In vivo effects of CB₂ receptor-selective cannabinoids on the vasculature of normal and arthritic rat knee joints

Abstract

Background and purpose:

Cannabinoids (CBs) are known to be vasoactive and to regulate tissue inflammation. The present study examined the in vivo vasomotor effects of the CB₂ receptor agonists JWH015 and JWH133 in rat knee joints. The effect of acute and chronic joint inflammation on CB₂ receptor-mediated responses was also tested.

Experimental approach:

Blood flow was assessed in rat knee joints by laser Doppler imaging both before and following topical administration of CB₂ receptor agonists. Vasoactivity was measured in normal, acute kaolin/carrageenan inflamed and Freund's complete adjuvant chronically inflamed knees.

Key results:

In normal animals, JWH015 and JWH133 caused a concentration-dependent increase in synovial blood flow which in the case of JWH133 was blocked by the selective CB₂ receptor antagonist AM630 as well as the transient receptor potential vanilloid-1 (TRPV1) antagonist SB366791. The vasodilator effect of JWH133 was significantly attenuated in both acute and chronically inflamed knees. Given alone, AM630 had no effect on joint blood flow.

Conclusion and implications:

In normal joints, the cannabinomimetic JWH133 causes hyperaemia via a CB₂ and TRPV1 receptor mechanism. During acute and chronic inflammation, however, this vasodilatatory response is significantly attenuated.

Introduction

Cannabis sativa is the source of at least 60 distinct alkaloids, which make up a group of compounds called cannabinoids. Cannabinoids can be categorized depending on whether they are synthetic, plant-derived (phytocannabinoids) or naturally produced in the body (endocannabinoids). The endocannabinoid system is believed to play an important role in health and disease where it can ameliorate the severity of disorders, such as pain, multiple sclerosis, schizophrenia and emesis (Di Marzo and Petrocellis, 2006; Pertwee, 2006a). Overproduction of endocannabinoids, however, could have a detrimental effect on the body, leading to problems such as obesity, infertility and inflammatory disorders. Thus, strategic control of the endocannabinoid system with selective agonists or antagonists may prove to have therapeutic value in the treatment of various diseases.

Two cannabinoid receptors have been cloned, viz. CB₁ and CB₂ (Matsuda et al., 1990; Munro et al., 1993). These G_i/o protein-coupled receptors are distributed throughout the body and are involved in the control of miscellaneous physiological processes, such as pain perception, appetite and vasoregulation. CB₁ receptors are predominantly found on nerve terminals in the central and peripheral nervous systems (Tsou et al., 1998; Pertwee, 1999a; Farquhar-Smith et al., 2000), although they have also been localized in non-neuronal tissues, such as the pituitary gland, spleen and immunocytes (Bouaboula et al., 1993). The primary location of CB₂ receptors is on immunocytes (Pertwee, 1999b, 2006a), but they have also been identified on peripheral nerves (Griffin et al., 1997) and in the central nervous system (Van Sickle et al., 2005). A plethora of cannabinoid ligands have been developed with fairly high selectivity for CB₁ and CB₂ receptors (Pertwee, 1999b, 2006b). Two selective CB₂ receptor agonists are JWH015 ((2-methyl-1-propyl-1H-indol-3-yl)-1-napthalenylmethanone) and JWH133 ((6aR,10aR)-3-(1,1-dimethylbutyl)-6a,7,10,10a-tetrahydro-6,6,9-trimethyl-6H-dibenzo[b,d]pyran). The K_i values for JWH015 are 383 and 13.8 nM for the CB₁ and CB₂ receptors, respectively (Chin et al., 1999), while JWH133 has K_i of 677 and 3.4 nM for the CB₁ and CB₂ receptors, respectively (Huffman et al., 1999). For antagonists, the aminoalkylindole AM630 (6-iodo-2-methyl-1-[2-(4-morpholinyl)ethyl]-1H-indol-3-yl](4-methoxyphenyl) methanone) has been shown to be an effective blocker of CB₂ receptors (Hosohata et al., 1997).

