Involvement of Cannabinoid Receptors in the Regulation of Neurotransmitter Release in the Rodent Striatum: A Combined Immunochemical and Pharmacological Analysis

Abstract

Despite the profound effect of cannabinoids on motor function, and their therapeutic potential in Parkinson's and Huntington's diseases, the cellular and subcellular distributions of striatal CB₁ receptors are not well defined. Here, we show that CB₁ receptors are primarily located on GABAergic (vesicular GABA transporter-positive) and glutamatergic [vesicular glutamate transporter-1 (VGLUT-1)- and VGLUT-2-positive] striatal nerve terminals and are present in the presynaptic active zone, in the postsynaptic density, as well as in the extrasynaptic membrane. Both the nonselective agonist WIN55212-2 [(R)-(+)-[2,3-dihydro-5-methyl-3[(4-morpholinyl)methyl] pyrrolo[1,2,3-de]-1,4-benzoxazinyl]-(1-naphthalenyl)methanone mesylate salt] (EC₅₀, 32 nm) and the CB₁-selective agonist ACEA [N-(2-chloroethyl)-5Z,8Z,11Z,14Z-eicosatetraenamide] inhibited [³H]GABA release from rat striatal slices. The effect of these agonists was prevented by the CB₁-selective antagonists SR141716A [N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide] (1 μm) and AM251 [1-(2,4-dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-1-piperidinyl-1H-pyrazole-3-carboxamide trifluoroacetate salt] (1 μm), indicating that cannabinoids inhibit the release of GABA via activation of presynaptic CB₁ receptors. Cannabinoids modulated glutamate release via both CB₁ and non-CB₁ mechanisms. Cannabinoid agonists and antagonists inhibited 25 mm K⁺-evoked [³H]glutamate release and sodium-dependent [³H]glutamate uptake. Partial involvement of CB₁ receptors is suggested because low concentrations of SR141716A partly and AM251 fully prevented the effect of WIN55212-2 and CP55940 [5-(1,1-dimethylheptyl)-2-[5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]phenol]. However, the effect of CB₁ agonists and antagonists persisted in CB₁ knock-out mice, indicating the involvement of non-CB₁,CB₁-like receptors. In contrast, cannabinoids did not modulate [³H]dopamine release or [³H]dopamine and [³H]GABA uptake. Our results indicate distinct modulation of striatal GABAergic and glutamatergic transmission by cannabinoids and will facilitate the understanding of the role and importance of the cannabinoid system in normal and pathological motor function.

Introduction

The profound motor and behavioral action of cannabinoids are thought to be mainly mediated by CB₁ receptors (Howlett et al., 1990; Compton et al., 1996). Indeed, CB₁ receptors are present in high density in the basal ganglia (Herkenham et al., 1990, 1991; Mailleux and Vanderhaeghen, 1992). The existence of CB₁ receptor mRNA transcripts in the caudate-putamen (Mailleux and Vanderhaeghen, 1992; Julian et al., 2003) suggests that CB₁ receptors are synthesized in striatal GABAergic efferent neurons (for review, see Romero et al., 2002).

The therapeutic role of the endocannabinoid system is now widely recognized (for review, see Di Marzo et al., 2004). Drugs acting at CB₁ receptors have therapeutic potential in Huntington's and Parkinson's diseases (for review, see Romero et al., 2002; Brotchie, 2003; Lastres-Becker et al., 2003; van der Stelt and Di Marzo, 2003). Namely, both endocannabinoid levels and CB₁ receptor density are selectively altered in the basal ganglia in animal models of Parkinson's and Huntington's diseases (Di Marzo et al., 2000c; Romero et al., 2000; Lastres-Becker et al., 2001a,b; Gubellini et al., 2002; Maccarrone et al., 2003) and in parkinsonian patients (Lastres-Becker et al., 2001a; Hurley et al., 2003). Furthermore, cannabinoids reduce levodopa-induced dyskinesia (Sieradzan et al., 2001; Ferrer et al., 2003).

The mechanisms of cannabinoid action in the striatum are still unclear. Cannabinoids, applied in vivo, increase striatal dopaminergic transmission (Malone and Taylor, 1999; Melis et al., 2000), likely via increasing the neuronal firing in the ventral tegmental area (Robbe et al., 2001) and in the substantia nigra. However, in vitro, they directly did not affect the release of dopamine (Szabo et al., 1999). Cannabinoids depress both GABAergic (Szabo et al., 1998) and glutamatergic [caudate-putamen (Gerdeman and Lovinger, 2001; Huang et al., 2001; Gerdeman et al., 2002; Brown et al., 2003; Ronesi et al., 2004), nucleus accumbens (Robbe et al., 2001, 2002)] synaptic currents through presynaptic CB₁ receptor activation. However, we (Köfalvi et al., 2003) and others (Hájos et al., 2001) have recently shown that cannabinoids still decrease glutamate release in the hippocampus in the CB₁ -/- mouse. Moreover, the endocannabinoid/endovanilloid agonist anandamide in the CB₁ -/- mice and the synthetic cannabinoid/vanilloid agonist arvanil in the wild-type mice depress spontaneous activity and cause catalepsy in a manner independent from CB₁ and TRPV₁ receptor activation (Di Marzo et al., 2000a,b). Furthermore, both anandamide and the synthetic cannabinoid agonist (R)-(+)-[2,3-dihydro-5-methyl(-3[(4-morpholinyl)methyl]pyrrolo[1,2,3-de]-1,4-benzoxazinyl]-(1-naphthalenyl)methanone mesylate salt (WIN55212-2) stimulate GTPγS binding in brain membranes from CB₁ -/- mice, although not in the basal ganglia (Di Marzo et al., 2000b; Breivogel et al., 2001; Monory et al., 2002). All of these observations support the notion that there are additional receptors for these ligands (Pertwee, 2004). Therefore, it is important to determine whether CB₁ receptors are the only receptors responsible for the presynaptic effect of cannabinoids in the striatum.

Thus, our goals were (1) to analyze the distribution of CB₁ receptors to define the colocalization of CB₁ immunoreactivity with markers of different types of nerve terminals and (2) to perform a pharmacological analysis of the effect of cannabinoids on the three main striatal neurotransmitters.

Materials and Methods

Animals

All studies were conducted in accordance with the principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the local Animal Care Committees of our institutes. CB₁ receptor homozygote knock-out mice were generated as described previously (Ledent et al., 1999). The genotype of the mice was tested by the conventional PCR technique on genomic DNA obtained from the tails of experimental animals immediately after the experiments.

Localization of the CB₁ receptor in rat striatal nerve terminals

Subsynaptic distribution of CB₁ receptor in the rat striatum. The solubilization of the presynaptic active zone and the extrasynaptic and postsynaptic fractions from rat striatal synaptosomes was performed according to the method described previously (Phillips et al., 2001), with minor modifications. We confirmed previously (Pinheiro et al., 2003) that this subsynaptic fractionation method allows a >90% effective separation of markers of the presynaptic active zone (syntaxin and synaptosome-associated protein of 25 kDa), postsynaptic density [postsynaptic density-95 (PSD-95) and NMDA receptor subunit 1], and extrasynaptic (synaptophysin) markers and can be used to assess the subsynaptic distribution of metabotropic receptors (Rebola et al., 2003).

