Astrocytes in Culture Produce Anandamide and Other Acylethanolamides*

Anandamide (arachidonylethanolamide) is an endocannabinoid that belongs to the acylethanolamide lipid family. It is produced by neurons in a calcium-dependent manner and acts through cannabinoid CB1 receptors. Other members of the acylethanolamide lipid family are also produced by neurons and act through G-protein-coupled receptors: homo-γ-linolenylethanolamide (HEA) and docosatetraenylethanolamide (DEA) act through CB1 receptors, palmitylethanolamide (PEA) acts through CB2-like receptors, and oleylethanolamide (OEA) acts through receptors that have not yet been cloned. Although it is clear that anandamide and other acylethanolamides play a major role in neuronal signaling, whether astrocytes also produce these lipids is unknown. We developed a chemical ionization gas chromatography/mass spectrometry method that allows femtomole detection and quantification of anandamide and other acylethanolamides. Using this method, we unambiguously detected and quantified anandamide, HEA, DEA, PEA, and OEA in mouse astrocytes in culture. Stimulation of mouse astrocytes with ionomycin, a calcium ionophore, enhanced the production of anandamide, HEA, and DEA, whereas PEA and OEA levels were unchanged. Endothelin-1, a peptide known to act on astrocytes, enhanced the production of anandamide, without affecting the levels of other acylethanolamides. These results show that astrocytes produce anandamide, HEA, and DEA in a calcium-dependent manner and that anandamide biosynthesis can be selectively stimulated under physiologically relevant conditions. The relative levels of acylethanolamides in astrocytes from rat and human were different from the relative levels of acylethanolamides in mouse astrocytes, indicating that the production of these lipids differs between species. Because astrocytes are known to express CB1 receptors and inactivate endocannabinoids, our finding strongly suggests the existence of a functional endocannabinoid signaling system in these cells.

Acylethanolamides (acyl-EAs) 1 are structurally related lipids that contain a saturated or unsaturated fatty acid moiety linked to ethanolamine (1). Several studies have shown that members of the acyl-EA lipid family may act as signaling molecules. For example, the prototypic endocannabinoid anandamide (arachidonylethanolamide, AEA) is produced by neurons in a stimulus-dependent manner and is released toward the extracellular space where it diffuses to activate CB1 receptors (2)(3)(4)(5)(6). Uptake into either neurons or astrocytes with subsequent hydrolysis inactivates AEA (2,7). Other acyl-EAs, including HEA, DEA, PEA, and OEA, are also present in brain and may act as signaling molecules (2,8,9). HEA and DEA act on CB1 receptors with potencies and efficacies similar to that of AEA (10 -12), PEA acts on CB2-like receptors (13,14), whereas OEA acts on a target that does not belong to the cannabinoid receptor family (15)(16)(17).
The molecular mechanism that underlies acyl-EA production is starting to be understood. Cultured neurons analyzed under basal conditions produce low amounts of AEA, but this production is dramatically increased after a stimulus such as a rise in intracellular calcium concentration ([Ca 2ϩ ] i ) (2,5,18). This calcium-dependent increase in production is likely mediated through the calcium-dependent increase in acyltransferase activity, the enzyme that gives rise to the precursor of AEA, arachidonylphosphatidylethanolamide, which is in turn directly cleaved by phospholipase D (1,2,9,8). AEA production is not restricted to neurons, because many different cell types have been shown to produce this lipid (13, 19 -22).
Whether astrocytes, the major glial cell type of the central nervous system, produce AEA was previously unknown, mainly because its presence in these cells has only been assessed with radioactive labeling, a method that has a poor detection limit (2,23). Also, it is unclear if an increase in [Ca 2ϩ ] i enhances the production of all acyl-EAs concomitantly. To address these two questions, we developed a chemical ionization gas chromatography/mass spectrometry (CI-GC/MS) method that allows the simultaneous and femtomole quantification of AEA, HEA, DEA, PEA, and OEA and determined their levels in astrocytes in culture under basal conditions and after ionomycin-induced increases in [Ca 2ϩ ] i . Furthermore, we measured acyl-EA levels in response to endothelin-1, a peptide that has been shown to stimulate phospholipase D in astrocytes (24), and determined if the production of AEA and other acyl-EAs differs between species.
