Neuron, Vol. 47, 795–801, September 15, 2005, Copyright ©2005 by Elsevier Inc. DOI 10.1016/j.neuron.2005.08.007 Report Disrupting Reconsolidation of Drug Memories Reduces Cocaine-Seeking Behavior Jonathan L.C. Lee,* Patricia Di Ciano, Kerrie L. Thomas, 1 and Barry J. Everitt Department of Experimental Psychology University of Cambridge Downing Street Cambridge CB2 3EB United Kingdom Summary Maladaptive memories that associate environmental stimuli with the effects of drugs of abuse are known to be a major cause of relapse to, and persistence of, a drug addictive habit. However, memories may be disrupted after their acquisition and consolidation by impairing their reconsolidation. Here, we show that infusion of Zif268 antisense oligodeoxynucleotides into the basolateral amygdala, prior to the reactiva- tion of a well-learned memory for a conditioned stim- ulus (CS)-cocaine association, abolishes the acquired conditioned reinforcing properties of the drug-associ- ated stimulus and thus its impact on the learning of a new cocaine-seeking response. Furthermore, we show that reconsolidation of CS-fear memories also requires Zif268 in the amygdala. These results dem- onstrate that appetitive CS-drug memories undergo reconsolidation in a manner similar to aversive mem- ories and that this amygdala-dependent reconsolida- tion can be disrupted to reduce the impact of drug cues on drug seeking. Introduction Drug cues are known to induce craving and relapse in abstinent humans (Childress et al., 1999; Ehrman et al., 1992; Gawin, 1991; O’Brien et al., 1998), as well as re- lapse to drug seeking in experimental animals (de Wit and Stewart, 1981; Fuchs et al., 1998; Meil and See, 1996; Weiss et al., 2000), and attempts to extinguish their powerful acquired properties have not generally been successful as a treatment strategy for drug addic- tion (Conklin and Tiffany, 2002; Di Ciano and Everitt, 2004). Several experimental models of drug addiction have shown that the formation of a conditioned stimu- lus (CS)-drug association during drug self-administra- tion training enables that CS subsequently to induce and maintain drug seeking for prolonged periods (Ar- royo et al., 1998; Goldberg, 1975), to retard the extinc- tion of drug seeking (Ciccocioppo et al., 2004; Ranaldi and Roberts, 1996), and to induce relapse to drug seek- ing in extinguished or abstinent animals (de Wit and Stewart, 1981; Fuchs et al., 1998; Grimm et al., 2001; Meil and See, 1996). Through predictive association with the drug’s effects, the CS acquires powerful and *Correspondence: jlcl2@cam.ac.uk 1 Present address: School of Biosciences, Cardiff University, Cardiff CF10 3US, United Kingdom. enduring conditioned reinforcing properties (Di Ciano and Everitt, 2004; Grimm et al., 2001), thereby enabling it to support prolonged periods of drug-seeking beha- vior (Arroyo et al., 1998) and the learning of new drug- seeking responses (Di Ciano and Everitt, 2004). Indeed, a specific test of the conditioned reinforcing properties of a CS is its ability to support new instrumental learn- ing (Mackintosh, 1974), and so an acquisition of new response with conditioned reinforcement procedure can be used to assess the impact that drug cues will have on drug seeking. Indeed, this procedure models directly one important aspect of human addictive beha- vior, namely the rapid acquisition of novel, flexible drug-seeking strategies (Di Ciano and Everitt, 2004). Therapeutically, there is great interest in reducing the impact that drug cues have on addictive behavior and relapse, and to prevent them from reinforcing new drug-seeking actions that lead to the compulsive drug- seeking characteristic of the addicted state (DSM-IV- TR, 2000). Though extinction of the CS-drug associa- tion has not proved to be effective in reducing drug seeking or relapse in either humans (Conklin and Tif- fany, 2002) or rats (Di Ciano and Everitt, 2004), the be- havioral impact of a CS may be greatly reduced by pre- venting the reconsolidation of the previously learned memories that are retrieved and reactivated by its pre- sentation (Nader et al., 2000). Reconsolidation impair- ments, characterized by amnesia that is critically de- pendent upon memory reactivation through CS exposure, have been observed in several forms of memory (Alber- ini, 2005; Dudai and Eisenberg, 2004). Importantly, re- consolidation of amygdala-dependent conditioned fear has been disrupted by intraamygdala infusion of the protein synthesis inhibitor anisomycin (Nader et al., 2000). The basolateral amygdala (BLA) is an important locus of CS-US associations in both aversive and appe- titive learning (Everitt et al., 2000; LeDoux, 2000), and BLA lesions impair the ability of drug-associated CSs to acquire drug seeking under a second-order schedule of reinforcement (Whitelaw et al., 1996), and to induce reinstatement of drug seeking following extinction in rats (Meil and See, 1997). The hypothesis that we tested here was that drug memories undergo BLA- dependent reconsolidation and that preventing this process would reduce the impact that drug cues have on drug seeking and relapse. A protein that may be necessary for the reconsolida- tion of amygdala-dependent CS-US associations is the product of the immediate-early gene Zif268 (also known as EGR1, NGFI-A, and Krox24), since its expres- sion is significantly upregulated in the BLA following reexposure to discrete CSs previously associated with either footshock (Hall et al., 2001) or self-administered cocaine (Thomas et al., 2003). Similar cellular imaging data demonstrated that hippocampal Zif268 expression was also highly correlated with the recall of context- shock, and not CS-shock, associations (Hall et al., 2001), and the knockdown of Zif268 protein in the hip- pocampus, through intracranial infusion of Zif268 anti- sense oligodeoxynucleotides (ASO), was subsequently