|Home | About | Journals | Submit | Contact Us | Français|
There is now compelling evidence that the excitatory amino acid neurotransmitter glutamate plays a pivotal role in drug addiction and alcoholism. As a result, there has been increasing interest in developing glutamate-based therapies for the treatment of addictive disorders. Receptors for glutamate are primarily divided into two classes: ionotropic glutamate receptors (iGluRs) that mediate fast excitatory glutamate transmission, and metabotropic glutamate receptors (mGluRs), which are G-protein coupled receptors that mediate slower, modulatory glutamate transmission. Most iGluR antagonists, while showing some efficacy in animal models of addiction, exhibit serious side effects when tested in humans. mGluR ligands, on the other hand, which have been advanced to testing in clinical trials for various medical conditions, have demonstrated the ability to reduce drug reward, reinforcement, and relapse-like behaviors in animal studies. mGluR ligands that have been shown to be primarily effective are Group I (mGluR1 and mGluR5) negative allosteric modulators and Group II (mGluR2 and mGluR3) orthosteric presynaptic autoreceptor agonists. In this review, we will summarize findings from animal studies suggesting that these mGluR ligands may be of potential benefit in reducing on-going drug self-administration and may aid in the prevention of relapse. The neuroanatomical distribution of mGluR1, mGluR2/3, and mGluR5 receptors and the pharmacological properties of Group I negative allosteric modulators and Group II agonists will also be overviewed. Finally, we will discuss the current status of mGluR ligands in human clinical trials.
Over the past two decades, however, a tremendous amount of evidence has accumulated suggesting that the excitatory amino acid neurotransmitter glutamate plays a pivotal role in drug addiction, drug self-administration and reward-related processes, and relapse [1–4]. Glutamate is the most abundant excitatory neurotransmitter in the central nervous system and governs many processes in the brain including fast and slow excitatory neurotransmission, control of basal neuronal activity, and synaptic plasticity.
Glutamate receptors fall into one of two categories: ligand-gated ion channels (i.e., ionotropic glutamate receptors, or iGluRs) which mediate fast excitatory neurotransmission, and G-protein coupled receptors (i.e., metabotropic glutamate receptors, or mGluRs) which mediate slower, modulatory neurotransmission. Numerous animal studies have shown that iGluR antagonists attenuate the rewarding and reinforcing effects of virtually all drugs of abuse, and can attenuate various forms of relapse-like behavior (see  for recent review). However, very few iGluR antagonists are without serious side effects in humans, which include memory loss, disorientation, and the production of symptoms of psychosis such as hallucinations and depersonalization. In light of these unwanted side effects, significant efforts in recent years have been undertaken to pharmacologically manipulate glutamate transmission with selective mGluRs ligands. Such ligands have been shown to not only be of potential benefit in the treatment of addiction, but other disorders of the central nervous system including chronic pain, Parkinson’s disease, depression, anxiety, epilepsy, and neurodegeneration (see refs [5–14] for reviews)
To date, eight different mGluR receptor subtypes have been cloned and characterized, and these receptors appear to have diverse neuroanatomical distributions as well as unique pharmacological and intracellular signaling properties [15–18]. The Group I family of mGluRs consists of mGluR1 and mGluR5 receptors, whereas the Group II family consists of mGluR2 and mGluR3 and the Group III family consists of mGluR4, mGluR6, mGluR7 and mGluR8.
Addiction is a multifaceted chronically relapsing disorder characterized by excessive drug intake, repeated unsuccessful attempts and stopping or reducing drug use, narrowing of the behavioral repertoire towards drug-seeking and self-administration, the emergence of symptoms of tolerance and withdrawal, and continued drug intake despite negative consequences . Although animal models of addiction such as the intravenous drug self-administration, reinstatement, and conditioned place preference paradigms closely resemble several aspects of addiction in humans, none of these paradigms completely encompasses all aspects of the human condition. There is a great need to develop or improve animal models of addiction so that they provide reliable measures of drug craving, voluntary abstinence, relapse following protracted withdrawal, devaluation of natural reinforcers in lieu of drug reinforcement, and impaired decision-making [20–23]. Nonetheless, currently employed animal models of addiction have yielded a vast amount of information on the neural substrates of drug reward and reinforcement, drug-seeking and relapse-like behaviors, as well as the neuroadaptations that accompany chronic drug exposure.
Studies employing the intravenous drug self-administration, reinstatement, and conditioned place preference paradigms have revealed that the mGluR ligands that show the greatest therapeutic potential as anti-addiction agents include mGluR1 and mGluR5 negative allosteric modulators and mGluR2/3 orthosteric agonists. The purpose of the present review is to summarize the existing preclinical literature on the effects of these mGluR ligands on drug and ethanol reward, reinforcement, and relapse-like behaviors. Following a brief overview of glutamatergic transmission, the neuroanatomical distribution of these receptors will be presented, followed by a summary of the pharmacological properties of mGluR1/5 antagonists and mGluR2/3 agonists. Next, following a brief overview of widely used animal models of addiction, we will then summarize the existing literature from animal studies on the efficacy of mGluR ligands in animal models of drug addiction and alcoholism. The review will conclude with an update on the status of mGluR ligands being tested in human clinical trials for other medical conditions, and a discussion of the possible clinical utility of mGluR ligands in the treatment of addiction.
