The G protein–coupled receptor 55 (GPR55) is a lysophosphatidylinositol (LPI) receptor that is also responsive to certain cannabinoids. Although GPR55 has been implicated in several (patho)physiologic functions, its role remains enigmatic owing mainly to the lack of selective GPR55 antagonists. Here we show that the compound CID16020046 ((4-[4-(3-hydroxyphenyl)-3-(4-methylphenyl)-6-oxo-1H,4H,5H,6H-pyrrolo[3,4-c]pyrazol-5-yl] benzoic acid) is a selective GPR55 antagonist. In yeast cells expressing human GPR55, CID16020046 antagonized agonist-induced receptor activation. In human embryonic kidney (HEK293) cells stably expressing human GPR55, the compound behaved as an antagonist on LPI-mediated Ca2+ release and extracellular signal-regulated kinases activation, but not in HEK293 cells expressing cannabinoid receptor 1 or 2 (CB1 or CB2). CID16020046 concentration dependently inhibited LPI-induced activation of nuclear factor of activated T-cells (NFAT), nuclear factor κ of activated B cells (NF-κB) and serum response element, translocation of NFAT and NF-κB, and GPR55 internalization. It reduced LPI-induced wound healing in primary human lung microvascular endothelial cells and reversed LPI-inhibited platelet aggregation, suggesting a novel role for GPR55 in platelet and endothelial cell function. CID16020046 is therefore a valuable tool to study GPR55-mediated mechanisms in primary cells and tissues.
Cannabinoids bind to and induce signaling via the cannabinoid 1 (CB1) and the cannabinoid 2 (CB2) receptors. Nevertheless, studies on cannabinoid receptor knockout mice suggested additional cannabinoid-sensitive targets (Mackie and Stella, 2006; Brown, 2007). One receptor for small lipid mediators and synthetic cannabinoids is the G protein–coupled receptor 55 (GPR55). GPR55 is highly abundant in the central nervous system as well as in intestine, bone marrow, spleen, platelets, and immune and endothelial cells (Sawzdargo et al., 1999; Ryberg et al., 2007; Waldeck-Weiermair et al., 2008; Pietr et al., 2009; Balenga et al., 2011a; Henstridge et al., 2011; Rowley et al., 2011). Moreover, GPR55 has been detected in a variety of cancer tissues and cancer cell lines (Ford et al., 2010; Andradas et al., 2011; Huang et al., 2011; Pineiro et al., 2011; Perez-Gomez et al., 2012). Several endogenous GPR55 signaling pathways have been described to date despite controversial findings concerning its agonists and antagonists (Balenga et al., 2011b). One consensus of several groups is that GPR55 couples to Gα13 or Gαq proteins in human embryonic kidney (HEK293) cells that transiently or stably express GPR55 (Ryberg et al., 2007; Lauckner et al., 2008; Henstridge et al., 2009, 2010; Schroder et al., 2010; Sharir and Abood, 2010). In addition, GPR55 has been reported to activate small GTPases (Ryberg et al., 2007; Henstridge et al., 2009; Balenga et al., 2011a) and to induce calcium release from intracellular stores (Oka et al., 2007; Henstridge et al., 2009, 2010; Brown, et al., 2011). Further downstream, GPR55 activation leads to the activation of several transcription factors, such as nuclear factor of activated T-cells (NFAT), nuclear factor κ of activated B cells (NF-κB), serum response element (SRE), cAMP response element–binding protein (CREB), and activating transcription factor 2 (Henstridge et al., 2009, 2010; Oka et al., 2010; Kargl et al., 2012a). In addition, mitogen-activated protein kinases, such as p38 and extracellular signal-regulated kinases (ERK1/2), are activated upon GPR55 stimulation (Henstridge et al., 2010; Oka et al., 2010).
The lipid l-α-lysophosphatidylinositol (LPI) was described as the first endogenous ligand for GPR55 by Oka et al. (2007). In addition, several synthetic CB1 receptor inverse agonists/antagonists, such as 1-(2,4-dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-(1-piperidyl)pyrazole-3-carboxamide (AM251), 1-(2,4-dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-4-morpholinyl-1H-pyrazole-3-carboxamide (AM281), and rimonabant (SR141716A; 5-(4-chlorophenyl)-1-(2,4-dichloro-phenyl)-4-methyl-N-(piperidin-1-yl)-1H-pyrazole-3-carboxamide), have been shown to activate GPR55 (Oka et al., 2007; Ryberg et al., 2007; Henstridge et al., 2009, 2010; Kapur et al., 2009; Yin et al., 2009; Brown et al., 2011). Although several cannabinoid ligands can activate GPR55, the receptor lacks the classic “cannabinoid binding pocket” (Kotsikorou et al., 2011).
Screening approaches have identified selective GPR55 agonists, such as [4-(3,4-dichloro-phenyl)-piperazin-1-yl]-(4′-fluoro-4-methanesulfonyl-biphenyl-2-yl)-methanone (GSK319197A) or GSK494581A (Brown et al., 2011; Kargl et al., 2012a), which generally appear to be inactive at CB1 and CB2 receptors and hence are promising tools to elucidate the pharmacologic, physiologic, and pathophysiologic functions of GPR55. The first such chemical series of synthetic GPR55 agonists to be described were the benzoylpiperazines. Importantly, benzoylpiperazines were independently identified at GlaxoSmithKline (Brown et al., 2011; Kargl et al., 2012a) and by the National Institutes of Health (NIH) Molecular Libraries Probe Identification program (Heynen-Genel et al., 2010b; Kotsikorou et al., 2011). The latter screen used β-arrestin fluorescent protein biosensors and identified further agonists as well as antagonists (Heynen-Genel et al., 2010a,b). However, there is so far only limited characterization of these antagonists in the peer-reviewed literature.
