The uptake by cells of 2-arachidonoylglycerol, an endogenous agonist of cannabinoid receptors


V. Di Marzo, Istituto per la Chimica di Molecole di Interesse Biologico, Via Toiano 6, 80072, Arco Felice (NA), Italy. Fax: + 39 081 804 1770, Tel.: + 39 081 853 4156, E–mail:


It is not yet clear if the endocannabinoid 2-arachidonoylglycerol (2-AG) is transported into cells through the same membrane transporter mediating the uptake of the other endogenous cannabinoid, anandamide (N-arachidonoylethanolamine, AEA), and whether this process (a) is regulated by cells and (b) limits 2-AG pharmacological actions. We have studied simultaneously the facilitated transport of [14C]AEA and [3H]2-AG into rat C6 glioma cells and found uptake mechanisms with different efficacies but similar affinities for the two compounds (Km 11.0 ± 2.0 and 15.3 ± 3.1 µm, Bmax 1.70 ± 0.30 and 0.24 ± 0.04 nmol·min−1·mg protein−1, respectively). Despite these similar Km values, 2-AG inhibits [14C]AEA uptake by cells at concentrations (Ki = 30.1 ± 3.9 µm) significantly higher than those required to either 2-AG or AEA to inhibit [3H]2-AG uptake (Ki = 18.9 ± 1.8 and 20.5 ± 3.2 µm, respectively). Furthermore: (a) if C6 cells are incubated simultaneously with identical concentrations of [14C]AEA and [3H]2-AG, only the uptake of the latter compound is significantly decreased as compared to that observed with [3H]2-AG alone; (b) the uptake of [14C]AEA and [3H]2-AG by cells is inhibited with the same potency by AM404 (Ki = 7.5 ± 0.7 and 10.2 ± 1.7 µm, respectively) and linvanil (Ki = 9.5 ± 0.7 and 6.4 ± 1.2 µm, respectively), two inhibitors of the AEA membrane transporter; (c) nitric oxide (NO) donors enhance the uptake of both [14C]AEA and [3H]2-AG, thus suggesting that 2-AG action can be regulated through NO release; (d) AEA and 2-AG induce a weak release of NO that can be blocked by a CB1 cannabinoid receptor antagonist, and significantly enhanced in the presence of AM404 and linvanil, thus suggesting that transport into C6 cells limits the action of both endocannabinoids.






anandamide membrane transporter


nitric oxide


Nω-nitro-l-arginine methyl ester


sodium nitroprusside







Anandamide (arachidonoylethanolamide, AEA [1]) and 2-arachidonoylglycerol (2-AG [2,3]), are two substances found in several animal tissues that have been proposed to act as endogenous agonists of cannabinoid receptor, e.g. endocannabinoids, and shown to mimick several of the pharmacological actions of cannabis psychoactive principle, Δ9-tetrahydrocannabinol (THC) (reviewed in [4–7]). The termination of the action at cannabinoid receptors by AEA and 2-AG is achieved through their uptake by cells and subsequent enzymatic degradation processes (reviewed in [8,9]). In the case of AEA, it was soon clear that uptake by cells occurs via diffusion through the cell membrane, facilitated by a saturable, temperature-dependent and selective transporter [10]. Such a transporter, the ‘anandamide membrane transporter’ (AMT), has been identified in most of the cells analyzed so far (reviewed in [11]), and inhibitors capable of enhancing AEA actions in vitro and in vivo have been developed [12,13]. Also, structure activity relationship studies have been carried out on the AMT with a large variety of AEA analogues (reviewed in [11,14–18]). It was established that at least one or no less than four cis double bonds in the fatty acyl chain are necessary in order to simply bind to the AMT or also being transported into cells, respectively [14]. On the other hand it was also observed that the ethanolamine ‘head’ of AEA could be substituted for more hindering groups, better if aromatic, and still yield compounds capable of binding to the AMT [15,16]. Finally, it was reported that the AMT is present in blood [19–22] as well as endothelial [23] cells, is activated by nitric oxide (NO) [20–23], and is inhibited by plant cannabinoids, THC and cannabidiol at micromolar concentrations [24]. The role of intracellular catabolism of AEA in driving, in part, the AMT was also pointed out [24,25].

