Physical and functional interaction between CB₁ cannabinoid receptors and β₂-adrenoceptors

Abstract

Background and purpose:

The CB₁ cannabinoid receptor and the β₂-adrenoceptor are G protein-coupled receptors (GPCRs) co-expressed in many tissues. The present study examined physical and functional interactions between these receptors in a heterologous expression system and in primary human ocular cells.

Experimental approach:

Physical interactions between CB₁ receptors and β₂-adrenoceptors were assessed using bioluminescence resonance energy transfer (BRET). Functional interactions between these receptors were evaluated by examining receptor trafficking, as well as extracellular signal-regulated kinase (ERK) and cyclic AMP response element binding protein (CREB) signalling.

Key results:

Physical interactions between CB₁ receptors and β₂-adrenoceptors were demonstrated using BRET. In human embryonic kidney (HEK) 293H cells, co-expression of β₂-adrenoceptors tempered the constitutive activity and increased cell surface expression of CB₁ receptors. Co-expression altered the signalling properties of CB₁receptors, resulting in increased Gα_i-dependent ERK phosphorylation, but decreased non-Gα_i-mediated CREB phosphorylation. The CB₁ receptor inverse agonist AM251 (N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide) attenuated β₂-adrenoceptor-pERK signalling in cells expressing both receptors, while the CB₁ receptor neutral antagonist O-2050 ((6aR,10aR)-3-(1-methanesulfonylamino-4-hexyn-6-yl)-6a,7,10,10a-tetrahydro-6,6,9-trimethyl-6H-dibenzo[b,d]pyran) did not. The actions of AM251 and O-2050 were further examined in primary human trabecular meshwork (HTM) cells, which are ocular cells endogenously co-expressing CB₁ receptors and β₂-adrenoceptors. In HTM cells, as in HEK 293H cells, AM251 but not O-2050, altered the β₂-adrenoceptor–pERK response.

Conclusion and implications:

A complex interaction was demonstrated between CB₁ receptors and β₂-adrenoceptors in HEK 293H cells. As similar functional interactions were also observed in HTM cells, such interactions may affect the pharmacology of these receptors in tissues where they are endogenously co-expressed.

This article is part of a themed issue on Cannabinoids. To view the editorial for this themed issue visit http://dx.doi.org/10.1111/j.1476-5381.2010.00831.x

Introduction

The CB₁ cannabinoid receptor is a rhodopsin-like, family A, G protein-coupled receptor (GPCR) that is widely expressed both within the CNS and the periphery. The CB₁ receptor (nomenclature follows Alexander et al., 2009) was originally described as a receptor for the primary psychotropic agent in the plant Cannabis sativa, Δ⁹-tetrahydrocannabinol (Matsuda et al., 1990), but has since been shown to also bind endogenous ligands including N-arachidonoylethanolamine or anandamide (AEA), and 2-arachidonoylglycerol (2-AG) (Devane et al., 1992; Mechoulam et al., 1995). CB₁ receptors are involved in a wide range of biological functions both in the CNS and the periphery. Within the CNS, they are present presynaptically and act to inhibit neurotransmitter release, while in the periphery CB₁ receptors are involved in the regulation of energy and metabolism, bone formation, embryo implantation, cardiovascular function and intraocular pressure (IOP) (Kunos et al., 2000; Wang et al., 2004; Szczesniak et al., 2006; Cota, 2007; Hashimotodani et al., 2007; Bab and Zimmer, 2008).

Functionally, CB₁ receptors have been reported to couple primarily to Pertussis toxin (PTx)-sensitive G_i/o proteins to inhibit adenylyl cyclase and voltage-gated Ca²⁺ channels, while activating mitogen-activated protein kinases (Demuth and Molleman, 2006). However, it has recently been shown that CB₁ receptors also couple to some degree with both G_s and G_q/11 proteins to activate adenylyl cyclase and increase intracellular Ca²⁺ respectively (Maneuf and Brotchie, 1997; Lauckner et al., 2005).

Like many other family A GPCRs, CB₁ receptors physically interact with other GPCRs to form both homodimers, as well as heterodimers with the D₂ dopamine receptor; the µ-, κ- and δ-opioid receptors; the orexin-1 receptor; and the A_2A adenosine receptor (Wager-Miller et al., 2002; Kearn et al., 2005; Mackie, 2005; Ellis et al., 2006; Rios et al., 2006; Carriba et al., 2007). These interactions have been shown to influence many aspects of CB₁ receptor function including ligand pharmacology, receptor trafficking and G protein coupling. Although, to date, these are the only GPCRs reported to physically interact with CB₁receptors, given the biological significance and widespread distribution of this receptor it is likely that CB₁ receptors interact with additional GPCRs.

The β₂-adrenoceptor is also a rhodopsin-like, family A, GPCR that is widely expressed in several tissue and cell types. β₂-adrenoceptors are predominantly coupled with G_s, but also to a lesser extent with G_i (Xiao et al., 1995). Like CB₁receptors, β₂-adrenoceptors have been shown to form homodimers and heterodimers with other family A GPCRs, including the β₁- and β₃-adrenoceptors, the prostaglandin EP₁ receptor and the µ-opioid receptor (Hebert et al., 1996; Angers et al., 2000; Mcvey et al., 2001; Lavoie et al., 2002; Breit et al., 2004; Mcgraw et al., 2006). The tissue distribution of β₂-adrenoceptors overlaps significantly with that of the CB₁ receptors, including parts of the cardiovascular system, reproductive tract, brain, eye and bone (Jampel et al., 1987a; Wanaka et al., 1989; Tsou et al., 1998; Stamer et al., 2001; Wang et al., 2004; Pacher and Hasko, 2008). Despite this overlapping distribution, possible direct physical and functional interactions between CB₁receptors and β₂-adrenoceptors have not been examined. Instead, most studies on the interactions between the cannabinoid and adrenergic systems have focused on inhibition of noradrenergic neurotransmission by presynaptic CB₁ receptors (Schlicker et al., 1997; Schultheiss et al., 2005; Pakdeechote et al., 2007; Tam et al., 2008).

One organ where interactions between CB₁ receptors and β₂-adrenoceptors may be of particular interest is the eye. Agonists targeting CB₁ receptors and antagonists targeting β₂-adrenoceptors in the eye are known clinically to decrease IOP (Hepler and Frank, 1971; Borthne, 1976; Pate et al., 1998; Mccarty et al., 2008). In fact, β₂-adrenoceptor antagonists are a front-line treatment for glaucoma, a blinding eye disease for which the major risk factor is elevated IOP. In humans, IOP is maintained by the balance of aqueous humour production in the ciliary body epithelium and outflow through trabecular meshwork and uveoscleral pathways (Woodward and Gil, 2004). Interestingly, both ciliary epithelial cells and trabecular meshwork cells co-express CB₁ receptors and β₂-adrenoceptors, and these receptors have been implicated in the regulation of both aqueous humour production and outflow (Jampel et al., 1987b; Wax et al., 1989; Straiker et al., 1999; Stamer et al., 2001; Njie et al., 2006). Thus, these ocular cells provide an ideal model for studying endogenous interactions between these two receptors.

In the present study, novel physical and functional interactions between CB₁ receptors and β₂-adrenoceptors were identified in human embryonic kidney (HEK) 293H cells. These interactions were found to influence both signalling and trafficking of the two receptors. The functional consequences of this CB₁/β₂-adrenoceptors interaction were then further examined in primary human trabecular meshwork (HTM) cells. Together, our observations in HEK 293H and HTM cells suggest complex cell type-specific physical and functional interactions between CB₁ receptors and β₂-adrenoceptors that may be relevant to the cells that co-express these two receptors in vivo.

Methods

Constructs

Human CB₁ cannabinoid receptor (CB₁) carboxy-terminal GFP² and Renilla luciferase (Rluc) constructs were generated by PCR; the CB₁ sequence was amplified without its stop codon from the Rc/CMV-CB₁ plasmid (from Tom Bonner, NIH, Bethesda, MD, USA) using forward (CGACGAATTCCAGCCTAATCAAAGACTGAGGTT) and reverse (TGACATGGATCCCACAGAGCCTCGGCAGAC) primers. The PCR product was digested with EcoRI and BamHI and inserted into the pGFP²-N3 and pRluc-N1 plasmids (PerkinElmer) to produce CB₁-GFP² and CB₁-Rluc respectively. Constructs of human β₂-adrenoceptors (β₂AR-GFP², or β₂AR-Rluc) and of the human ether-a-go-go related gene (HERG-GFP²) were prepared as previously reported (Lavoie et al., 2002; Dupre et al., 2007). The human metabotropic glutamate receptor 6 RcCMV(mGLuR6) and mGluR6-GFP² constructs were kind gifts from Dr Robert Duvoisin (Oregon Health, and Science University, Portland, OR, USA). The HA-tagged β₂-adrenoceptor in pcDNA3.1/Zeo(−) (Invitrogen Canada Inc., Burlingon, ON, Canada), HA-β₂AR(Zeo), was generated by inserting HA-β₂AR into the EcoRI and HindIII sites of pcDNA3.1/Zeo(−). The neomycin resistant HA-β₂AR(Neo) construct was generated by transferring the HA-β₂AR sequence from HA-β₂AR(Zeo) to pcDNA3.1(+) (Invitrogen Canada Inc.) using NheI and HindIII. To generate pTreHA-β₂AR, the HA-β₂AR sequence was cut out of HA-β₂AR(Neo) using NotI and inserted into pTRE2hyg (Clontech Laboratories Inc., Mountain View, CA, USA). The pTet-ON plasmid was from Clontech.

