《生物技术与人类》课程教学资源（经典阅读）Reversal of cocaine-evoked synaptic potentiation resets drug-induced adaptive behaviour

LETTER doi:10.1038/nature.10709 Reversal of cocaine-evoked synaptic potentiation resets drug-induced adaptive behaviour Vincent Pascoli,Marc Turiault'&Christian Ltscher.2 Drug-evoked synaptic plasticity is observed at many synapses and induced LTP in both types of MSNs after saline injection,became may underlie behavioural adaptations in addiction'.Mechanistic inefficient after cocaine treatment in DIR-MSNs (Supplementary investigations start with the identification of the molecular drug Fig.la-d).The D2R overexpression recently reported in drd2-eGFP targets.Cocaine,for example,exerts its reinforcing'and early mice interferes with neither the reported synaptic effects nor with the neuroadaptive effects by inhibiting the dopamine transporter, acute locomotor response to cocaine(Supplementary Fig.2).After a thus causing a strong increase in mesolimbic dopamine.Among cocaine injection,we were unable to induce HFS LTP regardless of the many signalling pathways subsequently engaged,phosphoryla- whether we recorded from MSNs in the shell or the core of the nucleus tion of the extracellular signal-regulated kinase(ERK)in the nucleus accumbens(Supplementary Fig.3),but because MSNs in the shell are accumbens'is of particular interest because it has been implicated in the immediate targets of the dopamine neurons of the medial ventral NMDA-receptor and type 1 dopamine (D1)-receptor-dependent tegmental area,which undergo the most significant changes in res- synaptic potentiations as well as in several behavioural adapta- ponse to a single cocaine injection',we focused on nucleus accumbens tions.A causal link between drug-evoked plasticity at identified shell neurons in the present study.Without distinguishing between synapses and behavioural adaptations,however,is missing,and the DIR-and D2R-MSNs,several reports have already suggested that benefits of restoring baseline transmission have yet to be demon- drug-evoked synaptic plasticity in the nucleus accumbens may strated.Here we find that cocaine potentiates excitatory trans- underlie drug-related behavioural adaptations (reviewed in ref.12). mission in D1-receptor-expressing medium-sized spiny neurons For example,when 5-7 daily cocaine injections were followed by a (DIR-MSNs)in mice via ERK signalling with a time course that 10-21-day withdrawal,an overall increase of the AMPA/NMDA parallels locomotor sensitization.Depotentiation of cortical nucleus receptor(AMPAR/NMDAR)ratioor GluA1/2 surface expression'45 accumbens inputs by optogenetic stimulation in vivo efficiently was observed,and both observations were reversed by a challenge restored normal transmission and abolished cocaine-induced injection of cocaine. locomotor sensitization.These findings establish synaptic potentia- The failure of HFS to induce LTP selectively in DIR-MSNs after tion selectively in DIR-MSNs as a mechanism underlying a core cocaine treatment may be due either to the impairment of LTP induc- component of addiction,probably by creating an imbalance tion or the occlusion of LTP expression.To distinguish between the between distinct populations of MSNs in the nucleus accumbens. two scenarios,we recorded miniature EPSCs(mEPSCs)in both cell Our data also provide proof of principle that reversal of types(Fig.le and Supplementary Fig.4a)and observed a significant cocaine-evoked synaptic plasticity can treat behavioural alterations increase in amplitude along with a modest change in frequency of caused by addictive drugs and may inspire novel therapeutic unitary events in DIR-MSNs(whereas these parameters remained approaches involving deep brain stimulation or transcranial unchanged in D2R-MSNs).Given that the paired pulse ratio also magnetic stimulation. remained unchanged (Fig.If and Supplementary Fig.4b),a post- We first tested whether cocaine treatment interfered with activity- synaptic mechanism underlying the increase of transmission and dependent long-term potentiation(LTP)in the nucleus accumbens hence an occlusion scenario is the most likely explanation. When excitatory afferents onto MSNs were challenged with a high- We characterized the induction criteria for HFS LTP in MSNs in frequency stimulation(HFS)train,LTP of the excitatory postsynaptic vitro and found that it depends on NMDAR activation (Fig.2a). currents(EPSCs)was observed(Fig.la).The magnitude of the LTP Applying the MEK(MAP ERK kinase)inhibitor U0126 for the dura- was halved in brain slices from mice that had received a single injection tion of the induction protocol also led to a complete block (Fig.2b) of cocaine 7 days before the recording.If the cocaine was injected a indicating that activation of the ERK pathway is essential for this form month before measuring synaptic plasticity ex vivo,the difference was of LTP,akin to NMDA-dependent LTP in the hippocampus and the no longer present(Fig.1b).MSNs of the nucleus accumbens fall into dorsal striatum7,the latter also depending on DiRs two classes ofabout equal proportions defined by the type of dopamine Given that our results indicate an occlusion of LTP,we hypothesized receptor expressed,with a small fraction of neurons(6-17%)that that cocaine could drive synaptic potentiation selectively in DIR- express both receptors.A possible explanation for the partial change MSNs via ERK activity.Indeed,a sharp increase in phosphorylated in LTP magnitude is therefore that cocaine exposure abolishes plasticity ERK is detected soon after cocaine exposure in DIR-MSNs,but not in selectively in one class.To test this,we attempted LTP induction ex vivo D2R-MSNs.In the nucleus ERK modulates gene expression,whereas after cocaine exposure in bacterial artificial chromosome(BAC)trans- in dendrites ERK is probably involved in the regulation of activity- genic mice expressing enhanced green fluorescent protein (eGFP) dependent spine dynamics,synaptic glutamate receptor insertion and either in D1 receptor (DIR)-or D2R-MSNs.We identified DIR- local dendritic protein synthesis.Toprovide in vivo evidence for ERK MSNs by a crossover strategy in which we recorded from green cells dependence of cocaine-evoked synaptic plasticity,we treated mice in drdla-eGFP mice and non-green cells in drd2-eGFP mice (and vice with SL327,a blood-brain barrier penetrant ERK pathway inhibitor, versa for D2R-MSNs).Because the two approaches to identify the cell before the saline or cocaine injection.We found that this manipula- type yielded very similar results,we pooled the data (Fig.Ic,d).The tion rescued HFS LTP in D1R-MSNs one week later (Fig.2c,d) main finding of this first experiment was that HFS,which reliably without modification of the acute locomotor response to cocaine Department of Basic Neurosciences,Medical Faculty.University of Geneva CH-1211 Geneva,Switzerland2Clinic of Neurology,Department of Clinical Neurosciences,Geneva University Hospital CH- 1211 Geneva,Switzerland. 5 JANUARY 2012 VOL 481I NATURE I71 2012 Macmillan Publishers Limited.All rights reserved

LETTER doi:10.1038/nature10709 Reversal of cocaine-evoked synaptic potentiation resets drug-induced adaptive behaviour Vincent Pascoli1 , Marc Turiault1 & Christian Lu¨scher1,2 Drug-evoked synaptic plasticity is observed at many synapses and may underlie behavioural adaptations in addiction1 . Mechanistic investigations start with the identification of the molecular drug targets. Cocaine, for example, exerts its reinforcing2 and early neuroadaptive effects3 by inhibiting the dopamine transporter, thus causing a strong increase in mesolimbic dopamine. Among the many signalling pathways subsequently engaged, phosphoryla￾tion of the extracellular signal-regulated kinase (ERK) in the nucleus accumbens4 is of particular interest because it has been implicated in NMDA-receptor and type 1 dopamine (D1)-receptor-dependent synaptic potentiation5 as well as in several behavioural adapta￾tions6–8. A causal link between drug-evoked plasticity at identified synapses and behavioural adaptations, however, is missing, and the benefits of restoring baseline transmission have yet to be demon￾strated. Here we find that cocaine potentiates excitatory trans￾mission in D1-receptor-expressing medium-sized spiny neurons (D1R-MSNs) in mice via ERK signalling with a time course that parallels locomotor sensitization. Depotentiation of cortical nucleus accumbens inputs by optogenetic stimulation in vivo efficiently restored normal transmission and abolished cocaine-induced locomotor sensitization. These findings establish synaptic potentia￾tion selectively in D1R-MSNs as a mechanism underlying a core component of addiction, probably by creating an imbalance between distinct populations of MSNs in the nucleus accumbens. Our data also provide proof of principle that reversal of cocaine-evoked synaptic plasticity can treat behavioural alterations caused by addictive drugs and may inspire novel therapeutic approaches involving deep brain stimulation or transcranial magnetic stimulation. We first tested whether cocaine treatment interfered with activity￾dependent long-term potentiation (LTP) in the nucleus accumbens. When excitatory afferents onto MSNs were challenged with a high￾frequency stimulation (HFS) train, LTP of the excitatory postsynaptic currents (EPSCs) was observed (Fig. 1a). The magnitude of the LTP was halved in brain slices from mice that had received a single injection of cocaine 7 days before the recording. If the cocaine was injected a month before measuring synaptic plasticity ex vivo, the difference was no longer present (Fig. 1b). MSNs of the nucleus accumbens fall into two classes of about equal proportions defined by the type of dopamine receptor expressed, with a small fraction of neurons (6–17%) that express both receptors9 . A possible explanation for the partial change in LTP magnitude is therefore that cocaine exposure abolishes plasticity selectively in one class. To test this, we attempted LTP induction ex