The uptake of 2-AG by cells was first observed in rat basophilic RBL-2H3 and mouse neuroblastoma N18TG2 cells, and was shown to be inhibited by unsaturated monoacylglycerols such as 2-oleoyl-, 2-linoleoyl and 2-linolenoyl-glycerols [26]. At a closer look, it was observed that although taken up in a temperature-dependent manner unmetabolized 2-AG, unlike AEA, would not greatly accumulate into cells and, in any case, not in a fashion sensitive to low temperature [23,27–29]. In studies carried out in RBL-2H3 and J774 cells [24,27,28] it was also observed that 2-AG does not inhibit efficiently the uptake by cells of [14C]AEA when the two substances are present at the same concentration, thus suggesting that the AMT would not recognize 2-AG as a substrate. These latter data were not confirmed by experiments carried out with human astrocytoma cells [14], rat cerebellar granule cells [16], and human umbilical vein endothelial cells [23]. In those experiments, however, the concentration of radiolabelled AEA used was much lower than the concentration of 2-AG. The uptake of 2-AG is followed by its enzymatic hydrolysis or esterification into phospholipids [23,27–29], two processes that appear to efficiently drive 2-AG diffusion through the cell membrane and, as assessed by the use of several inhibitors, occur also independently of each other. Very recently, a study carried out in human astrocytoma cells [30] confirmed that 2-AG is taken up by cells, and showed for the first time that this process can be inhibited by the previously developed AMT inhibitor, AM404 [12], thus re-opening the possibility that 2-AG is transported into cells via a facilitated diffusion mechanism, and, in particular, through the AMT.

In this study we have re-addressed the question of whether AEA and 2-AG are taken up by cells through the same mechanism by using C6 glioma cells, previously shown to express the AMT [11], and by studying the effect of both enhancers and inhibitors of the AMT on AEA and 2-AG uptake in intact cells. Furthermore, we wanted to understand whether 2-AG uptake by cells does limit 2-AG pharmacological actions. These issues are very important as 2-AG was shown to be a more efficacious cannabinoid receptor agonist than AEA [31,32], and modulators of the AMT may have important therapeutic applications, particularly if they are found to regulate simultaneously the effects of both AEA and 2-AG. We suggest that a 2-AG membrane transporter, similar or identical to the AMT but with an efficacy for 2-AG lower than that of the AMT for AEA, mediates 2-AG uptake by, and limits 2-AG action on, intact cells.

Materials and methods

Rat C6 glioma cells were a kind gift of D. G. Deutsch (New York University at Stony Brook, NY, USA). AEA was synthesized by us as described previously [19], and 2-AG was a generous gift of R. Mechoulam (Hebrew University, Jerusalem, Israel). Linvanil was synthesized as described in [15], whereas AM404 was purchased from Biomol. Nω-nitro-l-arginine methyl ester (l-NAME), sodium nitroprusside (SNP), S-nitroso-N-acetylpennicilamine (SNAP) and 3-morpholino-sydnonimine (SIN-1) were purchased from Sigma. SR141716A was a kind gift from Sanofi Recherche (Montpellier, France).

