Diarylureas Containing 5-Membered Heterocycles as CB₁ Receptor Allosteric Modulators: Design, Synthesis, and Pharmacological Evaluation

Abstract

Allosteric modulators have attracted significant interest as an alternate strategy to modulate CB₁ receptor signaling for therapeutic benefits that may avoid the adverse effects associated with orthosteric ligands. Here we extended our previous structure-activity relationship studies on the diarylurea-based CB₁ negative allosteric modulators (NAMs) by introducing five-membered heterocycles to replace the 5-pyrrolidinylpyridinyl group in PSNCBAM-1 (1), one of the first generation CB₁ allosteric modulators. Many of these compounds had comparable potency to 1 in blocking the CB₁ agonist CP55,940 stimulated calcium mobilization and [³⁵S]GTP-γ-S binding. Similar to 1, most compounds showed positive cooperativity by increasing [³H]CP55,940 binding, consistent with the positive allosteric modulator (PAM)-antagonist mechanism. Interestingly, these compounds exhibited differences in ability to increase specific binding of [³H]CP55,940 and decrease binding of the antagonist [³H]SR141716. In saturation binding studies, only increases in [³H]CP55,940 B_max, but not K_d, were observed, suggesting that these compounds stabilize low affinity receptors into a high affinity state. Among the series, the 2-pyrrolyl analogue (13) exhibited greater potency than 1 in the [³⁵S]GTP-γ-S binding assay and significantly enhanced the maximum binding level in the [³H]CP5,5940 binding assay, indicating greater CB₁ receptor affinity and/or cooperativity.

Graphical Abstract

Introduction

The cannabinoid type-1 (CB₁) receptor is one of the most abundant G protein-coupled receptors (GPCRs) in the central nervous system (CNS) and has long been regarded as a promising target for the treatment of many conditions such as obesity, drug addiction, pain, inflammation, gastrointestinal diseases, multiple sclerosis, psychosis, schizophrenia, and osteoporosis.^{1, 2} The CB₁ receptor antagonist/inverse agonist, rimonabant (SR141716), was previously approved in Europe for the treatment of obesity.³ It also exhibited clinical benefits in smoking cessation,⁴ and it attenuated drug seeking behaviors in animal studies,⁵ supporting the notion that CB₁ antagonism may be useful as a treatment for drug addiction.^6–8 As rimonabant was subsequently withdrawn from the market due to its adverse effects, allosteric modulators have attracted attention as an alternative means to target the CB₁ signaling pathway for therapeutic benefits.^9–12

PSNCBAM-1 (1) was one of the first CB₁ receptor allosteric modulators, identified from a high throughput screen in a yeast-based assay in 2007.¹³ 1 shared similar in vitro pharmacological properties to another CB₁ allosteric modulator, Org27569 (2) such as enhancing [³H]CP55,490 binding levels (positive allosteric modulation of agonist affinity) but inhibiting agonist-induced responses in functional assays (negative allosteric modulation of agonist function).^{14, 15} Termed positive allosteric modulator-antagonists, or PAM-antagonists, these unique allosteric modulators increase the affinity of the agonist for the receptor but simultaneously decrease functional efficacy of the co-bound agonist. ¹⁶ As PAM-antagonists ‘seek and destroy’ agonist-activated receptors, they offer unique advantages in contrast to orthosteric and typical NAMs such as better reversal of ongoing persistent agonism and favorable target coverage in vivo.¹⁶ In addition, unlike 2, 1 alone did not show agonist activities in several assays tested such as ERK phosphorylation.^{9, 14, 17} In vivo, 1 demonstrated efficacy in reducing food intake and body weight.¹³ Recently, our group has demonstrated that 1 and an analogue with greater metabolic stability, RTICBM-74 (3), attenuated prime-induced reinstatement of extinguished cocaine-seeking behavior in rats.¹⁸ Furthermore, we have shown that 1 attenuated tetrahydrocannabinol-induced anti-nociception, and its analogue RTICBM-28 (4) shifted the potency of THC in drug discrimination.¹⁹ The promising results from these in vivo studies encourage further development on these diarylurea compounds for the potential use in the treatment of CB₁-related disorders (Fig. 1).

Figure 1.

Structures of representative CB₁ negative allosteric modulators

In previous studies, our group reported that the pyrrolidinyl ring of 1 could be replaced by other alkylamino groups (e.g. RTICBM-15, 5) or abolished completely without compromising the CB₁ activity, and the entire 5-pyrrolidinylpyridinyl group could be replaced by substituted phenyl and pyridinyl rings.^{18, 20} In agreement with our results, Khurana et al recently reported that analogues with a pyrimidinyl ring in place of the pyridinyl ring retained CB₁ modulatory activity (6 and 7, Fig. 1).²¹ In order to further explore the structure-activity relationship (SAR) around this scaffold and develop potent and selective CB₁ NAMs as potential pharmacotherapies for treating CNS disorders, we designed and synthesized a series of analogues in which the 5-pyrrolidinylpyridinyl ring was replaced by other heterocyclic rings to introduce structural diversity (Fig. 2). Specifically, we have focused our effort on investigating the effects of substitution of various 5-membered heterocyclic rings in CB₁ receptor allostery, such as furan, thiophene, pyrrole, thiazole, and imidazole. These novel compounds were examined in calcium mobilization, [³⁵S]GTP-γ-S binding, [³H]CP55,940 and [³H]SR141716 radioligand binding assays to investigate their pharmacological properties as CB₁ allosteric modulators.

Figure 2.

Structural modifications of the diarylurea template in this study

Results and Discussion

Chemistry.

Scheme 1 depicts the synthetic routes used to prepare target compounds 8-12 and 16-18. For compounds 8-9, commercially available 3-furanylboronic acid or 3-thienylboronic acid underwent Suzuki coupling with 3-bromoaniline to provide the intermediates 20a-b, which reacted with 4-chlorophenyl isocyanate to afford the target compounds 8 and 9. For target compounds 10-12 and 16-18, the heterocyclic boronic acids were not commercially available and therefore the Suzuki coupling was carried out between the 3-nitrobenzene boronic acid and the corresponding heterocyclic bromide to give intermediates 21a-f. The nitro group was then reduced to furnish the corresponding anilines 22a-f by Raney-nickel catalyzed transfer hydrogenation using hydrazine in ethanol. Subsequently, anilines 22a-f were coupled with 4-chlorophenyl isocyanates to give the desired products 10-12 and 16-18.^{18, 20}

Scheme 1.

Reagents and conditions (a) aryl boronic acid, Pd(PPh₃)₄, DME, NaHCO₃, reflux, 16 h, 10-32% (b) 4-chlorophenyl isocyanate, chloroform, 50 °C, 16 h, 57-95% (c) aryl bromide, Pd(PPh₃)₄, DME, NaHCO₃, reflux, 16 h, 37-92% (d) hydrazine hydrate, Raney Ni, ethanol, 50 °C, 2 h, 46%-quant. yield.

Similarly, Suzuki coupling between N-Boc-2-pyrrole boronic acid and 1-bromo-3-nitrobenzene provided the intermediate 23, which was then reduced to aniline 24 under transfer hydrogenation conditions. Intermediate 24 was then reacted with 4-chlorophenyl isocyanate to yield the final product 15. Removal of the Boc group in 24 by refluxing in 5% w/v aqueous KOH solution gave the aniline 25, which was then coupling with 4-chlorophenyl isocyanate to afford the target compound 13 (Scheme 2).²² Notably, attempts to remove the Boc group of 24 under acidic conditions using acids such as 4N HCl in dioxane led to product degradation.

Scheme 2.

Reagents and conditions (a) 1-Bromo-3-nitrobenzne, Pd(PPh₃)₄, DME, NaHCO₃, reflux, 16 h, 51% (b) hydrazine hydrate, Raney Ni, ethanol, 50 °C, 2 h, 65% (c) 4-chlorophenyl isocyanate, chloroform, 50 °C, 16 h, 81-84% (d) aq. 5% KOH, reflux, 4 h, 78%.

