Endocannabinoid (eCB) signaling controls various central functions in mammals, such as nociception, feeding, energy homeostasis, mood, learning, memory, growth, development, and reward processes (1–6). The eCB system consists of cannabinoid receptors (CB1 and CB2), lipid-derived endogenous ligands [N-arachidonoylethanolamine or anandamide, 2-arachidonoylglycerol (2-AG)], and specific enzymes involved in the biosynthesis and degradation of eCB.

Many questions on eCB signaling still need to be answered. However, among the two endogenous ligands, anandamide and 2-AG, 2-AG appears to be the most potent endogenous ligand for brain cannabinoid CB₁ receptors (7). 2-AG acts as a full agonist of CB1 and CB2 receptors (8–10), and is synthesized on demand from 2-arachidonate-containing phosphoinositides by the action of diacylglycerol lipase (DAGLα and DAGLβ) in response to elevations of intracellular calcium. Unlike classic neurotransmission systems (e.g., monoaminergic, cholinergic, aminoacidergic), 2-AG signaling in the nervous system takes place in a retrograde fashion. Thus, stimulation of the postsynaptic neuron induces the biosynthesis of 2-AG that thus diffuses to the presynaptic terminal to act on CB1 receptors; due to its lipophilic nature, no synaptic vesicles for storage of 2-AG exist (4, 11, 12). Given these characteristics, 2-AG is considered to be a neuromodulator rather than a transmitter. CB1 activation of G_i/o proteins inhibits neurotransmission but the net inhibitory or excitatory effect of 2-AG signaling depends on the type of neurons involved in the process (4, 13–16). 2-AG is mainly degraded by monoacylglycerol lipase (MAGL), a membrane-associated, cytoplasm-facing soluble enzyme located pre-synaptically at axon terminals along with the CB1 receptor (17–19). Furthermore, about 15% of 2-AG is hydrolyzed by ABHD6 and ABHD12, and the remaining 1–2% by fatty acid amide hydrolase (FAAH) (20). The mechanisms of 2-AG neuronal reuptake are not completely known, but putative transporters have been described (3, 21).

Drugs of abuse affect brain eCB levels and it has been proposed that the activation of the eCB system is involved in many aspects of addiction (22). Drug-induced increases in eCB synthesis play a role in drug reward, and it has been suggested that eCBs are involved in long-term synaptic plasticity of neural substrates of motivation and reward in relation to addiction (23).

In human beings, a genetic disruption of eCB clearance is associated with drug abuse (24, 25). CB1 receptor activation affects the addictive properties of opioids, cocaine, alcohol, and nicotine (26). On the other hand, CB₁ KO mice show reduced ethanol and morphine self-administration (SA) and attenuated ethanol- and opiate-induced place conditioning (27–29). In rats, the CB₁ receptor inverse agonist/antagonist SR-141716A (Rimonabant) reduces ethanol and opiate SA (30–32) while in mice it reverses the behavioral and neurochemical effects of cocaine (33).

Δ9-Tetrahydrocannabinol (THC), the main active component of marijuana, is known to reduce the severity of opioid withdrawal in human beings and rodents, and acute inhibition of FAAH or MAGL alleviates symptoms of precipitated and spontaneous withdrawal in opioid- and THC-dependent mice (34, 35). The eCB system is involved in the rewarding effects of THC (22), as well as ethanol and opiates (36, 37). In particular, ethanol and heroin SA increases eCB levels in the nucleus accumbens (NAc) shell, suggesting a role for NAc eCB transmission in the reinforcing effects produced by these drugs (38). Finally, eCBs are SA by squirrel monkeys, meaning that they are effective reinforcers of drug-taking behavior (39, 40).

Dopamine (DA) neurotransmission is known to regulate a large number of motivated and addiction-related behaviors (41–43). Importantly, drugs of abuse of the most diverse pharmacological classes increase mesolimbic DA transmission in rats as well as in human beings preferentially or, depending on doses and conditions, selectively in the ventral striatum/NAc (42, 44–48). Accordingly, THC and the synthetic cannabinoid WIN 55,212-2 increase extracellular DA concentrations in the shell, but not in the core of the NAc, both when injected intravenously (iv) (49) and when SA by different strains of rats (50, 51). In addition, anandamide increases NAc shell DA and its effect is amplified by the FAAH inhibitor, URB597 (52). On the other hand, reduced DA transmission in the mesolimbic system is associated with spontaneous or rimonabant-precipitated THC withdrawal (53, 54).

In spite of this extensive literature, no studies are available on the reinforcing effects of 2-AG and its effects on DA transmission in rodents. In order to fill this gap, we studied intravenous 2-AG SA and monitored the effect of 2-AG on extracellular DA in the NAc shell and core via microdialysis in male Sprague-Dawley rats.

In the present study, we demonstrate for the first time that the eCB 2-AG is SA by Sprague-Dawley rats and stimulates DA transmission preferentially in the NAc shell as compared to the core. Hence, 2-AG possesses behavioral and DA releasing properties similar to those of food and drug rewards.