Experiments on the uptake of [14C]AEA and [3H]2-AG (5 mCi·mmol−1, radiolabelled on the ethanolamine and arachidonate moieties, respectively, and synthesized as described in [19] and [33]) were performed according two protocols. In the first case, varying concentrations for 5 min, or the same concentration (4 µm) for different intervals of time, of the two radioligands were added to intact confluent cells in 6-well plates in serum-free culture medium. Cells were then placed either at 37 °C or 4 °C. After the incubation, cells were placed on ice, and both the incubation medium and the cells were separately extracted with chloroform/methanol (2 : 1, v/v) or chloroform/methanol/serum-free medium (2 : 1 : 1, v/v/v), respectively. In the case of experiments with [3H]2-AG, the lipid extract from the cells was purified by thin layer chromatography on analytical plates (Merck) developed with chloroform/methanol/NH4OH (9 : 1 : 0.1, v/v/v) and the bands corresponding to [3H]2-AG (Rf = 0.7), [3H]arachidonic acid (Rf = 0.4) and [3H]phospholipids (Rf = 0.1) were scraped off the plates and radioactivity counted by scintillation spectrometry. The organic extract of the incubation media was lyophilized and counted by scintillation spectrometry. In the case of experiments with [14C]AEA, the aqueous and the lyophilized organic phase from both the incubation media and the cell extracts, containing [14C]ethanolamine formed from [14C]AEA hydrolysis and residual [14C]AEA, respectively, were counted by scintillation spectrometry.

In one set of experiments, cells were incubated with the same amount (20 000 d.p.m., 4 µm) of both [14C]AEA and [3H]2-AG. In this case, the two compounds found in the lyophilized organic phase of either the media or cell extracts were separated by TLC developed with the organic phase of a mixture of iso-octane/ethyl acetate/water/acetic acid 50 : 110 : 100 : 20 (v/v/v/v). Under these conditions AEA exhibits an Rf of 0.6 and 2-AG of 0.75.

The effect of endocannabinoids, AMT inhibitors and NO-donors on [14C]AEA and [3H]2-AG uptake was studied by coincubating the cells at 37 °C with these substances and either [14C]AEA or [3H]2-AG (4 µm), for 5 or 10 min in the case of inhibitors or NO-donors, respectively. Cells were then placed on ice, and the incubation medium extracted with chloroform/methanol (2 : 1, v/v). The organic phase was then lyophilized and radioactivity measured by scintillation counting of either the whole extract (in the case of [14C]AEA uptake) or of TLC purified fractions of the extract (in the case of [3H]2-AG uptake). For the study of 2-AG effect on AEA uptake, we also used two other procedures, identical to those described in [23] and [16], respectively, and differing from the procedure described above mainly by the use of a different concentration of labeled AEA. In both cases [3H]AEA (220 Ci·mmol−1, NEN Life Science Products) was used. In the first case we studied the effect on the uptake of 0.2 µm[3H]AEA of 2-AG (2 µm) as well as AM404 (10 µm), linvanil (10 µm) and the NO donors [23]. In the second case, the effect of various doses of 2-AG, AM404 and AEA on 0.2 nm[3H]AEA was studied [16].

The effect of AEA and 2-AG, or of NO-donors SNP, SNAP and SIN-1, on NO release from C6 cells was studied as described previously [23]. Briefly, generation of NO was determined by measuring accumulation of the stable end product nitrite (NO2) in culture supernatants [34]. C6 cells (5 × 106 cells per test) were treated with different compounds (or vehicle alone in the controls) for 15 min, then the nitrite levels were determined in the culture medium via spectrophotometric analysis, after using nitrate reductase (Alexis Corporation, Läufelfingen, Switzerland) and the acid-catalyzed di-azotation reaction with sulfanylamide and naphtylethylenediamine (Griess reaction) as described [34]. Nitrite levels in culture supernatants were within the linearity range of calibration curves made from a solution of sodium nitrite. In some experiments, the two AMT inhibitors AM404 and linvanil, the NO synthase inhibitor l-NAME, the CB1 antagonist SR141716A, or vehicle, where coincubated with AEA or 2-AG.

The expression of the CB1 receptor in C6 cells was assessed by RT-PCR amplification of total RNA, carried out as described in [28], and by binding studies performed as described in [35] with [3H]CP55 940 (0.4 nm, 126 Ci·mmol−1, NEN Life Science Products), which was displaced from C6 cell membranes by both SR141716A (1 µm) and AEA (1 pm– 1 µm).


