Sch35966 is a potent, selective agonist at the peripheral cannabinoid receptor (CB₂) in rodents and primates

Abstract

Background and purpose:

The peripheral cannabinoid receptor (CB₂) is expressed on peripheral immune cells and is thought to have a role in the immunosuppressive effects of cannabinoids. Historically, there have been few potent, CB₂-selective agonists to assess the contribution of CB₂ to this phenomenon. The studies presented here describe the synthesis of 8,10-bis[(2,2-dimethyl-1-oxopropyl)oxy]-11-methyl-1234-tetrahydro-6H-benzo[β]quinolizin-6-one (Sch35966), which binds with low nanomolar potency to CB₂ in both primates and rodents.

Experimental approach:

The affinity, potency and efficacy of Sch35966 and other cannabinoid ligands at CB₂ was assessed using competition binding assays vs [³H]CP55,940, [³⁵S]GTPγS exchange, cAMP accumulation and cell chemotaxis assays.

Key results:

We showed that Sch35966 has >450-fold selectivity for CB₂ binding vs the central cannabinoid receptor (CB₁) in primates (humans and cynomolgus monkeys) and rodents (rats and mice). Sch35966 is an agonist as it effectively inhibited forskolin-stimulated cAMP synthesis in CHO-hCB₂ cells, stimulated [³⁵S]GTPγS exchange and directed chemotaxis in cell membranes expressing CB₂. In all species examined, Sch35966 was more potent, more efficacious and more selective than JWH-015 (a commonly used CB₂-selective agonist).

Conclusions and implications:

Taken together, the data show that Sch35966 is a potent and efficacious CB₂-selective agonist in rodents and primates.

Introduction

The actions of cannabinoids such as Δ⁹-tetrahydrocannabinol (Δ⁹-THC), the endocannabinoids, anandamide and 2-arachidonyl glycerol (2-AG) and synthetic cannabinoids such as HU210 and (CP55,940) are mediated through activation of central cannabinoid receptor (CB₁) and peripheral cannabinoid receptor (CB₂) (Devane et al., 1992; Mechoulam et al., 1995; Sugiura et al., 1995; for review, see Howlett et al., 2002). Activation of either receptor subtype stimulates the inhibition of adenylyl cyclase (Bayewitch et al., 1995; Slipetz et al., 1995) and the activation of mitogen-activated protein kinase (Bouaboula et al., 1995, 1996). Originally cloned from rat brain (Matsuda et al., 1990; Gerard et al., 1991), CB₁ is expressed primarily in the central nervous system and mediates most of the psychotropic and analgesic effects associated with cannabinoid agonists. CB₂ was cloned from rat spleen and promyelocytic leukaemic HL60 cells (Munro et al., 1993) and is highly expressed in peripheral immune cells (Galiegue et al., 1995; Schatz et al., 1997). More recent work suggests that the receptor may also be expressed in neuronal cells involved in pain modulation and/or perception (Hanus et al., 1999; Malan et al., 2002; Hohmann et al., 2004; Ibrahim et al., 2005).

The recreational use of cannabinoids has been linked with diminished immune function (Kaminski, 1996, 1998; Klein et al., 1998) and there are studies suggesting that endocannabinoids may be immunomodulators (Cabral et al., 1995; Lee et al., 1995; Di Marzo et al., 1999). Owing to its expression pattern, the immunosuppressive effects of cannabinoids have been largely attributed to CB₂ activation. However, Smith et al. (2001b) recently showed that intracerebroventricular administration of the non-selective cannabinoid agonists HU210 or WIN55,212 to mice before an endotoxin challenge attenuated proinflammatory cytokine production and increased the levels of the anti-inflammatory cytokine interleukin (IL)-10. Co-administration of a CB₁-selective antagonist, SR141716A attenuated this reaction while the CB₂-selective antagonist SR144528 was ineffective in blocking the response.