Glutamate is the most abundant excitatory neurotransmitter in the brain, mediating as much as 70% of synaptic transmission within the central nervous system and reaching concentrations in the synaptic cleft as high as the mid-micromolar range. A diagram of a typical glutamatergic synapse is shown in Fig. 1. Glutamate is packaged into synaptic vesicles in the presynaptic terminal by vesicular glutamate transporters. Once released into the synaptic cleft, glutamate can bind to one of three different types of ionotropic glutamate receptors (iGluRs) located on the head of the postsynaptic dendritic spine: the N-methyl-D-aspartate (NMDA) receptor, the α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) receptor, and the kainic acid (kainate, KA) receptor. iGluRs are ligand-gated ion channels that mediate fast excitatory neurotransmission [24–27].
NMDA receptors are heterotetrameric protein complexes composed of at least one NR1 subunit (for which there are at least 8 splice variants) and a combination of NR2A-D and NR3A or 3B subunits [28–30]. In addition to being stimulated by glutamate, amino acids such as D-serine and glycine act as co-agonists at the NMDA receptor. The NR2 subunits contain the glutamate binding domain, whereas the NR1 subunit contains the glycine-binding domain. Under resting conditions, the NMDA receptor channel pore is blocked by Mg2+ ions, but once sufficient membrane depolarization has been established (i.e., by opening of AMPA receptor channels), the Mg2+ block is removed, allowing the influx of cations (primarily Ca2+ ions, but the NMDA receptor is also permeable to K+ and Na+ ions). Once thought to be exclusively located on neurons, NMDA receptors have recently been shown to be expressed on glial cells including microglia, astrocytes and oligodendrocytes . NMDA receptor subunits have also been found to exist on presynaptic terminals . The NMDA receptor has been extensively implicated in mediating neural plasticity as well as learning and memory processes [32–34].
AMPA receptors are also heterotetrameric protein complexes composed of various subunits termed GluR1–4 (also termed GluRA-D) and GluRδ1 and 2 . The mRNAs encoding AMPA subunits can be edited or alternatively spliced to form variants such as the flip and flop isoforms. Each GluR subunit contains a binding site for glutamate. Once activated, AMPA receptors are permeable to various cations including Ca2+, Na+ and K+, although the majority of AMPA receptors in the brain contain GluR2 subunits, which render the channel impermeable to Ca2+. It is believed that both NMDA and AMPA receptors are necessary for the induction of many forms of synaptic plasticity such as long-term potentiation (LTP) and long-term depression (LTD) [35–41].
Like NMDA and AMPA receptors, kainic acid (kainate, KA) receptors are also tetrameric protein complexes composed of various subunits. These subunits are termed GluR5–7 and KA1 and 2 . KA receptors can form homomeric tetramers composed entirely of GluR5, 6 or 7 subunits or heteromeric complexes containing GluR5 or KA subunits. KA receptors are permeable to Na+ and K+ ions and, like NMDA and AMPA receptors, contribute to excitatory postsynaptic currents. The role of KA receptors in synaptic plasticity is less well-defined, however, but KA receptors have been found to be localized presynaptically where they can modulate neurotransmitter release .
In addition to the iGluRs, glutamate can also bind to mGluRs, which are located either in the perisynaptic annulus or on presynaptic terminals. mGluRs are seven transmembrane domain G-protein coupled receptors (GPCRs) that mediate slower, modulatory glutamatergic transmission. mGluRs can be divided into three distinct groups, based on their pharmacological and signal transduction properties. Group I mGluR receptors (mGluR1 and mGluR5) activate the Gαq class of G-proteins, which stimulate one of several phospholipases (including phospholipase C), resulting in phosphoinositide (PI) hydrolysis and the formation of lipid signaling intermediates such as inositol triphosphate (IP3) and diacylglycerol (DAG), which in turn can activate various intracellular messengers including protein kinase C (PKC) [15, 17, 43]. Activation of Group I mGluR receptors also mobilizes calcium release from IP3 receptor-mediated intracellular stores, which can in turn activate other intracellular messengers such as calcium/calmodulin-dependent kinase II (CaMKII). Group I mGluRs, particularly mGluR5, are positively coupled to NMDA receptor function via PKC, and are structurally linked to these receptors as well as IP3-gated intracellular Ca2+ stores via the Homer family of proteins [44–48]. Group I mGluRs are rarely found presynaptically. Group II (mGluR2 and mGluR3) and Group III (mGluR4, mGluR6, mGluR7, and mGluR8) mGluRs, on the other hand, activate the Gαi class of G-proteins and are negatively coupled to adenylyl cyclase (AC) activity, and upon stimulation result in decreased intracellular levels of cyclic adenosine monophosphate (cAMP). Presynaptically localized Group II and Group III mGluRs, particularly mGluR2 and mGluR3, are thought to represent the classical inhibitory autoreceptor mechanism that suppresses excess glutamate release from the presynaptic terminal .
Glutamate is cleared from the extracellular environment by a family of sodium-dependent excitatory amino acid transporters (EAATs). This family of EAATs provides numerous mechanisms to prevent an excessive accumulation of extracellular glutamate, which if high enough concentrations are reached, can result in excitotoxicity. Once inside glial cells, glutamate is converted to glutamine by glutamine synthetase, where it is secreted back outside the glia and taken up by the presynaptic terminal for conversion back to glutamate by glutaminase. Conversely, glutamate can be transported from within glial cells to the extracellular environment by the cystine-glutamate exchanger (xc), and some investigators have proposed that this mechanism is critically involved in regulating extracellular glutamate levels [51–54].
