European Journal of Pharmacology 612 (2009) 61–68

Contents lists available at ScienceDirect

European Journal of Pharmacology

journal homepage: www.elsevier.com/locate/ejphar

Cardiovascular Pharmacology Time-dependent vascular actions of in the rat aorta

Saoirse E. O'Sullivan a,⁎, Yan Sun b, Andrew J. Bennett b, Michael D. Randall b, David A. Kendall b a School of Graduate Entry Medicine and Health, Derby City General Hospital, University of Nottingham, Derby DE22 3DT, United Kingdom b School of Biomedical Sciences, Queen's Medical Centre, University of Nottingham, Nottingham NG7 2UH, United Kingdom article info abstract

Article history: We have shown that the major active agent of sativa, Δ9-, activates Received 31 October 2008 peroxisome proliferator-activated gamma [PPARγ, O'Sullivan, S.E., Tarling, E.J., Bennett, A.J., Kendall, Received in revised form 18 February 2009 D.A., Randall, M.D., 2005c. Novel time-dependent vascular actions of delta9-tetrahydrocannabinol mediated Accepted 3 March 2009 by peroxisome proliferator-activated receptor gamma. Biochem. Biophys. Res. Commun. 337, 824–831]. The Available online 11 March 2009 aim of the present study was to investigate whether another pharmacologically active phytocannabinoid, cannabidiol, similarly activates PPAR . Functional vascular studies were carried out in rat aortae in vitro by Keywords: γ myography. PPARγ activation was investigated using reporter gene assays, a PPARγ competition-binding Cannabidiol assay and an adipogenesis assay. Cannabidiol caused time-dependent (over 2 h) vasorelaxation of pre- Aorta constricted aortae, sensitive to PPARγ antagonism (GW9662, 1 µM) and super oxide dismutase inhibition. Vasorelaxation The vascular effects of cannabidiol were not affected by endothelial denudation, nitric oxide synthase Peroxisome proliferator-activated receptor inhibition, pertussis toxin, cannabinoid CB1 or cannabinoid CB2 receptor antagonism, or capsaicin pre- gamma treatment. When aortae were contracted with U46619 in a Ca2+-free buffer, vasorelaxation to cannabidiol Calcium channel was substantially reduced. Furthermore, cannabidiol (1–30 µM) inhibited the contractile response to the re- introduction of Ca2+. In a reporter gene assay, cannabidiol increased the transcriptional activity of PPARγ. Cannabidiol was also found to bind to PPARγ and stimulate the differentiation of 3T3-L1 fibroblasts into adipocytes, a PPARγ-mediated response. These results show that cannabidiol binds to and activates PPARγ, which partially underlies the time-dependent vascular effects of cannabidiol. However, cannabidiol-induced vasorelaxation in the rat isolated aorta appears to be largely due to calcium channel inhibition. © 2009 Published by Elsevier B.V.

1. Introduction and cannabinoid CB2 (McPartland et al., 2007), although it has been shown to antagonise both cannabinoid CB1 and cannabinoid CB2 Cannabidiol is a major component of the cannabis plant, Cannabis receptors in the low nanomolar range (Thomas et al., 2007). sativa, but lacks the psychotropic effects of the better known active Cannabidiol is also suggested to be an antagonist of the GPR55 9 agent, Δ -tetrahydrocannabinol (THC). Cannabidiol is anxiolytic receptor (Ryberg et al., 2007). (Moreira et al., 2006; Zuardi et al., 2006; Resstel et al., 2006) and Peroxisome proliferator-activated receptors (PPARs) belong to a may antagonise the psychotropic effects of THC in the clinically family of nuclear receptors of which there are three isotypes: α, δ and available cannabis-based medicine, Sativex (Russo and Guy, 2006). In γ (Ferré, 2004). PPARs heterodimerise with the retinoid X receptor, addition to its behavioural effects, cannabidiol has been shown to be and bind to DNA sequences called PPAR response elements (PPREs), anti-inflammatory (Costa et al., 2007; Esposito et al., 2007), anti- which lead to the transcription of target genes upon activation. oxidant (García-Arencibia et al., 2007), neuroprotective (El-Remessy PPARs are primarily involved in the regulation of metabolism and et al., 2006), a potent inhibitor of cancer cell growth (Ligresti et al., energy homeostasis. For example, agonists of the PPARγ isoform 2006) and is beneficial in diabetes (El-Remessy et al., 2006; Weiss improve insulin sensitivity and are used in the management of type 2 et al., 2006). However, the molecular targets for cannabidiol are still diabetes (Ferré, 2004; Rangwala and Lazar, 2004). In addition, PPARγ unclear. Cannabidiol is reported to have a low affinity for both of the agonists have been shown to have positive cardiovascular effects, ‘ ’ classical G-protein-coupled cannabinoid receptors, cannabinoid CB1 which include increased availability of nitric oxide, reductions in blood pressure and attenuation of atherosclerosis (Bishop-Bailey, 2000; Hsueh and Bruemmer, 2004). Some of the beneficial effects of PPARγ ligands are brought about by anti-inflammatory actions, fl ⁎ Corresponding author. Tel.: +44 1332724643. including inhibition of pro-in ammatory cytokines, increasing anti- E-mail address: [email protected] (S.E. O'Sullivan). inflammatory cytokines, and inhibition of inducible nitric oxide