Palladium-catalyzed direct arylation of N-methylpyrrole with 3-iodoaniline gave the 3-(1-methyl-1H-pyrrol-2-yl)aniline 26,²³ which was then converted to the final product 14 upon treatment with 4-chlorophenyl isocyanate (Scheme 3).

Scheme 3.

Reagents and conditions (a) 3-iodoaniline, AcOK, Pd(OAc)₂, AcNMe₂, sealed tube, 150 °C, 20 h, 22% (b) 4-chlorophenyl isocyanate, chloroform, 50 °C, 16 h, 86-98% (c) (i) NaOMe, MeOH, rt, 5 h (ii) (MeO)₂CHCH₂NH₂, AcOH, 70 °C, 1 h (iii) aq. 6N HCl, MeOH, 70 °C, 3 h, 61% (d) hydrazine hydrate, Raney Ni, ethanol, 50 °C, 2 h, quant. yield.

For the preparation of compound 19, initial attempts of Suzuki coupling between 2-bromo-1H-imidazole and 3-nitrobenzene boronic acid failed to yield the desired product. Hence, the imidazole ring was constructed by reaction between 3-nitrobenzonitrile and aminoacetaldehyde dimethyl acetal to give the intermediate 27.²⁴ The target compound 19 was then obtained by the reduction of the nitro group leading to the aniline 28 and the subsequent reaction with 4-chlorophenyl isocyanate (Scheme 3).

Effects of compounds 8-19 in calcium mobilization assays.

All target compounds were first evaluated for their ability to inhibit the mobilization of intracellular calcium levels stimulated by CP55,940 in CHO cells overexpressing the promiscuous Gα16 protein (RD-HGA16 cells, Molecular Devices) and the human CB₁ or CB₂ receptor as previously described.²⁵ Table 1 lists the IC₅₀ values of the synthesized compounds against the EC₈₀ concentration of CP55,940 (100 nM) in the CB₁ calcium mobilization assay. 1 had an IC₅₀ value of 33 nM. When the 5-pyrrolidinylpyridinyl group was replaced with five-membered heterocyclic rings such as furan, thiophene, or methylthiophene, the CB₁ modulatory potency was retained (8-12, IC₅₀ = 36 - 67 nM). Replacement with pyrrole (13) or N-methylpyrrole (14) resulted in about 3-fold reduction in potencies (IC₅₀ = 165-169 nM). The presence of the bulky Boc group in 15 significantly decreased the potency (IC₅₀ > 3,000 nM). The three thiazolyl analogues had slightly decreased potency compared to 1 (16-18, IC₅₀ = 84 - 154 nM). Lastly, the potency was dampened by around 16-fold in the presence of the more polar group, imidazolyl (19, IC₅₀ = 529 nM). These results suggest that the CB₁ receptor can accommodate a diversity of heterocylic rings at this position.

Table 1.

Allosteric modulatory activities of compounds 8-19 in CB₁ and CB₂ calcium mobilization and CB₁ [³⁵S]GTP-γ-S binding assays.

		CB1		CB2
Compound	Ar	Calcium assay IC50 (nM)a	[35S]GTPγS binding assay IC50 (nM)b	Antagonist screen in calcium assay (%CP55,940 Emax)c
1		33 ± 8	115 (87, 155)	d
8		41 ± 9	441 (339, 574)	d
9		36 ± 1	164 (113, 239)	d
10		67 ± 6	573 (272, 1205)	d
11		40 ± 4	776 (324, 1858)	d
12		42 ± 6	372 (182, 759)	d
13		169 ± 34	40 (29, 55)	IC50 > 10 μM e
14		165 ± 14	665 (501, 883)	d
15		3444 ± 451	3236 (1480, 7070)	d
16		84 ± 6	251 (118, 530)	d
17		94 ± 9	331 (215, 509)	d
18		154 ± 14	1259 (562, 2818)	d
19		529 ± 98	343 (224, 524)	d

^aAgainst EC₈₀ (100 nM) of CP55,940 in stable human CB₁-CHO-RD-HGA16 cells. Values are the mean ± SEM of at least three independent experiments in duplicate.

^bTested in mouse cerebellar membranes against CP55,940 (100 nM). Values are expressed as mean (95% confidence interval) from at least three independent experiments in duplicate.

^cAgainst EC₈₀ (100 nM) of CP55,940 in stable human CB₂-CHO-RD-HGA16 cells. Compounds were tested at 10 μM final concentration in two independent experiments in duplicate.

^d< 50% inhibition.

^eCompound showed > 50% inhibition and was tested for potency.

None of the compounds showed significant CB₂ antagonist activity (< 50% inhibition of CP55,940 EC₈₀ concentration at 10 μM or IC₅₀ values >10 μM) (Table 1). All compounds were additionally tested for agonist activity at both CB₁ and CB₂ receptors. No significant CB₁ agonist activity (< 30% of CP55,940 E_max) or CB₂ agonist activity (< 30% of CP55,940 E_max) at concentrations of 10 μM was observed (Supporting Information).

Effects of diarylureas in [³⁵S]GTP-γ-S binding assay.

All compounds were then evaluated as NAMs at the CB₁ receptor in the [³⁵S]GTP-γ-S assay using the CB₁ agonist CP55,940 (Table 1). While 3-furanyl (8) showed similar potency in the calcium assay to 1, it was ~4 fold less potent in the GTP-γ-S assay. The 3-thiophenyl analogue (9, IC₅₀ = 164 nM) had comparable potency to 1 (IC₅₀ = 115 nM), and the 2-thiophenyl analogue (10) was slightly weaker (IC₅₀ = 573 nM). Introduction of a methyl group resulted in a 5-fold reduction in inhibitory potency in the 3-thiophenyl analogue (11), whereas the potency slightly increased in the 2-thiophenyl analogue ( 12 vs. 10). The lower potencies of both N-methylpyrrolyl (14) and N-Boc pyrrolyl (15) analogues were consistent with the results from the calcium assay. Similarly, the three thiazolyl analogues (16-18) shared the same trend in activities in both assays.

In general, the compounds were more potent in the calcium assay than in GTP-γ-S. However, the pyrrolyl analogue (13, IC₅₀ = 40 nM) appeared to be more potent than 1 in the GTP-γ-S binding assay even though it was weaker in the calcium assay. Similarly, the imidazolyl analogue (19) demonstrated better activity in the GTP-γ-S binding assay than the calcium assay. This difference in the potency ranking order between the two assays could result from the differences in affinity for human versus mouse CB₁ receptor and/or G protein subtype coupling involved in the response (e.g. overexpression of Gα₁₆ in the calcium assay).

Effects of diarylureas on [³H]CP55,940 radioligand binding.

Effects of representative allosteric modulators on equilibrium binding of 1 nM [³H]CP55,940 and 1 nM [³H]SR141716 are depicted in Figure 3. The curve-fit values of all target compounds are listed in Table 2.

Figure 3.

(A) Effects of representative diarylureas on the equilibrium binding levels of 1 nM of [³H]CP55,940, and (B) Effects of representative diarylureas on the equilibrium binding levels of 1 nM of [³H]SR141716.

Table 2.