Briefly, striata from 12 male Wistar rats were homogenized at 4°C with a Teflon-glass homogenizer in 15 ml of isolation solution [0.32 m sucrose, 0.1 mm CaCl₂, 1 mm MgCl₂, and 0.1 mm phenylmethylsulfonyl fluoride (PMSF)]. The concentration of sucrose was raised to 1.25 m by the addition of 75 ml of 2 m sucrose and 30 ml of 0.1 mm CaCl₂, and the suspension was divided in 10 ultracentrifuge tubes. The homogenate was overlaid with 8 ml of 1.0 m sucrose, 0.1 mm CaCl₂, and 5 ml of homogenization solution and centrifuged at 100,000 × g for 3 h at 4°C. Synaptosomes were collected at the 1.25/1.0 m sucrose interface, diluted 1:10 in cold 0.32 m sucrose with 0.1 mm CaCl₂, and pelleted (15,000 × g for 30 min at 4°C). Pellets were resuspended in 1 ml of 0.32 m sucrose with 0.1 mm CaCl₂, and a small sample was taken for gel electrophoresis. The solubilization procedure was also performed as described by Phillips et al. (2001), with minor modifications. Briefly, synaptosomal suspension was diluted 1:10 with cold 0.1 mm CaCl₂, and an equal volume of 2× solubilization buffer (2% Triton X-100 and 40 mm Tris, pH 6.0) was added to the suspension. The membranes were incubated for 30 min on ice with mild agitation, and the insoluble material (synaptic junctions) was pelleted (40,000 × g for 30 min at 4°C). The supernatant (extrasynaptic fraction) was decanted, and the proteins were precipitated with 6 vol of acetone at 20°C and recovered by centrifugation (18,000 × g for 30 min at 15°C). The synaptic junctions pellet was washed with solubilization buffer, pH 6.0, resuspended in 10 ml of 1% Triton X-100 and 20 mm Tris, pH 8.0, incubated for 30 min on ice with mild agitation, and centrifuged (40,000 × g for 30 min at 4°C), and the supernatant (presynaptic fraction) was processed as described above. PMSF (1 mm) was added to the suspension in all extraction steps. The pellets from the supernatants and the final insoluble pellet (postsynaptic fraction) were solubilized in 5% SDS, the protein concentration was determined by the bicinchoninic acid protein assay, and the samples were added to a 1/6 vol of 6× SDS-PAGE sample buffer before freezing at -20°C.

Western blot analysis was performed using 30 μg of each protein fraction, obtained as described above, which were loaded onto a 7.5% SDS-PAGE gel and then transferred onto polyvinylidene difluoride membranes. Membranes were blocked for 1 h at room temperature in 5% low-fat milk in Tris-buffered saline medium with 0.1% Tween 20 (Merck-Schuchardt, Munich, Germany). The membranes were probed with a rabbit antibody raised against the last 15 aa of the C-terminal of the rat CB₁ receptor (1:5000) overnight at 4°C. Detection was performed using the alkaline phosphatase-conjugated secondary antibody of goat anti-rabbit IgG (1:20,000; Amersham Biosciences, Little Chalfont, UK). Immunoblots were visualized using the enhanced chemifluorescence detection reagent (Amersham Biosciences). The specificity of CB₁ receptor antibody has been confirmed by the laboratory of origin and by the lack of immunostaining in the CB₁ receptor knock-out mice (Katona et al., 2001).

Localization of CB₁ receptor immunoreactivity in different nerve terminals of the rat and mouse striatum. For immunochemical analysis, synaptosomes from striata of male Wistar rats (6 weeks old) and male CD-1 mice (see above) were obtained through a discontinuous Percoll gradient, following the procedure described by Díaz-Hernandez et al. (2002), with minor modifications. Striata were homogenized in 0.25 m sucrose and 5 mm N-[tris[hydroxymethyl]methyl]-2-aminoethanesulfonic acid (TES), pH 7.4. The homogenate was spun for 3 min at 2000 × g at 4°C, and the resulting supernatant was spun again at 9500 × g for 13 min. Then the pellets were resuspended in 8 ml of 0.25 m sucrose and 5 mm TES, pH 7.4. Two milliliters of this synaptosomal suspension were placed onto 3 ml of Percoll discontinuous gradients containing 0.32 m sucrose, 1 mm EDTA, 0.25 mm dithiothreitol, and 3, 10, or 23% Percoll, pH 7.4. The gradients were centrifuged at 25,000 × g for 11 min at 4°C. Synaptosomes were collected between the 10 and 23% Percoll bands and diluted in 15 ml of HEPES-buffered medium (in mm: 140 NaCl, 5 KCl, 5 NaHCO₃, 1.2 NaH₂PO₄, 1 MgCl₂, 10 glucose, and 10 HEPES, pH 7.4).

The striatal synaptosomes were placed onto coverslips previously coated with poly-l-lysine, fixed with 4% paraformaldehyde for 15 min, and washed twice with PBS (in mm: 140 NaCl, 3 KCl, 20 NaH₂PO₄, 15 KH₂PO₄, pH 7.4). Permeabilization was performed in PBS containing 0.2% Triton X-100 for 10 min, and afterward the synaptosomes were incubated in PBS medium containing 3% bovine serum albumin (BSA) and 5% normal rat serum for 1 h. The synaptosomes were then washed twice with PBS and incubated with rabbit anti-CB₁ receptor and chicken anti-vesicular glutamate transporter (VGLUT)-1 (1:5000; Alpha Diagnostic, San Antonio, TX) and guinea pig anti-VGLUT-2 (1:5000; Chemicon, Temecula, CA), or guinea pig anti-vesicular GABA transporter (VGAT; 1:1000; Calbiochem, Darmstadt, Germany), or mouse antityrosine hydroxylase (Tyr-OH; 1:1000; Chemicon), or mouse antisynaptophysin (1:200; Sigma, St. Louis, MO) antibody for 1 h at room temperature. The synaptosomes were then washed three times with PBS/BSA (3%) and incubated for 1 h at room temperature with AlexaFluor-488 (green)-labeled goat anti-rabbit IgG antibodies (1:200; Molecular Probes, Leiden, The Netherlands) and goat anti-chicken or goat antiguinea pig or goat anti-mouse antibodies, all labeled with AlexaFluor-598 (red; 1:200 for all; Molecular Probes). After washing and mounting on slides with Prolong Antifade (Molecular Probes), the preparations were visualized in an Axiovert 200 (Zeiss, Oberkochen, Germany) inverted microscope equipped with a cooled CCD camera and analyzed with MetaFluor 4.0 software (Universal Imaging Corporation, West Chester, PA).

