Lipid Rafts Control Signaling of Type-1 Cannabinoid Receptors in Neuronal Cells

Several G protein-coupled receptors function within lipid rafts plasma membrane microdomains, which may be important in limiting signal transduction. Here we show that treatment of rat C6 glioma cells with the raft disruptor methyl-β-cyclodextrin (MCD) doubles the binding efficiency (i.e. the ratio between maximum binding and dissociation constant) of type-1 cannabinoid receptors (CB1R), which belong to the rhodopsin family of G protein-coupled receptors. In parallel, activation of CB1R by the endogenous agonist anandamide (AEA) leads to ∼3-fold higher [35S]GTPγS binding in MCD-treated cells than in controls, and CB1R-dependent signaling via adenylate cyclase, and p42/p44 MAPK is almost doubled by MCD. Unlike CB1R, the other AEA-binding receptor TRPV1, the AEA synthetase NAPE-PLD, and the AEA hydrolase FAAH are not modulated by MCD, whereas the activity of the AEA membrane transporter (AMT) is reduced to ∼50% of the controls. We also show that MCD reduces dose-dependently AEA-induced apoptosis in C6 cells but not in human CHP100 neuroblastoma cells, which mirror the endocannabinoid system of C6 cells but are devoid of CB1R. MCD reduces also cytochrome c release from mitochondria of C6 cells, and this effect is CB1R-dependent and partly mediated by activation of p42/p44 MAPK. Altogether, the present data suggest that lipid rafts control CB1R binding and signaling, and that CB1R activation underlies the protective effect of MCD against apoptosis.

One activity of AEA that has attracted growing interest is its ability to induce programmed cell death (apoptosis) in neuronal and peripheral cells (9), with therapeutic potential in cancer (10) and neurodegenerative diseases (11)(12)(13)(14). In a previous study we have shown that the pro-apoptotic activity of AEA occurs through activation of type 1 vanilloid receptors (now called transient receptor potential channel vanilloid receptor subunit 1, TRPV1), which are six transmembrane-spanning proteins with intracellular N and C termini (15). In fact, we have shown that AEA is a physiological agonist of TRPV1 (16), and thus can be also considered a true endovanilloid (17). Instead, activation of CB1R protects cells against AEA-induced apoptosis, suggesting that vanilloid and cannabinoid receptors regulate in opposite ways the apoptotic potential of AEA (18 -20).
In the last few years the concept of a protective role of CB receptors against (endo)cannabinoid-induced apoptosis has been extended to several cell types, from neuronal and immune cells to endothelial cells and astrocytes (reviewed in Refs. 9 and 10). Recently, two reports from Maruyama's group have shown that methyl-␤-cyclodextrin (MCD) blocks apoptosis induced in vitro by AEA in glioma cells (21) and hepatocytes (22). MCD is a membrane cholesterol depletor (23), and is widely used to disrupt the integrity of lipid rafts, i.e. subdomains of the plasma membrane that contain high concentrations of cholesterol and glycosphingolipids (24 -26). Thus, a novel role for membrane lipid rafts in AEA-induced apoptosis was suggested (21,22). In this line, there is growing evidence that G protein-coupled receptors (GPCR) may concentrate in lipid rafts (24 -26).
CB1R belongs to the rhodopsin family of GPCR (27,28), therefore we decided to investigate the effect of MCD on the proteins of the endocannabinoid system, and the implications for AEA-induced apoptosis. We have chosen rat C6 glioma cells, because they have a well characterized endocannabinoid system (18,21,29), and undergo AEA-induced apoptosis mediated by TRPV1 and counteracted by CB1R (18,19). We extended the study to human CHP100 neuroblastoma cells, which have the same ability as C6 cells to metabolize AEA, but are devoid of CB1R and hence are more sensitive to the pro-apoptotic activity of AEA (18). We did not further extend this study to 2-AG and the enzymes that degrade (30) and synthesize it (31), because 2-AG does not have pro-apoptotic activity toward C6 cells or CHP100 cells (18,21), in keeping with the observation that it does not activate TRPV1 receptors (17), or at least has a much lower potency than AEA (32 . Adenosine 5Ј-[␥-32 P]triphosphate (3000 Ci/mmol) was purchased from Amersham Biosciences, and N-piperidino-5-(4chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-3-pyrazole carboxamide (SR141716) was a kind gift of Sanofi-Synthelabo Recherche (Montpellier, France). 6-Dodecanoyl-2-dimethylamino-naphthalene (laurdan) was from Molecular Probes (Eugene, OR). Rabbit anti-CB1R polyclonal antibodies were from Cayman Chemicals (Ann Arbor, MI), and mouse anti-cytochrome c monoclonal antibodies (clone 7H8.2C12) were from BD PharMingen (San Diego, CA). Goat anti-rabbit or anti-mouse alkaline phosphatase conjugates (GAR-AP or GAM-AP, respectively) were purchased from Bio-Rad (33).