Allosteric modulatory activities of diarylureas 1 and 3-14 in the CB₁ [³H]CP55,940 and [³H]SR141716 binding assays ^a

Compound	Ar	[3H]CP55,940 pEC50 ± S.E.	[3H] CP55,940 Curve-fit top ± S.E.	[3H]SR141716 pIC50 ± S.E.	%[3H]SR141716 Binding a,b ± S.E.
1		5.69 ± 0.08	256.0 ± 5.59	4.91 ± 0.11	20.42 ± 3.05
8		5.13 ± 0.18	174.1 ± 8.86	4.76 ± 0.14	34.87 ± 2.43
9		5.34 ± 0.01	264.6 ± 9.81	4.95 ± 0.15	20.44 ± 2.13
10		5.24 ± 0.14	199.8 ± 8.78	4.54 ± 0.21	45.54 ± 0.15
11		5.41 ± 0.08	226.0 ± 5.82	4.08 ± 0.83	73.19 ± 5.77
12		4.88 ± 0.18	233.3 ± 20.53	4.64 ± 0.74	86.90 ± 5.87
13		5.86 ± 0.09	458.3 ± 16.25	4.85 ± 0.18	26.60 ± 5.68
14		5.25 ± 0.19	164.5 ± 8.61	4.75 ± 0.20	41.41 ± 3.26
15		4.83 ± 0.27	15.5 ± 24.12	5.07 ± 0.11	17.20 ± 4.12
16		4.97 ± 0.17	224.2 ± 15.85	4.59 ± 0.17	44.34 ± 1.42
17		5.19 ± 0.17	181.2 ± 8.15	5.53 ± 0.11	50.58 ± 2.04
18		5.29 ± 0.27	134.2 ± 4.47	4.78 ± 0.17	37.68 ± 0.76
19		4.99 ± 0.12	355.7 ± 27.39	3.91 ± 0.79	34.83 ± 8.94

^aCompounds tested in presence of 1 nM [³H]CP55,940 or 1 nM [³H]SR141716.

^bpercent specific [³H]SR141716 binding at 32 μM of test compound.

Similar to previously reported CB₁ allosteric modulators such as 1, most analogues increased specific binding of [³H]CP55,940, demonstrating positive cooperativity with the agonist radioligand. At the highest tested concentration of 10 μM, the 3-thiophenyl analogue (9) had relatively similar maximal increases in specific [³H]CP55,940 binding as compared to 1, whereas the 2-thiophenyl analogue (10) only exhibited approximately half the increase in specific binding. The 3-furanyl analogue (8) exhibited a similar increase in specific [³H]CP55,940 binding as compared to the 2-thiophenyl analogue (10), suggesting that both the position and nature of heteroatom are important for imparting cooperativity. Introduction of a methyl at the 5 position of the 3-thiophenyl (11) or 2-thiophenyl analogue (12) resulted in a slight lower increase in [³H]CP55,940 binding level compared to 9. Within the pyrrolyl series, the unsubstituted 2-pyrrolyl analogue (13, maximum binding level: 458%) increased the [³H]CP55,90 binding level more than 2-fold compared to 1 (maximum binding level: 256%), whereas N-methylpyrrol-2-yl (14) only slightly enhanced the level of radioligand binding. Interestingly, the N-Boc-2-pyrrolyl (15) significantly decreased the binding of the radioligand.

The trend of cooperativity in agonist binding positively correlates with their potency in the GTP-γ-S binding assay (r(13) = 0.65, p<0.05). For instance, 13 was the most potent compound of the series in the GTP-γ-S assay (IC₅₀ = 40 nM), and it also increased the maximal [³H]CP55,940 binding to the largest extent (458%). In contrast, consistent with its low activity in both the calcium and GTP-γ-S assays, 15 displayed no increase in [³H]CP55,940 binding but instead decreased the maximal level.

Among the thiazolyl analogues (16-18), the 4-thiazolyl analogue (16) increased the radioligand binding level nearly as much as 1. However, the positive effect was diminished in the case of the 5-thiazolyl (17), and markedly reduced in case of the 2-thiazolyl (18). The correlation between [³H]CP55,940 binding maximum level and potency in the GTP-γ-S binding assays seen in the pyrrole series was also observed in the thiazole series. On the other hand, the imidazolyl analogue (19, maximum binding level 355%) demonstrated stronger binding cooperativity with [³H]CP55,940 yet was less potent in the GTP-γ-S binding assay compared to 1, which may reflect a lower affinity (e.g. pEC₅₀ value appeared right shifted in comparison) despite greater cooperativity.

Effects of diarylureas on [³H]SR141716 radioligand binding.

As opposed to the [³H]CP55,940 binding studies where most compounds increase radioligand binding, most of the analogues reduced [³H]SR141716 binding in a similar fashion to previously reported CB₁ allosteric modulators 1 and 2.^{13, 26} These results are consistent with the PAM-antagonist mechanism.¹⁶ Cannabinoid agonists exhibit two distinct affinities for the CB₁ receptor,²⁷ which reflect two receptor states, one coupled to the cognate G protein (active) and the other uncoupled (inactive). Some reports suggest that allosteric regulation by G proteins does not affect the affinity of the CB₁ antagonist/inverse agonist SR141716 for the receptor, i.e. it binds with equal affinity to either receptor state,^{28, 29} and that SR141716 may sequester G proteins in a nonsignaling complex; however, Bouaboula et al. reported an increase in SR141716 affinity for the CB₁ receptor when it is uncoupled from G protein by coincubation with GTP-γ-S,³⁰ an effect that has been reported for other inverse agonists of G protein coupled receptors.³¹ Therefore, an increase in affinity of the modulator for the agonist high-affinity receptor state (as reflected by the ability of the modulator to increase binding of [³H]CP55,940) would also result in an increase in the ability of the modulator to reduce SR141716 binding. In other words, if SR141716 has a higher affinity for receptors inactive/uncoupled from G protein, then structural changes that increase the affinity of the modulator for the agonist high-affinity conformation will result in a concomitant increase in the modulator’s ability to reduce [³H]SR141716 binding and vice versa, since the distribution of receptor states would be shifted towards those with which SR141716 has higher affinity.

Interestingly, the 5-methylthiophenyl analogues (11 and 12) only exhibited a slight reduction in [³H]SR141716 binding that was significantly less than observed in the rest of the series. However, as discussed above, both 11 and 12 displayed a similar ability relative to 1 in increasing [³H]CP55,940 binding. It has suggested that the binding pocket for the structurally similar CB₁ allosteric modulator Org27569 overlaps with that of SR141716.³² Thus, it is possible that the binding site for these allosteric modulators also partially overlaps with that of SR141716, with the methyl group interfering with the ability of the modulator to occupy SR141716’s binding site. Alternatively, it could be that 11 and 12 have lower affinity for receptors uncoupled from their cognate G protein, for which SR141716 may have higher affinity.

Notably, calculated pIC₅₀ values from [³⁵S]GTP-γ-S binding studies were positively correlated with pEC₅₀ values from [³H]CP55,940 equilibrium binding (r(13) = 0.65, p<0.05) (Table 2) but not pIC₅₀ values from displacement binding of [³H]SR141716 (r(13) = 0.06, p=0.82). This suggests that affinity for the agonist-bound conformation, but not the antagonist, predicts the allosteric modulator’s inhibitory potency, which is consistent with positive cooperativity between the modulator and the probe agonist (PAM-antagonist).¹⁹

Effects of diarylureas on [³H]CP55,940 saturation binding.

Saturation binding experiments were conducted in the absence and presence (10 μM) of a subset of synthesized modulators to determine their effects on [³H]CP55,940’s B_max and K_d values. Under the vehicle condition, [³H]CP55,940 exhibited a B_max of 1.56 ± 0.09 pmol/mg, which was increased by co-incubation with each allosteric modulator (Table 3) as previously reported for 1.¹⁴ Compounds 11 and 17 exhibited comparable B_max and K_d values to 1. Intriguingly, 13, which exhibited a larger increase in specific [³H]CP55,940 binding than PSNCBAM-1 (Table 2), also produced a larger B_max in saturation experiments (Table 3). Additionally, 8, which exhibited a smaller increase in [³H]CP55,940 binding as compared to 1 (Table 2), also resulted in a lower B_max for [³H]CP55,940 in saturation binding (Table 3). In contrast, K_d values for [³H]CP55,940 were not significantly affected by the presence of the allosteric modulators. Considering the low range of concentrations for [³H]CP55,940, it is likely the B_max values observed in these saturation binding plots reflect the population of high affinity receptors that are coupled to the G proteins, as CP55,940 exhibits a K_i in the 100 nM range for the uncoupled low affinity receptor.³³ Thus, the increase in B_max following co-incubation with allosteric modulators reflects an increase in affinity of the non-coupled receptors resulting in a B_max shift rather than a K_d shift. These results combined with the elevated B_max values suggest that these compounds are not modulating high-affinity binding but rather altering the equilibrium of low- and high-affinity receptors. The enhanced B_max values and unchanged K_d values of these diarylureas correspond to the same trend observed for Org27569 previously.¹⁴ Therefore, differences in maximal increases observed in specific [³H]CP55,940 binding [Table 2; Figure 3A] between allosteric modulators likely reflect the proportion of low affinity receptors shifted into a high affinity state.

