INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE MAJOR PSYCHOACTIVE CONSTITUENT of Cannabis sativa, ⁹-tetrahydrocannabinol (THC) (10), exerts its cellular effects by binding and activating a cell-surface, cannabinoid-selective receptor (26). Two major subtypes of the cannabinoid receptor have been cloned and characterized; the first subtype is found primarily in the brain (CB1) (26), and the second is found primarily in cells of myeloid lineage (CB2) (27). Both receptor subtypes have the structural features typical of G protein-coupled receptors, including seven-transmembrane domains (26, 27). The cellular effects of CB1-receptor activation in neural cells, including inhibition of adenylyl cyclase (18), inhibition of voltage-operated N-type Ca²⁺ channels (3, 23), and activation of inward rectifier K⁺ current and K⁺ A-current (5, 16, 24) are inhibited by pertussis toxin, which ribosylates and inhibits G proteins of the G_i/G_o subtype. The CB1 receptor is also activated by a novel eicosanoid, N-arachidonylethanolamine (anandamide), which has been isolated from the brain and has been suggested to be an endogenous CB1 agonist (7).

There is evidence in the literature that marijuana and THC have a direct vasodilator effect on cerebral blood vessels. Marijuana smoking produces a dose-related increase in global cerebral blood flow (CBF) in humans that is not due to changes in sympathetic cerebrovascular regulation or PCO₂ and that is consistent with cerebral vasodilation (25). The increase in CBF is accompanied by a parallel increase in the velocity of flow through the middle cerebral artery (25). In addition, marijuana-intoxicated individuals show evidence of impaired cerebral autoregulation in response to a change in posture from supine to standing (25). The mechanism for this effect is not known; however, it is not accompanied by impaired peripheral sympathetic activity and suggests that cerebral arterioles are maximally dilated by THC and, therefore, are unable to respond to reduced perfusion pressure. A recent study has demonstrated that CB1 agonists dilate cerebral arterioles in an in situ perfusion preparation (9), supporting the hypothesis that the effects of marijuana on CBF are due to direct effects of the drugs on vascular tone. However, the mechanism(s) and the cell type in the vessel wall expressing the CB1 receptor responsible for the cerebrovascular effects of CB1 agonists are not known. The present studies were undertaken to determine whether cannabinoid receptors are expressed in cerebral vascular smooth muscle cells (VSMC), to investigate the electrophysiological bases of cannabinoid-induced cerebral vasodilation by determining the effect of agonists on voltage-activated L-type Ca²⁺ channel in cat cerebral VSMC, and to determine whether the CB1 receptor mediates the actions of cannabinoids on the cerebral vascular tone. The results of these studies demonstrate that cat cerebral VSMC express cDNA encoding the CB1 receptor, which, when stimulated by WIN-55,212-2 or anandamide, functions to dilate cat cerebral microvessels by inhibiting Ca²⁺ influx through a dihydropyridine-sensitive L-type Ca²⁺ channel in cerebral VSMC.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Vascular muscle cell dispersion. Adult mongrel cats of either sex were deeply anesthetized with Telazol (10 mg/kg im; Aveco, Fort Dodge, IA) and their brains removed. Cerebral microvessels [200-230 µm outer diameter (OD)] were dissected free of arachnoid, cut into small pieces, and placed in 2-ml vials containing an enzyme solution of the following composition (in mM): 134 NaCl, 5.2 KCl, 1.2 MgSO₄, 0.33 KH₂PO₄, 0.05 CaCl₂, 11 glucose, and 10 HEPES, with 100 U/ml collagenase (class II, Worthington), 0.5 mg/ml bovine serum albumin, and 1 mM trypsin inhibitor (Sigma, St. Louis, MO). The enzyme solution containing the arterial pieces was maintained at 37°C and pH 7.4 and was continually stirred at 10 rpm for 45 min. Supernatant fractions were collected every 5 min, and the dispersion of VSMC was examined under a microscope. The vascular smooth muscle origin of the cells was confirmed by indirect immunocytochemical staining. The freshly isolated vascular muscle cells did not show specific staining with fluorescein isothiocyanate-conjugated rabbit anti-human von Willebrand's factor (vWF), a marker for endothelial cells that cross-reacts with cat vWF (Zymed Laboratories, San Francisco, CA), whereas the cells exhibited high-intensity staining with anti--actin-Cy3-conjugated antibody (Sigma).

