CPL

Bacgs

Mest

Rx

Ce

Kg

Rccrtaemwttommlsggt(

pemGg

F

A

R

0d

eftriaxone Restores Glutamate Homeostasis andrevents Relapse to Cocaine Seeking

ori A. Knackstedt, Roberto I. Melendez, and Peter W. Kalivas

ackground: The cystine-glutamate exchanger is downregulated after chronic cocaine, resulting in reduced extracellular levels of nucleusccumbens glutamate. The importance of cocaine-induced loss of glutamate homeostasis is revealed by N-acetylcysteine restoringystine-glutamate exchange and attenuating reinstatement to cocaine seeking. Another regulator of extracellular glutamate is the gliallutamate transporter GLT-1. We hypothesized that cocaine self-administration reduces GLT-1 and that GLT-1 upregulation inhibits cocaineeeking.

ethods: We measured [3H] glutamate uptake and protein expression of GLT-1 and xCT, the catalytic subunit of the cystine-glutamatexchanger, following cocaine self-administration and 3 weeks of extinction training. We also examined the affect of ceftriaxone (previouslyhown to increase GLT-1) and N-acetylcysteine treatment on the expression of GLT-1 and xCT. Ceftriaxone was also tested for the capacityo inhibit cue- and cocaine-induced relapse.

esults: Cocaine self-administration reduced glutamate uptake and the expression of both GLT-1 and xCT. Ceftriaxone restored GLT-1 andCT levels and prevented cue- and cocaine-induced reinstatement of drug seeking. N-acetylcysteine also restored GLT-1 and xCT levels.

onclusions: These results indicate that glutamate transport and cystine-glutamate exchange may be coregulated and provide furthervidence that targeting glutamate homeostasis is a potential method for treating cocaine relapse.

ey Words: Cocaine, accumbens, cystine-glutamate exchange,lutamate uptake, GLT-1, self-administration

elapse to cocaine seeking in the animal model of reinstate-ment is strongly associated with changes in the extracel-lular levels of glutamate in the nucleus accumbens (NAcc)

ore, namely decreased basal levels of glutamate in the NAccore as well as enhanced extracellular glutamate levels inesponse to a cocaine challenge (1). Glutamate is released intohe extracellular space from synaptic and nonsynaptic sourcesnd is eliminated through a family of glutamate uptake transport-rs (2). The balance between synaptic and nonsynaptic gluta-ate release and elimination is termed “glutamate homeostasis,”hich modulates synaptic activity and plasticity by controlling

he stimulation of iontotropic and metabotropic glutamate recep-ors (3,4). The cystine-glutamate exchanger, which exchangesne extracellular cystine molecule for one intracellular glutamateolecule (5), accounts for most nonsynaptic extracellular gluta-ate in the nucleus accumbens (6), and its activity is downregu-

ated after chronic cocaine administration (7,8). The nutritionalupplement N-acetylcysteine restores the function of the cystine-lutamate exchanger, increases the basal levels of extracellularlutamate in the accumbens after withdrawal from cocaine, andhereby attenuates reinstatement to cocaine seeking in animals3,7,8) and cocaine cue reactivity in humans (9).

In addition to the cystine-glutamate exchanger, sodium-de-endent glutamate transport into glia is a critical regulator ofxtracellular glutamate concentrations; this type of transport isaintained primarily by the major glial glutamate transporter,LT-1 (EAAT2) (2). GLT-1 is responsible for 90% of total brainlutamate uptake (10). Given the apparent imbalance in gluta-

rom the Department of Neurosciences, Medical University of South, Caro-lina, Charleston, South Carolina.

ddress correspondence to Lori Knackstedt, Ph.D., Department of Neuro-sciences, Medical University of South Carolina, 173 Ashley Avenue, BSB403, Charleston, SC 29425; E-mail: knackst@musc.edu.

eceived May 6, 2009; revised Jul 1, 2009; accepted Jul 19, 2009.

