The hypothesis that antagonizing the positive reinforcing actions of drugs of abuse would prevent relapse and effectively treat drug dependence has received significant preclinical attention. Basic neuropharmacology has shown that all the behavioral effects of opiate drugs, including their positive reinforcing actions, can be reversed by the opiate antagonist naloxone (Di Chiara and North, 1992; Koob, 1992a). The opiate receptor subtype involved in heroin and morphine reinforcement appears to be largely the mu receptor. For cocaine, no specific competitive antagonist has been identified, but antagonism of dopamine receptors in the mesolimbic system appears to block competitively the reinforcing effects of cocaine (Woolverton and Johnson, 1992). Three of the five dopamine receptor subtypes have been implicated in these reinforcing effects, particularly the D-1 and D-3 receptors.
For benzodiazepines, a selective competitive antagonist has been characterized, but not studied in the context of benzodiazepine reinforcement, and it has little effect on ethanol reinforcement (Samson and Harris, 1992). Ethanol reinforcement can be blunted by antagonists and agonists to a number of neurotransmitter systems (Samson and Harris, 1992; Koob et al., 1994b), but none to date has proven wholly specific to or competitive with ethanol. Ethanol reinforcement can be decreased by GABA antagonists, dopamine antagonists, serotonin agonists, and opioid antagonists. Opiate antagonist effects appear to involve both mu and delta receptors, which has led to the introduction of naltrexone, the first new pharmacotherapeutic treatment for alcoholism in 40 years. Recent identification of a competitive THC antagonist will most certainly lead to its testing in reinforcement models. However, the clinical value of such an approach clearly still needs to be established, given the very limited success of opiate antagonists in treating opiate dependence.
Work on the development of antagonists for animal models of relapse (e.g., animal models for the conditioned positive and conditioned negative reinforcement associated with dependence) has only just begun (Koob, 1995). Limited studies suggest that dopamine antagonists can block the reinstatement induced by other drugs of abuse in the intravenous selfadministration reinstatement model. For the negative reinforcement associated with drug withdrawal, there is evidence that clonidine can block conditioned withdrawal from opiates (Kosten, 1994) and chlordiazepoxide can block conditioned withdrawal from ethanol (Baldo et al., 1995). Much more work is needed in this area, particularly in developing better mod-