The neuroanatomical localization of Group I and Group II mGluRs in the rodent brain, as assessed by immunohistochemical or in situ hybridization techniques, has revealed overlapping yet distinct patterns of expression of these receptors [18, 55–68] (see Table 1 for details). High levels of mGluR1 expression are found in the olfactory bulb, thalamus, hippocampus (excluding the CA1 region), lateral septum, superior colliculus and cerebellum. Moderate levels are found in the dorsal striatum, hypothalamus, pallidum, ventral midbrain, and cerebral cortex, and low levels are observed in the amygdala, medial septum, nucleus accumbens and brainstem.
In contrast to expression patterns of mGluR1, the expression of mGluR5 is highly concentrated in forebrain and limbic structures. High levels of mGluR5 expression are found in the olfactory bulb, anterior olfactory nuclei, olfactory tubercle, dorsal striatum, nucleus accumbens, lateral septum, hippocampal formation (CA1–CA3 regions and dentate gyrus), and inferior colliculus. More moderate levels of mGluR5 expression are observed in the cerebral cortex (with expression being more dense in superficial as compared to deeper layers), amygdala and caudal portions of the spinal trigeminal nucleus. The presence of mGluR5 is notably low or absent in regions of the hypothalamus, medial septum, ventral midbrain, pons and medulla, and the cerebellum only demonstrates a small amount of mGlu5 mRNA expression in Golgi cells.
To a moderate degree, the expression patterns of mGluR2/3 receptors in the rodent brain parallel those of mGluR5, although the overall abundance of mGluR2/3 receptors appears slighty reduced as compared with that of mGluR5 [5, 18]. Expression levels of mGluR2/3 receptors are high in the olfactory bulb and hippocampus, and moderate in the dorsal striatum, nucleus accumbens, amygdala, anterior thalamic nuclei, cerebral cortex and cerebellum. Low levels of mGluR2/3 are found in the pallidum, colliculi, ventral midbrain and hypothalamus.
Studies employing electron microscopy techniques have revealed that the vast majority of mGluR1 and mGlu5 receptors (i.e., >90%) appear to be located postsynaptically on the perisynaptic annulus of dendritic spines (). There is, however, some evidence that these receptors can indeed be localized to axon terminals in regions such as the hippocampus  and cerebral cortex [70, 71]. The function of presynaptic Group I mGluRs is unknown, but a recent study  demonstrated that they cooperatively interact with presynaptic NMDA receptors in a similar fashion to that observed in the postsynaptic membrane. mGluR2/3 receptors are highly expressed in glutamatergic presynaptic terminals where they function as inhibitory autoreceptors, but are also found on postsynaptic membranes as well.
Finally, although it is clear that high levels of Group I and Group II mGluRs are expressed in neurons, these receptors have also been detected in glial cells such as astrocytes of both human and rat origin [73–81], where they demonstrate similar signaling properties as observed in neurons (i.e., increases in PI hydrolysis upon stimulation) . The precise function of mGlu5 receptors in astrocytes is not currently known but may have a role in regulating ion channel function , glutamate transporter expression , reactive gliosis [81, 84] and calcium signaling .
In the following section, an overview of mGluR ligands (mGluR1 and mGluR5 negative allosteric modulators, and mGluR2/3 agonists) that are commonly used in preclinical research on addiction and other CNS disorders is presented. The structures of these compounds are shown in Fig. 2, and their chemical names and abbreviations are presented in Table 2.
While the term “antagonist” has been widely used when applied to ligands that inhibit the function of specific mGluRs (i.e., mGluR1 and mGluR5), it is more appropriate to refer to such ligands as “negative allosteric modulators” (NAMs), since they inhibit the function of the receptor at a site that is distal to the actual orthosteric ligand binding domain of the receptor, and only in the presence of the orthosteric ligand [6, 8, 11, 86–88]. Thus, many pharmacological inhibitors of Group I mGluRs are referred to in the literature as “antagonists”, while the term NAM is more technically correct. In addition, there is recent evidence to suggest that Group I NAMs may actually function as inverse agonists, since in cell-based assays they can inhibit the basal (constitutive) activity of Group I mGluRs in the absence of any orthosteric agonist or even when the glutamate binding domain has been removed or mutated [89–100]. Thus, it is possible that in near future the functional terminology if mGluR ligands may be changed yet again to reflect the inverse agonistic properties of these ligands.
One of the first selective mGluR1 NAMs to be developed was CPCCOEt [101, 102]. CPCCOEt exhibited an IC50 value of 6.5 µM for inhibition of glutamate-stimulated increases in intracellular calcium in cells expressing the human mGluR1b receptor, and did not displace [3H]glutamate binding, suggesting that it acted at an allosteric site separate from the glutamate-binding domain of the receptor. Another mGluR1 NAM, LY367385, was also characterized at approximately the same time . LY367385 exhibited an IC50 value of 8.8 µM for blockade of quisqualate-stimulated PI hydrolysis. Both CPCCOEt and LY367385 have been reported to be systemically active [104, 105].
Since the development of LY367385 and CPCCOEt, selective mGluR1 NAMs with greater potency that retain systemic bioavailability have been synthesized. One of these compounds is YM298198 , which has selective affinity for mGluR1 receptors in the low nanomolar range (Ki=19 nM, IC50=16 nM). Another mGluR1 NAM, JNJ16259685, has also been recently developed which selectively inhibits the activity of mGluR1 receptors at subnanomolar concentrations (Ki=0.34 nM, IC50 = 0.55 nM) [107, 108]. Remarkably, when administered subcutaneously, less than 0.1 mg/kg is needed to occupy central mGluR1 receptors.
