FLUORINATED CANNABINOID CB2 RECEPTOR LIGANDS: SYNTHESIS AND IN VITRO BINDING CHARACTERISTICS OF 2-OXOQUINOLINE DERIVATIVES

Abstract

Cannabinoid receptor 2 (CB2) plays an important role in human physiology and the pathophysiology of different diseases, including neuroinflammation, neurodegeneration, and cancer. Several classes of CB2 receptor ligands, including 2-oxoquinoline derivatives, have been previously reported. We report the synthesis and results of in vitro receptor binding of a focused library of new fluorinated 2-oxoquinoline CB2 ligands. Twelve compounds, 13-16 18, 19, 21-24, 27, and 28 were synthesized in good yields in multiple steps. Human U87 glioma cells expressing either hCB1 (control) or hCB2 were generated via lentiviral transduction. In vitro competitive binding assay was performed using [³H]CP-55,940 in U87hCB1 and U87hCB2 cells. Inhibition constant (Ki) values of compounds 13, 14, 15, 16, 18, 19, 21, 22, 23, 24, 27, and 28 for CB2 were >10000, 2.8, 5.0, 2.4, 22, 0.8, 1.4, >10000, 486, 58, 620, and 2400 nM, respectively, and those for CB1 were >10000 nM. Preliminary in vitro results suggest that six of these compounds may be useful for therapy of neuropathic pain, neuroinflammatory diseases and immune disorders. In addition, compound 19, with its subnanomolar Ki value, could be radiolabeled with ¹⁸F and explored for PET imaging of CB2 expression.

1. Introduction

The hemp plant, Cannabis sativa L., known as marijuana, has been used for centuries as a therapeutic agent and recreational drug.^1,2 The active components of marijuana and their derivatives are classified as cannabinoids.² Two subtypes of the mammalian cannabinoid receptors have been identified: the CB1 and CB2 receptors. The CB1 receptor is primarily located in the central nervous system, ^{3, 4} and the CB2 receptor is expressed in cells of the immune system, spleen, tonsils, and lymph nodes.^{3, 5, 6} The CB1 and CB2 receptors are both G-protein coupled receptors and are involved in the inhibition of adenylate cyclase.

CB2 is believed to be devoid of psychoactivity, and has significant anti-inflammatory functions,⁷ and inflammation is known to be a critical part of many types of neurodegenerative diseases.⁸ CB2 receptors have been implicated in a range of leukocyte functions, and studies of CB2 receptors in leukocytes have shown results consistent with an anti-inflammatory and immunosuppressive role. This has also been supported by demonstrations that CB2 regulates inflammation in a diverse range of animal models, including gastro-intestinal,⁹ acute hindpaw,¹⁰ and pulmonary inflammation.¹¹ CB2 expression in microglia in the brain was studied using human brain tissue taken from a patient with Alzheimer’s disease. Microglia clustering at β-amyloid plaques expressed both CB2 receptor proteins and fatty acid amine hydrolase.⁷ Ramirez et al.¹² replicated this finding and investigated the role of CB2 in the pathogenesis of a mouse model of Alzheimer’s disease. CB2 receptor expression was also found to be up-regulated in sub-populations of microglia in a rat model of Huntington’s disease.¹³ CB2 receptors have been reported to be over-expressed in several cancer cells, including breast cancer,¹⁴ skin cancer,¹⁵ prostate,¹⁶ and hepatocellular carcinoma.¹⁷ Experimental evidence has shown that cannabinoids inhibit the growth of tumor xenografts in mice, tumor angiogenesis, and directly induce apoptosis in neoplastic cells.^14,15

The CB2 receptor is an important target for therapeutic immune intervention, and, as such, research is currently focused on the development of CB2 selective ligands. WIN-55,212-2 was the first reported CB2 selective ligand, and it has been thoroughly investigated.¹ As a result, there are several different documented values for the binding affinity of WIN-55,212-2 at both CB1 and CB2 receptors.¹⁸ While WIN-55,212-2 is extensively used for evaluating the receptor binding of potential cannabinoid ligands, it is not very useful pharmacologically as a selective ligand for either cannabinoid receptor because it has high affinity for both receptors.^{1, 9}

Subsequent to WIN-55,212-2, several other CB2 selective ligands have been synthesized, including cannabimimetic indoles, such as 1-(2,3-dichlorobenzoyl)-2-methyl-3-(2-[1-morpholino]ethyl)-5-methoxyindole (L768242)¹⁹ and 1-propyl-2-methyl-3-(1-naphthoyl)indole (JWH-015).²⁰ Indole L768242 has exceptionally high selectivity for the CB2 receptor (146-fold) with significantly high affinity at CB2 (inhibition constant [Ki]=14 nM. The Huffman group has synthesized a variety of CB2 selective 1-deoxy and 1-methoxy-Δ⁸–THC analogues, which include 1-deoxy-3-(1′,1′-dimethylbutyl)-Δ⁸-THC (JWH-133), one of the most highly selective ligands for the CB2 receptor (Ki=677±132 nM at CB1 and Ki = 3.4±1.0 nM at CB2).²¹ 1-Methoxy-3-(1′,1′-dimethylhexyl)-Δ8-THC (JWH-229) is another highly selective ligand, with more than 170-fold selectivity for the CB2 receptor (Ki=3134±110 nM at CB1 and Ki=18±2 nM at CB2).²² This group has also synthesized a large number of indole derivatives to identify CB2-specific ligands.^1,2,6,23 A variety of cannabimimetic indole derivatives have also been developed by the pharmaceutical companies, including Bristol-Myers Squibb and Abbott Laboratories.^{24, 25}

In addition to traditional dibenzopyran-based cannabinoids ^{1, 2, 21, 22} and cannabimimetic indoles, ^{6, 24-26} a variety of potent cannabinoid ligands that are resorcinol derivatives ^{27, 28} have been synthesized and reported. Two series of resorcinol dimethyl ethers were also reported by Wiley and others.²⁸ One of these compounds, O-1966A has very high (220-fold) selectivity for the CB2 receptor combined with very low affinity for the CB1 receptor (Ki = 5,055±984 nM at CB1 and Ki = 23±2.1 nM at CB2). Mussinu et al. ²⁹ described several highly CB2 selective tricyclic pyrazoles based upon CB1 antagonists SR141716A (Rimonabant®) ³⁰ and the CB2 inverse agonist SR144528.^{31, 32} These pyrazoles are among the most highly selective ligands for the CB2 receptor to be reported to date, with selectivity ranging from 32-fold to 9810-fold depending on the substituent at the aromatic ring.

