) enantiomer of CBD has not been investigated, it is reasonable to assume that it follows the same pathways. Hence we prepared these CBD metabolites, their DMH homologues and some of the respective metabolites in the unnatural (+) series. In particular, in the present study we have examined whether the stereochemistry and the presence of certain chemical groups on the C-5′ and C-1 of CBD a.ect its capability of in¯uencing AEA inactivation via the AMT and FAAH. Furthermore, we have addressed the question of the possible molecular transducer of CBD by studying the possibility that this natural compound, its (+)-enantiomer and some of its synthetic analogues, interact with another proposed target for AEA, i.e. the vanilloid receptor type 1 (VR1) for capsaicin (Holzer, 1991, Figure 1). This protein is a ligand-, heat- and proton-activated non-speci®c cation channel acting as a molecular integrator of nociceptive stimuli (Tominaga et al., 1998). Recently, it was discovered that AEA is a full, albeit weak, VR1 agonist (Zygmunt et al., 1999; Smart et al., 2000) and that synthetic capsaicin analogues can interact with either CB1 receptors or the AMT, or both (Di Marzo et al., 1998). Thus, there appears to be some overlap between the ligand recognition properties of VR1 and CB1 receptors and, in particular, of VR1 and the AMT (De Petrocellis et al., 2000; Szallasi & Di Marzo, 2000).

Although VR1, via the release of in¯ammatory and algesic peptides, is involved in in¯ammatory hyperalgesia (Davis et al., 2000; Caterina et al., 2000), the stimulation of this receptor by capsaicin and some of its analogues leads to rapid desensitization, with subsequent paradoxical analgesic and anti-in¯ammatory e.ects (Holzer, 1991; Szallasi & Blumberg, 1999). As a consequence of this tachyphylactic e.ect, capsaicin, like CBD, has been used to treat arthritis (Lorton et al., 2000) and convulsions (Dib & Falchi, 1996). We report data suggesting that VR1 is a possible molecular target for CBD, and that inhibitors of the AMT can be developed by chemical modi®cation of this natural product.

Methods

Materials

The synthesis of some of the compounds assayed in this study will be described separately. CBD, whose structure and stereochemistry were described many years ago (Mechoulam & Shvo, 1963; Mechoulam & Gaoni, 1967), was isolated from hashish. (7)-5′-DMH-CBD, (+)-CBD and (+)-5′-DMH-CBD were prepared as described pre- viously (Baek et al., 1985; Leite et al., 1982). The synthesis of the CBD metabolite, (7)-7-hydroxy-CBD was recently reported (Tchilibon & Mechoulam, 2000). [14C]-AEA (5 mCi mmol71) was synthesized from [14C]-ethanolamine and arachidonoyl chloride as described (Devane et al., 1992b). Capsaicin, ionomycin and capsazepine were pur- chased from Sigma.

Cytosolic calcium concentration ([Ca2+]i) assay

Over-expression of human VR1 cDNA into human hem- bryonic kidney (HEK) 293 cells was carried out as described previously (Hayes et al., 2000). Cells were grown as monolayers in minimum essential medium supplemented with non-essential amino acids, 10% foetal calf serum and 0.2 mM glutamine, and maintained under 95%/5% O2/CO2 at 378C. The e.ect of the substances on [Ca2+]i was determined by using Fluo-3, a selective intracellular ¯uorescent probe for Ca2+ (Smart et al., 2000; De Petrocellis et al., 2000). One day  prior to experiments, cells were transferred into 6-well dishes coated with Poly-L-lysine (Sigma) and grown in the culture medium mentioned above. On the day of the experiment the cells (50 ± 60,000 per well) were loaded for 2 h at 258C with 4 mM Fluo-3 methylester (Molecular Probes) in DMSO containing 0.04% Pluoronic. After the loading, cells were washed with Tyrode pH=7.4, trypsinized, resuspended in Tyrode and transferred to the cuvette of the ¯uorescence detector (Perkin-Elmer LS50B) under continuous stirring. Experiments were carried out by measuring cell ¯uorescence at 258C (lEX=488 nm, lEM=540 nm) before and after the addition of the test compounds at various concentrations. Capsazepine (10 mM) was added 30 min before CBD. Data are expressed as the concentration exerting a half-maximal e.ect (EC50). The e.cacy of the e.ect was determined by comparing it to the analogous e.ect observed with 4 mM ionomycin. Response calibration was carried out by measur- ing the ¯uorescence intensity of intracellular ¯uo-3 with known extracellular [Ca2+] (Molecular Probes). The following equation was used to determine an ion dissociation constant (Kd) of 325 nM:

.Ca2..free . Kd.F ÿ Fmin.=.Fmax ÿ F.