In the late 1990s, researchers at SIBIA Neurosciences developed two compounds that were some of the first mGluR5-selective receptor NAMs to be characterized . These compounds, 6-methyl-2-(phenylazo)-3-pyrindol (SIB-1757) and (E)-2-methyl-6-(2-phenylethenyl)pyridine (SIB-1893) were originally isolated by high throughput screening assays for their ability to alter basal or glutamate-stimulated IP3 accumulation in various cell lines transfected with human cDNA for each of the 8 mGluR receptor subtypes. Since SIB-1757 and SIB-1893 had IC50 concentration in the low micromolar range, researchers began to modify their structure with hopes of finding more potent mGluR5 NAMs. In 1999, Gasparini and colleagues published the structure and initial characterization of 2-methyl-6-(phenylethynyl)-pyridine (MPEP) , which exhibited relative high solubility in aqueous solutions and good brain penetrance and potency. MPEP inhibits quisqualate-stimulated PI hydrolysis at an IC50 value of 36 nM (100 times less than that of SIB-1757 and SIB-1893) and displays no agonist or antagonist activity at other mGluR receptor subtypes at concentrations up to 30 µM. After peripheral administration in rats, MPEP dose-dependently occupies mGluR5 receptors in the brain (as measured by displacement of the selective mGluR5 radioligand [3H]methoxymethyl-MTEP) within 5 minutes of administration, with maximal occupation observed at doses of 10 mg/kg i.p or higher . In addition, >60% mGluR5 occupancy is observed 2 hr after peripheral administration [111, 112], indicating that its effects are somewhat long-lasting, although shorter durations of receptor occupancy have been observed in mice as compared to rats .
Unfortunately, several reports have been published that have brought the selectivity of MPEP for mGluR5 receptors into question. First, when applied to cultured rat cortical neurons at concentrations of 20 and 200 µM, O’Leary and colleagues [113, 114] found that MPEP significantly reduced NMDA receptor-mediated neurotoxicity via a direct action on NMDA receptors. Consistent with this, it was reported that MPEP displays antagonistic activity in cells expressing the NMDA receptor containing NR2B subunits at an IC50 value of 18 µM . This study also demonstrated that MPEP can inhibit monoamine oxidase A (MAO-A) activity at an IC50 value of 8 µM . Non-specific agonist-like effects of MPEP and SIB-1893 have also been reported in cells expressing recombinant human mGlu4 receptors . A recent report also demonstrated that MPEP displaced binding of [3H]nisoxetine (a selective norepinephrine transporter (NET) inhibitor) and inhibited norepinephrine uptake with in LLCPK cells transfected with various human monoamine transporters . However, as in the aforementioned studies, only concentrations of MPEP in the micromolar range exerted these effects or norepinephrine uptake. In addition, these findings have not been replicated in neuronal preparations. Nonetheless, this latter study by Heidbreder and colleagues also demonstrated that intraperitoneal (i.p.) administration of MPEP (30 mg/kg) increased extracellular levels of norepinephrine while decreasing firing rates of locus coeruleus neurons. These effects are intriguing, as they provide a neurochemical basis by which MPEP produces antidepressant-like effects in rodents [5, 118, 119]; however, it remains to be determined if MPEP-induced elevations of extracellular norepinephrine levels are mediated by a direct action of MPEP on neuronal NETs. Taken together, it appears that, like most neurotransmitter receptor antagonists, the selectivity of MPEP for mGluR5 appears to diminish at concentrations that far exceed its IC50 value (36 nM).
In an effort to reduce the off-target actions of MPEP and improve its potency and selectivity for mGluR5 receptors, Cosford and colleagues synthesized a series of MPEP analogues , one of which was 3-[(2-methyl-1,3-thiazol-4-yl)ethynyl]pyridine (MTEP). MTEP appears to be about twice as potent as MPEP in displacing radioligand binding from mGluR5 receptors in rat cortical membranes (Ki = 16 nM vs. 32 nM for MPEP). In addition, MTEP retains selectivity for mGluR5 receptors over mGluR1 receptors at concentrations up to 100 µM, and unlike MPEP, displays no antagonist activity at NR2Bcontaining NMDA receptors at concentrations up to 300 µM . However, MTEP does exert some antagonist activity at MAO-A with an IC50 value of 30 µM. In vivo experiments have shown that the ability of MTEP (16 mg/kg s.c.) to reduce anxiety-like behaviors is absent in mice lacking mGluR5 receptors , providing further evidence that MTEP is indeed selective for mGluR5. Other in vivo studies have shown that MTEP reaches CSF concentrations in the brain following a systemic oral administration of 30 mg/kg that are 5 times higher than those obtained after the same dose of MPEP (1 µM for MTEP vs. 0.2 µM for MPEP), although concentration in individual brain structures such as the hippocampus were not different . These data suggest that MTEP has either greater bioavailability after oral administration and/or greater solubility in CSF than MPEP. MTEP also appears to be more potent than MPEP at occupying mGluR5 receptors in vivo, with a dose of 10 mg/kg i.p. MPEP needed to occupy >75% of brain mGluR5 receptors, as opposed to only 3 mg/kg i.p. needed for MTEP . As is the case for MPEP, MTEP appears to be shorter-acting in mice than in rats, with occupancy of >75% of brain mGluR5 receptors being observed for 2 hrs at the aforementioned systemic doses in rats, while this level of receptor occupancy only lasted 15–30 min for these compounds in mice . Thus, MTEP appears to be a more potent, selective and orally bioavailable mGluR5 NAM than MPEP.
In the early 1980’s, a clinical report was published indicating that fenobam, a nonbenzodiazepine compound whose mechanism of action was unknown at the time, demonstrated clinical efficacy in reducing symptoms of anxiety in humans in the early 1980’s [121–123]. However, due to reports of adverse side effects by some patients, further clinical testing was discontinued. However, recently it has come to light that fenobam is actually a selective mGluR5 antagonist with moderate potency and efficacy (Ki=295 nM, IC50=134 nM) .