Cell Culture and Treatment and Determination of Apoptosis-Rat C6 glioma cells were cultured in Ham's F-12 medium (Flow Laboratories Ltd., Herts, UK), supplemented with 10% fetal calf serum as described (18). Cells were maintained at 37°C in a humidified atmosphere with 5% CO 2 and were fed every 3-4 days. Adherent human CHP100 neuroblastoma cells were cultured as previously reported (18) in a 1:1 mixture of MEM (Eagle's minimal essential medium plus Earle's salts) and Ham's F-12 media (Flow Laboratories Ltd.), supplemented with 15% heat-inactivated fetal bovine serum, sodium bicarbonate (1.2 g/liter), 15 mM Hepes buffer, 2 mM L-glutamine, and 1% non-essential aminoacids. CHP100 cells were cultured at 37°C in a 5% CO 2 atmosphere, and were trypsinized (0.05% trypsin, 0.02% EDTA) before treatment. Cholesterol depletion was performed by preincubating C6 cells or CHP100 cells for 30 min at 37°C with the indicated amounts of MCD, which removes cholesterol from the plasma membranes (21,24,34,35). After MCD pretreatment, cells were washed in phosphate-buffered saline (PBS) and then were treated with AEA (or vehicle in the controls) as detailed below. In some experiments, cholesterol repletion was achieved by treating C6 cells with 5 mM MCD, washing in PBS, and then incubating for 15 min at 37°C with 5 mM cholesterol-MCD complexes (36). Synthesis of the cholesterol-MCD complexes was carried out as reported (37). Cell viability was assessed by Trypan blue dye exclusion.
Apoptosis was estimated 48 h after treatment with AEA (or vehicle in the controls) (18), by using the cell death detection ELISA kit (Roche Applied Science), based on the evaluation of DNA fragmentation by an immunoassay for histone-associated DNA fragments in the cell cytoplasm (18). This method has been recently validated for C6 cells and CHP100 cells by comparison with cytofluorimetric analysis, performed in a FACScalibur Flow Cytometer (BD Biosciences) (18). This latter technique quantifies apoptotic bodies formation in dead cells by staining with propidium iodide (50 g/ml). Both C6 and CHP100 control cells contained less than 4.0 Ϯ 1.0 apoptotic bodies every 100 cells analyzed (18).
Cholesterol Quantitation and Analysis of Cell Membrane Fluidity-Membrane lipids were extracted from C6 cells or CHP100 cells (5 ϫ 10 6 per test) and cholesterol content was measured by means of cholesterol oxidase (kit from Roche Applied Science), as previously described (38). Membrane fluidity of the cells was determined by means of the fluorescent probe laurdan, as already described (38). Membrane fluidity is inversely proportional to the ratio of laurdan fluorescence at 440 nm versus that at 490 nm (F 440 /F 490 ): the higher the ratio, the lower the fluidity (39).
Receptor Binding Assays and AEA-stimulated [ 35 S]GTP␥S Binding-Membrane fractions were prepared from C6 cells (200 ϫ 10 6 test), that had been previously treated with MCD as described above, and were used to perform direct binding experiments of [ 3 H]CP55.940 by rapid filtration assays (18). Binding data were elaborated through nonlinear regression analysis, using the Prism 3 program ( (18,40). Western blot analysis of CB1R in the same membrane preparations (50 g/lane) used in binding assays was performed by 12% SDS-PAGE under reducing conditions, followed by electroblotting onto 0.45 m nitrocellulose filters (Bio-Rad) (18). Rabbit anti-CB1R antibodies (diluted 1:250) were used as first antibody, and goat anti-rabbit immunoglobulins conjugated with alkaline phosphatase (GAR-AP) as second antibody (at 1:2000 dilution) (18).
