Cannabinoid Receptor Agonists

In terms of chemical structure, established cannabinoid receptor agonists fall essentially into four main groups: classical, nonclassical, aminoalkylindole and eicosanoid (reviewed in Howlett et al. 2002; Pertwee 1999a).

- The classical group consists of dibenzopyran derivatives that are either cannabis-derived compounds (phytocannabinoids) or their synthetic analogues. Notable examples are the phytocannabinoids49-THC,48-THC and cannabinol (Fig. 1), and the synthetic cannabinoids, 11-hydroxy-48-THC-dimethylheptyl(HU-210), JWH-133, L-759633, L-759656, L-nantradol and desacetyl-L-nantradol (Figs. 4 and 5).

Cannabinoid Receptor Chart
3-(1',1'-dimethyl-cyclohexyl)-As-THC Fig. 4. The structures of five synthetic classical cannabinoids
Fig. 5. The structures of four nonclassical cannabinoids

- Nonclassical cannabinoids consist of bicyclic and tricyclic analogues of49-THC that lack a pyran ring; examples include CP55940, CP47497, CP55244 and HU-308 (Fig. 6). They are, therefore, closely related to the classical cannabinoids.

- In contrast, the aminoalkylindole group of cannabinoid receptor agonists (Fig. 7) have structures that are completely different from those of other cannabinoids. Indeed, results from experiments performed with wild-type and mutant CBi receptors (Chin et al. 1998; Petitet et al. 1996; Song and Bonner 1996; Tao and Abood 1998) suggest that fl-(+)-WIN55212 (WIN55212-2), the most widely investigated of the aminoalkylindoles, binds differently to the CB1 receptor than classical, nonclassical or eicosanoid cannabinoids, albeit it in a manner that still allows mutual competition between £-(+)-WIN55212 and non-aminoalkylindole cannabinoids for binding sites on the wild-type receptor.

- Members of the eicosanoid group of cannabinoid receptor agonists have markedly different structures both from the aminoalkylindoles and from classical and nonclassical cannabinoids. Important members of this group are the endo-cannabinoids, arachidonoylethanolamide (anandamide), O-arachidonoylethan-olamine (virodhamine), 2-arachidonoyl glycerol and 2-arachidonyl glyceryl

Anandamide Structure


Fig. 6. The structures of four nonclassical cannabinoids. The (+)-enantiomer of CP55940 is CP56667


Fig. 6. The structures of four nonclassical cannabinoids. The (+)-enantiomer of CP55940 is CP56667

Fig. 7. The structures of Ä-(+)-WIN55212, JWH-015, AM1241, L-768242 and BML-190

ether (noladin ether) (Fig. 3) and several synthetic analogues of anandamide, including .-(+)-methanandamide, arachidonyl-2'-chloroethylamide (ACEA), arachidonylcyclopropylamide (ACPA), O-689 and O-1812 (Fig. 8) (Howlett et al. 2002; Pertwee 1999a; Porter et al. 2002).

Fig. 8. The structures of eight structural analogues ofanandamide

Many cannabinoid receptor agonists exhibit marked stereoselectivity in pharmacological assays, reflecting the presence of chiral centres in these compounds (reviewed in Howlett et al. 2002). Classical and nonclassical cannabinoids with the same absolute stereochemistry as (-)-49-THC at 6a and 10a, trans (6aR, 10aR), are more active than their cis (6aS, 10aS) enantiomers, whilst R-(+)-WIN55212 is more active than S-(-)-WIN55212. Although anandamide does not contain any chiral centres, some of its synthetic analogues do. One of these is methanandamide, the R-(+)-isomer of which exhibits nine times higher affinity for CB1 receptors than the S-(-)-isomer (Abadji et al. 1994).

Several cannabinoid receptor agonists bind more or less equally well to CBi and CB2 receptors (Table 2), although they do exhibit different relative intrinsic activities at these receptors. Among these are HU-210, CP55940, R-(+)-WIN55212, (-)-49-THC, anandamide and 2-arachidonoyl glycerol (reviewed in Howlett et al. 2002; Pertwee 1999a).

- HU-210 has particularly high affinity for both CBi and CB2 receptors. It also exhibits high relative intrinsic activities at these receptors. Indeed, it is remarkably potent as a cannabinoid receptor agonist and exhibits an exceptionally long duration of action in vivo. The marked affinity and efficacy that HU-210 shows at cannabinoid receptors is due largely to the replacement of the pentyl side chain of48-THC with a dimethylheptyl group.

- CP55940 and R-(+)-WIN55212 have CB1 and CB2 relative intrinsic activities of the same order as those of HU-210 and, although they have lower CBi and CB2 affinities than HU-210, are still reasonably potent as they bind to these receptors at concentrations in the low nanomolar range.