RT-PCR, cloning, and sequencing. Total RNA was extracted from freshly dispersed cat cerebral VSMC using silica-based column chromatography (RNAeasy, Qiagen, Chatsworth, CA). RNA (2 µg) was reverse transcribed using commercial kits (Pharmacia First Strand cDNA Synthesis Kit) in a volume of 33 µl. Aliquots of RT reactions (5 µl) were amplified with KlenTaq DNA polymerase mix (Clontech Laboratories, Palo Alto, CA) using primers specific to the rat CB1 receptor (GenBank accession no. X55812): forward, 5'-ATGAAGTCGATCCTAGATGGCC-3'; reverse, 5'-TCACAGAGCCTCGGCGGACGTG-3' (26). Components of the reactions were mixed, overlaid with mineral oil, and heated to 95°C for 3 min; cycled for 10 cycles of 94°C for 30 s, 66°C for 30 s, and 68°C for 3 min; followed by 30 cycles of 94°C for 30 s, 61°C for 30 s, and 68°C 3 min; and ending with a 30-min extension at 72°C. Fifty microliters of each reaction were purified by electrophoresis on a 1.5% agarose gel (SeaKem GTG, FMC BioProducts), stained with ethidium bromide, and scanned using a Molecular Dynamics FluorImager. Bands of expected size were excised by using the scan printout as a guide, solubilized in NaI at 50°C for 10 min, and purified using a GlassMax column (GIBCO). Purified product (150 ng) was ligated into pCR2.1 using the T/A Cloning Kit (Invitrogen, Carlsbad, CA), transformed into Escherichia coli INVF', and plated on LB-kanamycin (50 µg/ml) plates containing X-Gal. Positive colonies were grown in LB-kanamycin overnight, and plasmid DNA was purified using the Qiaprep Spin Miniprep plasmid purification kit (Qiagen). Plasmid DNA (1 µg) was sequenced using dye-primer (M13 forward and reverse primers) Texas Red cycle sequencing (Amersham) on a Vistra model 725 automated DNA sequencer to confirm the identity of inserts. Four independent clones were identified that contained sequences with high homology to human, rat, and mouse CB1 receptors. Both strands of these four independent clones were sequenced repetitively using dye-terminator cycle sequencing on an ABI system (Perkin Elmer).

Western blot analysis. Cat and rat cerebral cortex microsomal fractions were prepared separately using standard differential centrifugation methods. Freshly dispersed cat or rat cerebral VSMC were pelleted by centrifugation and resuspended in buffer (150 mM NaCl, 1.0% NP-40, 0.5% deoxycholate, 0.1% SDS, 1 mM EDTA, and 50 mM Tris, pH 8.0) containing a protease inhibitor cocktail (PharMingen, San Diego, CA) consisting of 16 µg/ml benzamidine hydrochloride, 10 µg/ml phenanthroline, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 10 µg/ml pepstatin A, and 50 mM phenylmethylsulfonyl fluoride. Cells were solubilized by vigorous vortexing and sonication. Insoluble material was pelleted by centrifugation for 5 min at 14,000 g. Cerebellar microsomes (50 µg) and VSMC lysates (10 µg) were mixed separately with 2× Laemmli SDS sample buffer, boiled for 5 min, and loaded onto SDS-PAGE gels (4% stacking, 10% resolving; Bio-Rad Ready-gels). After fractionation, proteins were electrophoretically transferred to nitrocellulose membranes (Bio-Rad) and probed with the primary antibody anti-rat CB1 antibody (1-77) (1:1,000; kindly provided by Dr. Ken Mackie, University of Washington, Seattle, WA) and the secondary antibody goat anti-rabbit horseradish peroxidase (1:1,000; Bio-Rad). Prestained molecular mass marker proteins (Bio-Rad) were used to estimate sizes of immunoreactive protein bands.