Cell culture and transfection

Human embryonic kidney 293H cells (Invitrogen Canada Inc.) were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS). Transfections were carried out using the Lipofectamine 2000 transfection reagent (Invitrogen Canada Inc.) according to the manufacturer's protocol. Stable cell lines for various expression constructs were generated by selection with the appropriate antibiotics. Primary HTM cells were obtained from ScienCell (Carlsbad, CA, USA) and maintained in DMEM with 10% FBS. Cell cultures were split every 3–4 days before reaching confluency and were passaged up to 10 times before their experimental use.

Generation of stable cell lines

CB₁-GFP² was transfected into HEK 293H cells and stably expressing clones were selected using the antibiotic Zeocin. These CB₁-GFP² cells were then transfected with the Tet-ON plasmid and selected using G418. Finally, cells expressing CB₁-GFP² and Tet-ON were transfected with pTreHA-β₂AR and selected using hygromycin B to generate a cell line stably expressing CB₁-GFP, Tet-ON and pTreHA-β₂AR (CB₁-GFP²/TreHA-β₂AR cells). Using On-Cell Western, the addition of doxycycline (Dox) to CB₁-GFP²/TreHA-β₂AR cells was shown to produce a dose-dependent induction of HA-β₂AR expression (data not shown). A cell line stably expressing both CB₁-GFP² and HA-β₂AR was generated by transfecting HA-β₂AR(Neo) into CB₁-GFP² cells and selecting using both Zeocin and G418. Cells only expressing HA-β₂AR were generated using the HA-β₂AR(Zeo) plasmid and selected using Zeocin.

Bioluminescence resonance energy transfer

Bioluminescence resonance energy transfer (BRET) experiments were carried out using a combination of the GFP² BRET acceptor, and the DeepBlueC coelenterazine Rluc substrate (PerkinElmer, Waltham, MA, USA), as part of the previously described BRET² technique (Ramsay et al., 2002). Cells were transfected with the GFP² and Rluc constructs and cultured for 24–48 h before their use in BRET experiments. Cells were then washed twice with ice-cold phosphate-buffered saline (PBS) before being suspended in PBS supplemented with 1.0 g·L⁻¹ glucose, 10 mg·L⁻¹ benzamidine, 5 mg·L⁻¹ leupeptin and 5 mg·L⁻¹ soybean trypsin inhibitor (Roche Canada, Mississauga, ON, Canada). Cells were dispensed into a white 96-well plate and their GFP² emission was measured using a FLx800 fluorescence plate reader (BioTek Instruments Inc., Winooski, VT, USA) with excitation and emission filters of 485/20 and 528/20 nm respectively. BRET measurements were carried out using a Luminoskan Ascent plate reader (Thermo Scientific, Waltham, MA, USA) immediately following the addition of DeepBlueC coelenterazine substrate (PerkinElmer, Waltham, MA, USA) to a final concentration of 5 µM. All BRET measurements were taken by setting the plate reader to make dual luminescent emission measurements using 510 and 405 nm filters with the integration time set to 10 s and the photomultiplier tube voltage set to 1200. BRET measurements were then converted to BRET efficiencies (BRET_Eff) according to a previously described method (James et al., 2006). Briefly, BRET_Eff values were calculated by normalizing the ratio of luminescent emissions at 510/405 nm for each sample to the minimum and maximum 510/405 nm emission ratios obtained using empty Rluc and GFP²-Rluc fusion constructs respectively.

Immunofluorescence and confocal microscopy

Cells expressing HA and/or GFP² constructs or HTM cells were plated onto glass coverslips and maintained for 24–48 h. Cells were then treated as indicated in FBS-free DMEM before being fixed for 5 min in ice cold 100% methanol. After washing with PBS, cells were permeabilized with 100 µM digitonin, washed with PBS and blocked with 1% BSA in PBS. Cells were incubated with primary antibodies: monoclonal mouse anti-HA (Covance, Emeryville, CA, USA), or polyclonal mouse anti-β₂AR (Abnova, Neihu District. Taipei City, Taiwan) and polyclonal rabbit anti-CB₁ (Caymen Chemical, Ann Arbor, MI, USA) overnight at 4°C. After washing with PBS, coverslips were incubated with Cy3-conjugated anti-mouse IgG and/or fluorescein isothiocyanate-conjugated anti-rabbit IgG secondary antibodies for 1 h at room temperature. Coverslips were then washed again with PBS before being mounted on slides using Fluorsave reagent (Calbiochem, San Diego, CA, USA), and imaged using a Nikon Eclipse E800 microscope fitted with the D-Eclipse C1 confocal system (Nikon Canada Inc., Mississauga, ON, Canada). GFP² and fluorescein isothiocyanate were imaged using a 488 nm air-cooled argon laser (Spectra-Physics Lasers Inc., Mountain View, CA, USA), while Cy3 was imaged with a 543 nm He-Ne laser (JDS Uniphase, Milpitas, CA, USA).

In-Cell Western blot analyses

Phosphorylation of extracellular signal-regulated kinases 1 and 2 (ERK) and cyclic AMP response element binding protein (CREB) was assessed using a modified In-Cell Western protocol, as previously reported (Mcintosh et al., 2007). Briefly, HEK 293H cells expressing CB₁-GFP and/or HA-β₂AR constructs or HTM cells were plated in 96-well plates and cultured to confluency. Cell culture media was then replaced with FBS-free DMEM and cells were maintained for 24 h prior to experiments. Cells were then treated as indicated for either 5 or 10 min in ERK experiments, or 30 min in CREB experiments, before being fixed for 1 h with 4% paraformaldehyde. Rabbit anti-phospho ERK1/2 and goat anti-total ERK2 primary antibodies (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) were used to assess ERK phosphorylation, while goat anti-phospho-CREB and rabbit anti-total CREB primary antibodies (Santa Cruz Biotechnology Inc.) were used to assess CREB phosphorylation. Secondary antibodies used were a IRDye800CW-conjugated donkey anti-rabbit IgG (Rockland Immunochemicals Inc., Gilbertsville, PA, USA) and an Alexa Fluor680-conjugated donkey anti-goat IgG (Invitrogen Canada Inc.). Plates were scanned to measure their fluorescent emission from the IRDye800CW- and Alexa Fluor680-conjugated antibodies using the Odyssey infrared imaging system (Li-Cor Biotechnology, Lincoln, NE, USA). In order to convert these data to relative pERK and pCREB values, background fluoresence was first determined and then subtracted using wells of the plate that received only the secondary antibodies. The ratio of the pERK/ERK2 or pCREB/total CREB signals (with background subtracted) were then determined for each well and normalized to the ratios obtained from the appropriate vehicle or untreated wells in order to obtain relative pERK or pCREB values. Experiments were in all cases repeated several times, and within each experiment, each condition was repeated in 2–8 wells. The total numbers of individual wells used are presented as the ‘n’ number in each figure.

On-Cell Western blot analyses

Cell surface expression of CB₁-GFP² and HA-β₂AR was assessed using a modified In-Cell Western protocol utilizing non-permeablized cells (Miller, 2004). Cells were plated in 96-well plates and cultured for 24–48 h before experiments. Cells were treated as indicated in FBS-free DMEM and fixed in 4% paraformaldehyde. After washing with PBS, cells were blocked with 1% BSA in PBS at room temperature for 2 h. Primary antibodies directed at N-terminal epitopes of either the CB₁-GFP² or HA-β₂AR constructs were polyclonal rabbit anti-CB₁ (Cayman Chemical, Ann Arbor, MI, USA) or monoclonal anti-HA (Covance, Emeryville, CA, USA) antibodies respectively. Primary antibodies were diluted in 1% BSA in PBS and applied for 1 h at room temperature. After washing, cells were incubated in secondary antibodies: anti-rabbit IgG IRDye800CW and goat anti-mouse IgG Alexa Fluor680 (Invitrogen Canada Inc.), diluted in 1% BSA in PBS for 1 h at room temperature. Cells were washed with PBS then with distilled water before drying. Once dry, the plates were imaged using an Odyssey infrared imaging system (Li-Cor Biotechonology).

