The mGluRs comprise eight subtypes categorized into three groups according to similar properties, such as sequence homology, molecular structure, selective pathways of signal transduction, and pharmacological profiles (11–13). Group I receptors (mGluR1 and mGluR5) are coupled to G_q, stimulate phospholipase C and phosphoinositide hydrolysis (14,15), are predominantly located postsynaptically, and modulate neuronal signaling and cellular excitability (16). The Group II (mGluR2 and 3) and Group III (mGluR4, 6, 7, and 8) family of receptors couple to G_i, a pathway that inhibits cyclic adenosine monophosphate formation upon activation (16,17), and are predominantly located presynaptically but have also been localized postsynaptically (18) and function to regulate glutamate release to appropriate physiological levels (17). The mGluRs regulate a variety of neurobiological processes, such as learning and memory, synaptic plasticity, and modulation of ionotropic glutamate and γ-aminobutyric acid (GABA) receptor activity, as well as facilitating glutamate release (16,19–22). Low concentrations of ethanol selectively alter neuronal firing rates (23) and Ca²⁺ levels (24) mediated by mGluRs in vitro, which suggests that ethanol has direct actions on mGluR activity.

Although the neurobiological substrates that regulate the positive reinforcing effects of ethanol remain to be fully characterized, the mesolimbic system has received considerable attention because of its general involvement in reinforcement (25–28) and drug self-administration (29–32). Early studies focused on dopamine neurotransmission within this system, because acute injection of ethanol was found to increase extracellular dopamine levels in the nucleus accumbens (33). Infusion of dopamine agonists in the nucleus accumbens increases ethanol-reinforced responding, whereas D1 and D2 receptor antagonists produce decreases (29,34,35). Similarly, microinjection studies have shown that intra-accumbens systems that interact with dopamine—including GABA-A (36,37), opiate (38,39), and endocannabinoid CB1 (40) receptors—regulate ethanol reinforcement. Evidence also shows that ethanol injection increases glutamate levels in the nucleus accumbens (41) and NMDA receptor antagonist infusion in the nucleus accumbens decreases ethanol-reinforced responding (42,43), which underscores the potential importance of mesolimbic glutamate in this process.

Group I and II mGluRs are highly expressed in the mesocorticolimbic system. mGluR5s and mGluR2/3s are abundant in regions such as the nucleus accumbens, lateral septum, striatum, amygdala, and hippocampus (15,44–46). Alternatively, mGluR1s show low expression in most limbic brain regions but are highly expressed in the cerebellum where they regulate motor coordination (47). This differential mesocorticolimbic distribution pattern of mGluR subtypes suggests that mGluR5 and mGluR2/3 might regulate ethanol reinforcement, but mGluR1 might not. Indeed, systemically administered mGluR5 antagonists and mGluR2/3 agonists have been shown to modulate cocaine, nicotine, and ethanol reinforcement and drug-seeking behavior in reinstatement models (48–55). By contrast, mGluR1 receptor modulation of ethanol self-administration is attributable to nonspecific motor impairment (56,57). Furthermore, evidence from site-specific infusion studies shows that ethanol self-administration is modulated by activity of specific brain regions, including the nucleus accumbens and frontal cortex (29,35,36,42,58), that express high levels of mGluR5 and mGluR2/3. Accordingly, the present study was designed to examine the direct functional involvement of mGluR5 or mGluR2/3 expressed in mesocorticolimbic brain regions in ethanol reinforcement.

Because individuals with variation in the mGluR5 gene (GRM5) show significant risk for developing alcohol dependence (59), we sought to examine the functional involvement of mesocorticolimbic mGluR5 and mGluR2/3 in ethanol reinforcement in alcohol-preferring (P) rats, a prominent genetic model of high alcohol intake (60). The P-rat line has been found to fulfill the requirements of an animal model of alcoholism (61), because these rats voluntarily consume alcohol in quantities that produce significant blood alcohol concentrations, develop tolerance and dependence through voluntary drinking (62–64), and maintain preference for ethanol when palatable solutions are presented as an alternative (65). Briefly, P-rats were trained to self-administer ethanol (15% v/v) and then received surgical implantation of bilateral injector cannulae aimed at the nucleus accumbens core; however, no attempt was made to differentiate between subnuclei of the accumbens (e.g., core vs. shell), because the extent of drug diffusion is greater than the anatomical distance between these structures (66). After surgery, effects of intra-accumbens infusion of the mGluR5 antagonist 2-methyl-6-(phenylethynyl)pyridine hydrochloride (MPEP) and the mGluR2/3 agonist (1R,4R,5S,6R)-4-amino-2-oxabicyclo(3.1.0)hexane-4,6-dicarboxylic acid (LY379268) on the maintenance of ethanol self-administration were determined. The mGluR5 antagonist was also tested in the dorsomedial striatum and medial prefrontal cortex (mPFC) to further examine anatomical specificity of mGluR5 involvement in ethanol-reinforced responding. To assess reinforcer specificity, the mGluR5 antagonist was administered in the nucleus accumbens of P-rats trained to self-administer sucrose (.4% w/v) versus water. Finally, locomotor assessments were conducted after site-specific infusion of mGluR5 or mGluR2/3 compounds to determine whether reductions in self-administration were associated with nonspecific motor impairments.