Electrophysiology. Macroscopic Ca²⁺ currents were recorded at room temperature (22°C) using a conventional whole cell voltage-clamp technique (13) with Ba²⁺ as a charge carrier. The VSMC were dialyzed with the intracellular solution containing (in mM) 135 CsCl, 5 Mg-ATP, 5 HEPES, and 1 ethylene glycol-bis(-aminoethyl ether)-N,N,N',N'-tetraacetic acid, with pH adjusted to 7.2 with CsOH. The extracellular solution bathing the cells was composed of (in mM) 135 tetraethylammonium (TEA) chloride, 5 BaCl₂, 1 MgCl₂, 11 glucose, 10 HEPES, 5,4-aminopyridine, and 0.001 tetrodotoxin, with pH adjusted to 7.4 with TEA hydroxide. Inward Ca²⁺ currents were elicited every 1 s by 200-ms depolarizing pulses from a holding potential of 70 mV to test potentials between 50 and +50 mV in 10-mV increments. Whole cell Ca²⁺ currents were recorded through a List EPC-7 patch-clamp amplifier (Darmstadt, Germany) before and after addition to the bath of CB1 agonists in the presence or absence of the various blockers. Cell capacitance was determined from the current amplitude in response to a hyperpolarizing voltage pulse of 1 mV for 25 ms from a holding potential of 0 mV. Average cell membrane capacitance was 26 ± 3 pF (range 15-28 pF, n = 22). The inward Ca²⁺ current was normalized by dividing its amplitude by the cell capacitance (expressed in pA/pF). The signal was low-pass filtered at 1 kHz (3 dB frequency) by an eight-pole low-pass Bessel filter and digitized at a sampling frequency of 2.5 kHz. Data were analyzed using a pCLAMP software package (Axon Instruments).

Isolated vessel studies. Isolated cat cerebral microvascular segments (8-10 mm in length, 0.20- to 0.30-mm OD) were placed in a perfusion chamber, cannulated at both ends with glass micropipettes, and secured in place with 8-0 polyethylene sutures (Ethicon, Somerville, NJ) using a stereomicroscope. Side branches of the arteries were tied off with 10-0 polyethylene sutures. The arterial segments were superfused with physiological saline solution that was aerated with a 95% O₂-5% CO₂ gas mixture and maintained at 37°C and pH 7.35. The inflow cannula was connected in series with a volume reservoir and a pressure transducer (Gould, Cleveland, OH) to allow continuous monitoring of transmural pressure. Internal diameter of the arteries was measured using a video microscopy system composed of a television camera and a video micrometer, as previously described (11). After an equilibration period of 15 min, the cannulated arteries were pressurized to 90 mmHg and then maintained at this pressure throughout the course of the experiment. After an additional 15-min equilibration, the functional integrity of the vessels was verified by the contractile response to 60 mM KCl and the vasodilator response to the direct vasorelaxant sodium nitroprusside (1 µM) after preconstriction with serotonin (5 µM). After three successive washes, the arterial segments were allowed to equilibrate for an additional 15 min. The concentration-dependent relaxation responses of the cerebral arterial preparations to anandamide (10, 100, and 300 nM) or WIN-55,212-2 (10, 30, and 100 nM) were then determined in the absence and presence of the CB1 cannabinoid receptor antagonist SR-141716A (100 nM) after preconstriction of the vascular segments with 5 µM serotonin.

Statistics. Data are presented as means ± SE. Differences between groups were assessed using ANOVA or Student's t-test with a Bonferroni correction for multiple comparisons.