Release experiments

Four male Wistar rats (140-160 g; Gedeon Richter, Budapest, Hungary) or five male young adult mice (20-22 g) were decapitated under ether anesthesia, and the brains were quickly put into ice-cold Krebs' solution of the following composition (in mm): 115 NaCl, 4.7 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 2.5 CaCl₂, 25 NaHCO₃, and 10 glucose, oxygenated with 95% O₂ and 5% CO₂, pH 7.4. All striata were dissected rapidly and used for slice or synaptosomal experiments. For slice experiments, 400 μm slices were cut with a McIlwain tissue chopper (Bachofer, Reutlingen, Germany) and used for incubation (see below). A synaptosomal fraction of the striatum was prepared with slight modifications of the technique described previously (Cunha and Ribeiro, 2000; Köfalvi et al., 2003). The removed striata were homogenized in ice-cold 0.32 m sucrose solution (containing 1 mm EDTA, 1 mg/ml BSA, and 5 mm HEPES, pH 7.4) at 4°C and centrifuged at 2000 × g for 10 min. The supernatant was centrifuged at 13,000 × g for 12 min. The pellet was resuspended in ice-cold 45% (v/v) Percoll in Krebs' solution, pH 7.4, and centrifuged at 13,000 × g for 2 min to eliminate free mitochondria and glial debris. The top layer was washed twice at 13,000 × g for 2 min in oxygenated Krebs' solution at 4°C.

[³H]GABA and [³H]dopamine release experiments from striatal slices. The experiments were performed as described previously (Milusheva et al., 1996; Köfalvi et al., 2000), with slight modifications. Briefly, slices were incubated for 30 min at 37°C in the presence of 5 μCi of [³H]GABA or [³H]dopamine ([³H]DA) in 1 ml of Krebs' solution (Amersham Biosciences). In the case of [³H]GABA, the incubation solution contained 1 mm β-alanine to minimize [³H]GABA uptake into glial cells. After the incubation, four slices were transferred into a polypropylene perfusion chamber and presuperfused (washed) for 1 h. All solutions for [³H]GABA release experiments contained the GABA transaminase (GABA-T)/GAD inhibitor aminooxyacetic acid (200 μm), whereas for [³H]DA experiments, all solutions contained ascorbic acid (0.3 mm) and EDTA (0.03 mm).

After termination of the 1 h presuperfusion period, 3 min samples were collected for tritium assay (sample collection period). At 6 and 36 min after the start of the sample collection period, the release of [³H] transmitters was stimulated twice (EFS₁ and EFS₂) with electrical field stimulation (at 40 V, 2 Hz, 1 ms, 360 bipolar, square-wave pulses; except where noted otherwise), delivered by a Grass S88 Stimulator (Grass Medical Instruments, Quincy, MA) via a pair of platinum ring electrodes. Throughout the experiments, the temperature was maintained at 37°C. All drugs were introduced 20 min before the first or the second electrical field stimulation, as indicated later.

[³H]Glutamate release experiments from striatal synaptosomes. The experiments were performed with slight modifications of our previous study (Köfalvi et al., 2003). The synaptosomes were diluted to 1.5 ml with Krebs' solution and equilibrated with careful oxygenation (95% O₂ and 5% CO₂) at 37°C for 5 min, after which 4 μCi of [³H]glutamate (Amersham Biosciences) was added to the synaptosomes for 5 min. All solutions contained the GABA-T/GAD inhibitor aminooxyacetic acid (200 μm). Ninety-microliter aliquots (∼610 μg of protein) of the preloaded synaptosomes were transferred into 100 μl volume superfusion chambers of a 12-channel suprafusion system (Brandel, Gaithersburg, MD), trapped between two layers of Whatman (Maidstone, UK) GF/C filters, and superfused continuously at a rate of 0.5 ml/min until the end of the experiment. At 12 min after the start of the 45 min washout period, d-aspartate (50 μm; Sigma) was administered for 5 min to effectively reduce the Ca²⁺-independent release of glutamate evoked by high K⁺ (Terrian et al., 1991; Köfalvi et al., 2003). After termination of the 45 min washout, 3 min samples were collected for liquid scintillation assay. All experimental periods were performed at 37°C. At 6 and 36 min after the start of the sample collection period, the release of [³H]glutamate was stimulated twice (S_K1 and S_K2) with 25 mm K⁺ (isomolar substitution of Na⁺ by K⁺ in the buffer) for 3 min. The antagonists were given 20 min before S_K1 (except where noted otherwise), whereas the agonists were given 20 min before S_K2, both being present until the end of the experiments.

Validation of transmitter release. The tritium distribution in the effluent, under basal condition and stimulation, was analyzed by HPLC with fluorometric detection as described previously (Nakai et al., 1999). The majority of tritium represented the respective transmitters (GABA, glutamate, and dopamine), according to our previous findings. Tetrodotoxin (TTX) (1 μm; n = 4 in all cases) strongly inhibited the release of [³H]DA (by 76.5 ± 3.1%; p < 0.001 vs control) and [³H]GABA (by 77.0 ± 8.4%; p < 0.001 vs control), indicating that they are released in response to axonal activity. In the case of synaptosomal experiments, when a Ca²⁺-free buffer supplemented with 1 mm EGTA was superfused after the first stimulation with K⁺, the subsequent K⁺-evoked release of [³H]glutamate was reduced by 71.3 ± 6.1% (n = 4; p < 0.001 vs control) in the rat.

Uptake experiments. The uptake of [³H]GABA and [³H]DA into slices and [³H]glutamate into synaptosomes was investigated in the absence and presence of cannabinoid ligands. To study the uptake of GABA or DA, the incubation with the isotope was performed as described above, at 37°C. After a 30 min incubation, the slices were rinsed three times, weighed, and homogenized, and the tritium content of the homogenates was assayed and expressed in milligrams of tissue weight. For [³H]glutamate uptake experiments, after a 5 min equilibration with continuous gassing at 37°C, 50 μl aliquots of the preparation was transferred into 450 μl of incubation solution containing the drugs or their vehicle (vehicle control). After a 20 min incubation at 37°C, the uptake was terminated on ice, and the synaptosomes were washed three times at 15,000 × g for 5 min, diluted and sonicated in 10% trichloroacetic acid, and assayed for tritium. The uptake of glutamate was strongly dependent on sodium (8.0 ± 0.4% of control; n = 6; p < 0.001 when sodium was replaced by choline chloride) and on temperature (12 ± 2.3% of control; n = 6; p < 0.01 at 12°C).