Synthesis and Handling of Unlabeled and 2 H 4 -Labeled Acylethanolamide Standards-Unlabeled and 2 H 4 -labeled acylethanolamide ([ 2 H 4 ]acyl-EA) standards were synthesized as previously described (25,26), diluted in chloroform, and stored for a maximum of 6 months at Ϫ20°C in micro-reaction vials. To accelerate the evaporation of the chloroform in which the standards were diluted and to prevent lipid oxidation, we used N 2 gas and placed the micro-reaction vials on a heating plate at ϳ50°C, a procedure that does not affect the stability of acyl-EA. 2 For optimal GC/MS analysis, unlabeled and [ 2 H 4 ]acyl-EA standards, as well as acyl-EA purified from cells, were reacted with BSTFA for 30 min at room temperature to produce their trimethylsilyl (TMS) derivatives. Evaporating excess BSTFA stopped this reaction. To fully recover the dried acyl-EA TMS derivatives from micro-reaction vials, we added 1 ml of hexane, vortex-mixed the solution for several seconds, and stored the samples at Ϫ20°C for 1 h and up to several days.
GC/MS Analysis-For GC/MS analysis, acyl-EA TMS derivatives were dried, recovered in 2-4 l of hexane, and injected into a Varian CP3800 GC (split less mode) equipped with a 30-m column (fused-silica, CP-Sil 8 CB, low bleed). The capillary injector (model 1079) contained a Silico liner (5.4 mm, OD 3.4 mm) with a carbofrit to improve chromatographic resolution. The injector was temperature-programmed to clean its liner after each injection: i.e. 250°C for 2 min after injection, followed by 6 min at 300°C (with a split ratio of 1:200). Helium was used as a gas carrier (0.5 ml/min). Oven temperature was held at 150°C for 1 min after injection, followed by an increase from 150°C to 300°C at a rate of 20°C/min and held at 300°C for an additional 5.5 min. The transfer line was set at 300°C.
The GC was interfaced to a Varian Saturn 2000 mass spectrometer. Chemical ionization (CI) was performed by ejecting methanol into the trap (230°C) for a maximum reaction time of 40 ms. The emission current was 30 A, and the data were collected at 0.22 s/scan.
Quantification Limit of Isotope Dilution Calibration Curves-The quantification limits of our isotope dilution assay were determined by injecting 200 pmol of each [ 2 H 4 ]acyl-EA into the GC/MS (selected ion monitoring mode) and calculating the signal ratio between unlabeled acyl-EA and [ 2 H 4 ]acyl-EA, thus determining the signal ratio that corresponded to the zero values of each [ 2 H 4 ]acyl-EA (typically 1-2%). This was replicated 10 times to obtain the mean zero value of each [ 2 H 4 ]acyl-EA and its standard deviation. The sum of the mean zero value plus its standard deviation gave the ratio that established the quantification limit of each acyl-EA.
Mouse or Rat Astrocytes in Culture-Astrocytes in cultures were prepared as previously described (27). Briefly, 1-day-old C57BL/6 mice or Fisher 344 rat pups were ice-anesthetized, submerged in 70% ethanol, and decapitated according to the guidelines of the Institutional Animal Care and Use Committee of the University of Washington, Seattle, WA. Meninges were removed from the brains; the neopallia were dissected and enzymatically dissociated (1% trypsin, 0.05% DNase, 2 min). The resulting cells were centrifuged (200 ϫ g, 10 min), suspended in serum-supplemented culture media, and plated into 75cm 2 flasks pre-coated with 1 g/ml poly-L-ornithine (cells from the neopallia of two brains in 10 ml per flask). Serum-supplemented culture media was composed of Dulbecco's modified Eagle's medium supplemented with FBS (10%), HEPES (5 mM), NaHCO 3 (5 mM), penicillin (100 units/ml), and streptomycin (100 g/ml). After 1 h, adherent cells were washed with PBSglc and incubated with serum-supplemented culture media. The meninges were similarly prepared to obtain mixed cultures that contained fibroblasts.
After 3 days in culture, attached cells were rinsed once with phosphate-buffered saline containing high glucose (33 mM) (PBSglc) and reincubated with serum-supplemented culture media (10 ml). After 12-14 days in culture, floating and loosely attached microglial cells were manually shaken off and recovered onto coverslips for immunofluorescent microscopy (see below). Cells that remained attached were reincubated with serum-supplemented culture media, and repopulating microglial cells were removed every week for a total of 5-6 weeks until fewer microglial cells were observed. After 5-6 weeks, attached cells were rinsed once with PBSglc, detached (0.05% trypsin, 0.5 mM EDTA), and centrifuged (200 ϫ g, 5 min). Cells were plated onto either 100-mm culture dishes (3 ϫ 10 5 cells/dish) for GC/MS experiments, 35-mm culture dishes (6 ϫ 10 4 cells/dish) for the determination of protein content or coverslips (1.5 ϫ 10 4 cells/coverslip) for immunostaining. Culture dishes and coverslips were pre-coated with 1 g/ml poly-Lornithine. After 1-2 h, attached cells were rinsed once with PBSglc and incubated with serum-supplemented culture media.
Human Astrocytes in Culture-Human astrocytes were prepared as previously described (29). Mixed cultures were prepared from brains of human fetuses (12-15 weeks of gestation) that were kindly provided by the Birth Defects Research Laboratory of the University of Washington according to its Human Subject Division guidelines. Meninges were removed from the brains; the tissue was dissected into small blocks (Ϸ1 mm 3 ), enzymatically dissociated (0.25% trypsin, 0.05% DNase, 15 min), filtered through a 100-m mesh, and centrifuged (200 ϫ g, 10 min). The resulting cells were suspended in serum-supplemented culture media and plated into 175-cm 2 culture flasks pre-coated with 1 g/ml poly-Lornithine. Thereafter, cells were handled in a manner similar to that for mouse and rat astrocytes. After 6 -8 weeks, more than 95% of the cells expressed glial fibrillary acidic protein (GFAP), i.e. the mouse-anti-GFAP antibody (Roche Molecular Biochemicals, Indianapolis, MN) was visualized with Cy3-conjugated goat-anti-mouse IgG and immunofluorescence microscopy and cells were counted as described below.
Immunofluorescence Microscopy-For immunocytochemical characterization of mouse astrocytes in culture, we used astrocytes and fibroblasts that were serum-deprived for 18 -36 h and microglial cells that were serum-deprived by replacing media with serum-free MEM ϩ Cell-Gro (10%). Cells were rinsed with PBS, fixed in 2% paraformaldehyde for 10 min, and rinsed twice with PBS. Cells were permeabilized with saponin (0.1%) for 30 min and incubated overnight with primary antibodies at 4°C in the presence of goat serum (5%). We used: rabbit-anti-GFAP at 1:200 (Sigma) to label astrocytes, rabbit-anti-fibronectin (Sigma) at 1:200 to label fibroblasts, and rat-anti-MAC1 (Serotec) at 1:100 to label microglia. Bound primary antibodies were detected with fluorescein isothiocyanate-conjugated goat-anti-rabbit IgG (GFAP and fibronectin) at 1:100 or Texas Red-conjugated goat-anti-rat IgG (MAC1) (Jackson ImmunoResearch) at 1:10. Secondary antibodies were incubated for 1 h at room temperature. Stained cells were rinsed with PBS, counterstained with DAPI (1 g/ml) for 10 min, allowed to dry, and mounted on slides with Vectashield. Labeling was visualized with a Nikon optiphot-2 microscope. Photographs were taken with a Nikon 950 digital camera. Images of positive immunostaining versus controls were taken with identical fluorescent power and digital camera setups and processed identically with Adobe Photoshop 6.0.
Cell Type Quantification-To determine the actual quantity of astrocytes, microglial cells and fibroblasts present in culture of mouse astrocytes, we carried out three independent immunocytochemical characterizations of these cultures, made three digital images (20ϫ 2 N. Stella, L. Walter, and A. Witting, personal communication. magnification) of random areas of each immunostaining, and manually counted cell nuclei that were stained by DAPI to determine the total cell number present in the selected area (Ϸ75 cells per image). In the same area, we also manually counted the cells that were stained by anti-GFAP, anti-fibronectin, or anti-MAC1 to determine the number of astrocytes, fibroblasts, and microglia, respectively. Results are expressed as mean Ϯ S.E. (n ϭ 9 counts).
Incubation of Cells and Lipid Extraction-Culture dishes with confluent and serum-deprived astrocytes were placed in a shaking water bath at 37°C. Cells were kept in their culture media (9 ml) and stimulated by adding 1 ml of ionomycin for 5 min or endothelin-1 for 2.5 min (prepared in culture media with or without EGTA). Stimulation was stopped by placing dishes on ice and replacing 5 ml of media with 5 ml of ice-cold methanol. Fixed cells were scraped, and the resulting homogenates were transferred into ice-cold chloroform ( . Organic phases of samples were dried under N 2 , recovered in chloroform (2 ml), and purified by silica gel chromatography (30), eluting the sample with 2 ml of chloroform:methanol (9:1, v/v). Elutes were dried under N 2 , reconstituted in chloroform (100 l), and fractionated by normal-phase HPLC as previously described (26). Fractions eluted at 2.8 -5.0 min were collected into micro-reaction vials, evaporated to dryness, and derivatized with BSTFA for CI-GC/MS analysis. Between experiments, which typically consisted of 10 injections into the HPLC, we cleaned the HPLC column with 100% methanol at 1 ml/min for 10 min to prevent possible contamination and improve the recovery of DEA.
To determine whether cell membranes were intact, we measured lactate dehydrogenase activity release from mouse astrocytes into the incubation media. Lactate dehydrogenase activity was determined with the CytoTox 96 nonradioactive assay (Promega) according to manufacturer's instructions.
Statistics-Data were statistically analyzed using ANOVA followed by Dunnett's post-hoc test (InStat, GraphPad software) with p Ͻ 0.05 considered significant.

CI-GC/MS Analysis of Standard Acylethanolamides and [ 2 H 4 ]
Acylethanolamides-Standard acyl-EA and [ 2 H 4 ]acyl-EA were derivatized with BSTFA and analyzed by CI-GC/MS. The GC conditions specified under "Experimental Procedures" enabled the chromatographic separation of each acyl-EA, with retention times of (in minutes): 9.60 for PEA, 10.65 for OEA, 11.75 for AEA, 11.95 for HEA, and 13.70 for DEA (Fig. 1).
Retention times for [ 2 H 4 ]acyl-EA were ϳ0.03 min shorter than their corresponding acyl-EA, because of GC isotope discrimination (not shown). Thus, this method allowed the concomitant, yet individual, detection and analysis of AEA, HEA, DEA, PEA, and OEA.
The mass spectrum of AEA is shown in Fig. 2a. There were two predominant ions with high diagnostic value: the ion at m/z 420, corresponding to the protonated TMS molecule ([MϩH] ϩ ) and the ion at m/z 330 (base peak), which was produced by the neutral loss of the TMS alcohol ([MϩHϪ90] ϩ ). An additional informative ion was present at m/z 404, which corresponded to the neutral loss of methane ([MϩHϪ16] ϩ ). A minor ionic species at m/z 492 was present, corresponding to the protonated di-TMS molecule (the second TMS being added onto the nitrogen of the ethanolamine).
A similar mass spectrum pattern (i.e. a mass spectrum with predominant ions at [MϩH] ϩ and [MϩHϪ90] ϩ ) was obtained when we analyzed [ 2 H 4 ]AEA (Fig. 2b), HEA (Fig. 3a), Calibration Curves and Quantification Limits-CI does not cause extensive fragmentation of analytes but, rather, produces ions with high m/z that affords excellent structural identification and improved detection limits (31). We took advantage of the improved detection limit to develop an isotope dilution method. Calibration curves were built by injecting a fixed amount of each [ 2 H 4 ]acyl-EA (200 pmol) together with decreasing amounts of each acyl-EA (200, 100, 20, 10, 2, 1, and 0 pmol). When selectively monitoring the base peaks: i.e.
[MϩHϪ90] ϩ for AEA, HEA, and DEA, and [MϩH] ϩ for PEA and OEA, we found that the MS responses of each calibration curve were linear (r 2 Ն 0.98) and that their quantification limits were below 1 pmol (Table I). This result indicates that decreasing amounts of acyl-EA produce a linear decrease in the signals of the area ratio down to the femtomole range, which allows the unambiguous quantification of AEA, HEA, DEA, PEA, and OEA in the low picomole range in a biological matrix such as astrocytes in culture.
Immunocytochemical Characterization of Mouse Astrocytes in Culture-It is well known that cultures of astrocytes contain small amounts of other types of cells, such as microglial cells and fibroblasts (27). Microglial cells are macrophage-like cells that invade the central nervous system during development and are therefore present in the brain parenchyma of the pups used to prepare astrocytes in culture (32). Furthermore, microglial cells proliferate in the presence of serum (33,34). Remnants of meninges contain fibroblasts that also proliferate in the presence of serum (27). We sought to determine the relative amount of microglial cells and fibroblasts in cultures of mouse astrocytes that were grown for 5-6 weeks and serum-deprived. Using an antibody against GFAP, a protein specific to astrocytes, we determined that 94 Ϯ 2% of the cells in the cultures were astrocytes (Fig. 4A). An antibody against the macrophage marker MAC1 (Fig. 4b) stained 2 Ϯ 1% of the cells (Fig. 4d), whereas an antibody against the fibroblast marker fibronectin (Fig. 4c) stained 3 Ϯ 1% of the cells (Fig. 4e). Approximately 2% of the cells were not stained by these antibodies. Thus, the cultures used in this study were highly enriched in astrocytes.  (Table I).
Astrocytes in culture from different species have different phenotypes (35). To determine whether astrocytes from different species vary in their acyl-EA content, we prepared rat and human astrocytes in culture, analyzed their basal levels of AEA and other acyl-EAs, and compared it to that of mouse astrocytes in culture. Under basal conditions, rat astrocytes produced a different pattern of acyl-EA: PEA Ͼ Ͼ AEA Ͼ Ͼ OEA Ϸ HEA Ϸ DEA (Table II). Under basal conditions, human astrocytes in culture produced a pattern of acyl-EA that was reminiscent of the one found in rat astrocytes, with PEA Ͼ Ͼ AEA Ͼ Ͼ DEA Ϸ HEA Ϸ OEA (Table II). Thus, the relative levels of anandamide and other acyl-EAs in astrocytes vary between species.
C6 rat glioma cells had yet a different profile of acyl-EA amounts under basal conditions: PEA Ͼ Ͼ DEA Ϸ AEA Ͼ Ͼ OEA. HEA was not detectable.
Rise in Intracellular Calcium Increases the Production of Anandamide, HEA, and DEA in Mouse Astrocytes-We assessed whether increasing intracellular calcium with the calcium ionophore ionomycin increases the production of AEA and other acyl-EAs. To limit the impact of the variability between cell culture preparations on data interpretation, the effects of ionomycin were compared with the basal (i.e. vehicle-treated) within the same cell culture preparation (5). Ionomycin increased the production of AEA, HEA, and DEA, without affecting the amounts of PEA and OEA. The effect of ionomycin was prevented when calcium was chelated by EGTA (Fig. 5).
Because ionomycin is known to disrupt cell membranes, we treated mouse astrocytes with ionomycin for increasing time periods and assessed the release of lactate dehydrogenase (LDH). Addition of ionomycin to mouse astrocytes for 2.5, 5, or 10 min did not significantly affect the basal release of LDH (Fig. 6). Yet, a 4-to 5-fold increase in LDH activity was measured after 20 min of ionomycin treatment, as well as after 20 min of Triton treatment (Fig. 6). This result shows that cell membranes were intact after 5 min of ionomycin, which is the stimulus time point that we used to study acyl-EA production. . Data are expressed in "pmol per dish" to ascertain that the amount of each acylethanolamide per dish is above the quantification limit of the method presented herein and in "pmol per mg" of protein to allow their comparison with the data obtained with astrocytes from different species (see table 2). Dishes contained 1.43 mg Ϯ 0.11 of protein per dish (n ϭ 16). uli can elicit the changes observed using ionomycin, we measured the production of acyl-EAs in mouse astrocytes in culture that were stimulated with endothelin-1, a peptide that is present in brain and is known to stimulate phospholipase D activity in astrocytes (24). Endothelin-1 significantly increased the production of AEA, without affecting the amounts of HEA, DEA, PEA, and OEA. This effect was prevented when calcium was chelated by EGTA (Fig. 5).

DISCUSSION
The endocannabinoid signaling system is composed of 1) endocannabinoid ligands, with their biosynthesis pathways  a Amounts of acylethanolamide per dish were greater than the quantification limit, except for HEA in C6 cells, which was below quantification limit. and release mechanisms; 2) cannabinoid receptors, with their signal transduction mechanisms; and 3) endocannabinoid inactivation mechanisms, which encompass uptake and subsequent hydrolysis (36). In the central nervous system, each of the above components has been identified in neurons (2,37). Thus, neurons can signal between each other by using the endocannabinoid signaling system without involving other cell types of the central nervous system. It has been shown that astrocytes express CB1 receptors and inactivate endocannabinoids by uptake and subsequent hydrolysis (2,7,38,39). However, their ability to produce endocannabinoids remained unknown. In the present study we demonstrate that cultured astrocytes produce, in a calcium-dependent manner, the acyl-EAs AEA, HEA, and DEA, three endocannabinoids known to act on CB1 receptors with high potency and moderate efficacy (10 -12, 25, 37). Our results underscore the existence of a complete endocannabinoid signaling system in astrocytes, thus their ability to use this system to signal between each other or with surrounding neurons. Many of the mediators in the body, including eicosanoids, are present not as single entities but as large families of structurally related substances that act on different receptors. It seemed reasonable to expect that AEA was only the first representative of a larger family of bioactive lipids that might act on either CB1 receptors or different targets (10,11,25). Whether acyl-EAs are synthesized through a common calciumdependent pathway is unclear (1). Are different acyl-EAs produced by one type of phospholipase D that cleaves different precursors produced by one type of acyltransferase? Here, we show that an increase in [Ca 2ϩ ] i in mouse astrocytes enhances the production of AEA, HEA, and DEA, whereas the amount of PEA and OEA are unchanged. These results are consistent with recent studies showing the independent production of AEA and PEA/OEA in different cell types. Specifically, in mouse epidermal JB6 Pϩ cells in culture, UVB irradiation and serum deprivation increase the production of PEA/OEA, while slightly decreasing the basal amount of AEA (28). In rat cortical neurons in culture, co-activation of N-methyl-D-aspartate (NMDA) receptors and nicotinic receptors increases AEA production, without affecting the basal amounts of PEA/OEA (5). Furthermore, in the present study, we show that C6 glioma cells produce AEA and DEA, whereas HEA is undetectable. Finally, we demonstrate that endothelin-1 selectively increases the production of AEA. Together, these results emphasize the existence of independent biosynthesis pathways for the production of different acyl-EA. Because phospholipase D activity is thought to be independent of calcium and to directly cleave the precursors of acyl-EAs, we hypothesize that multiple isoforms of acyltransferase with distinct calcium-sensitivities exist and that each of these isoforms establishes independent biosynthesis pathways for different acyl-EA.
Astrocytes in culture prepared from different species may vary in their phenotype (35). For example, rat astrocytes in culture express CB1 receptors (40,41), whereas mouse astrocytes in culture do not (42). Here, we quantified the basal levels of acyl-EA in astrocytes from mouse, rat, and humans and found that PEA is abundantly produced in each species. However, although AEA, HEA, and DEA are equally abundant in mouse astrocytes, AEA is ϳ4 -5 times more abundant than HEA and DEA in both rat and human astrocytes. Although the significance of these differences is not known, our finding does suggest that AEA may be a predominant endocannabinoid in rat and human astrocytes.
What is the role of the endocannabinoid signaling system in astrocytes? The majority of the perivascular astrocytes in adult rat brain express CB1 receptors, with more than 70% of these receptors being localized on the plasmalemma of either their cell body or their filamentous glial processes (38). Activation of CB1 receptors on rat astrocytes in culture increases the rate of glucose oxidation and ketogenesis, two mechanisms involved in the energy supply of the brain (43,44). Because perivascular astrocytes are pivotally located and involved in supplying energy from blood to neurons in a stimulus-dependent manner (45), one hypothesis is that the endocannabinoid signaling system in astrocytes regulates the energy supply from blood to neurons. In agreement with this hypothesis, rat brain energetic metabolism is increased after exposure to AEA or ⌬ 9 -tetrahydrocannabinol (46).
It has been shown that AEA produces biological effects through CB1 receptors as well as through receptors that are distinct from CB1 receptor, i.e. "anandamide" receptors. Specifically, AEA inhibits gap junctions and intercellular calcium signaling between cultured mouse striatal astrocytes, an effect that is not prevented by CB1 receptor antagonists (47). It is thus possible that astrocytes produce AEA in a calcium-dependent manner to modulate energy metabolism by acting through CB1 receptors and regulate intercellular calcium signaling by acting through anandamide receptors.
Rat glioma cells express CB1 and CB2 receptors, whereas rat astrocytes only express CB1 receptor (40,48). The kinetic of AEA uptake into C6 cells is different from that measured with rat astrocytes in culture (7,49). Here, we show that C6 cells have a different pattern of acyl-EA amounts compared with rat astrocytes in culture. Together, these results suggest that transformation of astrocytes into glioma cells is likely associated with a profound change in their endocannabinoid signaling system.
In summary, our study shows that astrocytes in culture produce AEA and other acyl-EAs under basal conditions. Production of AEA, HEA, and DEA is increased by a calcium stimulus, whereas the levels of PEA and OEA are unaffected. Endothelin-1 selectively increases the production of AEA. This indicates that the production of different acyl-EAs can be independently increased and that independent biosynthesis pathways are involved in the production of different acyl-EAs. Finally, the acyl-EA production in astrocytes varies between species.