A novel family of benzaldazine compounds has been recently described that possesses distinct allosteric modulatory effects on mGluR5 receptor activity . One of these compounds is (3-methoxyphenyl)methylene]hydrazone-3-methoxybenzaldehyde (DMeOB), an mGluR5 NAM that inhibits glutamate-stimulated Ca2+ in mGluR5 transfected cells with an IC50 value of 3 µM without altering [3H]quisqualate binding to the glutamate binding domain of the receptor. The structure of this compound is shown in Fig. (2).
Most synthetic mGluR2/3 agonists that are used in preclinical addiction research are conformationally constrained analogues of glutamate that act at the glutamate binding site on the receptor. One of the first mGluR2/3 agonists to developed was LY354740 [125, 126], which exhibits an EC50 value of approx 5 and 24 nM in inhibiting forskolin-stimulated cAMP production in cells expressing human mGluR2 and mGluR3, respectively. Later, LY389795 and LY379268 were developed , which displayed slightly increased potency as compared to LY354740 in inhibiting forskolin-stimulated cAMP formation (EC50 values approximately 3 and 5 nM at human mGluR2 and mGluR3 receptors, respectively, for LY379268; EC50 values approximately 4 and 8 nM at human mGluR2 and mGluR3 receptors, respectively, for LY389795). More recently, the synthesis and in vitro and in vivo characterization of LY404039 were published [128, 129]. The potency of LY404039 for inhibiting cAMP formation is similar to that observed for LY354740. All of these compounds exhibit excellent brain penetrance and selectivity for mGluR2/3 over other mGluR subtypes.
In the latter half of the 20th century, numerous animal models of drug addiction were developed and have proved invaluable in elucidating the neural substrates of addictive behaviors. One of the most widely used methods to study drug addiction in animals is the intravenous self-administration paradigm. In this paradigm, an animal is trained to perform an operant task such as a nose poke or lever press in order to receive an intravenous infusion of a drug solution via an indwelling venous catheter. In the case of ethanol, each operant response results in the presentation of a dilute ethanol solution (usually 8–12% v/v) in a receptacle located near the lever or nose-poke orifice, where the animal can consume it orally. In the progressive ratio paradigm of drug self-administration, the behavioral demand (i.e., number of lever presses) placed on the animal in order for it to receive a single drug reinforcer is gradually increased until the animal stops responding (called the “breakpoint”). Many researchers use environmental cues such as stimulus lights or auditory tones that are paired with the drug presentation or delivery to allow for the formation of Pavlovian drug-stimulus associations, which can subsequently be used in second-order schedules of reinforcement (i.e., responding for the drug-associated stimulus in the absence of primary drug reinforcement) as well as in reinstatement of drug-seeking behavior, a widely used model of relapse. In the reinstatement model, following stabilization of drug self-administration patterns, the animal is subjection to extinction procedures where the drug is no longer available. After extinction criteria have been met, the animal is exposed to drug-associated cues, drug priming, or stressors, which generally cause a reinstatement of operant responding that previously resulted in drug delivery.
Another widely used animal model of drug addiction is the conditioned place preference (CPP) paradigm, where passively administration of a drug is repeatedly paired with a unique contextual environment, and over time the animal exhibits a preference for the drug-paired environment over an environment that has been paired with a neutral pharmacological stimulus (i.e., saline). The CPP paradigm has provided substantial information on the neural substrates of the rewarding effects of drugs of abuse [130–132]. A disadvantage of the CPP technique, however, is that it does not directly measure drug-seeking behavior, but rather the motivation for secondary reinforcers (i.e., drug-associated environments).
A third experimental paradigm used to model drug addiction in animals the sensitization paradigm, whereby repeated passive administration of a drug of abuse results in a progressive and long-lasting increase in the behavioral response to the drug (i.e., increased motor hyperactivity in response to repeated exposure to psychostimulants). It has been theorized that this increased behavioral response reflects an increase in the incentive salience of the drug [133, 134]. However, given that most human drug abusers develop a significant degree of tolerance to the effects of the abused substance, the relevance of the locomotor sensitization to drug addiction in humans is often subject to debate.
Although high concentrations of mGluR1 receptors and not typically found in forebrain and limbic regions known to be involved in addictive behaviors (see Table 1), a handful of studies have shown significant effects of selective mGluR1 NAMs on drug- or ethanol-related behaviors. For example, in rats trained to self-administer nicotine intravenously, it was shown that the selective mGluR1 NAM EMQMCM at a dose of 5 mg/kg inhibited cue and nicotine-induced reinstatement of nicotine-seeking behavior . However, higher doses of this compound were shown to inhibit cue-induced reinstatement of food-seeking, indicating that this compound at higher doses may have general inhibitory effects on appetitive responding. Nonetheless, these findings demonstrate a potential role for mGluR1 receptors in relapse-like behaviors related to nicotine addiction, which are especially important in light of findings that nicotine exposure up-regulates the expression of mGluR1 mRNA in addiction-related brain regions such as the ventral tegmental area (VTA) and amygdala . There also reports that EMQMCM inhibits the expression of behavioral (i.e., locomotor) sensitization to both morphine and cocaine [137, 138]. Clearly, additional studies are needed to further characterize the ability of selective mGluR1 NAMs on the drug reward, reinforcement and relapse with respect to other abused substances such as cocaine, methamphetamine and heroin.