006-3223/10/$36.00oi:10.1016/j.biopsych.2009.07.018

mate homeostasis associated with withdrawal from cocaine self-administration, we postulated that cocaine also induces changes inglutamate transport. Recently, this hypothesis was supported bymathematical modeling of glutamate homeostasis at glutamater-gic synapses in the nucleus accumbens that predicted bothcystine-glutamate exchange and sodium-dependent uptake wouldbe reduced by cocaine self-administration (11). Thus, it is possiblethat these two mechanisms of glutamate transport, glutamate uptakeand cystine-glutamate exchange, are coregulated. Supporting thispossibility, Bannai (12) reported that cystine-glutamate exchange isstimulated by increased activity of sodium-dependent glutamatetransport. Additionally, downregulated transporter activity leavesexcess glutamate in the extrasynaptic space, which has beenfound to inhibit the cystine-glutamate exchanger from exportingglutamate (13). The catalytic subunit of the cystine-glutamateexchanger is the protein xCT (5), and nicotine self-administrationwas recently shown to reduce the levels of both xCT and GLT-1in the nucleus accumbens (14).

Here, we measured glutamate uptake and levels of GLT-1 andxCT protein expression in nucleus accumbens tissue of rats thathad self-administered cocaine and underwent 3 weeks of extinc-tion training. The �-lactam antibiotic ceftriaxone increases theexpression and activity of GLT-1 (15,16), and we tested its abilityto attenuate the reinstatement of cocaine seeking and elevateGLT-1 levels following cocaine self-administration. We investi-gated the possibility of coregulation of the two glutamate trans-port systems by measuring the effects of ceftriaxone treatment onxCT expression and, conversely, the effect of N-acetylcysteinetreatment on GLT-1 expression in the NAcc of rats withdrawnfrom cocaine self-administration.

Methods and Materials

Rats were trained to self-administer cocaine for 2 weeks andthen underwent 3 weeks of extinction training. Glutamate uptakewas measured in nucleus accumbens tissue (see Supplement 1for a detailed description of uptake methods). A separate set ofanimals received ceftriaxone (200 mg/kg IP), N-acetylcysteine

(100 mg/kg IP), or vehicle (saline) daily for the last 7 days of

BIOL PSYCHIATRY 2010;67:81–84© 2010 Society of Biological Psychiatry

eTgGbS

wptIcegftd

R

altcsw

82 BIOL PSYCHIATRY 2010;67:81–84 L.A. Knackstedt et al.

w

xtinction training and were euthanized via rapid decapitation.he NAcc was dissected, and a membrane subfraction wasenerated. Western blotting was performed on this tissue forLT-1 and xCT proteins (for detailed description of the mem-rane subfractionation and Western blotting procedures, seeupplement 1).

Animals self-administered cocaine for 2 hours/day for 2eeks followed by extinction training for 3 weeks and until leverressing reached 25% of self-administration levels. During ex-inction, animals were treated with either ceftriaxone (200 mg/kgP) or vehicle for 5 to 12 days and then tested for both cue- andocaine-primed reinstatement of the drug-seeking response. Tonsure that ceftriaxone was not producing sedation, a separateroup of animals was injected with ceftriaxone (200 mg/kg IP)or 7 days and then tested for basal locomotion and the locomo-or response to a saline injection (see Supplement 1 for detailedescription of the behavioral methods).

esults

Sodium-dependent glutamate uptake in tissue slices from theccumbens was reduced in cocaine-trained compared with sa-ine controls over a physiologic range of glutamate concentra-ions (1 �mol/L, 3 �mol/L, 5 �mol/L; Figure 1A). This wasonfirmed in a separate set of cocaine and saline subjectshowing that reduced uptake resulted from decreased Vmax,ithout a difference in binding affinity (Kd � 2.6 �mol/L � .9