Cannabinoids are known to exert potent cardiovascular effects in vivo. For example, systemic administration of the endocannabinoid anandamide, which acts on CB₁ and CB₂ receptors, causes hypotension and bradycardia (Varga et al., 1996; Högestatt and Zygmunt, 2002). Moreover, anandamide has been shown to produce vasodilatation in the brain (Ellis et al., 1995), liver (Garcia et al., 2001), kidney (Deutsch et al., 1997), heart (Wagner et al., 2001; Ford et al., 2002) and mesentery (Randall et al., 1996; White and Hiley, 1998; Jarai et al., 1999). A number of studies using more selective cannabinoid analogues have confirmed that CB₁ receptor activation leads to relaxation of vascular smooth muscle and systemic hypotension (for review, see Randall et al., 2004). Indeed in the rat knee joint, local administration of the selective CB₁ receptor agonist, arachidonyl-2-chloroethylamide causes an increase in synovial blood flow (Baker and McDougall, 2004). Less is known about the vasomotor effects of CB₂ receptor agonists. Wagner et al. (2005) reported that in vitro perfusion of rat hearts with JWH133 caused an increase in coronary blood flow that was insensitive to AM630 antagonism. In vivo assessment of CB₂ receptor agonists on tissue blood flow have so far not been reported.

Evidence has been emerging in recent years that suggests that cannabinoids are able to bind to and activate the transient receptor potential vanilloid-1 (TRPV1) channel. TRPV1 channels are non-selective cation channels expressed on the axon terminals of a subgroup of primary afferent neurons (Caterina et al., 1997; Szallasi, 2002). In joints, activation of TRPV1 channels by local capsaicin treatment leads to inflammatory changes, including increased synovial blood flow (Karimian et al., 1995; Varga et al., 2005). Interestingly, the CB₁ receptor agonist arachidonyl-2-chloroethylamide has also been shown to act on TRPV1 channels to cause synovial hyperaemia in the rat knee (Baker and McDougall, 2004). Whether CB₂ receptors are similarly coupled to TRPV1 channels has never been explored.

The current study was undertaken to assess the vasomotor effects of the two CB₂ receptor agonists, JWH015 and JWH133, on knee joint blood flow in vivo. The role of TRPV1 channels in these vasoregulatory processes was also investigated. Further studies were carried out to examine the effect of joint inflammation on CB₂ receptor-mediated responses using animal models of acute and chronic arthritis.

Methods

All experimental procedures received prior approval from the University of Calgary Animal Care Committee, which is in accordance with the Canadian Council for Animal Care. Adult male Wistar rats (242–495 g) were obtained from Charles River Laboratories (Montreal, Quebec, Canada) and housed in pairs at room temperature under a 12 h light–dark cycle, with free access to water and standard rat chow. Animals were randomly assigned to one of three experimental groups: naïve control, acutely inflamed and chronically inflamed.

Induction of joint inflammation

Two distinct models of arthritis were used in the present investigation based on the different phases of inflammation. Thus, the kaolin/carrageenan arthritis model produces an acute inflammatory reaction within 24 h after induction, while adjuvant monoarthritis is more suitable to assess the chronic phase of joint inflammation. Prior to arthritis induction, knee joint diameters were measured using electronic digital callipers (Mitutoyo Instruments, Tokyo, Japan) oriented medio laterally across the joint line between the femoral condyle and the tibial plateau. Diameters were compared before and upon completion of inflammation development, which is at 24 h for the kaolin/carrageenan model and at 1 week for the adjuvant monoarthritis model.

In the acutely inflamed group of rats, inflammation was induced in the right knee joint by kaolin and carrageenan. Under isoflurane anaesthesia (2–5% isoflurane; 100% O₂ at 1 l min⁻¹), 0.2 ml of 2% kaolin was injected through a 27-gauge needle into the posterior and anterior synovial cavity, followed by repeated limb extensions and flexions for 10 min to ensure adequate dispersion of the suspension within the joint and to cause articular abrasion. Subsequently, 0.2 ml of 2% carrageenan was injected into the joint by the same procedure. These animals were allowed to recover for 24 h prior to blood flow assessment.