Drugs and chemicals. All chemicals were analytic grade. SR-141716A was the kind gift of Sanofi Researche (Montpelier, France). ATP (magnesium salt), sodium nitroprusside, BaCl₂, nifedipine, CdCl₂, TEA chloride, TEA hydroxide, 4-aminopyridine, and tetrodotoxin were all purchased from Sigma Chemical (St. Louis, MO). Anandamide was purchased from Biomol (Plymouth Meeting, PA), and R-(+)-WIN-55,212-2 was purchased from RBI (Natick, MA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Identification of cannabinoid CB1 receptor mRNA in cat cerebral VSMC. To determine whether cat cerebral VSMC express CB1 mRNA, we amplified cDNA encoding the CB1 receptor with the use of RT-PCR with oligonucleotide primers derived from the published sequence of rat CB1 (see RT, cloning, and sequencing) (26). The sequence of the amplified and cloned product showed >98% homology to the rat (26) and human (12) CB1 cannabinoid receptor (Fig. 1A). This finding provides evidence for the expression of the CB1 cannabinoid receptor mRNA in cat cerebral VSMC (GenBank accession no. U94342). The existence of the protein for the CB1 receptor in cerebral VSMC and the brain cortex was determined by Western blot analysis. Results of the Western immunoblot studies (Fig. 1B) revealed the expression of an immunoreactive band with an apparent molecular mass of 56 kDa in VSMC lysate and brain cortex microsomes obtained from cats or rats when probed with a polyclonal anti-rat CB1(1-77) antibody raised in rabbits. An additional band with a molecular mass of 61 kDa was also detected in the cat VSMC lysate. This 61-kDa band was not detected in either the VSMC lysate or cortical microsomes of the rat. It is possible that the CB1 receptor of cat cerebral VSMC has a different pattern of glycosylation and, therefore, a different molecular mass. It is likely that the 56-kDa protein is the nonglycosylated receptor. These results indicate that the protein for the CB1 receptor is expressed in cerebral arterial muscle cells and the cerebral cortex of the cat and rat.

View larger version (112K): [in this window] [in a new window]   Fig. 1.   A: deduced amino acid sequences of cat cerebral vascular smooth muscle cell (VSMC) cannabinoid CB1 receptor cDNA. Amino acids are indicated in a single-letter code. Comparisons show amino acid sequences of product amplified from reverse-transcribed mRNA extracted from cat cerebral VSMC (GenBank accession no. U94342) with previously reported rat and human CB1 receptor sequences. Identical amino acid residues are shown together in black boxes; sequences share >98% overall homology. B: Western blot analysis showing expression of immunoreactive bands in both cerebellar microsomes and VSMC lysate of cat and rat at ~56 kDa when probed with polyclonal anti-CB1(1-77) antibody.