Radioactivity assay and calculations. The radioactivity released from the preparations was measured with a Packard (Canberra, Australia) 1900 Tricarb liquid scintillation spectrometer, equipped with Dynamic Color Corrected DPM Option providing absolute activity (DPM) calculation and correction for different color quenching. The release of the transmitters was calculated as the percentage of the amount of radioactivity in the tissue at the sample collection time [fractional release (FR%)]. The tissue/synaptosomal tritium uptake was determined as the sum release plus the tissue/synaptosomal content after the experiment. The stimulation-evoked release of the transmitters was calculated by the area-under-the-curve method. All data represent mean ± SEM of n ≥ 4 observations. EC₅₀ values were calculated by fitting the data to sigmoidal logistic equations using the Prism 4.00 (Graph Pad, San Diego, CA) program. When the decrease in baseline was calculated, the tritium content of the last sample before the second stimulus was expressed as the percentage of the tritium content of the last sample before the first stimulus. Statistical significance was calculated by Student's t test or ANOVA, followed by Bonferroni's test for selected pairs of columns, as appropriate, and p < 0.05 was accepted as significant change.

Drugs. TTX, Na₂CNQX, AP-5, and bicuculline methobromide (all from Sigma) were dissolved in distilled water. WIN55212-2 and S(-)-[2,3-dihydro-5-methyl-3-[(4-morpholinyl)methyl]pyrrolo[1,2,3-de]-1,4-benzoxazinyl]-(1-naphthalenyl)methanone mesylate salt (WIN55212-3) (both from Sigma) were dissolved in 0.1 m HCl. Sulpiride (Sigma), 5-(1,1-dimethylheptyl)-2-[5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]-phenol (CP55940),N-(2-chloroethyl)-5Z,8Z,11Z,14Z-eicosatetraenamide(ACEA),Δ⁹-tetrahydrocannabinol(Δ⁹-THC),1-(2,4-dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-1-piperidinyl-1H-pyrazole-3-carboxamide trifluoroacetate salt (AM251) (all from Tocris, Bristol, UK), and N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1 H-pyrazole-3-carboxamide (SR141716A) (National Institute on Drug Abuse, Bethesda, MD) were dissolved in ethanol. The maximum concentrations of vehicles used had no significant effect on the release of the transmitters.

Results

Subsynaptic distribution of CB₁ receptor in the rat striatum

The method of solubilization of subsynaptic fractions allows the antibodies to access the presynaptic active zone and the postsynaptic density, which provided the opportunity to explore the subsynaptic CB₁ receptor distribution. This technique appears to be far more sensitive than classical immunogold electron microscopy for determining the localization of synaptic receptors (Phillips et al., 2001; Rebola et al., 2003). Western blot analyses of the subsynaptic fractions showed significant CB₁ immunoreactivity in all fractions. The total amount of synaptic CB₁ receptors from the initial fraction is denoted as “whole” in Figure 1, representing 100%. After fractionation, we found nearly 23% of the synaptic CB₁ receptors in the presynaptic active zone (Fig. 1, pre) and nearly 30% of them in the postsynaptic density (Fig. 1, post), where electron microscopy generally cannot localize them, likely because of antibody accessibility problems. The rest of the CB₁ receptor immunoreactivity was found in the extrasynaptic fraction of presynaptic and postsynaptic sides (Fig. 1A,B, extra).

Figure 1.

CB₁ receptors are located mainly extrasynaptically but also at the presynaptic active zone and at the postsynaptic density of rat striatal nerve terminals. A, Western blot (representative of 4 similar blots from different groups of animals) comparing CB₁ receptor immunoreactivity, corresponding to the 53 kDa band, in a fraction enriched in the presynaptic active zone (pre), in the postsynaptic density (post), in the nerve terminals portion outside the active zone (extra), and in the initial synaptosomal fraction (whole), from which fractionation was performed. These fractions were obtained by pH fractionation, after solubilization of purified striatal nerve terminals as described in Materials and Methods. Thirty micrograms of protein of each fraction were applied into the SDS-PAGE gel, and a CB₁ antibody was used at a 1:5000 dilution. B, Average distribution of CB₁ receptor immunoreactivity in subsynaptic compartments. The CB₁ receptor density was higher in the extrasynaptic fraction but was also present, to a lesser extent, in the presynaptic active zone and in the postsynaptic density.

Localization of CB₁ receptor immunoreactivity in different nerve terminals of the rat striatum

The protocol for separation of nerve terminals is designed to exclude contamination by postsynaptic elements. Nonetheless, we stained the nerve terminals for PSD-95, a postsynaptic marker protein, and observed a very low (∼0.01%) number of PSD-95-positive elements in synaptophysin-costained plates of synaptosomes (data not shown), establishing the specificity of the isolation procedure. Figure 2, A and B, illustrates that a high percentage (60%) of nerve terminals (identified by synaptophysin) display CB₁ receptor immunoreactivity, according to the well known strong CB₁ receptor expression in the striatum (Herkenham et al., 1990, 1991; Mailleux and Vanderhaeghen, 1992). Approximately 2000 synaptosomes were counted for each marker [i.e., VGAT (for GABAergic axon terminals; rat), VGLUT-1 and VGLUT-2 (for glutamatergic axon terminals; rat and mouse), and Tyr-OH (for catecholaminergic axon terminals; rat)] and for the CB₁ receptor. The strong CB₁ receptor density in GABAergic terminals is in agreement with previous morphological data (for review, see van der Stelt and Di Marzo, 2003). With respect to glutamatergic terminals, the presence of CB₁ receptors was expected based on electrophysiological observations (Gerdeman and Lovinger, 2001; Huang et al., 2001; Brown et al., 2003; Ronesi et al., 2004). Only a low percentage of CB₁ receptor immunoreactivity colocalized with Tyr-OH immunoreactivity, which is a marker for both noradrenergic and dopaminergic terminals.

Figure 2.

CB₁ receptors are present on GABAergic (from rat) and glutamatergic (from rat and mouse) and, to a lesser extent, on catecholaminergic (from rat) terminals. A, Representative double-labeling images of anti-CB₁ receptor (CB₁R) with anti-synaptophysin (marker of all nerve terminals), anti-VGAT (specific marker of GABAergic nerve terminals), anti-VGLUT-1 and anti-VGLUT-2 (specific markers of glutamatergic nerve terminals), and anti-Tyr-OH (specific marker of catecholaminergic nerve terminals). B, A summary of the extent of CB₁ receptor colocalization with the specific markers of each type of nerve terminal (mean ± SEM of n = 4-6 plates) after counting ∼2000 terminals for each marker.

[³H]GABA release from rat striatal slices

After the 1 h washout period, the basal [³H]GABA efflux amounted to 0.192 ± 0.010 FR% (n = 10), similar to that in the hippocampus (Katona et al., 1999; Köfalvi et al., 2000). The [³H]GABA release evoked by the first electrical stimulation (40 V, 2 Hz, 1 ms, 360 shocks, EFS₁) was 0.221 ± 0.041 FR%, whereas the EFS₂/EFS₁ value was 0.958 ± 0.091 (Fig. 3A-C). Although the nonselective cannabinoid agonist WIN55212-2 (0.01-1 μm) failed to affect the basal (resting) release, it concentration dependently and significantly diminished the evoked [³H]GABA release, with an EC₅₀ value of 32 nm (Fig. 3A,B). The maximal effect was obtained at 1 μm (42% inhibition; n = 13; p < 0.01). ACEA, the highly CB₁ receptor-selective agonist (1 μm), also inhibited the evoked [³H]GABA release by 27% (n = 8; p < 0.05) (Fig. 3B). AM251 (1 μm; n = 6) and SR141716A (1 μm; n = 4), the two CB₁ receptor-selective antagonists introduced before the first stimulation, prevented the inhibition by WIN55212-2 (1 μm), whereas they alone had no effect on the release (Fig. 3C). The ionotropic glutamate receptor antagonists AP-5 (50 μm) and CNQX (10 μm), all introduced before the first stimulation, did not affect the action of WIN55212-2 (1 μm) on the evoked [³H]GABA release (n = 8; p < 0.05) (Fig. 3C).