AEA-stimulated [ 35 S]GTP␥S binding was determined on membrane fractions (18) essentially as described (41). Membrane preparations were pre-incubated in 4 milliunits/ml adenosine deaminase (183 units/mg of protein, Sigma) for 10 min at 30°C. The binding of Determination of AEA Uptake, Synthesis, and Hydrolysis-The activity of the AMT in C6 cells was measured as described (42). Cells (2 ϫ 10 6 /test) were incubated for 15 min, at 37°C or 4°C, with 400 nM [ 3 H]AEA, then they were washed three times in 2 ml of PBS containing 1% bovine serum albumin and were finally resuspended in 200 l of PBS. Membrane lipids were then extracted (42), resuspended in 0.5 ml of methanol, mixed with 3.5 ml Sigma-Fluor liquid scintillation mixture for non-aqueous samples (Sigma), and radioactivity was measured in a LKB1214 Rackbeta scintillation counter (Amersham Biosciences). To discriminate non-carrier-mediated from carrier-mediated transport of AEA through cell membranes, [ 3 H]AEA uptake at 4°C was subtracted from that at 37°C (42). Michaelis-Menten constant (K m ) and maximum velocity (V max ) of the uptake of [ 3 H]AEA (0 -800 nM range) by AMT in C6 cells were determined by nonlinear regression analysis as reported (42). AMT activity was expressed as pmol of AEA taken up per min per mg of protein.  (18). FAAH activity was expressed as pmol arachidonate released per min per mg of protein.
Other Enzymatic Assays-C6 cells (5 ϫ 10 6 /test) were incubated for 15 min at 37°C with AEA and related compounds, then they were washed, homogenized and subjected to enzymatic assays. Forskolin (1 M)-stimulated adenylate cyclase (EC 4.6.1.1; AC) activity was determined according to the amount of cAMP (43), detected in cell extracts with the cAMP enzyme immunoassay kit, as detailed below. AC activity was expressed as pmol cAMP per min per mg of protein. The activity of p42/p44 MAPK (EC 2.7.1.37) was assayed in cell extracts by the phosphorylation of MAPK-specific peptide substrate at 30°C with adenosine 5Ј-[␥-32 P]triphosphate (44), using the Biotrak MAPK enzyme assay system (Amersham Biosciences). MAPK activity was expressed as pmol of phosphate per min per mg of protein. The effect of pertussis toxin (PTX) on enzymatic activities was determined by preincubating C6 cells for 3 h at 37°C with 5 g/ml PTX before addition of AEA, or vehicle in control experiments (43,44).
Determination of cAMP and Cytochrome c Release-C6 cells (5 ϫ 10 6 cells/test) were treated with 1 M forskolin in the presence of different compounds (or vehicle alone in the controls) for 15 min, then medium was discarded and trypsinized cells were homogenized as described (42). Cyclic AMP levels in acetylated cell extracts were determined by the Cayman Chemical cAMP Enzyme Immunoassay kit (Alexis Corporation, Lä ufelfingen, Switzerland), as reported (42). Cyclic AMP in the extracts was within the linearity range of the method, calibrated with acetylated cAMP as suggested by the manufacturer.
The amount of cytochrome c released into the cytosol of C6 cells or CHP100 cells (25 ϫ 10 6 cells/test) 8 h after treatment with AEA (18) was quantified by ELISA, previously validated for both cell lines by Western blot analysis (18). Cell extracts (25 g/well) were reacted with anticytochrome c monoclonal antibodies, diluted 1:250, and GAM-AP as second antibody at 1:2000 dilution. Color development of the alkaline phosphatase reaction was recorded at 405 nm, using p-nitrophenylphosphate as substrate (18). The absorbance values of the unknown samples were within the linearity range of the ELISA test, assessed by calibration curves with known amounts of cytochrome c (in the range 0 -500 ng/well).