The lipid ligand LPI; the CB1 receptor inverse agonists/antagonists SR141716A, AM251, and AM281; and selective GPR55 agonists were described to activate transcription factors in HEK-293 cells transiently or stably expressing human GPR55 (Ryberg et al., 2007; Henstridge et al., 2009, 2010; Brown et al., 2011; Kargl et al., 2012a). The phytocannabinoid cannabidiol (CBD) was reported to antagonize O-1602-mediated GPR55 activation in HEK-GPR55 cells and human osteoclasts (Ryberg et al., 2007; Whyte et al., 2009). In contrast, CBD had no effect on Ca2+ mobilization and β-arrestin recruitment assays in HEK-GPR55 cells (Kapur et al., 2009). In addition, Heynen-Genel et al. (2010a) did not identify CBD as a GPR55 antagonist in a small-molecule screen for GPR55 antagonists. Although CBD displays only low affinity for CB1 and CB2 receptors, CBD interacts with transient receptor potential (TRP) receptors, such as TRPV2 or TRPM8 (Qin et al., 2008; Pertwee et al., 2010). As such, CBD cannot be considered a specific GPR55 antagonist.
To further elucidate the role of GPR55 in physiology and pathobiology, selective agonists and antagonists are urgently needed.
Here we characterized the novel GPR55 antagonist 4-[4-(3-hydroxyphenyl)-3-(4-methylphenyl)-6-oxo-1H,4H,5H,6H-pyrrolo[3,4-c]pyrazol-5-yl] benzoic acid (CID16020046) in yeast cells and HEK293 cells stably expressing GPR55. In addition, we used CID16020046 as a tool to study the role of GPR55 in endothelial wound healing of primary human lung microvascular endothelial cells (HMVEC-Ls) and human platelet aggregation.
Materials and Methods
CID16020046 was obtained from Molport (Riga, Latvia) and dissolved in dimethylsulfoxide (DMSO). GSK319197A was provided by GlaxoSmithKline (Harlow, UK) and dissolved in DMSO. LPI was purchased from Sigma-Aldrich (Vienna, Austria) and dissolved in H2O.
Cell Culture, Transfections, and Stable Cell Lines
HEK293 cells were cultured in Dulbecco’s modified Eagle’s medium (Life Technology, Vienna, Austria) supplemented with 10% fetal bovine serum (Life Technology) at 37°C in 5% CO2, humidified atmosphere. HEK293 cells stably expressing the human 3× hemagglutinin (HA)-GPR55 (HEK-GPR55), human FLAG-CB1 (HEK-CB1), or human FLAG-CB2 (HEK-CB2) were previously described (Kargl et al., 2012a). All cells were serum-starved in Opti-MEM (Life Technology) before all experiments. Transient transfections were performed using Lipofectamine 2000 following the manufacturer’s instructions (Life Technology). HMVEC-Ls were purchased from Lonza and maintained in EGM-2 MV BulletKit medium (Lonza, Verviers, Belgium).
Yeast Exoglucanase Assay
Derivation of yeast strains and yeast reporter gene assays was described previously (Olesnicky et al., 1999; Brown et al., 2003, 2011). In brief, yeast strain YIG151 contains FLAG-GPR55 and Gpa1/Gα13 chimeric G-protein α-subunit, both integrated chromosomally. YIG151 cells were mixed with buffered growth media, pH 7.0, containing 10 mM 3-aminotriazole and lacking histidine and the Exg1p (exoglucanase) substrate fluorescein-d-glucopyranoside at 10 µM, preincubated with test antagonist or vehicle for 10 minutes, and incubated with test agonist in 384-well microtiter plates (50 μl/well; DMSO final concentration: 1%) for 21 hours. Fluorescein production was quantified using an Envision plate reader (PerkinElmer, Cambridge, UK), and curve-fitting was performed using GraphPad Prism (GraphPad Software, Inc., La Jolla, CA).
Intracellular Ca2+ Release Assays
96-Well FlexStation II Assay.
Agonist-mediated intracellular Ca2+ release was measured in a 96-well plate format (FlexStation II; Molecular Devices, Sunnyvale, CA), as previously described (Sedej et al., 2012). Briefly, HEK293, HEK-GPR55, and HEK-CB1 cells (40,000 cells per well) and HMVECs (10,000 cells per well) were seeded in black 96-well plates. Before each experiment, HEK293 cells were starved for 4 hours in Opti-MEM and HMVEC-Ls for 2 hours in EBM incomplete medium. Next, cells were loaded with the Ca2+ fluorophore (FLEX Calcium Assay Kit; Molecular Devices) for 60 minutes. A subset of cells was pretreated with vehicle or GPR55 antagonist CID16020046 for 10 minutes. Subsequently, cells were stimulated with increasing concentrations of the agonist diluted in assay buffer. Intracellular Ca2+ mobilization was measured immediately after agonist application and recorded in real time for 2 minutes at room temperature in a FlexStation-II System. On Ca2+ release from intracellular stores, the Ca2+ fluorophore was excited with a wavelength of 485 nm and emitted fluorescent light with a wavelength of 525 nm. Cell numbers were determined before Ca2+ measurement in a FlexStation-II (Molecular Devices). Ultimately, relative fluorescence unit readings for each well of the microplate were obtained and normalized to cell number.
