|Home | About | Journals | Submit | Contact Us | Français|
The importance of glutamate in the reinstatement of cocaine-seeking behavior has been established. New molecular and neurochemical adaptations in the glutamatergic system which drive cocaine relapse have been identified, such as the ability of CB1 receptor stimulation to reduce basal glutamate levels and the involvement of the GluR1 receptor subunit in reinstatement. Furthermore, it is apparent that similar glutamatergic neuroadaptations arise after self-administration of cocaine, heroin, nicotine and alcohol. For example, reinstatement to cocaine, nicotine and alcohol can be prevented both by stimulation of group II mGluR receptors as well as blockade of Group I mGluR receptors. The similarities in the neurochemistry behind relapse to these varied drug classes indicate that drugs that target the glutamate system could be effective at treating relapse to multiple types of drugs.
There is an extraordinarily high rate of recidivism for drug use even after long periods of abstinence have been attained . The reinstatement model of relapse has been developed as a means of studying relapse in experimental animals [2, 3]. Animals trained to self-administer drug are subsequently put through extinction training to reduce responding. When levels of responding are minimal, animals are presented with a stimulus that causes a “reinstatement” of the drug-contingent response. This pre-clinical model has been shown to have face validity for human relapse, based partially on the fact that the same factors that elicit relapse in the human population (drug priming injection, drug-associated cues, and stress) precipitate relapse in the reinstatement model (see review ). The reinstatement model has been used to identify brain structures involved in relapse and, although highly effective pharmacotherapeutic treatments have yet to emerge, this model has been judged a moderately effective tool for screening drugs to block relapse . One of the major discoveries to emerge from the use of this model is the role of glutamate release in the nucleus accumbens (NAC) core in the reinstatement of cocaine-seeking [5, 6, 7, 8]. The discovery that the infusion of AMPA agonists into the NAC core causes reinstatement  and that antagonism of AMPA receptors here prevented cocaine-  and cue-primed relapse  led to the identification of the PFC-NAC core glutamatergic projection as being essential for cocaine-primed reinstatement , and later for stress-induced reinstatement . A decrease in basal levels of glutamate was found to be present following cocaine self-administration accompanied by an enhanced release of glutamate during reinstatement .
In recent years, the involvement of glutamate in cue-, stress-, and drug-primed reinstatement has been further investigated in regards to relapse to cocaine as well as other drugs of abuse. The present review will focus on research conducted within the past two years that further investigates the nature of glutamate signaling and homeostasis in relapse to cocaine. Importantly, the last few years has seen the well-characterized role of glutamate in cocaine-seeking expanded to other drugs of abuse, and the role of glutamate in the relapse of heroin, alcohol, and nicotine will also be discussed. Based on this body of research, it follows that neuroplasticity in glutamate transmission may be the common factor in relapse for many types of drugs.
In the NAC, basal levels of extracellular glutamate are primarily regulated by the cystine-glutamate antiporter (system xc-), which exchanges extracellular cystine for intracellular glutamate . System xc- is down-regulated after repeated cocaine, accounting for the low basal levels of glutamate observed in the NAC of cocaine-withdrawn animals . N-acetylcysteine restores basal glutamate levels and prevents both cocaine-primed reinstatement as well as the increase in extracellular glutamate that accompanies reinstatement both when it is given acutely prior to reinstatement testing  and when it is only given during the self-administration portion of the experiment . These effects were further shown to be directly due to action of N-acetylcysteine on system xc- through use of CPG, the system xc- antagonist . In the reinstatement model, the mGluR2/3 antagonist LY341495 prevents N-acetylcysteine from inhibiting reinstatement of cocaine-seeking . Thus, it appears that increasing basal glutamate levels via NAC restores tone on the mGluR2/3 autoreceptors and dampens glutamate release. Double-blind pilot clinical trials found that N-acetylcysteine also decreases reactivity for cocaine cues  and cocaine use in cocaine-dependent humans .
The cannabinoid CB1 receptor is known to be a negative modulator of synaptic glutamate release  and the CB1 antagonist AM251 decreases cocaine-primed reinstatement while also inhibiting the increase in extracellular glutamate that accompanies reinstatement of cocaine-seeking . AM251 alone significantly increased extracellular glutamate in the NAC of rats that had self-administered cocaine and blocking mGluR2/3 receptors prevented the AM251-induced attenuation of cocaine-seeking. Thus, AM251 may be acting in a similar manner as N-acetylcysteine: elevating extracellular nonsynaptic glutamate to stimulate presynaptic inhibitory mGluR2/3 receptors and reduce synaptic glutamate release.