Iwamura et al. have synthesized several 2-oxoquinoline analogues with exceptionally high selectivity for the CB2 receptor.³³ Compound JTE-907 (29, Table 2, supplemental data) in particular, is described as an inverse agonist in vitro for the CB2 receptor and possesses anti-inflammatory properties in vivo. JTE-907 (29) has modest affinity (Ki = 35.9 nM) ¹ for the CB2 receptor, with an exceptionally high selectivity of 2,760-fold at this receptor. Structural variations in these 2-oxoquinolines include an N-methyl substituent at the quinoline ring; variable alkoxy substituents in the 6, 7, and 8 positions of the quinoline moiety and substituents at the 3-position that vary from carboxylic acid groups to various amide and ester groups.³⁴ Compound 30 (Table 2) in particular has very high CB2 receptor affinity (Ki = 0.014 nM) and extremely high selectivity (262,202-fold) for this receptor. Raitio et al. synthesized a new series of CB2 inverse agonists with responses comparable to SR144528,^{34, 35} based on the structure of 2-oxoquinoline JTE-907 (29, Table 2) in which the piperonylamide end has been replaced with a variety of aromatic amide structures (32-37 Table 2).³⁴

Several other classes of compounds, including tricyclic pyrazoles,^30-32 sulfamoyl benzamides,³⁶ triarylbis-sulfones, ³⁷ and arylsulfonamide,³⁸ have been synthesized and reported to be CB2 receptor ligands. However, among all the classes of CB2 ligands, 2-oxoquinoline derivatives appear to be the most efficient inverse agonists, with high binding affinity and selectivity for CB2 receptor. ^{34, 35}

In order to develop novel CB2-specific ligands for therapy of neurological diseases and non-invasive imaging of CB2 receptor expression, we have rationally designed a focused library of potential CB2-specific compounds. We report synthesis of a new library of fluorinated analogues of 2-oxoquinoline as CB2 receptor ligands and results of their in vitro receptor binding in cell lines genetically engineered to express the CB1 or CB2 receptors. These studies demonstrate that some of the fluorinated 2-oxoquinoline analogues have high binding efficiency and selectivity for CB2 receptor, and these compounds may be suitable for treatment of neuropathic pain, inflammation, and immune disorders.

2. Results and discussion

2.1. Chemistry

The scheme for synthesis of compounds 13, 14, 15, and 16 is shown in Fig. 1 (Synthetic Scheme 1).

Figure 1.

Scheme for synthesis of 7-methoxy-8-alkoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid 4-fluoro-benzylamides 13, 14 and 4-fluoro-phenethylamides 15 and 16.

a) DCM, -20°C, NO₂BF₄, 24h; b) DMF, K₂CO₃, MeI or 1-bromobutane, 80°C 24h; c) EtOH, Fe, HCl, H₂O, reflux 15 min; d) EtOH, CH₂(CO₂CH₃)₂, piperidine, AcOH, reflux 16h; e) EtOH, HCl/H₂O, 60°C 16h; f) Toluene, SOCl₂, reflux 3h; g) DCM, Et₃N, amine, RT 1h.

Compounds 2-12 were prepared following previously published methods^{34, 39, 40} as shown in the synthetic scheme 1 (Fig. 1) with minor modifications. Compound 2 was obtained in 80% yield and compounds 3 and 4 were obtained in 70% and 60% yields, respectively. Reduction of 3 and 4 using iron (Fe) powder and conc. HCl produced compounds 5 and 6 in 80% and 83% yields, respectively. Compounds 7 and 8 were obtained from 5 and 6, respectively, in 80% yields. The ethyl esters 7 and 8 were hydrolyzed by 2M HCl acid to produce 9 and 10 in 73% and 75% yields, respectively. Intermediate compounds 2-10 were fully characterized by spectroscopic methods, and the ¹H NMR spectra were consistent with those reported previously.³⁹ Detailed experimental procedures for compounds 2-12 are presented in the supplemental data.

Compounds 11 and 12 were prepared in situ and used without isolation and purification. Compounds 13, 14, 15, and 16 were prepared from the intermediate acid chlorides 11 and 12 by reaction with commercially available 4-fluorobenzyl amine and 4-fluoro-phenyl ethylamine (Scheme 1). The yields for 13, 14, 15 and 16 were 76%, 70%, 86% and 96%, respectively.

Compound 17 was synthesized in house according to a previously reported method ⁴¹ and reacted with the acid chlorides 11 and 12 as shown in Fig. 2. The desired products 18 and 19 were obtained in 60% and 85% yields, respectively.

Figure 2.

Scheme for synthesis of 7-methoxy-8-alkoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid 2-fluoro-2-phenethylamides 18 and 19.

a) Toluene, SOCl₂, reflux 3h; b) DCM, Et₃N, 17, RT 1h.

Synthesis of the compounds 20 and 21 is shown in Fig. 3 (Synthetic Scheme 3).

Figure 3.

Scheme for synthesis of 7-methoxy-8-butyloxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid 4-amino-phenethylamide 20, 7-methoxy-8-butoxyoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid 4-fluoro-phenethylamide 16 and 6-nitro-7-methoxy-8-butyloxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid 4-fluoro-phenethylamide 21.

a) DCM, Et₃N, 4-aminophenylethylamine, RT, 30 min; b) NaNO₂, pyridine.HF, RT 1h, 85°C 1h.

Compound 20 was prepared by reaction of 12 with 4-aminophenyl ethylamine and obtained in 80% yield. The experimental details are identical as those described for preparation of compounds 13-16 (Scheme 1). Compound 20 was used for an alternative synthesis of 16 and synthesis of 21.

Compound 16 was synthesized by two different methods: Method 1: The acid chloride 12 was reacted with 4-fluorophenyl ethylamine to produce 16 (as shown in Fig. 1). In this method, the yield of 16 was 97%; however, the method is not suitable for radiosynthesis using ¹⁸F, because ¹⁸F should be inserted into the benzene ring by a special method. Therefore, we explored an alternative method, as shown in Fig. 3, which should be suitable for radiosynthesis of 16. Method 2: In this method, diazotization of 20 was performed to produce an azo-salt compound in situ, which was converted to 16 by reaction with HF in pyridine. This reaction produced a mixture of two products, 16 and 21, in 40% and 36% yields, respectively. The formation of 21 is quite interesting; in addition to fluorination of the aromatic ring, C₆-hydrogen was substituted by a nitro group. Thus, compound 21 was obtained as a by-product during diazotization of 20, followed by reaction with HF in pyridine.

Synthesis of the compounds 22-24 is shown in Fig. 4 (Synthetic Scheme 4).

Figure 4.

Scheme for synthesis of 7-methoxy-8-alkoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid fluoro-alkylamides 22-24.

The acid chlorides 11 and 12 were prepared in situ from 9 and 10 as described in scheme 1, then coupled with the corresponding fluoroalkyl amines with two different carbon-chain lengths to produce compounds 22-24. The yields of these compounds were 90%-95%, almost quantitative.

Synthesis of compounds 27 and 28 is presented in Fig. 5 (Synthetic Scheme 5).

Figure 5.

Scheme for synthesis of 7-methoxy-8-butoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-(fluoroethyl-triazolalkyl)amides 27 and 28.

a) CH₂Cl₂, Et₃N rt; b) CuSO₄, Na-ascorbate, DMF, rt.

The acid chloride 12 was prepared in situ as shown in scheme 1, then coupled with the alkyne-amines following a previously published methodology ⁴² to produce 25 and 26. Both compounds 25 and 26 were obtained in quantitative yields. Fluroethylazide was prepared following a previously reported method ⁴³ and used without purification. Compounds 25 and 26 were then reacted with fluoroethylazide in the presence of CuSO₄ and Na-ascorbate at room temperature using click chemistry to produce 27 and 28. The yields in this step were 55% and 58%, respectively, for 27 and 28.

All new compounds were fully characterized by ¹H, ¹³C, and ¹⁹F NMR spectroscopy and high-resolution mass spectrometry. All compounds were analyzed by high-performance liquid chromatography (HPLC) for chemical purity and found to be greater than 98% pure. These compounds were used for an in vitro binding assay with the CB2 and CB1 receptors. HPLC chromatograms of the new compounds are presented in the supplemental data.

