CHROMENOPYRAZOLES: NON-PSYCHOACTIVE AND SELECTIVE CB₁ CANNABINOID AGONISTS WITH PERIPHERAL ANTINOCICEPTIVE PROPERTIES

Abstract

The unwanted psychoactive effects of cannabinoid receptor agonists have limited their development as medicines. These CB₁ mediated side effects are due to the fact that CB₁ receptors are largely expressed in the Central Nervous System (CNS). Since it is known that CB₁ receptors are also located peripherally, there is a growing interest in targeting cannabinoid receptors located outside the brain. A library of chromenopyrazoles designed in analogy to the classical cannabinoid cannabinol were synthesized, characterized and tested for cannabinoid activity. Radiolabeled binding assays were used to determine their affinities at CB₁ and CB₂ receptors. Structural features required for CB₁/CB₂ affinity and selectivity were explored using molecular modeling. Within the chromenopyrazoles series, some of them showed to be selective CB₁ ligands. These modeling studies suggest that CB₁ full selectivity over CB₂ can be accounted for the presence of a pyrazole ring in the structure. The functional activities of selected chromenopyrazoles were evaluated in isolated tissues. Behavioral tests, in vivo, were then carried on the most effective CB₁ cannabinoid agonist (13a). Chromenopyrazole 13a did not induce modifications in any of the tested parameters on the mouse cannabinoid tetrad, discarding CNS-mediated effects. This lack of agonistic activity in the CNS suggests that it does not readily cross the blood-brain barrier. Moreover, compound 13a can induce antinociception in a peripheral model of orofacial pain in rat. Taking into account the negative results obtained in the hot plate test, it could be suggested that the antinociception induced by 13a in the orofacial test may be mediated through peripheral mechanisms.

Introduction

The three major components of marijuana are Δ⁹-tetrahydrocannabinol (Δ⁹-THC), cannabidiol (CBD) and cannabinol (CBN).^[1] Unlike Δ⁹-THC,^[2] CBD and CBN are nonpsychotropic phytocannabinoids.^[3] Δ⁹-THC interacts with two well-characterized G protein-coupled receptors, CB₁ and CB1₂.^[4-5] The CB₁ receptors are localized with high density in the brain and are also found in peripheral tissues. On the other hand, the CB₂ receptors are expressed mainly in immune cells, although they were also found in CNS particularly under pathological circumstances. The activity of the cannabinoid receptors is elicited not only by phytocannabinoids, but also by synthetic ligands and endocannabinoids.^[6-9] The only cannabinoid receptor ligands prescribed so far are CB₁/CB₂ receptor agonists. Cesamet® (nabilone) and Marinol® (Δ⁹-THC) are used for the treatment of nausea and vomiting associated with cancer chemotherapy or as anti-emetic agents. Sativex® (Δ⁹-THC and CBD) is prescribed to relieve spasticity and pain of multiple sclerosis patients. However, preclinical data indicate that CB₁ and/or CB₂ receptor agonists are useful with diverse therapeutic applications that include pain relief, treatment of intestinal disorders, glaucoma, cancer proliferation and neurodegenerative diseases.^[10-12]

Δ⁹-THC and CBN are classical cannabinoids characterized by tricyclic terpenoids bearing a benzopyran moiety (Figure 1).^[13] Structure-activity relationships (SARs) of Δ⁹-THC and Δ⁸-THC analogues for CB₁/CB₂ receptors have been widely reported.^[14-19] It is well-established that C-1, C-3, and C-9 positions play a key role in binding affinity and pharmacological potency of THCs. Even though some of these classical cannabinoids have been reported to show significant selectivity for one of the two receptor types, structural variations within Δ⁹-THC generally resulted in derivatives with high affinity for both CB₁ and CB₂ receptors. Considerable effort has been directed toward the SARs of THCs; however less structural modifications have been made on the structure of cannabinol. Rhee et al. ^[20] reported binding evaluation and inhibition of adenylylcyclase data of a series of CBN derivatives. Unlike CBN that was found to be less potent than Δ⁹-THC, the 3-dimethylheptyl CBN analogue and the 9-hydroxymethyl CBN analogues showed more affinity and agonist potency than Δ⁹-THC at both CB₁ and CB₂ receptors. The presence of alkyl or aryl esters in position 9 of CBN resulted in weak CB₁ and CB₂ binding.^[21] More recently, cannabilactones have been reported by Khanolkar et al. ^[22] One of them exhibited high CB₂ receptor affinity with 500-fold selectivity over CB₁ receptor. Taking together all these considerations, it is clear that structural requirement for cannabinoid receptors binding of CBN series differs from THCs.

Figure 1.

Chemical structures of Δ⁹–THC, Δ⁸-THC and cannabinol derivatives.

The pharmacological properties of CBN have also received less attention than the THCs. Analgesic properties of CBN have been reported in different models of pain.23-26] [Thus, CBN generally needed higher doses than Δ⁹-THC to produce antinociception but it showed minimal psychomimetic effects.^[27-28] One of the main challenges at designing new cannabinoid ligands is to develop cannabinoids devoid of CNS side effects. In 1985 Press et al. ^[29] published benzopyrano[4,3-c]pyrazoles that did not show neuroleptic activity in the locomotor and catalepsy tests. We propose to explore this scaffold for cannabinoid ligands. In this context, we report here the contribution of a pyrazole ring in replacement of the cannabinol phenyl group to cannabinoid activity.

Chemistry

7-Alkyl-1,4-dihydro-4,4-dimethylchromenopyrazol-9-oles 7-15 were prepared from the corresponding resorcinol following Scheme 1. Resorcinol 5-(1′,1′-dimethylheptyl)-1,3-dihydroxybenzene (2) was previously synthesized by demethylation^[30] of 5-(1′,1′-dimethylheptyl)-1,3-dimethoxybenzene. The appropriate starting resorcinol reacted with 3,3-dimethylacrylic acid in methanesulfonic acid in presence of phosphorus pentoxide to form the 7-alkyl-5-hydroxy-2,2-dimethylchroman-4-ones 3 and 4 in microwave heating conditions and using the reagents described by Lim et al.^[31] For the α–formylation of 3-4, an alternative procedure to the overnight heating proposed by Press³² yielded the corresponding (Z)-7-alkyl-5-hydroxy-3-(hydroxymethylen)-2,2-dimetylchroman-4-ones 5 and 6 in 20 min using microwave irradiation. Then condensation of the β–ketoaldehydes 5 and 6 with the appropriate hydrazine gave the corresponding 7-alkyl-1(2),4-dihydro-4,4-dimethylchromeno[4,3-c]pyrazol-9-oles 7-15. From methyl- and ethylhydrazines, the two N¹- and N²-substituted pyrazole isomers (8a,b; 9a,b; 13a,b) were isolated with an approximate relative from 8:2 to 6:4 ratio (N¹:N²). However, when the β–ketoaldehyde 5 or 6 reacted with arylhydrazine only one isomer could be isolated corresponding to the N¹-aryl chromenopyrazole (10a, 14a and 15a). The fact that alkylhydrazines give a mixture of N¹- and N²-substituted 7-alkyl-1(2),4-dihydro-4,4-dimethylchromeno[4,3-c]pyrazol-9-oles can be explained by the reaction of N′-hydrazine but also of N-hydrazine with the aldehyde group of 5 and 6 giving upon cyclization compounds 8a,b; 9a,b and 13a,b. The nucleophilicity of N′-hydrazine is much lesser than that of N-hydrazine in the case of N′-aryl hydrazines and this leads to a single isomer as in the case of 10a, 14a and 15a.

Scheme 1.

Reagents and conditions: (i) 3,3-dimethylacrylic acid, CH₃SO₃H, P₂O₅, 70 °C, M.W., 10 min; (ii) NaH, THF, M.W., 46°C, 20 min then ethyl formate, THF, M.W., 46°C, 20 min; (iii) H₂N-NHR₂, EtOH, 16 h at room temperature or 10 min under M.W.

Biological results

Competitive binding studies

The compounds reported here were evaluated in vitro for their ability to displace [³H]-CP55,940 from human cannabinoid CB₁ and CB₂ receptors transfected into HEK293 EBNA cells. They were first subjected to a preliminary screening at a concentration of 40 μM. A complete dose-response curve was carried out for compounds that displaced the radioligand by more than 50% in the preliminary screenning. Table 1 lists the experimental binding affinities (K_i) from the respective displacement curves for hCB₁ and hCB₂ receptors.

Table 1.

Binding Affinity K_i Values of Chromenopyrazole Derivatives 7-15 at hCB₁ and hCB₂ Cannabinoid Receptors.^[a]

compd	R1	R2	hCB1 Ki (nM)	hCB2 Ki (nM)	CB1/CB2c
7	pentyl	H	4100±800	2010±500	0.5
8a	pentyl	1-methyl	4700±1200	3460±1000	0.7
8b	pentyl	2-methyl	22100±1410	>40000	-
9a	pentyl	1-ethyl	9610	>40000	-
9b	pentyl	2-ethyl	>40000	4450±1015	-
10a	pentyl	1-(3,4-dichlorophenyl)	607±151	>40000	-
11	1,1-dimethylheptyl	H	28.5±33.6	>40000	>1000
12b	1,1-dimethylheptyl	2-methyl	14.2±4.2	>40000	>1000
13a	1,1-dimethylheptyl	1-ethyl	4.5±0.8	>40000	>1000
13b	1,1-dimethylheptyl	2-ethyl	18.6±4.1	>40000	>1000
14a	1,1-dimethylheptyl	1-(3,4-dichlorophenyl)	514±355	270	0.5
15a	1,1-dimethylheptyl	1-(2,4-dichlorophenyl)	5.2±6.0	>40000	>1000
SR141716			7.3±0.9	NDb
WIN55,212-2			45.6±8.6	3.7±0.2

^[a]K_i values obtained from competition curves using as [3H]-CP55940 radioligand for hCB₁ and hCB₂ cannabinoid receptors. The values are expressed as mean ± SEM of at least three experiments.