Statistical analysis and curve fitting

All data are presented as mean ± SEM. Statistical analysis and curve fitting of the data were performed using Graphpad Prism v.4 (GraphPad Software Inc., San Diego, CA, USA). To fit data to dose–response curves, vehicle treatments were plotted at a concentration equal to one and a half log units less than the lowest drug treatment concentration then data were fitted to a sigmoidal dose–response curve with variable slope. Statistical significance for curve fits was determined by the F-test comparing global fits, by t-test when comparing means of only two groups, or by one-way, two-way, or repeated measures analysis of variance (anova), as appropriate, when comparing the means of multiple treatment groups. Tukey's post hoc analysis was used to determine differences among groups for one-way anova, while Bonferroni's post hoc analysis was used for two-way anova. P < 0.05 was considered statistically significant.

Materials

Pertussis toxin, hygromycin B and G418 sulphate were from Calbiochem. (R)-(+)-WIN 55,212-2 mesylate ((R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone mesylate), AM251 (N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide), AM630 (6-Iodo-2-methyl-1-[2-(4-morpholinyl)ethyl]-1H-indol-3-yl](4-methoxyphenyl)methanone), O-2050 ((6aR,10aR)-3-(1-methanesulfonylamino-4-hexyn-6-yl)-6a,7,10,10a-tetrahydro-6,6,9-trimethyl-6H-dibenzo[b,d]pyran), ICI 118,551 ((±)-erythro-(S*,S*)-1-[2,3-(dihydro-7-methyl-1H-inden-4-yl)oxy]-3-[(1-methylethyl)amino]-2-butanol hydrochloride) and CGP 20712 (1-[2-((3-carbamoyl-4-hydroxy)phenoxy)ethylamino]-3-[4-(1-methyl-4-trifluoromethyl-2-imidazolyl)phenoxy]-2-propan ol dihydrochloride) were from Tocris Bioscience (Ellisville, MO, USA). Zeocin and Opti-MEM were obtained from Invitrogen Canada Inc. FBS was from PAA laboratories Inc. (Etobicoke, ON, Canada). Restriction enzymes, DNA polymerases and other enzymes were from ferments Canada Inc. (Burlington, ON, Canada). All other chemicals and reagents were from Sigma-Aldrich Canada Ltd. (Oakville, ON, Canada).

Results

Physical interactions between CB₁ receptors and β₂-adrenoceptors in HEK 293H cells

Bioluminescence resonance energy transfer² (BRET²) was used to demonstrate an interaction between CB₁ receptors and the β₂-adrenoceptors in HEK 293H cells. BRET_Eff was measured from cells co-transfected with either CB₁-Rluc or β₂AR-Rluc, and one of CB₁-GFP², β₂AR-GFP², HERG-GFP² or mGluR6-GFP² (Figure 1A). When co-expressed with CB₁-Rluc, CB₁-GFP² and β₂AR-GFP² produced significantly increased BRET_Eff (P < 0.001) compared with either HERG-GFP² or mGluR6-GFP², two different membrane proteins not expected to interact with either CB₁ receptors or β₂-adrenoceptors. Similarly, when co-transfected with β₂AR-Rluc both CB₁-GFP² and β₂AR-GFP² produced significantly increased BRET_Eff compared with the HERG-GFP² (P < 0.001) and mGluR6-GFP² (P < 0.01) controls. In all cases, GFP² expression levels were equal to or less than those of the HERG-GFP² negative control (data not shown). These data confirm previous reports of both CB₁ receptor and β₂-adrenoceptor homodimerization (Hebert et al., 1996; Angers et al., 2000; Wager-Miller et al., 2002), but also suggest a novel physical interaction between CB₁ receptors and β₂-adrenoceptors.

Figure 1.

CB₁ cannabinoid receptors physically interact with β₂-adrenoceptors when expressed in 293H cells. (A) BRET_Eff values obtained from 293H cells transiently transfected with either CB₁-Rluc or β₂AR-Rluc and CB₁-GFP², β₂AR-GFP², mGluR6-GFP² or HERG-GFP². ***P < 0.001 compared with HERG-GFP² controls; n= 4–9. (B) BRET saturation curves for CB₁-Rluc with β₂AR-GFP², CB₁-GFP², HERG-GFP² and mGluR6-GFP². BRET_Eff is plotted against the ratio of GFP² fluorescence (obtained by directly exciting GFP²) and Rluc emission and the data were fitted to a rectangular hyperbola. (C) BRET_Eff values from 293H cells transfected with a fixed amount of CB₁-Rluc and β₂AR-GFP² (Control) and pcDNA or increasing amounts of HA-β₂AR or mGluR6. *P < 0.05 and ***P < 0.001 compared with control column; n= 4–10. (D) BRET_Eff values obtained from cells transfected with CB₁-Rluc and β₂AR-GFP² and treated for 15 min at room temperature with DMSO (0.05%), WIN (10 µM), AM251 (10 µM) or O-2050 (10 µM) prior to measuring BRET_Eff. *P < 0.05 compared with WIN; ^#P < 0.05 compared with O-2050; n= 9. AM251, N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide; BRET, bioluminescence resonance energy transfer; BRET_Eff, BRET efficiency; DMSO, dimethylsulphoxide; HERG, human ether-a-go-gorelated gene; mGluR6, metabotropic glutamate receptor 6; O-2050, (6aR,10aR)-3-(1-methanesulfonylamino-4-hexyn-6-yl)-6a,7,10,10a-tetrahydro-6,6,9-trimethyl-6H-dibenzo[b,d]pyran; Rluc, Renillaluciferase; WIN, WIN 55,212-2.

Bioluminescence resonance energy transfer saturation experiments were performed according to a previously described protocol (Roy et al., 2006). A fixed amount of the CB₁-Rluc construct was co-transfected with increasing amounts of β₂AR-GFP², CB₁-GFP², HERG-GFP² or mGluR6-GFP². BRET_Eff values were plotted against the ratio of GFP² fluorescent emission obtained by directly exciting GFP² (measuring GFP² emission in the absence of the Rluc coelenterazine substrate) to the Rluc bioluminescent emission and fitted to rectangular hyperbola curves (Figure 1B). Significantly different BRET₅₀ values (P < 0.05) of 0.6 ± 0.1 and 0.19 ± 0.07, and BRET_Max values (P < 0.001) of 0.53 ± 0.03 and 0.24 ± 0.03, were obtained from the saturation curves when β₂AR-GFP² and CB₁-GFP² were used as BRET acceptors respectively. These saturation curves demonstrate that there is a specific interaction between CB₁ receptors and β₂-adrenoceptors that is observed at low levels of receptor expression.

A BRET competition experiment was performed to demonstrate that the BRET signal between CB₁-Rluc and β₂AR-GFP² could be reduced in a dose-dependent manner by co-transfection with HA-β₂AR (Figure 1C). CB₁-Rluc/β₂AR-GFP² BRET_Eff was significantly reduced by co-transfection with 1 µg of HA-β₂AR cDNA (P < 0.05) and further reduced with 2 and 3 µg of HA-β₂AR plasmid (P < 0.001), but was not significantly affected by 3 µg of RcCMVmGluR6 (P > 0.05) plasmid, indicating the specificity of the CB₁/β₂-adrenoceptors interaction.

The effect of various CB₁ receptor ligands on the BRET_Eff for the CB₁-Rluc/β₂AR-GFP² pair was then assessed (Figure 1D). The CB₁ receptor inverse agonist AM251 resulted in a BRET_Eff that was significantly elevated compared with either the agonist WIN 55,212-2 (WIN) or the neutral antagonist O-2050 (P < 0.05). This finding suggests that AM251 modulates the CB₁/β₂-adrenoceptor heterodimer, either by altering the number of receptors interacting, or by altering the conformation of the dimer, such that it affects the orientation of the BRET donor and acceptor.

Functional interactions between CB₁ receptors and β₂-adrenoceptors in HEK 293H cells affect CB₁ receptor constitutive activity

When CB₁-GFP² was stably expressed in HEK 293H cells a punctate pattern of internalized CB₁ receptor distribution was observed in addition to a less intense pattern of receptors detected at the cell surface. Following addition of the CB₁ receptor inverse agonist AM251 (10 µM, 24 h), CB₁-GFP² expression was redistributed to the cell surface (Figure 2A). Using On-Cell Western blots, a quantitative measure of the effect of AM251 (10 µM, 24 h) was obtained, confirming a significant (P < 0.001, 360%) increase in CB₁-GFP² cell surface expression in response to treatment with the inverse agonist (Figure 2B). These results demonstrate that CB₁ receptors are constitutively active, resulting in a constitutive internalization of the receptor, which can be reversed by the inverse agonist AM251.