vivo after cocaine exposure in bacterial artificial chromosome (BAC) trans￾genic mice expressing enhanced green fluorescent protein (eGFP) either in D1 receptor (D1R)- or D2R-MSNs. We identified D1R￾MSNs by a crossover strategy in which we recorded from green cells in drd1a-eGFP mice and non-green cells in drd2-eGFP mice (and vice versa for D2R-MSNs). Because the two approaches to identify the cell type yielded very similar results, we pooled the data (Fig. 1c, d). The main finding of this first experiment was that HFS, which reliably induced LTP in both types of MSNs after saline injection, became inefficient after cocaine treatment in D1R-MSNs (Supplementary Fig. 1a–d). The D2R overexpression recently reported in drd2-eGFP mice10 interferes with neither the reported synaptic effects nor with the acute locomotor response to cocaine (Supplementary Fig. 2). After a cocaine injection, we were unable to induce HFS LTP regardless of whether we recorded from MSNs in the shell or the core of the nucleus accumbens (Supplementary Fig. 3), but because MSNs in the shell are the immediate targets of the dopamine neurons of the medial ventral tegmental area, which undergo the most significant changes in res￾ponse to a single cocaine injection11, we focused on nucleus accumbens shell neurons in the present study. Without distinguishing between D1R- and D2R-MSNs, several reports have already suggested that drug-evoked synaptic plasticity in the nucleus accumbens may underlie drug-related behavioural adaptations (reviewed in ref. 12). For example, when 5–7 daily cocaine injections were followed by a 10–21-day withdrawal, an overall increase of the AMPA/NMDA receptor (AMPAR/NMDAR) ratio13 or GluA1/2 surface expression14,15 was observed, and both observations were reversed by a challenge injection of cocaine. The failure of HFS to induce LTP selectively in D1R-MSNs after cocaine treatment may be due either to the impairment of LTP induc￾tion or the occlusion of LTP expression. To distinguish between the two scenarios, we recorded miniature EPSCs (mEPSCs) in both cell types (Fig. 1e and Supplementary Fig. 4a) and observed a significant increase in amplitude along with a modest change in frequency of unitary events in D1R-MSNs (whereas these parameters remained unchanged in D2R-MSNs). Given that the paired pulse ratio also remained unchanged (Fig. 1f and Supplementary Fig. 4b), a post￾synaptic mechanism underlying the increase of transmission and hence an occlusion scenario is the most likely explanation. We characterized the induction criteria for HFS LTP in MSNs in vitro and found that it depends on NMDAR activation (Fig. 2a). Applying the MEK (MAP ERK kinase) inhibitor U0126 for the dura￾tion of the induction protocol also led to a complete block (Fig. 2b), indicating that activation of the ERK pathway is essential for this form of LTP, akin to NMDA-dependent LTP in the hippocampus and the dorsal striatum16,17, the latter also depending on D1Rs18. Given that our results indicate an occlusion of LTP, we hypothesized that cocaine could drive synaptic potentiation selectively in D1R￾MSNs via ERK activity. Indeed, a sharp increase in phosphorylated ERK is detected soon after cocaine exposure in D1R-MSNs, but not in D2R-MSNs9 . In the nucleus ERK modulates gene expression, whereas in dendrites ERK is probably involved in the regulation of activity￾dependent spine dynamics, synaptic glutamate receptor insertion and local dendritic protein synthesis19. To provide in vivo evidence for ERK dependence of cocaine-evoked synaptic plasticity, we treated mice with SL327, a blood–brain barrier penetrant ERK pathway inhibitor, before the saline or cocaine injection. We found that this manipula￾tion rescued HFS LTP in D1R-MSNs one week later (Fig. 2c, d) without modification of the acute locomotor response to cocaine 1 Department of Basic Neurosciences, Medical Faculty, University of Geneva, CH-1211 Geneva, Switzerland. 2 Clinic of Neurology, Department of Clinical Neurosciences, Geneva University Hospital, CH- 1211 Geneva, Switzerland. 5 JANUARY 2012 | VOL 481 | NATURE | 71 ©2012 Macmillan Publishers Limited. All rights reserved

RESEARCH LETTER (Supplementary Fig.5).A crossover control design with the two mouse but the synapses involved have not been identified,and no causal link lines again showed no difference between the green cells of one line and has been established. the non-green cells of the other(Supplementary Fig.6). We then confirmed that a single injection of cocaine was sufficient Because inhibition of the ERK pathway blocks locomotor sensitiza- to cause locomotor sensitization to a second injection of the same tion to cocaine2021,and we found LTP in DIR-MSNs to be dependent dose22.This was the case when challenged a week but not a month on ERK,we reasoned that cocaine-evoked potentiation might be a after the initial cocaine injection (Supplementary Fig.7).The beha- cellular correlate of the behavioural adaptation.Moreover,ERK activa- vioural alteration therefore followed a time course similar to the tion is correlated both with AMPAR expression at the cell surface cocaine-evoked synaptic potentiation(Fig.la,b).If cocaine-evoked throughout the nucleus accumbens and with locomotor sensitization't potentiation is causally involved in locomotor sensitization,then depotentiating these synapses may reverse the behavioural alterations. a Recordings 6 Recordings To test this prediction experimentally,we injected channelrhodopsin 1 week 1 month◆ LD. i.p. (ChR2)-expressing adeno-associated virus(AAV)into the infralimbic injection injection cortex and implanted light guides into the nucleus accumbens to be 400 8 00 able to selectively activate in vivo the terminals of this major excitatory 300 300- input.Histological verification of the injection site did confirm robust H expression in the infralimbic cortex,along with sparser expression in 200 200 the prelimbic cortex(Fig.3a and Supplementary Fig.8).To validate 100 100 further this approach,we recorded photocurrents from infected cortical neurons and prepared slices of the nucleus accumbens shell 40 1020 30 1 102030 40 in which wide-field light exposure led to robust AMPAR-mediated Time (min) Time(min) EPSCs(Fig.3b).We next applied light pulses at 1 Hz for 10min(an established long-term depression(LTD)protocol to reduce synaptic Recordings Recordings transmission,see Methods).We found that this protocol strongly eek 1 week depressed transmission in nucleus accumbens slices from both saline and cocaine-treated mice(Fig.3c).Interestingly,the magnitude of the 400 400- depression was significantly larger in the latter,in line with an efficient 300 HFS 300 depotentiation added to the LTD.As the LTD/depotentiation was 200 NMDAR dependent and occurred without change of the paired pulse 200 ratio,it is probably mediated by a postsynaptic expression mechanism 100 100 and therefore constitutes an actual reversal of cocaine-evoked poten- tiation(Supplementary Fig.9). 10 2030 10 0102030 Time(min) Time(min) 个 6 Recordings D1R-MSNs 400- 400 1 week Saline w 300 HFS 8 300 HFS .p. injection ■●ocaine 200 200 100- 100 D1R-MSNs 0- oVehicle .AP5 0o Vehicle .U0126 0.8 Saline 10 0 2030 0 10203040 0.6 Time(min) Time(min) 0.4 02 Recordings Recordings -1 hour 1 week -1hour1week↓ ODTR-MSN 1R-MSNs▲D2R-MSN 0 102030 40 SL327 Cocaine SL327 Cocaine Amplitude (pA) Inter-event interval (s) 400 400 HES HES Figure 1 Cocaine disrupts HFS-induced LTP in DIR-MSNs of the nucleus 300 300 accumbens by potentiation of excitatory afferents.a-d,Graphs show 200 200 normalized EPSCs as a function of time and the overlay of averaged(20 trials) traces of AMPAR EPSCs before(black line)and after(grey line)HFS.Symbols 100 100 represent average of 6 trials.a,One week after cocaine injection,LTP was 0 0- halved (224+17.9%to 14914.7%,Student's t-test t=3.06).b,After 1 ~10 010203040 _10 010203040 month,no difference between treatments was detected(204+26.1%to Time(min) Time (min) 206+21.3%,t21=0.07).c,Using drdla-and drd2-eGFP mice,HFS reliably induced LTP in DIR-MSNs and D2R-MSNs after saline injection (206 16.0% Figure 2 HFS LTP and cocaine-evoked potentiation both depend on ERK and 229+24.1%,respectively,27=0.78).d,Cocaine abolished HFS LTP in activation.a,HFS LTP was blocked by an NMDAR antagonist (AP5,100 uM; D1R-MSNs but not in D2R-MSNs(115±8.5%and239±12.7%，respectively, control,20915%versus AP5,93+9%,t=6.78).b,Bath application of the t39=7.98).e,Sample traces,cumulative probability and mean values of MEK inhibitor(U0126,5 uM,15 min before until 10 min after HFS protocol) amplitude and frequency for mEPSCs recorded from DIR-MSNs one week also blocked LTP(control,234 17.8%versus U0126,10916.2% after injection of saline or cocaine.Cocaine significantly increases mean i3=5.047).c,Intraperitoneal administration of the MEK inhibitor(SL237, amplitude of mEPSCs in DIR-MSNs(t14=4.68;Kolmogorov-Smirnov test: 40 mgkg)in drdla-eGFP mice or drd2-eGFP mice 1 h before saline did not P0.05).f,Paired-pulse ratio (PPR,50 ms inter-stimulus interval) week later(224±31%and221±27%，respectively,to=0.09).d,SL237 measured in DIR-MSNs was not different a week after saline (Sal)or cocaine administered 1 h before cocaine restored HFS LTP in DIR-MSNs(200+22% (Coc)injection (1.7+0.11 and 1.40.1,respectively,t34 1.73).Scale bars: versus 23227%for D2R-MSNs,t22=0.93).Scale bars:10 ms and 20 pA. 20 ms and 20pA.n=8-30,*P<0.05.Error bars show s.e.m. n=6-13,*P<0.05.Error bars show s.e.m. VOL 481 5 JANUARY 2012 2012 Macmillan Publishers Limited.All rights reserved