Ca2+ Flux Flow Cytometric Assay.
Changes in intracellular Ca2+ levels in HEK293, HEK-GPR55, and HEK-CB1 cells after treatment with GPR55 agonists or antagonists were assessed by flow cytometry as previously described (Heinemann et al., 2003). Cells were starved in Opti-MEM for 4 hours, resuspended in 1 ml of wash buffer (without Ca2+ and Mg2+), and incubated with 5 µM acetoxymethyl ester of Fluo-3 in the presence of 2.5 mM probenecid for 60 minutes at room temperature in the dark. Cells were washed and resuspended in assay buffer (with Ca2+ and Mg2+). Flow cytometric analysis was performed on a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA). After baseline fluorescence had been recorded for 1 minute, cells were treated with the desired concentrations of agonists and measured for another 4 minutes. The changes in intracellular Ca2+ levels were detected as an increase in the fluorescence intensity of the [Ca2+]i-dependent signal of Fluo-3 at 526 nm, and data were normalized to baseline fluorescence.
ERK1/2 phosphorylation was detected as previously described (Kargl et al., 2012a). In brief, HEK293, HEK-GPR55, HEK-CB1, and HEK-CB2 cells were seeded in six-well plates, and confluent wells were serum-starved overnight. Then cells were incubated with prewarmed Opti-MEM containing vehicle (H2O or DMSO, final concentration of 0.025% (Merck, Vienna, Austria), LPI (Sigma-Aldrich), WIN55,212-2 (Tocris, Avonmouth, UK), or CID16020046 (Molport) or combinations thereof for 25 minutes at 37°C. Cells were washed, snap-frozen in liquid nitrogen, and lysed in IP-Buffer (10 mM Tris-HCl pH 7.4, 150 mM NaCl, 25 mM KCl, 1 mM CaCl2, 0.3% Triton X-100, 92 mg/ml sucrose and protease inhibitors (Roche Applied Science, Vienna, Austria). Lysates were resolved by SDS-PAGE (Life Technology) and transferred to a polyvinylidene difluoride membrane (Millipore, Vienna, Austria). Membranes were blocked in Tris-buffered saline/Tween 20 buffer [1 mM CaCl2, 136 mM NaCl, 2.5 mM KCl, 25 mM Tris-HCl, 0,1% (v/v) Tween 20] containing 5% milk, washed in Tris-buffered saline/Tween 20 without milk, and immunoblotted with rabbit anti-pERK1/2 (1:1000) or rabbit anti-tERK1/2 (1:1000) antibodies overnight at 4°C (New England Biolabs, Frankfurt, Germany). Membranes were incubated with horseradish peroxidase–conjugated goat anti-rabbit antibody (1:4000; Jackson ImmunoResearch, Suffolk, United Kingdom) for 2 hours at room temperature, and proteins were visualized with ECL Western Blotting Substrate (Fisher Scientific, Vienna, Austria). At least three independent blots were analyzed for quantification of phosphorylated ERK1/2 (pERK1/2) and total ERK1/2 (tERK1/2) levels using ImageJ Software (NIH, Bethesda, MD), and pERK1/2 was normalized to tERK1/2 levels.
Reporter Gene Assay
Transcription factor luciferase assays were carried out as previously described (Kargl et al., 2012a). Briefly, HEK293, HEK-GPR55, and HEK-CB1 cells were seeded in 96-well plates (40,000 cells/well) and transiently transfected with the cis-reporter plasmids (PathDetect; Stratagene, La Jolla, CA) for NFAT-luc (100–200 ng) SRE-luc (50 ng/well), NF-κB-luc (50 ng/well), or CREB (200 ng/well) using Lipofectamine 2000 (Life Technologies, Inc., Carlsbad, CA). Twenty-four hours post-transfection, cells were incubated with the indicated ligand concentrations for 4 hours in serum-free media at 37°C. The cell number was determined in a FlexStation-II (Molecular Devices). Luciferase activity was visualized using the Steadylite Plus Kit (Packard Instrument Company, Meriden, CT) and measured in a TopCounter (Top Count NXT; Packard) for 5 seconds. Luminescence values are given as relative light units. For reporter gene experiments, relative light units were normalized to the cell number.
HEK-GPR55 cells were seeded in 6-cm dishes and transfected with 8 µg of enhanced green fluorescent protein-p65 plasmid (NF-κB p65-GFP) or NFATc3-GFP plasmid (Henstridge et al., 2009; Waldeck-Weiermair et al., 2008). Twenty-four hours post-transfection, cells were seeded on coverslips and grown to 50% confluency in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. Cells were serum-starved in OPTI-MEM for another 4 hours before treatments. Subsequently, cells were pretreated with 2.5 µM GPR55 antagonist CID16020046 for 10 minutes or stimulated with 2.5 µM LPI for 15 minutes, and the experiment was terminated by fixing cells with 3.7% formaldehyde. Fixed cells were washed twice with Tris-buffered saline (135 mM NaCl, 25 mM Tris-HCl, 1 mM CaCl2, 2.5 mM KCl) and wet-mounted onto microscopy slides with Vectashield (Vector Laboratories, Inc., Burlingame, CA) containing a 4′,6-diamidino-2-phenylindole to visualize nuclei. Images were taken with an Olympus inverted IX70 fluorescence microscope (Olympus America, Center Valley, PA) equipped with a Hamamatsu Orca CCD camera (Hamamatsu Photonics, Hamamatsu, Japan).