Figure 2.

Co-expression of HA-β₂AR reduces the constitutive activity of CB₁-GFP² in 293H cells. (A) Confocal images of 293H cells stably expressing CB₁-GFP² treated for 24 h with 0.05% DMSO vehicle (left panel) or 10 µM AM251 (right panel). Scale bar is 20 µm. (B) On-Cell Western quantitative measure of CB₁ cell surface expression following 24 h AM251 treatment (10 µM) in 293H cells stably expressing CB₁-GFP². ***P < 0.001 compared with DMSO vehicle; n= 4–6. (C) Confocal images of 293H cells transfected with CB₁-GFP² and HA-β₂AR. Left panels show GFP² fluorescence, middle panels are anti-HA immunofluorescence utilizing a Cy3-conjugated secondary antibody, and the right panels are the merged images. Scale bar is 20 µm. (D) On-Cell Western quantitative measure of CB₁-GFP² cell surface expression in 293H cells stably expressing CB₁-GFP² and transiently transfected with pcDNA, HA-β₂AR, HERG-GFP² or mGluR6-GFP². ***P < 0.001 compared with pcDNA transfected cells; n= 4–18. (E) Basal pERK levels in CB₁-GFP²/TreHA-β₂AR cells without and with Dox pretreatment (10 µg·mL⁻¹, 24 h) to induce expression of HA-β₂AR. ***P < 0.001; n= 20. Inset shows On-Cell Western using an anti-HA primary antibody to measure HA-β₂ expression without or with Dox. (F) pERK levels in CB₁-GFP²/TreHA-β₂AR cells treated with 0.05% DMSO vehicle (open bars) or AM251 (1 µM, 10 min, solid bars) in cells without or with Dox pretreatment (10 µg·mL⁻¹, 24 h). ***P < 0.001 compared with respective vehicle controls; n= 20–34. AM251, N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide; DMSO, dimethylsulphoxide; Dox, doxycycline; ERK, extracellular signal-regulated kinase; HERG, human ether-a-go-gorelated gene; mGluR6, metabotropic glutamate receptor 6.

Co-expression of CB₁-GFP² with HA-β₂AR also resulted in increased localization of CB₁-GFP² at the cell surface. Immunofluorescence for the HA tag was carried out on HEK 293H cells transiently transfected with both CB₁-GFP² and HA-β₂AR (Figure 2C). When cells were successfully transfected with both receptor constructs, CB₁-GFP² expression was largely observed at the cell surface (lower panels of Figure 2C), but when CB₁-GFP² was expressed alone, the distribution was consistent with an internalized receptor (upper panels of Figure 2C). In order to measure the effect that HA-β₂AR co-expression had on CB₁-GFP² cell surface expression, On-Cell Western blots were used in CB₁-GFP² cells transiently transfected with HA-β₂AR (Figure 2D). Transient expression of HA-β₂AR in these cells resulted in a significant (P < 0.001, 130%) increase in CB₁-GFP² cell surface expression, while the expression of either HERG-GFP² or mGluR6-GFP² did not increase CB₁-GFP² surface expression.

To assess whether the increased CB₁-GFP² surface expression was the result of decreased constitutive activity of CB₁receptors, basal CB₁ pERK signalling was measured in CB₁-GFP²/TreHA-β₂AR cells treated with Dox to induce HA-β₂AR expression (Figure 2E). Induction of HA-β₂AR resulted in a significant decrease in the basal pERK level of these cells (P < 0.001). In order to show that this decrease in basal pERK was in fact the result of decreased constitutive activity of CB₁ receptors, the ability of AM251 to reduce basal pERK level was then measured in these cells in the absence or presence of co-expressed HA-β₂AR (Figure 2F). AM251 (1 µM, 10 min) significantly reduced basal pERK in the absence of HA-β₂AR (P < 0.001), but had no significant effect on cells co-expressing HA-β₂AR. These findings demonstrate that CB₁ receptors are constitutively active in 293H cells, but that co-expression of β₂-adrenoceptors attenuates this constitutive activity.

CB₁ receptors and β₂-adrenoceptors are co-internalized upon addition of a CB₁ receptor or β₂-adrenoceptor agonist

Co-transfection of CB₁-GFP² and HA-β₂AR in HEK 293H cells resulted in expression of both receptors primarily at the cell membrane (Figure 3A). When these cells were exposed to the β₂-adrenoceptor agonist, isoprenaline (10 µM, 30 min), internalization of not only HA-β₂AR but also CB₁-GFP² was observed. In contrast, when cells were transfected with mGluR6-GFP² and HA-β₂AR and treated with isoprenaline, only HA-β₂AR was internalized, while mGluR6-GFP² remained at the cell surface. On-Cell Western analyses were used for a quantitative measure of the ability of isoprenaline to co-internalize CB₁-GFP² in cells stably expressing CB₁-GFP² and transiently transfected with HA-β₂AR (Figure 3B). In the absence of HA-β₂AR there was no effect on the cell surface distribution of CB₁-GFP² following treatment with isoprenaline. However, isoprenaline treatment of CB₁-GFP² cells transfected with HA-β₂AR resulted in a significant decrease (P < 0.001, 49%) in CB₁-GFP² cell surface expression. As expected, HA-β₂AR cell surface expression was also significantly decreased (P < 0.001) in CB₁-GFP² cells transfected with HA-β₂AR following treatment with isoprenaline. These findings show that CB₁-GFP² is co-internalized with HA-β₂AR following treatment with the β-adrenoceptor agonist, isoprenaline.

Figure 3.

HA-β₂AR and CB₁-GFP² are co-internalized when exposed to either isoprenaline (ISO) or WIN. (A) Confocal images of 293H cells transiently transfected with CB₁-GFP² or mGluR6-GFP² and HA-β₂AR treated for 30 min with vehicle or 10 µM isoprenaline. Left panels show GFP² fluorescence, middle panels are anti-HA immunofluorescence utilizing a Cy3-conjugated secondary antibody, and right panels are merged images. Scale bar is 20 µm. (B) On-Cell Western quantitative measure of CB₁-GFP² or HA-β₂AR cell surface expression in 293H cells stably expressing CB₁-GFP² and transiently transfected with either pcDNA (CB₁-GFP² bars) or HA-β₂AR. Cells were treated with either H₂O vehicle (labelled CB₁-GFP² + HA-β₂) or isoprenaline (10 µM) for 30 min. ***P < 0.001 compared with CB₁-GFP² + HA-β₂AR vehicle-treated groups; n= 6. (C) Confocal images of HEK 293H cells transiently transfected with CB₁-GFP² and HA-β₂AR, upper panels are untreated controls, and lower panels are treated with WIN (10 µM, 30 min). Scale bar is 20 µm. (D) Relative pCREB levels in CB₁-GFP²/TreHA-β₂AR cells pretreated with Dox (10 µM, 24 h), treated for 1 h with either DMSO (0.05%) or WIN (10 µM), followed by 30 min treatment with isoprenaline (0–10 nM), or forskolin (10 µM). *P < 0.05 compared with respective 1 h DMSO treatments; n= 12–20. CREB, cyclic AMP response element binding protein; DMSO, dimethylsulphoxide; Dox, doxycycline; HEK, human embryonic kidney; mGluR6, metabotropic glutamate receptor 6; WIN, WIN 55,212-2.

Similar to the co-internalization observed when cells were co-transfected with CB₁-GFP² and HA-β₂AR and treated with the CB₁ receptor agonist WIN (10 µM, 30 min), internalization of both receptors was observed (Figure 3C). To further examine the functional significance of this co-internalization, the phospho-CREB (pCREB) signalling response to activation of β₂-adrenoceptors was assessed in CB₁-GFP²/TreHA-β₂AR cells pretreated with Dox (10 µM, 24 h) to induce HA-β₂AR expression, before being treated with WIN (10 µM, 1 h) to induce internalization (Figure 3D). In these cells, the ability of 1 nM isoprenaline to increase pCREB was attenuated by WIN treatment (P < 0.05), but the ability of 10 nM isoprenaline or 10 µM forskolin to increase pCREB were unaffected, indicating that the effect is not due to physiological antagonism of the pCREB signalling pathway. Taken together, these data demonstrate co-internalization of β₂-adrenoceptors with CB₁ receptors mediated by the agonist WIN, and that this co-internalization has functional consequences in the cross-desensitization of β₂-adrenoceptors.