The uptake of AEA and 2-AG by C6 glioma cells is saturable

We studied the uptake of [14C]AEA and [3H]2-AG by C6 cells at different time intervals, and different concentrations, both at 37 °C and 4 °C. We found that both compounds are sequestered from the incubation medium in a time- and temperature-dependent way and in a saturable fashion (Fig. 1a,b and Fig. 2a,b). It was possible to calculate the apparent Km and Bmax values for both [14C]AEA and [3H]2-AG, which were 11.0 ± 2.0 and 15.3 ± 3.1 µm, and 1.70 ± 0.30 and 0.24 ± 0.04 nmol·min−1·mg protein−1, respectively (Fig. 1). The values for [14C]AEA were quite different from those published by some of us for the same cells in previous studies [25,35]. However, these studies were performed using a different procedure, and cell lines that underwent less subculturing passages. On the other hand, the Km for AEA (10.7 µm), calculated by using the procedure described in [16], was identical to the one described above. While procedural differences can lead to remarkably different results (as discussed below), the biochemical features of cells may depend on the number of subculturing passages. For example, it was shown that MCF-7 cells respond to endocannabinoids in a dramatically different way when undergoing more than 15 subculturing passages [36] (see also below). In the present study, whose aim was a comparison between the uptake process of AEA and 2-AG, we always performed comparative experiments with the two compounds by using cells that had undergone the same number of subculturing passages.

Figure 1.

Concentration- and temperature-dependent uptake of 2-AG (A) and AEA (B) by C6 rat glioma cells. On the y-axis the amounts of labeled ligand disappeared from the incubation medium of C6 cells incubated at either 37 °C or 4 °C with increasing concentrations of either [3H]2-AG or [14C]AEA are shown. The net uptake (37 °C minus 4 °C) is also shown and data were used to calculate apparent Bmax and Km values. Each data point is the mean of three determinations. Standard deviation bars are not shown for the sake of clarity and were never higher than 5% of the means. The graphs are representative of two similar experiments.

Figure 2.

Time-dependent uptake of 2-AG (A) and AEA (B) by C6 rat glioma cells. On the y-axis the amounts of labeled metabolites either in the incubation medium or associated to cells are shown for either [3H]2-AG or [14C]AEA (4 µm), in the absence or presence of [14C]AEA or [3H]2-AG (4 µm), respectively. Each data point is the mean of three determinations. Data for arachidonic acid (AA) refer to the sum of the amounts found in both the incubation medium or associated to cells. Standard deviation bars are not shown for the sake of clarity and were never higher than 5% of the means. The graphs are representative of two similar experiments.

Endocannabinoid uptake by C6 glioma cells coincubated with both [14C]AEA and [3H]2-AG

The effects of AEA and 2-AG on each other’s uptake were determined by coincubating C6 glioma cells with [14C]AEA and [3H]2-AG for several time intervals and comparing their accumulation into cells, and the rate of their intracellular metabolism, to the same processes observed when incubating the cells with either compound alone (Fig. 2a,b). In the case of [14C]AEA, the presence of [3H]2-AG did not significantly change the amounts of the compound accumulated by cells, nor those hydrolyzed to [14C]ethanolamine (Fig. 2b). If anything, a trend for increasingly lower levels of [14C]AEA in the incubation medium was observed in the presence of [3H]2-AG. Conversely, [3H]2-AG accumulation was affected. Specifically, we observed: (a) higher levels of [3H]2-AG in the incubation medium, (b) less [3H]radioactivity incorporated into phospholipids, (c) less [3H]radioactivity incorporated into arachidonic acid released from [3H]2-AG hydrolysis, and (d) no significant effect on intact [3H]2-AG accumulated into cells (Fig. 2a).