Cells were grown on poly-d-lysine (Sigma-Aldrich) coated coverslips to 50% confluence, starved in Opti-MEM overnight, and antibody feeding experiments were performed essentially as described (Kargl et al., 2012b). In brief, living cells were fed with anti-HA-11 antibody (1:1000; Covence, Berkeley, CA) for 30 minutes at 37°C. Subsequently, cells were prestimulated with GPR55 antagonist CID16020046 for 10 minutes and/or stimulated with LPI for 45 minutes. Then cells were fixed in 3.7% formaldehyde, permeabilized in blotto (50 mM Tris-HCl, pH 7.5, 1 mM CaCl2, 0.3% Triton X-100 and 3% milk), and labeled with secondary antibodies [AlexaFluor 488-conjugated IgG1 against the HA-tag (1:1000; Life Technology)] for 20 minutes. Immunolabeled receptors were visualized by using an Olympus inverted IX70 fluorescence microscope.
Reverse-Transcription Polymerase Chain Reaction and Real-Time Polymerase Chain Reaction
Polymerase chain reaction (PCR) methods were conducted as previously described (Kargl et al., 2013). HMVEC-L cells were frozen in liquid nitrogen, and total RNA was extracted using QIAshredder and RNeasy Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. Reverse-transcription PCR was performed with 1 µg of total RNA and a high-capacity cDNA RT kit (Applied Biosystems/Life Technologies) for cDNA transcription. The reverse-transcription PCR program was set at 25°C for 10 minutes, followed by 2 hours at 37°C, with a terminal step at 85°C for 5 minutes. Quantitative PCR was performed using Fast SYBR Green PCR Master Mix (Applied Biosystems) following the manufacturer’s instructions. The following primers were used: GPR55 (forward: 5′-CCTCCCATTCAAGATGGTCC-3′ reverse: 5′- GACGCTTCCGTACATGCTGA-3′), CB1R (forward: 5′-CCTTCCTACCACTTCATCGGC-3′ reverse: 5′-CGTTGCGGCTATCTTTGCG-3′), CB2R (forward: 5′- GACCGCCATTGACCGATACC-3′ reverse: 5′-GGACCCACATGATGCCCAG-3′) and GAPDH (forward: 5′-ATGGGGAAGGTGAAGGTCG-3′ reverse: 5′-GGGGTCATTGATG-GCAACAATA-3′). The specificity of the PCR products was assessed by melting curve analyses, which showed only single amplified products, and agarose gel electrophoresis, which revealed fragments at the expected base-pair sizes. The pcDNA3.1 plasmids encoding GPR55, CB1, or CB2 genes were used for standard curve calculations, and absolute mRNA copy number was calculated.
Endothelial Wound-Healing Assay
HMVECs (50,000 cells per chamber) were grown to confluence on 1% gelatin-precoated Electric Cell-Substrate Impedance Sensing 8W1E polycarbonate arrays containing gold microelectrodes (Applied Biophysics, Troy, NY). Cells were serum-starved for 2 hours before performing impedance measurements at multiple frequencies by using the Electric Cell-Substrate Impedance Sensing system (Applied Biophysics). Confluent layers of endothelial cells were electrically wounded (20 seconds at 3000 μA and 100 kHz), resulting in severe electroporation and subsequent death of the cells situated on the electrodes. Impedance was continuously monitored to detect repopulation of the wound (Keese et al., 2004).
The study was approved by the Institutional Review Board of the Medical University Graz. Blood was drawn from healthy volunteers after they signed an informed consent form. Platelet-rich and platelet-poor plasma was prepared from citrated whole blood by centrifugation, and platelet aggregation was recorded using an APACT 4004 aggregometer (Haemochrom Diagnostika, Essen, Germany) as described previously (Philipose et al., 2010). First, the ADP concentration (2.5–20 µM) that induced near half-maximal aggregation was determined for each donor and was used for further experiments. Samples were then incubated with LPI (300 nM–3 µM–10 µM) for 5 minutes at 37°C, and platelet aggregation was induced by ADP. To determine the effect of the GPR55 antagonists CID16020046 on LPI-induced responses, samples were pretreated with CID16020046 (3 or 10 µM) or vehicle (DMSO diluted in saline) for 5 minutes at 37°C and then incubated with LPI (10 µM), followed by activation of platelets with ADP. Platelet aggregation was expressed as percentage of maximum light transmission, with nonstimulated platelet-rich plasma being 0% and platelet-poor plasma 100%.
Statistical analyses were performed using t tests or analysis of variance for comparisons between multiple groups, followed by a Bonferroni’s post hoc analysis using GraphPad Prism. A P value of <0.05 was considered as statistically significant.
CID16020046 Antagonizes Human GPR55 in Yeast.