Table 3.

Allosteric effects of diarylureas on [³H]CP55,940 saturation binding at 10 μM of test compound

Compound	[3H]CP55,940 Bmax (pmol/mg) ± S.E.	Bmax (pmol/mg) 95% CI	[3H]CP55,940 Kd (nM) ± S.E.	Kd (nM) 95% CI
Vehicle	1.56 ± 0.09	1.38 to 1.74	1.92 ± 0.30	1.31 to 2.52
1	4.53 ± 0.29	3.95 to 5.11	1.93 ± 0.33	1.27 to 2.60
8	3.22 ± 0.19	2.82 to 3.62	2.85 ± 0.41	1.98 to 3.71
11	3.64 ± 0.17	3.28 to 4.00	2.63 ± 0.31	1.98 to 3.29
13	5.87 ± 0.15	5.55 to 6.19	1.41 ± 0.10	1.20 to 1.63
17	3.57 ± 0.25	3.05 to 4.10	3.10 ± 0.51	2.02 to 4.17

Conclusions

In our previous study, we demonstrated that the 5-pyrrolidinylpyridinyl of 1 could be replaced by substituted phenyl and pyridinyl groups and modifications at this position led to CB₁ allosteric modulators with better metabolic stability and greater potency in vivo.¹⁸ This study further elucidates the structural requirements at the place of 5-pyrrolidinylpyridinyl group of 1. We showed that various five-membered heterocyclic rings could be accommodated at this position, resulting in the generation of interesting pharmacological activities at the CB₁ receptor. Replacement by furanyl, thiophenyl, or 5-methylthiophenyl groups (8-12) maintained similar allosteric modulating potency in the calcium mobilization assay against human CB₁ receptor. Notably, modification to 2-pyrrolyl group (13) significantly enhanced the potency in the GTP-γ-S assay, doubled the [³H]CP55,940 binding level in the competitive binding assay and markedly increased B_max in the [³H]CP55,940 saturation binding assay. These results indicate that 13 had greater affinity and/or cooperativity at the CB₁ receptor. The fact that K_d values in the [³H]CP55,940 saturation binding assay remained unchanged suggested that the these compounds modulated the CB₁ receptor by stabilizing agonist low-affinity receptors into an agonist high-affinity state. In the competitive [³H]SR141716 binding assay, while most synthesized diarylureas retained the ability to reduce SR141716 binding to CB₁ receptor as expected for PAM-antagonists, the 5-methylthiophenyl analogues (11 and 12) only slightly reduced [³H]SR141716 binding, possibly due to the partially overlapping binding pockets.³²

A number of CB₁ NAMs have been reported and their in vitro pharmacological properties have been thoroughly characterized to date, most of which appear to be PAM-antagonists. Since PAM-antagonists and agonists increase the other’s binding affinity to the receptor, this class of NAMs “seek out and destroy” agonist-bound receptor complexes, therefore providing the unique benefit of reversing ongoing persistent agonism and resulting in favorable in vivo target coverage.¹⁶ Despite these, the translation of in vitro to in vivo effects has not been well established with these CB₁ NAMs. For instance, the subtle attenuation of THC-induced anti-nociception by 1 and the small reduction in THC’s potency by 4 in drug discrimination assays were not seen with their structurally-related analogue 5.¹⁹ This may be attributed to pharmacokinetic profiles of these allosteric modulators or slower association kinetics commonly observed of allosteric modulators compared to the orthosteric ligands,³⁴ leading to less pronounced in vivo effects. Indeed, we have demonstrated that 3, a more metabolically stable analogue of 1, produced a more pronounced effect in attenuating reinstatement of cocaine self-administration.¹⁸ By replacing the 5-pyrrolidinylpyridinyl group of 1 with a diversity of five-membered heterocycles, this work illustrated that structural modification at this position may allow further fine-tuning of pharmacological properties to achieve desired therapeutic effects. Continued SAR studies and optimization effort by improving potency and pharmacokinetic properties and/or modifying receptor binding kinetics may lead to CB₁ allosteric modulators as improved in vivo probes and potential clinical therapeutics.

EXPERIMENTAL SECTION

Chemistry.

All solvents and chemicals were reagent grade. Unless otherwise mentioned, all reagents and solvents were purchased from commercial vendors and used as received. Flash column chromatography was carried out on a Teledyne ISCO CombiFlash Rf system using prepacked columns. Solvents used include hexanes, ethyl acetate (EtOAc), dichloromethane, and methanol. Purity and characterization of compounds were established by a combination of HPLC, TLC, mass spectrometry, and NMR analyses. ¹H and ¹³C NMR spectra were recorded on a Bruker Avance DPX-300 (300 MHz) spectrometer and were determined in CDCl₃, DMSO-d6, or CD₃OD with tetramethylsilane (TMS) (0.00 ppm) or solvent peaks as the internal reference. Chemical shifts are reported in ppm relative to the reference signal, and coupling constant (J) values are reported in hertz (Hz). Thin layer chromatography (TLC) was performed on EMD precoated silica gel 60 F254 plates, and spots were visualized with UV light or iodine staining. Nominal mass spectra were obtained using a Waters Alliance HT/Micromass ZQ system (ESI). High resolution mass spectra were obtained using Agilent 1290 Infinity UHPLC-6230 TOF system (ESI). All final compounds were greater than 95% pure as determined by HPLC on an Agilent 1100 system using an Agilent Zorbax SB-Phenyl, 2.1 mm × 150 mm, 5 μm column using a 15 minute gradient elution of 5-95% solvent B at 1 mL/min followed by 10 minutes at 95% solvent B (solvent A, water with 0.1% TFA; solvent B, acetonitrile with 0.1% TFA and 5% water; absorbance monitored at 220 and 280 nm).

General procedure A.

To a mixture of aryl bromide or aryl iodide (1 eq), boronic acid (1.1 eq) in dimethoxyethane (0.1 M) was added 1M aqueous NaHCO₃ solution (3 eq) followed by Pd(Ph₃)₄ (0.075 eq). The reaction mixture was refluxed overnight under nitrogen atmosphere. The reaction mixture was diluted with ethyl acetate, washed with a saturated NaHCO₃ solution and brine. The combined organic layers were dried over anhydrous MgSO₄ and filtered. The filtrate was concentrated in vacuo and the residue was purified by column chromatography (SiO₂, ethyl acetate/hexanes) to give the desired product.

3‐(Furan‐3‐yl)aniline (20a) was prepared from 3‐bromoaniline (0.19 ml, 1.74 mmol) and (furan‐3‐yl)boronic acid (0.21 g, 1.91 mmol) following the general procedure A as white solid (0.09 g, 32%). 1H NMR (300 MHz, CDCl₃) δ 7.66 - 7.71 (m, 1H), 7.45 (t, J = 1.70 Hz, 1H), 7.12 - 7.20 (m, 1H), 6.90 (td, J = 1.30, 7.58 Hz, 1H), 6.81 (t, J = 1.88 Hz, 1H), 6.66 (d, J = 1.13 Hz, 1H), 6.60 (ddd, J = 0.94, 2.45, 7.91 Hz, 1H), 3.69 (br. s., 2H). MS (ESI) m/z for C₁₀H₉NO [M+H]+: calcd: 160.1; found: 160.1.