Direct measurements of CB₁- and CB₂-specific effects have been hindered by the shortage of isotype-selective agonists. Indeed, until the recent isolation of 2-AG ether from porcine brain, there were no CB₁-selective agonists. 2-AG ether was characterized as an agonist with >140-fold selectivity for the CB₁ vs CB₂. Interestingly, there has been more success generating CB₂-selective agonists. For example, JWH-015 and JWH-133 are aminoalkylindole congeners of Δ⁹-THC that are relatively selective (28 to 200-fold selectivity over CB₁) and potent (K_i=13.8 and 3.4 nM, respectively) agonists (Showalter et al., 1996; Chin et al., 1999). L759633 and L759656 are aminoalkylindole congeners that have also been shown to be selective and potent CB₂ agonists (Ross et al., 1999). A bicyclic compound, HU-308, was recently described as a very selective CB₂ agonist (>400-fold selectivity over CB₁) although it has only moderate affinity for CB₂ (19–27 nM; Hanus et al., 1999). However, the CB₂-selective ligands described above are not widely accessible and until recently, only JWH-015 was available commercially. Data regarding the affinity, selectivity and efficacy of JWH-015 at CB₂ in human, nonhuman primates and rodents are lacking or incomplete.

In the studies described here, we assessed the utility of a benzoquinolizinone compound, 8,10-bis[(2,2-dimethyl-1-oxopropyl)oxy]-11-methyl-1234-tetrahydro-6H-benzo[β]quinolizin-6-one (Sch35966), as a CB₂-selective cannabinoid agonist in both rodents and primates. Using both functional and binding assays, we demonstrated that Sch35966 is both a potent and selective CB₂ agonist in rat, mouse, cynomolgus monkey and human.

Methods

Synthesis of Sch35966

Sch35966 was prepared from 3,5-bis(acetyloxy)benzoyl chloride and 2-(1,1-dimethoxyethyl)piperidine hydrochloride (Friary, 1990). The procedures described were used without further modification. The structure of Sch35966 is shown in Figure 1.

Figure 1.

The structure of Sch35966.

Cloning CB₂ from rat, mouse and cynomolgus monkey

RNA was isolated from lymph node and spleen samples using RNA STAT-60 (Tel-Test, Friendsword, TX, USA) according to manufacturer's protocol. RNA was isolated from cynomolgus blood using the PAXgene blood RNA kit by PreAnalytiX (Hombrechtikon, CH). All samples were DNase treated using DNase 1 (Roche Diagnostics Corp., Indianapolis, IN, USA) before cDNA synthesis. cDNA was generated using SuperScript First Strand Synthesis System for reverse transcription-PCR (Invitrogen, Carlsbad, CA, USA). Both oligo-dT (Invitrogen) and random hexomers (Promega, Madison, WI, USA) were used to prime first strand synthesis. Mouse CB₂ cDNA was generated by amplifying mouse spleen cDNA with the forward primer ACGGCTAGCGAGGGATGCCGGGAGACAGAAGT (Nhe1 site) and the reverse primer GACACGCGGCCGCCTAGGTGGTTTTCACATCAGCCTCTG (Not1 site). The amplified gene was cloned into pme18neoCD8-FLAG. To obtain the 5′- and 3′-prime ends of the rat CB₂ gene, primers (forward primer GGCCGGAGCTGACTTTCCTGGCCAGCGTGATCTTT and reverse primer GCCAGCCCAGTAGGTAGTCGTTGGGGATCA) were designed for the rapid amplification of cDNA ends (RACE) based on the existing CB₂ sequences. Marathon RACE amplification (Gibco-BRL; Bethesda, MD, USA) was performed on RNA purified from rat spleen RNA. The forward primer ACGGCTAGCGCGGGATGCCGGGAGCTGGAG (with an Nhe1 site) and a reverse primer GTTGGCGGCCGCTCAGCAATTGGAGCAGCCCTGGTGTCT (with a Not1 site) were designed based on the RACE sequences. Rat CB₂ cDNA was then amplified by PCR of rat spleen cDNA and cloned into the pME18neoCD8FLAG vector. Cyno CB2 sequence was determined using SMART RACE cDNA amplification kit (Clontech, Mountain View, CA, USA) using the forward primer GCCTTCTGCTCCATGCTGTGCCTCGTCA and the reverse primer GCAGGCGGAGAAGAGTGCACAACACGGC on cyno lymph node RNA. After sequencing of the RACE products, PCR primers (forward primer ACGGCTAGCGAGGAATGCTGGCTGACAGAGATA and reverse primer GTTGGCGGCCGCTCATCAGCAACTGGAGTGGTCTAG) were designed based on the sequence of the 5′ and 3′ end of cyno CB₂ to amplify the entire CB₂ cDNA from lymph node RNA. The cyno CB2 gene was then cloned into the pME18neoCD8FLAG vector.