(Supplementary Fig. 5). A crossover control design with the two mouse lines again showed no difference between the green cells of one line and the non-green cells of the other (Supplementary Fig. 6). Because inhibition of the ERK pathway blocks locomotor sensitiza￾tion to cocaine20,21, and we found LTP in D1R-MSNs to be dependent on ERK, we reasoned that cocaine-evoked potentiation might be a cellular correlate of the behavioural adaptation. Moreover, ERK activa￾tion is correlated both with AMPAR expression at the cell surface throughout the nucleus accumbens and with locomotor sensitization14, but the synapses involved have not been identified, and no causal link has been established. We then confirmed that a single injection of cocaine was sufficient to cause locomotor sensitization to a second injection of the same dose22. This was the case when challenged a week but not a month after the initial cocaine injection (Supplementary Fig. 7). The beha￾vioural alteration therefore followed a time course similar to the cocaine-evoked synaptic potentiation (Fig. 1a, b). If cocaine-evoked potentiation is causally involved in locomotor sensitization, then depotentiating these synapses may reverse the behavioural alterations. To test this prediction experimentally, we injected channelrhodopsin (ChR2)-expressing adeno-associated virus (AAV) into the infralimbic cortex and implanted light guides into the nucleus accumbens to be able to selectively activate in vivo the terminals of this major excitatory input. Histological verification of the injection site did confirm robust expression in the infralimbic cortex, along with sparser expression in the prelimbic cortex (Fig. 3a and Supplementary Fig. 8). To validate further this approach, we recorded photocurrents from infected cortical neurons and prepared slices of the nucleus accumbens shell in which wide-field light exposure led to robust AMPAR-mediated EPSCs (Fig. 3b). We next applied light pulses at 1 Hz for 10 min (an established long-term depression (LTD) protocol to reduce synaptic transmission, see Methods). We found that this protocol strongly depressed transmission in nucleus accumbens slices from both saline and cocaine-treated mice (Fig. 3c). Interestingly, the magnitude of the depression was significantly larger in the latter, in line with an efficient depotentiation added to the LTD. As the LTD/depotentiation was NMDAR dependent and occurred without change of the paired pulse ratio, it is probably mediated by a postsynaptic expression mechanism and therefore constitutes an actual reversal of cocaine-evoked poten￾tiation (Supplementary Fig. 9). Normalized EPSC (%) 400 300 200 100 0 –10 0 10 20 30 40 HFS U0126 400 300 200 100 0 –10 0 10 20 30 40 Normalized EPSC (%) HFS Vehicle a b d Normalized EPSC (%) 400 300 200 100 0 –10 0 10 20 30 40 HFS Vehicle AP5 D1R-MSNs D2R-MSNs 400 300 200 100 0 –10 0 10 20 30 40 Normalized EPSC (%) HFS c D1R-MSNs D2R-MSNs i.p. SL327 –1 hour 1 week Cocaine Recordings i.p. SL327 –1 hour 1 week Cocaine Recordings Time (min) Time (min) Time (min) Time (min) * * Figure 2 | HFS LTP and cocaine-evoked potentiation both depend on ERK activation. a, HFS LTP was blocked by an NMDAR antagonist (AP5, 100 mM; control, 209 6 15% versus AP5, 93 6 9%, t14 5 6.78). b, Bath application of the MEK inhibitor (U0126, 5 mM, 15 min before until 10 min after HFS protocol) also blocked LTP (control, 234 6 17.8% versus U0126, 109 6 16.2%, t13 5 5.047). c, Intraperitoneal administration of the MEK inhibitor (SL237, 40 mg kg21 ) in drd1a-eGFP mice or drd2-eGFP mice 1 h before saline did not modify the magnitude of HFS LTP in D1R- or D2R-MSNs when assessed one week later (224 6 31% and 221 6 27%, respectively, t10 5 0.09). d, SL237, administered 1 h before cocaine restored HFS LTP in D1R-MSNs (200 6 22% versus 232 6 27% for D2R-MSNs, t22 5 0.93). Scale bars: 10 ms and 20 pA. n 5 6–13, *P , 0.05. Error bars show s.e.m. 400 300 200 100 0 Normalized EPSC (%) HFS Saline Cocaine a b c d D1R-MSNs D2R-MSNs D1R-MSNs D2R-MSNs HFS HFS Saline Cocaine 1 week i.p. injection Recordings 400 300 200 100 Normalized EPSC (%) 0 HFS 400 300 200 100 0 0 10 20 30 40 Normalized EPSC (%) 400 300 200 100 0 Normalized EPSC (%) e 40302010 Amplitude (pA) 0 1.0 0.8 0.6 0.4 0 0.2 Cumulative probability 0 5 4321 Inter-event interval (s) Saline D1R-MSNs Cocaine 1 month i.p. injection Recordings 1 week i.p. injection Recordings 1 week i.p. injection Recordings 1 week i.p. injection Recordings Saline Cocaine D1R-MSNs PPR 3 2 1 0 Sal Coc f Time (min) 403020100 Time (min) 403020100 Time (min) 0 40302010 Time (min) mEPSCs frequency (Hz)8 6 4 2 0 * mEPSCs amplitude (pA)15 10 5 0 * * –10 –10 –10 –10 Figure 1 | Cocaine disrupts HFS-induced LTP in D1R-MSNs of the nucleus accumbens by potentiation of excitatory afferents. a–d, Graphs show normalized EPSCs as a function of time and the overlay of averaged (20 trials) traces of AMPAR EPSCs before (black line) and after (grey line) HFS. Symbols represent average of 6 trials. a, One week after cocaine injection, LTP was halved (224 6 17.9% to 149 6 14.7%, Student’s t-test t43 5 3.06). b, After 1 month, no difference between treatments was detected (204 6 26.1% to 206 6 21.3%, t21 5 0.07). c, Using drd1a- and drd2-eGFP mice, HFS reliably induced LTP in D1R-MSNs and D2R-MSNs after saline injection (206 6 16.0% and 229 6 24.1%, respectively, t27 5 0.78). d, Cocaine abolished HFS LTP in D1R-MSNs but not in D2R-MSNs (115 6 8.5% and 239 6 12.7%, respectively, t39 5 7.98). e, Sample traces, cumulative probability and mean values of amplitude and frequency for mEPSCs recorded from D1R-MSNs one week after injection of saline or cocaine. Cocaine significantly increases mean amplitude of mEPSCs in D1R-MSNs (t14 5 4.68; Kolmogorov–Smirnov test: P , 0.01) whereas the frequency was unaffected (t14 5 1.635; Kolmogorov– Smirnov: P . 0.05). f, Paired-pulse ratio (PPR, 50 ms inter-stimulus interval) measured in D1R-MSNs was not different a week after saline (Sal) or cocaine (Coc) injection (1.7 6 0.11 and 1.4 6 0.1, respectively, t34 5 1.73). Scale bars: 20 ms and 20 pA. n 5 8–30, *P , 0.05. Error bars show s.e.m. RESEARCH LETTER 72 | NATURE | VOL 481 | 5 JANUARY 2012 ©2012 Macmillan Publishers Limited. All rights reserved

LETTER RESEARCH Infralimbic mPFC Recordings o Saline Cocaine 1 week L.P. injection 100m 多100十 80- NAc shell 60- 40- 1H也 1 -Baseline 20- -NBOX -1000203040 Time (min) d Second i.p injection:Cocaine NS 47 3.000 1 week 3 2500- First i.p. injection: Laser 1 Hz 2,000 2. 1500 AAV-control o 1- 1000 AAV-ChR2 o Saline Cocaine 500- 0 0 Dav 1 Saline or cocaine: AAV-ChR2 AAV-control e Recordings D1R-MSNs Recordings 1 week 1week↓ L.D. iD.I ine cocaine Laser 1 Hz Laser 1 Hz D1R-MSNs AAV-ChR2 1.0- 400 0.8- HFS 8 300 0.6 10 200 5 100 0.2 E 0- 0. 0- 00203040506070 5 10 010203040 Amplitude(pA) Inter-event interval (s) Time (min) Figure 3 Reversal of cocaine-evoked potentiation abolishes locomotor cocaine versus saline pre-treatment on day 8 in AAV-control and P=0.203 sensitization to cocaine.a,Schematic illustration of the site infected with AAv cocaine versus saline pre-treatment on day 8 in AAV-ChR2 infected mice; coding for eGFP-ChR2 and of the bilateral cannula implantation with optic **P=0.0015 for locomotor response to cocaine on day 8 for AAV-ChR2 fibres inserted in the nudeus accumbens shell.Insets show AAV-eGFP-ChR2 versus AAV-control infected mice).No effect of light protocol on locomotor expression in the infralimbic cortex and high-resolution image ofaxonal inputs response to cocaine on day 8 when mice receive saline on day 1 was detected onto nuceus accumbens shell neurons(blue stains nuclei,X60 magnification). (P>0.05 AAV-ChR2 versus AAV control).The sensitization index(see b,Averaged current traces in response to light pulses(470 nm)of 4 ms and Methods)was significantly reduced in AAV-ChR2 infected mice treated with 100ms in infected cells in slices of the infralimbic medial prefrontal cortex the light stimulation protocol(4 ms pulse,1 Hz,10 min)45 min before the (mPFC;top).In nucleus accumbens (NAc)slices,light pulses (4 ms)evoked cocaine injection on day 8 (t2o =3.47).n=8-12,*P0.05).Scale bars:200 ms and 20pA. Multiple-way repeated-measures analysis of variance for matching data n=9-10,*P<0.05.f,The light stimulation protocol restored HFS LTP in yielded:interaction between day,virus and treatment F=13.41,P<0.001; DIR-MSNs from AAV-ChR2 but not AAV-control infected mice effect of day F(3s)=78.16,P<0.001;effect of virus F(1.35)=17.78,P<0.001 (191.0+21.2%and 93.0 9.1%,respectively,t12=4.26).Scale bars:10 ms and effect oftreatment F(3s)=88.65,P<0.001.Wilcoxon or t-test:*P<.001 for 20 pA.n=7,*P<0.05.Error bars show s.e.m. We next applied the above-validated protocol in vivo with the goal reduced their mean amplitude in DIR-MSNs(Fig.3e)to a level com- of establishing a causal link between cocaine-evoked plasticity and parable to baseline transmission(Fig.le)without any effect on paired behavioural sensitization.We placed light guides into the ventral pulse ratio(Supplementary Fig.10).In contrast,the mEPSCamplitude striatum,past the nucleus accumbens core,thus preferentially aiming and frequency were not modified in D2R-MSNs after cocaine(Sup- at the principal cortical input onto MSNs of the nucleus accumbens plementary Fig.11).Lastly,after cocaine treatment and optogenetic shell,that is,axons that have their origin in the infralimbic cortex depotentiation the HFS protocol resulted in LTP ex vivo (Fig.3f), (Supplementary Fig.8).When freely moving mice were treated with confirming the restoration of baseline transmission in DIR-MSNs. the optogenetic depotentiation protocol 45 min before the injection of Behavioural sensitization typically refers to the observation of the cocaine challenge at day 8,locomotor sensitization was completely increased locomotor responses with repeated injections of cocaine. erased.In control experiments,light stimulation did not affect the However,sensitization becomes apparent already after the second locomotor response at day 8 when the first injection was saline instead injection and is best observed about a week later(Fig.3).Although of cocaine or when a control virus was used (Fig.3d).To ensure that certainly not sufficient to induce addiction,such early forms of drug- the light stimulation restored normal transmission,we recorded ex induced adaptations are considered permissive building blocks for vivo mEPSCs and found that an effective light treatment significantly more definite behavioural alterations.If this is the case,the synaptic 5 JANUARY 2012 VOL 481 NATURE 73 2012 Macmillan Publishers Limited.All rights reserved