ww.sobp.org/journal

�mol/L [cocaine], � 2.4 �mol/L � .9 �mol/L [saline]; Figure 1B).Consistent with decreased Vmax, we observed significant down-regulation of GLT-1 expression in the accumbens membranesubfraction following cocaine self-administration (Figure 1C) butnot in the prefrontal cortex (Figure 1F). In the same samples usedto measure GLT-1, reduced expression of xCT was also measuredin the accumbens (Figure 1D) but not prefrontal cortex (althoughthe level of xCT trended down in the cocaine animals; Figure 1F).Both N-acetylcysteine and ceftriaxone treatment alone upregu-lated the expression of both xCT and GLT-1 (Figure 1C–1E). Ofnote, this ceftriaxone treatment regimen did not upregulateGLT-1 expression in drug-naive rats (Figure 1G). Similarly,treatment with N-acetylcysteine failed to upregulate GLT-1 indrug-naive rats (Figure 1H).

There was no difference in the number of cocaine infusionsattained over the course of the self-administration period for theceftriaxone- and vehicle-treated subjects (Figure 2A; for inactivelever presses see Figure S1C in Supplement 1); likewise, self-administration levels were no different between groups used forthe uptake assay, Western blots, or reinstatement tests (FigureS1A and S1B in Supplement 1). Ceftriaxone treatment had noeffect on extinction levels of lever pressing (Figure 2B); however,the lever-pressing behavior was largely extinguished by Day 5,when ceftriaxone administration began. Ceftriaxone treatmentsignificantly attenuated both cue- and cocaine-primed reinstate-ment (Figure 2C). Importantly, ceftriaxone potentiated sponta-

Figure 1. Cocaine self-administration reduces glutamatetransporter expression and function. (A) 3H glutamate up-take was significantly decreased in accumbens tissue slicesfollowing cocaine self-administration. A two-way analysis ofvariance (ANOVA) revealed significant main effects of cocainetreatment [F(1,26) � 11.36, p � .002] and glutamate concen-tration [F(2,26) � 14.73, p � .001]; *p � .05). (B) The Vmax ofglutamate uptake was reduced in cocaine subjects with nochange in Kd. A two-way ANOVA revealed a significant effectof cocaine treatment [F(1,49) � 11.21, p � .001] and concen-tration of glutamate [F(4,49) � 13.61, p � .001]; *p � .05. (C–E)Cocaine self-administration followed by 3 weeks of extinctiontraining significantly decreased GLT-1 and xCT protein ex-pression, and chronic treatment with N-acetylcysteine or ceftri-axone restored levels of both proteins. One-way ANOVAs con-firmed an effect of group for both GLT-1 [F(3,47) � 5.34, p �.003] and xCT [F(3,45) � 4.47, p � .008], *p � .05 comparedwith cocaine. S � saline; C � cocaine; C–C � cocaine �ceftriaxone; C–N � cocaine � N-acetylcysteine; M � markerladder. (F) Membrane fraction levels of GLT-1 and xCT werenot changed in the prefrontal cortex following cocaine self-administration. (G) Seven days of intraperitoneal (IP) ceftriax-one treatment (200 mg/kg) had no effect on accumbensGLT-1 levels in naive rats. (H) Seven days of IP N-acetylcyste-ine treatment (100 mg/kg) had no effect on accumbens GLT-1levels in naive rats. VEH, vehicle; CEF, ceftriaxone; NAC, N-ace-tylcysteine.

nilt

FcsmoipwtmidT2

L.A. Knackstedt et al. BIOL PSYCHIATRY 2010;67:81–84 83

eous motor activity in an unhabituated open field test, indicat-ng a lack of nonspecific sedation (Figure 2D). This increase inocomotion was most likely indicative of an enhanced response

igure 2. Ceftriaxone inhibits cue- and cocaine-primed reinstatement ofocaine seeking. (A) The mean number of cocaine infusions (.2 mg/infusion)elf-administered (no ceftriaxone was administered at this time). (B) The

ean number of active lever presses during extinction training; ceftriaxoner vehicle administration began on Day 5 (arrow). (C) Animals treated with

ntraperitoneal (IP) ceftriaxone (200 mg/kg) during extinction training dis-layed less active lever pressing in response to both cues previously pairedith cocaine infusion and a priming injection of IP cocaine (15 mg/kg). A

wo-way analysis of variance (ANOVA) indicated significant effects of treat-ent [F(1,3) � 35.19, p � .001], test condition [F(2,3) � 17.78, p � .001], and

nteraction [F(3,103) � 8.180, p � .005], *p � .05. (D) Ceftriaxone treatmentid not reduce locomotor activity. A two-way ANOVA revealed effects ofime [F(1,17) � 23.84, p � .001 and a Time � Group interaction [F(1,17) �.01, p � .04], *p � .05.