Cells and cell culture

The clonal Chinese hamster ovary (CHO)–hCB₂ cell line was generated by transfection of CHO-K1 cells with human cannabinoid type 2 receptor cDNA modified by placement of the haemagglutinin (HA) epitope on the N-terminus as described previously (Gonsiorek et al., 2000). Monolayer cultures were grown at 37°C in a humidified atmosphere (5% CO₂) in Dulbecco's modified Eagle F-12 medium containing L-glutamine and supplemented with 1% non-essential amino acids, 1% penicillin/streptomycin, 10% fetal bovine serum (FBS) (Gemini Bio-Products, Calabasas, CA, USA) and 0.2 mg ml⁻¹ hygromycin B, pH 7.4. Ba/F3 cell lines transfected to express human CB₁ (Ba/F3-hCB₁) or CB₂ (Ba/F3-hCB₂) were generated as described previously (Lunn et al., 2006). Ba/F3 cell lines transfected to express rat, mouse or cyno CB₂ were generated and maintained similarly. Experimental cultures were used 1–2 days after seeding. Cell culture medium (RPMI-1640) was purchased from Gibco-BRL (Grand Island, NY, USA).

Membrane preparation

Cell membranes from brains of mice, rats and cynomolgus monkeys or transfected cell lines were prepared as described previously (Hipkin et al., 1997). Briefly, brains were minced on ice using a razor blade, rinsed with ice-cold phosphate-buffered saline (Invitrogen Corp., Carlsbad, CA, USA) to remove excess blood and kept on ice. CHO–hCB₂ cells were harvested using cell dissociation buffer according to the manufacturer's instructions (Invitrogen Corp.), collected by centrifugation and used immediately or stored at −80°C. Transfected Ba/F3 cells were pelleted by centrifugation and used immediately or stored at −80°C. Cell pellets were resuspended and incubated on ice for 30 min in cell homogenization buffer (10 mM Tris-HCl, 5 mM ethylenediaminetetraacetic acid (EDTA), 3 mM ethylene glycol bis(β-aminoethylether)-N,N,N',N',-tetraacetic acid (EGTA), pH 7.6). Brain tissue was incubated with tissue homogenization buffer (10 mM Tris-HCl, 5 mM EDTA, 3 mM EGTA, 250 mM sucrose, pH 7.6). Both cell and tissue homogenization buffer were supplemented with 1 mM phenylmethylsulphonyl fluoride. Cells and brain tissue were then homogenized with 15–20 strokes at 900 r.p.m. with a Dounce homogenizer using stirrer type RZR1 polytron homogenizer (Caframo, Wiarton, Ontario, Canada). Intact cells and nuclei were removed by low-speed centrifugation (500 g for 5 min at 4°C). Membranes in the supernatant were pelleted by centrifugation at 100 000 g for 30 min at 4°C and then resuspended in gly–gly buffer (20 mM glycylglycine, 1 mM MgCl₂, 250 mM sucrose, pH 7.2) and stored at −80°C. Protein determinations were performed using the Bradford method (Bradford, 1976).

[³⁵S]GTPγS and [³H]CP55,940 membrane binding

Binding reactions were carried out in 96-well microplates in a final volume of 100 μl. Cell membranes (1–25 μg, in triplicate) were incubated in the presence or absence of various compounds for 60 min at room temperature in binding buffer (50 mM Tris, 100 mM NaCl, 5 mM MgCl₂, and 0.1% (1 mg l⁻¹) bovine serum albumin (BSA; factor V, lipid free), pH 7.4) containing 1–2 nM [³H]CP55,940 (SA=180 Ci mmol⁻¹). In some experiments, [³H]CP55,940 binding was done in GTPγS-binding buffer (20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulphonic acid (HEPES), 100 mM NaCl, 5 mM MgCl₂, and 0.2% (2 mg l⁻¹) BSA, pH 7.4), supplemented with the indicated concentrations of GDP and 0.1 nM nonisotopic GTPγS. [³H]CP55,940 binding was terminated by rapid filtration of the membranes through the microfiltration plates coated with 0.5% polyethylenimine (UniFilter GF/C filter plate; Packard, Meriden, CT, USA), using a Tomtek 96-well cell harvester (Hamden, CT, USA). The membranes were washed 10 times with ice-cold buffer (50 mM Tris, 3 mM MgCl₂, 1 mM EDTA, 0.1% BSA, pH 7.4). Membrane-bound radioligand was measured by liquid scintillation using a TopCount NXT Microplate Scintillation and Luminescence Counter (Packard, Meriden, CT, USA).