2.2. In vitro studies

2.2.1. CB1 and CB2 expressing U87 glioblastoma cells lines

To evaluate the specific binding of the aforementioned fluorinated compounds to hCB2, U87 cells expressing either hCB1 (control) or hCB2 were generated via lentiviral transduction. The red fluorescent co-reporter, mKateS158A, was used for cells sorting. The resulting populations were 84% and 98% positive for CB1 and CB2, respectively, (Fig. 6).

Figure 6.

Expression of CB1 and CB2 in transduced U87 cells. U87 cells transduced with hCB1 (upper row) or hCB2 (lower row) and the red fluorescent co-reporter mKateS158A (y-axis), were assessed for expression of hCB1 and hCB2 (x-axis) via flow cytometry, respectively. Density plots are shown for each. Two-step staining was performed with an allophycocyanine (APC) -conjugated goat anti-mouse secondary antibody plus either an isotype control antibody (left column), anti-hCB1 (middle column) or an anti-hCB2 (right column).

2.2.2. In vitro receptor binding assay

This study was performed on the compounds 13-16, 18, 19, 21-24, 27, and 28 using [³H]CP-55,940, [side chain-2,3,4-³H(N)], a radiolabeled CB2 receptor ligand, following a literature method.⁶ The percent inhibition against [³H]CP-55,940 and the Ki values were calculated using GraphPad Prism v4 software. The dissociation constant (Kd) and maximum binding (Bmax) for [³H]CP-55,940 were determined to be 4.9 nM and 791 fM/mg protein, respectively (Fig. 7A), in CB2-expressing cells, and these values were in agreement with previously reported data.⁶ The Kd and Bmax for [³H]CP-55,940 were determined to be 16.8 nM and 1773 fM/mg protein in CB1-expressing cells (Fig. 7B).

Figure 7.

A) Specific binding of [³H]CP-55,940 to CB2-expressing cell membranes, B) Specific binding of [³H]CP-55,940 to CB1-expressing cell membranes, C) competitive inhibition of [³H]CP-55,940 binding and Ki values for compounds 14-16, 18, 19, and 21.

Compounds 14-16, 18, 19, and 21 demonstrated high binding efficiency to CB2-expressing cell membrane preparations, as evidenced by inhibition of binding of [³H]CP-55,940 (Fig. 7C). Other compounds, 13, 22-24, 27, and 28, did not show good binding for either CB2 or CB1; in particular, compounds 13 and 22 did not have any affinity for the CB2 receptor. Compounds 23, 24, 27, and 28 demonstrated low levels of binding affinity for the CB2 receptor. Among these compounds, 23, 24, and 27, had better affinity than 28. Compounds 14-16, 18, 19, and 21 did not have any significant competitive binding to CB1 receptor-expressing cell membranes (Ki > 10 μM), suggesting that these compounds are not CB1-specific receptor ligands. By contrast, the Ki values of some of these compounds for the CB2 receptor are in the low nanomolar and sub-nanomolar range, which suggests that they have high affinity to the CB2 receptor. In terms of selectivity, compounds 14, 15, 16, 18, 19, and 21 have better selectivity for CB2, and 19 is the most selective one. Compound 21, with a nitro substitution at the C₆-position, did not affect any binding efficiency to the CB2 receptor, but the substitution increased its specificity for CB2. The Ki values for CB2 and CB1 for compounds 13-16, 18, 19, 21-24, 27, and 28 and their selectivity for CB2 are summarized in Table 1.

Table 1.

Inhibition constant (Ki) values of compounds 13-16, 18, 19, 21-24, 27, and 28 using [³H]CP-55,940 a radioligand for hCB1 and hCB2 cannabinoid receptors.

Comp	R1	R2	R3	R4	R5	hCB2 Ki, nM	hCB1 Ki, μM	Selectivity CB2/CB1
13	Methyl	H		CH3	H	>10000	>10	>1
14	Butyl	H		CH3	H	2.8±0.4	>10	>3570
15	Methyl	H		CH3	H	5.0±1.1	>10	>2000
16	Butyl	H		CH3	H	2.4±0.7	>10	>4166
18	Methyl	H		CH3	H	22±4.1	>10	>454
19	Butyl	H		CH3	H	0.8±0.3	>10	>12500
21	Butyl	NO2		CH3	H	1.4±0.4	>10	>7143
22	Methyl	H		CH3	H	>10000	>10	>1
23	Butyl	H		CH3	H	486±96	>10	>20.5
24	Butyl	H		CH3	H	58±14	>10	>172
27	Butyl	H		CH3	H	620±114	>10	>16
28	Butyl	H		CH3	H	2400±680	>10	>4

Multiple 2-oxo-quinoline derivatives containing various substituents at the R₁-, R₂-, and R₃-position have been previously reported^34,35 (Table 2, supplemental data), which showed that most of the compounds had Ki values of sub-nanomolar concentrations, ^{35, 44} much lower than those of our compounds. For example, compounds 29-31, 41-43 and 46-49 have Ki values 0.01-0.08 nM, which is > 10-fold lower than we found for our compounds. However, the Ki value of compound 29 was also reported to be 35.9 nM in another report ¹ as opposed to 0.087 nM in separate reports.^{35, 44} Furthermore, the Ki value of the potent CB2 specific indole derivative L-768242 has been reported to be 14 nM; ²³ and those of the traditional tetrahydrocannabibnoids (THC) JWH229 and JWH133 were reported to be 18.0 and 3.4 nM, respectively.²¹ All these Ki values for the indole derivative and traditional THC derivatives are comparable to those of our fluorinated 2-oxoquiniline derivatives. For further confirmation, we performed an in vitro binding assay of compound 20, which has an amino group at the para-position of the aromatic ring and a butyl group at the C₈-position, similar to the reported compound 39, which has a pentyl group at C₈.³⁴ The Ki value of 20 was 22.3 nM, comparable to that of our fluorinated compound 18 (22 nM). In one paper, the binding efficiency of 39 was reported in terms of −Log IC₅₀; we have also calculated the −Log IC₅₀ of compound 20 for comparison. The reported −LogIC₅₀ of 39 is 8.5±0.3, and that of our compound 20 is 8.8±0.4, which are comparable to those reported previously.³⁴ Therefore, Ki values of our compounds are reliable and consistent with previously reported results.

2.2.3. Structure-activity relationship

As described above (Table 1 and Fig. 7C), compounds 14-16, 18, 19 and 21 demonstrated high binding affinity in the low-to-subnanomolar concentration range and high selectivity to the CB2 receptor. Compounds 13 and 22 did not show any binding affinity for CB2; however, compounds 23, 24, 25, and 28 had some low levels of binding affinity for CB2. Compounds 13 and 14, with a 4-fluoro-benzyl amide group (R₃) have in common a one-carbon chain between the amide bond and aromatic ring; only alkyl substituents (methyl vs. butyl) at the C₈-position makes a significant difference in binding efficiency (Ki >10000 vs. 2.8 nM) between these two compounds. This suggests that the larger group at the C₈-position enhances the binding affinity. By contrast, compounds 15 and 16, with a two-carbon chain between the amide and the aromatic ring and a different alky substituent at the C₈-position, did not show a significant difference in binding efficiency. Therefore, the size of the substituent at C₈-position makes a significant difference in binding to CB2 for a compound with a methylene linker between the amide bond and phenyl ring. It has been reported that an ethylene linker between the amide bond and the phenyl ring (R₃) had 100-fold increase in potency compared with a methylene linker.³⁴ In the current series of newly synthesized fluorinated compounds, no such significant difference in binding efficiency was observed due to the different carbon-chain (methylene vs. ethylene) between the amide and aromatic ring. For example, the Ki values of compounds 14 with a methylene and 16 with an ethylene linker but a butyl group at C₈ are comparable (2.8 vs. 2.4 nM). A large difference in binding was observed between the methylene and ethylene linker only when the size of the substituent at the C₈-position was changed from the methyl to the butyl group.