Work done in the past few years has focused on elucidating the roles of the different glutamate receptor subtypes in reinstatement. Building on the early work by Cornish & Kalivas [5,9], it has been recently demonstrated that infusion of AMPA into both the core and the shell of the NAC induces reinstatement behavior . These authors also demonstrated that the knockdown of the AMPA receptor subunit GluR1 mRNA via antisense oligionucleotides blocked both AMPA- and cocaine-primed reinstatement, further solidifying the role of AMPA receptors in mediating reinstatement. Additionally, Backstrom & Hyytia found that systemic administration of two AMPA antagonists, CNQX and NBQX, was able to attenuate cue-induced reinstatement . CNQX was also able to attenuate cue-induced reinstatement when administered directly into the NAC core . However, a recent report by Bachtell et al. indicates the relationship of AMPA receptors to reinstatement may not be so straightforward . These authors over-expressed either a wild type (wt) or a pore-dead (pd) GluR1 subunit immediately prior to acquisition of cocaine self-administration or one day preceding a cue- or cocaine-primed reinstatement test. Over-expression of the wt-GluR1 was found to inhibit cocaine-primed reinstatement when delivered both prior to acquisition and the reinstatement test. Conversely, the over-expression of the pd-GluR1, which serves to reduce AMPA currents, enhanced cocaine-primed reinstatement when delivered at both time points tested. There was no effect on cue-induced reinstatement. The authors explain the discrepancies between these results and the evidence showing that reducing, and not facilitating, signaling through AMPA receptors blocks reinstatement [5, 20], with the idea that they have altered basal AMPA receptor function, versus the acute antagonism used by previous studies.
Acute blockade of AMPA receptors seems to block reinstatement [5,20,24] but blockade of NMDA receptors may enhance reinstatement. When the NMDA antagonist AP-5 was injected into the shell of the NAC, both low (3 μg) and high (30 μg) doses promoted the reinstatement of cocaine seeking-behavior. However, in the NAC core, only the high dose was able to produce reinstatement, the magnitude of which was substantially lower than that produced when either dose was injected into the shell . Contradictory reports have been made by Backstom & Hyytia [21, 22]. Systemic administration of the NMDA/glycine site antagonist L-701,324 was found to decrease cue-induced cocaine reinstatement while the competitive NMDA antagonist CGP-39551 had no effect on this measure [21, 22]. When AP-5 was injected directly into the NAC core, doses of 1 and 2 ug were capable of attenuating cue-induced reinstatement [21,22]. Thus, it is possible that low doses of AP-5 administered into the NAC core are capable of blocking reinstatement while high doses induce drug-seeking, possibly due to non-specific affects on other receptor sub-types. In the basolateral amygdala (BLA), NMDA receptors seem to play a role in the consolidation of associations between drug and conditioned stimuli which, when strengthened, can produce relapse .
The mGluR2/3 autoreceptors are capable of regulating glutamate release  and since there is an increase in glutamate levels during reinstatement, it follows that agonists at mGluR2/3 receptors inhibit reinstatement. Both systemic and intra-NAC LY379268 decreased cocaine-primed reinstatement, however, reinstatement of food-seeking was also attenuated [27,28]. In contrast, Bossert et al. found that intra-NAC LY379268 had no effect on reinstatement of sucrose-seeking [29, 30]. Stimulation of mGluR2/3 receptors with LY379268 reduces cocaine-primed reinstatement in non-human primate, without nonspecifically altering behaviors such as locomotion, grooming, vocalization and posture . Adewale and colleagues also found that the relapse-prevention effects of LY379268 could be blocked with the mGluR2/3 antagonist, LY341495, indicating a specific action at mGluR2/3 receptors. Post-synaptic mGluR5 receptors have also been targets for drugs to treat cocaine addiction. MPEP, a selective mGluR5 antagonist, attenuates cue-induced reinstatement of cocaine-seeking . The mGluR5 antagonist MTEP also blocks cue-induced reinstatement of cocaine-seeking .
In addition to recent advances in glutamate pharmacology and cocaine-seeking, evidence that glutamate neurotransmission in the NAC is altered by chronic cocaine administration has accumulated in the last two years. Thus, glutamatergic synapses appear to be in an LTP-like state as estimated by the ratio of AMPA to NMDA currents [32, 33]. This is associated with an increase in surface expression of the GluR1 AMPA receptor subunit [32, 33] and GluR1 surface expression is further elevated at 30 min following cocaine-induced relapse . In spite of NAC neurons being in an LTP-like state, the capacity to induce LTD in slices obtained from animals withdrawn from cocaine self-administration is blunted . Exactly how these physiological changes and loss of plasticity affect glutamate pharmacology requires further study, however, the upregulation of GluR1 is consistent with the increased locomotor response elicited by AMPA microinjection into the NAC of mice pretreated with chronic cocaine .