Figure 3.

Cannabinoid agonists inhibit the TTX-sensitive, electrically evoked release of [³H]GABA via CB₁ receptor activation in rat striatal slices. A, [³H]GABA release from striatal slices in the control condition and in the presence of the nonselective cannabinoid agonist WIN55212-2. After 1 h of washout, 3 min samples were collected, as indicated by the x-axis, and counted for tritium. The sample tritium content was expressed as the percentage of the actual tissue tritium content at the time of the sample collection (FR%). The slices were stimulated twice with a pair of platinum electrodes (at 40 V, 2 Hz, 1 ms, 360 bipolar, square-wave pulses), as indicated by the arrows (electrical field stimulation, EFS₁ and EFS₂). The nonselective CB₁ receptor agonist WIN55212-2 was applied, as indicated by the horizontal bar, 20 min before EFS₂. WIN55212-2 decreased the second stimulation-evoked release of [³H]GABA. B, WIN55212-2 concentration dependently attenuated the evoked release of [³H]GABA but did not modify its uptake. ACEA, a highly selective CB₁ receptor agonist, tested at 1 μm, also significantly inhibited the release of [³H]GABA. C, The effect of WIN55212-2 (expressed with the EFS₂/EFS₁ ratio) is prevented by the selective CB₁ receptor antagonists SR141716A (1 μm) and AM251 (1 μm), but not by coapplication of the AMPA/kainate receptor antagonist CNQX (10 μm) and the NMDA receptor antagonist AP-5 (50 μm). ^*p < 0.05; ^**p < 0.01. n ≥ 8 for all data points. CTRL, Control.

The effect of WIN55212-2 (1 and 30 μm) was tested on the uptake of [³H]GABA, to determine whether the decrease in the evoked release is attributable to increased uptake. After the 30 min incubation, the tritium uptake into the slices amounted to 27,085 ± 6741 DPM/mg tissue [vehicle control (CTRL), n = 10], which was unaltered in the presence of 1 μm WIN55212-2 (n = 6) (Fig. 3B). The cannabinoid agonist had no effect on the uptake of [³H]GABA even at 6 μm (data not shown) and 30 μm, which is supramaximal for the CB₁ receptor.

[³H]DA release experiments from slices

After the 1 h washout period, the basal [³H]DA efflux amounted to 0.551 ± 0.040 FR% (n = 10), similar to that found in our previous study (Milusheva et al., 1996). The same electrical field stimulation parameters were used as to evoke [³H]GABA release. The first stimulation-evoked release (EFS₁) was 2.794 ± 0.262 FR%, and the EFS₂/EFS₁ value was 0.757 ± 0.061 (Fig. 4A-C). WIN55212-2 (0.3-10 μm) and CP55940 (1-10 μm) did not significantly affect either the resting or the evoked [³H]DA release (Fig. 4A,B). To reveal whether a cannabinoid effect would appear at a different stimulation frequency, we also stimulated the slices at 0.5 Hz (EFS₁, 2.385 ± 0.400; EFS₂/EFS₁, 0.822 ± 0.108; n = 6) as well as at 10 Hz (EFS₁, 1.719 ± 0.266; EFS₂/EFS₁, 0.810 ± 0.055; n = 4). WIN55212-2 (1 μm; n = 6 and 4) did not modify the evoked release of [³H]DA at any of the frequencies tested. It is possible that the effect of the cannabinoids was hidden because of simultaneous activation of inhibitory and/or excitatory neurotransmission by field stimulation in the intact slice. Thus, we challenged the [³H]DA release with WIN55212-2 (1 μm) in the presence of the ionotropic glutamate receptor antagonists AP-5 (50 μm) and CNQX (10 μm) and the GABA_A receptor antagonist bicuculline (20 μm). However, no alteration of the evoked and resting [³H]DA release by WIN55212-2 was observed under this condition (Fig. 4C). CB₁ receptor activation may lead to the production of nitric oxide (for review, see Howlett and Mukhopadhyay, 2000), which can interact with the release of dopamine in the striatum (Kiss et al., 2004). However, WIN55212-2 (1 μm) had no effect on the release of dopamine in the presence of N-ω-nitro-l-arginine methyl ester (l-NAME) (100 μm), the water-soluble nitric oxide synthase inhibitor (Fig. 4C).

Figure 4.

Cannabinoids do not affect the TTX-sensitive, electrical field stimulation-evoked release and uptake of [³H]DA in rat striatal slices. A, B, Neither WIN55212-2 nor CP55940 modulated the basal outflow or evoked release of [³H]DA nor the uptake of [³H]DA under experimental conditions similar to those used for [³H]GABA (see Fig. 3 A, B). Note that WIN55212-2 also did not affect release if the frequency of stimulation was four times lower or five times higher (see Results). C, The lack of modulation by WIN55212-2 (expressed with the EFS₂/EFS₁ ratio) is not attributable to the presence of facilitatory or inhibitory polysynaptic mechanisms, because the blockade of ionotropic glutamate receptors by CNQX (10 μm) and AP-5 (50 μm) and by the GABA_A antagonist bicuculline (20 μm), respectively, did not reveal any WIN55212-2-mediated modulation, nor did the blockade of nitric oxide synthase by l-NAME (100 μm). Activation of CB₁ receptors by endogenous cannabinoids does not explain the lack of WIN55212-2 effect because the selective CB₁ receptor antagonist SR141716A (10 μm) applied 20 min before EFS₂ had no effect on the second stimulation-evoked release of [³H]DA. n ≥ 6 for all data points. CTRL, Control.

D₂ and CB₁ receptors can form heterodimers. Coactivation of this chimeric receptor by cannabinoids and dopamine increases cAMP accumulation (Glass and Felder, 1997; Jarrahian et al., 2004). This could also mask an inhibitory response mediated by lone CB₁ receptors. But WIN55212-2 (1 μm) did not change the evoked or resting [³H]DA release in the presence of the D₂ receptor antagonist sulpiride (3 μm) (Fig. 4C). Next, we tested whether the CB₁ receptors were already active (by constitutive activity or by endocannabinoids). The CB₁-selective antagonist SR141716A (at 10 μm, a concentration supramaximal to antagonize the CB₁ receptors), introduced after the first stimulation, altered neither the resting nor the evoked release of dopamine (Fig. 4C). Finally, WIN55212-2 at 6 μm did not modify the uptake of dopamine, which amounted to 15,776 ± 4745 DPM/mg tissue (vehicle CTRL, n = 6) (Fig. 4B). Thus, the lack of change in dopamine release could not have been caused by simultaneous reductions in uptake and release.