Co-expression of CB₁ receptors and β₂-adrenoceptors affects their ability to stimulate ERK and CREB phosphorylation

In-Cell Western analyses were used to assess levels of phospho-ERK (pERK) and pCREB in HEK 293H cells expressing CB₁-GFP² and/or HA-β₂AR. When cells expressing CB₁-GFP² alone were treated with WIN, increases in both pERK (5 min exposure) and pCREB (30 min exposure) were observed. The WIN-dependent pERK response was sensitive to PTx, while the pCREB response was not (data not shown). pERK dose–response measurement following WIN treatment in CB₁-GFP²/TreHA-β₂AR cells that were not induced with Dox (i.e. no HA-β₂AR expression) yielded a pEC₅₀ of 6.85 ± 0.04, E_max of 1.74 ± 0.02 and a Hill coefficient of 1.07 ± 0.11 (Figure 4A). Pretreatment of these cells for 48 h with Dox to induce HA-β₂AR expression, resulted in a significantly different WIN pERK dose–response curve (P < 0.001), with pEC₅₀values of 6.66 ± 0.03, 1.91 ± 0.02 for E_max and a Hill coefficient of 1.64 ± 0.20. However, the PTx sensitivity of the WIN–pERK response was not affected by co-expression of HA-β₂AR with CB₁-GFP² (data not shown). This demonstrates that co-expression of β₂-adrenoceptors enhances CB₁ receptor-dependent pERK signalling.

Figure 4.

Co-expression of HA-β₂AR and CB₁-GFP² alters receptor efficacy to activate ERK and CREB phosphorylation. (A) Dose–response curves for WIN pERK activation in CB₁-GFP²/TreHA-β₂AR cells pretreated for 48 h without or with Dox (10 µg·mL⁻¹); n= 8. Inset shows On-Cell Western using an anti-HA primary antibody in CB₁-GFP²/TreHA-β₂AR cells without or with Dox (10 µg·mL⁻¹, 48 h). (B) Dose–response curves for WIN pCREB activation in CB₁-GFP²/Tet-ON/HA-β₂AR cells pretreated for 48 h without or with Dox (10 µg·mL⁻¹); n= 4–8. (C) WIN–pCREB responses in 293H cells stably expressing CB₁-GFP² alone or CB₁-GFP² and HA-β₂AR. *P < 0.05 compared with 1.0 µM WIN in CB₁-GFP² cells; n= 3. (D) Isoprenaline–pCREB responses in 293H cells stably expressing HA-β₂AR and transiently transfected with pcDNA (solid bars) or CB₁-GFP² (open bars). **P < 0.01; P= 7–8. Inset is On-Cell Western using anti-CB₁ primary antibody of cells transfected with pcDNA or CB₁-GFP². CREB, cyclic AMP response element binding protein; Dox, doxycycline; ERK, extracellular signal-regulated kinase; WIN, WIN 55,212-2.

WIN treatment of CB₁-GFP²/TreHA-β₂AR cells that were not pretreated with Dox resulted in a pCREB dose–response curve with a pEC₅₀ of 6.82 ± 0.08, E_max of 1.79 ± 0.03 and a Hill coefficient of 1.32 ± 0.29 (Figure 4B). Pretreatment of these cells with Dox (10 µg·mL⁻¹, 48 h) resulted in a significantly different (P < 0.001) dose–response curve with values for pEC₅₀of 6.92 ± 0.28, for E_max of 1.51 ± 0.07 and a Hill coefficient of 0.69 ± 0.33. To determine the long-term effect of co-expression of CB₁ receptors and β₂-adrenoceptors, a cell line stably expressing both CB₁-GFP² and HA-β₂AR was employed. In these cells, the WIN-dependent pCREB response was nearly completely abolished at WIN concentrations of 0.3 and 1.0 µM, compared with cells expressing CB₁receptors alone (Figure 4C), indicating that co-expression of β₂-adrenoceptors inhibits the CB₁ receptor-pCREB signalling pathway.

When cells stably expressing HA-β₂AR were treated with isoprenaline, PTx-insensitive increases in both pCREB and pERK were observed (data not shown). In order to determine what influence co-expression of CB₁-GFP² had on the isoprenaline-stimulated pCREB and pERK responses, HA-β₂AR cells were transfected with either pcDNA vector control or CB₁-GFP² 48 h prior to treatment with isoprenaline. Cells transfected with CB₁-GFP² showed significantly greater (P < 0.01) isoprenaline-stimulated pCREB responses than those transfected with vector (Figure 4D), while there was no difference in the isoprenaline-stimulated pERK responses between cells transfected with pcDNA or CB₁-GFP² (data not shown). Again, the PTx insensitivity of the isoprenaline-induced pERK and pCREB responses was not affected by co-expression of CB₁-GFP² (data not shown).

Co-application of WIN and isoprenaline results in an additive response for pERK but not pCREB

Isoprenaline-stimulated pERK dose–response curves from CB₁-GFP²/TreHA-β₂AR cells pretreated with Dox to induce HA-β₂AR expression were generated in the presence of dimethylsulphoxide (DMSO) vehicle, 0.1 and 0.3 µM WIN (Figure 5A). The curves generated in the presence of 0.1 and 0.3 µM WIN were both significantly different from the DMSO vehicle curve (P < 0.001). Similar pEC₅₀ values of 7.57 ± 0.7, 7.48 ± 0.12 and 7.41 ± 0.14 were obtained for DMSO, 0.1 and 0.3 µM WIN respectively. Baseline pERK levels were increased from 0.91 ± 0.03 in DMSO-treated cells to 1.16 ± 0.04 and 1.30 ± 0.04 in 0.1 and 0.3 µM WIN-treated. Similarly, E_max values were also increased from 1.59 ± 0.04 in DMSO-treated cells to 1.84 ± 0.07 and 1.92 ± 0.07 in cells treated with 0.1 and 0.3 µM WIN respectively. This pattern of increased baseline pERK and E_max, but unchanged pEC₅₀ is consistent with additive WIN- and isoprenaline-induced pERK responses. In contrast, when 1 µM WIN was co-applied with isoprenaline (1 and 10 nM) to these cells, the pCREB response was unchanged compared with isoprenaline application alone (Figure 5B). To determine if the pCREB response was saturated, 10 µM forskolin was co-applied with isoprenaline (1 and 10 nM); it was found that at 1 but not 10 nM isoprenaline, forskolin significantly further elevated pCREB levels over isoprenaline treatment alone (P < 0.001). This demonstrates that WIN and isoprenaline show additive pERK but not pCREB responses in cells co-expressing these two receptors.

Figure 5.

Co-application of isoprenaline (ISO) and WIN results in an additive increase in ERK phosphorylation. (A) Dose–response curves for pERK activation following co-exposure of isoprenaline and 0.05% DMSO vehicle, 0.1 µM WIN or 0.3 µM WIN in CB₁-GFP²/TreHA-β₂AR pretreated for 24 h with 10 µg·mL⁻¹ Dox; n= 6. (B) pCREB responses in CB₁-GFP²/Tet-ON/HA-β₂AR cells pretreated for 24 h with Dox (10 µg·mL⁻¹) then co-exposed to isoprenaline (0, 1 and 10 nM) and DMSO vehicle (0.05%), WIN (1.0 µM) or forskolin (10 µM). *P < 0.05, ***P < 0.001 compared with DMSO, or isoprenaline alone groups, respectively; n= 10–28. CREB, cyclic AMP response element binding protein; DMSO, dimethylsulphoxide; Dox, doxycycline; ERK, extracellular signal-regulated kinase; WIN, WIN 55,212-2.

The CB₁ inverse agonist AM251 but not the neutral antagonist O-2050 inhibits the isoprenaline-stimulated pERK response

Dose–response curves for pERK were generated in CB₁-GFP² cells for the CB₁ receptor agonist WIN, the inverse agonist AM251, and the neutral antagonist O-2050 (Figure 6A). AM251 produced a dose-dependent decrease in pERK from baseline to 0.78 ± 0.03 with a pEC₅₀ of 8.2 ± 0.4, while O-2050 had no effect on basal pERK levels. Although both AM251 and O-2050 were capable of completely blocking the WIN–pERK response in these cells (data not shown), AM251 acts as an inverse agonist, while O-2050 acts as a neutral antagonist.

Figure 6.

The CB₁ inverse agonist AM251 but not the CB₁ neutral antagonist O-2050 attenuates isoprenaline (ISO)-induced ERK phosphorylation. (A) pERK dose–response curves for WIN, AM251 and O-2050 in 293H cells stably expressing CB₁-GFP². WIN exposures were for 5 min, while AM251 and O-2050 exposures were for 15 min; n= 6–16. (B) Isoprenaline pERK dose–response curves for CB₁-GFP²/Tet-ON/HA-β₂AR cells pretreated for 24 h with Dox (10 µg·mL⁻¹) and 15 min with 0.05 % DMSO (squares), AM251 (triangles) or O-2050 (inverted triangles); n= 9–12. Each curve is normalized to the pERK level in cells untreated with isoprenaline but pretreated with the DMSO, AM251 or O-2050. The effect of 20 min AM251 exposure on the basal pERK level in these cells is indicated by the open triangle. (C) Isoprenaline pCREB dose–response curves in CB₁-GFP²/TreHA-β₂AR cells pretreated for 24 h with Dox (10 µg·mL⁻¹) and 15 min with 0.05% DMSO, AM251 or O-2050; n= 6–9. AM251, N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide; CREB, cyclic AMP response element binding protein; DMSO, dimethylsulphoxide; Dox, doxycycline; ERK, extracellular signal-regulated kinase; O-2050, (6aR,10aR)-3-(1-methanesulfonylamino-4-hexyn-6-yl)-6a,7,10,10a-tetrahydro-6,6,9-trimethyl-6H-dibenzo[b,d]pyran; WIN, WIN 55,212-2.