Statistical Analysis-Data reported in this study are the means Ϯ S.D. of at least three independent experiments, each performed in duplicate. Statistical analysis was performed by the nonparametric Mann-Whitney U test, elaborating experimental data by means of the InStat 3 program (GraphPAD Software for Science).

Effect of MCD on CB1 Receptors and TRPV1 Receptors-
Treatment of rat C6 glioma cells with MCD caused a dose-dependent increase in the binding of [ 3 H]CP55.940, a CB1R and CB2R agonist (45). C6 cells express type-1 CB receptors (18,21), and consistently [ 3 H]CP55.940 binding was fully displaced by 0.1 M SR141716 (Fig. 1A), a selective CB1R antagonist (45). The increase in CB1R binding reached statistical significance at 1 mM MCD, and a maximum of ϳ250% of the control value at a dose of 2.5 mM (Fig. 1A). Therefore, the latter concentration was used to further investigate the effect of MCD. Nonlinear regression analysis of saturation curves like that shown in Fig.  1B (Fig. 1C) or MCD-treated cells (Fig. 1D), that yielded B max  (Table I). Incidentally, B max and K d values of CB1R in C6 cells are close to those reported for the same receptor from other sources (45). In addition, the saturation binding data of [ 3 H]CP55.940 were not better fitted by a two-site-binding model (not shown).
CB1R can undergo rapid (within minutes) internalization and recycling into the plasma membrane (46,47). In order to ascertain whether MCD could alter CB1 receptors recruitment in plasma membranes of C6 cells, Western blot analysis was performed. The results demonstrated that membrane levels of CB1R were identical in C6 cells untreated or treated with 2.5 mM MCD (Fig. 1B, inset), ruling out that MCD could increase the amount of CB1R present on the cell surface. Additionally, C6 cells that had been pretreated with 5 mM MCD and then with 5 mM cholesterol-MCD complexes showed the same binding ability as untreated controls (Fig. 1A), suggesting that the effect of MCD was reversible.
On the other hand binding of [ 3 H]RTX, a selective TRPV1 agonist (16,40), to C6 membranes was not affected by MCD at any dose ( Fig. 2A). Yet, [ 3 H]RTX binding was fully prevented by 1 M CPZ (Fig. 2A), a selective TRPV1 antagonist (16,40). In keeping with these data, saturation curves of [ 3 H]RTX binding (Fig. 2B) were not significantly affected by MCD and yielded B max and K d values almost superimposable on the controls (Table I), which are typical of TRPV1 receptors (40,48).
Effect of MCD on CB1-dependent Signaling Pathways-The main signaling pathways triggered by agonist binding to CB1 receptors include inhibition of adenylate cyclase (AC) and stimulation of MAPK, both mediated by G i/o proteins (for reviews see Refs. 27 and 28). Here, pretreatment of C6 cells with MCD was found to enhance dose-dependently the binding of [ 35 S]GTP␥S stimulated by 1 M AEA, an effect fully prevented by 0.1 M SR141716 (Fig. 3A). Dose dependence of MCD-induced increase of agonist-stimulated [ 35 Fig. 1A and Fig. 3A). In addition, dose response curves demonstrated that [ 35 S]GTP␥S binding in C6 cells pretreated with 2.5 mM MCD was enhanced dose-dependently by the agonist, reaching a maximum at the saturating concentration of 1 M (Fig. 3B).
Pretreatment with 2.5 mM MCD did not affect the basal activity of AC nor that of MAPK, but significantly potentiated the effect of 1 M AEA on these enzymes, inhibiting AC down to ϳ20% of the controls, and increasing MAPK up to ϳ500% (Fig.  4, A and B). The activity of AC and MAPK in AEA-treated cells not previously incubated with MCD was ϳ40% and ϳ300% of the basal levels, respectively (Fig. 4, A and B). Additionally, the effects of 1 M AEA on AC and MAPK were abolished by 0.1 M SR141716, or by 5 g/ml PTX, an inhibitor of G i/o proteins (43,44) (Fig. 4, A and B). Furthermore, dose response curves demonstrated that both AC activity and MAPK activity of C6 cells pretreated with 2.5 mM MCD were affected dose-dependently by agonist concentration, reaching a maximum at the saturating concentration of 1 M (Fig. 4, C and D). Taken together, these findings suggest that the increased binding to CB1R in MCD-treated cells led to enhanced agonist-induced CB1R signaling through G i/o proteins.