For chronic inflammation, rats were anaesthetized with 2–5% isoflurane (100% O₂ at 1 l min⁻¹), and a localized monoarthritis was induced as previously described (Donaldson et al., 1993; McDougall et al., 1995). The right knee joint was shaved and 0.2 ml of Freund's complete adjuvant (heat-killed Mycobacterium tuberculosis, 1 mg ml⁻¹) was injected through a 27-gauge needle into the knee joint with 0.1 ml being introduced into the posterior compartment of the joint and 0.1 ml being injected anteriorly. Animals were allowed to recover for 1 week prior to the blood flow experiments. Sham-injected animals were not tested as an intra-articular injection of sterile saline has been shown to have no effect on synovial blood flow or vasoreactivity (Karimian et al., 1995).

Surgical preparation of animals

On the day of experimentation, animals were deeply anaesthetized with urethane (25% stock solution, 2 g kg⁻¹, i.p.). Depth of anaesthesia was confirmed by a lack of a pinch withdrawal reflex applied to the hindpaw as well as an absence of a corneal blink reflex. A longitudinal incision was made in the neck and the trachea was isolated and cannulated to permit unobstructed breathing. The right carotid artery was then isolated and cannulated with heparinized saline (100 U ml⁻¹)-filled polyethylene tubing (Portex Fine Bore Tubing, 0.5 mm ID, 1.00 mm OD; SIMS Portex Ltd., Kent, UK). The carotid cannula was attached to a pressure transducer (1050 Pressure Transducer; Stoelting Co., Wood Dale, IL, USA) and mean arterial pressure (MAP) was recorded continuously throughout the experiment via a blood pressure monitor (Pressure Monitor BP-1; World Precision Instruments, Sarasota, FL, USA). The animal was then placed in dorsal recumbency on an electric-heating blanket (TR-100; Fine Science Tools Inc., Vancouver, British Columbia, Canada) and the internal body temperature was maintained at 37 °C as measured by a rectally inserted electronic thermometer. An ellipse of skin covering the anteromedial region of the right knee joint was removed and the underlying tissue was kept moist by regular superfusion of 37 °C physiological saline (0.9% NaCl).

Measurement of knee joint blood flow

Knee joint blood flow was measured by laser Doppler perfusion imaging (LDI) as previously described (Karimian et al., 1995; McDougall et al., 1995). The technique involved placing a laser Doppler perfusion imager (MoorLDI V.2; Moor Instruments Ltd., Devon, UK) 30 cm above the exposed rat knee and angling the scanner head to avoid reflectance artefact. A low-power (2 mW) red He–Ne laser (λ=633 nm) was directed onto the surface of the knee joint and by means of a motor-controlled mirror the laser beam scanned in a raster pattern across the joint capsule. Careful placement of opaque black cloth at the margins of the knee ensured that only the joint capsule was scanned. The image resolution was set at 42 pixels × 55 pixels with a scan speed of 4 ms pixel⁻¹. At each pixel point, photons in the incident laser beam underwent Doppler shift by virtue of erythrocytes flowing in the tissue. These Doppler-shifted photons were captured by a photodetector in the scanner head and a perfusion value was calculated based on the velocity and concentration of erythrocytes circulating in the microvasculature. A two-dimensional colour-coded image of tissue perfusion was then constructed and stored for later offline analysis.

To achieve a more accurate appreciation of the time course for cannabinoid activity, the imager was set to the single-point measurement mode to maximize the temporal resolution of the instrument. In these experiments, the laser beam remained stationary at a discrete locus on the surface of the joint capsule, and blood flow measurements were continuously acquired. The precise positioning of the laser beam was randomly chosen, but was usually placed in the vicinity of an observable microvascular network. A steady blood flow level was measured for about 1 min before a cannabinoid agonist was topically applied to the knee joint. Blood flow was then measured continually over the succeeding 5 min. In subsequent scanning experiments, the timeline for image acquisition was designed so as to coincide with the maximal vasomotor effect of the test agents.