Previous work has determined a role for mGluR5 in ethanol reinforcement as evidenced by reductions in ethanol self-administration after systemic administration of mGluR5 antagonists (48–51,56,71). This study shows that infusion of the mGluR5 antagonist MPEP (0 μg–10 μg) in the nucleus accumbens reduces ethanol-reinforced responding. The mGluR5 antagonism, or gene knockout, has been shown to potentiate ethanol's acute sedative-hypnotic effects and produce motor impairments after moderate alcohol doses (72–74); thus, examination of the temporal pattern of responding was critical for the interpretation of the total session reductions in ethanol self-administration. Intra-accumbens mGluR5 antagonism produced a general suppression of ethanol-reinforced responding throughout the 30-min self-administration session, indicating that MPEP effects were independent of consumed ethanol. Furthermore, the prolonged suppression of ethanol-reinforced responding across the 30-min session is consistent with the time course of mGluR5 occupancy after systemic MPEP injection (i.e., full receptor occupancy sustained for at least 1 hour (75). Finally, the reduction in ethanol self-administration is likely not related to drug substitution, because mGluR5 antagonism itself does not produce ethanol-like properties and does not alter the subjective properties of low ethanol dose (76–78).

The results of this study complement other evidence showing that glutamate transmission in the nucleus accumbens regulates ethanol reinforcement. Microinjection of the competitive NMDA receptor antagonist AP-5 into the nucleus accumbens of Wistar rats reduced ethanol-reinforced responding (42). The present study extends this work by showing that inhibition of mGluR5 but not mGluR2/3 activity specifically in the nucleus accumbens reduces the maintenance of ethanol-reinforced responding in P-rats that were self-administering relatively high doses of ethanol under control conditions (mean 1.04 g/kg/30-min session). Thus, it seems that disrupting glutamate neurotransmission either through blockade of postsynaptic ionotropic NMDA or metabotropic mGluR5 in the nucleus accumbens is sufficient to prevent the full expression of ethanol's reinforcing properties.

mGluR5s are widely expressed throughout the striatum (44), with similar expression in the nucleus accumbens and caudate putamen (76,79,80). However, infusion of MPEP in the dorsomedial caudate, contrary to the nucleus accumbens infusion, had no effect on ethanol self-administration. The dorsal striatum has been shown to regulate cocaine self-administration by dopamine and AMPA/kainate receptors (81) and ethanol self-administration by NMDA (NR2B) receptors (82). Furthermore, the dorsal striatum has been implicated in learning stimulus-response relations associated with habit formation (83). Given that compulsive/habitual drug use—a hallmark of drug addiction—can be regarded as maladaptive stimulus-response relations, the dorsal striatum has become a region of significant interest in the drug abuse field (84). Although the present findings suggest that mGluR5 within this region do not modulate the maintenance of ethanol-reinforced responding, a partial trend was observed at the highest dose of MPEP tested, suggesting that an extended dose range of MPEP or another mGluR5 antagonist such as 3-((2-methyl-1,3-thiazol-4-yl)ethynyl)pyridine might show efficacy. It will also be both interesting and important for future work to determine whether mGluR5 in the dorsal striatum regulate other aspects of ethanol-seeking behavior that rely on conditioned reinforcement, such as cue-induced reinstatement.