Characterization of macroscopic Ca²⁺ currents in cat cerebral VSMC. Inward Ca²⁺ currents carried by Ba²⁺ were recorded from freshly dispersed cat cerebral arterial muscle cells with high-Cs⁺ solution in the pipette and with bath solution containing TEA and tetrodotoxin (1 µM) to block outward K⁺ currents and inward Na⁺ currents, respectively. During step depolarization from a holding potential of 70 mV to test potentials from 50 to +50 mV, an inward Ca²⁺ current began to activate at 40 mV and peaked at +10 mV. The peak current-voltage relation over the range of more positive potentials started to reverse between +40 and +50 mV (Fig. 2A, c) and remained stable for >20 min. As shown in Fig. 2A, a and c, the inward Ca²⁺ current, elicited by a 200-ms depolarizing pulse from a holding potential of 70 mV to a test potential from 50 to +10 mV, was inhibited >96% by nifedipine (2 µM). This inward current was also completely blocked by 50 µM Cd²⁺ (data not shown). The high voltage threshold of channel activation and the sensitivity to blockade by either nifedipine or Cd²⁺ indicated that the inward current recorded from freshly dissociated cat cerebral arterial VSMC passes through a high voltage-activated, dihydropyridine-sensitive, L-type Ca²⁺ channel. We have previously shown (11) that the magnitude of this inward macroscopic current is not associated with substantial rundown or inactivation during alteration of the frequency of depolarization from slower to higher rates under identical experimental conditions.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   A: inhibition of L-type Ca2+ current by CB1 agonists anandamide (AEA) and WIN-55,212-2 (WIN) in cerebral VSMC. Bath application of either 300 nM AEA (a) or 100 nM WIN (b) markedly inhibited whole cell Ca2+ current elicited by depolarizing steps from 50 to +10 mV from a holding potential of 70 mV within 1-3 min. Nifed., nifedipine. Broken lines indicate zero current level. c: Effects of AEA (300 nM) and WIN (100 nM) on normalized mean peak current-voltage relation of Ca2+ currents in cerebral VSMC. WIN was more efficient in inhibiting mean peak Ca2+ current than AEA. Both AEA and WIN depressed mean peak current without shifting current-voltage curve. B: comparison of averaged percent maximum peak Ca2+ current inhibition evoked by increasing concentrations of either WIN (10, 30, and 100 nM) or AEA (10, 100, and 300 nM) in cat cerebral VSMC at a test potential of +10 mV. AEA was less potent in inhibiting L-type Ca2+ current at all concentrations studied. Data are means ± SE; nos. in parentheses represent no. of cells studied.

Effects of the CB1 agonists anandamide and WIN-55,212-2 on peak macroscopic L-type Ca²⁺ channel current in isolated cerebral VSMC. As depicted in Fig. 2, A and B, the application of 300 nM anandamide to the bath induced a 37 ± 4% (n = 12) maximum reduction in the amplitude of peak Ca²⁺ current elicited by 200-ms depolarizing pulses from a holding potential of 70 mV to test potentials from 50 to +10 mV in 10-mV increments. The inhibitory action of anandamide on Ca²⁺ current was not readily reversible on washout (data not shown); therefore, each cell was exposed to only one concentration of anandamide. The inhibitory action of anandamide on normalized peak Ca²⁺ current is concentration dependent; the peak Ca²⁺ current was reduced by 12 ± 3% at 10 nM (n = 5) and by 23 ± 5% at 100 nM (n = 5; Fig. 2B). Higher concentrations of anandamide (1 µM) did not cause further inhibition of the peak L-type Ca²⁺ current (data not shown).

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Time course of inhibitory effects of AEA and WIN on peak L-type Ca2+ current. Normalized peak L-type Ca2+ current (ICa) is plotted as a function of time. Inhibitory effects of both AEA (300 nM) and WIN (100 nM) remained stable over a period of 5 min. There is no apparent rundown or inactivation associated with normalized peak L-type current recorded from cerebral VSMC. Data points are means, and bars are SE.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Blockade of WIN- and AEA-induced inhibition of peak L-type Ca2+ current by cannabinoid (CB1)-receptor antagonist SR-141716A (SR) and pertussis toxin (PTX). Application of CB1 antagonist SR (100 nM, 3 min) had no effect on basal peak L-type Ca2+ current but completely prevented WIN (100 nM)- and AEA (300 nM)-induced inhibition of L-type Ca2+ current in cerebral VSMC (A and B, top). VSMC were pretreated with PTX (1 µg/ml) for 5-6 h before experiment (A and B, bottom). PTX pretreatment eliminated both WIN- and AEA-evoked inhibition of peak L-type Ca2+ current compared with control peak L-type Ca2+ current recorded from VSMC treated with vehicle.