Effect of various compounds on [14C]AEA uptake

Several doses of the two AMT inhibitors, AM404 and linvanil, and of AEA and 2-AG, and a single dose of the NO donors SNP (2.5 mm), SNAP (2.5 mm) and SIN-1 (0.5 mm), were tested on [14C]AEA uptake, and found to exert different effects. These optimal doses of the NO donors were determined in previous studies [20–23]. AM404, linvanil and AEA inhibited [14C]AEA uptake (Ki = 7.5, 9.5 and 13.4 µm, respectively), whereas 2-AG was much less potent (Ki = 30.1 µm) (Table 1) and the three NO donors slightly enhanced [14C]AEA uptake (111.0 ± 2.3, 116.6 ± 3.1 and 114.5 ± 2.6% control, respectively, P < 0.05 by anova, means ± SD, n = 3). No change in the presence of 0.5 mm l-NAME (which alone had no effect) was found in the effect of 2-AG on [14C]AEA uptake (Ki = 31.7 µm) (Table 1). When we changed the experimental protocol used for assaying AEA uptake, and used the one described in [23], we found that 2-AG (2 µm) did not significantly inhibit the uptake of a 10-fold lower amount of AEA (0.2 µm), whereas AM404 and linvanil (10 µm) did. In addition, the NO donors significantly stimulated AEA uptake (Fig. 3). When we repeated the experiment by using a much lower concentration of radiolabelled AEA (0.2 nm), we found that 2-AG did inhibit AEA uptake with a Ki of 4.6 ± 1.2 µm. Under these conditions the Ki values for AM404 and AEA (5.9 ± 1.1 and 12.3 ± 1.1 µm, respectively, means ± SD, n = 3) were slightly lower than those measured with a high concentration of radioligand.

Table 1.  K i values (µm) for the inhibition of either [3H]2-AG or [14C]AEA uptake by rat C6 cells. Data are means ± SD of n = 3 experiments. In the case of [14C]AEA uptake inhibition by 2-AG, the Ki value in the presence of 0.5 mm l-NAME is also shown. ND, not determined.
 [3H]2-AG uptake
(Ki, µm)
[14C]AEA uptake
(Ki, µm)
  1. a  P < 0.05 by anova vs 2-AG on 2-AG uptake, and vs AEA on AEA or 2-AG uptake.

AEA20.5 ± 3.213.4 ± 1.4
2-AG18.9 ± 1.830.1 ± 3.9a
2-AG + l-NAME (0.5 mm)ND31.7 ± 2.9
AM40410.2 ± 1.77.5 ± 0.7
Linvanil6.4 ± 1.29.5 ± 0.7
Figure 3.

Effect of various agents on the uptake of [3H]AEA by rat C6 cells, assayed according to the procedure described in [ 23 ]. Data are means ± SD of n = 3 experiments. Student’s t-test was used to compare means.

Effect of various compounds on [3H]2-AG uptake

Several doses of the two AMT inhibitors, AM404 and linvanil, and of AEA and 2-AG, and a single dose of the NO donors SNP (2.5 mm), SNAP (2.5 mm) and SIN-1 (0.5 mm), were tested also on [3H]2-AG uptake. AM404 and linvanil potently inhibited the uptake (Ki = 10.2 and 6.4 µm, respectively), whereas AEA and 2-AG were less active (Ki = 20.5 and 18.9 µm, respectively) (Table 1), and the three NO donors slightly enhanced [3H]2-AG uptake (114.7 ± 2.1, 116.2 ± 3.0 and 116.1 ± 2.4% control, respectively, P < 0.05 by anova, means ± SD, n = 3).