CID16020046 was originally described in PubChem to antagonize GPR55-mediated β-arrestin internalization (Data Source: Burnham Center for Chemical Genomics; Source Affiliation: Burnham Institute for Medical Research, La Jolla, CA; Network: the NIH Molecular Libraries Probe Production Centers Network; Assay Provider: Dr. Mary Abood, California Pacific Medical Center Research Institute). Yeast provide a useful host for functional expression of mammalian G protein–coupled receptors (GPCRs) because they can be engineered to remove endogenous receptors and to link activation of the heterologously expressed mammalian receptor to reporter genes (Dowell and Brown, 2002). In this way, GPR55 has previously been shown to respond to GPR55 agonists such as AM251, LPI, and GSK319197A in isolation from other GPCRs or secondary factors (Brown et al., 2011). Here, we used a yeast strain with significant basal (constitutive) exoglucanase activity due to expression of a FLAG-tagged version of human GPR55. CID16020046 (40 nM–10 μM) acted as an inverse agonist, inhibiting GPR55 constitutive activity with IC50 = 0.15 μM (IC50 from Fig. 1B). CID16020046 was tested for antagonist effects in combination with the agonists AM251, GSK522373A, and LPI. The half-maximal effective concentration of AM251 increased as a result of the presence of 10 μM CID16020046 by less than 2-fold (EC50 = 3.9 μM in the absence and 7.2 μM in the presence of CID16020046; Fig. 1A). However, the maximum asymptotes of agonist concentration-response curves to AM251 were significantly depressed by increasing concentrations of CID16020046 (Fig. 1A). GSK522373A is a benzoylpiperazine and close structural analog of GSK319197A and has been described previously as a GPR55 agonist (Brown et al., 2011). CID16020046 caused sequential rightward shift of the agonist concentration-response curve to GSK522373A. The extent of this curve-shift fitted closely to a linear regression model (r2 = 1.00) with a Hill slope of 0.73 (Schild analysis; Fig. 1D), giving an estimated pA2 = 7.3. Again, CID16020046 depressed the maximum effect of GSK522373A (Fig. 1C). LPI was tested up to a limiting concentration of 10 μM, since higher concentrations are toxic to yeast cells (Brown et al., 2011), and hence accurate EC50 values could not be determined. CID16020046 reduced or even abolished the agonist effect of 10 μM LPI on GPR55-expressing yeast cells. In conclusion, CID16020046 is an inverse agonist and antagonist of human GPR55, able to block the effects of multiple chemical classes of GPR55 agonist. The chemical structure of CID16020046 is shown in Fig. 1F.
LPI-Induced Ca2+ Signaling Is Inhibited by CID16020046.
Activation of GPR55 by several ligands can promote intracellular Ca2+ release (Oka et al., 2007; Lauckner et al., 2008; Henstridge et al., 2009, 2010). We tested whether LPI and GSK319197A can induce intracellular Ca2+ release in HEK-GPR55 cells using two different experimental approaches, with adherent cells using the FLEX system and with cells in suspension by flow cytometry. We observed intracellular Ca2+ release in HEK-GPR55 cells on stimulation with increasing concentrations of LPI (Fig. 2, A and D) and GSK319197A (Fig. 2D) in a concentration-dependent manner. To investigate the effect of CID16020046 on GPR55-mediated release of Ca2+ from intracellular stores, we pretreated HEK-GPR55 cells with increasing concentrations of CID16020046 for 15 minutes before exposure to 10 µM LPI (Fig. 2, A and G) or 1 µM GSK319197A (Fig. 2G) (IC50 values are shown in Table 1). CID16020046 inhibited LPI- and GSK319197A- induced Ca2+ mobilization (Fig. 2, A and G). The CB1 agonist (R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone (WIN55,212-2) induced intracellular Ca2+ release in HEK-CB1 cells, but 1 µM WIN55,212-2-induced Ca2+ release in HEK-CB1 cells was not altered in the presence of increasing concentrations of CID16020046 (Fig. 2, B and E). CID16020046 alone failed to induce intracellular Ca2+ release in HEK-GPR55 and HEK-CB1 cells (Fig. 2, A, B, and D). No effect of any ligand on intracellular Ca2+ levels was observed in control HEK293 cells at the concentrations tested (Fig. 2, C and F). These data indicate that CID16020046 acts as a GPR55 antagonist but has no effect in HEK293 and HEK-CB1 cells.
CID16020046 Inhibits GPR55-Mediated ERK1/2 Phosphorylation.
GPR55 activation induces ERK1/2 phosphorylation in several cellular systems (Whyte et al., 2009; Henstridge et al., 2010; Andradas et al., 2011; Kargl et al., 2012a; Perez-Gomez et al., 2012). To test whether CID16020046 acts as a selective GPR55 antagonist, we measured ERK1/2 phosphorylation in HEK293, HEK-GPR55, HEK-CB1, and HEK-CB2 cells in the absence or presence of CID16020046. As previously described (Kargl et al., 2012a), stimulation with 2.5 µM LPI for 25 minutes induced ERK1/2 phosphorylation in HEK-GPR55 cells (Fig. 3A); 2.5 µM of the GPR55 antagonist CID16020046 significantly inhibited the LPI-induced ERK1/2 phosphorylation. Treatment with vehicle and CID16020046 alone showed no ERK1/2 phosphorylation over background (Fig. 3A). 2.5 µM WIN55,212-2 induced ERK1/2 phosphorylation in HEK-CB1 and HEK-CB2 cells, but phospho-ERK1/2 signal was not altered when combined with 2.5 µM CID16020046 (Fig. 3B).
Effects of CID16020046 on GPR55-Mediated Transcription Factor Activation.
