Stimulation of Endocannabinoid Formation in Brain Slice Cultures through Activation of Group I Metabotropic Glutamate Receptors

.

Pharmacological studies suggest that 2-AG biosynthesis is initiated by activation of mGlu5 receptors, is catalyzed by phospholipase C (PLC) and 1,2-diacylglycerol lipase (DGL) activities, and is dependent on intracellular Ca 2ϩ ions.Realtime polymerase chain reaction and immunostaining analyses indicate that DGL-␤ is the predominant DGL isoform expressed in corticostriatal and hippocampal slices and that this enzyme is highly expressed in striatal neurons, where it is colocalized with PLC-␤1.The results suggest that 2-AG is a primary endocannabinoid mediator of mGlu receptor-dependent neuronal plasticity.
The endocannabinoid system serves important functions in the regulation of brain synaptic transmission (for review, see Freund et al., 2003).One prominent example of this function is a form of short-term synaptic plasticity termed depolarization-induced suppression of inhibition (DSI), in which depolarization of a postsynaptic neuron induces the transient suppression of neurotransmitter release from presynaptic nerve terminals impinging on that neuron (for review, see Alger, 2002).Based on electrophysiological and pharmacological studies, it has been proposed that DSI may be mediated through the Ca 2ϩ -dependent formation of an endocannabi-noid messenger, which might be produced postsynaptically and travel across the synaptic space to activate CB 1 receptors on adjacent axon terminals (Alger, 2002;Freund et al., 2003).
To fully understand the roles played by the endocannabinoid system in synaptic plasticity, it is essential to determine both the identity of the endocannabinoid(s) involved and the molecular mechanisms responsible for their production.In the present study, we have used direct biochemical analyses to characterize mGlu receptor-dependent endocannabinoid formation in organotypic cultures of rat corticostriatal and hippocampal slices.We found that activation of mGlu5 receptors, which are expressed at high levels in both striatum and hippocampus (Testa et al., 1994), rapidly stimulates the biosynthesis of 2-AG, but not anandamide, suggesting that 2-AG plays a key role in mGlu5 receptor-initiated signaling events.
Slice Cultures.We prepared organotypic slice cultures from Wistar rat pups, as described previously (Stoppini et al., 1991).In brief, the pups were sacrificed on postnatal day 7 by decapitation after halothane anesthesia.The brains were cut into 0.4-mm coronal slices using a vibratome (Campden Instruments, Leicestershire, UK) in a bath of ice-cold, high-glucose Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA).Corticostriatal or hippocampal slices were placed on Millicell culture inserts (Millipore, Billerica, MA) in six-well plates and covered with basal Eagle's medium with Earle's salts containing heat-inactivated horse serum (25%; v/v), L-glutamine (1 mM), and glucose (0.5%; w/v) supplemented with Earle's balanced salt solution and antibiotics (Invitrogen).Slices were incubated at 37°C with 5% CO 2 for 6 to 7 days before use.
Immunostaining and Western Blotting.We raised a rabbit polyclonal antibody against a glutathione S-transferase-linked peptide comprising residues 661 to 680 in the C-terminal sequence of rat DGL-␤ (ETEFSKILIGPKMLIDHMPD, GI:34870417).The antigen affinity-purified antibody was used at a dilution of 1:2500 for Western blotting and 1:500 for immunostaining, which were conducted as described previously (Jung et al., 2003).We used frozen brain sections prepared from 7 day-old rat pups for all the immunostaining experiments, except Fig. 1A, inset, in which we used corticostriatal slice cultures.Fixed sections were blocked for 4 h at room temperature with 4% normal goat serum in phosphate-buffered saline and subsequently incubated overnight at 4°C with primary antibody under the following conditions: monoclonal anti-neuronal class III ␤-Tubulin (TUJ1, 1:500; Covance, Berkeley, CA); monoclonal anti-V5 (1:5000; Invitrogen); monoclonal anti-PLC␤1 (1:200; Santa Cruz Biotechnology, Santa Cruz, CA) in 4% normal goat serum; and 0.2% Triton X-100 in phosphate-buffered saline.Sections were stained by using Elite ABC (Vector Laboratories, Burlingame, CA) and diaminobenzidine substrate (Vector Laboratories) kits and counterstained with hematoxylin before mounting.Alexa 488-labeled anti-rabbit or Alexa 568-labeled anti-mouse secondary antibodies (1:1000; Invitrogen) were used for fluorescence detection.For confocal microscopy, we used fluorescein isothiocyanate-labeled anti-rabbit or Texas Redlabeled anti-mouse secondary antibodies (1:150; Jackson ImmunoResearch Laboratories, West Grove, PA).Slides were mounted and images were captured using a confocal or fluorescence microscope equipped with a digital camera (Diagnostic Instruments, Sterling Heights, MI).For Western blotting, lysates from corticostriatal slices were prepared in a buffer containing 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.25% Nonidet P-40, and 2 mM EDTA supplemented with a mixture of protease inhibitors (Roche Diagnos-Activation of mGlu5 Receptors Stimulates 2-AG Biosynthesis tics, Indianapolis, IN).The slices were homogenized in lysis buffer and the homogenates were centrifuged at 14,000g for 10 min.Proteins (30 g) were separated on 4 to 15% SDS-polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membranes, and subjected to Western blotting.
Statistical Analyses.Results are expressed as the mean Ϯ S.E.M. of n separate experiments.Statistical significance was evaluated using Student's t test or, when appropriate, one-way analysis of variance followed by Dunnett's test.