The position of fluorine (in the aromatic ring vs. side chain) also made a difference in binding efficiency. For example, compound 19 had a much higher binding efficiency (Ki=0.8 nM) than compound 16 (Ki=2.4 nM). The major difference between these two compounds is the position of fluorine: in the side chain rather than in the aromatic ring. Compounds with fluorine in the side chain and different substituents at the C₈-position had significantly different binding efficiency; for example, when the R₁ was changed from the butyl to methyl group, the Ki increased from 0.8 nM for 19 to 22 nM for 18. This observation suggests that a larger group at the C₈-position (R₁) has better binding capability than a smaller group, irrespective of the other substituents, possibly due to the lipophylic nature of these compounds. Fluorine on the side chain rather than in the aromatic ring may also play a significant role in the π-stacking interactions between the ligand and aromatic amino acids of the receptor. Compounds with a 4-substituted-phenylethyl amide (R₃) with pentyl substituent at C₈-position (R₁) have been developed and reported earlier.^{34, 35} The presence of an electron withdrawing nitro group at the para-position of the aromatic ring has been shown to decrease potency through diminishing aromatic π-stacking interactions compared with other substituents at the para-position.³⁴ Thus, in our compounds, the presence of electronegative fluorine at the para-position of the aromatic ring may withdraw electron density of the π-electrons, reduce the stacking interaction, and diminish the potency of the compound compared with a fluorine in the benzylic position, which is consistent with the results in the literature.³⁴ As a result, compound 19 is 3-fold more potent than 16. Our results suggest that substitution at the C₆-position improves selectivity for CB2 over CB1. Compound 21 has a nitro group at the C₆-postion, and this substitution makes the compound more selective for CB2, maintaining a low Ki value. This result is also consistent with the literature, because other C₆-substituted compounds 30, 41, and 42 (Table 2) have been reported to be highly selective and sensitive CB2 ligands.^{35, 44}

Substitution of methyl for butyl at the C₈-position makes a significant difference in binding efficiency for the methylene-containing compounds (13 and 14), but no such difference was observed in the ethylene-containing compounds 15 and 16. By contrast, there was a small difference in binding efficiency between the compounds containing a methylene and ethylene linker with a butyl group at the C₈-position, i.e. an ethylene-containing compound had slightly better binding efficiency than those with a methylene-containing compound. There was a general trend that compounds containing a C₈-butyl substituent have better binding efficiency than those containing C₈-methyl. A nitro group at the C₆-position in compound 21 had a small influence that improved the binding efficiency compared with that of the C₆-unsubstituted compound 16.

Substitution at the amide bond (R₃) makes a significant difference in binding efficiency of these 2-oxo-quinoline derivatives. Compounds with a straight chain substituent, such as fluoroethyl, showed no binding affinity; however, binding affinity increased with increasing carbon-chain length, which is consistent with the hydrocarbon-substituted compounds reported previously.³⁴ To further explore the structure activity relationship with regard to π-stacking and charge or polarity on the R₃, we synthesized triazol-containing derivatives of 2-oxoquiniline. If these compounds had demonstrated improved binding characteristics to CB2, the rapid radiolabeling of these compounds could be achieved with click chemistry approaches using an alkyne derivative of 2-oxoquinoline and ¹⁸F-labeled fluoroethylazide. However, the binding efficiency of these compounds was quite low. Thus, compounds 27 and 28 have Ki values of 620 and 2400 nM, respectively. These results suggest that substitution at the amide bond is critical for the binding efficiency of these compounds. In terms of specificity, compound 19 has the highest specificity for the CB2 receptor, and compounds 21, 16, and 14 are the next selective compounds for the CB2 receptor. Thus, given their binding efficiency and specificity to the CB2 receptor, the selected compounds 14, 15, 16, 19, and 21 require further in vivo studies in animal models.

Previously, a few ligand docking, mutation, and modeling studies have been reported for the CB2 receptor, that have used the structure of rhodopsin as a template to construct a model of the CB2 receptor for use in docking studies.⁴⁵ Based on docking studies with CB2 selective ligands, it was reported that hydrophobic and aromatic stacking interactions were more important than hydrogen bonding for CP-55,940 and WIN-55,212-2, and numerous aromatic stacking interactions were observed for the binding of WIN-55,212-2 at the CB2 receptor site.⁴⁶ It appears that hydrophobic and aromatic π-stacking interactions play similar roles in the newly synthesized library of compounds in binding with the CB2 receptor, and our results are in agreement with those previously reported for CP-55,940 and WIN-55,212-2.⁴⁶

3. Conclusion

We have synthesized and tested a library of new fluorinated 2-oxoquinoline derivatives as CB2 receptor ligands. The results of our in vitro competitive radioligand binding studies suggest that some of these compounds are highly active and specific CB2 receptor ligands in the low nanomolar to sub-nanomolar affinity range and no significant binding to CB1 receptors. Therefore, these compounds may be useful for therapy of neuropathic pain, inflammation, and immune disorders and should be evaluated further in corresponding disease models in animals. In addition, compound 19 could be radiolabeled with ¹⁸F and explored for PET imaging of CB2 expression. The studies are currently in progress.

4. Experimental

4.1. Chemistry

4.1.1. Reagents and instrumentation

All reagents and solvents were purchased from Aldrich Chemical Co. (Milwaukee, WI), and used without further purification unless otherwise specified. Tritiated CP-55,950 ([³H]CP-55,940), with a specific activity of 174.6 mCi/μmol was purchased from PerkinElmer (USA). Thin-layer chromatography (TLC) was performed on pre-coated Kieselgel 60 F254 (Merck, Darmstadt, Germany) glass plates. Proton, ¹³C, and ¹⁹F NMR spectra were recorded on a Brucker 300 MHz spectrometer or a 600 MHz spectrometer using tetramethylsilane as an internal reference and hexafluorobenzene as an external reference, respectively, at The University of Texas M. D. Anderson Cancer Center. High-resolution mass spectra were obtained on a Brucker BioTOF II mass spectrometer at the University of Minnesota using the electrospray ionization (ESI) technique.

HPLC was performed with an 1100 series pump (Agilent, Germany), with a built-in UV detector operated at 214 or 254 nm, and a radioactivity detector with single-channel analyzer (Bioscan, Washington DC) with a C₁₈ reverse-phase analytical column (4.6×250 mm; Alltech, Econosil). An acetonitrile/water (MeCN/H₂O) solvent system with various composition (depending on compound), was used for quality control analyses at a flow of 1 mL/min.