Of the few studies that have been published on the effects of mGluR1 antagonists on ethanol-related behaviors, there exists some degree of controversy. Hodge and colleagues showed that the mGluR1 NAM CPCCOEt at doses up to 10 mg/kg i.p. had no effect on oral ethanol self-administration in ethanol-preferring P rats and C57BL/6J mice [139, 140]. In contras., Lominac et al.  showed that in C57BL/6J mice, similar doses of CPCCOEt reduced ethanol reinforcement, consumption, and expression of ethanol CPP while facilitating the motor-impairing effects of ethanol. Lominac and colleagues also showed that CPCCOEt suppressed acute ethanol-stimulated increases in extracellular levels of dopamine and glutamate in the nucleus accumbens, while potentiating the effects of acute ethanol on extracellular GABA in this region. The reasons underlying the discordant effects of CPCCOEt on ethanol reinforcement and consumption between these reports are currently unclear, but are likely due to numerous procedural differences between these studies, as discussed in . More recently, Hodge and colleagues have demonstrated that the highly potent mGluR1 NAM JNJ16259685 reduces operant ethanol self-administration at doses of 0.1, 0.3 and 1 mg/kg i.p.  as well as breakpoints for ethanol reinforcement in a progressive ratio paradigm . However, the high dose of JNJ16259685 also reduced sucrose-reinforced responding, and the authors noted that this compound also decreased horizontal locomotor activity . This, motor side effects of this particular compound may have contributed to the reductions in ethanol self-administration and breakpoints for ethanol reinforcement.
Perhaps the most influential publication suggesting a role of mGluR5 receptors in addiction was published by Chiamulera and colleagues in 2001 . These investigators demonstrated that mice carrying a targeted deletion of the mGluR5 receptor gene do not demonstrate hyperlocomotion in response to acute administration of various doses of cocaine (10, 20 or 40 mg/kg i.p.), and fail to acquire intravenous self-administration of cocaine. This phenomenon was not due to a deficit the ability to learn an operant task, as lever pressing for food was unaltered in these mice. These investigators further confirmed a role for mGluR5 receptors in cocaine self-administration by demonstrating that intravenous administration of MPEP also dose-dependently reduced cocaine self-administration, without producing non-specific motor effects (as indicated by a lack of alteration in the rate of lever pressing). It was also shown that mGluR5-deficient mice exhibited normal increases in extracellular dopamine levels in the nucleus accumbens following an acute injection of cocaine, and that radioligand binding to the dopamine transporter and D1 and D2 dopamine receptors was unaltered in the brains of these mice, all of which indicate normal mesolimbic dopamine function. This paper provided the first compelling evidence that either genetic or pharmacological inhibition of mGluR5 receptor function reduces cocaine self-administration as well as cocaine-induced hyperlocomotion, suggesting that inhibition of mGluR5 function may prove useful in treating cocaine addiction in humans. More recent studies have shown that mice lacking mGluR5 also exhibit reduced ethanol consumption, but in contrast to cocaine, exhibit increased behavioral sensitivity to ethanol [144, 145].
Using pharmacological methods, others have demonstrated that mGluR5 NAMs such as MPEP and MTEP attenuate intravenous self-administration of cocaine, nicotine, and heroin [146–154] and reduce voluntary ethanol consumption in a variety of rodent strains [105, 140, 142, 155–159] without altering food reinforcement [146, 148, 149, 152]. MPEP also reduces breakpoints for cocaine, nicotine and ethanol reinforcement under progressive ratio schedules of reinforcement [142, 160]; however, attenuation of breakpoints for food reinforcement by MPEP have also been reported , suggesting that MPEP may also reduce the reinforcing efficacy of natural rewards under conditions of increased behavioral demands. MPEP has also been shown to prevent reinstatement of cocaine, nicotine or ethanol-seeking behavior induced by acute drug priming [148, 151] or drug-associated environmental cues [149, 155, 161, 162]. MPEP also enhances the sedative/hypnotic properties of ethanol , and mGluR5 receptors appear to interact with adenosine A2A receptors in their ability to regulate ethanol self-administration . A recent microinjection study revealed that local infusion of MPEP into the nucleus accumbens core also attenuates cue-induced reinstatement of cocaine-seeking behavior .
MPEP at doses up to 9 mg/kg has also been shown to elevated thresholds for intracranial electrical self-stimulation [150, 166], indicating that negative allosteric modulation of mGluR5 receptors negatively influences brain reward function, which may underlie the ability of mGluR5 antagonists to suppress active drug self-administration. However, evidence against this hypothesis comes from reports that MPEP does not alter nicotine- or cocaine-induced lowering of brain reward stimulation thresholds [147, 166] or food reinforcement. An alternative explanation for the ability of mGluR5 NAMs to reduce drug or ethanol self-administration is that mGluR5 NAMs alter the discriminative stimulus (i.e., subjective) effects of various drugs of abuse, as has been demonstrated for cocaine , nicotine  and ethanol [168, 169].
Consistent with the findings of Chiamulera and colleagues , data from our laboratory have confirmed that pharmacological inhibition of mGluR5 receptor function with MPEP can reduce or completely prevent the hyperlocomotion produced by acute administration of cocaine or D-amphetamine in DBA/2J mice . These findings were later replicated by another group of investigators in rats . However, reductive effects of MPEP on amphetamine-induced locomotor activity were not observed in a different mouse strain (OF1/IC) after oral administration of similar doses of MPEP , suggesting that pharmacokinetic, bioavailability and/or genetic factors may influence the effects of mGluR5 NAMs on amphetamine-induced hyperlocomotion. Despite the findings that MPEP attenuates acute cocaine-stimulated hyperlocomotion, it has been shown that MTEP (at doses up to 10 mg/kg i.p.) does not prevent the expression of sensitization to the locomotor stimulant properties of repeatedly administered cocaine , but does prevent the expression of locomotor sensitization to morphine . Thus, neuroadaptations produced by repeated cocaine exposure may eliminate the ability of mGluR5 NAMs to attenuate the behavioral sensitizing effects of cocaine.