Initially, animals acquired 2-AG SA (25 μg/kg/20 μl infusion, unit dose) in single daily 1 h FR1 sessions for 5 weeks, which was then increased to FR2, under food restriction that maintained body weights at stable levels throughout these studies. Active nose pokes significantly increased over inactive nose pokes from the 6th SA session (acquisition phase), but no significant increase of nose pokes was observed under FR2. When 2-AG was replaced by vehicle (25th SA session, extinction phase), the responding rate as well as the number of injections slowly decreased. When the vehicle was replaced once again with 2-AG, SA behavior immediately recovered (reacquisition phase) (Figures 1A,C). In a second experiment, the effect of dose of 2-AG (12.5, 25, 50 μg/kg/20 μl infusion, unit dose) and the CB1 receptor-dependent SA behavior was assessed. Different doses of 2-AG showed a tendency toward a classic inverted-U-shape dose–response curve, although non-significant differences were observed between the 25 μg/kg dose and the other doses tested (Figure 3). The reinforcing effects of 2-AG in SA behavior were fully blocked by the administration of the CB1 receptor inverse agonist/antagonist SR-141716A (rimonabant, 1 mg/kg ip, 30 min before SA session) (Figure 2B). The average number of responses during each 2-AG SA session, however, was lower in the present study than the responses (either nose pokes or lever pressing) observed during SA studies using other cannabinoids or other common drugs of abuse. Thus, rats trained to SA WIN 55-212 (50, 56), nicotine (57, 58), cocaine (59, 60), or heroin (56, 61, 62) maintained higher rates of responding than what we observed for 2-AG SA. Nevertheless, our data are in agreement with a previous study showing that 2-AG is iv SA by non-human primates with either anandamide or nicotine SA history (40). Justinova et al. (40), however, observed that in squirrel monkeys the rate of responding maintained by 2-AG injections was similar to those observed in THC, anandamide, methanandamide, or cocaine SA studies (39, 63–65). In contrast with Justinova et al. (40), our experiments were carried out in naive animals. Thus, our findings indicate that 2-AG serves as an effective reinforcer and that its reinforcing properties are not dependent on previous SA of other drug reinforcers.

A second group of naive rats was used to estimate the neurochemical rewarding properties of 2-AG. Thus, 2-AG administered iv at 0.5 and 1.0 mg/kg doses was able to stimulate extracellular DA release preferentially in the NAc shell as compared to the core. The increase of DA in the NAc shell was about 25% over basal value, peaking at 20 min after the injection of 2-AG. DA stimulation in this area was absent after the lowest dose tested (0.1 mg/kg iv), and the DA levels were comparable to those after vehicle injections. Interestingly, the higher dose tested (1.0 mg/kg iv) elicited a biphasic DA response with a second rise in DA levels 60 min after 2-AG injection. Notably, our data show for the first time that 2-AG stimulates DA neurotransmission in the mesolimbic DA system. This feature is a common characteristic of all the drugs abused by human beings and it is believed to be crucial in the activation of brain rewarding pathways (66, 67), to be an anticipatory message for rewarding events (41) and to promote incentive learning (68) specially the DA increase in the NAc shell (42). This 2-AG-dependent increase of DA in the NAc shell most likely does not depend on the direct activation of DA neurons in the VTA, but, as in the case of THC and other cannabinoid agonists, is determined by the activation of CB1 receptors located on presynaptic glutamatergic and/or GABAergic terminals onto VTA DA neurons (69–71).

Previous studies have demonstrated that DA transmission increases in a dose-dependent fashion after administration of natural (THC), synthetic (WIN-55212), or endogenous (anandamide) cannabinoids and that this stimulation of DA in the NAc shell peaks at about 50% over basal levels. The doses used in these previous studies are comparable to those used in our study (49, 50, 52, 72) but they were accompanied by higher rates of SA behavior (39, 63, 64). In the present study, we did not observe a similar effect most likely due to the chemically unstable nature of 2-AG that quickly converts to 1-AG (38, 73). This, together with MAGL activity, could be responsible for the lack of high brain levels of 2-AG. Notably, in the present study, these properties of 2-AG were carefully taken into consideration during the preparation of the solutions for iv administration; 2AG content in administered solutions was determined daily before microdialysis or SA sessions by HPLC–MS/MS analysis.

Even though the rate of responding and DA stimulation were lower for 2-AG than those observed after administration of common drugs of abuse, our study demonstrates that the eCB system stimulation, via the mesolimbic DA system (microdialysis experiments) or by 2-AG itself (SA experiments), encodes rewarding events. The present microdialysis data, however, show that passive administration as a single intravenous bolus of 2-AG increased DA levels in the NAc shell at doses of 5–10-fold higher than the 2-AG intake reached by rats during SA behavior. A possible explanation might be the sensitization of the effect of 2-AG on DA transmission in the NAc shell of animals chronically exposed to 2-AG during the acquisition of 2-AG SA. Further experiments of SA coupled to microdialysis performed with the same experimental protocol previously published by our group (50) will serve to directly correlate the preferential increase of DA in the NAc shell to the 2-AG SA behavior observed in our experiments. On the other hand, the results obtained in 2-AG SA experiments may be also due to the food restriction regimen. Many studies show that the reduction of body weight by 20–25% dramatically enhances drug-induced SA (see (74) for review). Similarly, the suppression of natural growth used in the present study, although not associated with loss of weight, may have increased the motivation for 2-AG, suggesting that 2-AG SA may not be crucial in standard condition (e.g., ad libitum access to food). If so, the decrease of active responding for 2-AG after the administration of rimonabant could be also due to the appetite suppression induced by this drug (75).

Consistent with other studies (40, 52), the main significance of our findings is not that 2-AG may pose a risk of abuse liability, but that the activation of the eCB system together with the DA system may potentiate the reinforcing effects or the abuse potential of many drugs of abuse, supporting the hypothesis that an increase in eCB tone may be involved in drug-seeking behavior (76). Recent studies show that the manipulation of eCB levels in the brain can be selectively and successfully achieved by eCB clearance inhibitors (2, 77–79). Altogether, the above evidence from previous work, in conjunction with our current results, suggests that the eCB system plays a role in reward and reinforcement that makes it a good candidate as a therapeutic target in behavioral disturbances of motivation and reward.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.