Further, we investigated whether GPR55-mediated transcription factor activation can be modulated by CID16020046. Previously, we showed that various transcription factors (i.e., NFAT, SRE, NF-κB and CREB) can be activated via GPR55 (Henstridge et al., 2009, 2010; Kargl et al., 2012a). Here we demonstrate that both the endogenous GPR55 agonist LPI and the selective GPR55 agonist GSK319197A can induce NFAT (Fig. 4A) and NF-κB (Fig. 4E) activation as well as SRE induction (Fig. 4C) in HEK-GPR55 cells. In contrast, CID16020046 alone did not induce GPR55-mediated transcription factor activation (Fig. 4, A, C, and E). Pretreatment with CID16020046 led to a concentration-dependent decrease in GPR55-mediated NFAT activation (Fig. 4B), NF-κB activation (Fig. 4F), and SRE induction (Fig. 4D) in response to 1 μM LPI or GSK319197A (IC50 values are shown in Table 1). These data indicate that CID16020046 can inhibit GPR55-mediated transcription factor activation in HEK-GPR55 cells.
We next tested whether CID16020046 could antagonize CB1-mediated CREB activation. Stimulation of HEK-CB1 with WIN55,212-2 led to concentration-dependent CREB activation (Fig. 4G). Preincubation of HEK-CB1 cells with increasing concentrations of CID16020046 did not result in decreased CB1-mediated CREB activation after stimulation with 1 µM WIN55,212-2 (Fig. 4G). No effect of CID16020046 alone was observed in HEK-CB1 cells (Fig. 4G) and empty HEK293 cells (unpublished data).
To visualize the nuclear translocation of transcription factors in response to activation of cells with LPI, we transiently expressed NFATc3-GFP or NF-κB-p65-GFP in HEK-GPR55 cells. In untreated and vehicle-treated cells, NFATc3-GFP and NF-κB-p65-GFP localized predominantly in the cytosol (Fig. 5, A and G), indicating that NFAT and NF-κB signaling cascades are not activated in the absence of GPR55 agonists. Upon stimulation with 2.5 µM LPI (Fig. 5, B and H) or vehicle + 2.5 µM LPI (Fig. 5, E and K), NFATc3-GFP and NF-κB-p65-GFP rapidly translocated into the nucleus. NFATc3-GFP and NF-κB-p65-GFP were observed in the cytosol upon incubation with 2.5 µM CID16020046 prior to 2.5 µM LPI stimulation (Fig. 5, F and L), indicating that CID16020046 blocks the GPR55-mediated NFAT and NF-κB activation. These data demonstrate that CID16020046 antagonizes GPR55-mediated activation and nuclear translocation of transcription factors but has no effect on CB1-mediated CREB activation.
CID16020046 Inhibits GPR55 Internalization.
GPR55 has been described to internalize rapidly on agonist stimulation in several cell models (Henstridge et al., 2009, 2010; Kapur et al., 2009; Kargl et al., 2012a,b). In general, GPCR antagonists inhibit receptor internalization. Here we set out to investigate whether CID16020046 has the ability to inhibit LPI-induced GPR55 internalization. Antibody feeding experiments in live HEK-GPR55 cells showed that GPR55 is predominantly located on the cell surface under nonstimulated conditions (Fig. 6A). As expected, GPR55 internalized after treatment with 2.5 µM LPI or vehicle + 2.5 µM LPI for 45 minutes (Fig. 6, B and E). Pretreatment with 2.5 µM CID16020046 for 10 minutes inhibited LPI-induced GPR55 internalization (Fig. 6F). Vehicle (DMSO) or CID16020046 alone did not induce GPR55 internalization (Fig. 6, C and D).
Secondary Pharmacology (Selectivity) Profile of CID16020046.
CID16020046 was tested in a panel of routinely run assays to determine its activity against a broad spectrum of GPCRs, ion channels, enzymes (including kinases), and nuclear receptors, 36 targets in total (Supplemental Table 1). The most potent activities detected were for inhibition of phosphodiesterases PDE3A and PDE4B (pIC50 = 5 ± 0.01 and 4.8 ± 0.05, respectively; mean ± S.D., n = 2). CID16020046 was observed to have weak activities close to the top concentrations tested in several other assays: these were for inhibition of acetylcholinesterase (pIC50 = 4.4 ± 0.11), antagonism of the μ-opioid receptor (pIC50 = 4.6 ± 0.01), and blockade of KCNH2, the hERG channel pIC50 = 4.6; n = 1). In all other assays, CID16020046 was inactive up to the highest concentration tested of 25 μM (i.e., pXC50 < 4.6) to 100 μM (i.e., pXC50 < 4.0). Taken together, these data support the conclusion that CID16020046 is selective for GPR55 and give confidence that effects of CID16020046 observed in complex primary cell systems and tissues can be mediated via GPR55.
Functional Characterization of CID16020046 in Primary HMVEC-Ls.
We next investigated whether inhibition of GPR55 in HMVEC-Ls with CID16020046 has an effect on cellular function that involves GPR55. It was previously described that human dermal microvascular endothelial cells express GPR55 and that this receptor might be involved in angiogenesis and endothelial wound-healing capacity (Zhang et al., 2010). Here, we tested whether another primary human endothelial cell line, HMVEC-Ls, express GPR55 and whether this receptor may have a similar physiologic relevance in this cell line. In fact, HMVEC-Ls expressed both GPR55 (Fig. 7A) and CB1 receptors, but only very little CB2 receptor mRNA was detected (Fig. 7A). We then tested whether LPI can induce Ca2+ mobilization and enhance endothelial wound healing in HMVEC-Ls and whether these effects are blocked by pretreatment with CID16020046.