Experimental protocol

All test reagents were placed in a water bath and gently warmed to 37 °C. A control LDI scan of the knee was taken and then the test drug was topically applied to the surface of the joint capsule as a 0.1 ml bolus. The CB₂ receptor agonists tested in this study were JWH015 (10⁻¹⁴–10⁻¹¹ mol) and JWH133 (10⁻¹⁴–10⁻¹² mol). Blood flow scans of the joint were then taken at 0, 1.5, 2, 3 and 5 min following drug application. A cumulative concentration–response curve to each agonist was generated in separate groups of animals and sequential doses of agonist were applied to the joint following each 5-min scan. In other experiments, each dose of the CB₂ receptor agonist was co-administered with the selective CB₂ receptor antagonist AM630 (10⁻⁸ mol; 0.1 ml bolus topical). The concentration of AM630 was based on previous successful antagonism in the rat cardiovascular system (Ford et al., 2002). To determine whether the transient potential vanilloid receptor-1 (TRPV1) is involved in CB₂ receptor-mediated vasoregulation, additional experiments were undertaken in which animals were pretreated with the TRPV1 receptor antagonist SB366791. At 20 min following an intraperitoneal injection of SB366791 (500 μg kg⁻¹), concentration–response effects of JWH133 were assessed. This treatment regimen with SB366791 has previously been shown to successfully block TRPV1 receptors in the rat knee joint (Varga et al., 2005). It should be pointed out that JWH015 proved to be a weak, non-selective agonist for the CB₂ receptor and as such was neither used in the SB366791 experiments nor tested in the inflammatory models.

At the conclusion of the experiment, the animal was killed by anaesthetic overdose (pentobarbital sodium, 240 mg intracardiac) and a final scan of the joint was performed. This dead scan constituted a ‘biological zero' measurement, which corresponded to tissue optical noise and residual Brownian motion. This ‘biological zero', which was typically about 5–10% of the basal perfusion level, was subtracted from all image perfusion values prior to data processing.

Image analysis and statistics

Knee joint LDI scans were analysed by proprietary perfusion image-processing software (Moor Instruments Ltd.). For each image, an average blood flow value for the entire anteromedial knee joint capsule was calculated and expressed in arbitrary perfusion units (PUs). Blood flow changes in response to drug administration were expressed as percentage change in perfusion between control and test scans. The maximal vasomotor effect (E_max) and the EC₅₀ were calculated from a sigmoidal nonlinear regression curve fit. All data sets conformed to a Gaussian distribution and as such were tested with parametric statistical analyses using GraphPad Prism software (GraphPad Software, San Diego, CA, USA). ANOVA was used to assess significant differences in blood flows within (one-way ANOVA) and between (two-way ANOVA) different animal groups. A Bonferroni post hoc test was used to determine whether groups of data were significantly different at a specific concentration of cannabinoid. A significance level of P<0.05 was used for each test and all reported values represent means±s.e.m. for ‘n' observations.

Drugs and reagents

JWH015, JWH133 and AM630 were all purchased from Tocris Bioscience, Ellisville, MO, USA; SB366791, Freund's complete adjuvant, λ-carrageenan, kaolin, urethane and heparin were all obtained from Sigma-Aldrich Ltd., Oakville, Ontario, Canada. SB366791 and all cannabinoid agonists and antagonists were dissolved in dimethyl sulphoxide and cremophor, diluted to their working concentrations, aliquoted and stored at either 4 °C or −20 °C as required. The final concentration of vehicle was maintained at 2% dimethyl sulphoxide and 1% cremophor for each concentration. This vehicle formulation has previously been found to have no significant effect on joint blood flow (Baker and McDougall, 2004).