Effect of MCD on AEA Synthesis, Uptake, and Degradation-We further investigated the effect of MCD on the other  proteins of the endocannabinoid system in C6 cells. MCD did not affect the activity of NAPE-PLD (Fig. 5A), which is the checkpoint in AEA synthesis (8), whereas it reduced the activity of the AMT (Fig. 5B) without affecting that of the AEAhydrolase FAAH (Fig. 5C). As expected, the activity of AMT and that of FAAH were fully inhibited by the corresponding selective inhibitors VDM11 (10 M) and MAFP (100 nM) (16). The lack of specific inhibitors of NAPE-PLD did not allow to further extend the inhibition experiments. In addition, kinetic analysis of AMT in control C6 cells (not shown) yielded K m and V max values of 160 Ϯ 18 nM and 45 Ϯ 5 pmol/min per mg of protein respectively, in keeping with a previous report (18). Interestingly, in C6 cells treated with 2.5 mM MCD the values of K m and V max of AMT were 140 Ϯ 20 nM and 25 Ϯ 3 pmol/min per mg of protein, respectively. Effect of MCD on Membrane Properties-In an attempt to clarify the mechanism of MCD effect, we measured the cholesterol content and membrane fluidity of C6 cells. MCD is a cholesterol depletor (23), and indeed treatment with this compound led to a dose-dependent decrease in membrane cholesterol content (Table II). This effect was paralleled by a dose-dependent decrease in the fluorescence ratio of laurdan (Table  II), which is an index of membrane fluidity: the higher the ratio, the lower the fluidity (39). At the concentration of 2.5 mM, MCD reduced cholesterol content and fluorescence ratio of C6 cell membranes to 34 and 60% of the control values, indicating that membranes of MCD-treated cells were more fluid than those of controls. In addition, C6 cells that had been pretreated with 5 mM MCD and then with 5 mM cholesterol-MCD complexes showed almost the same cholesterol content and fluorescence ratio as untreated controls (Table II), demonstrating that the effect of MCD was reversible. Much the same was observed with the binding ability of C6 cells (Fig. 1A).
These findings extend previous reports showing that cholesterol-MCD complexes can fully reverse the effects of MCD (36,37). Furthermore, human CHP100 neuroblastoma cells showed the same basal level of cholesterol (3.20 Ϯ 0.30 nmol/10 6 cells) and fluorescence ratio (2.00 Ϯ 0.25) as C6 cells, and both parameters were changed by MCD much alike to those shown in Table II for C6 cells. Incidentally, the extent of cholesterol depletion in C6 cell plasma membranes after MCD treatment is in keeping with a recent report on cardiac myocytes (34), while the values of cholesterol content and laurdan fluorescence of CHP100 cells corroborate a previous study (38).
Effect of MCD on AEA-induced Apoptosis in C6 Glioma Cells and in CHP100 Neuroblastoma Cells-CHP100 cells have the same ability as C6 cells to metabolize AEA and to bind it via TRPV1, yet they are devoid of CB1 receptors (18). In a previous study we have shown that AEA induces apoptosis in both cell types by binding to TRPV1 receptors, and that the lack of CB1R makes CHP100 cells much more sensitive than C6 cells to this pro-apoptotic agent (18). Indeed, 10 M AEA induced ϳ14% apoptotic bodies formation in C6 cells after 48 h, whereas it caused ϳ25% apoptotic bodies formation in CHP100 cells at a concentration of 1 M (18). These findings were confirmed in this study (compare Fig. 6, A and B).
Interestingly, pretreatment of C6 cells with MCD caused a dose-dependent decrease in apoptotic bodies formation, down to a minimum of ϳ25% of the controls at 2.5 mM (Fig. 6A). Also co-incubation with 10 M CPZ, a selective TRPV1 antagonist (16,40), reduced apoptosis of C6 cells (ϳ40% of controls), and this effect of CPZ was additive to that of MCD (Fig. 6A). Conversely, co-incubation with 1 M SR141716 almost doubled the pro-apoptotic activity of AEA in controls, as already shown (18), and prevented the anti-apoptotic effect of MCD at all doses used (Fig. 6A). Unlike C6 cells, MCD had no effect on AEA-induced apoptosis of CHP100 cells under the same experimental conditions, whereas 10 M CPZ was as effective in reducing apoptotic bodies formation in these cells as it was in C6 cells (compare Fig. 6, A and B). In addition, 1 M SR141716 failed to prevent apoptosis in CHP100 cells (Fig. 6B).