We next applied the above-validated protocol in vivo with the goal of establishing a causal link between cocaine-evoked plasticity and behavioural sensitization. We placed light guides into the ventral striatum, past the nucleus accumbens core, thus preferentially aiming at the principal cortical input onto MSNs of the nucleus accumbens shell, that is, axons that have their origin in the infralimbic cortex (Supplementary Fig. 8). When freely moving mice were treated with the optogenetic depotentiation protocol 45 min before the injection of the cocaine challenge at day 8, locomotor sensitization was completely erased. In control experiments, light stimulation did not affect the locomotor response at day 8 when the first injection was saline instead of cocaine or when a control virus was used (Fig. 3d). To ensure that the light stimulation restored normal transmission, we recorded ex vivo mEPSCs and found that an effective light treatment significantly reduced their mean amplitude in D1R-MSNs (Fig. 3e) to a level com￾parable to baseline transmission (Fig. 1e) without any effect on paired pulse ratio (Supplementary Fig. 10). In contrast, the mEPSC amplitude and frequency were not modified in D2R-MSNs after cocaine (Sup￾plementary Fig. 11). Lastly, after cocaine treatment and optogenetic depotentiation the HFS protocol resulted in LTP ex vivo (Fig. 3f), confirming the restoration of baseline transmission in D1R-MSNs. Behavioural sensitization typically refers to the observation of increased locomotor responses with repeated injections of cocaine. However, sensitization becomes apparent already after the second injection and is best observed about a week later22 (Fig. 3). Although certainly not sufficient to induce addiction, such early forms of drug￾induced adaptations are considered permissive building blocks for more definite behavioural alterations. If this is the case, the synaptic a Laser 1 Hz D1R-MSNs Infralimbic mPFC Second i.p. injection: 1 D1R-MSNs 8 c 1 c 8 c 1 s 8 c 1 c 8 c * NS 4 3 2 1 0 * 100 ms200 pA NAc shell NBQX Baseline 50 pA 20 ms 1 week i.p. injection Recordings 100 80 60 40 Normalized EPSC (%) 403020100 Time (min) Saline Cocaine –10 1 Hz 20 0 b c 1 week First i.p. injection: Cocaine d Saline Cocaine AAV-control AAV-ChR2 Locomotion (1/4 turns) 3,000 2,500 2,000 1,500 1,000 500 0 Sensitization index Saline or cocaine: Day: s AAV-control AAV-ChR2 Laser 1 Hz 1 week i.p. cocaine Recordings AAV-control AAV-ChR2 e mEPSCs amplitude (pA)15 10 5 0 1.0 0.8 0.6 0.4 0 0.2 Cumulative probability mEPSCs frequency (Hz) 0 5 4321 Inter-event interval (s) 302010 70 Amplitude (pA) 0 40 50 60 8 6 4 2 0 f Laser 1 Hz 1 week i.p. cocaine Recordings HFS 400 300 200 100 0 Normalized EPSC (%) 403020100 Time (min) AAV-control AAV-ChR2 * * * ** –10 Figure 3 | Reversal of cocaine-evoked potentiation abolishes locomotor sensitization to cocaine. a, Schematic illustration of the site infected with AAV coding for eGFP-ChR2 and of the bilateral cannula implantation with optic fibres inserted in the nucleus accumbens shell. Insets show AAV-eGFP-ChR2 expression in the infralimbic cortex and high-resolution image of axonal inputs onto nucleus accumbens shell neurons (blue stains nuclei, 360 magnification). b, Averaged current traces in response to light pulses (470 nm) of 4 ms and 100 ms in infected cells in slices of the infralimbic medial prefrontal cortex (mPFC; top). In nucleus accumbens (NAc) slices, light pulses (4 ms) evoked AMPAR EPSCs (blocked by NBQX 20 mM, green trace, bottom). c, Low￾frequency stimulation (light pulses of 4 ms at 1 Hz for 10 min) induced depression of AMPAR EPSCs evoked with light in nucleus accumbens slices from mice injected with saline or cocaine one week previously (48 6 5.7% and 27 6 4.5%, respectively,t14 5 2.851). n 5 8, *P , 0.05. d, Quarter turns done by mice in the circular corridor for 60 min after injection represent locomotion (scatter plots of individual score and bars of mean 6 s.e.m. are shown). Multiple-way repeated-measures analysis of variance for matching data yielded: interaction between day, virus and treatment F(1,35) 5 13.41, P , 0.001; effect of day F(1,35) 5 78.16, P , 0.001; effect of virus F(1,35) 5 17.78, P , 0.001; effect of treatment F(1,35) 5 88.65, P , 0.001. Wilcoxon ort-test: *P , 0.001 for cocaine versus saline pre-treatment on day 8 in AAV-control and P 5 0.203 cocaine versus saline pre-treatment on day 8 in AAV-ChR2 infected mice; **P 5 0.0015 for locomotor response to cocaine on day 8 for AAV-ChR2 versus AAV-control infected mice). No effect of light protocol on locomotor response to cocaine on day 8 when mice receive saline on day 1 was detected (P . 0.05 AAV-ChR2 versus AAV control). The sensitization index (see Methods) was significantly reduced in AAV-ChR2 infected mice treated with the light stimulation protocol (4 ms pulse, 1 Hz, 10 min) 45 min before the cocaine injection on day 8 (t20 5 3.47). n 5 8–12, *P , 0.05. NS, not significant. e, Acute slices of cocaine- and laser-treated AAV-ChR2 or AAV-control infected mice were prepared on day 8 (45 min after light stimulation). Sample traces, cumulative probability and mean values of amplitude and frequency for mEPSCs recorded from D1R-MSNs are shown. mEPSC amplitude was decreased in AAV-ChR2 compared to AAV-control infected mice (t17 5 5.08; Kolmogorov–Smirnov: P , 0.001). mEPSC frequency was unaltered (t17 5 1,45; Kolmogorov–Smirnov: P . 0.05). Scale bars: 200 ms and 20 pA. n 5 9–10, *P , 0.05. f, The light stimulation protocol restored HFS LTP in D1R-MSNs from AAV-ChR2 but not AAV-control infected mice (191.0 6 21.2% and 93.0 6 9.1%, respectively, t12 5 4.26). Scale bars: 10 ms and 20 pA. n 5 7, *P , 0.05. Error bars show s.e.m. LETTER RESEARCH 5 JANUARY 2012 | VOL 481 | NATURE | 73 ©2012 Macmillan Publishers Limited. All rights reserved

RESEARCH LETTER potentiation onto DIR-MSNs should still be observed after chronic We have explored and gained insight into the molecular mechan- cocaine exposure,for example after 5 daily injections followed by a isms of synaptic adaptations to develop a strategy for reversal of 10-day withdrawal.To confirm this hypothesis we recorded mEPSCs cocaine-evoked potentiation with the goal of normalizing behaviour at the end of this protocol,as well as I h and 24h after a challenge We chose an optogenetic depotientiaton of inputs from the infralimbic injection (Supplementary Fig.12).This experiment confirmed that cortex to the nucleus accumbens shell because of the strong anatomical excitatory synapses onto DIR-MSNs were selectively potentiated at connection and the functional implication of this projection in the end of the withdrawal period. cocaine-seeking behaviour It is an appealing idea that behavioural We then tested whether,after this chronic exposure to cocaine,light adaptation,which closely reflects the potentiation of excitatory trans- stimulation was still effective in reversing locomotor sensitization. mission onto DIR-MSNs,is due to an imbalance of the two classes of After the 5 days of injection we imposed 10 days of withdrawal,and MSNS26 treated the mice with light stimulation 45 min before injecting a Several studies have already reported that pharmacological and challenge dose.This completely reversed locomotor sensitization molecular manipulations of key players in synaptic plasticity in the (Fig.4a).When a second challenge injection was administered 24h nucleus accumbens can affect adaptive behaviours associated with later,sensitized responses were again only observed in control mice.As addictive drug exposure.For example,the inhibition of calcium- a control,we did not observe an acute effect on locomotor behaviour permeable AMPARs,a hallmark oflate-stage cocaine-evoked synaptic during light stimulation(Supplementary Fig.14). plasticity in the nucleus accumbens27 and the viral expression of a Lastly,to estimate how long the effect of the light stimulation lasted, peptide that impairs GluAl trafficking?,reduce cue-induced cocaine we tested for behavioural sensitization 5 days after the intervention seeking and cocaine-primed reinstatement,respectively. (Fig.4b).No significantlocomotor sensitization was observed even when We provide proof of principle that optogenetic manipulations can the mice were challenged at this extended time after light stimulation. be used to reverse cocaine-evoked synaptic plasticity and thus abolish Our results identify NMDAR-and ERK-dependent LTP in DIR- locomotor sensitization.Although light stimulation fully resets MSNs of the nucleus accumbens as a form of synaptic plasticity required locomotor behaviour,sensitization begins to reappear after a few days, for locomotor sensitization to cocaine.ERK activation probably consti- suggesting that several treatment sessions may be required to obtain tutes a general feature ofaddictive drugs,because in all brain regions that long-lasting effects.This is not surprising,as chronic cocaine exposure receive dopamine inputs tetrahydrocannabinol,amphetamines, also induces a number of additional adaptive changes including struc- morphine and nicotine also activate ERK signalling Through tural remodelling (for example,increase in spines")and alterations of cocaine-driven ERK phosphorylation,potentiation is induced selec- gene expression tively in DIR-MSNs,which leads to the occlusion of HFS-driven LTP. Sensitization to cocaine-associated stimuli has been linked with incentive saliency0 and may explain the exceptionally strong motiva- These findings are in line with observations that ERK activation may control AMPAR trafficking directly,an effect that may also be tion of addicts to obtain the drug.With chronic use,early adaptive maintained over days through activation of ERK nuclear targets lead- changes such as those described here may build up to enhance craving during cocaine withdrawals.Successful