^bND:not determined.

^cSelectivity ratio CB₁ versus CB₂.

The first series to be examined were the 2,4-dihydrochromen[4,3-c]pyrazol-9-oles and 1,4-dihydrochromen[4,3-c]pyrazol-9-oles bearing a n-pentyl side chain (7-10a). Excluding 10a that binds weakly to the CB₁ cannabinoid receptor, the binding data for 7, 8a, 8b, 9a and 9b (Table 1) clearly show that they failed to bind to either CB₁ or CB₂ receptors. As reported in the literature,[³³] the introduction of C1′-alkyl substituent to Δ⁹-THC, Δ⁸-THC and cannabinol derivatives leads to enhancement in affinity for both receptors, CB₁ and CB₂. Interestingly, the 1,1-dimethylheptyl chromenopyrazole derivatives 11, 12b, 13a, 13b, and 15a showed significant to high affinity for CB₁ (CB₁: K_i = 4.5-28.5 nM) while they did not bind to the CB₂ receptor at all (CB₂: K_i > 40000 nM). So far such CB₁ receptor selectivity has never been observed among the cannabinoid ligands with a classical cannabinoid structure. Thus, chromenopyrazoles 11, 12b, 13a, 13b, and 15a showed an optimal CB₁ selectivity compared to CB₁/CB₂ binding data reported for Δ⁹-THC, Δ⁸-THC and cannabinol derivatives.

With respect to the substitution modification on the pyrazole ring, the K_i value remained substantially comparable in the 1,1-dimethylheptyl analogues, except for the N¹-3,4-dichlorophenyl moiety which compound (14a) exhibited reduced affinity for both receptors, CB₁ and CB₂ , with a loss of CB₁ selectivity.

Isolated tissues assays

The functional activity of 11, 13a and 13b was tested on mouse vas deferens, a tissue commonly used to study and characterize cannabinoid effects, where CB₁ cannabinoid receptors are expressed, as previously described.^[34]

Compounds 11, 13a and 13b inhibited electrically evoked contractile response of this tissue. In agreement with its high affinity for the CB₁ type, 13a exhibited the highest effectiveness (Figure 2). These three ligands were less effective than WIN55212-2 (WIN); however, their effect was comparable to that of arachidonyl-2-chloroethylamide (ACEA), a CB₁ selective agonist commonly used to characterize cannabinoid effects. Considering that compound 13a showed the most interesting profile as potential CB₁ agonist, the inhibition of its effect was tested adding to the organ bath the cannabinoid antagonist AM251, 10 min before the addition of the tested compound. As it is illustrated in Figure 3, the effect of 13a was clearly decreased by the selective CB₁ antagonist AM251. Moreover, the CB₂ antagonist AM630 was also tested but it did not block 13a effect (data not shown).

Figure 2.

Effect of WIN55, 212-2 (WIN), arachidonyl-2-chloroethylamide (ACEA) and compounds 11, 13a and 13b in mouse vas deferens. Lines show the means % ± S.E.M. (n = 6-8) of modification of the electrically induced contraction of the mouse vas deferens by addition of increasing concentrations of vehicle (Control), WIN, ACEA or the new compounds 11, 13a and 13b. The * represent the significant difference versus Control: *p<0.05, **p < 0.01, ***p< 0.001 (Two-ways ANOVA followed by Bonferroni′s post-hoc test).

Figure 3.

Lines show the mean % ± S.E.M. (n = 6) inhibition of the electrically induced contraction of the mouse vas deferens induced by addition of increasing concentrations of compound 13a in control tissues or in tissues incubated with AM251. The * represent the significant difference versus control tissues: *p<0.05 (Two-ways ANOVA followed by Bonferroni′s post-hoc test).

In vivo bioassays: Cannabinoid tetrad

Psychoactive cannabinoids dose-dependently modify spontaneous activity, antinociceptive response, rectal temperature, and catalepsy in mice.^[35] Effects of WIN and 13a on the cannabinoid tetrad were evaluated. Intraperitoneal (i.p.) administration of WIN (2.5 and 5 mg/kg) induced antinociceptive effect in the hot-plate test, hypothermia, catalepsy and reduction of the locomotor activity, whereas compound 13a (5 and 10 mg/kg) did not modify any of the signs of cannabinoid tetrad (Figure 4). This result suggests that 13a lacks of significant central effects.

Figure 4.

Effect of WIN 55,212-2 (WIN) (2.5 and 5 mg/kg) and compound 13a (5 and 10 mg/kg) in the mouse cannabinoid tetrad. Mice were tested for analgesia on a hot-plate (A), rectal temperature (B), catalepsy on a ring (C) and locomotor activity on the rota-rod test (D). ** p<0.01, *** p<0.001 vs vehicle (Veh). One-way ANOVA, n ≥10

In vivo bioassays: Orofacial pain model

Hyper tonic saline (HS) injection in the masseter of rats produced a paw shaking behavior. This nociceptive behavior was reduced by the intraperitoneal (i.p.) administration of compound 13a (1 and 3 mg/kg), suggesting that 13a is active in this pain model (Figure 5), acting at CB₁ receptors located outside the CNS.

Figure 5.

Antinociceptive effect of compound 13a (1 and 3 mg/kg), i.p.administered, 30 min before HS injection. Bars show the total number of shakes (mean ± SEM). * p<0.05, ** p<0.01 vs vehicle (Veh) (One-way ANOVA, n ≥6)

Molecular Modeling

The CB₁ cannabinoid receptor selectivity observed for the 1,1-dimethylheptyl chromenopyrazoles offers us the opportunity to explore structural features required for CB₁/CB₂ selectivity using molecular modeling.

Conformational analysis of the N-H-chromenopyrazole 11 and the two N-CH₃-chromenopyrazole regioisomers 13a and 13b were first performed to determine the global minimum energy conformation of each and other minimum energy conformations. With respect to the N-H-chromenopyrazole 11, two tautomers (11a and 11b in Figure 6) were considered. Even though the conformational analysis of 11a showed that it was more stable than 11b by 1.9 kcal/mol, both tautomers were taken into consideration for docking studies.

Figure 6.

Molecular structure of the two tautomers 11b and 11a.

Figure 7 illustrates the global minimum energy conformers of tautomers 11a and 11b, and regioisomers 13a and 13b.

Figure 7.

Minimum energy conformers of tautomers 11a, 11b, 13a, and 13b.

The global energy minima of 11a, 11b, 13a and 13b were then docked using a model of the activated state (R*) of the cannabinoid receptors CB₁ and CB₂.^[36-37] These models include the extracellular and intracellular loops, the N-terminus (truncated in CB1) and the C-terminus, including the intracellular helix portion of each receptor, termed Helix 8. CB₁ and CB₂ receptor docking studies were performed in the same binding site described for HU-210^[38] and for AM-841^[39-40] respectively.

Chromenopyrazole tautomer 11b/ CB₁R* Cannabinoid Receptor Docking Studies

The energy minimized 11b/CB₁R* complex is illustrated in Figure 8. Lys3.28(192) was used as the primary interaction site in CB₁ docking studies reported here.^[41] The phenolic oxygen of 11b is engaged in hydrogen bond with Lys3.28(192) as reported for HU-210/CB₁R* binding model [Hydrogen bond (N-O) distance = 2.78 Å and (N-H--O) angle = 150°]. The N²-pyrazole nitrogen is involved in hydrogen bond with Ser7.39(383) [Hydrogen bond (N-O) distance = 3.15 Å and (O-H--N) angle = 142°]. The ligand 11b exhibits the greatest pair-wise interaction energy with Lys3.28(192) (−11.88 kcal/mol), followed by Leu7.43(387) (−5.64 kcal/mol), Cys7.42(386) (−5.61 kcal/mol) and Asn7.45(389) (−5.43 kcal/mol). Coulombic energy dominates the overall pair-wise energy of interaction in the case of Lys3.28(192), while van der Waals energy is predominant for Leu7.43(387), Cys7.42(386) and Asn7.45(389). The interaction with Ser7.39(383) was found to be only −4.48 kcal/mol indicating a weak hydrogen bond with the N²-pyrazole nitrogen. The tautomer 11b also has significant interactions with Asp2.50(163) (−4.99 kcal/mol) with van der Waals and coulombic contributions, and with Val3.32(196) (−4.97 kcal/mol), whose interaction is dominated by van der Waals interactions. The energy difference between the initially docked 11b conformation and the final conformation in the energy-minimized complex was found to be 6.69 kcal/mol at the Hartree-Fock (HF) 6-31G* level. The overall interaction energy for 11b at CB₁ was found to be −56.52 kcal/mol (see Table S1 in Supporting Information).

Figure 8.

Binding site of 11b (left, in pink) and 11a (right, in pink) in the CB₁R* model. The amino acid residues interacting with the ligand are shown in grey. Yellow dashed lines indicate hydrogen bonding interactions.

Chromenopyrazole tautomer 11a/ CB₁R* Cannabinoid Receptor Docking Studies

The docking of the tautomer 11a in CB₁R* receptor model revealed a similar occupation of the binding site with similar hydrogen bonds involving Lys 3.28(192) and Ser7.39(383) as shown in Figure 8. As for 11b/CB₁R* complex, 11a has the greatest pair-wise interaction energy with Lys3.28(192) (−11.02 kcal/mol), and significant interactions with Cys7.42(386) (−6.22 kcal/mol), Asp2.50(163) (−6.03 kcal/mol), Leu7.43(387) (−5.20 kcal/mol), Asn7.45(389) (−5.13 kcal/mol) and Val3.32(196) (−4.40 kcal/mol). However, the interaction energy with Ser7.39(383) (−10.88 kcal/mol) in the 11a/CB₁R* complex was found much stronger than for 11b. The energy difference between the initially docked 11a conformation and the final conformation in the energy-minimized complex was found to be 6.51 kcal/mol at the HF 6-31G* level. The overall interaction energy for 11a at CB₁ was found to be −63.92 kcal/mol (see Table S1 in Supporting Information).

Taken together, the energies of interaction of 11a and 11b at CB₁ suggest that 11a is the preferred tautomeric form for 11 binding at CB₁.

Chromenopyrazole 13b/ CB₁R* Cannabinoid Receptor Docking Studies

In the energy minimized 13b/CB₁R* complex as illustrated in Figure 9, 13b forms two hydrogen bonds with Lys 3.28(192). The first involved Lys 3.28(192) as hydrogen donor to the phenolic oxygen of 13b [Hydrogen bond (N-O) distance = 2.75 Å and (N-H--O) angle = 152°]). The second interaction involved Lys 3.28(192) hydrogen bonding with the pyrazole N² nitrogen [Hydrogen bond (N-N) distance = 3.10 Å and (N-H--N) angle = 132°]. This pyrazole N² nitrogen also forms a hydrogen bond with Ser7.39(383) [Hydrogen bond (N-O) distance = 3.15 Å and (O-H--N) angle =134°]. However, this interaction is weaker than in the CB₁R* complex with 11 and 13a. The ligand 13b has its greatest pair-wise interaction energy with Lys3.28(192) (−14.34 kcal/mol, mainly coulombic energy), followed by Asn7.45(389) (−5.19 kcal/mol), Leu7.43(387) (−5.13 kcal/mol) and Cys7.42(386) (−4.76 kcal/mol) which are predominantly van der Waals energy. The 13b/CB₁R* complex also has significant interactions with Val3.32(196) (−4.72 kcal/mol) and Asp2.50(163) (−4.70 kcal/mol). The energy difference between the initially docked 13b conformation and the final conformation in the energy-minimized complex was found to be 5.13 kcal/mol at the HF 6-31G* level. The overall interaction energy for 13b at CB₁ was found to be −59.92 kcal/mol (see Table S2 in Supporting Information).

Figure 9.

Binding site of 13a (left, in pink) and 13b (right, in pink) in the CB₁R* model. The amino acid residues interacting with the ligand are shown in grey. Yellow dashed lines indicate hydrogen bonding interactions.

Chromenopyrazole 13a/ CB₁R* Cannabinoid Receptor Docking Studies

The main interactions of the energy minimized 13a/CB₁R* complex are shown on Figure 9. As observed for 11a, 11b and 13b, Lys 3.28(192) forms hydrogen bond with the phenolic oxygen of 13a [Hydrogen bond (N-O) distance = 2.70 Å and (N-H--O) angle = 166°]. The N² of the pyrazole of 13a hydrogen bonds with Ser7.39(383), as a hydrogen bond acceptor [ Hydrogen bond (N-O) distance = 2.80 Å and (O-H--N) angle = 160°]. It is interesting to note that an additional hydrogen bond between the pyran ring oxygen and Cys7.42(386) was revealed in the 13a/CB₁R* complex [Hydrogen bond (S-O) distance = 2.91 Å and (S-H-O) angle = 172°]. The significant increase in pair-wise interaction energy with Cys7.42(386) (−7.01 kcal/mol) compared to the other complexes presented here (11a,11b and 13b) is a consequence of this additional hydrogen bond. The ligand 13a exhibits greatest interaction energy with Lys3.28(192) (−13.83 kcal/mol), followed by the hydrogen bond interaction with Ser7.39(383) (−7.47 kcal/mol) and Phe2.57(170) (−5.30 kcal/mol). The Phe2.57(170) interaction has significant van der Waals and coulombic contributions and seems to have arisen from the interaction of the phenolic ring hydrogens with the aromatic ring of the phenylalanine. The 13a/CB₁R* complex also has significant interactions with Leu7.43(387) (−7.43 kcal/mol), Asp2.50(163) (−6.22 kcal/mol), Asn7.45(389) (−4.97 kcal/mol) and Val3.32(196) (−4.94 kcal/mol). The energy difference between the initially docked 13a conformation and the final conformation in the energy-minimized complex was found to be 7.39 kcal/mol at the HF 6-31G* level. The overall interaction energy for 13a, at CB₁ was found to be −69.16 kcal/mol (see Table S2 in Supporting Information).

Taken together with results for 13b above, docking studies indicate that the interaction of 13a at CB₁ is more energetically favorable than those of its positional isomer.

Chromenopyrazoles 11a/, 11b/, 13a/ and 13b/ CB₂R* Cannabinoid Receptor Docking Studies

Compounds11a, 11b, 13b and 13a were also docked in the previously reported advanced CB₂R* model^[37,40,42] at the AM-841 binding site^[40]. The CB₂ receptor model contains a salt bridge between D275 (Asp275) in the EC-3 loop and K3.28 Lys3.28(109). Docking studies of the chromenopyrazoles 11a, 11b, 13a and 13b revealed a steric clash between the pyrazole moiety of the structures and the ionic lock as illustrated in Figure 10.

Figure 10.

Binding site of 13b (left, in pink) and 13a (right, in pink) in the CB₂R* model. The D275-K3.28(109) ionic lock is shown in grey. Red circle indicate steric clash.

This result suggests that the presence of the pyrazole plays a key role in the selectivity of compounds 11 and 13 for CB₁cannabinoid receptor.

Discussion and Conclusion

Even though the first generation of classical cannabinoids showed potent activity in vivo, they lacked CB₁ or CB₂ selectivity and most of them are psychoactive. These CB₁ mediated side effects are due to the fact that CB₁ receptors are largely expressed in the CNS. Thus, the unwanted psychoactive effects of cannabinoid receptor agonists have limited their development as medicines. Since it is known that CB₁ receptors are also located peripherally,^[43] there is a growing interest in targeting cannabinoid receptors located outside the brain. For this reason, it is important to develop new nonpsychoactive cannabinoids that do not cross the blood-brain barrier but act on peripherally located cannabinoid receptors. The chromenopyrazoles presented here were designed in analogy to CBN, which is a CB₁/CB₂ cannabinoid ligand. The binding data show that the 1,1-dimethylheptyl side chain on this scaffold is necessary for high affinity. The ligand binding studies resulted in K_i values of 4.5 - 28.5 nM for the 1,1-dimethylheptyl analogues (11, 12b, 13a, 13b, and 15a) at hCB₁ receptors. Worth of note is that these 1,1-dimethylheptyl analogues did not have any affinity for hCB₂ receptors (K_i > 40000 nM). Unlike the major members of the classical cannabinoid family that lack CB₁ or CB₂ full selectivity, chromenopyrazoles structure have a determinant influence on CB₁ selectivity. These results suggest that such selectivity can be accounted by the presence of a pyrazole ring in the structure. However, the substituent on the pyrazole may play a major role in the binding to both receptors. Thus, we observed that 1-(3,4-dichlorophenyl) substituent (14a) disturbed significantly the CB₁ receptor selectivity with a loss of affinity for CB₁ and a moderate affinity for CB₂ receptor.

As assessed by modeling studies, the 1,1-dimethylheptylchromenopyrazoles 11a, 11b, 13a and 13b revealed a similar occupation of the HU-210/CP55,940 binding site in the CB₁ receptor model. The phenolic hydroxyl group of 11a, 11b, 13a and 13b is crucial for the interaction with CB₁R*, due to a hydrogen bond with Lys3.28(192). This residue has been shown to be critical for the binding of both classical and endocannabinoids.^[41] Furthermore, interaction with Ser7.39(383) was found as the main one for the pyrazole moiety, in particular for the 13a/CB₁R* complex. This residue (Ser7.39(383)) has been reported to be crucial for the binding of the CB₁ agonist, CP55,940.^[38]

While K3.28 in CB₁ has been reported to be critical for the binding of non-classical, classical and endocannabinoids,^[41] mutation of the equivalent residue in CB₂, K3.28(109) has no effect on the binding of any cannabinoid ligand.^[44] An important sequence divergence in the EC-3 loop of the CB₂ model (TTLSDQVKK) vs CB₁ model (GKMNKLIKT) has been reported.^[40] The CB₂R* model suggests a salt bridge formed between Asp275 and Lys3.28(109) that makes K3.28 unavailable for ligand interaction. In our CB₂R* dock, this salt bridge causes a steric clash with the pyrazole moiety of 11a, 11b, 13a or 13b due to the rigidity and planarity of the heterocycle. Therefore, these results suggest that the pyrazole moiety in these compounds is responsible for the CB₁ selectivity over CB₂.

Compounds 13a and 13b are positional isomers of one another differing in the placement of an ethyl substituent (N¹-, 13a; N²-, 13b). It is interesting to note that the interaction energies of 13a and 13b with CB₁ calculated here follow the same trend as their CB₁ affinities (13a, E_int = −69.16 kcal/mol, K_i = 4.5nM vs 13b, E_int = −59.92 kcal/mol, K_i = 18.6nM)_.

The chromenopyrazoles 11, 13a and 13b were tested in functional in vitro assays, where from all of them, 13a acted as the most effective cannabinoid agonist. Its effect was significantly and almost completely inhibited by the CB₁ antagonist AM251 but not by the CB₂ antagonist AM630.

Then, 13a was selected to carry on behavioral tests, in vivo, in which it did not induce modifications in any of the tested parameters on the mouse cannabinoid tetrad, discarding CNS-mediated effects. This lack of agonistic activity in the CNS suggests that they do not readily cross the blood-brain barrier.

To study other possibility of antinociception, compound 13a was tested in another pain model. We chose a model of orofacial pain in rat (nocive stimulation of the masseter by injection of hypertonic saline); it is known that, in this model, other drugs, such as opioids, can induce antinociception by peripheral mechanisms.^[45-46] In this test, 13a was able to reduce the nociceptive response. It is interesting to remark the different result obtained using this test with respect to the results recorded when the hot plate test was performed. From these data it could be suggested that the antinociception induced by compound 13a in the orofacial test may be mediated through peripheral mechanisms.

Experimental Section

Chemistry

Commercially available starting materials and reagents were used as supplied. Reactions conducted under anhydrous condition were performed under nitrogen atmosphere in solvents dried over CaCl₂ (THF) or Na/benzophenone (THF). Microwave-mediated syntheses were performed using 800W Ethos Synth microwave (Milestone Inc.) and CEM Biotage microwave. Column chromatography was performed using silica gel 60 (230-400 mesh). Analytical HPLC/MS analysis was performed on a Waters 2695 HPLC system equipped with a Photodiode Array 2996 coupled to Micromass ZQ 2000 mass spectrometer (ESI-MS), using a Waters X-bridge C18 column (3.5 μm, 2.1 × 100 mm) and 30 min gradient A: MeCN/0.08% formic acid, B: H₂O/0.05% formic acid visualizing at λ = 254 nm. ¹H and ¹³C NMR spectra were recorded on a Bruker 300 (300 and 75 MHz) at 25°C. Samples were prepared as solutions in deuterated solvent and referenced to internal nondeuterated solvent peak. Chemical shifts were expressed in ppm (δ) downfield of tetramethylsilane. Elemental analyses were determined by LECO CHNS-932. Melting points were determined on MP70 Reichert Jung Thermovar apparatus and are uncorrected. The purity of final compounds was determined by HPLC/MS analyses using above-mentioned instrument and by elemental analyses performed with LECO CHNS-932 analyzer.

5-(1,1-Dimethylheptyl)-1,3-dihydroxybenzene^[30] (2)

To a solution of 5-(1,1-dimethylheptyl)-1,3-dimethoxybenzene (0.45 g, 1.9 mmol) in dry CH₂Cl₂ was added boron tribromide (1M in CH₂Cl₂) (19 mL, 19 mmol) at −16 °C under nitrogen in the dark. The reaction mixture was allowed to warm to room temperature and was stirred for 20 h. Then MeOH was added carefully at 0 °C until pH = 7. The solvent was removed in vacuo and the crude was purified by chromatography on silica gel (EtOAc) to give the title compound as a white solid (0.33 g, 73%); mp: 88-91 °C (98 °C)^[30]; ¹H NMR (CDCl₃): δ = 6.35-6.45 (m, 2H, 4-H), 6.16 (m, 1H, 2-H), 4.65 (bs, 2H, OH), 1.45-1.50 (m, 2H, 2′-H), 1.22 (s, 6H, CH₃), 1.21-1.10 (m, 6H, 3′-H, 4′-H, 5′-H,), 1.03 (bs, 2H, 6′-H), 0.84 (t, J = 6.5 Hz, 3H, CH₃); ¹³C NMR (CDCl₃): δ = 163.2 (3-C), 154.0 (5-C), 111.2 (4-C), 101.3 (2-C)), 45.0 (2′-C), 35.3, 31.1, and 23.9 (3′-C, 4′-C, 5′-C) , 29.8 (8′-C), 25.1 (6′-C), 14.8 (7′-C); HPLC/MS: [A, 20% - 80%], Rt = 14.0 min, (90%); MS (ES⁺, m/z) 237 (100%) [M+H]⁺; Anal. calcd. for C₁₅H₂₄O₂: C 76.23, H 10.24, found: C 76.10, H 10.18.

5-Hydroxy-2,2-dimethyl-7-pentylchroman-4-one^[29] (3)

To a mixture of phosphorus pentoxide (0.23 g, 1.6 mmol) and methanesulfonic acid (4.6 mL, 72 mmol) was added olivetol (0.48 g, 2.7 mmol) and 3,3-dimethylacrylic acid (0.27 g, 2.7 mmol) under nitrogen at room temperature. Then the mixture was irradiated by microwave at 70 °C in a sealed reactor for 10 min. The mixture was poured onto water/ice then it was extracted with CH₂Cl₂. The organic layer was dried over MgSO₄. After removal of the solvent, the crude was purified by chromatography on silica gel (EtOAc) to give the title compound as a orange oil (0.34 g, 48%); ¹H NMR (CDCl₃): δ = 6.32 (d, J = 2.4 Hz, 1H, 8-H), 6.26 (d, J = 2.4 Hz, 1H, 6-H), 2.98 (t, J = 7.1 Hz, 2H, 1′-H), 2.66 (s, 2H, 3-H), 1.48-1.53 (m, 2H, 2′-H), 1.44-1.38 (m, 2H, 3′-H), 1.41 (s, 6H, OC(CH₃)₂), 1.30-1.40 (m, 2H, 4′-H), 0.87 (t, J = 6.9 Hz, 3H, 5′-H); ¹³C NMR (CDCl₃): δ = 191.2 (4-C), 161.9 (5-C), 159.7 (8a-C), 147.6 7-C), 110. 2 (8-C), 109.9 (6-C), 100.1 (4-C), 76.4 (2-C), 48.3 (3-C), 33.3 (1′-C), 30.1 (3′-C), 28.1 (2′-C), 24.8 (OC(CH₃)₂), 20.7 (4′-C), 12.2 (5′-C); HPLC/MS: [A, 20% - 80%], Rt = 17.6 min, (90%); MS (ES⁺, m/z) 263 (100%) [M+H]⁺; Anal. calcd. for C₁₆H₂₂O₃: C 73.25, H 8.45, found: C 72.98, H 8.71.

7-(1,1-Dimethylheptyl)-5-hydroxy-2,2-dimethylchroman-4-one (4)

Prepared from 2 (0.37 g, 1.6 mmol), phosphorus pentoxide (0.18 g, 1.3 mmol), methanesulphonic acid (2.67 mL, 54 mmol) and 3,3-dimethylacrylic acid (1.99 g, 19.9 mmol) following the procedure described for 3. Yellowish oil (0.34 g, 40%); ¹H NMR (CDCl₃): δ = 6.45 (d, J = 1.6 Hz, 1H, 8-H), 6.37 (d, J = 1.6 Hz, 1H, 6-H), 2.71 (s, 2H, 3-H), 1.52-1.59 (m, 2H, 2′-H), 1.46 (s, 6H, OC(CH₃)₂), 1.22 (s, 6H, C(CH₃)₂), 1.19 (bs, 6H, 3′-H, 4′-H, 5′-H), 1.05 (bs, 2H, 6′-H), 0.87 (m, 3H, 7′-H); ¹³C NMR (CDCl₃): δ = 197.9 (4-C), 163.0 (5-C), 161.7 (8a-C), 156.0 (7-C), 107.0 (8-C), 106.2 (6-C), 105.7 (4-C), 79.1 (2-C), 48.5 (3-C), 41.0 (2′-C), 39.2 (1′-C), 32.1, 27.1 and 23.0 ( 3′-C, 4′-C and 5′-.C), 30.3 (C(CH₃)₂), 28.8 (OC(CH₃)₂), 25.0 (6′-C), 14.5 (7′-C); HPLC/MS: [A, 80% - 100%], Rt = 4.9 min, (97%); MS (ES⁺, m/z) 318 (97%) [M+H]⁺; Anal. calcd. for C₂₀H₃₀O₃: C 75.43, H 9.50, found: C 75.52, H 9.64.

5-Hydroxy-3-(hydroxymethylen)-2,2-dimethyl-7-pentyl-chroman-4-one^[29] (5)

To a solution of the dihydrochroman-4-one 3 (0.17 g, 0.7 mmol) in dry THF was added dry 95% sodium hydride (0.14 g, 6 mmol) under nitrogen. The reaction mixture was irradiated by microwave at 46 °C in a sealed reactor for 20 min. Then ethyl formate (0.96 mL, 12 mmol) was added and the mixture was irradiated for an additional 20 min at 46 °C. The solvent was then evaporated under reduced pressure. Water was added to the residue. The aqueous solution was extracted with Et₂O then it was neutralized with 1M HCl and extracted with CH₂Cl₂. The combined organic layers were dried anhydride MgSO₄ and evaporated under reduced pressure. Column chromatography on silica gel (Hex/EtOAc, 1/1) afforded the titled compound as a yellowish oil (0.12 g, 60%); ¹H NMR (CDCl₃): δ = 13.48 (d, J = 11.1 Hz, 1H, CHOH), 11.36 (s, 1H, 5-OH), 7.34 (d, J = 11.1 Hz, 1H, 10-H), 6.32 (d, J = 1.3 Hz, 1H, 8-H), 6.21 (d, J = 1.3 Hz, 1H, 6-H), 2.49 (t, J = 7.6 Hz, 2H, 1′-H), 1.58-1.65 (m, 2H, 2′-H), 1.56 (bs, 6H, 9-H), 1.30-1.19 (m, 4H, 3′-H and 4′-H), 0.88 (t, J = 7.2 Hz, 3H, 5′-H); ¹³C NMR (CDCl₃): δ = 189.9 (4-C), 162.3 (5-C), 161.9 (CHOH), 159.3 (8a-C), 155.9 (7-C), 114.7 (3-C), 109.8 (8-C), 108.7 (6-C), 105.5 (4-C), 78.7 (2-C), 37.1 (1′-C)), 31.8 (3′-C), 30.4 (2′-C), 28.6 (9-C), 22.9 (4′-C), 14.4 (5′-C). MS (ES⁺, m/z) 291 (100%) [M+H]⁺; HPLC/MS: [A, 80% - 100%], Rt = 1.5 min, (99%); MS (ES⁺, m/z) 318 (97%) [M+H]⁺; Anal. calcd. for C₁₇H₂₂O₄: C 70.32, H 7.64, found: C 70.47, H 7.31.

7-(1,1-Dimethylheptyl)-5-hydroxy-3-(hydroxymethylen)-2,2-dimethylchroman-4-one (6)

Prepared from 4 (0.17 g, 0.5 mmol), sodium hydride (0.16 g, 6.7 mmol) and ethyl formate (1.44 mL, 2.7 mmol) following the procedure described for 5. Column chromatography on silica gel (Hex/EtOAc, 2/1). Yellowish oil (0.16 g, 76%); ¹H NMR (CDCl₃): δ = 13.48 (d, J = 11.6 Hz, 1H, CHOH), 11.28 (s, 1H, 5-OH), 7.34 (d, J = 11.6 Hz, 1H, CHOH), 6.46 (d, J = 1.6 Hz, 1H, 8-H), 6.35 (d, J = 1.6 Hz, 1H, 6-H), 1.58 (bs, 6H, OC(CH₃)₂), 1.22 (bs, 6H, C(CH₃)₂), 1.12-1.28 (m, 6H, 3′-H, 4′-H and 5′-H), 1.03 (m, 2H, 6′-H), 0.83 (t, J = 6.7 Hz, 3H, 7′-H); ¹³C NMR (CDCl₃): δ = 189.8 (4-C), 163.1 (5-C), 162.0 (CHOH), 161.9 (8a-C), 159.1 (7-C), 114.8 (3-C), 107.8 (6-C), 106.6 (8-C), 105.2 (4-C), 78.7 (2-C), 44.4 (2′-C), 39.2 (1′-C), 32.1, 30.3 and 25.0 (3′-C, 4′-H and 5′-H), 28.8 (C(CH₃)₂), 28.6 (OC(CH₃)₂), 23.0 (6′-C), 14.5 (7′-C); HPLC/MS: [A, 20% - 80%], Rt = 22.7 min, (100%); MS (ES⁺, m/z) 347 (100%) [M+H]⁺; Anal. calcd. for C₂₁H₃₀O₄: C 72.80, H 8.73, found: C 73.07, H 8.64.

1,4-Dihydro-4,4-dimethyl-7-pentylchromen[4,3-c]pyrazol-9-ol (7)

A solution of 5 (40 mg, 0.13 mmol) and anhydrous hydrazine (10 mg, 0.26 mmol) in EtOH (3 mL) was irradiated under microwave for 10 min. The solvent was evaporated and the crude residue was chromatographied on silica gel (Hex/EtOAc, 2/1) to obtain 7 as a white solid (20 mg, 59%); mp: 133-137 °C; ¹H NMR (CDCl₃): δ = 8.52 (bs, 1H, OH), 7.31 (s, 1H, 3-H), 6.44 (d, J = 1.5 Hz, 1H, 6-H), 6.37 (d, J = 1.5 Hz, 1H, 8-H), 2.53 (t, J = 7.1 Hz, 2H, 1′-H), 1.90-1.76 (m, 8H, OC(CH₃)₂ and 2′-H), 1.41-1.31 (m, 4H, 3′-H and 4′-H), 0.86 (t, J = 6.9 Hz, 3H, 5′-H); ¹³C NMR (CDCl₃): δ = 156.1 (9-C), 153.9 (5a-C), 146.3 (7-C), 142.7 (9b-C), 123.4 (3-C), 120.2 (3a-C), 109.4 (6-C), 108.6 (8-C), 101.7 (9a-C), 77.0 (4-C), 36.6 (1′-C), 31.9 (3′-C), 30.9 (2′-C), 29.9 (OC(CH₃)₂), 22.9 (4′-C), 14.4 (5′-C); HPLC/MS: [A, 10% - 100%], Rt = 5.9 min, (98%). MS (ES⁺, m/z) 287 (97%) [M+H]⁺; Anal. calcd. for C₁₇H₂₂N₂O₂: C 71.30, H 7.74, found: C 71.01, H 7.47.

1,4-Dihydro-1,4,4-trimethyl-7-pentylchromeno[4,3-c]pyrazol-9-ol^[29] (8a) and 2,4-dihydro-2,4,4-trimethyl-7-pentylchromeno[4,3-c]pyrazol-9-ol^[29] (8b)

To a solution of 5 (10 mg, 0.03 mmol) in EtOH was added methylhydrazine (7.00 μL, 0.13 mmol). The reaction mixture was stirred at room temperature for 16 h. The solvent was evaporated under reduced pressure. The crude oil was purified by column chromatography on silica gel (Hex/EtOAc, 2/1) to isolate the two isomers 8a and 8b. 8a was obtained as an orange oil (4 mg, 39%); ¹H NMR (CDCl₃): δ = 7.32 (s, 1H, 3-H), 6.50 (d, J = 1.3 Hz, 1H, 6-H), 6.27 (d, J = 1.3 Hz, 1H, 8-H), 4.11 (s, 3H, N-CH₃), 2.50 (t, J = 7.7 Hz, 2H, 1′-H), 1.55-1.44 (m, 8H, 2′-H and OC(CH₃)₂), 1.31-1.25 (m, 4H, 3′-H and 4′-H), 0.89 (m, 3H, 5′-H); ¹³C NMR (CDCl₃): δ = 154.5 (9-C), 149.9 (5a-C), 145.4 (7-C), 132.0 (9b-H), 130.9 (3-C)), 123.1 (3a-C), 111.4(8-C), 109.5 (6-C), 76.4 (4-C), 41.1 (N-CH₃), 38.7 (1′-C), 35.7 (3′-C), 31.4 (2′-C), 27.3 (OC(CH₃)₂), 22.5 (4′-C), 14.0 (5′-C); HPLC/MS: [A, 20% - 80%], Rt = 15.1 min, (95%); MS (ES⁺, m/z) 301 (100%) [M+H]⁺; Anal. calcd. for C₁₉H₂₆N₂O₂: C 72.58, H 8.33, found: C 72.81, H 8.26. 8b was obtained as a yellowish oil (6 mg, 58%); ¹H NMR (CDCl₃): δ = 8.31 (s, 1H, OH), 7.09 (s, 1H, 3-H), 6.42 (d, J = 1.1 Hz, 1H, 6-H), 6.34 (d, J = 1.1 Hz, 1H, 8-H), 3.90 (s, 3H, N-CH₃), 2.50 (t, J = 7.6 Hz, 2H, 1′-H), 1.59 (bs, 6H, OC(CH₃)₂), 1.50-1.60 (m, 2H, 2′-H), 1.30-1.18 (m, 4H, 3′-H and 4′-H), 0.95-0.84 (m, 3H, 5′-H); ¹³C NMR (CDCl₃): δ = 153.6 (9-C), 153.1 (5a-C), 145.3 (7-C), 142.6 (9b-C), 124.0 (3a-C), 120.3 (8-C), 108.7 (6-C), 108.5 (9a-C), 101.5(9a-C), 76.4(4-C), 38.8 (N-CH₃), 36.2 (1′-C), 31.5 and 22.5 (3′-C and 4′-C), 30.7 (2′-C), 29.6 (OC(CH₃)₂), 14.0 (5′-C); HPLC/MS: [A, 20% - 80%], Rt = 19.9 min, (96%); MS (ES⁺, m/z) 301 (100%) [M+H]⁺; Anal. calcd. for C₁₉H₂₆N₂O₂: C 72.58, H 8.33, found: C 72.36, H 8.49.

1-Ethyl-1,4-dihydro-4,4-dimethyl-7-pentylchromeno[4,3-c]pyrazol-9-ol (9a) and 2-Ethyl-2,4-dihydro-4,4-dimethyl-7-pentylchromeno[4,3-c]pyrazol-9-ol (9b)

A solution of 5 (40 mg, 0.15 mmol) and ethylhydrazine oxalate (20 mg, 0.15 mmol) in EtOH (3 mL) was irradiated under microwave for 10 min. The solvent was evaporated and the crude residue was chromatographied on silica gel (Hex/EtOAc, 1/1) to isolate the two isomers 9a and 9b. 9a was obtained as a yellowish oil (6.0 mg, 12%); ¹H NMR (CDCl₃): δ = 9.35 (bs, 1H, OH), 7.38 (s, 1H, 3-H), 6.46 (s, 1H, 6-H), 6.45 (s, 1H, 8-H), 4.48 (q, J = 7.1 Hz, 2H, NCH₂CH₃), 2.50 (t, J = 7.6 Hz, 2H, 1′-H), 1.66-1.52 (m, 8H, 2′-H and OC(CH₃)₂), 1.43 (t, J = 7.1 Hz, 3H, NCH₂CH₃), 1.28-1.19 (m, 4H, 3′-H and 4′-H), 0.88 (t, J = 6.7 Hz, 3H, 5′-H); ¹³C NMR (CDCl₃): δ = 155.5 (9-C), 151.8 (5a-C), 145.6 (7-C), 132.9 (9b-C), 131.7 (3-C), 122.9 (3a-C), 110.4 (6-C), 109.7 (8-C), 103.1 (9a-C), 76.04(4-C), 48.0 (NCH₂CH₃), 35.7 (1′-C), 31.4 and 22.5 (3′-C and 4′-C), 30.4 (2′-C), 27.2 (OC(CH₃)₂), 15.9 (NCH₂CH₃), 14.0 (5′-C); HPLC/MS: [A, 60% - 100%], Rt = 1.6 min, (100%); MS (ES⁺, m/z) 315 (100%) [M+H]⁺; Anal. calcd. for C₂₃H₃₄N₂O₂: C 74.55, H 9.25, found: C 74.63, H 9.19. 9b was obtained as a yellow solid (0.04 g, 61%); m.p. = 145-149 °C; ¹H NMR (CDCl₃): δ = 8.38 (s, 1H, OH), 7.12 (s, 1H, 3-H), 6.42 (d, J = 1.3 Hz, 1H, 6-H), 6.30 (d, J = 1.3 Hz, 1H, 8-H), 4.17 (q, J = 7.1 Hz, 2H, NCH₂CH₃), 2.50 (t, J = 7.1 Hz, 2H, 1′-H), 1.58-1.53 (m, 8H, 2′-H and OC(CH₃)₂), 1.48 (t, J = 7.1 Hz, 3H, NCH₂CH₃), 1.30-1.19 (m, 4H, 3′-H and 4′-H), 0.95-0.81 (m, 3, 5′-H); ¹³C NMR (CDCl₃): δ = 153.5 (9-C), 153.0 (5a-C), 145.2 (7-C)), 142.4 (9b-C), 122.3 (3-C), 119.8 (3a-C), 108.7 (8-C), 108.4 (6-C), 101.6 (9a-C), 76.7 (4-C), 47.0 (NCH₂CH₃), 36.2 (1′-C), 31.4 and 22.5 (3′-C and 4′-C), 30.7 (2′-C), 29.6 (OC(CH₃)₂), 15.4 (NCH₂CH₃), 14.0 (5′-C); HPLC/MS: [A, 60% - 100%], Rt = 2.3 min, (100%); MS (ES⁺, m/z) 315 (100%) [M+H]⁺; Anal. calcd. for C₂₃H₃₄N₂O₂: C 74.55, H 9.25, found: C 74.23, H 9.41.

1-(3,4-Dichlorophenyl)-1,4-dihydro-4,4-dimethyl-7-pentylchromeno[4,3-c]pyrazol-9-ol (10b)

Prepared from 5 (14 mg, 0.05 mmol) and 3,4-dichlorophenylhydrazine hydrochloride (10 mg, 0.05 mmol) following the procedure described for 7. Column chromatography on silica gel (Hex/EtOAc, 2/1) afforded 10b as an orange solid (7 mg, 31%); mp: 120-124°C; ¹H NMR (CDCl₃): δ = 7.57 (d, J = 2.4 Hz, 1H, 2-Hphenyl), 7.49 (s, 1H, 3-H), 7.44 (d, J = 8.7 Hz, 1H, 5-H_phenyl), 7.24 (dd, J = 2.4 Hz, J = 8.7 Hz, 1H, 6-H_phenyl), 6.52 (d, J = 1.2 Hz,1H, 6-H), 6.10 (d, J = 1.2 Hz, 1H, 8-H), 2.53-2.45 (m, 2H, 1′-H), 1.60-1.51 (m, 8H, 2′-H and OC(CH₃)₂), 1.30-1.22 (m, 4H, 3′-H and 4′-H), 0.93-0.88 (m, 3H, 5′-H); ¹³C NMR (CDCl₃): δ = 154.5 (9-C), 150.0 (5a-C), 146.3 (7-C), 142.0 (1-C_phenyl), 137.9 (3-C), 132.9 (3-C_phenyl), 132.2 (9b-C), 131.0 (4-C_phenyl), 129.9 (5-C_phenyl), 126.0 (2-C_phenyl), 124.9 (3a-C), 123.5 (6-C_phenyl), 111.2 (8-C), 109.7 (6-C), 102.3 (9a-C), 77.2 (4-C), 36.7 (1′-C), 31.5 and 22.5 (3′-C and 4′-C), 30.3 (2′-C), 27.2 (OC(CH₃)₂), 14.0 (5′-C); HPLC/MS: [A, 60% - 100%], Rt = 2.3 min, (100%); MS (ES⁺, m/z) 431 (100%) [M+H]⁺.

1,4-Dihydro-4,4-dimethyl-7-(1,1-dimethylheptyl)chromeno[4,3-c]pyrazol-9-ol (11)

Prepared from 6 (16 mg, 0.05 mmol) and anhydrous hydrazine (0.01 mL, 0.32 mmol) following the procedure described for 7. Column chromatography on silica gel (Hex/EtOAc, 2/1) afforded 11 as an orange oil (7 mg, 41%); ¹H NMR (CDCl₃): δ = 7.32 (bs, 1H, NH), 6.58 (d, J = 1.5 Hz, 1H, 6-H), 6.51 (d, J = 1.5 Hz, 8-H), 6.48 (s, 1H, 3-H), 1.63 (bs, 6H, OC(CH₃)₂), 1.58-1.52 (m, 2H, 2′-H), 1.25 (s, 6H, C(CH₃)₂), 1.18 (bs, 6H, 3′-H, 4′-H and 5′-H), 1.12-1.05 (m, 2H, 6′-H), 0.83 (t, J = 6.7 Hz, 3H, 7′-H); ¹³C NMR: δ = (CDCl₃) 153.7 (9-C), 153.5 (5a-C), 153.4 (7-C), 144.1 (9b-C), 129.1 (3-C), 123.4 (3a-C), 106.8 (8-C), 106.5 (6-C), 101.7 (9a-C), 77.0 (OC(CH₃)₂), 44.9 (2′-C), 38.4 (C(CH₃)₂), 32.2, 30.4, and 30.0 (3′-C, 4′-C and 5′-C), 29.3 (C(CH₃)₂), 25.0 (OC(CH₃)₂), 23.1 (6′-C), 14.5 (7′-C); HPLC/MS: [A, 80% - 100%], Rt = 3.1 min, (100%); MS (ES⁺, m/z) 343 (100%) [M+H]⁺; Anal. calcd. for C₂₁H₃₀N₂O₂: C 73.65, H 8.83, found: C 74.01, H, 8.59.

7-(1,1-Dimethylheptyl)-2,4-dihydro-2,4,4-trimethylchromeno[4,3-c]pyrazol-9-ol (12b)

Prepared from 6 (30 mg, 0.09 mmol) and methylhydrazine (0.02 mL, 0.34 mmol) following the procedure described for 7. Column chromatography on silica gel (Hex/EtOAc, 2/1) afforded 12b as a yellow solid (13 mg, 42%); mp > 300°C; ¹H NMR (CDCl₃): δ = 8.23 (s, 1H, OH), 7.09 (s, 1H, 3-H), 6.57 (d, J = 1.6 Hz, 1H, 6H), 6.48 (d, J = 1.6 Hz, 1H, 8-H), 3.90 (s, 3H, NCH₃), 1.63 (s, 6H, OC(CH₃)₂), 1.65-1.54 (m, 2H, 3′-H), 1.24 (s, 6H, C(CH₃)₂), 1.18-1.09 (m, 6H, 3′-H, 4′-H and 5′-H), 1.10-1.04 (m, 2H, 6′-H), 0.83 (t, J = 6.7 Hz, 3H, 7′-H); ¹³C NMR (CDCl₃): δ = 152.2 (9-C), 151.8 (5a-C), 151.6 (7-C), 141.5 (9b-C), 122.9 (3-C), 119.3 (3a-C), 105.5 (6-C), 105.3 (8-C), 100.2 (9a-C), 76.4 (OC(CH₃)₂), 43.5 (NCH₃), 38.9 (2′-C), 38.0 (1′-C), 31.8, 30.0 and 29.7 (3′-C, 4′-C and 5′-C), 28.9 (8′-C), 24.6 (OC(CH₃)₂), 22.6 (6′-C), 14.1 (7′-C); HPLC/MS: [A, 80% - 100%], Rt = 5.9 min, (100%); MS (ES⁺, m/z) 357 (100%) [M+H]⁺.

7-(1,1-Dimethylheptyl)-1-ethyl-1,4-dihydro-4,4-dimethylchromeno[4,3-c]pyrazol-9-ol (13a) and 7-(1,1-dimethylheptyl)-2-ethyl-2,4-dihydro-4,4-dimethylchromeno[4,3-c]pyrazol-9-ol (13b)

Prepared from 6 (35 mg, 0.1 mmol) and ethylhydrazine oxalate (15.00 mg, 0.1 mmol) following the procedure described for 9a/9b. Column chromatography on silica gel (Hex/EtOAc, 2/1) allowed isolating the two isomers 13a and 13b. 13a was obtained as a yellowish oil (5.00 mg, 18%); ¹H NMR (CDCl₃): δ = 8.77 (bs, 1H, OH), 7.37 (s, 1H, 3-H), 6.58 (d, J = 1.6 Hz, 2H, 6-H), 6.49 (d, J = 1.6 Hz, 1H, 8-H), 4.67 (q, J = 7.1 Hz, 2H, NCH₂CH₃), 1.57 (s, 6H, OC(CH₃)₂), 1.58-1.47 (m, 2H, NCH₂CH₃), 1.44 (t, J = 7.1 Hz, 3H, 2′-H), 1.23 (s, 6H, C(CH₃)₂), 1.17 (bs, 6H, 3′-H, 4′-H and 5′-H), 1.07 (bs, 2H, 6′-H), 0.84-0.78 (m, 3H, 7′-H); ¹³C NMR (CDCl₃): δ = 154.6 (9-C), 153.1 (5a-C), 151.67 (7-C), 133.1 (3-C), 132.2 (9b-C), 123.4 (3a-C), 108.8 (8-C), 108.0 (6-C), 103.2 (9a-C), 76.7 (OC(CH₃)₂), 48.4 (NCH₂CH₃), 44.8 (2′-C), 38.1 (C(CH₃)₂), 32.1, 30.4 and 25.01 (3′-C, 4′-C and 5-C), 29.0 (C(CH₃)₂), 27.7 (OC(CH₃)₂), 23.0 (6′-C), 16.3 (2′-C), 14.4 (7′-C); HPLC/MS: [A, 80% - 100%], Rt = 3.08 min, (99%); MS (ES⁺, m/z) 371 (100%) [M+H]⁺; Anal. calcd. for C₂₃H₃₄N₂O₂: C 74.55, H 9.25, found: C 74.63, H 9.19. 13b was obtained as a white solid (23 mg, 61%); m.p.: 160-164 °C; ¹H NMR (CDCl₃): δ = 8.32 (s, 1H, OH), 7.13 (s, 1H, 3-H), 6.57 (d, J = 1.5 Hz, 8-H), 6.48 (s, 1H, J = 1.5 Hz, 6-H), 4.16 (q, 2H, J = 7.2 Hz, NCH₂CH₃), 1.60 (s, 6H, OC(CH₃)₂), 1.56 (bs, 2H, 2′-H), 1.50 (t, 2H, J = 7.2 Hz, NCH₂CH₃), 1.24 (s, 6H, C(CH₃)₂), 1.17 (bs, 6H, 3′-H, 4′-H and 5′-H), 1.01-0.96 (m, 2H, 6′-H), 0.86-79 (m, 3H, 7′-H); ¹³C NMR (CDCl₃): δ = 153.6 (9-C), 153.2 (5a-C), 152.9 (7-C), 142.7 (9b-C), 122.8 (3-C), 120.3 (3a-C), 106.9 (6-C), 106.7 (8-C), 101.7 (9a-C), 76.4 (OC(CH₃)₂), 47.4 (NCH₂CH₃), 44.9 82(2′-C), 32.2, 30.4 and 30.1 (3′-C, 4′-C and 5′-C), 29.3 (C(CH₃)₂), 25.0 (OC(CH₃)₂), 23.1 (6′-C), 15.9 (NCH₂CH₃), 14.5 (7′-C); HPLC/MS: [A, 80% - 100%], Rt = 5.6 min, (98%); MS (ES⁺, m/z) 371 (100%) [M+H]⁺; Anal. calcd. for C₂₃H₃₄N₂O₂: C 74.55, H 9.25, found: C 74.23, H 9.41.

1-(3,4-Dichlorophenyl)-7-(1,1-dimethylheptyl)-1,4-dihydro-4,4-dimethylchromeno[4,3-c]pyrazol-9-ol (14a)

Prepared from 6 (17.0 mg, 0.05 mmol) and 3,4-dichlorophenylhydrazine hydrochloride (10.0 mg, 0.05 mmol) following the procedure described for 7. Column chromatography on silica gel (Hex/EtOAc, 2/1) afforded 14a as an orange solid (9 mg, 40%); mp: 124-126 °C; ¹H NMR (CDCl₃): δ = 7.65 (d, J = 2.3 Hz, 1H, 2-H_phenyl), 7.50 (s, 1H, 3-H), 7.43 (d, J = 8.5 Hz, 1H, 5-H_phenyl), 7.25 (dd, J = 2.3 Hz, J = 8.5 Hz, 1H, 6-H_phenyl), 6.66 (d, J = 1.6 Hz, 1H, 6-H), 6.24 (d, J = 1.6 Hz, 1H, 8-H), 1.68 (s, 6H, OC(CH₃)₂), 1.60-1.54 (m, 2H, 2′-H), 1.22 (s, 6H, C(CH₃)₂), 1.20-1.09 (m, 6H, 3′-H, 4′-H and 5′-H), 1.08-1.02 (m, 2H, 6′-H), 0.84 (t, J = 6.5 Hz, 3H, 7′-H); ¹³C NMR (CDCl₃): δ = 154.6 (9-C), 154.0 (5a-C), 150.0 (7-C), 142.3 (1-C_phenyl), 135.3 (3-C), 133.2 (3-C_phenyl), 132.2 (9b-C), 130.4 (4-C_phenyl), 126.5 (5-C_phenyl), 125.4 (2-C_phenyl), 123.9 (6-C_phenyl), 109.6 (6-C), 108.0 (8-C), 102.4 (9a-C), 76.2 (OC(CH₃)₂), 44.7 (2′-C), 38.3 (C(CH₃)₂), 32.1, 30.3 and 25.0 (3′-C, 4′-C and 5′-C), 28.9 (C(CH₃)₂), 27.7 (OC(CH₃)₂, 23.0 (6′-C), 14.5 (7′-C); HPLC/MS: [A, 80% - 100%], Rt = 5.6 min, (98%); MS (ES⁺, m/z) 487 (100%) [M+H]⁺.

1-(2,4-Dichlorophenyl)-7-(1,1-dimethylheptyl)-1,4-dihydro-4,4-dimethylchromeno[4,3-c]pyrazol-9-ol (15a)

Prepared from 6 (0.05 mg, 0.16 mmol) and 2,4-dichlorophenylhydrazine hydrochloride (0.13 g, 0.63 mmol) following the procedure described for 7. Column chromatography on silica gel (Hex/EtOAc, 2/1) afforded 15a as an orange oil (0.06 g, 75%); ¹H NMR (CDCl₃): δ = 7.43 (s, 1H, 3-H), 7.39 (d, J = 2.2 Hz, 1H, 3-H_phenyl), 7.27 (dd, J = 2.2 Hz, J = 8.7 Hz, 1H, 5-H_phenyl), 7.21 (d, J = 8.7 Hz, 1H, 6-H_phenyl), 6.56 (d, J = 1.6 Hz, 1H, 6-H), 6.12 (d, J = 1.6 Hz, 1H, 8-H), 1.64 (s, 6H, OC(CH₃)₂), 1.46-1.41 (m, 2H, 2′-H), 1.16-1.08 (m, 12H, 3′-H, 4′-H, 5′-H and C(CH₃)₂), 1.12-0.98 (m, 2H, 6′-H), 0.82 (t, J = 6.9 Hz, 3H, 7′-H); ¹³C NMR (CDCl₃): δ = 154.3 (9-C), 153.6 (5a-C), 151.3 (7-C), 140.3 (1-C_phenyl), 135.3 (2-C_phenyl), 135.0 (4-C_phenyl) , 134.4 (3-C), 133.0 (9b-C), 129.8 (3-C_phenyl), 129.6 (5-C_phenyl), 127.3 (6-C_phenyl), 123.4 (3a-C), 108.8 (6-C), 107.6 (8-C), 102.6 (9a-C), 77.6 (OC(CH₃)₂), 44.7 (2′-C), 39.1 (C(CH₃)₂), 32.1 , 30.3 and 24.9 (3′-C, 4′-C and 5′-C), 28.9 (C(CH₃)₂), 28.0 (OC(CH₃)₂), 23.0 (6′-C), 14.5 (7′-C); HPLC/MS: [A, 20% - 80%], Rt = 19.0 min, (99%); MS (ES⁺, m/z) 487 (100%) [M+H]⁺.

Biological studies

Binding Evaluation

Membranes from transfected cells with human CB₁ or CB₂ expressed cannabinoid receptors (RBHCB1M400UA and RBXCB2M400UA) were supplied by Perkin-Elmer Life and Analytical Sciences (Boston, MA). The CB₁ receptor membrane proteins concentration was 2.33 pmol/mg or 3.60 pmol/mg depending on the batch and the protein concentration was 8.0 mg/ml. The CB₂ receptor membrane protein concentration was 5.20 pmol/mg or 6.20 pmol/mg and the protein concentration was 4.0 mg/ml or 3.6 mg/ml depending on the batch. The commercial membranes were diluted (approximatively1:20) with the binding buffer (50 mM TrisCl, 5 mM MgCl₂.H₂O, 2.5 mM EDTA, 0.5 mg/mL BSA and pH = 7.4 for CB₁ binding; 50 mM TrisCl, 5 mM MgCl₂.H₂O, 2.5 mM EGTA, 1 mg/mL BSA and pH = 7.5 for CB₂ binding). The final membrane protein concentration was 0.4 mg/mL of incubation volume and 0.2 mg/mL of incubation volume for the CB₁ and the CB₂ receptor assays, respectively. The radioligand used was [³H]-CP55940 (PerkinElmer) at a concentration of membrane K_D × 0.8 nm, and the final volume was 200 μL for CB₁ binding and was 600 μL for CB₂ binding. 96-Well plates and the tubes necessary for the experiment were previously siliconized with Sigmacote (Sigma).

Membranes were resuspended in the corresponding buffer and were incubated with the radioligand and each compound (10⁻⁴-10⁻¹¹ M) for 90 min at 30 °C. Non-specific binding was determined with 10 μM WIN55212-2 and 100 % binding of the radioligand to the membrane was determined by its incubation with membrane without any compound. Filtration was performed by a Harvester® filtermate (Perkin-Elmer) with Filtermat A GF/C filters pretreated with polyethylenimine 0.05%. After filtering, the filter was washed nine times with binding buffer, dried and a melt-on scintillation sheet (Meltilex™ A, Perkin Elmer) was melted onto it. Then, radioactivity was quantified by a liquid scintillation spectrophotometer (Wallac MicroBeta Trilux, Perkin-Elmer). Competition binding data were analyzed by using GraphPad Prism program and K_i values are expressed as mean ± SD of at least three experiments performed in triplicate for each point.

Isolated tissues assays

Compounds 11, 13a and 13b were evaluated on the mouse vas deferens preparation. This is a nerve-smooth muscle preparation that serves as a highly sensitive and quantitative functional in vitro bioassay for cannabinoid receptor agonists. These ligands induce concentration-related decrease in the amplitude of electrically evoked contractions of the vas deferens by acting on naturally expressed prejunctional neuronal cannabinoid receptors to inhibit release of the contractile neurotransmitters, noradrenaline and ATP, that is provoked by the electrical stimulation.^47-48 For this study, male ICR mice weighing 25–30 g were used. Mouse vas deferens were isolated as described by Hughes.^[49] Tissues were suspended in a 10 mL organ bath containing 5 mL of Krebs solution (NaCl 118; KCl 4.75; CaCl₂ 2.54; KH₂PO₄ 1.19; MgSO₄ 1.2; NaHCO₃ 25; glucose 11 mM) that was continuously gassed with 95% O₂ and 5% CO₂. Tissues were kept under 0.5 g of resting tension at 37 °C and were electrically stimulated through two platinum ring electrodes. They were subjected to alternate periods of stimulation (trains of five rectangular pulses of 70 V, 15 Hz and 2 ms duration each were applied every minute) and rest (10 min). The isometric force was monitored by computer using a MacLab data recording and analysis system.

The effect of the synthetic cannabinoid agonists arachidonyl-2-chloroethylamide (ACEA), WIN 55,212-2 and that of the new compounds 11, 13a and 13b (10⁻⁷–1.8 × 10⁻⁵ M) was tested by constructing concentration–response curves for them in a step by step manner. Curves were carried out by the following protocol: ACEA, WIN 55,212-2 or the new compounds were added at a dose to the organ bath 50 min after the beginning of electrical stimulation and their effect on the electrically induced contractions was evaluated 10 min after their addition. Then, the electrical stimulation was stopped, Krebs solution was replaced and the following dose of the compounds was added. This protocol was repeated for every dose.

To test the involvement of the CB₁ and CB₂ receptors in the effect of 13a, it was tested in tissues incubated with the cannabinoid antagonists AM251 or AM630 (10⁻⁶ M) respectively. Concentration–response curves for the new compound were constructed in a step by step manner as follows: AM251 or AM630 was added to the organ bath 50 min after the beginning of electrical stimulation and 10 minutes later, a dose of 13a was added and its effect on the electrically induced contractions was tested 10 min later. Then, the electrical stimulation was stopped, Krebs solution was replaced and the cannabinoid receptor antagonist was added again to test the effect of the following concentration of the new compound. This protocol was repeated for every dose of 13a. Results have been expressed as % of inhibition, taking the mean amplitude of the last five contractions before the first addition of the agonist as 100%. Each tissue was employed to construct only one concentration–response curve.

In vivo behavioural studies

Behavioural testing of cannabinoids is performed in order to assess psychoactive drug potential, central side effects, as well as medicinal potential. The compound 13a was evaluated in tests of CNS activity, using the mouse cannabinoid tetrad. The potential antinociceptive effect was also evaluated in an orofacial pain model, induced by hypertonic saline (HS).

Animals

ICR male mice (25-30 g) and Wistar male rats (250-300 g), purchased at Harlan, S.A., were used in cannabinoid tetrad and orofacial pain model, respectively. Animals were supplied with food and water “ad libitum” and were housed in a temperature-controlled room at 23 ± 1°C under a standard 12/12-h light/dark cycle (08:00-20:00 h); they were housed in the test room for at least two days before experimentation. Throughout the experimental procedure, the international ethics standards for pain-inducing experiments in laboratory animals^[50] and the European Communities Council Directive of 24 November 1986 (86/609 EEC, Nov 24, 1986) were followed. All animal procedures were reviewed and approved by the Animal Care and Use Committee of Rey Juan Carlos University.

Drugs

WIN 55,212-2 and the compound 13a were dissolved in ethanol 1 mg:1 ml and subsequently in ethanol and Tween 80 (1:2), after which the ethanol was evaporated and saline solution added to reach the final concentration.^[47] All solutions were made fresh before each experiment.

Cannabinoid tetrad

The classical cannabinoid tetrad was performed to study SNC side effects; this test evaluates antinociception, hypothermia, catalepsy and locomotor activity in the same animal 20 minutes after cannabinoid administration.^[51] Separated groups of mice (n ≥ 10) were intraperitoneally treated with vehicle, WIN55,212-2 (2.5 and 5 mg/kg) and compound 13a (5 and 10 mg/kg). Tests were consecutively conducted with 5 minutes of interval between them.

Antinociception

The hot-plate test was carried out using a hot-plate at 55 °C as nociceptive stimulus. The latency time of licking of the front paw was taken as an index of nociception. The latency was measured before treatment (control latency) and after every treatment (latency after treatment). The cut-off time was 30 s and analgesia was quantified with the formula of the Maximum Possible Effect (M.P.E.), expressed as a percentage: % M.P.E. = (Latency after treatment - Control Latency) / (Cut-off time - Control latency) × 100.

Hypothermia

Core temperatures in mice were measured using a P6 thermometer and a lubricated rectal probe (CIBERTEC, Spain) inserted into the rectum at a constant depth of 1 cm. Data were recorded before and 30 min after treatment.

Catalepsy

Catalepsy was measured using a modified “ring test”, originally described by Pertwee.^[51] Mice were placed on a rubber coated metal ring (6 cm diameter) fixed horizontally at a height of 30 cm. The amount of time in which the mouse is immobile after placement on the ring is recorded for 5 min, and it is considered as an index of catalepsy.

Locomotor activity

Motor coordination was assessed using the rota-rod test (Cibertec, Madrid, Spain), in which mice were required to walk against the motion of an rotating drum with a constant speed of rotation of 10 revolutions/min over 5 min. The time (in s) taken to fall down was recorded as the latency. Animals were trained to the rota-rod test before the pharmacological assay. On the day of the drug test rota-rod latencies were measured immediately before the drug or vehicle was given and 30 min after the drug injection. In all experiments we used a 300-s cut-off time and to this time is given a value of 100 % of locomotor activity.

Orofacial pain model

The injection of 100 μl of HS (5% NaCl) in the masseter of lightly anesthetised rats produces ipsilateral hindpaw shaking behaviour that is accepted as an index of muscle nociception.^{[46, 52-53]} Separated groups of rats (n ≥ 10) were intraperitoneally (i.p.) treated with the vehicle, and compound 13a (1-3 mg/kg) and 30 min after HS was injected in the masseter. Shaking behaviour was quantified by counting the total number of shakes in a 2-min period after the intramuscular injection of HS. To count the number of shakes, the experiments were recorded on video and then played back in slow motion.

Molecular modeling

Amino acid numbering

The numbering scheme for Class A GPCRs suggested by Ballesteros and Weinstein^[54] was employed here. In this system, the most highly conserved residue in each TMH is assigned a locant of 0.50. This number is preceded by the TMH number and followed in parenthesis by the sequence number. All other residues in a TMH are numbered relative to this residue.

Conformational analysis of 11 and 13a/13b

Global minimum energy conformations of 11 and 13a/13b were determined with Spartan ′04 as follows: the structure of each molecule was built from the fragment library available in the program. Then, ab initio energy minimizations of each structure (HF 6-31G*) were performed. A conformational search was next performed using Spartan ′04 (Monte Carlo method) followed by a minimization of the energy of each conformer at the semiempirical PM3 level. For this search, selected bonds were allowed to rotate: C-O bond in the phenolic ring, the first two C-C bonds of the dimethylheptyl chain, and the N-C bond in the ethyl substituent of the pyrazole, in the case of 13a and 13b. Representative conformers according to their geometry were selected for ab initio energy minimization (HF 6-31G*), the global minimum energy conformer of each was used in docking studies.

Docking with CB₁R*

Binding site anchoring interactions within the receptor for each ligand were based on earlier published docking studies for HU210.^[38] Lys3.28(192) was used as the primary interaction site for the phenolic hydroxyl of each chromenopyrazole. Mutation of this residue in CB1 results in the loss of binding of classical, non-classical and endocannabinoids, suggesting that interaction with K3.28 is crucial for binding of this class of ligand.^[41] The energy of the ligand/CB₁R* TMH bundle complex was minimized using the OPLS2005 force field in Macromodel 9.1 (Schrödinger Inc., Portland, OR). An 8.0-Å extended nonbonded cutoff (updated every 10 steps), a 20.0- Å electrostatic cutoff, and a 4.0-Å hydrogen bond cutoff were used in each stage of the calculation. 7000 steps of conjugate gradient minimization in 500-step (6 times) increments followed by 1000-step (4 times) increments were employed in a distance-dependent dielectric. A 100 kJ/mol restraint was placed on all φ and ψ angles in TMH1-7 and Hx 8 and a 50 kJ/mol restraint was placed on Lys3.28(192)-phenolic OH hydrogen bond.

Energy expense assessments for docked ligands

To calculate the energy difference between the global minimum energy conformer of each compound and its final conformation after energy minimization of the ligand/receptor complex, rotatable bonds in the global minimum energy conformation were driven to their corresponding value in the final docked conformation and the single point energy of the resultant structure was calculated at the HF 6-31G* level using Jaguar (implemented in Maestro 8.5, Schrödinger).

Docking in CB₂R*

Global minimum conformations of each ligand were superimposed on HU210 in its complex with the CB₂R* model (unpublished results). Benzopyran atoms of 14 and 16a/b were selected for superimposition with benzopyran ones of HU210. The conformation of the 1,1-dimethylheptyl side chain in the HU210/ CB₂R* complex was used to overlay the 1,1-dimethylheptyl side chain position of 11 and 13a/b. After the superimposition, HU210 was removed. The pyrazole ring of the chromenopyrazoles ligands all had steric clashes with the Asp275 /Lys3.28(109) salt bridge. Changes in side chain dihedrals of these two residues were attempted to relieve the steric overlaps, however it was impossible to relieve this clash without disrupting this ionic lock.

Assessment of Pair-wise Interaction Energies

After defining the atoms of each ligand as one group (Group 1) and the atoms corresponding to a residue that lines the binding site in the final ligand/CB₁ R complex as another group (Group 2), Macromodel (version 8.6, Schrödinger, LLC, New York, NY) was used to output the pair-wise interaction energy (coulombic and Van der Waals) for a given pair of atoms. The pairs corresponding to Group 1 (ligand) and Group 2 (residue of interest) were then summed to yield the interaction energy between the ligand and that residue.