LPI (10 µM) induced intracellular Ca2+ release in HMVEC-L (Fig. 7B), and this effect was inhibited upon pretreatment with 25 and 50 µM CID16020046 (Fig. 7C). Further, we observed that enhanced endothelial wound healing after electric wounding of the monolayer was augmented by 0.1 µM LPI (Fig. 7D) compared to vehicle (H2O, Fig. 7D), indicating that GPR55 activation may be involved in migration of cells and hence in wound closure. Pretreatment with 1 µM CID16020046 (Fig. 7E) abolished the LPI-induced stimulation of wound healing in HMVEC-Ls compared with vehicle control (Fig. 7E). In summary, these data show that CID16020046 is a specific inhibitor of GPR55 function in HMVEC-L cells, which endogenously express this receptor.
CID16020046 Reverses the Inhibitory Effect of LPI on ADP-Induced Platelet Aggregation.
Human platelets were previously shown to express the mRNA for GPR55 (Henstridge et al., 2011; Rowley, et al., 2011). Therefore, we tested whether GPR55 is involved and LPI is capable of modifying platelet function. Up to concentrations of 10 µM, LPI itself did not induce platelet aggregation (n = 4; unpublished data). To investigate the modulatory effects of LPI on platelet aggregation in response to a known inducer, platelets were stimulated with concentrations of ADP (2.5–20 µM) that gave half-maximal aggregatory responses. Under these conditions, LPI significantly attenuated the ADP-induced platelet aggregation by about 20–25% (Fig. 8, A and C). This effect was completely reversed by the GPR55 antagonist CID16020046 (10 µM; Fig. 8, B, and C). At this concentration, the antagonist by its own did not significantly alter ADP-induced platelet aggregation (n = 4; unpublished data). This is a preliminary indication that GPR55 may be involved in the regulation of platelet function.
CID16020046 originates from a high-throughput screen of approximately 300,000 compounds for molecules able to block GPR55-mediated internalization of a β-arrestin-GFP biosensor (http://pubchem.ncbi.nlm.nih.gov/summary/summary.cgi?cid=16020046). This screen was performed by the Abood group at the Burnham Institute as part of the NIH Molecular Library Screening Center Network (MLSCN) program. In this screen, multiple chemical series of both GPR55 agonists and antagonists were identified (Heynen-Genel et al., 2010a,b). CID16020046 does not appear to be one of the GPR55 antagonists chosen by the Abood group for more detailed analysis (Heynen-Genel et al., 2010a,b) but was of interest to us nonetheless because it coincided with a chemical series that had been identified independently in a separate high-throughput screen performed at GlaxoSmithKline, which used yeast as expression host to provide a gene-reporter assay (unpublished data) (Brown et al., 2011). Here, we present further corroboration of the activity of CID16020046 as a GPR55 antagonist, performed in a third laboratory (Medical University of Graz) working independently. We also provide data from GlaxoSmithKline to show that CID16020046 also antagonizes GPR55 in the yeast assay, as expected.
Independent confirmation of newly identified GPR55 ligands is of paramount importance because the literature around this receptor contains many inconsistencies and contradictions. Several compounds have been described as antagonists and used to support a hypothesis of functional expression of GPR55, even though these compounds do not consistently block the effect of acknowledged GPR55 agonists such as AM251 and LPI in many cell systems. Furthermore, many of the tools used in these studies are not selective, having known effects at other targets. For example, the reported GPR55 antagonist CBD interacts with TRP receptors (Qin et al., 2008; Pertwee et al., 2010), whereas we show here that CID16020046 is inactive at the TRPV4 channel. The complexity of GPR55 pharmacology and the lack of selective ligands have made it challenging in the extreme to determine convincingly the physiologic role of GPR55. This area has been reviewed extensively by others and will not be elaborated on here (Pertwee et al., 2010; Sharir and Abood, 2010; Balenga et al., 2011b; Henstridge, 2012).
We show that CID16020046 antagonizes agonist-mediated GPR55 activation in yeast cells (Fig. 1) and that this compound inhibits GPR55-mediated Ca2+ release from intracellular stores (Fig. 2; Table 1) as well as transcription factor activation (Figs. 4 and 5; Table 1) in HEK293 cells stably expressing GPR55. CID16020046 is selective for GPR55 over CB1, since it had no effect on WIN55,212-2-induced Ca2+ release and did not affect CB1-mediated CREB activation (Figs. 2 and 4). LPI-induced ERK1/2 phosphorylation was blocked upon pretreatment with CID16020046 (Fig. 3), while the antagonist had no effect on CB1- and CB2-mediated ERK1/2 phosphorylation (Fig. 3). In addition, internalization of GPR55 was inhibited by CID16020046 pretreatment (Fig. 6). Importantly, CID16020046 had similar antagonistic effects on both the physiologic ligand LPI and the synthetic agonists GSK319197A and GSK522373A. The calculated IC50 values for the antagonist effect of CID16020046 at GPR55 were 0.21 μM (Ca2+ mobilization; 10 μM LPI), 0.48 μM (NFAT activation; 1 μM LPI), and 0.71 μM (NF-κB activation; 1 μM LPI) (see Table 1). Although these are not expected to be numerically identical, they are consistent with both the calculated pA2 value (7.3, equating to 54 nM) from the yeast assay using GSK522373A as agonist and the IC50 of inverse agonism in yeast (0.145 μM). Blockade of SRE induction appears to be somewhat an outlier, with IC50 values for CID16020046 of 2.0 μM (1 μM LPI). It seems reasonable to expect a functional potency for CID16020046 in blocking LPI-evoked GPR55-mediated effects in the approximate range 0.1–1 μM. The effect of CID16020046 on the shape of agonist concentration-response curves (depression of the maximum asymptote) is suggestive of a noncompetitive mechanism, which may be confirmed once validated radioligands become available.