interventions that reverse ing to gene regulation2. these changes in animal models could inspire novel treatments for human addiction,a disease with a high social burden.Indeed,novel protocols of deep brain stimulation or transcranial magnetic stimu- ●ocaine b Cocaine lation may induce forms of synaptic plasticity that reverse drug-evoked Saline or aller Saline or challen adaptations,thus curbing the risk of relapse. Saline cocaine Saline cocaine 025 15 224 AAM-control o● 45m%n METHODS SUMMARY AAV-control o AAV-ChR2 Laser 1 Hz AAV-ChR2 All experiments were reviewed by the institutional ethics committee and approved Laser 1 Hz Salin ocaine Saline by the relevant authorities of the Canton of Geneva.C57BL/6 or heterozygous BAC transgenic female and male mice,in which eGFP expression was driven by 1,600- 1.600- either DIR (drdla-eGFP)or D2R(drd2-eGFP)gene regulatory elements were injected intraperitoneally with saline or cocaine.We then prepared brain slices 1,200 1.200 for electrophysiological recordings as previously described or placed the mice in 800 800 an apparatus to quantify locomotor behaviour.Locomotor sensitization or synaptic plasticity was monitored at the various time points after the cocaine 400 400- injection.A two-injection or five-injection protocol to induce locomotor sensitiza- 0- 0000-0 0- 0900000 tion was used as previously described2.Standard surgical procedures were used to 21012346 15 012345 15 infect mice with ChR2-AAV or a control AAV (0.5 ul)in the infralimbic medial Time (days) Time(days) prefrontal cortex while the light guides were aimed at both nucleus accumbens (shell).In depotentiation experiments,a 473-nm solid-state laser was used to carry Figure 4 Optogenetic depotentiation resets behavioural sensitization out the in vivo light stimulation protocol in awake mice (600 pulses of 4ms induced by chronic cocaine injections.a,After 5 daily cocaine injections a duration at 1 Hz,10-20 mW),45 min or 5 days before behavioural testing or ex robust locomotor sensitization is measured in AAV-control and AAV-ChR2 vivo electrophysiology recordings,respectively. infected mice.The light stimulation protocol (45 min before a challenge injection of cocaine on day 15)abolished the locomotor sensitized response in Full Methods and any associated references are available in the online version of the paper at www.nature.com/nature. AAV-ChR2 infected mice (one-way analysis of variance:effect of treatment F)=88.65,P<0.001.Post-hoc comparison by Bonferroni test yielded: Received 10 June;accepted 10 November 2011. *P<0.001 for cocaine versus saline pre-treatment in AAV control; **P<0.001 for AAV-ChR2 versus AAV-control infected mice,pre-treated Published online 7 December 2011. with cocaine(n=13-15).b,Light stimulation protocol applied on day 10 still 1. Luscher,C.Malenka,R.C.Drug-evoked synaptic plasticity in addiction:from reduced the locomotor sensitized response to a cocaine challenge injection on molecular changes to circuit remodeling.Neuron 69,650-663(2011). day 15 (one-way analysis of variance:effect of treatment F(2.45)=15.31, 2 Chen,R.et al Abolished cocaine reward in mice with a cocaine-insensitive P<0.001.Post-hoc comparison by Bonferroni test yielded:*P<0.001 for dopamine transporter.Proc.Nat/Acad.Sci.USA 103,9333-9338(2006). cocaine versus saline pre-treatment in AAV-control infected mice;**P=0.003 3. Brown,M.T.etal.Drug-driven AMPAreceptor redistribution mimicked by selective for AAV-ChR2 versus AAV-control infected mice,pre-treated with cocaine. dopamine neuron stimulation.PLoS ONE 5,e15870(2010). 4 Valjent,E,Pages,C.Herve,D.Girault,J.A.Caboche,J.Addictive and non- No sensitization in AAV-ChR2 (P=0.27,compared to saline pre-treatment; addictive drugs induce distinct and specific patterns of ERK activation in mouse n =11-21).Error bars show s.e.m. brain.Eur.J.Neurosci.19,1826-1836 (2004). I VOL JANUARY 2012 2012 Macmillan Publishers Limited.All rights reserved

potentiation onto D1R-MSNs should still be observed after chronic cocaine exposure, for example after 5 daily injections followed by a 10-day withdrawal. To confirm this hypothesis we recorded mEPSCs at the end of this protocol, as well as 1 h and 24 h after a challenge injection (Supplementary Fig. 12). This experiment confirmed that excitatory synapses onto D1R-MSNs were selectively potentiated at the end of the withdrawal period. We then tested whether, after this chronic exposure to cocaine, light stimulation was still effective in reversing locomotor sensitization. After the 5 days of injection we imposed 10 days of withdrawal, and treated the mice with light stimulation 45 min before injecting a challenge dose. This completely reversed locomotor sensitization (Fig. 4a). When a second challenge injection was administered 24 h later, sensitized responses were again only observed in control mice. As a control, we did not observe an acute effect on locomotor behaviour during light stimulation (Supplementary Fig. 14). Lastly, to estimate how long the effect of the light stimulation lasted, we tested for behavioural sensitization 5 days after the intervention (Fig. 4b). No significant locomotor sensitization was observed even when the mice were challenged at this extended time after light stimulation. Our results identify NMDAR- and ERK-dependent LTP in D1R￾MSNs of the nucleus accumbens as a form of synaptic plasticity required for locomotor sensitization to cocaine. ERK activation probably consti￾tutes a generalfeature of addictive drugs, because in all brain regions that receive dopamine inputs tetrahydrocannabinol, amphetamines, morphine and nicotine also activate ERK signalling4 . Through cocaine-driven ERK phosphorylation, potentiation is induced selec￾tively in D1R-MSNs, which leads to the occlusion of HFS-driven LTP. These findings are in line with observations that ERK activation may control AMPAR trafficking directly, an effect that may also be maintained over days through activation of ERK nuclear targets lead￾ing to gene regulation23. We have explored and gained insight into the molecular mechan￾isms of synaptic adaptations to develop a strategy for reversal of cocaine-evoked potentiation with the goal of normalizing behaviour. We chose an optogenetic depotientiaton of inputs from the infralimbic cortex to the nucleus accumbens shell because of the strong anatomical connection and the functional implication of this projection in cocaine-seeking behaviour24,25. It is an appealing idea that behavioural adaptation, which closely reflects the potentiation of excitatory trans￾mission onto D1R-MSNs, is due to an imbalance of the two classes of MSNs26. Several studies have already reported that pharmacological and molecular manipulations of key players in synaptic plasticity in the nucleus accumbens can affect adaptive behaviours associated with addictive drug exposure. For example, the inhibition of calcium￾permeable AMPARs, a hallmark of late-stage cocaine-evoked synaptic plasticity in the nucleus accumbens27 and the viral expression of a peptide that impairs GluA1 trafficking28, reduce cue-induced cocaine seeking and cocaine-primed reinstatement, respectively. We provide proof of principle that optogenetic manipulations can be used to reverse cocaine-evoked synaptic plasticity and thus abolish locomotor sensitization. Although light stimulation fully resets locomotor behaviour, sensitization begins to reappear after a few days, suggesting that several treatment sessions may be required to obtain long-lasting effects. This is not surprising, as chronic cocaine exposure also induces a number of additional adaptive changes including struc￾tural remodelling (for example, increase in spines29) and alterations of gene expression23. Sensitization to cocaine-associated stimuli has been linked with incentive saliency30 and may explain the exceptionally strong motiva￾tion of addicts to obtain the drug. With chronic use, early adaptive changes such as those described here may build up to enhance craving during cocaine withdrawal30. Successful interventions that reverse these changes in animal models could inspire novel treatments for human addiction, a disease with a high social burden. Indeed, novel protocols of deep brain stimulation or transcranial magnetic stimu￾lation may induce forms of synaptic plasticity that reverse drug-evoked adaptations, thus curbing the risk of relapse. METHODS SUMMARY All experiments were reviewed by the institutional ethics committee and approved by the relevant authorities of the Canton of Geneva. C57BL/6 or heterozygous BAC transgenic female and male mice, in which eGFP expression was driven by either D1R (drd1a-eGFP) or D2R (drd2-eGFP) gene regulatory elements were injected intraperitoneally with saline or cocaine. We then prepared brain slices for electrophysiological recordings as previously described3 or placed the mice in an apparatus to quantify locomotor behaviour. Locomotor sensitization or synaptic plasticity was monitored at the various time points after the cocaine injection. A two-injection or five-injection protocol to induce locomotor sensitiza￾tion was used as previously described22. Standard surgical procedures3were used to infect mice with ChR2-AAV or a control AAV (0.5 ml) in the infralimbic medial prefrontal cortex while the light guides were aimed at both nucleus accumbens (shell). In depotentiation experiments, a 473-nm solid-state laser was used to carry out the in vivo light stimulation protocol in awake mice (600 pulses of 4 ms duration at 1 Hz, 10–20 mW), 45 min or 5 days before behavioural testing or ex vivo electrophysiology recordings, respectively. Full Methods and any associated references are available in the online version of the paper at www.nature.com/nature. Received 10 June; accepted 10 November 2011. Published online 7 December 2011. 1. Lu¨ scher, C. & Malenka, R. C. Drug-evoked synaptic plasticity in addiction: from molecular changes to circuit remodeling. Neuron 69, 650–663 (2011). 