; where Fmin and Fmax are the ¯uorescence intensities of ¯uo-3 without or with maximal [Ca2+], and F is the ¯uorescence intensity with an intermediate [Ca2+]. Average FEM/FEX was 200 and this value was increased by 60+7% in the presence of 4 mM ionomycin.

VR1 receptor binding assays

The a.nity of CBD and (+)-CBD for human VR1 receptors was assessed by means of displacement assays carried out with membranes (50 mg tube71) from HEK- hVR1 cells, prepared as described previously (Ross et al., 2001), and the high a.nity VR1 ligand [3H]-resiniferatoxin (48 Ci mmol71, NEN-Dupont), using the incubation condi- tions described previously (Ross et al., 2001). Under these conditions the Kd and Bmax for [3H]-resiniferatoxin were 0.5 nM and 1.39 pmol mg71 protein. The Ki for the displacement of 1 nM [3H]-resiniferatoxin by increasing concentrations of CBD and (+)-CBD was calculated from the IC50 values (obtained by GraphPad Software) using the Cheng ± Pruso. equation. Speci®c binding was calculated with 1 mM resiniferatoxin (Alexis Biochemicals) and was 78.1+3.7%.

Cannabinoid CB1 and CB2 receptor binding assays

These methods have been described previously by Devane et al. (1992a) for CB1, and Bayewitch et al. (1996) for CB2. For CB1 receptor binding assays synaptosomal membranes from rat brains were used. Sabra male rats weighing 250 ± 300 g were decapitated and their brains, without the brain stem, were quickly removed. Synaptosomal membranes were prepared from the brains by centrifugations and gradient centrifugation after their homogenization. The synaptosomal proteins thus obtained were used in the binding assay. The CB2 receptor binding assays were performed with transfected cells. COS-7 cells were transfected with plasmids containing CB2 receptor cDNA, and crude membranes were prepared. The high a.nity CB1/CB2 receptor ligand, [3H]-HU-243, with a dissociation constant of 45 pM, was incubated with synaptosomal membranes (3 ± 4 mg), for CB1 assays, or transfected cells, for CB2 assays, for 90 min at 308C with the di.erent cannabidiol derivatives or with the vehicle alone, and then centrifuged at 13,000 r.p.m. for 6 min. Bound and free radioligand were separated by centrifugation. The data were normalized to 100% of speci®c binding, which was determined with 50 nM unlabelled HU-243. All experiments were repeated 2 ± 3 times and each point performed in triplicate. The Ki values were determined with a GraphPad Prism program version 2.01 (San Diego, CA, U.S.A.) and using the Cheng ± Pruso. equation.

Anandamide cellular uptake assay

The e.ect of compounds on the uptake of [14C]-AEA by rat basophilic leukaemia (RBL ± 2H3) cells was studied by using 3.6 mM (10,000 c.p.m.) of [14C]-AEA as described previously (Bisogno et al., 1997). Cells were incubated with [14C]-AEA for 5 min at 378C, in the presence or absence of varying concentrations of the inhibitors. Residual [14C]-AEA in the incubation media after extraction with CHCl3/ CH3OH 2 : 1 (by vol.), determined by scintillation counting of the lyophilized organic phase, was used as a measure of the AEA that was taken up by cells (De Petrocellis et al., 2000). Data are expressed as the concentration exerting 50% inhibition of AEA uptake (IC50) calculated by GraphPad.

Fatty acid amide hydrolase assay

The e.ect of CBD and its analogues on the enzymatic hydrolysis of AEA was studied as described previously (Bisogno et al., 1997), using cell membranes from mouse neuroblastoma (N18TG2) cells, incubated with com- pounds and [14C]-AEA (9 mM) in 50mM Tris-HCl, pH 9, for 30 min at 378C. [14C]-Ethanolamine produced from [14C]-AEA hydrolysis was measured by scintillation counting of the aqueous phase after extraction of the incubation mixture with 2 volumes of CHCl3/CH3OH 2 : 1 (by vol.). Data are expressed as the concentration exerting 50% inhibition of AEA uptake (IC50), calculated by GraphPad.