For [³⁵S]GTPγS-binding experiments, membranes in GTPγS-binding buffer were incubated in the presence or absence of cannabinoids, 0.3 nM [³⁵S]GTPγS (SA=1250 Ci mmol⁻¹ (NEN Boston, MA, USA) and the indicated concentrations of GDP. Membrane-bound [³⁵S]GTPγS was measured using WGA-SPA beads (300 μg) at 60 min by use of a 1450 Microbeta Trilux counter (Wallac, Gaithersburg, MD, USA) as described previously (Chou et al., 2002).

Cyclic AMP accumulation assay

Assays were performed as described previously (Gonsiorek et al., 2000). Briefly, cells, seeded in 96-well plates, were chilled on ice and washed two times with cold F-12 nutrient mixture (HAM) medium, containing 10 mM HEPES and 0.2% BSA, pH 7.4. Cells were then incubated for 15 min at 37°C in the above medium supplemented with 200 μM isobutylmethylxanthine (cAMP assay media), 5 μM forskolin and the indicated concentrations of cannabinoids. The media was removed and the cells lysed with 0.1 N HCl and rapid freezing. Intracellular cAMP in thawed lysates was measured by cAMP Enzyme Immunoassay (Biomol Research Laboratories, Plymouth Meeting, PA, USA) according to manufacturer's instructions. The results are expressed as a fraction of forskolin-stimulated cAMP accumulation measured in the absence of cannabinoids.

Cell chemotaxis assays

Cellular migration was measured as described previously (Lunn et al., 2006). Briefly, cannabinoids diluted in assay buffer (phenol red free-RPMI supplemented with 10% FBS) were dispensed (30 μl) into the bottom wells of disposable microchemotaxis plates (ChemoTx 101-5 sp; Neuroprobe Inc., Gaithersburg, MD, USA). Cell aliquots (25 μl; 50 000 Ba/F3-hCB₂ cells) in assay buffer were then applied to filters (5 μm pore size) in the top plate. After incubation for 90 min at 37°C, migrated cells were collected in the bottom well by centrifugation and transferred to wells of a flat-bottom Microlite (1+) luminometer plate (Thermo Electron Corporation, Waltham, MA, USA). Eighty microlitres of assay buffer and 100 μl of CellTiter Glo Reagent (Promega) were added per well, incubated for 10 min and the luminescence intensity was measured using a luminometer (Thermo Electron Corporation) at an excitation time of 100 ms. We found that a linear relationship exists between luminescence intensity and cell number (data not shown). Relative migration is expressed as a percentage of total input cell number. Data are presented as the mean of triplicate determinations.

Data analysis

Data are presented as mean values±s.e.m. of at least three independent experiments, each of which was performed in triplicate. Nonlinear regression analysis of saturation data and of concentration–response data was performed using Prism 2.0c software (GraphPad Software, San Diego, CA, USA) to calculate K_D, B_max, IC₅₀ and EC₅₀ values. IC₅₀ values were converted to apparent K_i values by the method of Cheng and Prusoff (1973) using the K_D values for [³H]CP55,940 determined from saturation experiments.

Materials

Sf9 membranes exogenously expressing Gα_i3, β₁γ₂ and hCB₂ (7–14 pmol mg⁻¹) or hCB₁ (0.7 pmol mg⁻¹) were purchased from NEN Life Sciences. HU210 ((6aR)-trans-3-(1,1-dimethylheptyl)-6a,7,10,10a-tetrahydro-1-hydroxy-6,6-dimethyl-6H-dibenzo[b,d]pyran-9-methanol) was purchased from BIOMOL Research Laboratories. AM-630 (6-iodopravadoline) and JWH-015 (2-methyl-1-propyl-1H-indol-3-yl-1-naphthalenylmethanone) was purchased from Tocris Bioscience (Ellisville, Missouri, USA). CP55,940 ((1R,3R,4R)-3-[2-hydroxy-4-(1,1-dimethylheptyl)phenyl]-4-(3-hydroxypropyl)cyclohexan-1-ol) was purchased from NEN. WIN55,212-2 ((R)-(+)-2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo-[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthaleneylmethanone mesylate) and SR144528 (N-[(1S)-endo-1,3,3-trimethyl bicycle[2,2,1] heptane-2-yl]-5-(4-chloro-3-methylphenyl)-1-(4-methylbenzyl)-pyrazole-3carboxamide) were synthesized in our Department of Chemistry at SPRI. All other reagents were of the best grade available and purchased from common suppliers.