[³H]Glutamate release experiments from rat striatal synaptosomes

Because astrocytes are equipped with cannabinoid receptors (Di Marzo et al., 2002) and can take up and calcium dependently release glutamate (for review, see Nishizaki, 2004), we performed the following experiments in separated nerve terminals. This protocol allowed us to selectively investigate the presynaptic effects in the absence of extrasynaptic and glia mechanisms.

The basal [³H]glutamate release amounted to 7.19 ± 0.39 FR% (CTRL, n = 16), the [³H]glutamate efflux evoked by the first K⁺ depolarization (S_K1) was 8.50 ± 0.59 FR%, and the S_K2/S_K1 value was 0.823 ± 0.049, data similar to that obtained in the hippocampus (Köfalvi et al., 2003) (Fig. 5A-C). The nonselective cannabinoid agonists CP55940 (0.1-10 μm; EC₅₀, 0.58 μm; maximal effect, 40.2% inhibition), Δ⁹-THC (1-30 μm; 1.46 μm; 25.2%), and WIN55212-2 (0.3-30 μm; 3.38 μm; 53.7%) concentration dependently attenuated the 25 mm K⁺-evoked [³H]glutamate release (n = 4-6) (Fig. 5A-C). It is noteworthy that WIN55212-3, the enantiomer of WIN55212-2 being inactive at the CB₁ receptor, had no significant effect (n = 4) (Fig. 5B). This enantiomer was applied at 6 μm, at which the original compound exerted maximal inhibition. In our previous study (Köfalvi et al., 2003), we found concentration-dependent inhibition of the 25 mm K⁺-evoked [³H]glutamate release by SR141716A, the CB₁-selective antagonist in the hippocampus. A similar inhibition also occurs in striatal synaptosomes (0.6-30 μm;EC₅₀, 3.01 μm; maximal effect, 45.9%). In addition, a structurally related CB₁ antagonist, AM251 (0.6-30 μm; EC₅₀, 3.94 μm; maximal effect, 57.0%), also decreases glutamate release (Fig. 5B). This effect, which is unique for the release of glutamate, makes the antagonists as potent and effective glutamate release inhibitors as WIN55212-2, the agonist.

Figure 5.

Cannabinoids attenuate the Ca²⁺-dependent, 25 mm K⁺-evoked release and the temperature- and Na⁺-dependent uptake of [³H]glutamate in rat striatal synaptosomes. A, B, Arrows indicate the 3-min-long stimulation by high K⁺ superfusion (S_K1 and S_K2). The diagrams show that (in order of potency) CP55940, Δ⁹-THC, SR141716A, WIN55212-2, and AM251 (i.e., both agonists and antagonists of the CB₁ receptor) concentration dependently inhibited the evoked release of [³H]glutamate. WIN55212-2 (6 μm) and AM251 (30 μm) displayed the greatest efficacy. WIN55212-3, the enantiomer inactive at the CB₁ receptor, had no effect at 6 μm. C, SR141716A (SR) attenuated by only one-third the effect of WIN55212-2 (WIN) or CP55940 (CP), but AM251 (AM) fully prevented the action of WIN55212-2 and CP55940. As expected, the evoked [³H]glutamate release was not sodium channel (i.e., TTX) dependent. D, WIN55212-2 inhibited the uptake of [³H]glutamate as well, which was prevented only by AM251. However, the effect of WIN55212-2 was nonstereoselective (i.e., non-CB₁ receptor-mediated). AM251 more potently and effectively inhibited the uptake of [³H]glutamate than WIN55212-2. ^*p < 0.05; ^**p < 0.01. Only one point per ligand was marked for significance for the sake of simplicity. n ≥ 6 for all data points. CTRL, Control.

Next, we tested the effect of WIN55212-2 and CP55940 at the concentrations that had evoked the maximal effect, in the presence of SR141716A and AM251, the CB₁-selective antagonists, at a concentration that did not decrease glutamate release (1 μm). SR141716A partially reversed the inhibition of CP55940 (3 μm) and of WIN55212-2 (6 μm), but the effect of agonists remained significant. However, their effects on the evoked release were completely prevented by AM251 (Fig. 5C).

WIN55212-2 and AM251 at the low micromolar range inhibit veratridine (i.e., sodium channel activation)-evoked glutamate release from whole-brain synaptosomes by direct, competitive blockade of voltage-dependent Na⁺ channels (Nicholson et al., 2003; Liao et al., 2004). In our model, as expected, the 25 mm K⁺-evoked [³H]glutamate release was not TTX (1 μm) (i.e., voltage-gated sodium channel) dependent (Fig. 5C); therefore, the effect of the cannabinoid ligands was independent of sodium channel blockade.

WIN55212-2 slightly but concentration dependently attenuated the resting [³H]glutamate outflow, with a maximal effect of 24% at 6 μm (n = 6; p < 0.05) (Fig. 5A). The same was observed with CP55940 (21% at 10 μm; n = 6; p < 0.05), whereas Δ⁹-THC was devoid of any effect on the resting [³H]glutamate outflow at the concentrations tested. This suggests that some cannabinoid ligands might affect glutamate transporters independently of activation of the CB₁ receptor. To test this hypothesis, we challenged the uptake of glutamate by WIN55212-2. The [³H]glutamate uptake into striatal synaptosomes amounted to 571,127 ± 9282 DPM/50 μl (∼340 μg of protein; n = 24). WIN55212-2 (6 μm) concentration dependently attenuated the uptake of [³H]glutamate with the EC₅₀ value of 2.62 μm (Fig. 5D). SR141716A (1 μm) alone had no significant effect on the uptake and did not reverse the uptake inhibitory action of WIN55212-2 (Fig. 5D). AM251 concentration dependently attenuated the uptake of [³H]glutamate, and again 1 μm AM251 fully prevented the effect of WIN55212-2. WIN55212-3, the enantiomer of WIN55212-2 inactive at the CB₁ receptor, had the same efficacy to attenuate [³H]glutamate uptake (i.e., the action of cannabinoids on [³H]glutamate uptake might not be CB₁ receptor mediated) (Fig. 5D).

Localization of CB₁ receptor expression on glutamatergic nerve terminals of wild-type mouse striatum

Because the pharmacological profile of the cannabinoid receptor underlying the attenuation of glutamate efflux was not entirely identical to that responsible for the reduction of [³H]GABA release, we also explored the localization and function of CB₁ receptors in striatal nerve terminals of wild-type and CB₁ -/- mice. Overall, 1684 immunopositive nerve terminals from three wild-type animals, littermates of the CB₁ -/- mice, were counted. Interestingly, 78.8 ± 8.5% of VGLUT-1- and VGLUT-2-positive nerve terminals were endowed with CB₁ immunoreactivity, indicating an ∼40% greater colocalization than in the rat (Fig. 2A,B).