Results
Activation of Group I mGlu Receptors Increases 2-AG Levels in Brain Slice Cultures.To measure receptor-dependent changes in brain endocannabinoid levels, we used primary cultures of corticostriatal or hippocampal slices prepared from 7-day-old rat pups (Stoppini et al., 1991).The slice cultures maintained an essentially normal morphology for at least 7 days in vitro, as assessed by light microscopy and immunostaining for the neuron-specific marker class III ␤-Tubulin (Fig. 1A), and contained detectable amounts of 2-AG.A representative HPLC/MS tracing illustrating the presence of a diagnostic ion for this compound (m/z ϭ 401, [MϩNa] ϩ ) in extracts of corticostriatal slices is reported in Fig. 1B. Figure 1B also shows a second HPLC component, which was identified as 1(3)-AG from its retention time and mass-to-charge ratio (m/z ϭ 401, [MϩNa] ϩ ).Because 1(3)-AG arises from the nonenzymatic isomerization of 2-AG, which occurs during sample preparation, we included it in our calculations of total 2-AG levels (Stella et al., 1997).On average, corticostriatal slices contained Ϯ 18.5 pmol of 2-AG per mg protein (n ϭ 44), whereas hippocampal slices contained 440.0 Ϯ 35.2 pmol of 2-AG per mg protein (n ϭ 8).Anandamide was also detectable in slice extracts, although its levels were lower than those of 2-AG [6.4 Ϯ 0.5 and 13.0 Ϯ 0.9 pmol/mg of protein in corticostriatal (n ϭ 44) and hippocampal (n ϭ 8) slices, respectively].
DGL Is Expressed in Slice Cultures.To gather further information on the molecular mechanism of 2-AG production, we examined whether corticostriatal and hippocampal slice cultures express DGL, an intracellular lipid hydrolase that catalyzes the first committed step in neuronal 2-AG biosynthesis (Stella et al., 1997).Two DGL isoforms have been molecularly cloned, DGL-␣ and DGL-␤ (Bisogno et al., 2003).Real-time PCR analyses revealed that both isoforms are present in corticostriatal slices (Fig. 4A), where DGL-␤ mRNA was approximately 150 times more abundant than
We confirmed the high expression of DGL-␤ in slice cultures using an affinity-purified polyclonal antibody, which we raised using a peptide antigen comprising 20 amino acid residues of the rat DGL-␤ C terminus.The antibody recognized a protein with an apparent molecular mass of approx-imately 70 kDa on SDS-polyacrylamide gel electrophoresis, as expected for DGL-␤ (Bisogno et al., 2003; Fig. 5A).The band disappeared after preabsorption with the immunizing peptide (Fig. 5A).To further characterize the specificity of our antibody, we used human embryonic kidney 293 cells that heterologously expressed a modified DGL-␤ containing a V5-His tag fused to the protein's C terminus.Double immunofluorescence staining with anti-DGL-␤ and anti-V5 antibodies followed by confocal imaging showed colocalization of the two signals in cytosol and plasma membrane (Fig. 5B).Preabsorption with the antigen selectively abrogated the DGL-␤ signal, confirming its specificity (Fig. 5B).Additional immunostaining studies revealed the presence of immunoreactive DGL-␤ throughout the striatum, cortex, and hippocampus of 7-day-old pups (Fig. 5C).In particular, anti-DGL-␤ antibody selectively stained neuronal elements in the striatum (Fig. 5D) that were also stained by an antibody that recognizes the 65-kDa isoform of glutamic acid decarboxylase (GAD-65), a marker of GABAergic neurons (data not shown).The DGL-␤ staining was eliminated by preabsorption with DGL-␤ peptide (Fig. 5D) and was absent when the primary antibody was omitted (data not shown).
Finally, because of the postulated role of PLC-␤ in 2-AG formation, we sought to determine whether this enzyme was colocalized with DGL-␤.Double immunofluorescence labeling confirmed the presence of PLC-␤1 in striatum (Hernan-   Jung et al. dez-Lopez et al., 2000) (Fig. 5E) and revealed that expression of this protein largely coincides with that of DGL-␤ (Fig. 5F).