Results

Functional characterization of Sch35966

To assess if Sch35966 is an agonist, inverse agonist or neutral antagonist, we tested the effect of known cannabinoid agonists (HU210, WIN55,212-2), Sch35966 and an inverse agonist (SR144528) on forskolin-stimulated cAMP accumulation in CHO–hCB₂ cells, (Figure 2) which we had previously used to characterize endocannabinoid pharmacology (Gonsiorek et al., 2000). As shown in Figure 2a, Sch35966 inhibited forskolin-stimulated cAMP accumulation with an efficacy similar to HU210 and WIN55,212-2, although the latter ligands were more potent. It is difficult to conclude if Sch35966 is a full or partial agonist due to the high constitutive activity of CB₂, which suppressed the endogenous adenylyl cyclase activity in these cells. The constitutive activity of CB₂ was apparent upon co-incubation of these cells with SR144528 which relieved this tonic suppression as witnessed by a dramatic increase in forskolin-stimulated cAMP levels. In parallel, we assessed the effect of these ligands on forskolin-stimulated cAMP accumulation in CHO–hCB₂ cells preincubated with pertussis toxin (PTX) to inactivate the Gi transducers (Figure 2b). Interestingly, incubation of PTX-pretreated cells with cannabinoid agonists resulted in stimulation of cAMP levels. The stimulatory effects of the inverse agonist were largely abolished. These data suggest that hCB₂ expressed in our cell line interacts with both inhibitory (G_i) and stimulatory G proteins (G_s). A similar observation was made for CB1 in the brain (Glass and Felder, 1997).

Figure 2.

The effect of cannabinoids on cAMP accumulation in CHO–hCB₂ cells. Cells in 96-well plates were pretreated in the absence (left) or presence (right) of 100 ng ml⁻¹ PTX for 24 h (as described in the Methods section). Cells were washed and incubated for 15 min at 37°C in cAMP assay buffer containing 5 μM forskolin and the indicated concentrations of HU210, Sch35966, WIN 55,212-2 or SR144528. Following incubation, intracellular cAMP was measured by enzyme immunoassay. Data represent the mean fraction of forskolin-stimulated cAMP (pmol/well)±range of triplicate determinations from a representative experiment.

Lastly, we tested the chemotactic response of Ba/F3-hCB₂ in response to Sch35966 and HU210 as cannabinoids have been shown to stimulate cell chemotaxis via CB₂ (Sacerdote et al., 2000; Kishimoto et al., 2003; Lunn et al., 2006). Both Sch35966 and HU210 stimulated Ba/F3-hCB₂ cell chemotaxis with the classic bell-shaped concentration–response curve (data not shown).

We next assessed the pharmacology of Sch35966 using [³⁵S]GTPγS-exchange assays (as described in the Methods section). Breivogel et al. (1998) previously established that relative to full agonists, the efficacy of partial agonists at the cannabinoid CB₁ to stimulate [³⁵S]GTPγS exchange was more susceptible to inhibition with elevated GDP concentrations. Therefore, we assessed the effect of increasing concentrations of GDP on the [³⁵S]GTPγS exchange in CHO–hCB₂ membranes in response to 100 nM HU210, CP55,940, WIN55,212-2, SR144528, AM-630, Sch35966 or 1 μM JWH-015 (Figure 3a) and expressed the effect of the ligands relative to basal binding (Figure 3b). CP55,940 was the most efficacious ligand tested in stimulating GTPγS exchange. While HU210 appeared to be (at least) as effective as CP55,940 with ⩽1 μM GDP, its relative efficacy declined in the face of higher GDP levels. Sch35966, WIN55,212-2 and JWH-015 were less efficacious. As would be expected, the inverse agonists SR144528 and AM-630 decreased constitutive GTPγS exchange. From these experiments, we selected 5 μM GDP and 0.3 nM GTPγS for the more extensive assessment of the potency, efficacy and affinity of HU210, CP55,940, Sch35966 and JWH-015 at hCB₂. Representative data are shown in Figure 4. Consistent with the data from the GDP titration experiments (Figure 3), CP55,940 was a slightly more efficacious agonist than HU210 and Sch35966 although the latter compound was less potent. In this particular set of data, Sch35966 stimulated GTPγS exchange to levels which are equivalent to that seen with HU210, although taken together the results from our experiments suggest that HU210 is slightly more efficacious than Sch35966 (Figure 3b; data not shown). HU210, CP55,940 and Sch35966 were all more potent than JWH-015 in stimulating GTPγS exchange. The binding affinities in buffer containing 5 μM GDP and 0.3 nM nonisotopic GTPγS were in general agreement with the functional potencies.