3‐(Thiophen‐3‐yl)aniline (20b) was prepared from 3‐bromoaniline (0.19 ml, 1.74 mmol) and (thiophen‐3‐yl)boronic acid (0.25 g, 1.91 mmol) following the general procedure A as white solid (0.03 g, 10%). ¹H NMR (300 MHz, CDCl₃) δ 7.38 - 7.42 (m, 1H), 7.32 - 7.37 (m, 2H), 7.15 - 7.22 (m, 1H), 7.00 (td, J = 1.25, 7.68 Hz, 1H), 6.92 (t, J = 1.88 Hz, 1H), 6.63 (ddd, J = 0.94, 2.26, 7.91 Hz, 1H), 3.71 (br. s., 2H). MS (ESI) m/z for C₁₀H₉NS [M+H]⁺: calcd: 176.1; found: 176.3.

2‐(3‐Nitrophenyl)thiophene (21a) was prepared from 2‐bromothiophene (0.10 ml, 1.00 mmol) and 3-nitrophenylboronic acid (0.18 g, 1.10 mmol) following the general procedure A as yellow solid (0.18 g, 92%). ¹H NMR (300 MHz, CDCl₃) δ 8.42 (t, J = 1.88 Hz, 1H), 8.06 - 8.13 (m, 1H), 7.89 (d, J = 7.72 Hz, 1H), 7.53 (t, J = 8.01 Hz, 1H), 7.42 (dd, J = 0.85, 3.67 Hz, 1H), 7.37 (dd, J = 0.75, 5.09 Hz, 1H), 7.12 (dd, J = 3.77, 5.09 Hz, 1H).

2‐Methyl‐4‐(3‐nitrophenyl)thiophene (21b) was prepared from 4‐bromo‐2‐methylthiophene (0.10 ml, 1.00 mmol) and 3-nitrophenylboronic acid (0.18 g, 1.10 mmol) following the general procedure A as yellow solid (0.09 g, 42%). ¹H NMR (300 MHz, CDCl₃) δ 8.37 (t, J = 1.88 Hz, 1H), 8.09 (td, J = 1.06, 8.24 Hz, 1H), 7.85 (d, J = 7.72 Hz, 1H), 7.52 (t, J = 8.01 Hz, 1H), 7.32 (d, J = 1.51 Hz, 1H), 7.08 (s, 1H), 2.54 (s, 1H).

2‐Methyl‐5‐(3‐nitrophenyl)thiophene (21c) was prepared from 4‐bromo‐2‐methylthiophene (0.10 ml, 1.00 mmol) and 3-nitrophenylboronic acid (0.18 g, 1.10 mmol) following the general procedure A as yellow solid (0.09 g, 42%). ¹H NMR (300 MHz, CDCl₃) δ 8.36 (t, J = 1.98 Hz, 1H), 8.05 (ddd, J = 0.94, 2.12, 8.24 Hz, 1H), 7.79 - 7.84 (m, 1H), 7.50 (t, J = 8.01 Hz, 1H), 7.22 (d, J = 3.58 Hz, 1H), 6.75 - 6.79 (m, 1H), 2.50 - 2.56 (m, 3H).

4‐(3‐Nitrophenyl)‐1,3‐thiazole (21d) was prepared from 4‐bromo‐1,3‐thiazole (0.09 ml, 1.00 mmol) and 3-nitrophenylboronic acid (0.18 g, 1.10 mmol) following the general procedure A as yellow solid (0.12 g, 58%). ¹H NMR (300 MHz, CDCl₃) δ 8.93 (d, J = 1.88 Hz, 1H), 8.78 (t, J = 1.98 Hz, 1H), 8.26 - 8.31 (m, 1H), 8.17 - 8.22 (m, 1H), 7.72 (d, J = 2.07 Hz, 1H), 7.62 (t, J = 8.01 Hz, 1H).

5‐(3‐Nitrophenyl)‐1,3‐thiazole (21e) was prepared from 5‐bromo‐1,3‐thiazole (0.09 ml, 1.00 mmol) and 3-nitrophenylboronic acid (0.18 g, 1.10 mmol) following the general procedure A as yellow solid (0.10 g, 46%). ¹H NMR (300 MHz, CDCl₃) δ 8.86 (s, 1H), 8.43 (t, J = 1.88 Hz, 1H), 8.17 - 8.24 (m, 2H), 7.88 - 7.94 (m, 1H), 7.58 - 7.66 (m, 1H).

2‐(3‐Nitrophenyl)‐1,3‐thiazole (21f) was prepared from 4‐bromo‐1,3‐thiazole (0.09 ml, 1.00 mmol) and 3-nitrophenylboronic acid (0.18 g, 1.10 mmol) following the general procedure A as yellow solid (0.08 g, 37%). ¹H NMR (300 MHz, CDCl₃) δ 8.80 (t, J = 1.88 Hz, 1H), 8.24 - 8.32 (m, 1H), 7.87 - 7.96 (m, 2H), 7.43 - 7.47 (m, 2H).

tert‐Butyl 2‐(3‐nitrophenyl)‐1H‐pyrrole‐1‐carboxylate (23) was prepared from 1-bromo-3-nitrobenzene (0.20 g, 1 mmol) and N-Boc-2-pyrroleboronic acid (0.23 g, 1.1 mmol) following the general procedure A as yellow solid (0.15 g, 51%). ¹H NMR (300 MHz, CDCl₃) δ 8.12 - 8.28 (m, 2H), 7.69 (d, J = 6.22 Hz, 1H), 7.47 - 7.58 (m, 1H), 7.36 - 7.44 (m, 1H), 6.22 - 6.35 (m, 2H), 1.40 (s, 9H).

2‐(3‐Nitrophenyl)‐1H‐imidazole (27). To a solution of 3-nitrobenzonitrile (0.50 g, 3.37 mmol) in anhydrous methanol (17 ml) was added sodium methoxide (0.18 g, 3.37 mmol). The reaction mixture was stirred at room temperature for 5 h. Acid acetic (0.39 ml, 6.82 mmol) and aminoacetaldehyde dimethyl acetal (0.37 ml, 3.37 mmol) were then added and the reaction mixture was heated at 70 °C with stirring for 1 h. After cooling to room temperature, methanol (2.25 ml) and 6N aqueous HCl (1.7 ml) solution were added to the reaction mixture. The reaction temperature was subsequently raised to 70 °C for 1 h. After cooling to room temperature, the solvent was removed under reduced pressure. Saturated aqueous potassium carbonate was added slowly until pH 8-10. The desired product precipitated and was collected by filtration as white solid (0.39 g, 61%). ¹H NMR (300 MHz, CD₃OD) δ 7.21 - 7.27 (m, 1H), 6.66 - 6.77 (m, 2H), 6.13 - 6.24 (m, 1H), 5.64 - 5.73 (m, 2H). MS (ESI) m/z for C₉H₇N₃O₂ [M+H]⁺: calcd: 190.1; found: 190.3.

General procedure B.

To a solution of nitrobenzene derivative (1 eq) in ethanol (0.1 M) was added hydrazine hydrate (15 eq). The reaction was stirred at 50 °C for 15 min and an excess of Raney nickel slurry in water (1.2 eq) was added slowly. After 1 h, the bubbling ceased, the mixture was cooled to room temperature and filtered through Celite. The filtrate was condensed under reduced pressured and the residue was either used for the next step without purification or purified by column chromatography (SiO₂, ethyl acetate/hexanes) to afford the desired product.

3‐(Thiophen‐2‐yl)aniline (22a) was prepared from 21a (0.18 g, 0.89 mmol) following the general procedure B as yellow solid (0.10 g, 64%). ¹H NMR (300 MHz, CDCl₃) δ 7.20 - 7.29 (m, 2H), 7.10 - 7.18 (m, 1H), 6.98 - 7.08 (m, 2H), 6.92 (s, 1H), 6.59 (dd, J = 2.07, 7.91 Hz, 1H), 3.69 (br. s., 2H). MS (ESI) m/z for C₁₂H₉Cl₂N [M+H]⁺: calcd: 238.0; found: 238.1.