Effects of anandamide and WIN-55,212-2 on cerebral arterial tone. The vasodilator effects of the endogenous CB1 ligand anandamide and the cannabimimetic drug WIN-55,212-2 were determined in either intact or endothelium-denuded cat cerebral microvascular segments pressurized to 90 mmHg and preconstricted with 5 µM serotonin. Under control conditions, the basal diameter of the cerebral microvascular segments averaged 214 ± 25 µm (n = 6) and decreased to 160 ± 10 µm when stimulated with 5 µM serotonin. Cumulative addition of either anandamide (10, 100, and 300 nM) or WIN-55,212-2 (10, 30, and 100 nM) to the bath induced concentration-related relaxation of the arterial segments (Fig. 5, A and B). Similar to the electrophysiological findings, anandamide was less potent than WIN-55,212-2 in inducing relaxation of the arterial segments. The vasodilator effects of anandamide and WIN-55,212-2 were greatly attenuated after pretreatment of the cerebral arterial preparations with the CB1-receptor antagonist SR-141716A (100 nM) (Fig. 5, A and B; n = 6, P < 0.05). Anandamide induced a similar degree of relaxation in cerebral arterial preparations contracted with high extracellular KCl (60 mM) (data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Effect of CB1-receptor antagonist SR (100 nM) on AEA (10, 100, and 300 nM)- and WIN (10, 30, and 100 nM)-induced relaxations (means ± SE, n = 6) of cannulated and pressurized (90 mmHg) cat cerebral arterial segments preconstricted with serotonin (5 µM). Graphs show concentration-dependent relaxation of cerebral arterial segments induced by AEA (A) and WIN (B) in absence or presence of CB1 antagonist SR. * CB1 antagonist significantly prevented (P < 0.05, n = 6) relaxations induced by either AEA or WIN at all concentrations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In these studies we have demonstrated that isolated cerebral VSMC of the cat express the protein and the mRNA for CB1 cannabinoid receptor. Furthermore, external application of the CB1 agonist WIN-55,212-2 produced potent and profound inhibition of Ca²⁺ current recorded in cat cerebral VSMC. We have identified the Ca²⁺ current carriers as L-type Ca²⁺ channels on the basis of a characteristically high threshold of activation, slow activation and inactivation kinetics, and sensitivity to blockade by nifedipine. The properties of this L-type Ca²⁺ channel current are similar to those reported in a variety of tissues (2, 32). The effect of WIN-55,212-2 was mimicked by the putative endogenous cannabinoid ligand anandamide (7) and reversed by the CB1-receptor antagonist SR-141716A (33). Inhibition of the L-type Ca²⁺ current by either anandamide or WIN-55,212-2 was abolished by pretreatment of the cells with pertussis toxin, which is consistent with the findings of others (3, 19, 23) and supports the involvement of a G protein-mediated signaling pathway. We have also demonstrated that activation of the CB1 receptor results in dilation of isolated pressurized cerebral vessels preconstricted with serotonin or high KCl. These findings suggest that CB1 agonists mediate the inhibition of Ca²⁺ influx through a G protein-mediated pathway, resulting in a decrease in intracellular Ca²⁺ available to sustain smooth muscle cell contraction (1) and decreased vessel tone. Because the extent of cerebral microvascular tone is one of the mechanisms by which CBF is regulated (39), these results are consistent with the hypothesis that the endogenous cannabinoid agonist plays a potential role in the regulation of regional CBF.

Several other investigators have demonstrated that activation of the CB1 receptor results in decreased Ca²⁺ transients in hippocampal neurons (40) and neuroblastoma cells expressing the endogenous CB1 receptor (3, 23), in AtT20 cells (24), and in superior cervical ganglion cells (28) transfected with the CB1 receptor. In these studies, CB1-receptor agonists produce inhibition of Ca²⁺ current through N-type (23, 28, 40) or P/Q-type Ca²⁺ channels (24, 40) but not through L-type Ca²⁺ channels (23, 24). CB1 receptor-mediated inhibition of Ca²⁺ channel opening required an intact G protein signaling cascade because pretreatment of the cells with pertussis toxin (3, 23, 28), N-ethylmaleimide (22), or guanosine 5'-O-(2-thiodiphosphate) (28) abolished the cannabinoid agonist response. The precise signaling mechanism, i.e., whether CB1 receptor-mediated inhibition is due to a direct interaction of a component of the G protein with the channel or occurs as a result of an as yet undefined change in a second messenger, is an open question. One current hypothesis is that G subunits bind directly to a domain with the sequence QXXER in the 1-2 linker region of N-, P-, and Q-type Ca²⁺ channels (4).