Functional specificity of nucleus accumbens mGluR5 was further evaluated by examining the effects of MPEP infusion in the mPFC on the maintenance of ethanol self-administration. The mPFC was tested for several reasons. First, this brain region has been shown to modulate ethanol self-administration (58,85,86), and low doses of ethanol increase extracellular glutamate levels in the mPFC of alcohol-preferring Lewis rats (87). Second, mGluR5 regulate firing rate of mPFC neurons (88), which is a neural correlate of reward prediction (89). Third, the mPFC sends glutamatergic projections to the nucleus accumbens, and inactivation of the mPFC reduces the firing rate of nucleus accumbens neurons in response to reward-predictive cues (90). Thus, inhibition of postsynaptic mGluR5 in the mPFC might alter ethanol self-administration via local mechanisms or by dampening mPFC glutamatergic inputs to the nucleus accumbens. Results showed, however, that infusion of the mGluR5 antagonist MPEP in the mPFC had no effect on ethanol self-administration. This was evident even though a higher MPEP dose range was tested in the mPFC compared with the nucleus accumbens. These findings suggest that mGluR5 in the mPFC are not involved in the regulation of the maintenance of ethanol self-administration. However, because the mPFC regulates response to predictive cues, mGluR5 in this brain region might influence behaviors that are regulated by associative learning, such as cue-induced reinstatement of ethanol seeking.

Specificity of nucleus accumbens mGluR5 was further confirmed by testing the role of mGluR2/3 in ethanol self-administration. Intra-accumbens mGluR2/3 activation reduced ethanol self-administration. This was evident by a reduction in total session ethanol-reinforced responding and a pattern of suppression similar to that observed after mGluR5 antagonism (i.e., general suppression of responding throughout the session). These results are consistent with evidence showing that intra-accumbens infusion of the mGlu2/3 agonist APDC reduced ethanol-drinking in a 4-bottle home-cage procedure by C57Bl/6 J and DBA2/J mice (91). However, results from the present study showed that LY379268 also reduced spontaneous motor activity, suggesting that reductions in ethanol-reinforced responding were likely due to a motor impairment. This finding is consistent with the findings of a study by Backstrom and Hyytia (92) in which a systemically administered mGluR2/3 agonist produced a significant reduction in ethanol self-administration at a dose that was accompanied by a motor impairment (but see [93]). Given that LY379268 decreased ethanol self-administration only at a dose that produced nonspecific reductions in locomotor activity, these findings indicate that mGluR2/3 activation in the nucleus accumbens does not selectively modulate the maintenance of ethanol self-administration. However, because systemically administered mGluR2/3 agonists exhibit efficacy in other behavioral procedures, such as models of relapse, anxiety, and stress reactivity during abstinence (93–95), future work might determine a more selective role of intra-accumbens mGluR2/3 in these pathological behaviors.

To determine whether mGluR5 in the nucleus accumbens produces a reduction in the maintenance of self-administration of a rewarding substance in general, intra-accumbens mGluR5 antagonism was assessed in rats trained to self-administer sucrose versus water. Baseline level of sucrose (.4% w/v) self-administration was similar to the baseline ethanol self-administration, consistent with our previous findings (56,57). Accordingly, differences in self-administration behavior between the sucrose and ethanol group cannot be attributed to differential effects of the mGluR5 antagonist on response rate. Intra-accumbens antagonism was without effect on sucrose self-administration, even as an mGluR5 antagonist dose range higher than for ethanol self-administration was tested. This finding suggests specificity of the effects of mGluR5 antagonism to ethanol reinforcement and also confirms the lack of antagonist-induced nonspecific motor effects. Furthermore, the dissociation of mGluR5 involvement in ethanol versus sucrose reinforcement is similar to the findings of another study in which motivation to self-administer ethanol but not sucrose was reduced after systemic mGluR5 antagonism (57). Overall, some studies have shown reductions in food-reinforced behavior after mGluR5 antagonism (96,97), whereas other studies have not (53,98,99). However, the present results are consistent with another study that showed no change in sucrose self-administration after intra-accumbens mGluR5 antagonism (100). Together, this data pattern supports a role for intra-accumbens mGluR5 in the selective modulation of ethanol self-administration.

One issue that should be discussed when considering mGlu receptor subtype specificity relates to the selectivity of mGlu receptor compounds tested. Studies have emerged showing that the mGluR5 antagonist MPEP modulates the activity of other receptor systems. For instance, MPEP has been shown to blunt NMDA-evoked currents (101), inhibit the norepinephrine transporter (102), and act as a positive allosteric modulator of mGlu4 receptors (103). The mGluR2/3 agonist LY379268 has recently been shown to stimulate dopamine D2 receptors (104). Moreover, MPEP infused in the nucleus accumbens inhibits intra-accumbens mGluR5 agonist-induced GABA release in the ventral pallidum (105). Because these receptor systems have been shown to modulate ethanol self-administration or voluntary ethanol-drinking (29,82,106–108), it is possible that mGluR5 inhibition in the nucleus accumbens reduced ethanol self-administration via nonspecific pharmacological effects or regulation of associated neural circuit(s).

Supplementary Methods