Statistics

Means were compared by means of analysis of variance followed by Bonferroni’s test (ANOVA, StatMostTM soft- ware, DataMost Corp.).

Results

Effect of CBD analogues on human vanilloid VR1 receptors The e.ects upon [Ca2+]i in HEK± hVR1 cells of CBD, (+)- CBD, and (7)-7-hydroxy-5′-DMH-CBD (HU-317) are shown in Figure 2. These three compounds all induced an increase in [Ca2+]i and behaved as full agonists as compared to capsaicin, although only the two former compounds exerted this e.ect with an EC50510 mM and independently of  their stereochemistry. In fact, the potency (EC50 3.2+0.4) and e.cacy (max. e.ect 68.5+3.1% of the e.ect of 4 mM ionomycin) of (+)-CBD were indistinguishable from those of CBD (EC50 3.5+0.3, max. e.ect 64.1+3.9%, means+s.e.- mean, n=4). The e.cacy of both compounds was almost identical to that of a maximal concentration of capsaicin (70.2+3.5% at 10 mM, mean+s.e.mean, n=4), which was however 100 fold more potent (EC50=26+9 nM).

The other six CBD analogues examined in this study were all inactive or very weakly active on [Ca2+]i. The e.ect of CBD was abolished by the VR1 receptor antagonist capsazepine (10 mM, Figure 2) and could not be observed in wild-type HEK cells (data not shown). This strongly suggests that this e.ect, like that of capsaicin, was due to stimulation of VR1 receptors. Binding assays for the displacement of [3H]- resiniferatoxin from HEK± hVR1 cell membranes by CBD and (+)-CBD con®rmed this hypothesis, and showed that CBD and (+)-CBD compete for the binding of [3H]- resiniferatoxin with Ki values (3.6+0.2 and 3.0+0.3 mM, respectively, means+s.d., n=3) similar to the EC50 values for the e.ect on [Ca2+]i. Furthermore, capsaicin (0.1 mM) and CBD (10 mM) exhibited cross-desensitization of their e.ect on [Ca2+]i, providing evidence consistent with these two compounds acting at the same receptor. A 1 h pre-exposure to 0.1 mM capsaicin reduced the e.ect of 10 mM CBD from 66.7+3.4 to 11.7+1.5% of the e.ect of 4 mM ionomycin, whereas a 1 h pre-exposure to 10 mM CBD reduced the e.ect of 0.1 mM capsaicin from 68.1+4.1 to 22.3+3.5% (means+  s.e.mean, n=4, P50.01 by ANOVA). Co-treatment of cells with both capsaicin (10 mM) and CBD (10 mM) did not produce any additive e.ect (72.2+4.1% of the e.ect of ionomycin, mean+s.e.mean, n=4).

Affinity of CBD analogues for cannabinoid CB1 and CB2 receptors

Of the nine compounds tested, the (7) analogues were all weakly active or inactive (Ki410 mM) in binding assays for CB1 and CB2 receptor a.nity (Table 1), with the exception of (7)-7-hydroxy-5′-DMH-CBD, which exhibited a weak a.nity for CB2 receptors (Ki=0.7 mM). By contrast, of the three (+)-analogues tested, (+)-5′-DMH±CBD and (+)-7- hydroxy-5′-DMH±CBD behaved as high a.nity CB1 receptor ligands (Ki=17.4 and 2.5 nM), and were 10 ± 20 fold less active as CB2 receptor ligands (Ki=211 and 44.0 nM). Effect of CBD analogues on the anandamide membrane transporter Of the nine analogues tested, only (+)- and (7)-CBD, (+)- and (7)-5′-DMH-CBD and (+)- and (7)-7-hydroxy-5′- DMH±CBD inhibited the uptake of [14C]-AEA from RBL± 2H3 cells with IC50 values lower than 25 mM (Figure 3 and Table 1). Of these six compounds, the (+)-enantiomers were signi®cantly (P50.05 by ANOVA) and consistently more active than the (7)-enantiomers, and, in particular, (+)-5′-DMH-CBD and (+)-7-hydroxy-5′-DMH-CBD were as potent as the AMT inhibitor, AM404 (Khanolkar & Makriyannis, 1999) (IC50=10.0, 7.0 and 8.1 mM for the two compounds and AM404, respectively). (7)-7-hydroxy-5′- DMH-CBD and (7)-5′-DMH-CBD were also almost as potent as AM404 (IC50=12.5 and 14.0 mM). The IC50 of CBD (22.0 mM) was higher than that previously reported for this compound in the same cell line (11.4 mM; Rakhshan et al., 2000).