2. Chen, R. et al. Abolished cocaine reward in mice with a cocaine-insensitive dopamine transporter. Proc. Natl Acad. Sci. USA 103, 9333–9338 (2006). 3. Brown, M. T. et al.Drug-driven AMPA receptor redistribution mimicked by selective dopamine neuron stimulation. PLoS ONE 5, e15870 (2010). 4. Valjent, E., Pages, C., Herve, D., Girault, J. A. & Caboche, J. Addictive and non￾addictive drugs induce distinct and specific patterns of ERK activation in mouse brain. Eur. J. Neurosci. 19, 1826–1836 (2004). Saline 45 min Saline or cocaine a b Laser 1 Hz * 15 Cocaine challenge Saline Cocaine AAV-control AAV-ChR2 * * * * * * –2 0 1 2 3 4 5 –1 Saline Saline or cocaine Laser 1 Hz 15 Cocaine challenge Saline Cocaine AAV-control AAV-ChR2 –2 –1 0 1 23 4 5 10 543210–1–2 Time (days) 15 1,600 1,200 800 400 0 Locomotion (1/4 turns) * * * 543210–1–2 Time (days) 15 1,600 1,200 800 400 0 Locomotion (1/4 turns) * ** ** Figure 4 | Optogenetic depotentiation resets behavioural sensitization induced by chronic cocaine injections. a, After 5 daily cocaine injections a robust locomotor sensitization is measured in AAV-control and AAV-ChR2 infected mice. The light stimulation protocol (45 min before a challenge injection of cocaine on day 15) abolished the locomotor sensitized response in AAV-ChR2 infected mice (one-way analysis of variance: effect of treatment F(2,41) 5 88.65, P , 0.001. Post-hoc comparison by Bonferroni test yielded: *P , 0.001 for cocaine versus saline pre-treatment in AAV control; **P , 0.001 for AAV-ChR2 versus AAV-control infected mice, pre-treated with cocaine (n 5 13–15). b, Light stimulation protocol applied on day 10 still reduced the locomotor sensitized response to a cocaine challenge injection on day 15 (one-way analysis of variance: effect of treatment F(2,45) 5 15.31, P , 0.001. Post-hoc comparison by Bonferroni test yielded: *P , 0.001 for cocaine versus saline pre-treatment in AAV-control infected mice; **P 5 0.003 for AAV-ChR2 versus AAV-control infected mice, pre-treated with cocaine. No sensitization in AAV-ChR2 (P 5 0.27, compared to saline pre-treatment; n 5 11–21). Error bars show s.e.m. RESEARCH LETTER 74 | NATURE | VOL 481 | 5 JANUARY 2012 ©2012 Macmillan Publishers Limited. All rights reserved

LETTER RESEARCH 5. Pascoli,V.et al Cyclic adenosine monophosphate-independent tyrosine activated protein kinase signal transduction cascade.J.Neurosci.19,8685-8695 phosphorylation of NR2B mediates cocaine-induced extracellular signal- (1999) regulated kinase activation.Biol Psychiatry 69,218-227(2011). 22.Valient.E et al Mechanisms of locomotor sensitization to drugs of abuse in a two- 6. Lu,L,Koya,E.Zhai,H.,Hope,B.T.Shaham,Y.Role of ERK in cocaine addiction. injection protocol.Neuropsychopharmacology 35,401-415(2010). Trends Neurosci.29,695-703(2006). 23.Brami-Cherrier,K.Roze,E,Girault,J.A.,Betuing.S.&Caboche,J.Role of the ERK/ 7. Girault,J.A.Valjent E.,Caboche.J.Herve,D.ERK2:a logical AND gate critical for MSK1 signalling pathway in chromatin remodelling and brain responses to drugs drug-induced plasticity?Curr.Opin.Pharmacol 7,77-85(2007). of abuse.J.Neurochem.108,1323-1335(2009). 8. Thomas,M.J.Kalivas,P.W.Shaham,Y.Neuroplasticity in the mesolimbic 24.Sesack,S.R.,Deutch,A.Y.,Roth,R.H.Bunney,B.S.Topographical organization dopamine system and cocaine addiction.Br.J.Pharmacol.154,327-342(2008). 9. of the efferent projections of the medial prefrontal cortex in the rat:an anterograde Bertran-Gonzalez,J.et al.Opposing patterns of signaling activation in dopamine D and Da receptor-expressing striatal neurons in response to cocaine and tract-tracing study with Phaseolus vulgaris leucoagglutinin.J.Comp.Neurol.290, 213-242(1989). haloperidol.J.Neurosci.28,5671-5685(2008). 25.LaLumiere,R.T.,Niehoff,K.E Kalivas,P.W.The infralimbic cortex regulates the 10.Kramer,P.F.et al.Dopamine D2 receptor overexpression alters behavior and physiology in Drd2-EGFP mice.J.Neurosci.31,126-132 (2011). consolidation of extinction after cocaine self-administration.Leam.Mem.17, 11.Lammel,S.,lon,D.1 Roeper,J.Malenka,R.C.Projection-specific modulation of 168-175(2010). dopamine neuron synapses by aversive and rewarding stimuli.Neuron 70, 26.Lobo,M.K.et al.Cell type- pecific loss of BDNF signaling mimics optogenetic control of cocaine reward.Science 330,385-390 (2010). 855-8522011】 12.Wolf,M.E.The Bermuda Triangle of cocaine-induced neuroadaptations.Trends 27.Conrad,K.L.et al.Formation of accumbens GluR2-lacking AMPA receptors Neurosci..33.391-398(2010). mediates incubation of cocaine craving.Nature 454,118-121 (2008). 13.Kourrich,S.Rothwell,P.E.Klug.J.R.Thomas,M.J.Cocaine experience controls 28.Anderson,S.M.et al.CaMKll:a biochemical bridge linking accumbens bidirectional synaptic plasticity in the nucleus accumbens.J.Neurosci.27, dopamine and glutamate systems in cocaine seeking.Nature Neurosci.11, 7921-7928(2007. 344-353(2008). 14.Boudreau,A.C.Wolf,M.E.Behavioral sensitization to cocaine is associated with 29.Robinson,T.E Kolb,B.Structural plasticity associated with exposure to drugs of increased AMPA receptor surface expression in the nucleus accumbens. abuse.Neuropharmacology 47 (suppL 1),33-46(2004). J.Neurosci..25,9144-9151(2005). 30.Vanderschuren,L.J.Pie ce,R.C.Sensitization processes in drug addiction.Cur. 15.Boudreau,A.C.Reimers,J.M.,Milovanovic,M.Wolf,M.E.Cell surface AMPA Top.Behav.Neurosci.3,179-195(2010) receptors in the rat nucleus accumbens increase during cocaine withdrawal but internalize after cocaine challenge in association with altered activation of Supplementary Information is linked to the online version of the paper at mitogen-activated protein kinases.J.Neurosci.27,10621-10635(2007) www.nature.com/nature 16.English,J.D.Sweatt,J.D.Arequirement for the mitogen-activated protein kinase cascade in hippocampal long term potentiation.J.Biol.Chem.272,19103-19106 Acknowledgements We thank P.Kalivas and the members of the Luscher laboratory for suggestions on the manuscript.This work is supported by the Swiss National (1997). Science Foundation(C.L.)and "Synapsy",a National Competence Center in Research 17.Xie,G.Q.et al.Ethanol attenuates the HFS-induced,ERK-mediated LTP in a dose dependent manner in rat striatum.Alcohol.Clin.Exp.Res.33,121-128(2009). (NCCR)of the Swiss Confederation on the synaptic basis of mental disorders. 18.Shen,W.Flajolet,M.,Greengard,P.Surmeier,D.J.Dichotomous dopaminergic Author Contributions V.P.carried out all electrophysiology experiments and was control of striatal synaptic plasticity.Science 321,848-851(2008). helped by M.T.with the behavioural experiments.C.L.designed the study and wrote the 19.Zhu,J.J.,Qin,Y.,Zhao,M.,Van Aelst,L Malinow,R.Ras and Rap control AMPA receptor trafficking during synaptic plasticity.Cel/110,443-455(2002) manuscript together with V.P.and M.T 20.Valjent,E.et al.Regulation of a protein phosphatase cascade allows convergent Author Information Reprints and permissions information is available at dopamine and glutamate signals to activate ERK in the striatum.Proc.Natl Acad. www.nature.com/reprints.The authors declare no competing financial interests Sci.USA102,491-496(2005) Readers are welcome to comment on the online version of this article at 21.Pierce,R.C.Pierce-Bancroft,A.F.Prasad,B.M.Neurotrophin-3 contributes to www.nature.com/nature.Correspondence and requests for materials should be the initiation of behavioral sensitization to cocaine by activating the Ras/mitogen- addressed to C.L.(christian.luscher@unige.ch). 5 JANUARY 2012 VOL 481I NATURE 75 2012 Macmillan Publishers Limited.All rights reserved

5. Pascoli, V. et al. Cyclic adenosine monophosphate-independent tyrosine phosphorylation of NR2B mediates cocaine-induced extracellular signal￾regulated kinase activation. Biol. Psychiatry 69, 218–227 (2011). 6. Lu, L., Koya, E., Zhai, H., Hope, B. T. & Shaham, Y. Role of ERK in cocaine addiction. Trends Neurosci. 29, 695–703 (2006). 7. Girault, J. A., Valjent, E., Caboche, J. & Herve, D. ERK2: a logical AND gate critical for drug-induced plasticity? Curr. Opin. Pharmacol. 7, 77–85 (2007). 8. Thomas, M. J., Kalivas, P. W. & Shaham, Y. Neuroplasticity in the mesolimbic dopamine system and cocaine addiction. Br. J. Pharmacol. 154, 327–342 (2008). 9. Bertran-Gonzalez, J. et al. Opposing patterns of signaling activation in dopamine D1 and D2 receptor-expressing striatal neurons in response to cocaine and haloperidol. J. Neurosci. 28, 5671–5685 (2008). 10. Kramer, P. F. et al. Dopamine D2 receptor overexpression alters behavior and physiology in Drd2-EGFP mice. J. Neurosci. 31, 126–132 (2011). 11. Lammel, S., Ion, D. I., Roeper, J. & Malenka, R. C. Projection-specific modulation of dopamine neuron synapses by aversive and rewarding stimuli. Neuron 70, 855–862 (2011). 12. Wolf, M. E. The Bermuda Triangle of cocaine-induced neuroadaptations. Trends Neurosci. 33, 391–398 (2010). 13. Kourrich, S., Rothwell, P. E., Klug, J. R. & Thomas, M. J. Cocaine experience controls bidirectional synaptic plasticity in the nucleus accumbens. J. Neurosci. 27, 7921–7928 (2007). 14. Boudreau, A. C. & Wolf, M. E. Behavioral sensitization to cocaine is associated with increased AMPA receptor surface expression in the nucleus accumbens. J. Neurosci. 25, 9144–9151 (2005). 15. Boudreau, A. C., Reimers, J. M., Milovanovic, M. & Wolf, M. E. Cell surface AMPA receptors in the rat nucleus accumbens increase during cocaine withdrawal but internalize after cocaine challenge in association with altered activation of mitogen-activated protein kinases. J. Neurosci. 27, 10621–10635 (2007). 16. English, J. D. & Sweatt, J. D. A requirement for the mitogen-activated protein kinase cascade in hippocampal long term potentiation. J. Biol. Chem. 272, 19103–19106 (1997). 17. Xie, G. Q. et al. Ethanol attenuates the HFS-induced, ERK-mediated LTP in a dose￾dependent manner in rat striatum. Alcohol. Clin. Exp. Res. 33, 121–128 (2009). 18. Shen, W., Flajolet, M., Greengard, P. & Surmeier, D. J. Dichotomous dopaminergic control of striatal synaptic plasticity. Science 321, 848–851 (2008). 