Previous studies have shown that the anti-apoptotic effect of CB1R activation in human and rat astrocytes involves activation of phosphatidylinositol 3-kinase (PI3K) and of protein kinase B (PKB), followed by PI3K/PKB-dependent stimulation of MAPK (49,50). These observations apply also to C6 glioma cells, where the PI3K inhibitor wortmannin (49) and the MAPK inhibitor PD98059 (44), used at 20 M and 200 nM respectively (49,50), significantly enhanced apoptosis induced by 10 M AEA (Table III). Interestingly, the pro-apoptotic effect of wortmannin and that of PD98059 were not additive (Table III), strengthening the hypothesis that they act along the same signaling pathway (50). In keeping with these data, 20 M PD98059 abolished the anti-apoptotic effect of MCD pretreatment, resembling the effect of 1 M SR141716 (Fig. 6A).
Unlike C6 cells, wortmannin and PD98059 were ineffective on programmed death of CHP100 cells induced by 1 M AEA  (Table III), and 20 M PD98059 did not modify apoptosis of these cells pretreated with MCD (Fig. 6B). Overall, the antiapoptotic effect of MCD seems to depend on the activation of CB1R, which triggers the PI3K/PKB-dependent MAPK activation. Therefore, the shutdown of CB1R or of MAPK minimizes the anti-apoptotic effect of MCD treatment.
We have shown previously that a critical event along the chain leading to AEA-induced apoptosis of C6 cells and CHP100 cells via TRPV1 receptors is the disruption of mitochondrial integrity followed by the release of cytochrome c (18). These events have been shown to underlie the AEA-induced apoptosis of different cell types (20,(51)(52)(53), and are in common to the apoptotic pathway triggered by the TRPV1 agonist capsaicin (54). In keeping with our previous report (18), 1 M AEA was found to induce a ϳ2-fold higher cytochrome c release in CHP100 cells than in C6 cells at a dose of 10 M (Fig. 6C). Remarkably, MCD was shown to block dose-dependently the AEA-induced cytochrome c release in C6 cells, while it was ineffective in CHP100 cells under the same experimental conditions (Fig. 6C). In addition, the effects of 1 M SR141716, 10 M CPZ or 20 M PD98059 on AEA-induced cytochrome c release from mitochondria of C6 cells paralleled those on apoptosis (compare Fig. 6, A and D). DISCUSSION In this investigation we report new evidence that cholesterol depletion by MCD enhances the binding of CB1 receptors, and subsequent G protein-dependent signaling through AC and MAPK. We also show that the effect of MCD is reversible, and we demonstrate that MCD treatment protects CB1Rexpressing C6 cells against AEA-induced apoptosis. In fact, enhanced CB1R signaling underlies the protective effect of MCD via increased MAPK activity and decreased release of mitochondrial cytochrome c. Furthermore, MCD treatment reduces AMT activity without affecting other proteins of the endocannabinoid system.
The observation that cholesterol depletion is accompanied by a 2-fold enhancement of the binding efficiency of CB1R in C6 cells seems interesting, because recent evidence suggests that a specific groove in the transmembrane helix 6 enables CB1R to recognize AEA and related (endo)cannabinoids within the lipid bilayer, rather than at the cellular surface (55). Therefore, it is conceivable that the change in membrane fluidity due to cholesterol depletion (Table II)  ulate: (i) the development of Alzheimer's disease (56), (ii) the toxicity of immunodeficiency virus type 1 (HIV-1) glycoprotein gp120 (38), and (iii) the plasticity and efficiency of synapses (57). Interestingly, endocannabinoids have been shown to be involved in all these processes through CB1R-dependent mechanisms (2,9). In the same line, it is noteworthy that cholesterol depletion by MCD pretreatment under the same conditions used here has been recently shown to increase (up to 9-fold) hyaluronan binding to U87 glioma cells (35).