Results

CB₂ receptor effects in normal knee joints

LDI-determined basal blood flow in normal rat knees was 421±26.8 PU (n=52). Application of the CB₂ receptor agonists JWH015 and JWH133 onto the surface of normal rat knees caused a marked but transient increase in synovial blood flow. Continuous measurement of joint blood flow at a single discrete location indicated that vasodilatation was rapid in onset in response to CB₂ receptor activation, with maximal blood flow being reached approximately 1.5 min after drug application (Figure 1). Synovial blood flow subsequently returned to control levels at around the 4-min time point. Concentration–response curves were constructed for JWH015 and JWH133 and both agonists showed a concentration-dependent effect (Figure 2). The calculated E_max for JWH015 was 18.5±3.2% with an EC₅₀ of 1.9 × 10⁻¹⁴ mol. JWH133 on the other hand had an E_max of 36.4±5.3% and an EC₅₀ of 7.9 × 10⁻¹⁴ mol. The selective CB₂ receptor antagonist AM630 had no effect on JWH015-mediated vasodilatation (P=0.81, two-way ANOVA; n=9–10; Figure 2a). Conversely, co-administration of AM630 with JWH133 significantly inhibited the vasodilatatory effect of the cannabinoid (P<0.0001, two-way ANOVA; n=11–38; Figure 2b).

Figure 1.

Single-point measurements of knee joint blood flow at a discrete locus on the joint capsule showing typical responses to topical application of the CB₂ receptor agonists JWH015 (a) and JWH133 (b). Blood flow gradually rose, with maximal vasodilatation occurring 1.5 min after drug administration. Joint blood flow subsequently returned to control levels by about 4 min.

Figure 2.

Concentration–response effects of JWH015 (a) and JWH133 (b) on synovial blood flow. The selective CB₂ receptor antagonist AM630 (10⁻⁸ mol) was unable to block the vasodilator effect of JWH015, but significantly attenuated the hyperaemic response of JWH133 (^***P<0.001, ^**P<0.01 vs control; two-way ANOVA followed by Bonferroni post hoc test; n=11–38). Data are presented as mean percent change in perfusion±s.e.m.

Administration of 10⁻¹² mol JWH133 without previous application of the lower concentrations of the cannabinoid still elicited around a 40% increase in joint blood flow (data not shown), indicating that tachyphylaxis was not occurring in these experiments.

Since JWH015 only produced a weak, non-selective effect on synovial blood flow, the subsequent experiments were only carried out with the more efficacious JWH133.

Role of TRPV1 in CB₂ receptor-mediated vasodilatation

To determine if the TRPV1 ion channel is involved in the activity shown by JWH133, animals were treated with the TRPV1 antagonist SB366791. As shown in Figure 3, the vasodilator effect of JWH133 was blocked in SB366791-treated animals. A two-way ANOVA confirmed that the JWH133 concentration–response curves in SB366791-treated vs non-treated animals were statistically different from each other (P<0.0001; n=12–38).

Figure 3.

Vasodilator effect of JWH133 is attenuated in animals pretreated with the TRPV1 antagonist SB366791 (500 μg kg⁻¹). ^***P<0.001, ^**P<0.01, ^*P<0.05 two-way ANOVA followed by Bonferroni post hoc test; n=12–38. Data are shown as means±s.e.m.

Joint oedema following inflammation

Comparison of joint diameters before and after inflammation induction confirmed oedema formation in the animal models of arthritis. In kaolin/carrageenan acutely inflamed knees, joint diameter increased from 9.0±0.07 to 15.1±0.44 mm, while chronically inflamed adjuvant monoarthritic knees showed an increase in knee diameter from 9.0±0.13 to 12.2±0.42 mm. Both increases were found to be statistically significant (P<0.0001, paired Student's t-test; n=10–13 knees).

Assessment of JWH133 in inflamed knee joints

Basal blood flow to acutely inflamed knees was 566±65 PU (n=10). Administration of JWH133 to acutely inflamed rat knee joints failed to elicit any significant change in synovial blood flow (Figure 4). A two-factor ANOVA confirmed that the vasodilator effect of JWH133 seen in normal joints was abrogated in the acute inflammatory model (P<0.0001, two-way ANOVA; n=8–12).