[³H]Glutamate release experiments from wild-type and CB₁ homozygote null-mutant mouse striatal synaptosomes

The basal and 25 mm K⁺-evoked [³H]glutamate release from the striatal synaptosomes of the rat and wild-type and CB₁ -/- mice is illustrated in Figure 6A. Although the resting and evoked [³H]glutamate efflux were similar in the rat and in the wild-type mouse, the basal release was slightly and significantly higher, whereas the evoked release was slightly and significantly lower in the CB₁ -/- mouse. Interestingly, WIN55212-2 appeared to be less potent and efficacious to inhibit the release of [³H]glutamate in the wild-type mice (Fig. 6C) than in the rat striatum or in the mouse hippocampus (Köfalvi et al., 2003). In contrast, CP55940 inhibited the evoked release by 26% already at 3 μm. AM251 (1 μm) applied 20 min before the first stimulation significantly decreased the S_K2/S_K1 ratio in wild-type mice, and in the presence of AM251, WIN55212-2 (20 μm) did not further inhibit the evoked release of [³H]glutamate (Fig. 6C). Nevertheless, the inhibitory effect of both cannabinoid agonists, as well as that of AM251, persisted in the CB₁ -/- mouse (Fig. 6D). Furthermore, AM251 occluded the effect of WIN55212-2 in the CB₁ receptor knock-out mice. This is in contrast to the hippocampus, where another CB₁ receptor-selective antagonist, SR141716A, applied at 1-5 μm did not prevent the inhibition produced by agonists (Köfalvi et al., 2003). WIN55212-2 (20 μm; 24%; n = 6; p < 0.01) and CP55940 (3 μm; 17%; n = 6; p < 0.01) also significantly inhibited the basal release of [³H]glutamate in the CB₁ null-mutant mouse. WIN55212-2 (20 μm) strongly and significantly inhibited the uptake of [³H]glutamate in both mouse types (Fig. 6B). For the sake of comprehension, the effect of cannabinoid agonists and antagonists on the uptake and the stimulation-evoked release of [³H]glutamate is summarized in Table 1.

Figure 6.

Modulation of the uptake and release of [³H]glutamate in striatal synaptosomes of wild-type and CB₁ receptor knock-out (-/-) mice. A, Properties of [³H]glutamate release in the rat and the two mouse types. The release is slightly but significantly altered in the CB₁ -/- mouse compared with the wild-type mouse. B, WIN55212-2 inhibits the uptake of [³H]glutamate in both mouse types. C, D, WIN55212-2 (although less potently and effectively) and CP55940 attenuate the Ca²⁺-dependent, 25 mm K⁺-evoked release of [³H]glutamate from the wild-type CD-1 and the CB₁ null-mutant mouse striatal synaptosomes. In the presence of AM251, given 20 min before S_K1 and being present in all solutions used, the S_K2/S_K1 ratio significantly diminished, indicating differences between the CD-1 mouse and the Wistar rat, although the likely time-dependent underlying mechanism is unclear. No further inhibition of [³H]glutamate release by WIN55212-2 (20 μm) is observed in the presence of AM251 either in wild-type or CB1-/-mice. ^*p < 0.05; ^**p < 0.01; ^***p < 0.001 versus drug-freec ontrols (CTRL). n ≥ 8 for all data points.

Table 1.

Effect of cannabinoid ligands on the uptake (U) and the stimulation-evoked release (ER) of [³H]glutamate from striatal synaptosomes tested in different animal species

Treatments	Species	U	ER
WIN55212-2	R, WT	↓	↓
WIN55212-3	R	↓	→
WIN55212-2 plus SR141716A	R	↓/→	↓/→
WIN55212-2 plus AM251	R, WT	→	→
SR141716A	R	→	↓
AM251	R	↓	↓
WIN55212-2	KO	↓	↓
WIN55212-2 plus AM251	KO	→	→

R, Rat; WT, wild-type mouse; KO, CB₁ −/− mouse.

Discussion

Activation of CB₁ receptors attenuates neuronal activity by presynaptic inhibition of neurotransmitter release and postsynaptic hyperpolarization (for review, see Freund et al., 2003). We now used a highly sensitive immunochemical method that allows determining the quantitative distribution of CB₁ receptors among subsynaptic elements in isolated nerve terminals. Our novel finding (i.e., the localization of CB₁ receptors in the presynaptic active zone) facilitates our understanding of how CB₁ receptors likely attenuate neurotransmitter release via inhibition of active zone molecular targets (i.e., of N- and P-type calcium channels and release machinery functions) (for review, see Jarvis and Zamponi, 2001). The CB₁ receptor immunoreactivity, found extrasynaptically, which may indicate recycling and/or newly synthesized pools of the CB₁ receptors, is concordant with our previous electron microscopy findings in the hippocampus, where presynaptic CB₁ receptors were found primarily in extrasynaptic membranes of GABAergic boutons (Katona et al., 1999, 2000). In the present study, receptors in the extrasynaptic fraction may comprise postsynaptic receptors outside the postsynaptic density as well, and we found CB₁ receptors also in the postsynaptic density. The presence of postsynaptic CB₁ receptors, which may control the excitability of the dendrites, is consistent with a previous immunocytochemical observation in the striatum (Rodríguez et al., 2001) and other functional data (Childers and Deadwyler, 1996; Bacci et al., 2004). Together, the distribution of CB₁ receptors in the striatum is similar to that of hippocampal A₁ adenosine receptors, which were found in all subsynaptic fractions, although strategically in the active zone (Rebola et al., 2003).

Modulation of GABAergic transmission

Our present finding (i.e., the widespread CB₁ receptor expression in GABAergic terminals) is in agreement with previous reports (Herkenham et al., 1990, 1991; Mailleux and Vanderhaeghen, 1992; Fusco et al., 2004). WIN55212-2 inhibited the evoked release of [³H]GABA with a potency and efficacy similar to those of our and others' previous findings in the hippocampus and striatum (Katona et al., 1999, 2000; Szabo et al., 1998). Moreover, ACEA, a selective CB₁ receptor agonist, also significantly inhibited the release of GABA. The effect of WIN55212-2 was completely prevented by the CB₁-selective antagonists AM251 and SR141716A. Similar to the hippocampus (Katona et al., 1999, 2000), the inhibition of GABA release was not attributable to a decrease in the excitatory inputs or to an increase in reuptake. Thus, nanomolar concentrations of cannabinoid agonists inhibit the release of GABA with a pharmacological profile fully consistent with the activation of CB₁ receptors, regardless of the brain area (striatum vs hippocampus), ligands (e.g., SR141716A, AM251), and techniques (neurochemistry vs electrophysiology) used.