4.1.2. Chemical synthesis

4.1.2.1. Preparation of 7-methoxy-8-alkoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-(4-fluorobenzyl)amide: 13, 14

Both compounds 13 and 14 were synthesized following the same methodology; a representative method is described here. To a solution of the 4-fluorobenzylamine (0.187 mmol) in DCM (3.0 mL) was added triethylamine (37 μL), and the mixture was cooled to 0°C. The acid chloride 11 or 12 (0.280 mmol) in DCM (0.5 mL) was added slowly and the mixture was stirred for 10 min; then the cooling bath was removed, and the mixture was stirred for an additional 20 min at RT. The solvent was removed under reduced pressure and the residue was purified by flash chromatography on a silica gel column using 5% MeOH/CH₂Cl₂. The solvent was evaporated on a rotary evaporator, and the product 13 was obtained as a white solid (powder) in 76% yield. ¹H NMR (CDCl₃) δ:10.04 (t, 1H, NH), 9.36 (s, 1H, NH), 8.92 (s, 1H, C₄-H), 7.50 (d, J=8.4 Hz, 1H, C₅-H), 7.36 (m, 2H, aromatic), 7.03 (m, 2H, aromatic), 6.97 (d, J=8.4 Hz, 1H, C₆-H), 4.66 (d, J = 6.6 Hz, 2H, N-CH₂), 4.02 (s, 3H, OMe), 3.99 (s, 3H, OMe). ¹³C NMR (CDCl₃) δ: 163.66, 162.28, 162.05 (d, J_1-F=243.8 Hz), 154.51, 145.13, 134.47, 134.45, 133.28 (d, J_4-F=5.0 Hz), 129.33 (d, J_3-F=8.8 Hz), 125.63, 119.49, 115.38 (d, J_2-F=21.3 Hz), 114.29, 109.11, 61.08, 56.32, 42.81.¹⁹F NMR (CDCl₃) δ: -114.66 (septet,AA’MM’X, J =15.0 Hz, J=9.0 Hz, J=6.0 Hz, J=3.0 Hz). High-resolution MS: calculated for C₁₉H₁₇FN₂O₄ (M+H) 356.1172, found 356.1169.

Compound 14 was purified on a silica gel column eluted with 70% EtOAc/Hexane. The product 14 was obtained as a white solid in 50% yield. ¹H NMR (CDCl₃) δ: 8.55 (s, 1H, C₄-H), 7.60 (d, J=8.4 Hz, 1H, C₅-H), 7.42–7.37 (m, 3H, aromatic), 7.10-7.04 (m, 3H, aromatic), 4.70 (d, J=6.6 Hz, 2H, NCH₂), 4.27 (t, J=6.6 Hz, 2H, OCH₂), 4.11 (s, 3H, OMe), 4.10 (s, 3H, OMe), 1.85–1.80 (m, 2H), 1.58–1.55 (m, 2H), 0.98 (t, J=7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 163.67, 162.21, 162.05 (d, J_1-F=243.8 Hz), 154.51, 145.13, 134.47, 134.45, 133.28 (d, J_4-F=5.0 Hz), 129.33 (d, J_3-F=8.8 Hz), 125.63, 119.49, 115.38 (d, J_2-F=21.3 Hz), 114.29, 109.13, 73.54, 56.27, 42.79, 32.21, 32.21, 19.13, 13.82. ¹⁹F NMR (CDCl₃) δ: -114.4 6 (septet,AA’MM’X, J=15.0 Hz, J=9.0 Hz, J=6.0 Hz, J=3.0 Hz). High-resolution MS: calculated for C₂₂H₂₃FN₂O₄ (M+H) 398.1642, found 398.1649.

4.1.2.2. Preparation of 7-methoxy-8-alkoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-(4-fluorophenylethyl)amide: 15, 16

Both compounds 15 and 16 were synthesized following the same methodology; a representative method is described here. To a solution of the 4-fluorophenyl ethylamine (0.187 mmol) in DCM (3.0 mL) was added triethylamine (37 μL), and the mixture was cooled to 0°C. The acid chloride 11 or 12 (0.280 mmol) in DCM (0.5 mL) was added slowly and the mixture was stirred for 10 min; then the cooling bath was removed, and the mixture was stirred for an additional 20 min at room temperature. The solvent was removed under reduced pressure, and the residue was purified by flash chromatography on a silica gel column using 5% MeOH/CH₂Cl₂. The solvent was evaporated on a rotary evaporator and the product 15 was obtained as a white solid in 86% yield. ¹H NMR (CDCl₃) δ: 9.70 (s, 1H, NH), 9.32 (s, 1H, NH), 8.88 (s, 1H, C₄-H), 7.50 (d, J = 8.4 Hz, 1H, C₅-H), 7.22–7.28 (m, 2H, aromatic), 6.95-7.05 (m, 3H, aromatic), 4.01 (s, OCH₃), 3.99 (s, 3H, OCH₃), 3.72 (q, 2H, J=7.2 Hz, N-CH₂), 2.94 (t, J=7.2 Hz, 2H, benzylic). ¹³C NMR (CDCl₃) δ: 163.66, 162.28, 162.05 (d, J_1-F=243.8 Hz), 154.51, 145.13, 134.47, 134.45, 133.28 (d, J_4-F=5.0 Hz), 129.33 (d, J_3-F=8.8 Hz), 125.63, 119.49, 115.38 (d, J_2-F=21.3 Hz), 114.29, 109.11, 61.08, 56.32, 56.2, 42.81. High-resolution MS: calculated for C₂₀H₁₉N₂FNaO₄ (M+Na) 393.1227, found 393.1244.

Compound 16 was obtained in 96% yield. ¹H NMR (CDCl₃) δ: 9.65 (s, 1H, NH), 9.36 (s, 1H, NH), 8.85 (s, 1H, C₄-H), 7.43 (d, J=8.4 Hz, 1H, C₅-H), 7.22–7.20 (m, 2H, aromatic), 6.96-7.00 (m, 2H, aromatic), 6.92 (d, J=9.0 Hz, 1H, C₆-H), 4.12 (t, J=6.6 Hz, 2H, OCH₂), 3.95 (s, 3H, OMe), 3.70–3.67 (m, 2H, N-CH₂), 2.90 (t, J=7.2 Hz, 2H, benzylic), 1.80–1.75 (m, 2H), 1.52–1.45 (m, 2H), 0.97 (t, J=7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 163.5, 162.4, 162.2, 160.7, 154.4, 144.9, 134.8 (2C), 133.4, 132.4, 130.2 130.1, 125.2, 119.3, 115.3, 115.2, 114.3, 109.1, 73.5, 56.2, 41.0, 34.9, 32.1, 19.1, 13.8. ¹⁹F NMR (CDCl₃) δ: -116.9. High-resolution MS: calculated for C₂₃H₂₅N₂FNaO₄ (M+Na) 435.1718, found 435.1788.

4.1.2.3. Preparation of 2-fluoro-2-phenylethylamine: 17

Compound 17 was prepared following a literature method ⁴¹ with modifications. Briefly, to a stirred solution of mandelonitrile (1.33 g, 1.0 mmol, 1.0 equiv) in dry DCM (5 mL) at 0°C was added diethylaminosulfure trifluoride (DAST) (1.77 g, 1.1mmol, 1.1 equiv) in DCM (2 ml). The resulting mixture was stirred at 0°C and stirred for 30 min, diluted with DCM (15 ml), and treated with saturated NaHCO₃. The organic layer was collected and dried over MgSO₄. After removal of the solvent under vacuum, the crude product was chromatographed on a silica gel column and eluted with 10% EtOAc/hexane to afford α-fluorophenylacetonitrile as yellow oil (0.88 g) in 65% yield. The ¹H NMR spectrum was consistent with the literature, ⁴¹ and ¹⁹F NMR showed a doublet at -167.4 ppm with J=48 Hz. Finally, the cyano-group was reduced to the corresponding amine by treatment with 1M borane-THF solution in THF at 0°C for 40 min and quenched the reaction mixture with ethanolic hydrochloric acid, then concentrated under reduced pressure. The desired 2-fluorophenylethyl amine was triturated with CH₃CN and then recrystalized from ethanol/acetonitrile to afford 17 as colorless crystals in 49% yield. ¹H NMR (CDCl₃) δ: 7.50 (m, 5 H, aromatic), 5.8 (dddd, J=48.0 Hz, 8.4 Hz, 7.2 Hz,4.5 Hz, 3.0 Hz, 1 H, benzylic), 3.50 (m, 2 H, methylene). ¹⁹F NMR (CDCl₃) δ: -183 (dm). MS: 154 (M+H).