Figure 4.

Effect of joint inflammation on JWH133-mediated vasodilatation. The hyperaemic effect of JWH133 was significantly reduced in both acute and chronically inflamed knees (^***P<0.001, ^*P<0.05 two-way ANOVA followed by Bonferroni post hoc test; n=8–38). Data are presented as means±s.e.m.

In chronically inflamed joints, the hyperaemic effect of JWH133 was also significantly attenuated compared to normal control knees (P<0.0001, two-way ANOVA; Figure 4). Basal blood flow to adjuvant monoarthritic knees was 513±47.5 PU (n=10).

Effect of AM630 on normal, acutely and chronically inflamed knees

Topical administration of the selective CB₂ receptor antagonist AM630 (10⁻⁸ mol) had no effect on synovial blood flow in normal (P=0.29), acutely inflamed knees (P=0.81) or chronically inflamed (P=0.05) rat knee joints (Figure 5). Thus, AM630 does not appear to be acting as an inverse or partial agonist to increase blood flow in this model system.

Figure 5.

Effect of topical application of the selective CB₂ antagonist AM630 (10⁻⁸ mol) on knee joint blood flow. AM630 had no effect on joint blood flow in normal, acutely inflamed or chronically inflamed knees. Data are mean % of control values±s.e.m.

Blood pressure effects

As shown in Table 1, local administration of all of the test compounds had no significant effect on MAP. Thus, all reported vasomotor changes are a direct effect of the drugs on articular blood vessels and are not a secondary consequence of systemic changes in blood pressure.

Table 1.

MAP taken before (control) and at 1.5 min following topical application of the CB2 receptor agonists JWH015 and JWH133, and the CB2 receptor antagonist AM630

Drugs	MAP (mm Hg)	n values
JWH015 (mol)
Control	77±5	10
10−14	75±4	10
Control	75±5	10
10−13	73±4	10
Control	74±4	10
10−12	73±3	10
Control	71±3	10
10−11	69±2	10
JWH133 (mol)
Control	89±3	41
10−14	87±4	38
Control	90±4	38
5 × 10−14	88±5	28
Control	88±5	28
10−13	82±5	23
Control	82±5	24
5 × 10−13	75±6	11
Control	85±5	28
10−12	89±6	19
AM630 (mol)
Control	94±5	17
10−8	93±6	17
SB366971 (μg kg−1)
Control	94±9	8
500	93±9	10

The effect of the TRPV1 antagonist SB366791 on MAP is also shown (values are expressed as means±s.e.m.).

Abbreviation: MAP, mean arterial pressure.

Discussion

Cannabinoids acting via the CB₂ receptor are known to heavily influence the activity and function of the nervous system; however, the effect of selective CB₂ receptor agonists on the cardiovascular system are less well understood. The present study found that the synthetic CB₂ receptor agonists JWH015 and JWH133 produced a significant increase in synovial blood flow when applied locally to the joint. These hyperaemic responses were a direct effect of the drugs on the joint microenvironment and not a secondary systemic reaction since the agonists were restricted to the periphery and had no influence on mean arterial blood pressure (see Table 1). Both agonists exhibited similar potencies; however, JWH133 was found to be more efficacious and showed greater selectivity for the CB₂ receptor than JWH015, whose mild vasodilator effect was not blocked by the selective CB₂ receptor antagonist AM630. The difference in responsiveness of these two drugs is likely related to the comparatively low affinity of JWH015 for the CB₂ receptor (Chin et al., 1999; Huffman et al., 1999). Thus, JWH133 is the pharmacological tool of choice for in vivo vasomotor assessment and, as such, follow-up experiments investigating the mechanism of action of cannabinoids and the effect of joint inflammation on CB₂ receptor responses was confined to this particular agonist.