Unlike CB1R, TRPV1 receptors were not modulated by cholesterol depletion, an observation which extends recent thermodynamic data in TRPV1-transfected HEK cells (58). Of interest is also the fact that MCD pretreatment reduced AMT activity (Fig. 5B), in keeping with a report, which appeared during the preparation of this manuscript (59). The observation that raft disruption reduced V max without affecting K m of AMT of C6 cells seems to favor the endocytic hypothesis of McFarland et al. (59), who proposed that the AEA transporter may be internalized upon MCD treatment, thus reducing the number (and hence the V max ) of transporter molecules on the cell surface. However, the lack of anti-AMT antibodies did not allow the quantification of AMT protein in plasma membranes. At any rate, it seems noteworthy that inhibition of AMT may contribute to enhance AEA activity at CB1 receptors, since reduced intracellular transport of AEA should increase its concentration at the extracellular binding site of CB1R.
There is a growing body of evidence that GPCRs, like CB1R, function in the context of lipid rafts, and that integrity of these compartments may be important in limiting signaling (26,60,61). Indeed, AEA-induced stimulation of the binding of GTP␥S to C6 cells was enhanced dose-dependently by MCD, more (ϳ330%) than the CP55.940 binding (ϳ250%) under the same conditions (compare Figs. 1A and 3A). Therefore, it can be suggested that raft disruption may improve not only ligand binding to CB1R but also the subsequent interaction of activated CB1R with G proteins. This effect could be explained also with a faster motion of proteins within a more fluid lipid bilayer. In keeping with this observation, G protein-dependent signaling of CB1R via AC and MAPK activity was in turn enhanced by cholesterol depletion (Fig. 4, A and B).
There is also converging evidence that activation of CB1R protects neurons, astrocytes, and many peripheral cells against apoptosis (reviewed in Refs. 9 and 10). Interestingly, two recent reports from the same group have shown that MCD stops apoptosis induced in vitro by AEA in glioma cells (21) and hepatocytes (22). Our results suggest that CB1R activation can fully account for the protective effect of MCD against apoptosis induced by AEA via TRPV1 receptors. This conclusion is supported by the following lines of evidence: (i) the CB1R antagonist SR141716 counteracted the MCD protection in CB1Rexpressing C6 cells, and almost doubled apoptotic bodies formation independently of the dose of MCD (Fig. 6A); (ii) MCD or SR141716 had no effect on AEA-induced apoptosis of CHP100 cells, which have the same endocannabinoid system of C6 cells but CB1R (Fig. 6B); (iii) MCD protected mitochondria from AEA-induced cytochrome c release in C6 cells, while it was ineffective on mitochondria of CHP100 cells (Fig. 6C).
Of further interest is the finding that CB1R-dependent protection depends on the activation of MAPK, which is downstream the PI3K/PKB pathway (49,50,62). In turn, this involves activation of downstream targets like p90 ribosomal S6 kinase (RSK), Bad, and IkB kinase, which finally stops the execution of the apoptotic program by caspase 9 (50). It seems noteworthy that also a lower cytochrome c release from mitochondria induces a lower caspase activity (63,64), overall suggesting that two pathways triggered by CB1R (i.e. MAPK activation and cytochrome c release reduction) curb the proapoptotic potential of AEA by limiting the caspase activation. In the same line, depletion of cholesterol in lipid rafts has been shown to inhibit anti-tumor drug-induced apoptosis (65). This raises the suggestive hypothesis that lipid rafts alteration might be a way to modulate CB1R-dependent signaling, thus controlling cell survival within the central nervous system. This concept might be exploited for the treatment of endocannabinoid-related diseases such as cancer (10), brain injury (12), and neurodegenerative disorders (13).
In conclusion, this investigation demonstrates that lipid rafts can modulate binding and signaling of CB1R, which are enhanced by cholesterol depletion. In addition, it gives a biochemical background to the anti-apoptotic effect of MCD, showing that it depends on the protection exerted by activated CB1R through stimulation of MAPK and reduction of cytochrome c release from mitochondria. Finally, it is noteworthy that