0014-2999/$ – see front matter © 2009 Published by Elsevier B.V. doi:10.1016/j.ejphar.2009.03.010 62 S.E. O'Sullivan et al. / European Journal of Pharmacology 612 (2009) 61–68 synthase (iNOS) expression (see Széles et al., 2007 for a recent vasorelaxant effect of a single concentration of cannabidiol or review). vehicle control (0.1% ethanol) on induced tone was assessed as the PPARs have a large ligand binding pocket and are pharmacologi- reduction in tone over time. For each experimental protocol, vehicle cally promiscuous, being activated by a number of natural and control and cannabidiol-treated experiments were performed in synthetic ligands including, as reported recently, (see adjacent segments of the same artery. Net relaxation was calculated O'Sullivan, 2007 for a recent review). Specifically, it has been reported as the vasorelaxant effect of cannabidiol minus the vasorelaxant that , an analogue of a THC metabolite, binds to and effect of vehicle (in adjacent aortic segments) over time. In some increases the transcriptional activity of PPARγ with anti-inflammatory experiments, the vascular effects of cannabidiol were investigated actions (Liu et al., 2003). Similarly, the endocannabinoids in un-contracted vessels. and 2-arachidonoylglycerol (2-AG) have anti-inflammatory actions To assess the contribution of PPARγ activation, some experi- sensitive to PPARγ antagonism (Rockwell and Kaminski, 2004; ments were performed in the presence of the PPARγ antagonist Rockwell et al., 2006). Anandamide has also been shown to directly GW9662 (1 µM) added 10 min prior to pre-contraction. To bind to PPARγ (Bouaboula et al., 2005; Gasperi et al., 2007). We investigate the role of endothelium-derived relaxing factors in recently showed that THC increases the transcriptional activity of the time-dependent vasorelaxation to cannabidiol, some vessels PPARγ and leads to vascular responses that are inhibited by a PPARγ were denuded of their endothelium by abrasion with a human hair. antagonist (O'Sullivan et al., 2005c, 2006). As many of the effects of The role of endothelium-derived nitric oxide (NO) was investigated G cannabidiol treatment are consistent with those of PPARγ activation, using the nitric oxide synthase inhibitor, N -nitro-L-arginine and since other cannabinoid compounds have been shown to activate methyl ester (L-NAME, 300 µM, present throughout, O'Sullivan PPARγ, the possibility exists that PPARγ activation might underlie et al., 2005c). To establish a potential role for superoxide dismutase some of the pharmacological effects of cannabidiol. (SOD) activity, some experiments were performed in the presence In vitro, , a synthetic analogue of cannabidiol, of the SOD inhibitor diethyldithiocarbamate (DETCA, 3 mM), added causes vasorelaxation of rat isolated mesenteric arteries via the 30 min prior to pre-contraction of arteries (O'Sullivan et al., putative ‘endothelial’ , potassium channel acti- 2005c). vation and calcium channel blockade (Offertáler et al., 2003; Ho and To assess any possible contribution of cannabinoid receptor Hiley, 2003). In vivo, abnormal cannabidiol causes hypotension and activation, some experiments were performed in the presence of the mesenteric vasodilatation in wild-type mice and in mice lacking both cannabinoid CB1 receptor antagonist, AM251 (1 µM), or the cannabi- cannabinoid CB1 and cannabinoid CB2 receptors (Járai et al., 1999). noid CB2 receptor antagonist, AM630 (1 µM), both added 10 min before However, to date, the vascular responses to cannabidiol (as opposed to contracting vessels. To assess if cannabidiol acts at a G(i/o)-protein- abnormal cannabidiol) remain uninvestigated. We have previously coupled receptor, some vessels were incubated for 2 h with 200 ng/ml shown that THC causes time-dependent, endothelium-dependent, pertussis toxin (PTX, O'Sullivan et al., 2005b), and vasorelaxation to PPARγ-mediated vasorelaxation of the rat isolated aorta (O'Sullivan cannabidiol or vehicle assessed. The involvement of TRPV1 receptors on et al., 2005c). Therefore, the aim of the present study was to sensory nerves was assessed by incubating vessels for 1 h with the investigate whether similar time-dependent, PPARγ-mediated vasor- TRPV1 agonist capsaicin (10 µM) to deplete the sensory nerves of elaxation to cannabidiol occurs in the rat aorta. vasoactive neurotransmitters (Zygmunt et al., 1999; O'Sullivan et al., Our studies have shown for the first time that cannabidiol causes 2005a). vasorelaxation of the rat isolated aorta, partly mediated by PPARγ. To investigate the involvement of K+ channel activation in We have demonstrated that cannabidiol increases the transcrip- vasorelaxation to cannabidiol, some experiments were performed in tional activity of PPARγ,bindstoPPARγ and causes a PPARγ- vessels pre-contracted with a high-K+ (60 mM) buffer to inhibit mediated response, adipogenesis (Mueller et al., 2002). These data potassium efflux. To test whether cannabidiol causes vasorelaxation in confirm that cannabidiol is a PPARγ ligand, and suggest that PPARγ arteries in Ca2+-free conditions, some vessels were contracted with activation may underlie some of the pharmacological effects of U46619 in Ca2+-free Krebs and the vasorelaxant effects of cannabidiol cannabidiol. investigated. To investigate the effects of THC on Ca2+ influx,

concentration–response curves to calcium chloride (CaCl2, 1 µM to 2. Methods and materials 100 mM) were performed in the presence and absence of cannabidiol (1, 10 and 30 µM, 10 min incubation). The vessels were first allowed to 2.1. In vitro vascular studies equilibrate in Ca2+-free buffer, and were then bathed in Ca2+-free, high-K+ buffer. 10 min after adding cannabidiol, a concentration–

Male Wistar rats (250–350 g) were stunned by a blow to the response curve to re-introduced CaCl2 was constructed. back of the head and killed by cervical dislocation under Schedule 1 to the Animals Act 1986 (Scientific Procedures). The aortae were 2.2. Cell culture removedrapidlyandplacedintocoldmodified Krebs–Henseleit buffer (composition, mM: NaCl 118, KCl 4.7, MgSO4 1.2, KH2PO4 1.2, Human embryonic kidney (HEK293) cells and 3T3L1 fibroblasts NaHCO3 25, CaCl2 2, and D-glucose 10). The aortae were dissected were grown in Dulbecco's modified Eagle's medium with 10% fetal free of adherent connective and adipose tissue and cut into rings 3– calf serum with antibiotics (penicillin, streptomycin and gentami-

4 mm long, and mounted on fixed segment support pins using the cin) in 5% CO2. Multi Myograph system (Model 610 M, Danish Myo Technology, Denmark). Once mounted, all vessels were kept at 37 °C in modified 2.3. Transactivation assay