It should be noted that the pharmacological tools currently available for cannabinoid research all possess some limits to selectivity. For example, JWH015, JWH133 and AM630 all show a mild affinity for the CB₁ receptor. Thus, it is entirely feasible that the vasodilator responses observed in this study are being mediated by CB₁ receptors and not CB₂ receptors as reported. Attempts to block CB₁ receptors with an antagonist such as AM251 were not carried out in this study, as AM251 also has a mild affinity for CB₂ receptors. A further consideration is that the JWH compounds may be acting via the putative GPR55 cannabinoid receptor (Johns et al., 2007; Ryberg et al., 2007) or indeed via other, so far undefined, orphan receptors. Until robust, highly selective drugs become available and all potential cannabinoid receptor subtypes have been identified, the customary caveat with respect to the precise cannabinoid receptors involved in JWH-mediated vasomotor control should be considered.

The site of action of the tested cannabinoids is not readily identifiable in this in vivo model. The general viewpoint that CB₂ receptors are localized primarily on immunocytes would indicate that these cells could be involved in the vasomotor effects of JWH015 and JWH133. Indeed, cannabinoids have been shown to activate human leukocytes, leading to the release of vasodilating agents, such as nitric oxide and pro-inflammatory chemokines (Stefano et al., 1996; Sugiura et al., 2004). This, however, would not account for the increased blood flow observed in normal joints that do not contain an appreciable number of immunocytes. Alternatively, the cannabinoids could be acting directly on the vascular smooth muscle to cause vasorelaxation. The endocannabinoid anandamide produces vasodilatation by opening large-conductance calcium-activated potassium channels (Grainger and Boachie-Ansah, 2001; White et al., 2001; Wagner et al., 2005), which causes hyperpolarization of smooth muscle cells and vascular relaxation. The vascular endothelium has also been implicated in the vasodilatatory effect of cannabinoids. A number of in vitro studies have shown that cannabinoids can act either as an endothelially derived hyperpolarizing factor or by stimulating the endothelium to cause the secondary release of vasodilator chemicals, such as nitric oxide (Randall et al., 1996; Deutsch et al., 1997; Stefano et al., 2000). A final putative mechanism by which cannabinoids alter tissue blood flow is via their modulatory effects on sensory and sympathetic neurotransmission. Cannabinoids are able to inhibit noradrenaline release from postganglionic sympathetic efferents (Ishac et al., 1996; Varga et al., 1996; Niederhoffer and Szabo, 1999), thereby attenuating sympathetic vascular tone and leading to hyperaemia. Articular blood vessels are also innervated by sensory nerves containing vasoactive neurotransmitters, such as substance P, vasoactive intestinal peptide and calcitonin gene-related peptide (Bjurholm et al., 1990; Abramovici et al., 1991; McDougall et al., 1997), all of which have been shown to increase synovial blood flow (Lam and Ferrell, 1993; McDougall et al., 1995, 1999; McDougall and Barin, 2005). These neuropeptides are released from vasosensory nerves in response to activation of the capsaicin-sensitive TRPV1 channel. Interestingly, anandamide and its synthetic analogues are thought to act on TRPV1 channels to cause the secondary release of pro-inflammatory neurotransmitters, which in turn cause vasodilatation of neighbouring blood vessels (Zygmunt et al., 1999; Ralevic et al., 2000; Smith and McQueen, 2001; Baker and McDougall, 2004). In the present study, blockade of TRPV1 channels with SB366791 significantly attenuated the vasodilator effect of JWH133. This is the first reported evidence showing that TRPV1 channels are essential for CB₂ receptor functional activity. Similar observations have been described for CB₁ receptor-mediated vasomotor control where the hyperaemic effect of the CB₁ receptor agonist arachidonyl-2-chloroethylamide was attenuated following TRPV1 channel blockade (Baker and McDougall, 2004). It appears, therefore, that both CB₁ and CB₂ receptors are somehow functionally coupled to TRPV1 channels to orchestrate vasoregulation in the knee joint. Whether JWH133 binds directly to TRPV1 receptors to cause secondary release of pro-inflammatory neuropeptides or whether some intracellular pathway exists to allow cross-talk between TRPV1 and CB₂ receptors remains to be resolved. While any combination of the mechanisms discussed above may be responsible for the vasodilator effect of the JWH compounds described here, other pathways cannot be discounted. Indeed, such is the complex nature of cannabinoid biology that several studies report contradictory evidence as to the mechanism of action of these eicosanoids on the vasculature (for reviews, see Mendizabal and Adler-Graschinsky, 2003; Randall et al., 2004). Future studies using highly selective cannabinoid agonists in different vascular beds will hopefully reveal the precise mechanism by which CB₂ receptor activation leads to vascular smooth muscle relaxation.