To illustrate the power of the newly available GPR55-selective antagonist CID16020046, we show a preliminary investigation into the putative signaling properties of GPR55 in human cells endogenously expressing the receptor, including primary cells: first in microvascular lung endothelial cells (HMVEC-Ls), second in human platelets. GPR55 mRNA levels are expressed in HMVEC-Ls, and LPI (10 µM) caused intracellular Ca2+ release from HMVEC-L intracellular stores, and CID16020046 could block this effect. However, higher concentrations of CID16020046 were required to block HMVEC-L Ca2+ release (25 and 50 µM; 10 µM was ineffective) than had been observed to block Ca2+ release in the recombinant assay (IC50 = 0.21), and in this concentration range, CID16020046 may have nonspecific effects, for example, at phosphodiesterases (Supplemental Table 1). Therefore, further experiments are required to confirm that GPR55 in HMVEC-Ls couples to Ca2+ release. However, we found much lower concentrations of LPI to be effective in enhancing endothelial wound healing in the HMVEC-L model. Moreover, the degree of enhancement attributable to LPI was largely abolished by 1 µM CID16020046, a concentration in keeping with its potency in recombinant assays. Taken together, these data support functional expression of GPR55 by HMVEC-L cells. We also provide preliminary evidence that ADP-induced platelet aggregation may be attenuated by 10 µM LPI and that this effect is reversed by CID16020046 (3 and 10 μM). Therefore, it appears that platelets may be another cell type in which GPR55, and its ligand LPI, play a role.
Accumulating evidence suggests that GPR55 plays a role in diverse physiologic and pathologic processes (Henstridge et al., 2011). For instance, it has been reported that GPR55 and CB2 receptors are coexpressed and cross-talk at the level of small GTPases in human neutrophils. GPR55 thereby enhances the migratory capacity of neutrophils while it limits their bactericidal functions, such as reactive oxygen species production and degranulation (Balenga et al., 2011a). It has also been reported that GPR55 is highly expressed in malignant human tumors, and its expression level is directly correlated to the aggressiveness of the tumors (Andradas et al., 2011; Henstridge et al., 2011; Perez-Gomez et al., 2012). GPR55 was shown to be expressed in human osteoclasts and that its activation could stimulate osteoclast polarization and resorption in vitro (Whyte et al., 2009). In addition, GPR55 knockout mice were resistant to inflammatory and neuropathic pain (Staton et al., 2008), and GPR55 may play a role in the inflammatory responses of microglial cells (Pietr et al., 2009). None of these studies were able to access CID16020046 or the other new GPR55 antagonists to validate and support their findings, but the prospect for the future is for a much more well-founded understanding of GPR55 function, especially whether medicines targeting GPR55 could be therapeutically beneficial. CID16020046 and at least one of the GPR55 antagonists characterized by the Abood group, ML-193, are commercially available from Molport. Since ML-193 and CID16020046 are chemically distinct, a powerful approach would be to use them in combination. If, for example, LPI-evoked effects in primary cells and tissues were blocked by both these agents, and at concentrations consistent with the potency of each antagonist observed in recombinant assays, this would constitute strong evidence that GPR55 is indeed functionally important. We have not so far accessed ML-193, but the reproducibility of both synthetic agonists [benzylpiperazines (Brown et al., 2011)] and antagonists (CID16020046) arising out of the Molecular Library Screening Center Network screen give confidence about its usefulness.
In summary, CID16020046 is a selective GPR55 antagonist that is commercially available and allows a range of applications to study GPR55 pharmacology and signaling properties. In combination with the other selective agonist and antagonist tools recently described, it offers the prospect of a hugely improved understanding of the physiologic and pathophysiologic role of this complex and controversial receptor.
The authors thank Veronika Pommer, Martina Ofner, Viktoria Konya, and Wolfgang Platzer for technical support.
Participated in research design: Kargl, Heinemann.
Conducted experiments: Kargl, Brown, Andersen, Dorn.
Performed data analysis: Kargl, Brown, Heinemann
Wrote or contributed to the writing of the manuscript: Kargl, Brown, Schicho, Waldhoer, Heinemann.
- Received February 9, 2013.
- Accepted May 1, 2013.
↵1 Current affiliation: Hagedorn Research Institute, Novo Nordisk A/S, Gentofte, Denmark.
This work was supported by funds from the Austrian Science Fund [Grants P18723 (to M.W.), P22521 (to A.H.), P22771, and P25633 (to R.S.)]; the Jubilaeumsfonds of the Austrian National Bank [Grants 12552 (to M.W.), 13487, 14263 (to A.H.), and 14429 (to R.S.)]; the Lanyar Stiftung (to M.W.); the START-Funding Program of the Medical University of Graz, Ph.D. program Molecular Medicine of the Medical University of Graz, and an EMBO short-term fellowship (all to J.K.).
- cannabinoid 1 receptor
- cannabinoid 2 receptor
- 4-[4-(3-hydroxyphenyl)-3-(4-methylphenyl)-6-oxo-1H,4H,5H,6H-pyrrolo[3,4-c]pyrazol-5-yl] benzoic acid
- cAMP response element–binding protein
- extracellular-signal regulated kinase
- G protein–coupled receptor
- G protein –coupled receptor 55
- human embryonic kidney cells
- human lung microvascular endothelial cells
- nuclear factor of activated T-cells
- nuclear factor κB
- polymerase chain reaction
- serum response element
- transient receptor potential
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics