Carrier‐mediated transport and enzymatic hydrolysis of the endogenous cannabinoid 2‐arachidonylglycerol

The human astrocytoma cell line CCF-STTGI accumulates [3H]2-AG through an Na+- and energy-independent process, with a Km of 0.7 ± 0.1 μM. Non-radioactive 2-AG, anandamide or the anandamide transport inhibitor 4-hydroxyphenyl arachidonamide inhibit [3H]2-AG uptake with half-maximal inhibitory concentrations (IC50) of 5.5 ± 1.0 μM, 4.2 ± 0.3 μM and 1.8 ± 0.1 μM, respectively. A variety of lipid transport substrates and inhibitors interfere with neither [3H]2-AG nor [3H]anandamide uptake. These results suggest that 2-AG and anandamide are internalized in astrocytoma cells through a common carrier-mediated mechanism. After incubation with [3H]2-AG, radioactivity is recovered in phospholipids, monoacylglycerols (unmetabolized [3H]2-AG), free fatty acids ([3H]arachidonate) and, to a minor extent, diacylglycerols and triacylglycerols. Arachidonic acid (100 μM) and triacsin C (10 μM), an acyl-CoA synthetase inhibitor, prevent incorporation of [3H]arachidonic acid in phospholipids and significantly reduce [3H]2-AG transport. Thus, the driving force for 2-AG internalization may derive from the hydrolysis of 2-AG to arachidonate and the subsequent incorporation of this fatty acid into phospholipids.

Anandamide is released from neurons in an activitydependent manner [2,6] and is inactivated by transport into cells followed by intracellular hydrolysis. Anandamide transport is mediated by a high-af®nity, Na -independent carrier and is selectively inhibited by 4-hydroxyphenyl arachidonamide (AM404) [8,9]. Hydrolysis of anandamide to arachidonic acid and ethanolamine is catalyzed by a membrane-bound amidohydrolase enzyme [10±12], which has been puri®ed and molecularly cloned [13,14]. 2-AG may be produced by cleavage of 1,2-diacylglycerol generated by phospholipase C (PLC) acting on phosphatidylinositol bisphosphate [15]. The mechanisms mediating 2-AG inactivation are still only partially understood, but the ability of 2-AG to inhibit [ 3 H]anandamide transport in human astrocytoma cells [16] and to serve as a substrate for anandamide amidohydrolase [17,18] suggest that 2-AG and anandamide may share a common inactivation route.
As a test of this hypothesis, here we have investigated the biological disposition of [ 3 H]2-AG in human astrocytoma cells. Our results suggest that 2-AG and anandamide may be internalized in these cells via a common carriermediated process. By contrast, the intracellular hydrolysis of 2-AG to arachidonic acid and glycerol may be catalyzed by an enzyme activity distinct from anandamide amidohydrolase.
Transport assay: Human astrocytoma CCF-STTG1 cells (American Type Culture Collection, Rockville, MD) were grown in 24-well plates with RPMI 1640 medium containing FBS (10%) and glutamine (1 mM) [16]. Con¯uent cultures were rinsed with Tris±Krebs' buffer at 378C, incubated for 4 min with buffer containing [ 3 H]2-AG (for kinetic analyses: 28±1600 nM; for inhibition assays: 30 nM; 100 mCi/mmol; New England Nuclear, custom synthesized) either at 378C or 0±48C, rinsed with buffer containing bovine serum albumin (BSA, 0.1%), and disrupted by sonication in buffer containing Triton X-100 (1%). Radioactivity in samples was measured by liquid scintillation counting. In some experiments, anandamide transport was measured using [ 3 H]anandamide (30 nM, 220 Ci/mmol, New England Nuclear). For inhibition assays, the cells were preincubated for 10 min with test compounds at appropriate concentrations. The same concentrations of test compounds were also added to the ®nal incubations, which were carried out for an additional 4 min period. Concentrations required for half-maximal inhibition of transport (IC 50 ) were determined by non-linear least square ®tting of the data, using the software GraphPad Prism (GraphPad Software, San Diego, CA).
Statistical analyses: Statistically signi®cant differences were determined by ANOVA, followed by Bonferroni's multiple comparison test.

RESULTS
[ 3 H]2-AG accumulation in astrocytoma cells is rapid and saturable: Lineweaver±Burk analyses of this accumulation yield a Michaelis constant (K m ) of 0.7 AE 0.1 ìM and a maximal accumulation rate (V max ) of 28 AE 6 pmol min À1 mg À1 protein (n 9).
Non-radioactive 2-AG inhibits [ 3 H]2-AG accumulation in a concentration-dependent manner, with a half-maximal inhibitory concentration (IC 50 ) of 5.5 AE 1.0 ìM (n 6). At 100 ìM, non-radioactive 2-AG reduces [ 3 H]2-AG accumulation in cells to 24 AE 1% of control (control: 8381 AE 393 d.p.m./well; 2-AG: 2017 AE 87 d.p.m./well; n 6). We considered this value to be an estimate of the non-speci®c association of [ 3 H]2-AG with the cells, and used it in subsequent experiments as a background value to calculate potency and ef®cacy of test compounds.

DISCUSSION
Human astrocytoma CCF-STTG1 cells internalize [ 3 H]anandamide by a transport mechanism functionally indistinguishable from that found in rat brain neurons and astrocytes [8,16]. Therefore, we used these cells to test the hypothesis that [ 3 H]2-AG may be a substrate for the anandamide transporter. The kinetic of [ 3 H]2-AG accumulation in astrocytoma cells is consistent with a single high-af®nity transport process. In the same cells and under identical conditions, [ 3 H]anandamide transport displays similar characteristics (K m of 0.6 AE 0.1 ìM and V max of 14.7 AE 1.5 pmol min À1 mg À1 protein) [16]. Several lines of evidence indicate that [ 3 H]2-AG accumulation in astrocytoma cells is mediated by a transport system akin to that involved in anandamide internalization. First, non-radioactive anandamide inhibits [ 3 H]2-AG accumulation while, conversely, 2-AG inhibits [ 3 H]anandamide transport (IC 50 of 18.5 AE 0.7 ìM) [16]. Second, the anandamide transport inhibitor AM404 is also effective, and even more potent in interfering with [ 3 H]2-AG uptake than anandamide or 2-AG. Third, a variety of substrates and inhibitors of lipid transport systems have no effect on the accumulation of either [ 3 H]2-AG or [ 3 H]anandamide [8,16]. Finally, anandamide and 2-AG transport are both Na and energy independent [8,16]. There are, however, a number of differences between anandamide and 2-AG inactivation. [ 3 H]Anandamide transport is not affected by arachidonic acid (Fig. 1a) and, vice versa, [ 3 H]arachidonic acid transport is not affected by anandamide (data not shown), indicating that distinct membrane mechanisms mediate the internalization of these two lipid compounds [8,16]. By contrast, arachidonic acid substantially reduces the accumulation of [ 3 H]2-AG (Fig. 1b). A possible interpretation of this ®nding is that arachidonic acid may directly interfere with the putative carrier involved in [ 3 H]2-AG uptake, implying the existence of different transport mechanisms for [ 3 H]2-AG and [ 3 H]anandamide (sensitive and insensitive, respectively, to arachidonic acid). An alternative possibility is that arachidonic acid may inhibit [ 3 H]2-AG accumulation indirectly, for example by interfering with the intracellular disposition of [ 3 H]2-AG. To evaluate these alternative possibilities, we investigated the fate of [ 3 H]2-AG in astrocytoma cells. The routes of 2-AG metabolism in cells are still only partially understood, but two enzymes are thought to be primarily involved: anandamide amidohydrolase and monoacylglycerol lipase [17,18,21]. Both enzymes catalyze the hydrolytic cleavage of 2-AG to arachidonate and glycerol. Free arachidonate is short-lived in cells, being rapidly converted to arachidonyl-coenzyme A (CoA) and then incorporated into phospholipids [22]. Blockade of anandamide amidohydrolase activity completely abrogates [ 3 H]anandamide degradation [8,16], but it does not affect [ 3 H]2-AG hydrolysis, suggesting that monoacylglycerol lipase may be a major player in 2-AG degradation in intact cells [23]. The transport inhibitor AM404 dramatically reduces the intracellular accumulation of [ 3 H]2-AG whereas administration of arachidonic acid has no effect on intracellular [   H]anandamide uptake are insensitive to several substrates and inhibitors of lipid transporters and Na -and energyindependent. By contrast, the intracellular degradation of 2-AG may differ substantially from that of anandamide: whereas anandamide hydrolysis is blocked by BTNP, an inhibitor of amidohydrolase activity, 2-AG hydrolysis is insensitive to this drug. This indicates that 2-AG breakdown in intact cells may depend on other hydrolase enzymes, such as monoacylglycerol lipase [23]. The role of hydrolysis in 2-AG disposition may differ from anandamide in another important way. We found that treating the cells with arachidonic acid or the acyl-CoA synthetase inhibitor, triacsin C, prevents the transport of 2-AG, but not that of anandamide. A plausible interpretation of these results is that 2-AG hydrolysis into free arachidonic acid and the subsequent incorporation of arachidonic acid into membrane phospholipids may be a primary driving force for 2-AG internalization. This difference between the biological dispositions of anandamide and 2-AG might be exploited for the development of selective inactivation inhibitors.