Figure 3.

The effect of GDP concentration on [³⁵S]GTPγS exchange in response to cannabinoids in CHO–hCB₂ membranes. Membranes expressing hCB₂ (4 μg) prebound to WGA-SPA beads were incubated in GTPγS binding buffer (as described in the Methods section) containing 0.1 nM [³⁵S]GTPγS and the indicated concentration of GDP in the absence or presence of cannabinoids at either 100 nM (HU210, CP55,940, WIN 55,212, SR144528) or 1 μM (Sch35966, JWH-015) for 60 min at room temperature. The mean binding±range of triplicate determinations from a representative experiment is shown on the left panel. The same data are presented in the right panel expressed as total/basal binding at each GDP concentration (B/B₀). Membrane-associated radioactivity was measured by WGA-SPA scintillation.

Figure 4.

The effect of cannabinoids on [³⁵S]GTPγS exchange and [³H]CP55,940 binding in CHO–hCB₂ membranes. Membranes (4 μg) were incubated at 30°C for 60 min in binding buffer containing 5 μM GDP, the indicated concentrations of HU210, CP55,940, WIN 55,212 or Sch35966. In GTPγS exchange assays, the incubation contained 0.1 nM [³⁵S]GTPγS (open symbols, solid lines). In competition binding assays, the incubation contained 0.1 nM nonisotopic GTPγS and 1–2 nM [³H]CP55,940 (closed symbols, broken lines), Following filtration, the membrane-associated radioactivity was measured by liquid scintillation. Data (expressed as B/B₀) represent the mean specific binding±range of triplicate determinations from a representative experiment (n=2–4).

The effect of Sch35966 on [³H]CP55,940 binding and [³⁵S]GTPγS exchange in membranes expressing primate or rodent CB₂

We measured the binding affinities and intrinsic efficacies of Sch35966, HU210, JWH-015 using membranes from a cell line transfected to express CB₂ from human (Ba/F3-hCB₂), cynomolgus monkey (Ba/F3-cynoCB₂), rat (Ba/F3-rCB₂) or mouse (Ba/F3-mCB₂). Saturation analysis (Figure 5) showed that [³H]CP55,940 bound with high affinity to CB₂ from mouse (mCB₂ K_D=0.5 nM), rat (rCB₂ K_D=0.6 nM) and cynomolgus monkey (cynoCB₂ K_D=0.7 nM). In GTPγS exchange assays (Figure 6, Table 1), Sch35966 was less potent than HU210 and demonstrably more potent than JWH-015. In competition binding assays (Figure 7 and Table 2), HU210 bound with highest affinity in all species (0.40–0.60 nM) followed by Sch35966 (2–7 nM). JWH-015 bound CB₂ with considerably lower affinity (∼30–400 nM).

Figure 5.

Saturation binding isotherms for mouse, rat and monkey CB₂. Membranes from Ba/F3 cells transfected to express CB₂ from the mouse (mCB₂), rat (rCB₂) or monkey (cynoCB₂) were incubated for 30 min at 30°C in binding buffer (as described in the Methods section) with the indicated concentrations of [³H]CP55,940 in the presence or absence of excess unlabelled ligand. Following filtration, the membrane-associated radioactivity was measured by liquid scintillation. Data represent the mean total binding±range of triplicate determinations from two independent experiments.

Figure 6.

The effect of cannabinoids on [³⁵S]GTPγS exchange in membranes-expressing rodent or monkey CB₂. Membranes from Ba/F3 cells transfected to express CB₂ from the mouse (mCB₂), rat (rCB₂) or monkey (cynoCB₂) were incubated at 30°C for 60 min in binding buffer containing 5 μM GDP and the indicated concentrations of HU210, Sch35966 or JWH-015. The incubation contained 0.1 nM [³⁵S]GTPγS. Following filtration, the membrane-associated radioactivity was measured by liquid scintillation. Data represent the mean specific binding±range of triplicate determinations from a representative experiment (n=2–4).

Table 1.