- (-)-49-THC has lower CBi and CB2 affinities and relative intrinsic activities than HU-210, CP55940 or fl-(+)-WIN55212. Whilst it behaves as a partial agonist at both these receptor types, it exhibits less efficacy at CB2 than at CB1 receptors to the extent that in one bioassay system it has been found to behave as a CB2 receptor antagonist (Bayewitch et al. 1996). (-)-49-THC can also produce CB1 receptor antagonism. Thus, it has been found to oppose CB1 receptor activation by the higher efficacy agonist, 2-arachidonoyl glycerol, in hippocampal cultures that may have contained neurons with rather low CB1 receptor density (Kelley and Thayer 2004). This it did with an IC50 of 42 nM, which is close to its reported CB1 Ki values (Table 2).

- Anandamide resembles (-)-49-THC in its affinity for CB1 receptors, in behaving as a CB1 and CB2 receptor partial agonist (Gonsiorek et al. 2000; Hillard 2000; Mackie et al. 1993; Savinainen et al. 2001; Sugiura et al. 1996,2000) and in having lower CB2 than CB1 intrinsic activity (reviewed in Howlett et al. 2002; Pertwee 1999a). It has also been found that, like (-)-49-THC, anandamide can behave as a CB2 receptor antagonist in at least one bioassay system (Gonsiorek et al. 2000). In contrast to £-(+)-WIN55212, which has slightly higher CB2 than CB1 affinity, anandamide binds marginally more readily to CB1 than to CB2 receptors.

- 2-Arachidonoyl glycerol is known to activate both CB1 and CB2 receptors. It binds about equally well to both receptor types (Table 2) and has been reported to exhibit greater CB1 intrinsic activity but less CB1 potency than CP55940 and greater CB1 intrinsic activity and potency than anandamide (Gonsiorek et al. 2000; Savinainen et al. 2001, 2003; Sugiura et al. 1996). This endocannabi-noid also has greater CB2 potency than anandamide or 1-arachidonoyl glycerol (Gonsiorek et al. 2000; Sugiura et al. 2000).

One recently developed synthetic cannabinoid receptor agonist that interacts almost as well with CB2 as with CB1 receptors (Tables 1 and 2) is BAY 38-7271 (De Vry and Jentzsch 2002; Mauler et al. 2002, 2003). This compound has a structure that is not classical, non-classical, aminoalkylindole or eicosanoid (Fig. 9).

Phytocannabinoids other than49-THC that are known to activate cannabinoid receptors are (-)-48-THC and cannabinol (reviewed in Pertwee 1999a). Of these, (-)-48-THC resembles (-)-49-THC both in its CB1 and CB2 receptor affinities (Table 2) and in its relative intrinsic activity at the CB1 receptor (Gérard et al. 1991; Howlett and Fleming 1984; Matsuda et al. 1990). Cannabinol also behaves as a partial agonist at CB1 receptors but has even less relative intrinsic activity than (-)-49-THC (Howlett 1987; Matsuda et al. 1990; Petitet et al. 1997, 1998). Whilst there is one report that cannabinol activates CB2 receptors in the cyclic AMP assay more effectively than49-THC (Rhee et al. 1997), there is another that in the GTPyS binding assay, it behaves as a CB2 receptor inverse agonist (MacLennan et al. 1998).

As to the endocannabinoid virodhamine, Porter et al. (2002) have shown that this activates both CB1 and CB2 receptors. Their experiments with transfected cells yielded CB1 and CB2 EC50 values in the GTPyS binding assay of 1.9 and 1.4 ^M, respectively, for this endocannabinoid, indicating it to be less potent than anandamide, 2-arachidonoyl glycerol or £-(+)-WIN55212. The CB2 intrinsic

1',1 '-Dimethylheptyl-A3-THC-11 -oic acid (Ajulemic acid)

Fig. 9. The structures of BAY 38-7271, JTE-907, ajulemic acid and Q-1057

1',1 '-Dimethylheptyl-A3-THC-11 -oic acid (Ajulemic acid)

Fig. 9. The structures of BAY 38-7271, JTE-907, ajulemic acid and Q-1057

activity of virodhamine matched that of anandamide which, however, behaved as a full agonist in this investigation, suggesting that the CB2 expression level of the cell line used may have been rather high. In contrast, the CB1 intrinsic activity of virodhamine was less than that of anandamide, and indeed it was found that virodhamine could attenuate anandamide-induced activation of CB1 receptors. No binding data are yet available for virodhamine.

Turning now to potent cannabinoid receptor agonists that interact more readily with CB1 or CB2 receptors, a number of these have been developed. The starting point for all current CB1-selective agonists has been anandamide. Thus, results from binding experiments have shown that it is possible to enhance the marginal CB1 selectivity exhibited by anandamide by replacing a hydrogen atom on the 1' or 2 carbon with a methyl group to form £-(+)-methanandamide or 0-689 (Fig. 8) (Abadji et al. 1994; Showalter et al. 1996). As well as increasing CB1 selectivity, insertion of a methyl group on the 1' or 2 carbon of anandamide increases resistance to the hydrolytic action of fatty acid amide hydrolase (FAAH) (Abadji et al. 1994; Adams et al. 1995). Anandamide analogues that exhibit particularly marked CB1-selectivity in binding assays are ACEA, ACPA and a cyano analogue of methanandamide (0-1812) (Table 2; Fig. 8). All three behave as potent CB1 receptor agonists (Di Marzo et al. 2001; Hillard et al. 1999). 0-1812 appears to lack significant susceptibility to hydrolysis by FAAH, presumably because it resembles £-(+)-methanandamide in having a methyl group attached to its 1'-carbon. ACEA and ACPA, which do not have the 1'-carbon methyl substituent of £-(+)-methanandamide, show no sign of reduced susceptibility to enzymic hy-