The inhibition of L-type Ca²⁺ currents by CB1-receptor agonists in cat cerebral microvessel VSMC has many biochemical features in common with CB1 receptor-mediated inhibition of Ca²⁺ current in cell lines and neurons. In particular, the CB1 agonist WIN-55,212-2 is a very potent inhibitor of Ca²⁺ current, with an IC₅₀ in the range of 10⁸ M, and the effect is pertussis toxin sensitive. However, the Ca²⁺ current in the cat cerebral VSMC is carried predominantly through L-type Ca²⁺ channels, and our data demonstrate that activation of the CB1 receptor in cerebral VSMC inhibits Ca²⁺ influx through these L-type Ca²⁺ channels. This finding is somewhat surprising because there are only a few examples of coupling of heptahelical receptors to L-type Ca²⁺channels via pertussis toxin-sensitive G proteins and because the CB1 receptor itself does not affect L-type Ca²⁺ currents in other cells (23, 24). However, G_o proteins are involved in the coupling of muscarinic receptors to L-type Ca²⁺ channels in GH4C1 pituitary cells, in which L-type rather than N-type Ca²⁺ channels subserve stimulus-secretion coupling (15, 17, 37). Several reports (8, 15) have suggested that the L-type Ca²⁺ channel is also subject to direct inhibition by a pertussis toxin-sensitive G protein in neuronal cells. To our knowledge, this is the first demonstration of the inhibition of cerebral VSMC L-type Ca²⁺ channels by the activation of heptahelical receptors acting through a pertussis toxin-sensitive G protein. However, it is not clear from the present studies whether the mechanisms by which activation of the CB1 receptor inhibits N- and P/Q-type Ca²⁺ channels in neurons and L-type Ca²⁺ channels in cerebral VSMC are the same.

Although no single study has investigated the structure-activity relationships among more than two CB1-receptor agonists, previous studies of the effect of CB1-receptor activation on Ca²⁺ channel currents revealed that WIN-55,212-2 is the most efficacious of the CB1-receptor agonists investigated (22-24, 28, 40). The intrinsic activity of anandamide is less than that of WIN-55,212-2 in N18 cells (22), superior cervical ganglion cells transfected with CB1 cDNA (28), and cerebral VSMC (present study) but equal to that of WIN-55,212-2 in hippocampal neurons expressing endogenous CB1 receptor (40). In other studies, both THC (3) and its synthetic congener, CP-55940 (28, 35), also have intrinsic efficacies lower than that of WIN-55,212-2. The most likely explanation for this discrepancy, as suggested by Mackie and colleagues (40), is that anandamide and the THC analogs have a lower absolute intrinsic activity but can produce full agonist effect in cellular systems in which the expression levels of the CB1 receptor and/or G protein effectors are high.

There are several reports in the literature (9, 41) that cannabinoids vasodilate cerebral and other vascular beds and elicit hypotensive action in anesthetized animals. Ellis and colleagues (9) have reported that low concentrations of THC and anandamide produce dilation of rabbit cerebral arterioles in situ. Whereas these results are consistent with our finding that cerebral VSMC Ca²⁺ channels are inhibited by CB1 agonists, indomethacin inhibited the vasodilator effects of both anandamide and THC in this model. This supports the alternative hypothesis that CB1-receptor activation increases the production of vasodilator prostanoids. Consistent with this hypothesis is the previous finding (36) that anandamide and THC increase the release of arachidonic acid from rat cortical astrocytes and that this release is inhibited by both the CB1-receptor antagonist SR-141716A and pretreatment with pertussis toxin, indicative of the involvement of the CB1 receptor in the effect in a G protein-mediated pathway. In light of the multiple cell types that express the CB1 receptor, it is entirely possible that multiple mechanisms are involved in the regulation of cerebrovascular tone by the CB1 receptor.