Isoprenaline-mediated pERK dose–response curves were produced in CB₁-GFP²/TreHA-β₂AR cells pretreated for 24 h with Dox to induce HA-β₂AR expression then 15 min with either DMSO vehicle, AM251 or O-2050 (Figure 6B). Following DMSO pretreatment, the isoprenaline dose–response curve had a pEC₅₀ of 8.57 ± 0.15 and an E_max of 1.71 ± 0.05. This dose–response curve was significantly altered by pretreatment with AM251 (P < 0.001) but not O-2050. AM251 pretreatment did not affect the pEC₅₀ producing a value of 8.70 ± 0.28, but did result in a much lower E_max value of 1.45 ± 0.06. These findings demonstrate that the CB₁ receptor inverse agonist AM251, but not the neutral antagonist O-2050 attenuates pERK signalling mediated by β₂-adrenoceptors in HEK 293 cells.

Isoprenaline-mediated pCREB dose–response curves were also generated from CB₁-GFP²/TreHA-β₂AR cells pretreated for 24 h with Dox and 15 min with either DMSO vehicle, AM251 or O-2050 (Figure 6C). The isoprenaline pCREB dose–response curve with DMSO pretreatment had a pEC₅₀ of 8.06 ± 0.07 and E_max of 1.98 ± 0.05. Neither pretreatment with AM251 or O-2050 significantly altered the isoprenaline pCREB dose–response curve in these cells. In addition, the β₂-adrenoceptor inverse agonist timolol had no effect on CB₁ receptor-mediated pERK or pCREB responses (data not shown).

Interactions between pERK signalling induced by CB₁ receptors and β₂-adrenoceptors in HTM cells

Immunofluorescence studies in HTM cells demonstrated clear labelling of endogenous CB₁ receptors and β₂-adrenoceptors (Figure 7A) when compared with secondary antibody-only controls (data not shown). CB₁ receptor and β₂-adrenoceptor expression was detected at the cell membrane in HTM cells, with co-localization of the two receptors apparent.

Figure 7.

Interactions between CB₁ receptor and β₂-adrenoceptor pERK signalling in HTM cells. (A) Confocal immunofluorescence images of HTM cells using an anti-CB₁ primary antibody with a FITC-conjugated secondary antibody (left panel), an anti-β₂-adrenoceptor primary antibody with a Cy3-conjugated secondary antibody (middle panel) and the merged image (right panel). Scale bar is 20 µM. (B) Isoprenaline (ISO, 1 and 10 nM, 30 min) and forskolin (10 µM, 30 min) pCREB responses in HTM cells following 1 h DMSO or WIN (10 µM) pretreatment. **P < 0.01 compared with corresponding isoprenaline concentration pretreated with DMSO; n= 18–24. (C) pERK responses in HTM cells following a 10 min exposure to either aqueous vehicle and isoprenaline (100 nM) treatments, or to 0.05% DMSO vehicle and WIN (10 µM). Pretreatment with ICI (1 µM), CGP (1 µM), AM251 (1 µM) or AM630 (1 µM) was for 15 min and with PTx (100 ng·mL⁻¹) for 24 h before isoprenaline or WIN application. PTx, ICI, CGP and AM251 bars are expressed as relative pERK level normalized to vehicle-treated cells that were also pretreated with the same antagonist or toxin to eliminate any affect these compounds may have had on the basal pERK levels. **P < 0.01 compared with vehicle, ***P < 0.001 compared with vehicle or DMSO; n= 8–27. (D) pERK dose–response in HTM cells following 10 min co-application of isoprenaline with 0.05% DMSO or 1 µM WIN; n= 13–16. (E) Isoprenaline pERK dose–response curves in HTM cells following 15 min pretreatment with either 0.05% DMSO or 1 µM AM251, followed by 5 min treatment with isoprenaline; n= 12–24. (F) Isoprenaline pERK dose–response curves in HTM cells following 15 min pretreatment with 0.05% DMSO or 1 µM O-2050 and 5 min exposure to isoprenaline; n= 12–16. AM251, N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide; CREB, cyclic AMP response element binding protein; DMSO, dimethylsulphoxide; ERK, extracellular signal-regulated kinase; HTM, human trabecular meshwork; O-2050, (6aR,10aR)-3-(1-methanesulfonylamino-4-hexyn-6-yl)-6a,7,10,10a-tetrahydro-6,6,9-trimethyl-6H-dibenzo[b,d]pyran; PTx, Pertussis toxin; WIN, WIN 55,212-2.

In order to assess the role CB₁/β₂-adrenoceptors heterodimerization might play in HTM cells, cross-desensitization of pCREB signalling induced by activation of β₂-adrenoceptors, following treatment with the CB₁ receptor agonist WIN was measured (Figure 7B). In these cells following a 1 h pretreatment with WIN (10 µM) the isoprenaline–pCREB response was significantly (P < 0.01) reduced. In contrast, WIN pretreatment had no effect on the forskolin-stimulated pCREB in these cells, indicating that the effect is not due to a non-specific effect on the pCREB signalling pathway by CB₁ receptors. These findings indicate that, as in HEK 293 cells, the CB₁ receptor agonist WIN produces a cross-desensitization of β₂-adrenoceptors.

To demonstrate function of CB₁ receptors and β₂-adrenoceptors in HTM cells, the abilities of adrenergic and CB receptor ligands to affect pERK levels were examined using In-Cell Western analysis (Figure 7C). In these cells, isoprenaline treatment (100 nM) produced a significant decrease in basal pERK (P < 0.001) that was blocked by the selective β₂-adrenoceptor antagonist ICI 118,551 (1 µM, P < 0.001), but not by the selective β₁-adrenoceptor antagonist CGP 20712 (1 µM), or by PTx (100 ng·mL⁻¹). In contrast, WIN treatment (10 µM) of HTM cells resulted in a significant increase in pERK (P < 0.001) that was blocked by the selective CB₁ receptor inverse agonist AM251 (1 µM, P < 0.001) as well as by PTx (100 ng·mL⁻¹, P < 0.001), but not by the selective CB₂ receptor antagonist AM630 (1 µM). Treatment of HTM cells with AM251 did not alter basal pERK levels in HTM cells, suggesting CB₁ receptors are not constitutively coupled to pERK signalling in these cells (data not shown). Together these findings demonstrate that both CB₁ and β₂AR are functionally expressed in HTM cells, and that activation of these receptors are capable of modulating pERK.

Isoprenaline-mediated pERK dose–response curves in HTM cells were generated in the presence of either DMSO vehicle or WIN (10 µM) (Figure 7D). The two isoprenaline dose–response curves were significantly different (P < 0.001) with values for pEC₅₀ of 8.49 ± 0.08 and 8.91 ± 0.12; baseline pERK of 1.00 ± 0.01 and 1.28 ± 0.02; E_max of 0.75 ± 0.01 and 0.97 ± 0.02 for the curves in the presence of DMSO and WIN respectively. These data show that the WIN increase and the isoprenaline decrease in pERK inhibit each other to produce no net response when WIN and isoprenaline are applied together.

Isoprenaline-mediated pERK dose–response curves were generated in HTM cells following 15 min pretreatment with either DMSO vehicle, AM251 (1 µM),\ or O-2050 (1 µM) (Figure 7E and F). Pretreatment with AM251 resulted in a significantly altered isoprenaline dose–response curve in HTM cells (P < 0.001), which was unaffected by pretreatment with O-2050 (P > 0.05). Specifically, AM251 pretreatment resulted in an isoprenaline dose–response curve with a pEC₅₀ of 9.10 ± 0.65 and a Hill coefficient of −0.34 ± 0.25, compared with the pEC₅₀ of 8.02 ± 0.16 and Hill coefficient of −1.01 ± 0.32 for the DMSO vehicle-pretreated cells. These curves demonstrate that, as observed in HEK 293 cells, the CB₁ receptor inverse agonist AM251 but not the CB₁ receptor neutral antagonist O-2050 was capable of altering the pERK response to β₂-adrenoceptors in HTM cells.

Discussion and conclusion

The ability of rhodopsin-like family A GPCRs to interact with each other as dimers or higher order oligomers has generated considerable interest in recent years. These complexes have received much attention, not only because they are formed by many different GPCRs, but also because they appear to influence nearly every aspect of GPCR function (Terrillon and Bouvier, 2004; Pfleger and Eidne, 2005). The findings presented here demonstrating that the CB₁ receptor and β₂-adrenoceptor BRET pair produce significantly increased BRET_Eff that is saturable and can be diluted by an untagged β₂-adrenoceptor construct strongly suggests that CB₁ receptors and β₂-adrenoceptors can physically interact with each other. Similarly, the observed increased BRET_Eff values for CB₁/CB₁ receptor and β₂/β₂-adrenoceptor BRET pairs demonstrates that consistent with previous reports both of these receptors can also form homodimers in these cells (Hebert et al., 1996; Angers et al., 2000; Wager-Miller et al., 2002). In comparing the BRET saturation curves for the CB₁ receptor homodimer with the CB₁/β₂-adrenoceptor heterodimer, it is apparent that the BRET₅₀ value for the homodimer is lower than that of the heterodimer. This suggests that if these two receptors are expressed at similar levels, the CB₁ receptor will preferentially form the homodimer over the heterodimer, although caution should be exercised in drawing such direct comparisons with two distinct BRET pairs (Mercier et al., 2002). In addition, the observation that treatment with the CB₁ receptor inverse agonist AM251 altered BRET_Eff compared with either the agonist WIN or neutral antagonist O-2050, further supports the conclusion of a specific interaction, and implies that AM251 either facilitates increased interactions between the two receptors or alters the conformation of the heterodimer in such a way as to increase the BRET_Eff (Ayoub and Pfleger, 2009).

Functional interactions between CB₁ receptors and β₂-adrenoceptors were assessed by first examining the trafficking of these two receptors. When expressed in HEK 293 cells, CB₁ receptors are highly constitutively active, resulting in constitutive internalization of the receptor (Leterrier et al., 2004; Ellis et al., 2006; Bohn, 2007). This is consistent with the punctate expression pattern observed when CB₁-GFP² was expressed in HEK 293H cells, which reverts to a cell surface expression pattern upon addition of the CB₁ receptor inverse agonist AM251. In contrast, when β₂-adrenoceptors are expressed in HEK 293 cells, they are predominantly localized to the cell surface (von Zastrow and Kobilka, 1992; Sunaguchi et al., 2003). This non-overlapping distribution of CB₁-GFP² and β₂-adrenoceptors indicates that in order for these receptors to physically interact with each other, the subcellular expression of one receptor must be affected by the presence of the other. Our results demonstrate that when cells co-expressed CB₁-GFP² and HA-β₂AR, CB₁-GFP² localization was indeed shifted towards the cell surface compared with when CB₁-GFP² was expressed alone, thus allowing for the physical interaction to take place. One possible mechanism for this would be that β₂-adrenoceptors tempered the constitutive activity of CB₁ receptors, as has been previously shown to occur in the β₁/β₂-adrenoceptor heterodimer (Zhu et al., 2005). This could explain why β₂-adrenoceptors shift CB₁ receptor expression towards the cell membrane, which does not seem to be the case when CB₁ receptors interact with other GPCRs. For example, heterodimerization of CB₁ receptors with the orexin-1 receptor did not affect the internalized localization of CB₁ receptors, but instead altered the distribution of the orexin-1 receptor (Ellis et al., 2006). In addition, when CB₁ receptors and another interacting GPCR, the µ-opioid receptor (Rios et al., 2006), were co-expressed neither CB₁ nor µ-opioid receptor localization was affected, CB₁ receptors remaining largely internalized while the µ-opioid receptors stayed at the cell surface (Ellis et al., 2006; Canals and Milligan, 2008).

Further support for the notion that CB₁/β₂-adrenoceptor dimerization attenuates CB₁ receptor constitutive activity comes from the observation that basal pERK levels in CB₁-GFP²/TreHA-β₂AR cells were decreased following the induction of HA-β₂AR expression, as was the ability of the CB₁ receptor inverse agonist AM251 to decrease the basal pERK level. In addition, the fact that either CB₁ or β₂-adrenoceptor agonists were able to produce co-internalization of both receptors, a phenomenon reported for several other GPCR dimers (Terrillon and Bouvier, 2004), indirectly suggests that β₂-adrenoceptors suppress constitutive activity of CB₁ receptors. Specifically, as WIN produced co-internalization of β₂-adrenoceptors and the constitutive trafficking of GPCRs generally follows a similar mechanism to agonist-induced receptor internalization (Leterrier et al., 2004; Marion et al., 2004), it should be expected that constitutive internalization of CB₁ receptors would produce co-internalization of β₂-adrenoceptors, as was the case when CB₁ receptors were co-expressed with the orexin-1 receptor (Ellis et al., 2006). The fact that this was not observed, and that instead CB₁ receptors were redistributed towards the cell surface, suggests that CB₁ receptors are more likely to have a reduced constitutive internalization when β₂-adrenoceptors are present.

G protein-coupled receptor heterodimerization may also influence the signalling pathways activated by the receptors present in the complex. In HEK 293H cells stably expressing CB₁-GFP² the cannabinoid agonist WIN produced both a PTx-sensitive increase in pERK and a PTx-insensitive increase in pCREB, indicating that CB₁-GFP² couples to both G_i/o and non-G_i/o pathways in these cells. Based on previous reports that in the presence of PTx, CB₁ activates G_s to increase cAMP production (Maneuf and Brotchie, 1997; Jarrahian et al., 2004; Kearn et al., 2005), it is likely that the observed non-G_i/o pCREB–WIN response in these cells occurs via G_s. The fact that the induction of HA-β₂AR expression in cells stably expressing CB₁-GFP² increased both the E_max and Hill coefficient of the WIN-mediated pERK dose–response, while decreasing the E_max and Hill coefficient of the WIN-mediated pCREB response suggests that HA-β₂AR alters the G protein coupling preference of CB₁ receptors. Specifically, the presence of β₂-adrenoceptors shifts the G_i/o to G_s coupling ratio of CB₁ receptors towards increased G_i/o, but decreased G_s coupling. GPCR heterodimerization has previously been shown to alter G protein coupling and similar signalling effects have previously been reported for both CB₁ receptors as well as β₂-adrenoceptors (Breit et al., 2004; Kearn et al., 2005; Mcgraw et al., 2006). It therefore is plausible that the physical interaction between CB₁ receptors and β₂-adrenoceptors directly affects CB₁ receptor–G protein coupling, although it also cannot be ruled out that the shift in CB₁ receptor coupling could be the result of G_s sequestration by β₂-adrenoceptors (Vasquez and Lewis, 2003).

Similar to the changes observed in CB₁ receptor signalling, β₂-adrenoceptor signalling pathways were also affected by the presence of CB₁ receptors. Co-expression of CB₁ receptors and β₂-adrenoceptors resulted in an increase in the isoprenaline-stimulated pCREB response with no change in the isoprenaline-mediated pERK response. The enhanced pCREB response is consistent with increased β₂-adrenoceptor–G_s coupling that may result either directly from the physical interaction with CB₁ receptors, or indirectly from decreased β₂-adrenoceptor–G_i coupling caused by CB₁ receptor–G_i sequestration (Vasquez and Lewis, 1999).

Co-application of WIN and isoprenaline to cells expressing both CB₁-GFP² and HA-β₂AR resulted in an additive effect on pERK levels, but not on pCREB. The additive pERK response might be expected as WIN and isoprenaline activate pERK through different signalling pathways, WIN through PTx-sensitive G_i/o and isoprenaline through PTx-insensitive G_s or β-arrestin (Demuth and Molleman, 2006; Shenoy et al., 2006). However, the lack of an additive effect on pCREB is more interesting. Although activation of CB₁ receptors has previously been shown to increase CREB phosphorylation (Casu et al., 2005), the underlying pathway has not been examined. Our observations suggest that the WIN-mediated pCREB response is not G_i/o-mediated, as it was insensitive to PTx, and that instead it is likely to be the result of CB₁ receptor–G_s coupling. As β₂-adrenoceptors also activates pCREB via G_s, the lack of an additive effect may be due to overlap in the two pathways. However, the fact that the pCREB response is not additive even at sub-maximal concentrations of isoprenaline suggests that instead it may arise from physiological antagonism resulting from CB₁ receptor-induced activation of G_i, resulting in inhibition of adenylyl cyclase.

Another possible functional consequence of GPCR heterodimerization is altered receptor pharmacology of one or both receptors in the complex (Milligan, 2004; Terrillon and Bouvier, 2004). Our results demonstrate that the CB₁ receptor inverse agonist AM251 inhibits β₂-adrenoceptor-mediated pERK signalling, but had no effect on pCREB signalling, while the CB₁ neutral antagonist O-2050 did not affect either β₂-adrenoceptor-mediated pERK or pCREB activation. Recently, a similar result was reported where the CB₁ receptor inverse agonist SR141716A enhanced pERK signalling of the µ-opioid receptor, while the neutral antagonist O-2050 had no effect (Canals and Milligan, 2008). In this study the result was attributed to the inverse agonist blocking constitutive activity of CB₁ receptors and not to CB₁/µ-opioid receptor heterodimerization because the two receptors were not detected in the same subcellular location. However, in the present study AM251 inhibited rather than enhanced the β₂-adrenoceptor-mediated pERK response, suggesting that the AM251 effect is not simply the result of blocking CB₁ receptor constitutive activity. Instead, our results are best explained by the concept that in a heterodimer, one GPCR can allosterically modulate the second receptor in the complex (Milligan and Smith, 2007). As an inverse agonist, such as AM251, will drive CB₁ receptors from a constitutively active to an inactive state, it is conceivable that such a conformational change would, through heterodimerization, allosterically influence β₂-adrenoceptors and their subsequent signalling. If so, it follows that the CB₁ receptor neutral antagonist O-2050 should not alter the conformation of CB₁ receptors, as a neutral antagonist shows no preference for either the active or inactive states of the receptor, and therefore should not influence β₂-adrenoceptor signalling. Additional support for this conclusion may be drawn from the observation that AM251 treatment resulted in a change in the BRET_Eff compared with O-2050. Although alterations in BRET_Eff by ligands may be caused by changes in the number of receptors interacting as heterodimers, such ligand-mediated changes in BRET_Eff are now believed to be more representative of conformational changes within the heterodimer (Ayoub et al., 2002; Ayoub and Pfleger, 2009). Therefore, our observation that AM251 produces a change in BRET_Eff compared with O-2050 supports the notion that this inverse agonist is allosterically modulating the function of β₂-adrenoceptors through the CB₁/β₂-adrenoceptor heterodimer.

CB₁ receptors and β₂-adrenoceptors are co-expressed in many tissues and cells, including parts of the cardiovascular system, female reproductive tract, brain, eye and bone (Jampel et al., 1987a; Wanaka et al., 1989; Tsou et al., 1998; Stamer et al., 2001; Wang et al., 2004; Pacher and Hasko, 2008). However, few studies have directly examined interactions between these two receptors in vivo. Instead, most work on interactions between the cannabinoid and adrenergic systems has focused on the presynaptic ability of CB receptors to inhibit noradrenergic neurotransmission (Schlicker et al., 1997; Schultheiss et al., 2005; Pakdeechote et al., 2007; Tam et al., 2008). We therefore examined the potential impact of a CB₁/β₂-adrenoceptor interaction on pERK signalling in primary HTM cells.

The trabecular meshwork represents the primary route of aqueous humour outflow in the human eye and as a result represents a key therapeutic target for the regulation of IOP and for the development of novel treatments for glaucoma (Kaufman et al., 1999; Ferrer, 2006). HTM cells have been shown to express β₂-adrenoceptors, activation of which leads to increased cAMP levels and facilitates increased aqueous humour outflow (Jampel et al., 1987b; Erickson-Lamy and Nathanson, 1992). In addition, trabecular meshwork cells also express CB₁ receptors, activation of which appears to increase aqueous humour outflow through activation of pERK (Straiker et al., 1999; Njie et al., 2006, Njie et al., 2008). Consistent with this previous work, our demonstration that HTM cells express both CB₁ receptors and β₂-adrenoceptors, that these receptors are functionally active in pERK signalling, and that they are co-localized at the cell membrane, suggests that any physical or functional interaction between these receptors may be relevant in HTM, and as a result in the regulation of aqueous humour outflow.

To determine if the co-internalization of CB₁ receptors and β₂-adrenoceptors observed in HEK 293H cells was also important to the function of these receptors in HTM cells, the ability of WIN to desensitize the isoprenaline-induced pCREB response in these cells was examined. As in HEK 293H cells, when HTM cells were pretreated with WIN, their isoprenaline–pCREB response was attenuated, while their response to forskolin was unaffected. Such cross-desensitization has previously been reported for other GPCR heterodimers (Pfeiffer et al., 2002) and supports the concept that CB₁ receptors and β₂-adrenoceptors form heterodimers in these cells, and further, that this heterodimer leads to cross-desensitization of β₂-adrenoceptors.

To examine the functional interaction between CB₁ receptors and β₂-adrenoceptors in HTM cells, the ability of these receptors to affect pERK signalling was examined. In these cells, β₂-adrenoceptor activation resulted in a PTx-insensitive decrease in pERK, while CB₁ receptor activation led to a PTx-sensitive increase in pERK. As seen in the HEK 293H cells, co-application of the CB₁ receptor and β₂-adrenoceptor agonists to HTM cells produced an additive response; although given that the CB₁ receptor response was positive while the β₂-adrenoceptor response was negative in HTM cells, the overall results was net inhibitory. Also similar to that observed in HEK 293H cells, it was found that the CB₁ receptor inverse agonist AM251 but not the neutral antagonist O-2050 altered the pERK response to the β₂-adrenoceptor agonist isoprenaline. The combination of an increased pEC₅₀, yet decreased Hill coefficient for the isoprenaline–pERK response when AM251 was present resulted in an increased ability of lower doses of isoprenaline to reduce pERK levels. Although this potentiation of the isoprenaline–pERK response by AM251 in HTM cells would seem to conflict with the inhibition AM251 observed in HEK 293H cells, as the HEK 293H–β₂AR response was a positive pERK response, while the HTM–β₂AR pERK response was negative it is conceivable that inhibition of a positive response and potentiation of a negative one could in fact be mediated by a similar mechanism.

Another interesting finding from our study is that the natures of ligand-induced effects differ depending on the cellular context. Other studies have demonstrated that both β₂-adrenoceptor agonists and inverse agonists can activate the ERK pathway in HEK cells through multiple mechanisms (Azzi et al., 2003; Baker et al., 2003; Shenoy et al., 2006). However, to our knowledge, our findings in HTM cells are the first demonstration that a β₂-adrenoceptor agonist can inhibit ERK phosphorylation, given the proper cellular context. Our observation that the CB₁ receptor inverse agonist AM251 alters β₂-adrenoceptor signalling in cells endogenously expressing both CB₁ receptors and β₂-adrenoceptors may be particularly relevant in vivo due to the interest in CB₁ receptor inverse agonists for the clinical treatment of obesity, metabolic syndrome, type II diabetes and addiction (Xie et al., 2007; Le foll et al., 2008; Van diepen et al., 2008; Vemuri et al., 2008). Our results suggest that these CB₁ receptor inverse agonists may exert their effects not only by directly blocking CB₁ receptors, but also by indirectly affecting β₂-adrenoceptor function in cells that co-express these receptors. In keeping with this, AM251 was found in one study to block the regional haemodynamic response to lipopolysaccharide in rats, a response that was attributed to β-adrenoceptor activation and was also blocked by the selective β₂-adrenoceptor antagonist ICI 118,551 (Gardiner et al., 2005). In addition, the recent observation that CB₁ receptors are expressed in murine cardiomyocytes (Mukhopadhyay et al., 2007) suggests that some of the cardiovascular effects of CB₁ receptor inverse agonists (Van diepen et al., 2008) may be the result of indirect actions on β₂-adrenoceptors in addition to the direct blockade of CB₁ receptors.

The present study demonstrates a physical and functional interaction between CB₁ receptors and β₂-adrenoceptors. These results suggest a more complex picture than previously surmised in which interactions between the cannabinoid and the adrenergic systems were primarily attributed to CB₁ receptor-mediated presynaptic inhibition of noradrenergic transmission. We suggest, in light of our findings and the wide biological and pharmacological importance of the β-adrenoceptor and endocannabinoid systems, a re-evaluation of the nature of previously reported functional interactions between the cannabinoid and β-adrenoceptor systems in cells and tissues that co-express these two receptors.

Glossary

Abbreviations:

2-AG

2-arachidonoylglycerol

AEA

N-arachidonoylethanolamine, anandamide

BRET

bioluminescence resonance energy transfer

BRET_Eff

BRET efficiency

CREB

cyclic AMP response element binding protein

DMEM

Dulbecco's modified Eagle's medium

DMSO

dimethylsulphoxide

Dox

doxycycline

ERK

extracellular signal-regulated kinase

FBS

fetal bovine serum

GPCR

G protein-coupled receptor

HERG

human ether-a-go-go related gene

HTM

human trabecular meshwork

IOP

intraocular pressure

mGluR6

metabotropic glutamate receptor 6

PTx

Pertussis toxin

Rluc

Renilla luciferase

WIN

WIN 55,212-2

Statement of conflict of interests

The authors report no conflict of interest.