Effect of CBD analogues on fatty acid amide hydrolase

Only (+)- and (7)-CBD, and (7)-7-hydroxy-CBD exhibited a IC505100 mM for the inhibition of [14C]-AEA hydrolysis by N18TG2 cell membrane preparations (Figure 4 and Table 1), which express high levels of FAAH. The (7)-enantiomer was signi®cantly more potent than the (+)-enantiomer, CBD being also the most potent compound found (IC50=27.5 mM). However, none of the CBD analogues tested can be considered a potent inhibitor of FAAH (i.e. with an IC50420 mM). The activity of CBD in this study was higher than that previously reported for anandamide hydrolysis by mouse brain (Watanabe et al., 1996; 1998), where however only a very high concentration of the compound (160 mM) was used.

Discussion

In this study we investigated whether CBD, now being considered as a possible therapeutic agent (Straus, 2000), is capable of interacting with the recently cloned vanilloid VR1 receptor. In fact, some of the pharmacological actions of CBD are similar to those of natural (e.g. capsaicin) and synthetic agonists of VR1. Although stimulation of VR1 receptors leads to vasodilation and in¯ammation, capsaicin and its long chain analogues exert anti-in¯ammatory e.ects by rapidly desensitizing VR1 receptors to the action of nociceptive stimuli and causing depletion of sensory vasoactive neuropeptides (Szallasi & Blumberg, 1999). CBD also induces anti-in¯ammatory e.ects, a possible explanation for this property being its capability of modulating the release of anti-in¯ammatory or pro-in¯ammatory mediators (Srivas- tava et al., 1998, Malfait et al., 2000). CBD and capsaicin also have in common anti-convulsive and anti-rheumatoid- arthritis e.ects (Consroe et al., 1981; Dib & Falchi, 1996; Malfait et al., 2000; Lorton et al., 2000). Here we found that CBD, compared to capsaicin, is a full, although weak, agonist of human VR1 at concentrations that might be  attained after administration of this compound at the doses often used in vivo (10 ± 50 mg kg71 in men), and lower than those required for CBD to bind to cannabinoid receptors. CBD desensitized VR1 to the action of capsaicin, thus opening the possibility that this cannabinoid exerts an anti- in¯ammatory action in part by desensitization of sensory nociceptors. Future studies with capsazepine (which antag- onizes capsaicin e.ects in rats (Di Marzo et al., 2001) but not always in mice (Di Marzo et al., 2000b), and VR1 `knockout’ mice (Davis et al., 2000; Caterina et al., 2000), should test the involvement of VR1 in the pharmacological actions of CBD. We found that insertion in CBD of a DMH instead of an n-pentyl chain on the C-5′, or of a carboxyl function instead of the methyl group on the C-1, abolishes the capability of the cannabinoid to induce a VR1-mediated functional response, whereas insertion of both a hydroxy-group on the C-7 and of a 5′-DMH group decreases the potency but not the e.cacy of CBD. By contrast, inversion of the stereochemistry does not modify the activity of CBD. These data suggest that the C-1 methyl and the aromatic `A’ ring, which is chemically similar to the vanillyl moiety of capsaicin (Figure 1), are more important than the chiral part of CBD for its interaction with VR1. That CBD binds to the same site as capsaicin is suggested by the ®nding that both compounds displace [3H]-resiniferatoxin from its speci®c binding sites in membranes from cells over-expressing VR1 receptors (this study and Ross et al., 2001). However, while capsaicin exhibits higher potency than a.nity for vanilloid receptors (Szallasi & Blumberg, 1999), CBD is as active in the [3H]-resiniferatoxin binding assay as in the hVR1 functional assay. This suggests that CBD is less capable than capsaicin to induce a VR1-mediated functional response at low concentrations, even though the e.cacy of high concentra- tions of both compounds is the same. It should be noted that the cells used here to assess the functional activity at VR1 express high levels of this receptor, and that the potency and e.cacy of CBD in native cells containing lower amounts of vanilloid receptors might be lower than those observed here.

The endocannabinoid AEA is thought to exert anti- in¯ammatory and neuroprotective actions (Di Marzo et al., 2000a; Hansen et al., 1998). As it was found to inhibit the re- uptake and hydrolysis of AEA in vitro (Rakhshan et al., 2000; Watanabe et al., 1996), it is possible that CBD acts in part by interfering with AEA inactivation, thereby enhancing the putative tonic inhibitory action of AEA on in¯ammation. The fact that the pharmacological actions of CBD are not in¯uenced by CB1/CB2 receptor antagonists should not be taken as evidence against this hypothesis, since it is now established that AEA also acts upon non-cannabinoid receptor targets, including VR1 receptors and TASK-1 K+ channels (Zygmunt et al., 1999; Maingret et al., 2001). Here we con®rmed that CBD inhibits AEA transporter-mediated uptake by cells and enzymatic hydrolysis. We also found that analogues of CBD are inhibitors of the AMT, and that this property is more pronounced with (+)-enantiomers, or when the C-7 and C-5′ are derivatized with a hydroxyl- and a DMH group, respectively. The most potent inhibitor found was (+)-7-hydroxy-5′-DMH-CBD. However, this compound exhibited high a.nity for CB1 and CB2 receptors. Also (+)- 5′-DMH-CBD was more active as a CB1 and CB2 receptor ligand than as an AMT inhibitor. By contrast, (7)-7- hydroxy-5′-DMH-CBD and (7)-5′-DMH-CBD, which were almost as potent as AM404 against the AMT, but, unlike AM404, had low a.nity for the two cannabinoid receptors subtypes and no activity at VR1, may represent metabolically stable and relatively selective pharmacological tools for the study of AEA inactivation in vitro. The (7)-5’DMH-CBD is obtained by a facile, high yield synthesis (Baek et al., 1985) and may ®nd application as therapeutic agent for those disorders where AEA exerts an endogenous tone with bene®cial e.ects. The novel AMT inhibitors developed here should be tested also on the cellular uptake of palmitoy- lethanolamide, a natural anti-in¯ammatory AEA congener (Lambert & Di Marzo, 1999), since a recent study showed that CBD inhibits the facilitated transport of this compound into RBL± 2H3 cells (Jacobsson & Fowler, 2001).

We have mentioned above that the (+)-enantiomers of the CBD analogues tested here on AEA cellular uptake were more potent inhibitors than the (7)-enantiomers. A certain enantio-speci®city for the interaction with the AMT has been noted previously also for AEA analogues (see Khanolkar & Makriyannis, 1999, for review). It was also noted that the same stereochemical preference existed for the interaction of these compounds with FAAH, whereas the interaction with CB1 receptors followed the opposite enantio-selectivity (Khanolkar & Makriyannis, 1999). We noted that CBD inhibits FAAH more potently than the (+)-enantiomer. By contrast, all but one of the (+)-CBD analogues tested exhibited much higher a.nity for CB1 receptors than their (7)-enantiomers. Thus, for CBD analogues, the stereoche- mical requisites for the interaction with the AMT and CB1 receptors are the same, and they may be opposite to those necessary for the interaction with FAAH. The binding data were indeed unexpected. In the tetrahydrocannabinol series the (7) (3R,4R) enantiomers bind to CB1 and have pharmacological activity in various typical cannabinoid assays, while the (+) (3S,4S) enantiomers are essentially inactive (Mechoulam et al., 1988; Howlett et al., 1990; Little et al., 1989; Jarbe et al., 1989). In the CBD series of compounds we observed here the opposite situation. The reason for this dichotomy is unknown. Further studies investigating the potency and e.cacy of the compounds in functional assays of CB1 receptor-mediated activity need to be performed in order to fully assess the cannabimimetic activity of the compounds in the (+)-CBD series. In conclusion, the present study has provided novel insights into the possible mechanism(s) of action of the natural cannabinoid CBD by identifying in VR1 receptors a novel potential molecular target for this compound. Furthermore, we have shown that potent inhibitors of AEA cellular uptake can be developed from certain chemical modi®cations of CBD, and have con®rmed that CBD can act in principle also by inhibiting AEA inactivation. Future studies will be needed to address the question of whether vanilloid receptors or endogenous cannabinoids contribute to the anti-in¯ammatory and neuroprotective actions of CBD.

We thank the U.S. National Institute on Drug Abuse (grant DA 9789, to R. Mechoulam), the Israel Science Foundation (to R. Mechoulan), the Yeshaya Horowitz Association (to R. Mechou- lam) and the MURST (3933 to V. Di Marzo) for support.