Effect of AEA and 2-AG on NO release by C6 cells

Rat glioma C6 cells contain CB1 cannabinoid receptors [37], whose stimulation is coupled with NO release [23,38]. Treatment of C6 cells with either AEA (1 µm) or 2-AG (1 µm) led to a small, albeit significant, release of NO (Fig. 4). This effect of the endocannabinoids was blocked by both the NO synthase inhibitor l-NAME (0.5 mm) and by the CB1 antagonist SR141716A (0.1 µm). Lower levels of CB1 receptors were found in the C6 cells utilized in this study as compared to previous studies carried out with the same cells [35,37], and as assessed by both RT-PCR (Fig. 5) and binding assays (specific binding of 0.4 nm[3H]CP55 940 was 80 ± 1% of the total binding and = 12.7 ± 1.3 fmol·mg protein−1; Ki for AEA displacement of [3H]CP55 940 was 30 ± 4 nm, a value compatible with previous data in other cells and tissues). This finding may explain the little effect of AEA and 2-AG on NO release observed here. At any rate, this effect was significantly enhanced by two substances, AM404 and linvanil, shown here and previously [12,15,23,30] to block AEA and 2-AG uptake by cells, thus suggesting that re-uptake mechanisms contribute to minimize the actions at CB1 receptors not only of AEA [12] but also of 2-AG. Both AM404 and linvanil had no effect per se on the release of NO from C6 cells (not shown).

Figure 4.

Effect of various agents on nitric oxide (NO) release from rat C6 cells. Data are means ± SD of n = 3 experiments. Student’s t-test was used to compare means.

Figure 5.

Expression of a rat CB1 receptor transcript in C6 rat glioma cells. Faint, but detectable, amounts of a transcript of the size (675 bp) expected from the rat CB1 mRNA sequence using the appropriate oligoprobes [28] was detected when total RNA from C6 had been retro-transcripted into cDNA (lane 4). No transcript was observed with non retrotranscribed RNA (lane 2) thus indicating the absence of contaminating DNA in the RNA preparation from C6 cells. A 100-bp DNA ladder and the housekeeping gene transcript (β-actin, 325 bp) are shown in lane 1 and 3.


Discrepant data have been reported in studies where the effect of 2-AG on the uptake by cells of AEA was examined. In RBL-2H3 cells and J774 macrophages very little inhibition was observed [27,28]. In rat cerebellar granule cells and human astrocytoma cells as well as in human umbilical vein endothelial cells 2-AG exerted a stronger inhibition of AEA uptake [14,16,23]. There are several possible reasons for these discrepancies. (a) 2-AG is very probably a mono-arachidonoylglycerol mixture in vivo because of acyl migration occurring during the uptake assay. In one case where mono-arachidonoylglycerols were found to inhibit AEA uptake it was shown that the 2-isomer was more potent than the 1(3)-isomers [14]. Migration from sn-2 to sn-1(3) positions is more rapid than sn-1(3) to sn-2 migration. The actual relative ratio of the 2-AG/1(3)-AG isomers is impossible to control and may depend on the experimental conditions used for uptake experiments. Thus, experimental conditions may determine how potently 2-AG inhibits AEA uptake. (b) In previous reports, the effect of 2-AG on AEA facilitated transport has been studied by using concentrations of radiolabelled AEA that were either similar [27,28] or much lower [14,16] than those of 2-AG. Whenever a nm concentration of radiolabelled AEA was used in the assay, 2-AG inhibited AEA uptake by cells rather potently (Ki < 20 µm). Here, by using a 4-µm concentration of [14C]AEA, again we observed an inhibition of AEA uptake only with concentrations of 2-AG (Ki = 30.1 µm) significantly higher than those necessary to unlabelled AEA to inhibit [14C]AEA uptake, or to either AEA or 2-AG to inhibit [3H]2-AG uptake. When we used the experimental protocol described in [23], we found that 2-AG did not inhibit the uptake of a 20-fold lower amount of AEA (0.2 µm), whereas AM404 and linvanil did. Finally, when we used a 20,000-fold lower concentration of radiolabeled AEA (0.2 nm) and the conditions described in [16], we found that 2-AG inhibited AEA uptake with a Ki of 4.6 µm. These findings confirm that the potency of 2-AG as an inhibitor of AEA uptake depends on the experimental protocol, but also suggest that the AMT in C6 cells is not very efficacious with 2-AG as this substance becomes as potent as other inhibitors only when a very low concentration of ligand is used. (c) Finally, it is known that activation of CB1 cannabinoid receptors stimulates NO release, and that NO activates the AMT [20–23]. Therefore, in principle, it is possible that the inhibitory effect by 2-AG of AEA uptake may be masked by a NO-mediated stimulatory effect on the AMT. However, we showed here that, although 2-AG does stimulate NO release in C6 glioma cells, and this effect is blocked by the NO synthase inhibitor l-NAME, 2-AG does not inhibit [14C]AEA uptake more effectively in the presence of l-NAME.