o a novel environment, a phenomenon that would not influence

reinstatement behavior because reinstatement was measured in afamiliar environment (operant chamber).

Discussion

In summary, the Vmax of glutamate uptake was decreased inthe nucleus accumbens of rats with a history of cocaine self-administration and 3 weeks of extinction training (Figure 1B).The decreased uptake was associated with reduced expressionof the glutamate transporter GLT-1 (Figure 1C). The Food andDrug Administration–approved, �-lactam antibiotic ceftriaxoneincreases glutamate uptake and the transcription of GLT-1(15,16), and we showed that it restores NAcc levels of both GLT-1and xCT in rats trained to self-administer cocaine (Figure 1C and1D). Moreover, another compound known to inhibit cocaineseeking, N-acetylcysteine, also restored GLT-1 and xCT levels inthe accumbens (Figure 1C and 1D). These results point to acoregulation of expression of the proteins responsible for thetwo glutamate transport systems.

The restoration of these proteins critical for glutamate ho-meostasis by ceftriaxone resulted in an attenuation of both cue-and cocaine-primed reinstatement (Figure 2C). This reduction indrug-seeking behavior was not due to a nonspecific affect ofceftriaxone on locomotor behavior (Figure 2D). Because it haspreviously been shown that decreased levels of extracellularglutamate in the accumbens core predisposes rats to the rein-statement of cocaine seeking (1,7), it seems paradoxical thatceftriaxone-induced upregulation of sodium-dependent gluta-mate transport prevented reinstatement. However, ceftriaxonetreatment increased the expression of both GLT-1 and xCT(Figure 1C and 1D), indicating that the increase in uptake wasoffset by an increase in the efflux of nonsynaptic glutamatethrough enhanced cystine-glutamate exchange. It is possible thatthe attenuation of reinstatement observed here is due solely to therestoration of xCT expression, an idea that is supported by theprevious finding that, by normalizing basal glutamate levels inthe accumbens, N-acetylcysteine restores glutamatergic tone onmGluR2/3 receptors, thus preventing the overflow of glutamatethat drives cocaine-primed reinstatement (17). However, it islikely that the upregulation of GLT-1 observed here is anessential factor in the attenuation of reinstatement, serving todampen the overflow of extracellular glutamate during reinstate-ment. In fact, a computer model of the accumbens synapsefollowing cocaine self-administration requires a downregulationin sodium-dependent glutamate uptake in addition to a decreasein basal levels of glutamate and mGluR2/3 function to accountfor the increase in extracellular glutamate observed duringreinstatement (11). It is important to note that this study wasconducted in animals undergoing extinction training, which mayhave contributed to the observed changes in glutamate ho-meostasis. However, this seems unlikely given that daily noncon-tingent cocaine administration is also known to alter measures ofglutamate homeostasis, such as lowering basal extracellular gluta-mate content and reducing cystine-glutamate exchange (7). Inconclusion, these data indicate that the restoration of glutamatehomeostasis by normalizing glutamate uptake and cystine-gluta-mate exchange with ceftriaxone inhibits cocaine seeking inanimals and poses drugs regulating glutamate homeostasis aspotential therapies in human addiction.

This research was supported by NIDA Grant No. DA 026010awarded to Lori Knackstedt and NIDA Grant Nos. DA 015369

and DA 12,513 awarded to Peter Kalivas.

www.sobp.org/journal

t

o

84 BIOL PSYCHIATRY 2010;67:81–84 L.A. Knackstedt et al.

w

The authors report no biomedical financial interests or po-ential conflicts of interest.