19. Zhu, J. J., Qin, Y., Zhao, M., Van Aelst, L. & Malinow, R. Ras and Rap control AMPA receptor trafficking during synaptic plasticity. Cell 110, 443–455 (2002). 20. Valjent, E. et al. Regulation of a protein phosphatase cascade allows convergent dopamine and glutamate signals to activate ERK in the striatum. Proc. Natl Acad. Sci. USA 102, 491–496 (2005). 21. Pierce, R. C., Pierce-Bancroft, A. F. & Prasad, B. M. Neurotrophin-3 contributes to the initiation of behavioral sensitization to cocaine by activating the Ras/mitogen￾activated protein kinase signal transduction cascade. J. Neurosci. 19, 8685–8695 (1999). 22. Valjent, E. et al. Mechanisms of locomotor sensitization to drugs of abuse in a two￾injection protocol. Neuropsychopharmacology 35, 401–415 (2010). 23. Brami-Cherrier, K., Roze, E., Girault, J. A., Betuing, S. & Caboche, J. Role of the ERK/ MSK1 signalling pathway in chromatin remodelling and brain responses to drugs of abuse. J. Neurochem. 108, 1323–1335 (2009). 24. Sesack, S. R., Deutch, A. Y., Roth, R. H. & Bunney, B. S. Topographical organization of the efferent projections of the medial prefrontal cortex in the rat: an anterograde tract-tracing study with Phaseolus vulgaris leucoagglutinin. J. Comp. Neurol. 290, 213–242 (1989). 25. LaLumiere, R. T., Niehoff, K. E. & Kalivas, P. W. The infralimbic cortex regulates the consolidation of extinction after cocaine self-administration. Learn. Mem. 17, 168–175 (2010). 26. Lobo, M. K. et al. Cell type-specific loss of BDNF signaling mimics optogenetic control of cocaine reward. Science 330, 385–390 (2010). 27. Conrad, K. L. et al. Formation of accumbens GluR2-lacking AMPA receptors mediates incubation of cocaine craving. Nature 454, 118–121 (2008). 28. Anderson, S. M. et al. CaMKII: a biochemical bridge linking accumbens dopamine and glutamate systems in cocaine seeking. Nature Neurosci. 11, 344–353 (2008). 29. Robinson, T. E. & Kolb, B. Structural plasticity associated with exposure to drugs of abuse. Neuropharmacology 47 (suppl. 1), 33–46 (2004). 30. Vanderschuren, L. J. & Pierce, R. C. Sensitization processes in drug addiction. Curr. Top. Behav. Neurosci. 3, 179–195 (2010). Supplementary Information is linked to the online version of the paper at www.nature.com/nature. Acknowledgements We thank P. Kalivas and the members of the Lu¨ scher laboratory for suggestions on the manuscript. This work is supported by the Swiss National Science Foundation (C.L.) and ‘‘Synapsy’’, a National Competence Center in Research (NCCR) of the Swiss Confederation on the synaptic basis of mental disorders. Author Contributions V.P. carried out all electrophysiology experiments and was helped by M.T. with the behavioural experiments. C.L. designed the study and wrote the manuscript together with V.P. and M.T. Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of this article at www.nature.com/nature. Correspondence and requests for materials should be addressed to C.L. (christian.luscher@unige.ch). LETTER RESEARCH 5 JANUARY 2012 | VOL 481 | NATURE | 75 ©2012 Macmillan Publishers Limited. All rights reserved

RESEARCH LETTER METHODS presence of the GFP of BAC transgenic mice by using a fluorescent microscope Mice.C57BL/6 or heterozygous BAC transgenic mice,in which eGFP expression Olympus BX50WI,fluorescent light U-RFL-T).The holding potential was was driven by either DIR (drdla-eGFP)or D2R (drd2-eGFP)gene regulatory -70mV,and the access resistance was monitored by a hyperpolarizing step of elements were backcrossed in C57Bl/6 mice for three to four generations,were -14mV with each sweep,every 10s.Experiments were discarded if the access used.Mice were housed in groups of 3-4 except for those implanted with guide resistance varied by more than 20%.Synaptic currents were evoked by stimuli(50- cannulae,in which case animals were housed separately.All animals were kept in a 100 us)at 0.1 Hz through bipolar stainless steel electrodes placed at the cortex- temperature-and hygrometry-controlled environment with a 12 h light/12 h dark nucleus-accumbens border.The internal solution contained (in mM):140 k-glu- cycle.Mice were injected intraperitoneally with 20mgkg cocaine,40mg kg conate,5 KCL,130 CsCl,10 HEPES,0.2 EGTA,2 MgCl,4 Na2ATP,0.3 NaGTP SL327 (dissolved in 25%DMSO)or 0.9%saline (injection volume 10 mlkg). and 10 sodium creatine-phosphate.Currents were amplified(Multiclamp 700B, Immediately after injection,mice were placed in the locomotor recording Axon Instruments),filtered at 5kHz and digitized at 20kHz (National apparatus for 1h.All procedures were approved by the Institutional Animal Instruments Board PCI-MIO-16E4,Igor,WaveMetrics).The liquid junction Care and Use Committee of the University of Geneva. potential was small(-3 mV),and therefore traces were not corrected.All experi- Locomotor sensitization.Locomotor activity was measured as the number of ments were carried out in the presence of picrotoxin (100 uM).LTP was induced quarter turns entirely crossed by a mouse in a circular corridor.Locomotor chamber apparatus was placed under a video tracking system (Any-maze,Stoelting)and 1 rad uimes measurements were made automatically by the software.After 3 days ofhabituation Miniature EPSCs were recorded in the presence of tetrodotoxin (0.5 uM).The to the test apparatus,mice underwent the experimental procedure,which consisted frequency,amplitudes and kinetic properties of these currents were then analysed of two sessions of 60 min separated by 1 week (or 1 month),called day 1 and day 8 using the Mini Analysis software package (v.4.3,Synaptosoft).Cocaine or light (or day 30).During the day I session,mice received saline or cocaine and were illumination of ChR2-induced changes in cumulative miniature EPSC amplitude placed immediately in the corridor for 60 min.One week or I month later (day 8 or and inter-event interval distribution were analysed for statistical significance using day 30),a second session was performed during which all mice were injected with the nonparametric two-sample Kolmogorov-Smirnov test (KyPlot)with a con- cocaine before being placed in the circular corridor for 60 min.The light stimulation servative critical probability level of P<0.05 protocol (600 pulses of 4ms at 1 Hz)was done 45 min before the second cocaine Paired pulse ratio(PPR)was calculated by dividing the second evoked EPSC by injection.To compare the effects of various times after the first injections or various the first with a 50-ms interval in between. virus infections,locomotor activity in response to the second cocaine injection was AMPAR EPSCs evoked with ChR2 stimulation by 4-ms light pulses (LED normalized to the mean locomotor activity of saline-pre-treated mice and the Thorlabs)were recorded in the same conditions as electrically evoked synaptic sensitization index was calculated by dividing the normalized locomotor response currents.Low-frequency stimulation(LFS;1 Hz for 10min)was applied with light to the second injection by the normalized response to the first injection.Locomotor pulses and the magnitude of LTD was determined by comparing average EPSCs sensitization was also evaluated during challenge sessions that followed a chronic that were recorded 20-30 min after induction to EPSCs recorded immediately treatment (5 days of cocaine 15mgkg,10 days withdrawal).The light stimu- before induction. lation protocol was done 45 min or 5 days before the challenge injection ofcocaine. In vivo stimulation of infralimbic cortex projections in the nucleus accumbens Virus stereotaxic injection of ChR2-AAV or control AAV.AAVI viruses pro- shell.Virus injected and cannulated animals were allowed a minimum of I week to duced at the University of North Carolina(Vector Core Facility)were injected into recover and to express the virus.473-nm solid-state lasers(GMP,CH)were used to the infralimbic cortex of 15-20g wild-type or BAC transgenic mice.Anaesthesia carry out the in vivo stimulation protocol in awake mice.A fibre optic (Thorlabs) was induced and maintained with isoflurane(Baxter AG).The animal was placed in was customized to enable the mouse to move freely during stimulation.Briefly,the a stereotaxic frame(Angle One)and craniotomies were performed using stereotaxic plastic cap of a dummy cannula(Plastics One)was hollowed out and a hole of coordinates (anterio-posterior,+1.9;medio-lateral,+0.3;dorso-ventral,2.4-2.6). sufficient diameter for the fibre optic to pass through made in the top.This was Injections of AAVI viruses (0.5ul)were carried out using graduated pipettes threaded onto the fibre optic,one end of which was stripped to leave a 200um (Drummond Scientific Company),broken back to a tip diameter of 10-15 um,at external diameter.The fibre was then lowered into the guide cannula on the mouse a rate of~0.05 ul min.In all experiments the viruses were allowed a minimum of and the hollowed-out dummy cannula cap screwed onto the guide cannula.An 3 weeks to incubate before any other procedures were carried out.As a control, FC/PC rotative fibre-optic rotary joint(Doric lenses)was used to release torsion in some mice were injected with an AAV containing only GFP. the fibre caused by the animal's rotation.The fibre was connected to the laser, Cannula implantation.Following anaesthesia and craniotomy over the nucleus which delivered 4ms pulses at I Hz for 10 min,an established LTD protocol at accumbens,two holes were drilled around the craniotomy and screws were placed excitatory synapses in the nucleus accumbens".All stimulations were carried in the holes.Two weeks after viral injections,guide cannulae(Plastics One)were out in the mouse home cage (except in the experiment shown in Supplementary lowered slowly into position using stereotaxic coordinates (bilaterally anterio- Fig.10,in which stimulation was performed during locomotor recordings in the posterior,+1.5;medio-lateral,+1.6;dorso-ventral 4.1;15 angle)and cemented circular corridor)45 min or 5 days before behavioural testing or ex vivo electro- in place using dental cement (Lang Dental MFG Company)to encase the base of physiology recordings. the guide cannulae and the screws.Once the cement had dried,a dummy cannula 31.Mameli,M.etal Cocaine-evoked synaptic plasticity:persistence in the VTAtriggers (Plastics One)was placed inside each guide cannula to prevent infection. adaptations in the nucleus accumbens.Nature Neurosci.12,1036-1041 (2009) Slice electrophysiology.Coronal 200-250-um slices of mouse forebrain were 32.Gong.S.et al.Targeting Cre recombinase to specific neuron populations with prepared in cooled artificial cerebrospinal fluid (ACSF)containing (in mM): bacterial artificial chromosome constructs.J.Neurosci.27,9817-9823(2007) NaCl 119,KCl 2.5,MgCl 1.3,CaCl2 2.5,Na2HPO 1.0,NaHCOs 26.2 and glucose 33.Kombian,S.B.Malenka,R.C.Simultaneous LTP of non-NMDA-and LTD of 11,bubbled with%Oz and%CO2.Slices were kept at 32-34C in a recording NMDA-receptor-mediated responses in the nucleus accumbens.Nature 368, chamber superfused with 2.5mlmin ACSF.Visualized whole-cell voltage. 242-246(1994). 34.Pennartz,C.M.Ameerun,R.F.,Groenewegen,H.J.Lopes da Silva,F.H.Synaptic clamp recording techniques were used to measure holding and synaptic responses plasticity in an in vitro slice preparation of the rat nucleus accumbens.Eur.J. of MSNs of the nucleus accumbens shell,identified in some experiments by the Neurosci.5,107-117(1993). 2012 Macmillan Publishers Limited.All rights reserved