The only two studies where the effect of AEA on 2-AG uptake was examined gave contrasting results. In J774 macrophages, AEA did not inhibit the uptake of 4 µm[3H]2-AG [28], whereas a strong effect by AEA on the accumulation of 30 nm[3H]2-AG in human astrocytoma cells was found [30]. Here, we showed that AEA inhibits [3H]2-AG uptake also when using 4 µm of the radioligand. Indeed, when incubated with both radiolabeled endocannabinoids in equal concentrations, intact C6 cells take up AEA at the same rate and 2-AG at a lower rate, compared to what is observed when the two substances are incubated separately. These data may suggest that AEA can be taken up by both the AMT and a 2-AG transporter, whereas 2-AG is mostly taken up by the latter mechanism. Alternatively, a single AMT may mediate the uptake of both compounds, but in this case its efficacy with 2-AG would again be much lower than with AEA.

As shown in previous studies [23,27–30], we observed here that when C6 cells are incubated with [3H]2-AG, a part of the radioactivity is time-dependently incorporated into free arachidonate and part into membrane phospholipids. We have previously shown that 2-AG esterification into phospholipids occurs also independently of its hydrolysis to arachidonic acid. Furthermore, 2-AG hydrolysis is catalyzed by FAAH and at least one additional hydrolase [27–29,39]. If the uptake of a solute by cells is driven uniquely by the gradient of concentration across the membrane, its accumulation into cells will reach equilibrium when there is no difference between its free concentrations outside and inside the cell. Thus, intracellular metabolism drives, at least in part, the facilitated transport of the solute. In the case of AEA, metabolism is slower than uptake [10], although it contributes in part to driving the AMT [24,25]. Hence, intact AEA accumulation into cells and its sensitivity to low temperature can be observed in most cell types. However, in the case of 2-AG, whose multipathway metabolism (Scheme 1) is almost simultaneous to its uptake, little cell accumulation of the unmetabolized compound is often observed ([27–29] and this study). Furthermore, this accumulation is undistinguishable from that observed at low temperatures, which is due to temperature-insensitive, passive diffusion. Substances such as AEA that, as for low temperature, inhibit both the transport and the intracellular metabolism of 2-AG (Scheme 1) will also have little effect on the amounts of intact 2-AG accumulated into cells (Fig. 2b).

Figure Scheme 1. .

Uptake and intracellular metabolism of 2-arachidonoylglycerol (2-AG): effect of anandamide (AEA) and low temperature. The uptake of 2-AG by cells is followed by rapid and efficient intracellular metabolism, which prevents a large accumulation of the compound in the cells. 2-AG intracellular metabolism consists of: (1) hydrolysis by fatty acid amide hydrolase (FAAH) [27,39]; (2) hydrolysis by monoacylglycerol lipases [28,29,39]; and (3) direct esterification into membrane phospholipids (PL) [27–29]. Arachidonic acid (AA) produced from 2-AG hydrolysis is also esterified into phospholipids [27–30]. AEA inhibits at the same time both 2-AG uptake, by competing for the same site on membrane transporter, and 2-AG hydrolysis. Through arachidonic acid produced from AEA hydrolysis, AEA can inhibit also the formation of [3H]phospholipids subsequent to [3H]2-AG hydrolysis [29]. Lowering the temperature to 4oC also inhibits both the uptake and intracellular metabolism of 2-AG. Dashed arrows denote inhibition; solid arrows denote reversible processes. The catabolic reactions of AEA and 2-AG occur inside the cell.