Work from our laboratory has also demonstrated that MPEP, at doses up to 20 mg/kg i.p., prevents the development of cocaine CPP in mice , suggesting that mGluR5 receptors are also involved in the acute conditioned rewarding effects of cocaine. When used as the conditioning drug, our group and others have found that MPEP alone does not produce CPP or CPA in mice or rats [173–175], indicating that MPEP does not possess intrinsic rewarding or aversive effects. However, we found that MPEP did not inhibit the development of CPP to other drugs of abuse such as amphetamine, morphine, nicotine or ethanol , suggesting that the ability of MPEP to inhibit the development of cocaine CPP was not due to the production of a learning deficit, resulting in an inability of the animal to associate the subjective effects of the drug with the conditioning environment. However, others have found that MPEP attenuates the expression of CPP for amphetamine , ethanol  and morphine [174, 175], but interestingly, not cocaine . Together, these data suggest that mGluR5 receptors are involved in some, but not all, aspects of the conditioned rewarding effects of various drugs of abuse. With respect to cocaine and ethanol, there appears to be a disconnect between the role of mGluR5 receptors in the development versus expression of drug CPP. These data underscore the notion that pharmacological manipulations of CPP can yield results that are highly divergent from those obtained using operant self-administration paradigms.
There are a handful of studies indicating that mGluR5 NAMs may be beneficial in the treatment of opiate withdrawal symptoms. Both MPEP and MTEP were shown to ameliorate the behavioral signs of withdrawal from morphine [138, 177, 178]. However MPEP was recently shown to actually exacerbate the somatic signs of nicotine withdrawal , and reduces handling-induced convulsions induced by ethanol withdrawal in female but not male mice [145, 179]. Thus, mGluR5 NAMs may be of potential therapeutic benefit in the treatment of symptoms of opiate withdrawal but may be of limited use in amelioration of nicotine or ethanol withdrawal symptoms.
Stimulation of inhibitory presynaptic mGluR2/3 autoreceptors, which results in decreased release of glutamate from the presynaptic terminal, is another pharmacological method to dampen glutamate transmission. One of the first demonstrations that mGluR2/3 agonists might possess therapeutic efficacy in the treatment of drug addiction was published in 1997 by Helton and colleagues, who showed that LY354740 suppressed behavioral signs of nicotine withdrawal in rats . However, it was not until several years later when it was demonstrated that mGluR2/3 agonists reduce active drug reinforcement and drug-seeking behavior. In 2004, Baptista and colleagues showed that LY379268 (at doses up to 3 mg/kg i.p.) suppressed self-administration of cocaine but not a conventional reinforcer (sweetened condensed milk) in rats . These effects on intravenous cocaine self-administration have subsequently been replicated by other investigators in both rats and squirrel monkeys [182, 183]. However, while Baptista et al. showed that LY379268 also attenuated cue-induced reinstatement of cocaine-seeking behavior, this compound also suppressed reinstatement of seeking of the conventional reinforcer. This study suggests that LY379268 has selective effects on cocaine reinforcement but its effects on natural reinforcers become apparent under cue-induced reinstatement conditions.
It has since been shown that mGluR2/3 agonists also attenuate the reinforcing effects of nicotine and ethanol [184, 185] as well as cue-, stress- or drug-primed reinstatement [183, 184, 186]. Interestingly, however, Rodd and colleagues  demonstrated that a different mGluR2/3 agonist, LY404039, suppressed cue-induced reinstatement of ethanol-seeking behavior but not active ethanol self-administration. The reason for the lack of effect on LY404039 on active ethanol self-administration are unclear, but may be due to the use of a high ethanol-preferring P rat strain as opposed to outbred rats used in other studies. Systemic administration of LY379268 has also been shown to attenuate the phenomenon of “incubation” of cocaine craving , which is characterized by a progressive increase in the magnitude of cue-induced reinstatement over time following extinction [189, 190].
Some of the aforementioned effects of mGluR2/3 agonists on drug and ethanol reinforcement or reinstatement must be interpreted with caution, however, since higher doses (3–5 mg/kg i.p.) of LY379268 have been shown to decrease spontaneous locomotor activity , food reinforcement  but see  as well as responding for conventional reinforcers such as sweetened solutions .
Studies imploying intracerebral microinjection techniques have revealed some of the neural circuitry involved in regulation of drug-reinforcement or reinstatement of drug-seeking behavior by Group II mGluRs. Infusion of LY379268 into the ventral tegmental area attenuates nicotine self-administration  as well as contextual cue-induced reinstatement of heroin-seeking . Infusion of this ligand into the nucleus accumbens core attenuates cocaine, nicotine, and food reinforcement [182, 185], while infusion into the nucleus accumbens core or shell reduces context-induced reinstatement of heroin seeking without affecting sucrose seeking . Finally, it was recently demonstrated that infusions of LY379268 into the central nucleus of the amygdala attenuates the “incubation” of cocaine craving . Thus, not surprisingly, regions of the extended amygdala and mesolimbic reward circuitry are critical for regulation of drug-seeking and relapse-like behaviors by mGluR2/3 agonists.