Modulation of dopaminergic transmission

Our present study allowed a direct visualization of CB₁ receptors for the first time on catecholaminergic terminals in the striatum. Only a small percentage of these terminals were endowed with CB₁ receptors, which corresponds to the minor expression of CB₁ receptor mRNA in the substantia nigra (Julian et al., 2003) and other dopaminergic nuclei (Matsuda et al., 1993) and suggests that cannabinoids might not modulate dopamine release directly. Accordingly, WIN55212-2 and CP55940 failed to modify the release or uptake of dopamine, even when GABAergic, glutamatergic, dopaminergic, nitrergic, and endogenous cannabinergic inputs were excluded. These data are in concordance with the results obtained by measuring endogenous dopamine levels (Szabo et al., 1999; de Lago et al., 2004). Together, these observations do not support a major role for direct cannabinoid control of striatal dopamine release (van der Stelt and Di Marzo, 2003).

Modulation of glutamatergic transmission

We visualized a strong CB₁ receptor expression in glutamatergic terminals, colabeled for VGLUT-1 and VGLUT-2, in the rat and mouse striatum. WIN55212-2 and CP55940 attenuated the basal [³H]glutamate outflow, thought to occur mainly via membrane transporters. WIN55212-2, WIN55212-3, and AM251, but not SR141716A, also inhibited [³H]glutamate uptake, even in the CB₁ -/- mouse. The possible underlying mechanisms are detailed in supplemental Figure 1 (available at www.jneurosci.org as supplemental material).

We also showed that cannabinoid ligands concentration dependently inhibited K⁺-evoked, Ca²⁺-dependent [³H]glutamate release. Electrophysiological recordings in the striatum have recently delineated a presynaptic inhibition of excitatory transmission by cannabinoids at a low micromolar range (Gerdeman and Lovinger, 2001; Huang et al., 2001; Gerdeman et al., 2002; Brown et al., 2003; Ronesi et al., 2004). The potency of WIN55212-2 and Δ⁹-THC were similar to those in our present study and to the EC₅₀ value of WIN55212-2 in our previous study in the hippocampus (3.47 μm). Furthermore, the effect of WIN55212-2 and CP55940 was prevented by the CB₁ antagonist AM251 (1 μm), whereas the CB₁-inactive enantiomer WIN55212-3 had no effect. Together, these results suggest that presynaptic CB₁ receptor activation inhibits Ca²⁺-dependent glutamate release in rat striatal synaptosomes.

On the other hand, a part of our present findings, in particular the persisting inhibitory effect of WIN55212-2 and CP55940 in the CB₁ -/- mice, cannot be explained by the exclusive activation of CB₁ receptors in the striatum. Indeed, cannabinoids were shown to interact with other targets to inhibit transmitter release (for review, see Di Marzo et al., 2002; De Petrocellis et al., 2004; Pertwee, 2004). SR141716A is known to block voltage-dependent Ca²⁺ channels and K_Ca channels (EC₅₀, ≈3-5 μm) (Shen and Thayer, 1998; White and Hiley, 1998; Bukoski et al., 2002). This can explain the inhibitory action of SR141716A and its structural analog AM251 on the K⁺-evoked, Ca²⁺-dependent [³H]glutamate release but not that of WIN55212-2, which stereoselectively inhibited glutamate release although it does not stereoselectively inhibit Ca²⁺ channels. Importantly, WIN55212-2 and AM251 were recently shown to block sodium channels (Liao et al., 2004; Nicholson et al., 2003). However, in our model, K⁺-evoked [³H]glutamate release is TTX insensitive (i.e., the activity of sodium channels does not contribute to [³H]glutamate release) (Fig. 5C), therefore the effect of WIN55212-2 cannot be explained by sodium channel blockade either. Conversely, the finding that AM251 prevented the effect of cannabinoids in the CB₁ -/- mice is a clear indication for the involvement of a non-CB₁ cannabinoid receptor. However, the pharmacological phenotype of this non-CB₁ receptor seems to be very similar to the classical CB₁ receptor; therefore, we call it “CB₁-like” receptor. Hence, although both SR141716A and AM251 are known as selective CB₁ receptor antagonists, their selectivity toward a yet unknown and uncloned receptor cannot be determined. For similar reasons, it is difficult to determine the exact contribution of CB₁ and CB₁-like receptors in the rat and wild-type mice.

The question arises whether this CB₁-like receptor is identical to the putative non-CB₁ cannabinoid receptors demonstrated in other studies using CB₁ -/- mouse. WIN55212-2 and anandamide in the low micromolar range inhibit glutamate release in the hippocampus of CB₁ -/- mouse (Hájos et al., 2001; Hájos and Freund, 2002a,b; Köfalvi et al., 2003); however, these actions were insensitive to AM251, opposite from our study. WIN55212-2 and anandamide, but not CP55940, stimulate [³⁵S]GTPγS binding in the whole-brain and cerebellar membranes (Di Marzo et al., 2000b; Breivogel et al., 2001; Monory et al., 2002) but not in the basal ganglia (Breivogel et al., 2001; Monory et al., 2002) of CB₁ -/- mice; therefore, this pathway is also unlikely to mediate the actions described in our study. On the other hand, the identity of the CB₁-like receptor with those non-CB₁, non-VR₁ receptors that were shown to mediate the motor depressant effect of anandamide and other vanillyl compounds (Di Marzo et al., 2000a,b) is still an open possibility, which needs additional investigation. Presumably, this CB₁-like receptor functions regardless of the presence of CB₁ receptors in glutamatergic nerve terminals and can fully compensate the lack of CB₁ receptors in the CB₁ -/- mice, perhaps via developmental upregulation. This idea is also supported by the fact that Δ⁹-THC, which failed to affect locomotor activity in the CB₁ -/- mouse (Di Marzo et al., 2000b), was less efficacious in our study, possibly because it activates mostly CB₁ receptors.

Endocannabinoid levels and CB₁ receptor expression are altered in animal models and patients of Huntington's and Parkinson's diseases (for review, see van der Stelt and Di Marzo, 2003). This illustrates the therapeutic potential of the striatal endocannabinoid system modulation. Opposite from other previous findings (Sieradzan et al., 2001; Ferrer et al., 2003), a recent study with cannabis (Δ⁹-THC) did not reveal alteration of akinesia in parkinsonian patients (Carroll et al., 2004). Thus, it is possible that a CB₁-like rather than the CB₁ receptor (presumably the only receptor of the two that Δ⁹-THC activates) is responsible for the beneficial effects of other cannabinoids in Parkinson's disease. On the other hand, the therapeutic value of CB₁ receptor antagonists in Huntington's disease (van der Stelt and Di Marzo, 2003) can be better explained by the findings of our study (i.e., by the simultaneous facilitation of GABAergic transmission and CB₁ receptor-independent inhibition of glutamate release that could brake the glutamatergic excitotoxicity).

In summary, cannabinoids can modulate motor function via direct inhibition of striatal GABA and glutamate release. Moreover, the demonstration of a pharmacologically fully CB₁-like effect in the CB₁ -/- mice indicates that the conventional cannabinoid ligands, routinely used to identify CB₁ receptors, might be inadequate and should provoke a systemic reevaluation using CB₁-deficient mice of previous results based solely on the pharmacological identification of the CB₁ receptor.