Discussion
The main finding of the present study is that activation of glutamate mGlu5 receptors stimulates 2-AG formation in rat corticostriatal and hippocampal slice cultures.mGlu5 receptors belong to the group I mGlu receptor subfamily, which includes the mGlu1 and mGlu5 subtypes (Conn and Pin, 1997).mGlu5 receptors are highly expressed in the striatum, nucleus accumbens, and hippocampus and more moderately in the neocortex (Testa et al., 1994).Within the striatum, these receptors are localized to GABAegic projection neurons (Testa et al., 1994), where they may contribute to both shortand long-term forms of synaptic plasticity (Gerdeman et al., 2003;Gubellini et al., 2004) as well as to the addictive properties of psychostimulant drugs (Chiamulera et al., 2001).On the other hand, mGlu1 receptors are predominantly expressed in the cerebellum, amygdala, and brainstem (Testa et al., 1994).Thus, the brain distribution of these receptor subtypes tallies well with our results, which show that DHPG-induced 2-AG formation in corticostriatal slices is prevented by the selective mGlu5 receptor antagonist MPEP.Two points are important, however.First, the role of striatal mGlu1 and mGlu5 receptors should not be understood as a simple dichotomy, because mGlu1 receptors are also expressed in the striatum (Kerner et al., 1997) and might be involved in local forms of long-term depression (Gubellini et al., 2001;Sung et al., 2001).Second, in other brain areas mGlu1 receptors participate in endocannabinoid-mediated plasticity (Brown et al., 2003;Galante and Diana, 2004;Azad et al., 2004), suggesting that they might be responsible for the generation of 2-AG in these areas.Thus, it would be interesting to test whether mGlu1 receptor activation triggers 2-AG release in mGlu1-rich structures such as the cerebellum or the amygdala (Azad et al., 2004;Galante and Diana, 2004).
Group I mGlu receptors are linked through G q/11 to the activation of PLC-␤, which catalyzes the hydrolysis of membrane phosphatidylinositol bisphosphate to produce the pleiotropic intracellular second messengers inositol 1,4,5trisphosphate and 1,2-DAG (Conn and Pin, 1997).The effects of 1,2-DAG, which are mostly mediated by protein kinase C activation, are terminated through DAG kinase-mediated phosphorylation of 1,2-DAG to phosphatidic acid (Topham and Prescott, 1999).Cells that express DGL can convert 1,2-DAG to 2-AG and other unsaturated 2-acylglycerols (Prescott and Majerus, 1983).Our experiments indicate that nonselective pharmacological inhibitors of PLC and DGL, which were previously shown to block mGlu-mediated plasticity (Chevaleyre and Castillo, 2003;Galante and Diana, 2004), prevent DHPG-induced 2-AG formation.The results also show that the ␤ isoform of DGL is expressed in striatal neurons, where it colocalizes with the ␤1 isoform of PLC, an essential component of mGlu-and endocannabinoid-dependent plasticity in hippocampal neurons (Hashimotodani et al., 2005).Together, these findings support the view that the PLC-␤-DGL-␤ pathway is responsible for mGlu5-induced 2-AG generation in corticostriatal slice cultures.Future experiments should further test this hypothesis, determining in particular the roles of different PLC-␤ and DGL isoforms in depolarization-and receptor-dependent 2-AG formation.
In the striatum, endocannabinoid-mediated changes in synaptic strength are dependent on intracellular Ca 2ϩ (Gerdeman et al., 2002;Robbe et al., 2002) and mGlu5 receptor activation mobilizes Ca 2ϩ from internal stores (Robbe et al., 2002;Mao and Wang, 2003).Our results, showing Ca 2ϩ chelators block that mGlu5-induced 2-AG biosynthesis, are consistent with these data.There are, however, several forms of endocannabinoid-mediated plasticity that do not require Ca 2ϩ for their expression, such as those triggered by muscarinic acetylcholine receptors in the hippocampus (Maejima et al., 2001;Kim et al., 2002;Chevaleyre and Castillo, 2003;Azad et al., 2004;Galante and Diana, 2004).The identity of the endocannabinoid mediator(s) released by activation of such receptors and the molecular mechanism underlying Ca 2ϩ -independent endocannabinoid release are important questions, which should be addressed in the future.
In conclusion, our experiments provide the first biochemical demonstration that mGlu5 receptors are linked to the biosynthesis of 2-AG, but not anandamide, in the rat brain.These findings should help differentiate the functions served by these two endocannabinoid lipids in synaptic modulation.

TABLE 1 Levels
of anandamide