Glutamate has been shown to be important for context-induced reinstatement of heroin-seeking. Infusing the mGluR2/3 agonist LY379268 (0.3 or 1.0 μg) into the NAC shell dose-dependently attenuated reinstatement while these doses when injected into the NAC core had no effect. However, a higher dose (3.0 μg) of LY379268 infused into the core was able to block reinstatement . These results are in agreement with the finding that both the NAC core and shell are part of the circuitry of heroin relapse . Microdialysis sampling of glutamate levels in the NAC core during both heroin- and cue-primed reinstatement indicates that glutamate levels rise during reinstatement and consequently, blocking AMPA/kainite receptors in the core attenuates heroin seeking primed by both cue and drug . Reinstatement of heroin-seeking and the accompanying rise in extracellular glutamate in the NAC are blocked by inhibiting prelimbic cortex projections to the NAC core  indicating that the same glutamatergic projection is involved in relapse to both cocaine- and heroin-seeking. In agreement with this finding is that N-acetylcysteine treatment blocks both heroin- and cue-primed reinstatement for up to one month following the last injection of N-acetylcysteine that was administered repeatedly during extinction training . Interestingly, extinction training was facilitated by the daily N-acetylcysteine injection.
In a similar manner to both heroin and cocaine, the mGluR2/3 agonist LY379268 dose-dependently attenuates reinstatement of alcohol. Systemically administered LY379268 blocked both stress- and cue-induced reinstatement of alcohol-seeking . Also in a similar manner to cocaine, antagonism of mGluR5 receptors prevents cue-induced reinstatement both when administered alone  and when sub-threshold doses are combined with an adenosine A2A antagonist . The combination of mGluR5 and A2A receptor agonists enhance glutamate release in the rat striatum , thus it follows that the combination of antagonists at these receptors decreases glutamate release. Cue-induced reinstatement of alcohol-seeking was recently linked to increased phosphorylated extracellular signal-regulated kinase (p-ERK1/2) expression in the NAC shell and BLA . The ERK1/2 pathway is downstream from mGluR5 receptors (as well as others such as NMDA) and the blockade of mGluR5 receptors with MPEP and subsequent attenuation of reinstatement of alcohol-seeking was associated with a decrease in p-ERK1/2 expression in both the shell and the BLA. Lamotrigine inhibits the release of glutamate , and systemic administration attenuates cue-induced reinstatement of alcohol seeking , further supporting a role for glutamate in relapse to alcohol-seeking.
Relapse to nicotine seeking is also influenced by glutamate release. Stimulation of mGluR2/3 receptors with systemically administered LY379268 blocks cue-induced nicotine relapse; however it also blocked food-seeking . Systemic administration of the mGluR1 antagonist EMQMCM (5 mg/kg) blocked both cue- and nicotine-primed reinstatement and only blocked food-seeking at a dose (10 mg/kg) much higher than the doses effective at blocking nicotine relapse . This finding is in agreement with previous work showing that mGluR5 antagonists block nicotine-  and cue-primed reinstatement . Nicotine self-administration may produce similar neuroadaptations in glutamate transmission akin to the effects of cocaine self-administration. Thus, nicotine self-administration reduces mGluR2/3 function and expression of the catalytic subunit of system xc- [47,51]. Accordingly, N-acetylcysteine was found to reduce the number of cigarettes smoked in a double-blind trial involving four weeks of drug administration .
Based on the work presented here, glutamate is involved in relapse to multiple types of addictive drugs (also see ). The majority of work in this field has been done with cocaine, thus providing more insight into the glutamatergic signaling mechanisms underlying cocaine relapse relative to other drugs. However, the research presented here indicates that there are commonalities between relapse to cocaine, alcohol, nicotine and heroin. Cocaine and heroin share a common circuitry of relapse, with the glutamatergic projection from the prelimbic cortex to the nucleus accumbens core being essential for reinstatement. Stimulation of mGluR2/3 receptors is effective at blocking reinstatement of cocaine, nicotine and alcohol, presumably by reducing the release of glutamate from pre-synaptic terminals and providing more evidence that it is the release of glutamate that drives relapse. Similarly, antagonizing group I mGluR receptors blocks reinstatement of cocaine-, nicotine- and alcohol-seeking. Given that different and overlapping brain structures contribute to reinstatement for different drug classes, the drug-induced glutamatergic neuroadaptations and effects of glutamate pharmacology on reinstatement that have been identified primarily in the NAC may generalize to other key structures, such as the amygdala or ventral tegmental area. Nonetheless, the last two years of research consistently points to an important role for glutamate transmission in the NAC in the reinstatement of drug-seeking, regardless of drug class, and correspondingly identifies different aspects of glutamate transmission and homeostasis as potential pharmacotherapeutic targets in treating addiction.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.