A number of studies (6, 21, 29-31, 43) in other vascular beds have demonstrated that anandamide acts as a vasorelaxant through a variety of diverse mechanisms. Administration of anandamide to anesthetized rats produces a triphasic effect on blood pressure (21, 41, 42): first, a brief period of hypotension accompanied by a profound reduction in heart rate; second, a brief period of hypertension; and third, a period of hypotension. The third period is consistent with inhibition of transmitter release from sympathetic terminals and appears to be mediated by a cannabinoid receptor because it is reversed by SR-141716A. Interestingly, the third period of hypotension is not apparent in conscious rats, possibly because sympathetic outflow is diminished and the pressor effect predominates (21, 38). A recent study (43) has provided evidence that activation of the CB1 receptor by an endogenous ligand produced during hemorrhagic shock contributes to the profound hypotension that occurs. This evidence supports the hypothesis that the endogenous ligand is anandamide and that the cellular source is the macrophage. Although the mechanism for the vasodilation was not explored in depth, changes in nitric oxide synthase activity are not likely to be involved.

Anandamide produces vasorelaxation in isolated arterial segments through a mechanism(s) that is not clear and that may vary depending on the vessel bed. In rat mesenteric artery, anandamide-induced relaxation is not consistent with an effect on the CB1 receptor because its effect is not mimicked by the high-affinity CB1-receptor agonists HU-210 or WIN-55212-2 (29) and is either partially inhibited by relatively high (i.e., >1 µM) concentrations of SR-141716A (31, 44, 45) or not inhibited by SR-141716A at all (29). The mechanism for this effect is not clear, although there is evidence that anandamide inhibits the release of Ca²⁺ from caffeine-sensitive stores (45) and opens large-conductance K⁺ channels in vascular smooth muscle cells (29). In bovine coronary arteries, relaxation in response to anandamide is abolished by diazomethylarachidonyl ketone, an inhibitor of anandamide amidohydrolase, suggesting that anandamide serves as an arachidonic acid donor in this vessel (30). In the kidney, low concentrations of anandamide stimulate the release of nitric oxide from endothelial cells, and anandamide vasorelaxation is sensitive to both SR-141716A and the nitric oxide synthase inhibitor nitro-L-arginine methyl ester, suggesting a role for CB1 receptor in the regulation of nitric oxide synthase activity in this tissue (6). In summary, there are several mechanistic paradigms that explain anandamide-mediated vasodilation in various tissues and under various physiological and pathological conditions. It is remarkable that, despite the variety of mechanistic explanations, the physiological effect that is consistently observed is vasorelaxation.

The physiological role of CB1-receptor regulation of cerebral vascular tone remains to be determined. However, the studies reported here suggest that one function of the endogenous ligand for the CB1 receptor is to increase regional CBF. Anandamide has been isolated from the brain and has the functional characteristics of a CB1-receptor agonist (7, 22), which has led to the suggestion that it is the endogenous ligand of the CB1 receptor (7). Although the regulation of anandamide production in the brain is not completely understood, there is strong evidence that brain concentrations of anandamide increase during severe brain ischemia (20, 34). Hansen and co-workers (14) have reported that neurons exposed to an azide produce anandamide, which suggests that dying neurons synthesize anandamide. In light of the actions of anandamide on the regulation of Ca²⁺ influx into cerebral VSMC, one could speculate that anandamide released from ischemic cells acts via CB1 receptors of VSMC to increase blood flow to ischemic regions of the brain. Furthermore, these findings suggest that the role of the CB1 receptor in the normal brain is more fundamental than previously appreciated. It is likely that the endogenous ligand of the CB1 receptor targets multiple cell types in the brain and plays a role in fundamental processes including the regulation of regional CBF.