Krebs–Henseleit buffer and gassed with 5% CO2 in O2. The aortae were stretched to an optimal passive tension of 9.8 mN tension Confluent HEK293 cells (90%) were transiently transfected with (O'Sullivan et al., 2005c). Vessels were allowed to equilibrate and 2 µg of a 3xPPRE-TK-Luc luciferase reporter construct (a kind gift the contractile integrity of each was tested by its ability to contract from Prof. Walter Wahli at the Universite de Lausanne), 500 ng to 60 mM KCl by at least 4.9 mN. Vessels were contracted with a PPARγ (also from Prof. Walter Wahli) and 500 ng retinoid X receptor combination of U46619 (10–100 nM, a thromboxane prostanoid (a kind donation from Prof. David Mangelsdorf at the Southwestern receptor agonist), and the α-adrenoceptor agonist methoxamine Medical Center, Dallas, Texas), using polyethyleneimine (Poly- (1–5 µM) to increase tension as previously described (O'Sullivan et sciences Inc.). 6 h after transfection, cells were treated for 24 h al., 2005c). When a stable contraction was maintained, the with the PPARγ agonist rosiglitazone or cannabidiol. Cells were S.E. O'Sullivan et al. / European Journal of Pharmacology 612 (2009) 61–68 63

2.6. Statistical analysis

In each protocol, the number of animals in each group is repre- sented by n, and values are expressed as mean± S.E.M. The dif- ference between cannabidiol-treated and vehicle-treated vessels under each experimental protocol was analysed by paired Student's t-test (Fig. 1A). Other data, including net vasorelaxant values, were compared, as appropriate, by analysis of variance (ANOVA) with statistical significance between manipulations and controls deter- mined by Dunnett's post-hoc test (Figs.2,3,4A,B,and5A, Tables 1 and 2). The concentration of vasorelaxant giving half-maximal

responses (EC50)andmaximalresponses(Rmax)wereobtained from concentration–response curves fitted to a sigmoidal logistic equation using the GraphPad Prism package [Y =Bottom+(Top− Bottom)/(1+10((LogEC50 − X) ⁎ Hillslope)), where X is a logarithm of agonist concentration and Y istheresponsethatstartsfromthe Bottom and goes to the Top in a sigmoid shape. Maximal responses

(Rmax)andpEC50 (negative logarithm of the EC50)valuesare expressed as mean±S.E.M. (Fig. 4C).

Fig. 1. A, The mean vasorelaxant response of cannabidiol vs vehicle (EtOH) over 2 h in a pre-constricted rat aorta. B, The net vasorelaxant effect of cannabidiol (that is, the vasorelaxant effect of cannabidiol minus the vasorelaxant effect of vehicle in the same artery). Data are given as means with error bars representing S.E.M. (n=8–15).

harvested and assayed for luciferase activity (as a measure of reporter gene transcription) using a Luciferase Reporter Assay (Promega; according to the manufacturer's instructions). Protein content of the cells was determined using the BioRad protein assay according to the manufacturer's instructions.

2.4. PPARγ binding

Binding to the PPARγ-ligand binding domain was measured using the PolarScreen™ PPARγ competitor assay kit (Invitrogen). Briefly, the competitors were diluted in reaction buffer in a black 384 shallow well plate (NUNC). The purified PPARγ-ligand binding domain and fluorescent PPARγ ligand (Fluormone™ PPAR Green) were then added. Samples were incubated at room temperature for 2 h and fluorescence polarisation obtained by reading the plate using a PerkinElmer Envision 2102 Multi-label Reader with an excitation wavelength of 485 nm and emission wavelength of 535 nm.

2.5. Adipocyte differentiation assay

3T3L1 fibroblasts were grown until confluent and then treated with 1 µM dexamethasone and 5 µg/µl insulin, to induce differentia- tion, for 2 days at 10% CO2. Subsequently, cells were kept in maintenance medium consisting of Dulbecco's modified Eagle's medium with 10% fetal calf serum supplemented with insulin (5 µg/

µl) at 10% CO2. During this time, cells were treated for 7 days with either increasing concentrations of cannabidiol or rosiglitazone (10 µM). Cells were stained with Oil Red O, to determine lipid Fig. 2. The effects of PTX treatment (A, 200 ng/ml, 2 h), the cannabinoid CB1 receptor antagonist AM251 (1 µM) and the cannabinoid CB2 receptor antagonist AM630 (B, 1 µM), accumulation (Mueller et al., 2002; O'Sullivan et al., 2005c), and and capsaicin pre-treatment (C, 10 µM, 1 h) on the net vasorelaxation to cannabidiol. Data counterstained with Cresyl Blue. are given as means with error bars representing S.E.M. 64 S.E. O'Sullivan et al. / European Journal of Pharmacology 612 (2009) 61–68

only significantly different from vehicle at 105 and 120 min (2 h, cannabidiol 31.3± 6.9% relaxation, n =9, P b0.05, Fig. 1A). At 5 µM, the vascular effect of cannabidiol was significantly different from vehicle at all time-points (2 h, cannabidiol 51.9 ±5.4% relaxation, n =8, P b0.001, Fig. 1A). Similarly, 10 µM cannabidiol caused significant time-dependent relaxation of the rat aorta compared to vehicle control at all time-points (2 h, vehicle 18.0±2.5% vs cannabidiol 70.2 ±3.4% relaxation, n=15, P b0.0001, Fig. 1A). After 2 h, the net relaxation (the vasorelaxant effect of cannabidiol minus the vasorelaxant effects of vehicle/time in adjacent segments of aorta) was; 100 nM, 3.6±2.6; 1 µM, 18.1±7.3; 5 µM, 37.1±5.4;

10 µM, 52.2 ±3.1% relaxation (Fig. 1B). The pEC50 for the time- dependent vasorelaxant effects of cannabidiol is −5.3±0.4 (around 5 µM). Cannabidiol had no effect on vascular tone in aortae not contracted pharmacologically (2 h, vehicle −0.02±0.01 g cf. cannabidiol −0.03±0.01 g, n =7). Pre-treating arteries with PTX (200 ng/ml, 2 h) had no effect on the vascular response to cannabidiol (Table 1, Fig. 2A). Neither the

cannabinoid CB1 receptor antagonist AM251 (1 µM) nor the cannabinoid CB2 receptor antagonist AM630 (1 µM) had any

Fig. 3. The effects of the PPAR antagonist (GW9662,1 µM, A), removing the endothelium and inhibiting nitric oxide synthase (L-NAME, 300 µM, B), and inhibition of SOD activity (DETCA, 3 mM, C) on the net vasorelaxation to cannabidiol. Data are given as means with error bars representing S.E.M. *Pb0.05, **Pb0.01, compared to control (ANOVA).

2.7. Drugs

All drugs were supplied by Sigma Chemical Co. (Poole, UK) except where stated. GW9662 (2-Chloro-5-nitro-N-phenylbenzamide), AM251 (N-(Piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophen yl)-4- methyl-1H-pyrazole-3-carboxamide) and AM630 (6-Iodo-2-methyl-1- [2-(4-morpholinyl)ethyl]-1H-indol-3-yl](4-methoxyphenyl)metha- none) were obtained from Tocris (UK). Cannabidiol was generously supplied by GW Pharmaceuticals. Acetylcholine, methoxamine, U46619 (9,11-dideoxy-9alpha,11alpha-methanoepoxy-prosta-5E,13E-dien-1- oic acid) and PTX were dissolved in distilled water. L-NAME was dissolved in the Krebs–Henseleit solution. Cannabidiol, capsaicin and verapamil were dissolved in ethanol to a stock concentration of 10 mM with further dilutions made in distilled water. GW9662, AM251 and AM630 were dissolved in DMSO to 10 mM, with further dilutions made in distilled water.

3. Results

+ 3.1. Time-dependent effects of cannabidiol in the aorta Fig. 4. The vasorelaxant effects of cannabidiol in aortae contracted with (A) a high-K Krebs solution and (B) U46619 in a Ca2+-free Krebs–Henseleit buffer. C, The effects of increasing concentrations of cannabidiol (1–30 µM, 10 min) on the contractile response fi 2+ At 100 nM, the vascular effect of cannabidiol was not signi cantly to calcium in a Ca -free, high potassium buffer. Data are given as means with error different to vehicle (Fig. 1A). At 1 µM, the effect of cannabidiol was bars representing S.E.M. ⁎Pb0.05, ⁎⁎Pb0.01, compared to control (ANOVA). S.E. O'Sullivan et al. / European Journal of Pharmacology 612 (2009) 61–68 65

or absence of the PPARγ antagonist GW9662 (Fig. 4B). The potency and maximal contractile response to the re-introduction of calcium in calcium-free, high potassium Krebs–Henseleit solution was signifi- cantly reduced by cannabidiol in a concentration-dependent manner (1 µM to 30 µM; see Table 2, Fig. 4C). The maximal vasorelaxant response to 10 µM cannabidiol was similar to that of the calcium channel blocker, verapamil (cannabidiol 70±3% relaxation, n=15vs 10 µM verapamil 64±9% relaxation).

3.2. Transactivation assays

To determine whether cannabidiol stimulates PPARγ, transactiva- tion assays were performed in HEK293 cells transiently over- expressing PPARγ and retinoid X receptor α in combination with a luciferase reporter gene (3xPPRE TK luc). In these assays, the synthetic PPARγ agonist rosiglitazone (10 µM) significantly stimulated the transcriptional activity of PPARγ compared to vehicle-treated cells (148±7 vs 319±7 relative luciferase activity (per ng/ml protein), P b0.01, Fig. 5A). Cannabidiol also significantly stimulated the transcriptional activity of PPARγ compared to untreated cells at 10 µM (305±18 relative luciferase activity, Pb0.01) and 20 µM (470± 37 relative luciferase activity, Pb0.01) in a concentration-dependent manner.

3.3. PPARγ binding

To determine whether cannabidiol directly binds to the PPARγ ligand binding domain, a fluorescence polarisation assay was performed. As shown in Fig. 5B, binding of the PPARγ fluorescent ligand was inhibited in a concentration-dependent manner by

rosiglitazone (IC50 ~35 nM), ajulemic acid (IC50 ~600 nM) and cannabidiol (IC50 ~5 µM). Fig. 5. Transactivation assays were performed in homologous cells transiently over- expressing PPARγ and retinoid X receptor in combination with a luciferase reporter 3.4. Induction of adipocyte differentiation gene (3xPPRE TK luc). A, Compared with untreated cells, a significant increase in luciferase activity was seen after treatment with rosiglitazone (10 µM), and with 10 µM and 20 µM cannabidiol. B, Rosiglitazone, ajulemic acid and cannabidiol displaced the 3T3L1 cells were cultured until confluent and then treated for fluorescent Fluormone™ PPAR Green ligand from the PPAR-LBD/Fluormone™ PPAR 8 days with either cannabidiol or rosiglitazone. Cells were fixed and Green complex, resulting in a reduction of polarisation value. Data are given as means stained with Oil Red O to identify fat droplets, the presence of which ⁎⁎ fi b with error bars representing S.E.M. denotes a signi cant difference (P 0.01) indicates differentiation of fibroblasts into adipocytes. As shown in compared to vehicle control (ANOVA). Fig. 6A, untreated cells showed some signs of differentiation, but the majority of cells retained their spindle shape with little Oil Red O staining. Rosiglitazone (Fig. 6B) induced differentiation of 3T3 L1 cells to adipocytes, as evidenced by large amounts of Oil Red O staining significant effect on the time-dependent vascular responses to indicating fat droplet accumulation within the cytoplasm. In the cannabidiol (Table 1, Fig. 2B). Similarly, pre-treating arteries with the TRPV1 agonist capsaicin (10 µM, 1 h) had no effect on the time- dependent vascular response to cannabinoid (Table 1, Fig. ). In the presence of the PPARγ receptor antagonist GW9662 (1 µM), Table 1 Mechanisms underlying the vasorelaxant effect of cannabidiol. the net vasorelaxant effect of cannabidiol (the vascular effect of cannabidiol in the presence of GW9662 minus the vasorelaxant effect 1 h % relaxation 2 h % relaxation n of vehicle in the presence of GW9662) was significantly reduced from Cannabidiol 43.4±3.4 52.2±3.1 15 a b 45 min onwards (Fig. 3A, Table 1). The vasorelaxant effects of & GW9662 27.3±5.2 33.0±5.7 12 Endothelium denuded 40.3±8.7 53.8±10.7 9 cannabidiol were not different in endothelium-denuded and control & L-NAME (300 µM) 38.3±3.4 46.6±5.7 9 aortae (Table 1, Fig. 3B). Similarly, in the presence of the nitric oxide & DETCA (3 mM) 25.9±2.8b 40.0±7.4 6 synthase inhibitor, L-NAME (300 µM), the net vasorelaxant effect of & AM251 (1 µM) 47.5±6.2 55.4±5.5 12 cannabidiol was not different to that observed in control conditions & AM630 (1 µM) 39.0±4.6 47.7±4.4 8 (Table 1, Fig. 3B). In the presence of the SOD inhibitor, DETCA, the & PTX (200 ng/ml, 2 h) 47.0±7.7 54.7±7.2 10 fi & Capsaicin (10 µM, 1 h) 44.2±4.5 42.6±4.7 13 vasorelaxant effect of cannabidiol was signi cantly reduced from High K+ contracted 33.6±3.3 51.5±4.9 11 45 min to 105 min (Table 1, Fig. 3C). U46619 contracted (Ca2+ free) 14.9±2.1b 9.3±2.5b 9 When arteries were contracted with a high potassium buffer, there U46619 contracted (Ca2+ free) 15.8±3.5b 20.7±9.0b 6 was a small decrease in the net vasorelaxant effect of cannabidiol & GW9662 (1 µM) compared with control at 45 min (Pb0.05, Fig. 4A), but no difference Residual relaxation (the vasorelaxant effect of cannabidiol minus the vasorelaxant in the net relaxant effect of cannabidiol after 2 h (Table 1). In vessels in effect of vehicle in an adjacent segment of artery under the same experimental conditions) to cannabidiol after 1 or 2 h under various experimental conditions. which tone was induced with U46619 in calcium-free buffer, the a b fi fi (P 0.05, ANOVA) denotes a signi cant difference between vasorelaxation to vasorelaxant effect of cannabidiol was signi cantly blunted at all cannabidiol under control conditions, and those under experimental conditions time-points compared with control (Table 1, Fig. 4B) in the presence (bPb0.01). 66 S.E. O'Sullivan et al. / European Journal of Pharmacology 612 (2009) 61–68

Table 2 We demonstrated that that the time-dependent effects of THC The effects of cannabidiol on the contractile response to calcium. were abolished in the presence of a SOD inhibitor, DETCA (O'Sullivan

pEC50 Rmax (g increase in tension) n et al., 2005c). It was similarly observed in the present study that Vehicle control 3.49±0.08 2.49±0.06 11 DETCA inhibited the time-dependent vasorelaxant effects of canna- 1 µM cannabidiol 3.22±0.14 2.10±0.13a 8 bidiol, suggesting that increased SOD activity promotes vasorelaxation 10 µM cannabidiol 2.87±0.11b 1.86±0.11b 11 through reductions in reactive oxygen species. This is in agreement b b 30 µM cannabidiol 2.24±0.10 1.38±0.11 6 with other works showing PPARγ ligands cause the induction of Cu/ a (Pb0.05, ANOVA) denotes a significant difference between vehicle and cannabidiol- Zn-SOD (Hwang et al., 2005). b treated aortae ( Pb0.01). Some of the vasorelaxant effects of cannabinoids are due to activation of other target sites (see Randall et al., 2004 for a review), presence of cannabidiol, concentration-dependent fat droplet accu- and we explored whether the vasorelaxant response to cannabidiol mulation was apparent at all concentrations tested (Fig. 6C–F). might be partially mediated by any of these. We found that neither the

cannabinoid CB1 nor cannabinoid CB2 receptor antagonists had any 4. Discussion effect on vasorelaxation to cannabidiol. These data are supported by our finding that the vasorelaxant effect of cannabidiol was not affected

The aim of the present study was to establish whether the by the G(i/o)-protein inactivator PTX, and suggest that the time- phytocannabinoid, cannabidiol, causes time-dependent vasorelaxa- dependent vasorelaxant effects of cannabidiol are not mediated tion of the rat aorta similar to THC (O'Sullivan et al., 2005c), and through as yet unidentified G(i/o)-protein-coupled cannabinoid whether this is due to PPARγ activation. We have shown, for the first receptors. It is also unlikely that cannabidiol acts through the time, that cannabidiol causes a time-dependent vasorelaxant effect proposed G-protein-coupled endothelial cannabinoid receptor as the that is partially inhibited by a PPARγ antagonist. We have also shown vasorelaxant response to cannabidiol was not sensitive to either that cannabidiol causes significant activation of the transcriptional removal of the endothelium or PTX treatment, as previously shown for activity of PPARγ, binds directly to PPARγ and causes the induction of ligands of this receptor (Mo et al., 2004; Begg et al., 2003). adipogenesis, a PPARγ ligand property. Together, these data strongly Additionally, we previously have demonstrated that sensitivity to suggest that cannabidiol is a functional PPARγ agonist. the endothelial cannabinoid receptor antagonist, O-1918, only occurs We have previously shown that THC produces time-dependent in small resistance arteries of the mesenteric bed and not in conduit vasorelaxation of isolated aortae mediated by PPARγ (O'Sullivan et al., arteries like the aorta (O'Sullivan et al., 2004). It is also not likely that 2005c). We have now investigated whether similar responses are the vasorelaxant response to cannabidiol is mediated by the proposed produced by cannabidiol, and have shown that cannabidiol (max- cannabinoid receptor, GPR55, as cannabidiol is not suggested to be an imum net relaxation 52%), causes a time-dependent progressive agonist at this receptor (Ryberg et al., 2007). The time-dependent vasorelaxant effect similar to that of THC (maximum net relaxation vasorelaxant response to cannabidiol was not inhibited by pre- 47%) and rosiglitazone, (maximum net relaxation 70%) (Cunnane incubation with capsaicin, to deplete sensory neurotransmitters, et al., 2004; O'Sullivan et al., 2005c). To test whether this response is making it unlikely that cannabidiol acts via TRPV1 in the aorta. due to PPARγ activation, some experiments were performed in the Some of the vasorelaxant effects of endocannabinoids and THC are presence of the PPARγ antagonist, GW9662, which caused a brought about by potassium channel activation and calcium channel significant reduction in the vasorelaxant response to cannabidiol, as inhibition (Randall et al., 1997; Ho and Hiley, 2003; O'Sullivan et al., previously observed for THC and rosiglitazone (O'Sullivan et al., 2005b). The potential involvement of K+ channel activation by 2005c). cannabidiol was therefore investigated by assessing the vasorelaxant

Fig. 6. Induction of adipocyte differentiation by rosiglitazone and cannabidiol. 3T3 L1 cells were cultured until confluent and then treated for 8 days with either cannabidiol or rosiglitazone (10 µM). Cells were fixed and stained with Oil Red O to identify fat droplets, which stain red, to indicate differentiation of fibroblasts to adipocytes, and counterstained with Cresyl Blue. Vehicle-treated cells (A) showed some signs of differentiation, but the majority of cells retained their spindle shape and blue cytoplasmic staining was still visible. Rosiglitazone (B) induced differentiation to adipocytes, with large numbers of red fat droplets. Cannabidiol induced adipocyte differentiation at all concentrations tested (100 nM to 10 µM), as evidenced by positive Oil Red O staining (C–F). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) S.E. O'Sullivan et al. / European Journal of Pharmacology 612 (2009) 61–68 67 response to cannabidiol in vessels pre-contracted with a high It is generally recognised that there are several side effects potassium buffer to inhibit potassium flux. Under these conditions, a associated with PPARγ ligands, including weight gain, oedema and small reduction in the vasorelaxant response to cannabidiol was increased plasma lipoproteins (Gelman et al., 2007). New PPARγ observed at 45 min, suggesting that activation of potassium channels agonists that do not possess these side effects are currently being leading to hyperpolarisation and vascular relaxation is not the investigated, and it is suggested that partial or low affinity agonists primary mechanism by which cannabidiol acts. In further experi- may be beneficial (Gelman et al., 2007). Cannabidiol may, therefore, ments, some aortae were contracted with U46619 in a calcium-free prove to have therapeutic utility as a low affinity agonist of PPARγ. buffer, a condition in which contraction is independent of calcium In summary, these data provide strong evidence, for the first time, influx (Williams and Roberts, 2003). In these experiments, the that cannabidiol is a PPARγ agonist, and suggest a novel means by vasorelaxant effect of cannabidiol was considerably reduced, suggest- which the diverse effects of cannabidiol could be brought about. In ing the mechanism by which cannabidiol acts involves calcium. To light of the evidence that PPARγ ligands have beneficial effects in type investigate this further, we performed concentration–response curves 2 diabetes, the cardiovascular system and, potentially, in a wide 2+ to CaCl2 in Ca -free buffers in the absence and presence of variety of other disorders including cancer, gastro-inflammatory cannabidiol, and it was found that cannabidiol significantly inhibited disorders and many skin diseases, our data provide a further rationale the contractile response to calcium in a concentration-dependent for the investigation of natural cannabinoids as therapeutic agents. manner. Taken together, these data suggest that the vasorelaxant effect of cannabidiol is mediated by calcium channel inhibition. While Acknowledgements the possibility exists that the inhibition of Ca2+ channels may be as a consequence of PPARγ activation, this is unlikely, as cannabidiol We would like to thank Dr Richard Roberts for the use of a inhibited the contractile response to calcium after only 10 min myograph. SO'S was previously supported by a Leverhulme Trust Early incubation. Interestingly, other PPARγ agonists have been reported to Career Fellowship. cause non-PPARγ-mediated acute relaxation of resistance arteries via blockade of Ca2+ channels (Heppner et al., 2005). To summarise the previously unreported vascular effects of References cannabidiol in the rat aorta, we have shown that cannabidiol causes time-dependent vasorelaxation that is neither endothelium- nor NO- Ambrosio, A.L., Dias, S.M., Polikarpov, I., Zurier, R.B., Burstein, S.H., Garratt, R.C., 2007. Ajulemic acid, a synthetic non-psychoactive cannabinoid acid, bound to the ligand dependent, does not involve cannabinoid CB1, cannabinoid CB2 binding domain of the human peroxisome proliferator activated receptor gamma. receptors or TRPV1 channel activation, but appears to be partly due J. Biol. Chem. 282, 18625–18633. to inhibition of calcium channels. The time-dependent vasorelaxant Begg, M., Mo, F.M., Offertaler, L., Bátkai, S., Pacher, P., Razdan, R.K., Lovinger, D.M., Kunos, G., 2003. G protein-coupled endothelial receptor for atypical cannabinoid ligands effects of cannabidiol were also antagonised by the PPARγ antagonist, modulates a Ca2+-dependent K+ current. J. Biol. Chem. 278, 46188–46194. GW9662, and by the SOD inhibitor, DETCA. Bishop-Bailey, D., 2000. Peroxisome proliferator-activated receptors in the cardiovas- Increasing evidence has indicated that cannabinoids are capable of cular system. Br. J. Pharmacol. 129, 823–834. Bouaboula, M., Hilairet, S., Marchand, J., Fajas, L., Le Fur, G., Casellas, P., 2005. binding to, activating and causing PPAR-mediated responses (see Anandamide induced PPARgamma transcriptional activation and 3T3-L1 preadi- O'Sullivan, 2007). We have shown that the major active ingredient of pocyte differentiation. Eur. J. Pharmacol. 517, 174–181. cannabis, THC, activates PPARγ (O'Sullivan et al., 2005c, 2006). To test Costa, B., Trovato, A.E., Comelli, F., Giagnoni, G., Colleoni, M., 2007. The non-psychoactive cannabis constituent cannabidiol is an orally effective therapeutic agent in rat whether cannabidiol similarly activates PPARγ, a number of experi- chronic inflammatory and neuropathic pain. Eur. J. Pharmacol. 556, 75–83. ments were performed. Transactivation assays employed homologous Cunnane, S.E., Chan, Y.Y., Randall, M.D., 2004. Rosiglitazone-induced vasorelaxation in cells transiently over-expressing PPARγ and retinoid X receptor α in the rat aorta. Proc Br Pharm Soc at http://www.pa2online.org/Vol2Issue2abst096P. El-Remessy, A.B., Al-Shabrawey, M., Khalifa, Y., Tsai, N.T., Caldwell, R.B., Liou, G.I., 2006. combination with a luciferase reporter gene. In these assays, the Neuroprotective and blood-retinal barrier-preserving effects of cannabidiol in synthetic PPARγ agonist rosiglitazone significantly stimulated the experimental diabetes. Am. J. Pathol. 168, 235–244. transcriptional activity of PPARγ. We also found that cannabidiol Esposito, G., Scuderi, C., Savani, C., Steardo Jr, L., De Filippis, D., Cottone, P., Iuvone, T., significantly increased the transcriptional activity of PPARγ at 10 µM Cuomo, V., Steardo, L., 2007. Cannabidiol in vivo blunts beta-amyloid induced neuroinflammation by suppressing IL-1beta and iNOS expression. Br. J. Pharmacol. and 20 µM. In a competition-binding assay, cannabidiol displaced a 151, 1272–1279. fluorescent PPARγ ligand with an IC50 of approximately 5 µM, Ferré, P., 2004. The biology of peroxisome proliferator-activated receptors: relationship – indicating a much lower affinity than was found for either rosiglita- with lipid metabolism and insulin sensitivity. Diabetes 53 (Suppl 1), S43 50. fi García-Arencibia, M., González, S., de Lago, E., Ramos, J.A., Mechoulam, R., Fernández- zone or ajulemic acid. As further con rmation that cannabidiol is a Ruiz, J., 2007. Evaluation of the neuroprotective effect of cannabinoids in a rat PPARγ ligand, we examined the ability of cannabidiol to stimulate fat model of Parkinson's disease: importance of antioxidant and cannabinoid receptor- cell differentiation in 3T3 L1 fibroblasts, a well-known property of independent properties. Brain Res. 1134, 162–170. Gasperi, V., Fezza, F., Pasquariello, N., Bari, M., Oddi, S., Agro, A.F., Maccarrone, M., 2007. PPARγ ligands (Mueller et al., 2002). At all concentrations tested, from Endocannabinoids in adipocytes during differentiation and their role in glucose 100 nM to 10 µM, cannabidiol was found to stimulate the differentia- uptake. Cell Mol. Life Sci. 64, 219–229. tion of fibroblasts to adipocytes, shown by the presence of fat droplets Gelman, L., Feige, J.N., Desvergne, B., 2007. Molecular basis of selective PPARgamma modulation for the treatment of type 2 diabetes. Biochim. Biophys. Acta 1771,1094–1107. stained red with Oil Red O. Taken together, these data show for the Heppner, T.J., Bonev, A.D., Eckman, D.M., Gomez, M.F., Petkov, G.V., Nelson, M.T., 2005. first time that cannabidiol is a PPARγ ligand capable of binding to and Novel PPARgamma agonists GI 262570, GW 7845, GW 1929, and pioglitazone increasing the transcriptional activity of PPARγ. We have also decrease calcium channel function and myogenic tone in rat mesenteric arteries. Pharmacology 73, 15–22. demonstrated PPARγ-mediated effects of cannabidiol both in the Ho, W.S., Hiley, C.R., 2003. Vasodilator actions of abnormal-cannabidiol in rat isolated vasculature and in adipocytes. small mesenteric artery. Br. J. Pharmacol. 138, 1320–1332. Recent crystallography studies using ajulemic acid suggest that this Hsueh, W.A., Bruemmer, D., 2004. Peroxisome proliferator-activated receptor gamma: – compound occupies about 30% of the ligand binding cavity of PPARγ implications for cardiovascular disease. Hypertension 43, 297 305. Hwang, J., Kleinhenz, D.J., Lassegue, B., Griendling, K.K., Dikalov, S., Hart, C.M., 2005. and forms polar contacts mainly with the ω-loop, and not the C- Peroxisome proliferator-activated receptor-gamma ligands regulate endothelial terminal helix H12, as has been observed for other ligands (Ambrosio membrane superoxide production. Am. J. Physiol. Cell Physiol. 288, C899–905. et al., 2007). This may explain why cannabinoids such as anandamide, Járai, Z., Wagner, J.A., Varga, K., Lake, K.D., Compton, D.R., Martin, B.R., Zimmer, A.M., Bonner, T.I., Buckley, N.E., Mezey, E., Razdan, R.K., Zimmer, A., Kunos, G., 1999. ajulemic acid and cannabidiol are less potent compared to synthetic Cannabinoid-induced mesenteric through an endothelial site distinct ligands. However, it should be noted that, despite the low affinity of from CB1 or CB2 receptors. Proc. Natl. Acad. Sci. 96, 14136–14141. cannabidiol in the PPARγ binding assay, similar increases in transcrip- Ligresti, A., Moriello, A.S., Starowicz, K., Matias, I., Pisanti, S., De Petrocellis, L., Laezza, C., Portella, G., Bifulco, M., Di Marzo, V., 2006. Antitumor activity of plant cannabinoids tional activity and adipogenesis were observed when compared to with emphasis on the effect of cannabidiol on human breast carcinoma. rosiglitazone, suggesting that cannabidiol has good efficacy at PPARγ. J. Pharmacol. Exp. Ther. 318, 1375–1387. 68 S.E. O'Sullivan et al. / European Journal of Pharmacology 612 (2009) 61–68

Liu, J., Li, H., Burstein, S.H., Zurier, R.B., Chen, J.D., 2003. Activation and binding of Randall, M.D., Kendall, D.A., O'Sullivan, S.E., 2004. The complexities of the cardiovas- peroxisome proliferator-activated receptor gamma by synthetic cannabinoid cular actions of cannabinoids. Br. J. Pharmacol. 142, 20–26. ajulemic acid. Mol. Pharmacol. 63, 983–992. Rangwala, S.M., Lazar, M.A., 2004. Peroxisome proliferator-activated receptor gamma in McPartland, J.M., Glass, M., Pertwee, R.G., 2007. Meta-analysis of cannabinoid ligand diabetes and metabolism. Trends Pharmacol. Sci. 25, 331–336. binding affinity and receptor distribution: interspecies differences. Br. J. Pharmacol. Resstel, L.B., Joca, S.R., Moreira, F.A., Corrêa, F.M., Guimarães, F.S., 2006. Effects of 152, 583–593. cannabidiol and diazepam on behavioral and cardiovascular responses induced by Mo, F.M., Offertáler, L., Kunos, G., 2004. Atypical cannabinoid stimulates endothelial cell contextual conditioned fear in rats. Behav. Brain Res. 172, 294–298. migration via a Gi/Go-coupled receptor distinct from CB1, CB2 or EDG-1. Eur. J. Rockwell, C.E., Kaminski, N.E., 2004. A cyclooxygenase metabolite of anandamide causes Pharmacol. 489, 21–27. inhibition of interleukin-2 secretion in murine splenocytes. J. Pharmacol. Exp. Ther. Moreira, F.A., Aguiar, D.C., Guimarães, F.S., 2006. Anxiolytic-like effect of cannabidiol in 311, 683–690. the rat Vogel conflict test. Prog. Neuropsychopharmacol. Biol. Psychiatry 30, Rockwell, C.E., Snider, N.T., Thompson, J.T., Vanden Heuvel, J.P., Kaminski, N.E., 2006. 1466–1471. Interleukin-2 suppression by 2-arachidonyl glycerol is mediated through peroxi- Mueller, E., Drori, S., Aiyer, A., Yie, J., Sarraf, P., Chen, H., Hauser, S., Rosen, E.D., Ge, K., some proliferator-activated receptor gamma independently of cannabinoid Roeder, R.G., Spiegelman, B.M., 2002. Genetic analysis of adipogenesis through receptors 1 and 2. Mol. Pharmacol. 70, 101–111. peroxisome proliferator-activated receptor gamma isoforms. J. Biol. Chem. 277, Russo, E., Guy, G.W., 2006. A tale of two cannabinoids: the therapeutic rationale for 41925–41930. combining tetrahydrocannabinol and cannabidiol. Med. Hypotheses 66, 234–246. Offertáler, L., Mo, F.M., Bátkai, S., Liu, J., Begg, M., Razdan, R.K., Martin, B.R., Bukoski, R.D., Ryberg, E., Larsson, N., Sjögren, S., Hjorth, S., Hermansson, N.O., Leonova, J., Elebring, T., Kunos, G., 2003. Selective ligands and cellular effectors of a G protein-coupled Nilsson, K., Drmota, T., Greasley, P.J., 2007. The orphan receptor GPR55 is a novel endothelial cannabinoid receptor. Mol. Pharmacol. 63, 699–705. cannabinoid receptor. Br. J. Pharmacol. 152, 1092–1101. O'Sullivan, S.E., 2007. Cannabinoids go nuclear: evidence for activation of peroxisome Széles, L., Töröcsik, D., Nagy, L., 2007. PPARgamma in immunity and inflammation: cell proliferator-activated receptors. Br. J. Pharmacol. 152, 576–582. types and diseases. Biochim. Biophys. Acta 1771, 1014–1030. O'Sullivan, S.E., Kendall, D.A., Randall, M.D., 2004. Heterogeneity in the mechanisms of Thomas, A., Baillie, G.L., Phillips, A.M., Razdan, R.K., Ross, R.A., Pertwee, R.G., 2007. vasorelaxation to anandamide in resistance and conduit rat mesenteric arteries. Cannabidiol displays unexpectedly high potency as an antagonist of CB1 and CB2 Br. J. Pharmacol. 142, 435–442. receptor agonists in vitro. Br. J. Pharmacol. 150, 613–623. O'Sullivan, S.E., Kendall, D.A., Randall, M.D., 2005a. Vascular effects of delta 9- Weiss, L., Zeira, M., Reich, S., Har-Noy, M., Mechoulam, R., Slavin, S., Gallily, R., 2006. tetrahydrocannabinol (THC), anandamide and N-arachidonoyldopamine (NADA) Cannabidiol lowers incidence of diabetes in non-obese diabetic mice. Autoimmu- in the rat isolated aorta. Eur. J. Pharmacol. 507, 211–221. nity 39, 143–151. O'Sullivan, S.E., Kendall, D.A., Randall, M.D., 2005b. The effects of delta 9-tetrahydrocannabinol Williams, R., Roberts, R.E., 2003. Role of calcium, ERK-MAP kinase, and RHO-kinase in in rat mesenteric vasculature, and its interactions with the endocannabinoid thromboxane receptor-mediated vasoconstriction in the porcine isolated ear artery. anandamide. Br. J. Pharmacol. 145, 514–526. Br. J. Pharmacol. 140, 38P. O'Sullivan, S.E., Tarling, E.J., Bennett, A.J., Kendall, D.A., Randall, M.D., 2005c. Novel time- Zuardi, A.W., Crippa, J.A., Hallak, J.E., Moreira, F.A., Guimarães, F.S., 2006. Cannabidiol, a dependent vascular actions of delta9-tetrahydrocannabinol mediated by peroxi- constituent, as an antipsychotic drug. Braz. J. Med. Biol. Res. 39, some proliferator-activated receptor gamma. Biochem. Biophys. Res. Commun. 337, 421–429. 824–831. Zygmunt, P.M., Petersson, J., Andersson, D.A., Chuang, H., Sørgård, M., Di Marzo, V., O'Sullivan, S.E., Kendall, D.A., Randall, M.D., 2006. Further characterization of the time- Julius, D., Högestätt, E.D., 1999. Vanilloid receptors on sensory nerves mediate the dependent vascular effects of delta9-tetrahydrocannabinol. J. Pharmacol. Exp. Ther. vasodilator action of anandamide. Nature 400, 452–457. 317, 428–438. Randall, M.D., McCulloch, A.I., Kendall, D.A., 1997. Comparative pharmacology of endothelium-derived hyperpolarizing factor and anandamide in rat isolated mesentery. Eur. J. Pharmacol. 333, 191–197.