4.1.2.4. Preparation of 7-methoxy-8-alkoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-(2-fluoro-2-phenyl ethyl)amide: 18, 19

Both compounds 18 and 19 were prepared using the same methodology, a representative method is described here. To a solution of the 2-fluoro-2-phenylethylamine (0.187 mmol) in DCM (3.0 mL) was added triethylamine (37 μL), and the mixture was cooled to 0°C. The acid chloride 11 or 12 (0.280 mmol) in DCM (0.5 mL) was added slowly and the mixture was stirred for 10 min; then the cooling bath was removed and the mixture was stirred for an additional 20 min at RT. The solvent was removed under reduced pressure, and the residue was purified by flash chromatography on a silica gel column using 5% MeOH/CH₂Cl₂. The solvent was evaporated on a rotary evaporator, and the product 18 was obtained as a white solid (powder) in 75% yield. ¹H NMR (CDCl₃) δ: 10.00 (t, J= 6 Hz, 1H, NH), 9.32 (s, 1H, NH), 8.89 (s, 1H, C₄-H), 7.35-7.55 (m, 6H, aromatic H & C₅-H), 6.97 (d, J=9.0 Hz, 1H, C₆-H), (dddd, J=48.0 Hz, 8.4 Hz, 7.2 Hz,4.5 Hz, 3.0 Hz, 1 H, benzylic), 4.05-4.20 (m,1H, N-CH2), 4.03 (s, 3H, OMe), 4.00 (s, 3H, OMe), 3.58-3.84 (m, 1H, N-CH2). ¹³C NMR (CDCl₃) δ: 163.4, 162.1, 154.1, 144.7, 144.6, 133.4, 132.3, 129.5 (2C), 129.3, 125.2, 120.0, 115.7 (2C), 114.2, 109.0, 73.7, 56.3, 41.4, 35.0, 29.8, 28.1, 22.4, 13.8. ¹⁹F NMR (CDCl₃) δ: -182.71 (dt, J=47.6 Hz, J=30.8 Hz, J=14.0 Hz). High-resolution MS: calculated for C₂₀H₁₉N₂FNaO₄ (M+Na) 393.1227, found 393.1258.

Compound 19 was purified on silica gel column using 70% EtOAc/Hexane. The product 19 was obtained as a white solid in 50% yield. ¹H NMR (CDCl₃) δ: 10.00 (t, J= 6 Hz, 1H, NH), 9.17 (s, 1H, NH), 8.88 (s, 1H, C₄-H), 7.28-7.49 (m, 6H, aromatic & C₅-H), 6.95 (d, J=9.0 Hz, 1H, C₆-H), 5.70 (dm, J_HF=48 Hz 1H, benzylic), 4.14-4.19 (m, 3H, O-CH₂ & N-CH₂), 3.99 (s, 3H, O-CH₃), 3.65-3.85, (m, 1H, N-CH2), 1.76-1.84 (m, 2H), 1.50–1.57 (m, 2H), 1.02 (t, J = 7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 163.96, 162.02, 154.45, 145.06, 137.78, 137.63, 133.56, 132.42, 128.67, 128.37, 125.64, 125.58, 125.36, 119.35, 114.2, 109.09, 92.71 (d, J_1-F=172.5 Hz), 73.56, 56.30, 45.40 (d, J_2-F=24.0 Hz), 32.26, 19.15, 13.8. ¹⁹F NMR (CDCl₃) δ: -183.11 (dt, J=47.6 Hz, J=30.8 Hz, J=14.0 Hz). High-resolution MS: calculated for C₂₃H₂₅N₂FNaO₄ (M+Na) 435.1696, found 435.1702.

4.1.2.5. Preparation of 7-methoxy-8-butyloxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid 4-amino-phenethylamide: 20

Compound 20 was prepared by similar method as described for compounds 13-16 (scheme 1). Briefly, the acid chloride 12 (1.0 equiv) in anhydrous DCM was added drop-wise to a solution of 4-aminophenylethylamine (1.5 equiv) and triethylamine (2.0 equiv) in dry DCM at 0°C. The ice bath was removed in 15 min and the reaction mixture stirred at room temperature (RT) for 16 h. The reaction was quenched in a brine solution and the product was extracted with DCM, dried over MgSO₄, and then concentrated under vacuum at ambient temperature. The crude residue was chromatographed on a silica gel column eluting with 10% MeOH/CH₂Cl₂ to give the desired product 20 in 96% yield. ¹H NMR (CDCl₃) δ: 9.67 (s, 1 H), 9.13 (s, 1 H), 8.88 (s, 1H), 7.48 (d, J=8.4 Hz, 1 H), 7.10 (d, J=7.8 Hz, 2 H), 6.95 (d, J=8.4 Hz, 1 H), 6.69 (d, J=7.2, 2 H), 4.16 (t, J=6.6 Hz, 2 H), 4.00 (s, 3 H), 3.70 – 3.69 (m, 2 H), 3.61 (bs, 2 H), 2.86 (t, J=7.2 Hz, 2 H), 1.83 – 1.80 (m, 2 H), 1.55 – 1.52 (m, 2 H), 1.02 (t, J=7.2 Hz, 3 H). ¹³C NMR (CDCl₃) δ: 163.4, 162.1, 154.1, 144.7, 144.6, 133.4, 132.3, 129.5 (2C), 129.3, 125.2, 120.0, 115.7 (2C), 114.2, 109.0, 73.7, 56.3, 41.4, 35.0, 29.8, 28.1, 22.4, 13.8. MS: calculated for C₂₂H₂₅N₃O₄ (M+Na) 395.1643, found 395.1648.

4.1.2.6. Preparation of 7-methoxy-8-butyloxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-(4-fluorphenethyl)amide and 6-nitro-7-methoxy-8-butyloxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-(4-fluorphenethyl)amide: 16, 21

Compound 16 was prepared by two different methods.

Method 1: This method is a direct coupling of the 4-fluorophenyl ethylamine with the acid chloride 12, similar to the preparation of other compounds, 13-16. Method 2: To a solution of 7-methoxy-8-butyloxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid p-amino-phenethylamide 20 (20 mg, 1 equiv) in 125 μL of pyridine•HF was added 5.5 mg of sodium nitrite (0.08 mmol, 1.5 equiv). The reaction mixture was stirred at RT for 1 h and then at 85°C for an additional hour. The reaction mixture was quenched with ice-water (10 mL), and the product was extracted with ethyl acetate, dried over MgSO₄, and then the solution was concentrated under vacuum. The crude residue was chromatographed on silica gel column eluting with 2:1 EtOAc/hexane to give 8.0 mg of the desired product 16 (40% yield) and 8 mg of its 6-nitro derivative 21 (36% yield). The NMR and mass spectra of 16 produced by this method were identical with those of 16 obtained from method 1.