Changes in JWH133-mediated vasodilatation in acute and chronically inflamed knees

When applied topically to acutely inflamed knee joints, JWH133 failed to elicit any significant change in articular blood flow across the concentration range tested. Similarly in chronically inflamed knees, the vasodilator effect of the cannabinoid was conspicuously attenuated compared to normal control joints. One possible explanation for this inability of JWH133 to elicit a hyperaemic response in arthritic knees could be that the synovial microvasculature is already maximally vasodilated such that supplementary smooth muscle relaxation is unattainable in these models. Studies testing other vasodilators in these same inflammatory models, however, have successfully demonstrated that arthritic joint blood vessels possess a vasodilator reserve, indicating that vascular smooth muscle relaxation is still possible in these joints (McDougall and Barin, 2005; Zhang and McDougall, 2006). An alternative reason for the lack of a vasodilatatory response to JWH133 in arthritic knees is that there could be an alteration in articular CB₂ receptor expression and/or sensitivity in inflamed joints; however, this possibility requires further experimental assessment. CB₂ receptor expression on macrophages, microglia, dendritic cells and splenocytes is known to be downregulated following exposure to an inflammatory stimulus (Lee et al., 2001; Carlisle et al., 2002; Matias et al., 2002). Inflammation can cause overactivity of the endocannabinoid system, leading to increased release of anandamide and/or 2-arachidonoylglycerol (Varga et al., 1998; D'Argenio et al., 2006; Oka et al., 2006), and prolonged exposure of cannabinoid receptors to endogenous and synthetic cannabinoid ligands can cause internalization of these G_i/o protein-coupled receptors (Fan et al., 1996; Rubino et al., 2000; Shoemaker et al., 2005). In other organ systems, such as the gut, however, there is evidence of cannabinoid receptor upregulation despite elevated endocannabinoid levels (Izzo et al., 2001; Massa et al., 2004; Kimball et al., 2006). It appears, therefore, that the regulation of cannabinoid receptor expression during inflammation depends upon the tissue involved as well as the nature of the inflammatory stimulus. Interestingly, fatty acid amide hydrolase, the enzyme responsible for the degradation of endocannabinoids, is more active in inflamed gut compared to control animals (Izzo et al., 2001). This continuous turnover of endocannabinoids during intestinal inflammation implies that endocannabinoids cannot accumulate in the tissue and therefore cannabinoid receptor downregulation is averted. Future examination of fatty acid amide hydrolase activity, cannabinoid receptor expression and endocannabinoid levels in inflamed joints is required to help to explain the altered vasomotor responses to JWH133 demonstrated in the present investigation.

In conclusion, the present investigation clearly demonstrated the functional presence of CB₂ receptors in rat knee joints, whose activation leads to a vasodilator response. Vanilloid TRPV1 channels are essential for the hyperaemic action of JWH133 and further studies are required to determine the molecular and biochemical pathways that link TRPV1 and CB₂ receptors. The attenuation of CB₂ receptor-mediated vasodilatation in acute and chronically inflamed joints suggests an alteration in CB₂ receptor expression or sensitivity following an arthritic insult.

Abbreviations

CB

cannabinoid

LDI

laser Doppler perfusion imaging

MAP

mean arterial pressure

PU

perfusion unit

TRPV1

transient receptor potential vanilloid-1

Conflict of interest

The authors state no conflict of interest.