Supplementary material cited in this article is availablenline.

1. McFarland K, Lapish CC, Kalivas PW (2003): Prefrontal glutamate releaseinto the core of the nucleus accumbens mediates cocaine-inducedreinstatement of drug-seeking behavior. J Neurosci 23:3531–3537.

2. Danbolt NC (2001): Glutamate uptake. Prog Neurobiol 65:1–105.3. Moussawi K, Pacchioni A, Moran M, Olive MF, Gass JT, Lavin A, et al.

(2009): N-acetylcysteine reverses cocaine-induced metaplasticity. NatNeurosci 12:182–189.

4. Mulholland PJ, Carpenter-Hyland EP, Hearing MC, Becker HC, Wood-ward JJ, Chandler LJ (2008): Glutamate transporters regulate extrasyn-aptic NMDA receptor modulation of Kv2.1 potassium channels. J Neuro-sci 28:8801– 8809.

5. McBean GJ (2002): Cerebral cystine uptake: A tale of two transporters.Trends Pharmacol Sci 23:299 –302.

6. Baker DA, Xi ZX, Shen H, Swanson CJ, Kalivas PW (2002): The origin andneuronal function of in vivo nonsynaptic glutamate. J Neurosci22:9134 –9141.

7. Baker DA, McFarland K, Lake RW, Shen H, Tang XC, Toda S, et al. (2003):Neuroadaptations in cystine-glutamate exchange underlie cocaine re-lapse. Nat Neurosci 6:743–749.

8. Madayag A, Lobner D, Kau KS, Mantsch JR, Abdulhameed O, Hearing M,

et al. (2007): Repeated N-acetylcysteine administration alters plasticity-dependent effects of cocaine. J Neurosci 27:13968 –13976.

ww.sobp.org/journal

9. LaRowe SD, Myrick H, Hedden S, Mardikian P, Saladin M, Mcrae A, et al.(2007): Is cocaine desire reduced by N-acetylcysteine? Am J Psychiatry164:1115–1117.

10. Haugeto O, Ullensvang K, Levy LM, Chaudhry FA, Honore T, Nielsen M, etal. (1996): Brain glutamate transporter proteins form homomultimers.J Biol Chem 271:27715–27722.

11. Pendyam S, Mohan A, Kalivas PW, Nair SS (2009): Computational modelof extracellular glutamate in the nucleus accumbens incorporates neu-roadaptations by chronic cocaine. Neuroscience 158:1266 –1276.

12. Bannai S (1986): Exchange of cystine and glutamate across plasmamembrane of human fibroblasts. J Biol Chem 261:2256 –2263.

13. Murphy TH, Miyamoto M, Sastre A, Schnaar RL, Coyle JT (1989): Gluta-mate toxicity in a neuronal cell line involves inhibition of cystine trans-port leading to oxidative stress. Neuron 2:1547–1558.

14. Knackstedt LA, Larowe S, Mardikian P, Malcolm R, Upadhyaya H, HeddenS, et al. (2008): The role of cystine-glutamate exchange in nicotine de-pendence in rats and humans. Biol Psychiatry.

15. Lee SG, Su ZZ, Emdad L, Gupta P, Sarkar D, Borjabad A, et al. (2008):Mechanism of ceftriaxone induction of excitatory amino acid transport-er-2 expression and glutamate uptake in primary human astrocytes.J Biol Chem 283:13116 –13123.

16. Rothstein JD, Patel S, Regan MR, Haenggeli C, Huang YH, Bergles DE, etal. (2005): Beta-lactam antibiotics offer neuroprotection by increasingglutamate transporter expression. Nature 433:73–77.

17. Moran MM, McFarland K, Melendez RI, Kalivas PW, Seamans JK (2005):Cystine/glutamate exchange regulates metabotropic glutamate recep-tor presynaptic inhibition of excitatory transmission and vulnerability to

cocaine seeking. J Neurosci 25:6389 – 6393.