Comparison of functional potencies (EC₅₀) of cannabinoid ligands in stimulating [³⁵S]GTPγS exchange in membranes expressing monkey, rat or mouse CB₂

Ligand	Monkey	Rat	Mouse
HU-210	0.32±0.17	0.28±0.22	0.24±0.09
Sch35966	1.5±0.98	6203±7597	2.4±1.4
JWH-015	106±41	4407±3090	537±253

Abbreviation: CB₂, peripheral cannabinoid receptor.

Potency data (nM) are presented as mean±s.e./(range) values from two or three independent experiments performed in triplicate.

Figure 7.

The effect of cannabinoids on [³H]CP55,940 binding to human, rat and monkey CB₁ and CB₂. Membranes from Ba/F3 cells transfected to express rCB₂, mCB2, cynoCB₂ or hCB₁ or from rat or monkey brain were incubated at 30°C for 60 min in binding buffer containing 2 nM [³H]CP55,940 and the indicated concentrations of HU210, Sch35966 or JWH-015. Following filtration, the membrane-associated radioactivity was measured by liquid scintillation. Data represent the mean specific binding±range of triplicate determinations from a representative experiment (n=2–4).

Table 2.

Comparison of binding affinities (K_i) of cannabinoid ligands to CB₂ in human, monkey, rat and mouse

Ligand	Human	Monkey	Rat	Mouse
HU-210	0.45±0.01	0.53±0.1	0.41±0.12	0.50±0.09
Sch35966	6.8±2.3	5.4±0.4	2.4±0.5	4.8±1.6
JWH-015	180±70	34±6	340±111	373±176

Abbreviation: CB₂, peripheral cannabinoid receptor.

Binding constants (nM) are presented as mean±s.d. values from at least three independent experiments performed in triplicate.

To assess CB₂ selectivity, we measured the binding affinity of these compounds to hCB₁ (Ba/F3-hCB₁ membranes), rCB₁ (rat brain membranes) and cynoCB₁ (cynomolgus monkey brain membranes) by competition with [³H]CP55,940 (Figure 7). Again, HU210 bound with high affinity in all of the membrane preparations (0.7–1.1 nM; Table 3). Both Sch35966 and JWH-015 displaced [³H]CP55,940 from CB₁ with low affinity. On the basis of its higher affinity for CB₂, Sch35966 is more CB₂ selective than is JWH-015 in all species tested.

Table 3.

Comparison of binding affinities (K_i) of cannabinoid ligands to human, monkey and rat CB₁

Ligand	Human	Monkey	Rat
HU-210	0.7±0.11	0.94±0.03	1.00±0.18
Sch35966	2633±829	5100±718	3000±982
JWH-015	2300±442	4400±1227	1257±400

Binding constants (nM) are presented as mean±s.e. values from at least three independent experiments performed in triplicate.

Discussion

The studies presented here describe the characterization of a novel benzoquinolizinone, Sch35966, which binds with low nM potency to CB₂ in both primates (human and cynomolgus monkey) and rodents (rat and mouse). Further, we demonstrated that Sch35966 has >450-fold selectivity for CB₂ binding vs the CB₁. Sch35966 is an agonist as it effectively inhibited forskolin-stimulated cAMP synthesis in CHO–hCB₂ cells and stimulated [³⁵S]GTPγS exchange in cell membranes expressing human, monkey, rat and mouse CB₂. In all species examined, Sch35966 was more potent and selective than JWH-015 (a commonly used CB₂-selective agonist).

Benzoquinolizinones have been used to treat a number of medical conditions. The compounds were identified early on as novel partial agonists for the benzodiazepine receptors (Jenck et al., 1992) and developed for the treatment of anxiety disorders. More recent examples of this class of compound can be seen in the development of a class of non-steroidal inhibitors of human steroid 5α-reductases 1 and 2 (Ferrali et al., 2005), progressed for the treatment of prostatic hyperplasia, acne and alopecia. We identified the benzoquinolizinone Sch35966 by screening a group of compounds using a recombinant CB₂ membrane preparation in a ligand-binding assay, while seeking to identify a novel class of cannabinoid CB₂ ligands. The interest in developing cannabinoid CB₂-specific agonists was based on the hypothesis that the relatively high expression of the receptor mRNA in immune cells vs its expression in the CNS could allow the generation of an immunomodulatory cannabinoid compound that lacked the psychoactive effects mediated by the central CB₁. Data supporting this approach continue to be demonstrated – in a recent publication, Correa et al. (2005) showed that a cannabinoid CB₂-specific agonist inhibited lipopolysaccharide/interferon-γ (IFN-γ) induced IL-12p40 and enhanced IL-10 release.

In spite of the extensive in vitro data supporting an immunomodulatory role for cannabinoid agonists, in recent clinical trials no immunological deficits were observed following short-term dosing with plant cannabinoid agonists. Abrams et al. (2003) showed that Δ⁹-THC had no significant effect on the number of peripheral CD4⁺ and CD8⁺ cells, and no effect on viral load in patients using cannabis to control AIDS-related wasting syndrome. Likewise, Katona et al. (2005) found no effect on serum IFN-γ, IL-10, IL-12 or C-reactive protein in patients participating in CAMS, the multicentre randomized placebo-controlled trial on the effect of Δ⁹-THC on the symptoms of multiple sclerosis. In addition, trials for the only cannabinoid compounds approved for clinical use, marinol and sativex, have reported no drug-related immunological deficits in an immunocompromised patient population (Guy and Stout, 2005). It is important to note that the lack of profound immunological effects upon administration of Δ⁹-THC may well reflect its lack of activity as an agonist at hCB₂ (Bayewitch et al., 1995, 1996; Govaerts et al., 2004; for review, see Howlett et al., 2002). Interestingly, Smith et al. (2000, 2001a) demonstrated that in vivo cytokine regulation in mice challenged with endotoxin involved the central CB₁. Studies in models of thioglycolate-induced or staphylococcus enterotoxin A-induced peritoneal inflammation demonstrated roles for both CB₁ and CB₂ ligands in mediating inflammation (Smith et al., 2001b). Our experiments have also suggested that cannabinoid CB₂ inverse agonists may mediate immune cell motility following an immune insult (Lavey et al., 2005; Lunn et al., 2006), supporting the idea that CB₂ agonists may not be the only productive strategy for cannabinoid-based immunoregulation.

In recent years, another CB₂-mediated physiological effect has been identified, further intensifying the search for novel CB₂-specific agonists. CB₂ agonists have been found to be effective in a number of animal models of pain, including formalin-induced and capsaicin-induced pain (Hanus et al., 1999; Hohmann et al., 2004), inflammatory pain (Clayton et al., 2002; Quartilho et al., 2003), neuropathic pain (Malan et al., 2002; Ibrahim et al., 2005) and a hindpaw incision model of postoperative pain (Labuda et al., 2005). The mechanism by which CB₂ mediates pain probably varies between the different classes of pain and remains an active area of investigation. Because of the highly selective expression of CB₂ in peripheral immune cells, most models target these cells. Malan et al. (2003) suggested that the CB₂ agonists could act on immune cells proximal to the site of insult, blocking mediator release. Ibrahim et al. (2005) showed that the pain response was blocked by μ-receptor antagonists or anti-β-endorphin, suggesting the modulation of β-endorphin production by neighbouring tissue was involved. With increasing evidence for CB₂ expression in neural tissue (Van Sickle et al., 2005), other non-peripheral models for the role of CB₂ in pain have been proposed. Beltramo et al. (2006) showed that CB₂-specific agonists reduce capsaicin-induced calcitonin gene-related peptide (CGRP) release from cultures of spinal cord microglia cells. This effect appeared to be mediated by the CB₂, as SR144528 induced a rightward shift of the agonist dose–response curve. In addition, studies in CB₁-depleted mice ruled out a role for the central CB₁. These studies support a potential role for CB₂-specific agonists as a novel new class of drug for the induction of pain relief without psychoactive effects (Malan et al., 2003).

In conclusion, the benzoquinolizinone compound Sch35966 is a novel agonist at the CB₂. Sch35966 potently activates CB₂ from both primates (human and cynomolgus monkey) and rodents (rat and mouse) with >450-fold selectivity vs the CB₁. In all species examined, Sch35966 was more potent and selective than JWH-015 (a commonly used CB₂-selective agonist).

Abbreviations

2-AG

2-arachidonyl glycerol

CB₁

central cannabinoid receptor

CB₂

peripheral cannabinoid receptor

CHO

Chinese hamster ovary

FBS

fetal bovin serum

Δ⁹-THC

Δ⁹-tetrahydrocannabinol

Conflict of interest

The authors are or were employees of the Schering-Plough Research Institute.