METHODS Mice. C57BL/6 or heterozygous BAC transgenic mice, in which eGFP expression was driven by either D1R (drd1a-eGFP) or D2R (drd2-eGFP) gene regulatory elements were backcrossed31 in C57Bl/6 mice for three to four generations, were used. Mice were housed in groups of 3–4 except for those implanted with guide cannulae, in which case animals were housed separately. All animals were kept in a temperature- and hygrometry-controlled environment with a 12 h light/12 h dark cycle. Mice were injected intraperitoneally with 20 mg kg21 cocaine, 40 mg kg21 SL327 (dissolved in 25% DMSO) or 0.9% saline (injection volume 10 ml kg21 ). Immediately after injection, mice were placed in the locomotor recording apparatus for 1 h. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Geneva. Locomotor sensitization. Locomotor activity was measured as the number of quarter turns entirely crossed by a mouse in a circular corridor. Locomotor chamber apparatus was placed under a video tracking system (Any-maze, Stoelting) and measurements were made automatically by the software. After 3 days of habituation to the test apparatus, mice underwent the experimental procedure, which consisted of two sessions of 60 min separated by 1 week (or 1 month), called day 1 and day 8 (or day 30). During the day 1 session, mice received saline or cocaine and were placed immediately in the corridor for 60 min. One week or 1 month later (day 8 or day 30), a second session was performed during which all mice were injected with cocaine before being placed in the circular corridor for 60 min. The light stimulation protocol (600 pulses of 4 ms at 1 Hz) was done 45 min before the second cocaine injection. To compare the effects of various times after the first injections or various virus infections, locomotor activity in response to the second cocaine injection was normalized to the mean locomotor activity of saline-pre-treated mice and the sensitization index was calculated by dividing the normalized locomotor response to the second injection by the normalized response to thefirst injection. Locomotor sensitization was also evaluated during challenge sessions that followed a chronic treatment (5 days of cocaine 15 mg kg21 , 10 days withdrawal). The light stimu￾lation protocol was done 45 min or 5 days before the challenge injection of cocaine. Virus stereotaxic injection of ChR2-AAV or control AAV. AAV1 viruses pro￾duced at the University of North Carolina (Vector Core Facility) were injected into the infralimbic cortex of 15–20 g wild-type or BAC transgenic mice. Anaesthesia was induced and maintained with isoflurane (Baxter AG). The animal was placed in a stereotaxicframe (Angle One) and craniotomieswere performed using stereotaxic coordinates (anterio-posterior, 11.9; medio-lateral, 60.3; dorso-ventral, 2.4–2.6). Injections of AAV1 viruses (0.5 ml) were carried out using graduated pipettes (Drummond Scientific Company), broken back to a tip diameter of 10–15 mm, at a rate of ,0.05 ml min21 . In all experiments the viruses were allowed a minimum of 3 weeks to incubate before any other procedures were carried out. As a control, some mice were injected with an AAV containing only GFP. Cannula implantation. Following anaesthesia and craniotomy over the nucleus accumbens, two holes were drilled around the craniotomy and screws were placed in the holes. Two weeks after viral injections, guide cannulae (Plastics One) were lowered slowly into position using stereotaxic coordinates (bilaterally anterio￾posterior, 11.5; medio-lateral, 61.6; dorso-ventral 4.1; 15u angle) and cemented in place using dental cement (Lang Dental MFG Company) to encase the base of the guide cannulae and the screws. Once the cement had dried, a dummy cannula (Plastics One) was placed inside each guide cannula to prevent infection. Slice electrophysiology. Coronal 200–250-mm slices of mouse forebrain were prepared in cooled artificial cerebrospinal fluid (ACSF) containing (in mM): NaCl 119, KCl 2.5, MgCl 1.3, CaCl2 2.5, Na2HPO4 1.0, NaHCO3 26.2 and glucose 11, bubbled with 95% O2 and 5% CO2. Slices were kept at 32–34 uC in a recording chamber superfused with 2.5 ml min21 ACSF. Visualized whole-cell voltage￾clamp recording techniques were used to measure holding and synaptic responses of MSNs of the nucleus accumbens shell, identified in some experiments by the presence of the GFP of BAC transgenic mice by using a fluorescent microscope (Olympus BX50WI, fluorescent light U-RFL-T). The holding potential was 270 mV, and the access resistance was monitored by a hyperpolarizing step of 214 mV with each sweep, every 10 s. Experiments were discarded if the access resistance varied by more than 20%. Synaptic currents were evoked by stimuli (50– 100 ms) at 0.1 Hz through bipolar stainless steel electrodes placed at the cortex– nucleus-accumbens border. The internal solution contained (in mM): 140 k-glu￾conate, 5 KCl, 130 CsCl, 10 HEPES, 0.2 EGTA, 2 MgCl2, 4 Na2ATP, 0.3 Na3GTP and 10 sodium creatine-phosphate. Currents were amplified (Multiclamp 700B, Axon Instruments), filtered at 5 kHz and digitized at 20 kHz (National Instruments Board PCI-MIO-16E4, Igor, WaveMetrics). The liquid junction potential was small (23 mV), and therefore traces were not corrected. All experi￾ments were carried out in the presence of picrotoxin (100 mM). LTP was induced by using the following HFS protocol: 100 pulses at 100 Hz repeated 4 times at 0.1 Hz paired with depolarization at 0 mV32,33. Miniature EPSCs were recorded in the presence of tetrodotoxin (0.5 mM). The frequency, amplitudes and kinetic properties of these currents were then analysed using the Mini Analysis software package (v.4.3, Synaptosoft). Cocaine or light illumination of ChR2-induced changes in cumulative miniature EPSC amplitude and inter-event interval distribution were analysed for statistical significance using the nonparametric two-sample Kolmogorov–Smirnov test (KyPlot) with a con￾servative critical probability level of P , 0.05. Paired pulse ratio (PPR) was calculated by dividing the second evoked EPSC by the first with a 50-ms interval in between. AMPAR EPSCs evoked with ChR2 stimulation by 4-ms light pulses (LED, Thorlabs) were recorded in the same conditions as electrically evoked synaptic currents. Low-frequency stimulation (LFS; 1 Hz for 10 min) was applied with light pulses and the magnitude of LTD was determined by comparing average EPSCs that were recorded 20–30 min after induction to EPSCs recorded immediately before induction. In vivo stimulation of infralimbic cortex projections in the nucleus accumbens shell.Virus injected and cannulated animals were allowed a minimum of 1 week to recover and to express the virus. 473-nm solid-state lasers (GMP, CH) were used to carry out the in vivo stimulation protocol in awake mice. A fibre optic (Thorlabs) was customized to enable the mouse to move freely during stimulation. Briefly, the plastic cap of a dummy cannula (Plastics One) was hollowed out and a hole of sufficient diameter for the fibre optic to pass through made in the top. This was threaded onto the fibre optic, one end of which was stripped to leave a 200 mm external diameter. The fibre was then lowered into the guide cannula on the mouse and the hollowed-out dummy cannula cap screwed onto the guide cannula. An FC/PC rotative fibre-optic rotary joint (Doric lenses) was used to release torsion in the fibre caused by the animal’s rotation. The fibre was connected to the laser, which delivered 4 ms pulses at 1 Hz for 10 min, an established LTD protocol at excitatory synapses in the nucleus accumbens34. All stimulations were carried out in the mouse home cage (except in the experiment shown in Supplementary Fig. 10, in which stimulation was performed during locomotor recordings in the circular corridor) 45 min or 5 days before behavioural testing or ex vivo electro￾physiology recordings. 31. Mameli,M. et al.Cocaine-evoked synaptic plasticity: persistence in the VTA triggers adaptations in the nucleus accumbens. Nature Neurosci. 12, 1036–1041 (2009). 32. Gong, S. et al. Targeting Cre recombinase to specific neuron populations with bacterial artificial chromosome constructs. J. Neurosci. 27, 9817–9823 (2007). 33. Kombian, S. B. & Malenka, R. C. Simultaneous LTP of non-NMDA- and LTD of NMDA-receptor-mediated responses in the nucleus accumbens. Nature 368, 242–246 (1994). 34. Pennartz, C. M., Ameerun, R. F., Groenewegen, H. J. & Lopes da Silva, F. H. Synaptic plasticity in an in vitro slice preparation of the rat nucleus accumbens. Eur. J. Neurosci. 5, 107–117 (1993). RESEARCH LETTER ©2012 Macmillan Publishers Limited. All rights reserved