3‐(5‐Methylthiophen‐3‐yl)aniline (22b) was prepared from 21b (0.09 g, 0.42 mmol) following the general procedure B as white solid (0.08 g, quant.). ¹H NMR (300 MHz, CDCl₃) δ 7.12 - 7.17 (m, 2H), 6.93 - 7.02 (m, 2H), 6.87 (d, J = 1.88 Hz, 1H), 6.59 (dd, J = 1.51, 7.91 Hz, 1H), 3.68 (br. s., 2H), 2.50 (s, 3H). MS (ESI) m/z for C₁₁H₁₁NS [M+H]⁺: calcd: 190.1; found: 190.2.

3‐(5‐Methylthiophen‐2‐yl)aniline (22c) was prepared from 21c (0.09 g, 0.42 mmol) following the general procedure B as colorless liquid (0.08 g, quant.). ¹H NMR (300 MHz, CDCl₃) δ 7.09 - 7.16 (m, 1H), 7.05 (d, J = 3.39 Hz, 1H), 6.93 - 6.99 (m, 1H), 6.86 (t, J = 1.88 Hz, 1H), 6.67 - 6.72 (m, 1H), 6.56 (ddd, J = 0.94, 2.26, 7.91 Hz, 1H), 3.65 (br. s., 2H), 2.48 (s, 2H). MS (ESI) m/z for C₁₁H₁₁NS [M+H]⁺: calcd: 190.1; found: 190.3.

3‐(1,3‐Thiazol‐4‐yl)aniline (22d) was prepared from 21d (0.12 g, 0.58 mmol) following the general procedure B as colorless liquid (0.06 g, 61%). ¹H NMR (300 MHz, CDCl₃) δ 8.86 (d, J = 1.88 Hz, 1H), 7.48 (d, J = 1.88 Hz, 1H), 7.26 - 7.35 (m, 2H), 6.69 (dd, J = 1.22, 2.35 Hz, 1H), 6.23 (dd, J = 2.26, 7.91 Hz, 1H), 3.71 (br. s., 2H). MS (ESI) m/z for C₉H₈N₂S [M+H]⁺: calcd: 177.1; found: 177.5.

3‐(1,3‐Thiazol‐5‐yl)aniline (22e) was prepared from 21e (0.09 g, 0.46 mmol) following the general procedure B as colorless liquid (0.06 g, 73%). ¹H NMR (300 MHz, CDCl₃) δ 8.72 (s, 1H), 8.03 (s, 1H), 7.13 - 7.23 (m, 1H), 6.97 (dd, J = 0.94, 7.72 Hz, 1H), 6.88 (d, J = 3.77 Hz, 1H), 6.66 (td, J = 1.06, 8.05 Hz, 1H), 3.80 (br. s., 2H). MS (ESI) m/z for C₉H₈N₂S [M+H]⁺: calcd: 177.1; found: 177.4.

3‐(1,3‐Thiazol‐2‐yl)aniline (22f) was prepared from 21f (0.08 g, 0.37 mmol) following the general procedure B as colorless liquid (0.03 g, 46%). ¹H NMR (300 MHz, CDCl₃) δ 7.84 (d, J = 3.39 Hz, 1H), 7.29 - 7.35 (m, 3H), 7.22 (t, J = 8.01 Hz, 1H), 6.72 - 6.77 (m, 1H), 3.80 (br. s., 2H). MS (ESI) m/z for C₉H₈N₂S [M+H]⁺: calcd: 177.1; found: 177.3.

tert‐Butyl 2‐(3‐aminophenyl)‐1H‐pyrrole‐1‐carboxylate (24) was prepared from 23 (0.147 g, 0.51 mmol) following the general procedure B as white solid (0.09 g, 65%). ¹H NMR (300 MHz, CDCl₃) δ 7.21 - 7.37 (m, 1H), 7.06 - 7.18 (m, 1H), 6.70 - 6.78 (m, 1H), 6.57 - 6.68 (m, 2H), 6.18 (d, J = 10.74 Hz, 2H), 3.63 (br. s., 2H), 1.37 (s, 9H). MS (ESI) m/z for C₁₅H₁₈N₂O₂ [M+H]⁺: calcd: 259.1; found: 259.5.

3‐(1H‐Pyrrol‐2‐yl)aniline (25). A solution of 24 (0.06 g, 0.23 mmol) in 5% aqueous potassium hydroxide (23 ml) was refluxed for 4 h. After cooling to room temperature, the mixture was poured into water and extracted with dichloromethane. The organic phase was dried with anhydrous magnesium sulfate, filtered, concentrated in vacuo to the desired product as white solid (0.03 g, 78%). ¹H NMR (300 MHz, CDCl₃) δ 8.41 (br. s., 1H), 7.05 - 7.22 (m, 1H), 6.73 - 6.94 (m, 3H), 6.40 - 6.60 (m, 2H), 6.28 (s, 1H), 3.67 (br. s., 2H). MS (ESI) m/z for C₁₀H₁₀N₂ [M+H]⁺: calcd: 159.1; found: 159.2.

3‐(1‐Methyl‐1H‐pyrrol‐2‐yl)aniline (26). To a solution of 3-iodoaniline (0.24 ml, 2 mmol) in N,N-dimethylacetamide (8 ml) in a sealed tube was added N-methylpyrrole (0.36 ml, 8 mmol), potassium acetate (0.39 g, 8 mmol), and palladium (II) acetate (0.005 g, 0.02 mmol). The reaction mixture was stirred at 150 °C for 20 h. The reaction mixture was then diluted with ethyl acetate, washed three times with water and once with brine. The organic layer was dried with anhydrous magnesium sulfate, filtered, and concentrated under reduced pressure. The residue was purified by column chromatography (SiO₂, ethyl acetate/ hexanes) to give the desired product was yellow liquid (0.08 g, 22%). ¹H NMR (300 MHz, CDCl₃) δ 7.17 (t, J = 7.82 Hz, 1H), 6.76 - 6.81 (m, 1H), 6.69 (td, J = 2.10, 8.24 Hz, 2H), 6.61 (ddd, J = 0.75, 2.31, 8.05 Hz, 1H), 6.16 - 6.21 (m, 2H), 3.68 (br. s., 2H), 3.64 (s, 3H). MS (ESI) m/z for C₁₁H₁₂N₂ [M+H]⁺: calcd: 173.1; found: 173.5.

3‐(1H‐Imidazol‐2‐yl)aniline (28) was prepared from 27 (0.30 g, 1.58 mmol) following the general procedure B as white solid (0.25 g, quant.). ¹H NMR (300 MHz, CDCl₃) δ 7.05 - 7.23 (m, 5H), 6.69 (d, J = 8.10 Hz, 1H), 3.74 (br. s., 2H). MS (ESI) m/z for C₉H₉N₃ [M+H]⁺: calcd: 160.1; found: 160.2.

General procedure C.

To a solution of aryl amine (1 eq) in anhydrous chloroform (0.04 M) was added 4-chlorophenyl isocyanate (1eq) at room temperature. The reaction mixture was then heated at 60 °C for 16 h. The precipitated product was filtered and thoroughly washed with dichloromethane.

3‐(4‐Chlorophenyl)‐1‐[3‐(furan‐3‐yl)phenyl]urea (8) was prepared from 20a (0.03 g, 0.16 mmol) following the general procedure C as white solid (0.05 g, 95%). ¹H NMR (300 MHz, DMSO-d₆) δ 8.86 (s, 1H), 8.73 (s, 1H), 8.12 (s, 1H), 7.75 (t, J = 1.70 Hz, 1H), 7.65 (s, 1H), 7.47 - 7.52 (m, 2H), 7.29 - 7.36 (m, 4H), 7.20 - 7.28 (m, 1H), 6.88 (d, J = 0.94 Hz, 1H). HRMS (ESI) m/z for C₁₇H₁₃ClN₂O₂ [M+H]⁺: calcd: 313.0738; found: 313.0744.