Compound 21 was isolated as a by-product in 36% yield during the preparation of 16 using method 2. ¹H NMR (CDCl₃) δ: 9.63 (s, 1H, NH), 9.48 (s, 1H, NH), 9.29 (s, 1H, C₄-H), 7.76 (s, 1H, C₅-H), 7.22 – 7.19 (m, 2H, aromatic), 6.98 (t, J=8.4 Hz, 2H, aromatic), 4.32 (t, J=7.2 Hz, 2H, O-CH₂), 4.03 (s, 3H, OMe), 3.71 – 3.68 (m, 2H, N-CH₂), 2.91 (t, J=7.2 Hz, 2H, benzylic), 1.81 – 1.76 (m, 2H), 1.51 – 1.45 (m, 2H), 0.98 (t, J=7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 162.4 (2C), 160.8 (2C), 151.3, 141.2, 140.1, 136.4, 134.7 (2C), 133.4 130.2, 130.1, 123.2, 115.3, 115.2, 108.0, 107.3, 74.2, 56.8, 41.2, 34.8, 29.7, 19.0, 13.7. ¹⁹F NMR (CDCl₃) δ: -116.8. High-resolution MS: calculated for C₂₃H₂₄N₃FO₆Na (M+Na)480.1718, found 480.1593.

4.1.2.7. Preparation of 7-methoxy-8-alkoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-(fluoroalkyl)amide: 22-24

Compounds 22-24 were synthesized following the same methodology as described in Scheme 4; a representative method is described here. To a solution of the 2-fluoroethylamine (0.187 mmol) in DCM (3.0 mL) was added triethylamine (37 μL), and the mixture was cooled to 0°C. The acid chloride 11 or 12 (0.280 mmol) in DCM (0.5 mL) was added slowly, the mixture was stirred for 10 min; then the cooling bath was removed, and the mixture was stirred for an additional 30 min at RT. The solvent was removed under reduced pressure, and the residue was purified by flash chromatography on a silica gel column using 5% MeOH/CH₂Cl₂. The solvent was evaporated on a rotary evaporator and the product 22 was obtained as a white solid (powder) in 90% yield. ¹H NMR (CDCl₃) δ: 9.94 (bt, 1H, NH), 9.55 (s, 1H, NH), 8.88 (s, 1H, C₄-H), 7.48 (d, J=9.0 Hz, 1H, C₅-H), 6.97 (d, J=9.0 Hz, 1H, C₆-H), 4.65 (dt, J_1-F = 47.4 Hz, J₂=4.8 Hz, 2H, F-CH₂), 3.99 (s, 6H, OMe), 3.85 (q, J₁=10.5 Hz, J₂ = 5.4 Hz, 1H, N-CH₂), 3.76 (q, J₁=10.5 Hz, J₂=5.4). ¹³C NMR (CDCl₃) δ: 164.01, 162.26, 154.55, 145.03, 133.40, 133.30, 125.63, 119.34, 114.23, 109.03, 82.45 (d, J_1-F=161.3 Hz), 61.10, 56.32, 40.00 (d, J_3-F=21.1 Hz). ¹⁹F NMR (CDCl₃) δ: -223.06 (m). High-resolution MS: calculated for C₁₄H₁₅FN₂O₄Na (M+Na) 317.0914, found 317.0909.

Compound 23 was also purified on a silica gel column eluted with 70% EtOAc/Hexane and obtained as a white solid in 92% yield. ¹H NMR (CDCl₃) δ: 9.91 (bt, 1H, NH), 9.17 (s, 1H, NH), 8.88 (s, 1H, C₄-H), 7.47 (d, J=9.0 Hz, 1H, C₅-H), 6.96 (d, J=9.0 Hz, 1H, C₆-H), 4.60 (dt, J_1-F=47.6 Hz, J₂=5.7 Hz, 2H, F-CH₂), 4.17 (t, J=6.6 Hz, 2H, OCH₂), 3.99 (s, 3H, OMe), 3.62 (q, J₁=12.9 Hz, J₂ = 6.9 Hz, 2H, N-CH₂), 1.80 (m, 2H), 1.58 (m, 2H), 0.98 (t, J = 7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 164.00, 162.08, 154.45, 145.03, 133.55, 132.38, 125.36, 119.28, 114.20, 108.07, 82.43 (d, J_1-F=168.3 Hz), 73.54, 56.28, 39.97(d, J_2-F=21.1 Hz), 32.21, 19.13, 13.83. ¹⁹F NMR (CDCl₃) δ: -223.18 (m). High-resolution MS: calculated for C₁₇H₂₁FN₂O₄ Na (M+Na) 336.1485, found 336.1485.

Compound 24 was obtained as a white solid in quantitative yield after purification. ¹H NMR (CDCl₃) δ: 9.75 (bt, 1H, NH), 9.16 (s, 1H, NH), 8.87 (s, 1H, C₄-H), 7.47 (d, J=9.0 Hz, 1H, C₅-H), 6.95 (d, J=9.0 Hz, 1H, C₆-H), 4.60 (dt, J_1-F=47.6 Hz, J₂=5.7 Hz, 2H, F-CH₂), 4.17 (t, J=6.6 Hz, 2H, OCH₂), 3.99 (s, 3H, OMe), 3.62 (q, J₁=12.9 Hz, J₂=6.9 Hz, 2H, N-CH₂), 2.13-2.00 (m, 2H), 1.81 (m, 2H), 1.52 (m, 2H), 1.00 (t, J=7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 169.79, 162.15, 154.36, 144.36, 133.45, 132.38, 125.32, 119.43, 114.26, 109.07, 82.01 (d, J_1-F=164.3 Hz), 73.54, 56.27, 35.92 (d, J_3-F=5.2 Hz), 32.22, 30.55 (d, J_2-F=19.6 Hz), 19.13, 13.83. ¹⁹F NMR (CDCl₃) δ: -226.82 (m). High-resolution MS: calculated for C₁₈H₂₃FN₂O₄Na (M+Na) 373.1540, found 373.1573.

4.1.2.8. Preparation of 7-methoxy-8-butoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-(alkyl-alkyne)amide: 25, 26

Compounds 25 and 26 were prepared by the same method as described in Scheme 5 following a previously published method.⁴² The acid chloride 12 (0.280 mmol) in DCM (0.5 mL) was added slowly to a solution of either 2-propyne amine or 3-butyne amine (0.187 mmol) in DCM (3.0 mL), which was cooled to 0°C, and the mixture was stirred for 10 min. The cooling bath was removed, and the reaction mixture was stirred for an additional 3 h at RT. The solvent was removed under reduced pressure, and the residue was purified by flash chromatography on a silica gel column using 70% EtOAc/Hexane. Compound 25 was obtained in quantitative yield; its ¹H NMR spectrum was consistent with that reported in the literature.⁴²

Compound 26 was obtained as a white solid in quantitative yield. ¹H NMR (CDCl₃) δ: 9.90 (bt, 1H, NH), 9.18 (s, 1H, NH), 8.87 (s, 1H, C₄-H), 7.46 (d, J=9.0 Hz, 1H, C₅-H), 6.95 (d, J=9.0 Hz, 1H, C₆-H), 4.15 (t, J=6.9 Hz, 2H, OCH₂), 3.99 (s, 3H, OMe), 3.65 (q, J₁=12.9 Hz, J₂=6.6 Hz, 2H, N-CH₂), 2.54 (ddd, J=9.3 Hz, J = 6.9 Hz, J=2.7 Hz, 2H), 2.06 (t, J=2.7 Hz, 2H), 1.81 (m, 2H), 1.52 (m, 2H), 1.00 (t, J = 7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 163.67, 162.11, 154.37, 144.93, 133.47, 132.36, 125.33, 119.44, 114.25, 109.06, 81.70, 73.54, 69.76, 56.28, 38.36, 32.24, 19.51, 19.13, 13.83. High-resolution MS: calculated for C₁₉H₂₂N₂O₄Na (M+Na) 365.1477, found 365.1475.