An alternative explanation [27] to the lack of temperature sensitivity of 2-AG intracellular levels would be that there is no 2-AG transporter, and that the diffusion of 2-AG through the cell membrane is uniquely driven by its rapid intracellular metabolism. This latter explanation, however, is now not supported by the finding that 2-AG uptake is inhibited by the AMT inhibitors, AM404 and linvanil ([30] and this study) and enhanced by NO donors (this study). In fact, AM404, linvanil and NO do not appear to affect 2-AG metabolism, and their effect on 2-AG uptake can be explained only by the interference with a ‘2-AG transporter’. Interestingly, AM404 and linvanil were shown here to enhance both AEA and 2-AG stimulation of NO release, a CB1 receptor-mediated effect, thus also providing for the first time evidence that facilitated transport of 2-AG plays an important role in limiting its actions. The inhibitors counteracted AEA and 2-AG uptake with a similar potency, whereas the NO donors used in this study enhanced to the same extent the uptake of both compounds by C6 cells. These findings suggest that mechanism(s) with similar sensitivity to inhibitors and activators mediate(s) the facilitated transport of endocannabinoids.

The data presented here can be interpreted in two ways. (a) One single transporter mediates the facilitated diffusion across C6 cell plasma membrane of both AEA and, with a significantly lower efficacy, 2-AG; (b) Two distinct, although functionally similar, mechanisms exist, one selective for AEA vs 2-AG, and the other that is capable of transporting both AEA and 2-AG with similar efficacies. We observed that, despite the fact that AEA and 2-AG are taken up by cells with similar apparent Km values, AEA inhibits 2-AG uptake at concentrations significantly lower than those needed to 2-AG to inhibit AEA uptake, which might support the existence of two distinct transporters. This finding, however, could be also due to the fact that AEA, which is taken up by cells more rapidly than 2-AG, inhibits 2-AG uptake by inhibiting also 2-AG intracellular metabolism (Scheme 1). By contrast, 2-AG is likely to inhibit AEA uptake by interfering only with the AMT. The possibility of the two transporters is in agreement with all the experimental data described in this article, although one might question the likelihood and need of two mechanisms, both equally sensitive to the agents tested here. It is possible that splicing of the same gene or the same gene transcript leads to more than one transporter. It is also possible that cells use for AEA a specific transporter to avoid interference by 2-AG, which is more abundant than AEA in animal tissues. Only the molecular characterization of the AMT, its over-expression in cells and, most importantly, its genetic deletion will provide a definitive answer as to whether 2-AG transport into cells is mediated or not by this protein. Whatever interpretation is given to the data presented here, our finding that 2-AG uptake by C6 cells is less efficacious than AEA uptake explains why 2-AG is poorly released from neuronal cells as compared to AEA [33], as it was shown that in neurons facilitated transport through the cell membrane mediates also AEA release from cells [11].

In conclusion, this study, while addressing the question of whether the same or distinct mechanisms mediate the diffusion into cells of AEA and 2-AG, showed that the use of different experimental conditions leads to contradictory results in studies on endocannabinoids uptake, thus possibly explaining some of the discrepancies found in the literature on this subject. More importantly, for the first time this study: (a) provides evidence for the existence of a ‘2-AG/AEA membrane transporter’ that, even if proven to be distinct from the AMT, is still likely to have similar molecular and regulatory features; (b) shows that inhibitors of the uptake of 2-AG enhance a CB1 receptor-mediated effect of this endocannabinoid, i.e. the release of NO [40]; and (c) demonstrates that agents that induce NO release can also enhance the uptake of 2-AG by cells. These findings raise the previously unexplored possibility that the actions of 2-AG might be regulated, both pharmacologically and physiologically, by modulating its re-uptake by cells.


This work was supported by the Ministero dell’Università e della Ricerca Scientifica e Tecnologica (PRIN Program to A. F. A, and 3933 to V. D. M.).