As depicted in Fig. 1, Group I mGluR receptors, particular mGluR5, are structurally linked to NMDA receptors by various scaffolding proteins (such as Homer, Shank and GKAP), and positively modulate NMDA receptor function (reviewed elsewhere in [48, 193, 194]). In accord with this well-established positive coupling between Group I mGluRs and NMDA receptors, there is a wealth of evidence suggesting that these receptors play an important role in NMDA-dependent synaptic plasticity and learning and memory processes [33, 195]. As a result, numerous studies have shown that pharmacological inhibition of Group I mGluR function or genetic deletion of the mGluR1 or mGluR5 receptor gene produces performance decrements in various learning and memory tasks in rodents such as the Morris water maze [196, 197], radial arm maze [198–203], 5-choice serial reaction time task [204, 205], contextual and auditory conditioning paradigms [196, 206–211], spatial alternation tasks [212, 213], and passive/inhibitory avoidance tasks [214–217]. However, there are reports that inhibition of mGluR5 function has no effect on performance in a spatial reward-finding task [174, 218], the acquisition of food reinforcement  or in the development of the conditioned rewarding effects of morphine, ethanol, nicotine or amphetamine in the CPP paradigm . Thus, inhibition of Group I mGluR receptor function does not eliminate all forms of learning and memory, and many observed effects may be dependent on the type of paradigm used or the particular phase of the task examined (acquisition, expression, consolidation, etc.).
Along these lines, some investigators have reported that mGluR1 NAMs actually enhance performance on some learning or short-term memory or conditioning tasks [219–221]. For example, EMQMCM has been shown to improve working memory and reduce impulsive choice-making in rodents , providing support for the notion that mGluR1 NAMs may be potential cognitive enhancing agents. However, given the number of other studies showing deleterious effects of mGluR5 NAMs on learning and memory tasks in rodents, it is possible that such ligands, if ever brought to clinical testing in humans for the treatment of addiction, may cause dose-related unwanted side effects such as amnesia.
The data from the animal studies reviewed in this article strongly suggest that attenuation of glutamatergic neurotransmission by Group I mGluR NAMs or Group II mGluR agonists may be novel and effective therapeutics in the treatment of addiction to cocaine, opiates, nicotine, or ethanol. Perhaps one of the most attractive aspects of these ligands is their apparent beneficial effects in preclinical models of anxiety, depression and chronic pain, conditions that are often co-morbid with substance use disorders.
However, to date there are no clinical studies on the efficacy of Group I mGluR NAMs or Group II mGluR agonists in treating addictive disorders. This lack of clinical data makes it difficult to speculate what the precise therapeutic scope of use of such ligands would antagonists might be in human drug addicts. Would such ligands be most efficacious in primarily reducing basal drug intake or the more problematic phenomenon of relapse? If the latter is the case, will such ligands be most effective in preventing relapse triggered by drug-associated environmental cues, stress, the drug itself, or spontaneous relapse? Will mGluR ligands reduce the incidence drug craving, which is primarily a subjective measure that is difficult to model in animals? Could such drugs be used in in-patient populations during acute detoxification as well as in outpatient populations with psychosocial support networks? Are there any demographic characteristics, biomarkers, or pharmacogenetic screens that might be used to predict responsiveness to these mGluR ligands?
The answers to these questions may be generated sometime in the near future. Addex Pharmaceuticals (Geneva, Switzerland) has recently completed Phase IIa clinical trials on its mGluR5 NAM ADX10059 for the treatment of migraine, anxiety and gastroesophageal reflux disease, and is currently conducting Phase I clinical trials on another mGluR5 NAM, ADX48621, for depression and anxiety . Similarly, Eli Lilly (Indianpolis, IN) has conducted clinical studies with positive outcomes on its mGluR2/3 agonist LY354740 in the treatment of anxiety [224, 225] and its LY404309 pro-drug LY2140023 in the treatment of schizophrenia . Should these compounds eventually gain approval by the Food and Drug Administration, there is a strong scientifically-based rationale for testing these ligands in the treatment of drug addiction and alcoholism.
Other unanswered questions about the potential use of mGluR ligands in the treatment addiction pertain to their neurobiological actions. Does the mere acute suppression of glutamate transmission by inhibition of postsynaptic Group I mGluRs or stimulation of presynaptic mGluR2/3 autoreceptors mediate the ability of mGluR ligands to attenuate drug reinforcement and relapse? Or are there neuroadaptations at the cellular and molecular levels (i.e., changes in intracellular enzyme activity, gene expression, or synaptic plasticity) induced by these compounds that underlie their potential efficacy? For example, while numerous studies have shown that mGluR5 NAMs reduce ethanol consumption, data from the author’s laboratory has shown that the epsilon isoform of protein kinase C is an important downstream signaling target that is required for the ability of mGluR5 NAMs to reduce ethanol intake . Further research is needed to identify cellular and molecular signaling components that may be essential for the potential anti-addictive properties of Group I NAMs and Group II agonists. In addition, the neuroanatomical loci where these drugs act to produce their potential beneficial effects have only partially been characterized and found to be located within the mesolimbic dopamine reward circuitry [165, 182, 191, 192]. Additional neuroanatomical mapping of the effects of these mGluR ligands on drug self-administration and reinstatement is needed to further characterize the neural circuitry involved in regulation of addictive behaviors by mGluR ligands.
M.F.O. is supported by a Public Health Service grants AA013852 from the National Institute on Alcohol Abuse and Alcoholism and DA024355 from the National Institute on Drug Abuse. The author has no conflicts of interest to declare.