4.1.2.9. Preparation of 2-fluorethylazide

This compound was prepared according to a previously reported method.⁴³ Briefly, 2-fluoroethyl bromide (0.028 mmol) was reacted with sodium azide (0.029 mmol) in DMF (10 mL) for 48 h at RT. The solution was filtered and used without further purification.

4.1.2.10. Preparation of 7-methoxy-8-butoxy-2-oxo-1,2-dihydroquinoline-3-carboxylic acid-((1-(2-fluoroethyl)-1H-1,2,3-triazol)alkyl)amide: 27, 28

Compounds 27 and 28 were prepared from 25 and 26, respectively, following a method similar to that described in scheme 5. A representative procedure is described here. To a solution of fluoroethylazide (1.5 mL, 2.8 mM, in DMF) was added propargyl amide 25. In a separate flask was dissolved copper (II) sulfate pentahydrate (25.0 mg, 0.10 mmol) in water (1.0 mL) followed by addition of sodium ascorbate (56 mg, mmol). This solution was transferred to the above reaction mixture and stirred at RT for 12 h. The solvent was removed under a stream of air; the residue was redissolved in DCM and subjected to column chromatography using 10% MeOH/CH₂Cl₂. The solvent was evaporated on a rotary evaporator and the product 27 was obtained as a white solid in 55% yield. ¹H NMR (CDCl₃) δ: 10.00 (bt, 1H, NH), 9.14 (s, 1H, NH), 8.89 (s, 1H, C₄-H), 7.58 (s, 1H, triazole-H), 7.49 (d, J=9.0 Hz, 1H, C₅-H), 6.60 (d, J=9.0 Hz, 1H, C₆-H), 4.88 (t, J=4.7 Hz 1H, N-CH), 4.81 (m, 2H), 4.70 (m, 2H), 4.60 (m, 1H, F-CH), 4.16 (t, J=6.6 Hz, 2H, OCH₂), 3.99 (s, 3H, OMe), 1.81 (m, 2H), 1.52 (m, 2H), 1.00 (t, J=7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 163.84, 161.93, 154.45, 145.84, 145.03, 133.54, 132.39, 125.35, 123.20, 119.34, 114.19, 109.09, 81.51 (d, J_1-F=171.0 Hz), 73.55, 56.29, 50.50 (d, J_2-F=20.0 Hz), 35.26, 32.22, 19.12, 13.33. ¹⁹F NMR (CDCl₃) δ: -221.6 (m). High-resolution MS: calculated for C₂₀H₂₄FN₅O₄Na (M+Na) 440.1710, found 440.1694.

Compound 28 was obtained as a white solid in 58% yield. ¹H NMR (CDCl₃) δ: 9.79 (bt, 1H, NH), 9.17 (s, 1H, NH), 8.86 (s, 1H, C₄-H), 7.58 (s, 1H, triazole-H), 7.46 (d, J=9.0 Hz, 1H, C₅-H), 6.95 (d, J=9.0 Hz, 1H, C₆-H), 4.87 (t, J=4.7 Hz 1H, N-CH), 4.70 (m, overlapped, 2H, N-CH-CH-F), 4.60 (m, 1H, F-CH), 4.15 (t, J=6.6 Hz, 2H, OCH₂), 3.98 (s, 3H, OMe), 3.82 (q, J₁=12.9 Hz, J₂=6.9 Hz, 2H, N-CH₂), 3.10 (t, J=6.9 Hz, 2H), 1.81 (m, 2H), 1.52 (m, 2H), 1.00 (t, J=7.2 Hz, 3H, CH₃). ¹³C NMR (CDCl₃) δ: 163.71, 162.07, 154.36, 145.74, 144.86, 133.45, 132.37, 125.31, 122.33, 119.49, 114.24, 109.06, 81.61 (d, J_1-F=172.0 Hz), 73.53, 56.28 50.41 (d, J_2-F=20.0 Hz), 33.04, 32.22, 26.02, 19.12, 13.33. ¹⁹F NMR (CDCl₃) δ: -221.4 (m). High-resolution MS: calculated for C₂₁H₂₆FN₅O₄Na (M+Na) 454.1867, found 454.1869.

4.2. In vitro studies

4.2.1. Engineering and flow cytometric evaluation of human CB1 and CB2-expressing U87 glioblastoma cells

pCR4-TOPO-hCB1 and pPCR-Script Amp SK(+)-hCB2, encoding the human CB1 (accession number NM016083) and hCB2 (accession number BC069722) were purchased from Open Biosystems (Huntsville, AL), respectively. CB1 and CB2 were subcloned into pDONR222 via BP reaction following amplification with attB1 and attB2 flanked primers and then recombined into a Gateway-adapted lentivirus encoding an internal mscv LTR (Ref = http://www.ncbi.nlm.nih.gov/pubmed/20730500) and the red fluorescent protein mKateS158A downstream of an emcv IRES (pLV4312-CB2). Virus was packaged in 293-FT cells using pMD2.G (VSV.G env) and pCMV-deltaR8.91 and concentrated 50x using Amicon Ultra-15 100,000 NMWL centrifugal concentration units (Millipore, Billerica, MA). Concentrated viral supernatants were used to transduce the human glioblastoma cell line U87 (ATTC, Manassas, VA) via spinfection for 2 hr at 2200 RPM/30°C. CB1+/mKateS158A+ and CB2+/mKateS158A+ U87 cells were sorted on a FACSAria cell sorter (BD Biosciences, San Jose, CA) based on expression of the co-reporter mKateS158A. Cell surface expression of hCB1 and hCB2 was assessed via flow cytometry on a FACSCalibur (BD Biosciences, San Jose, CA). Primary antibodies specific for CB1 (Clone 368302) and CB2 (Clone 352114) were obtained from R&D Systems (Minneapolis, MN). Secondary staining was performed using a goat anti-mouse allophycocyanin conjugated polyclonal antibody (Jackson ImmunoResearch, West Grove, PA).

4.2.2. Receptor binding assay

This study was performed on the compounds 13-16, 18, 19, 21-24, 27, and 28 using [³H]CP-55,940, [side chain-2,3,4-³H(N)], a radiolabeled CB2 receptor ligand, as follows. U87 cells (transduced with CB2 receptor) were cultured on Petri-dishes until 80%–90% confluence was achieved. Then cells were washed with PBS, scraped, and centrifuged. The cell pellet was homogenized in the binding buffer (50 mM TRIS-HCl, 5 mM MgCl2, and 2.5 mM EDTA (pH=7.4). The homogenate was divided into aliquots (30-50 μg protein) and then incubated with a fixed concentration of the radiolabeled ligand [³H]CP-55,940 (2.8 nM) and different concentrations of the compounds for 1 h at 4°C. After incubation, aliquots were filtered through GF/C filters, washed with TRIS buffer (4 mL × 5), and radioactivity on the filters was counted in a scintillation counter. The Kd and Bmax for [³H]CP55940 were calculated to be 4.9 nM and 791 fmol/mg protein, respectively. The percent inhibition against [³H]CP-55,940 and the Ki were calculated using GraphPad Prism v4 software. The binding assay for CB1 receptor was performed following the same methodology as that described for the CB2 receptor binding assay.

Supplementary Material

01. Supplemental data.

Experimental details on compounds 2-12, HPLC chromatograms of the new compounds, and a list of compounds with Ki values previously published in the literature (with references), associated with this article can be found in the online version of the article.