Table 4. Ki values of certain other ligands for the in vitro displacement of [3H]CP55940 or [3H]HU243a from CB1 - and CB2-specific binding sites


K i value (nM)

Ki value (nM)


CBi -selective ligands in order of decreasing CBi /CB2 selectivity




Goutopoulos et al. 2001





Showalter et al. 1996




Showalter et al. 1996

Ligands without any marked CB1 or CB2 selectivity

Ajulemicacid (CT-3)



Rhee et al. 1997




Rhee et al. 1997




Krishnamurthy et al. 2003




Rhee et al. 1997




Rhee et al. 1997




Rhee et al. 1997




Showalter et al. 1996



Bisogno et al. 2001




Rhee et al. 1997




Huffman et al. 1996




Papahatjis et al. 2002




Ross et al. 1999b

cis (6aS, 10aS)-3-(1',1'-DMH)-



Showalter et al. 1996

11-hydroxy-48-THC (HU-211)




Showalter et al. 1996

CB2-selective ligands in order of increasing CB1/CB2 selectivity




Showalter et al. 1996




Huffman et al. 1996





Iwamura et al. 2001




Gallant et al. 1996




Huffman et al. 1999

1-deoxy-48-THC (JWH-139)




Huffman et al. 2002





Huffman et al. 1999

See Figs. 1,4,5,7,8,9,11 and 12 for the structures of some of the compounds listed in this table. DMH, dimethylheptyl; ND, not determined; THC, tetrahydrocannabinol. bWith phenylmethylsulphonyl fluoride (PMSF) in order to inhibit enzymic hydrolysis. cBinding to rat cannabinoid receptors on transfected cells or on brain (CB1) or spleen tissue (CB2). d Binding to mouse brain (CB1) or spleen tissue (CB2).

eSpecies unspecified. All other data from experiments with human cannabinoid receptors.

drolysis. Although insertion of this group into ACEA does markedly reduce the susceptibility of this molecule to FAAH-mediated hydrolysis, it also decreases the affinity of ACEA for CBi receptors by about 14-fold (Jarrahian et al. 2000). R-N-(1-

methyl-2-hydroxyethyl)-2-R-methyl-arachidonamide, which also exhibits marked CBi-selectivity in binding assays (Table 4), has less metabolic stability than R-(+)-methanandamide (Goutopoulos et al. 2001). Another CB1-selective agonist of note is the endocannabinoid 2-arachidonyl glyceryl ether (Hanus et al. 2001), the CB1 intrinsic activity of which has been reported to match that of CP55940 and to be less than that of 2-arachidonoyl glycerol. 2-Arachidonyl glyceryl ether exhibits less potency at CB1 receptors than either CP55940 or 2-arachidonoyl glycerol (Savinainen et al. 2001,2003; Suhara et al. 2000,2001).

The best CB2-selective agonists to have been developed to date are all non-eicosanoid cannabinoids (Howlett et al. 2002; Ibrahim et al. 2003; Pertwee 1999a). They include the classical cannabinoids, L-759633, L-759656 and JWH-133, the non-classical cannabinoid HU-308, and the aminoalkylindole AM1241 (Figs. 5,6 and 7). All these ligands bind more readily to CB2 than to CB1 receptors (Table 2) and have also been shown to behave as potent CB2-selective agonists in functional bioassays (Hanus et al. 1999; Ibrahim et al. 2003; Pertwee 2000; Ross et al. 1999a).

One other cannabinoid receptor agonist of note is 3-(5'-cyano-1',1/-dimethyl-pentyl)-1-(4-N-morpholinobutyryloxy)-48-THC hydrochloride (0-1057). Thus, unlike all established cannabinoid receptor agonists, this is readily soluble in water and yet, compared to CP55940, its potency in the cyclic AMP assay is just 2.9 times less at CB1 receptors and 6.5 times less at CB2 receptors (Pertwee et al. 2000). The finding that it is possible to solubilize a cannabinoid and yet retain pharmacological activity has important implications for cannabinoid delivery not only in the laboratory but also in the clinic. As to structure-activity relationships for cannabinoid receptor agonists, the salient features of these have been well described elsewhere (Howlett et al. 2002; Pertwee 1999a). Recent findings of special interest are that the CB1 and CB2 affinities of 48-THC can be greatly enhanced both by replacing its C3 pentyl side chain with a 1/,1'-dimethyl-1'-cyclohexyl moiety (Fig. 4; Table 4) (Krishnamurthy et al. 2003) and by changing this side chain from pentyl to heptyl and introducing a cyclopropyl group at the 1' position (Fig. 4; Table 4) (Papahatjis et al. 2002).

0 0

Post a comment