3‐(4‐Chlorophenyl)‐1‐[3‐(thiophen‐3‐yl)phenyl]urea (9) was prepared from 20b (0.03 g, 0.15 mmol) following the general procedure C as white solid (0.03 g, 63%). ¹H NMR (300 MHz, DMSO-d₆) δ 8.85 (s, 1H), 8.76 (s, 1H), 7.75 - 7.79 (m, 2H), 7.63 (dd, J = 2.92, 4.99 Hz, 1H), 7.45 - 7.51 (m, 3H), 7.30 - 7.36 (m, 5H). HRMS (ESI) m/z for C₁₇H₁₃ClN₂OS [M+H]⁺: calcd: 329.0510; found: 329.0504.

3‐(4‐Chlorophenyl)‐1‐[3‐(thiophen‐2‐yl)phenyl]urea (10) was prepared from 22a (0.10 g, 0.67 mmol) following the general procedure C as white solid (0.13 g, 57%). ¹H NMR (300 MHz, DMSO-d₆) δ 8.84 (s, 2H), 7.82 (s, 1H), 7.55 (d, J = 5.09 Hz, 1H), 7.50 (d, J = 8.85 Hz, 2H), 7.44 - 7.47 (m, 1H), 7.28 - 7.36 (m, 5H), 7.14 (dd, J = 3.67, 4.99 Hz, 1H). HRMS (ESI) m/z for C₁₇H₁₃ClN₂OS [M+H]⁺: calcd: 329.0510; found: 329.0508.

3‐(4‐Chlorophenyl)‐1‐[3‐(5‐methylthiophen‐3‐yl)phenyl]urea (11) was prepared from 22b (0.03 g, 0.17 mmol) following the general procedure C as white solid (0.04 g, 70%). ¹H NMR (300 MHz, CD₃OD) δ 7.73 (br. s., 1H), 7.61 (d, J = 7.54 Hz, 1H), 7.39 - 7.48 (m, 4H), 7.32 - 7.38 (m, 2H), 7.24 - 7.31 (m, 4H), 2.51 (s, 3H). HRMS (ESI) m/z for C₁₈H₁₅ClN₂OS [M+H]⁺: calcd: 343.0666; found: 343.0666.

3‐(4‐Chlorophenyl)‐1‐[3‐(5‐methylthiophen‐2‐yl)phenyl]urea (12) was prepared from 22c (0.04 g, 0.19 mmol) following the general procedure C as white solid (0.05 g, 77%). ¹H NMR (300 MHz, CD₃OD) δ 7.71 (s, 1H), 7.44 (d, J = 8.85 Hz, 2H), 7.21 - 7.32 (m, 5H), 7.16 (d, J = 3.58 Hz, 1H), 6.74 (d, J = 2.64 Hz, 1H), 2.49 (s, 3H). HRMS (ESI) m/z for C₁₈H₁₅ClN₂OS [M+H]⁺: calcd: 343.0666; found: 343.0665.

3‐(4‐Chlorophenyl)‐1‐[3‐(1H‐pyrrol‐2‐yl)phenyl]urea (13) was prepared from 25 (0.03 g, 0.18 mmol) following the general procedure C as white solid (0.05 g, 81%). ¹H NMR (300 MHz, DMSO-d₆) δ 11.26 (br. s., 1H), 8.87 (br. s., 1H), 8.67 (br. s., 1H), 7.67 (s, 1H), 7.47 - 7.56 (m, 2H), 7.30 - 7.39 (m, J = 5.70 Hz, 2H), 7.22 - 7.29 (m, 3H), 6.84 (s, 1H), 6.44 (s, 1H), 6.12 (s, 1H). ¹³C NMR (75 MHz, DMSO-d₆) δ 152.5, 139.8, 138.7, 133.6, 131.1, 129.0, 128.6, 125.3, 119.7, 119.3, 117.5, 115.9, 113.7, 109.0, 105.5. HRMS (ESI) m/z for C₁₇H₁₄ClN₃O [M+H]⁺: calcd: 312.0898; found: 312.0894.

3‐(4‐Chlorophenyl)‐1‐[3‐(1‐methyl‐1H‐pyrrol‐2‐yl)phenyl]urea (14) was prepared from 26 (0.08 g, 0.45 mmol) following the general procedure C as white solid (0.14 g, 98%). ¹H NMR (300 MHz, DMSO-d₆) δ 8.86 (s, 1H), 8.77 (s, 1H), 7.56 (s, 1H), 7.50 (d, J = 8.85 Hz, 2H), 7.30 - 7.36 (m, 4H), 7.03 - 7.08 (m, 1H), 6.83 (t, J = 2.17 Hz, 1H), 6.15 (dd, J = 1.88, 3.58 Hz, 1H), 6.04 - 6.08 (m, 1H), 3.66 (s, 3H). ¹³C NMR (75 MHz, DMSO-d₆) δ 152.5, 139.6, 138.6, 133.4, 128.8, 128.6, 125.4, 124.3, 121.6, 119.8, 117.8, 116.6, 108.3, 107.3, 34.9. HRMS (ESI) m/z for C₁₈H₁₆ClN₃O [M+H]⁺: calcd: 326.1055; found: 326.1049.

tert‐Butyl 2‐[3‐({1‐[(4‐chlorophenyl)amino]ethenyl}amino)phenyl]‐1H‐pyrrole‐1‐carboxylate (15) was prepared from 24 (0.03 g, 0.10 mmol) following the general procedure C as white solid (0.04 g, 84%). ¹H NMR (300 MHz, DMSO-d₆) δ 8.83 (s, 1H), 8.75 (s, 1H), 7.46 - 7.51 (m, 2H), 7.31 - 7.39 (m, 3H), 7.23 - 7.30 (m, 1H), 6.94 (d, J = 7.54 Hz, 1H), 6.26 - 6.30 (m, 1H), 6.22 - 6.26 (m, 1H), 1.31 (s, 9H). ¹³C NMR (75 MHz, DMSO-d₆) δ 152.4, 148.8, 138.9, 138.6, 134.4, 134.2, 128.6, 128.0, 125.3, 122.5, 119.7, 118.6, 117.0, 114.1, 110.7, 83.5, 27.1. HRMS (ESI) m/z for C₂₂H₂₂ClN₃O₃ [M+H]⁺: calcd: 412.1422; found: 412.1423.

3‐(4‐Chlorophenyl)‐1‐[3‐(1,3‐thiazol‐4‐yl)phenyl]urea (16) was prepared from 22d (0.03 g, 0.19 mmol) following the general procedure C as white solid (0.05 g, 74%). ¹H NMR (300 MHz, DMSO-d₆) δ 9.20 (d, J = 1.88 Hz, 1H), 8.85 (s, 2H), 8.12 - 8.15 (m, 1H), 8.11 (d, J = 1.88 Hz, 1H), 7.59 (d, J = 7.35 Hz, 1H), 7.48 - 7.53 (m, 2H), 7.31 - 7.45 (m, 5H). HRMS (ESI) m/z for C₁₆H₁₂ClN₃OS [M+H]⁺: calcd: 330.0462; found: 330.0460.

3‐(4‐Chlorophenyl)‐1‐[3‐(1,3‐thiazol‐5‐yl)phenyl]urea (17) was prepared from 21e (0.03 g, 0.16 mmol) following the general procedure C as white solid (0.04 g, 66%). ¹H NMR (300 MHz, DMSO-d₆) δ 9.09 (s, 1H), 8.88 (s, 2H), 8.26 (s, 1H), 7.83 (s, 1H), 7.50 (d, J = 8.85 Hz, 2H), 7.30 - 7.41 (m, 5H). HRMS (ESI) m/z for C₁₆H₁₂ClN₃OS [M+H]⁺: calcd: 330.0462; found: 330.0454.

3‐(4‐Chlorophenyl)‐1‐[3‐(1,3‐thiazol‐2‐yl)phenyl]urea (18) was prepared from 22f (0.03 g, 0.16 mmol) following the general procedure C as white solid (0.04 g, 72%). ¹H NMR (300 MHz, DMSO-d₆) δ 8.96 (s, 1H), 8.87 (s, 1H), 8.22 (s, 1H), 7.93 (d, J = 3.20 Hz, 1H), 7.80 (d, J = 3.01 Hz, 1H), 7.48 - 7.59 (m, 3H), 7.40 - 7.47 (m, 2H), 7.30 - 7.39 (m, 2H). HRMS (ESI) m/z for C₁₆H₁₂ClN₃OS [M+H]⁺: calcd: 330.0462; found: 330.0464.

3‐(4‐Chlorophenyl)‐1‐[3‐(1H‐imidazol‐2‐yl)phenyl]urea (19) was prepared from 28 (0.13 g, 0.82 mmol) following the general procedure C as white solid (0.22 g, 86%). ¹H NMR (300 MHz, DMSO-d₆) δ 12.52 (br. s., 1H), 8.84 (d, J = 11.30 Hz, 2H), 8.07 (br. s., 1H), 7.52 (br. s., 3H), 7.44 (br. s., 1H), 7.34 (d, J = 7.16 Hz, 3H), 7.03 - 7.20 (m, 2H). ¹³C NMR (75 MHz, DMSO-d₆) δ 152.4, 145.5, 139.8, 138.7, 131.4, 129.1, 128.6, 125.4, 119.7, 118.6, 118.0, 115.1. HRMS (ESI) m/z for C₁₆H₁₃ClN₄O [M+H]⁺: calcd: 313.0851; found: 313.0846.

Calcium Mobilization Assay.

CHO-RD-HGA16 cells (Molecular Devices, CA) stably expressing the human CB₁ receptor were plated into 96-well black-walled assay plates at 25,000 cells/well in 100 μL of Ham’s F12 (supplemented with 10% fetal bovine serum, 100 units of penicillin/streptomycin, and 100 μg/mL Normocin) and incubated overnight at 37 °C, 5% CO₂. Calcium 5 dye (Molecular Devices, CA) was reconstituted according to the manufacturer’s instructions. The reconstituted dye was diluted 1:40 in prewarmed (37 °C) assay buffer (1x HBSS, 20 mM HEPES, 2.5 mM probenecid, pH 7.4 at 37 °C). Growth medium was removed, and the cells were gently washed with 100 μL of prewarmed (37 °C) assay buffer. The cells were incubated for 45 min at 37 °C, 5% CO₂ in 200 μL of the diluted Calcium 5 dye solution. For antagonist assays to determine IC₅₀ values, the EC₈₀ concentration of CP55,940 was prepared at 10x the desired final concentration in 0.25% BSA/0.5% DMSO/0.5% EtOH/assay buffer, aliquoted into 96-well polypropylene plates, and warmed to 37 °C. Serial dilutions of the test compounds were prepared at 10x the desired final concentration in 2.25% BSA/4.5% DMSO/4.5% EtOH/assay buffer. After the dye loading incubation period, the cells were pretreated with 25 μL of the test compound serial dilutions and incubated for 15 min at 37 °C. After the pretreatment incubation period, the plate was read with a FLIPR Tetra (Molecular Devices, CA). Calcium-mediated changes in fluorescence were monitored every 1 s over a 90 s time period, with the Tetra adding 25 μL of the CP55,940 EC₈₀ concentration at the 10s time point (excitation/emission: 485/525 nm). Relative fluorescence units (RFU) were plotted against the log of compound concentrations. For agonist screens, the above procedure was followed except that cells were pretreated with 2.25% BSA/4.5% DMSO/4.5% EtOH/assay buffer and the Tetra added single concentration dilutions of the test compounds prepared at 10x the desired final concentration in 0.25% BSA/0.5% DMSO/0.5% EtOH/assay buffer. Test compound RFUs were compared to the CP55,940 E_max RFUs to generate % E_max values. For the CB₂ agonist and antagonist assays, the same procedures were followed except that stable human CB₂-CHO-RD-HGA16 cells were used.

Membrane preparation.

Cerebella from male ICR mice (6-8 weeks old; Enviga International, Indianapolis, IN) were dissected on ice, flash frozen in liquid N₂ and stored at −80°C until the day of the experiment. On the day of the experiment, cerebella were mechanically homogenized (Brinkmann Polytron) in membrane buffer (50 mM Tris, 3 mM MgCl2, 0.2 mM EGTA, 100 mM NaCl, pH 7.4) on ice, and then centrifuged for 10 min at 40,000 g at 4°C. The supernatant was discarded and the pellet was suspended in membrane buffer, homogenized, and centrifuged again for 10 min at 40,000 g. The pellet was resuspended in membrane buffer and protein quantified by Bradford method.

Radioligand Binding.

For equilibrium radioligand binding, reactions were carried out in assay buffer (membrane buffer containing 1 mg/ml bovine serum albumin; BSA) and initiated by addition of cerebellar membranes (10 μg protein) and incubated for 90 min at 30°C. For all radioligand binding experiments, non-specific binding was determined by excess cold ligand (1 μM). Total bound of radioligand was less than 10% of total added (minimal ligand depletion). For competition curves, either 1 nM [³H]SR141716 (NIDA Drug Supply Program, Bethesda, MD) or 1 nM [³H]CP55,940 (NIDA) were incubated with multiple concentrations of the allosteric modulators. For Saturation binding of [³H]CP55,940, reactions were carried out by incubating the radioligand at nominal concentrations of 0.032, 0.1, 0.32, 0.56, 1, 3.2, and 5.6 nM with a fixed concentration (10 μM) of allosteric modulator. Radioligand binding reactions were terminated by vacuum filtration through Perkin Elmer GF/C 96 well filter plates followed by 3 rinses of wash buffer.

Agonist-stimulated [³⁵S]GTPγS binding.

For receptor signaling, membranes (10 μg protein) were preincubated in assay buffer for 10 min with 3 units/ml adenosine deaminase then incubated for 60 min at 30°C with 30 μM GDP and 0.1 nM [³⁵S]GTPγS (Perkin Elmer Life Sciences, Boston, MA). Non-specific binding was determined by adding 30 μM unlabeled GTPγS. Concentration response curves for allosteric modulators were conducted in the presence of CP55,940 (100 nM) to calculate IC₅₀ values.

Data Analysis

For calcium mobilization experiments, data were fit to a three-parameter logistic curve to generate IC₅₀ values (GraphPad Prism 6.0, CA). For [³⁵S]GTP-γ-S experiments, data were normalized to maximal CP55,940 (100 nM) stimulation in the absence of test compound (i.e., vehicle = 100%). Curve fits were accomplished using GraphPad Prism 6.0 and data were fit to three-parameter nonlinear regression, with bottom and top constrained to >0 and = 100, respectively, for IC₅₀ calculation. For equilibrium radioligand binding data, concentration response curves for the allosteric modulators using a single fixed concentration of radioligand were fit to 3 parameter non-linear regression for determination of EC₅₀ and curve top ([³H]CP55,940 binding) and IC₅₀ ([³H]SR141716 binding). For saturation binding, data were fit to a one-site specific binding using GraphPad Prism 6.0 to calculate [³H]CP55,940’s B_max and K_d.

ABBREVIATIONS

CB₁

cannabinoid 1 receptor

CB₂

cannabinoid 2 receptor

CNS

central nervous system

FLIPR

fluorometric imaging plate reader

GPCR

G-protein-coupled receptor

HPLC

high performance liquid chromatography

IC₅₀

half-maximum inhibitory concentration

MS

mass spectrometry

NAM

negative allosteric modulator

NMR

nuclear magnetic resonance

PAM-Antagonist

positive allosteric modulator - Antagonist

SAR

structure–activity relationship

TLC

thin-layer chromatography