The ability of memantine (1-amino-3,5-dimethyl-adamantane) to block the expression and maintenance of morphine dependence was examined in mice. When administered to morphine-dependent mice 45 min prior to naloxone challenge, memantine (7.5-30 mg/kg IP) in a dose-dependent manner reduced jumping behavior (a manifestation of the expression of dependence). The ability of memantine to attenuate naloxone-precipitated jumping was reversed by administration of glycine, an observation consistent with electrophysiological studies indicating that memantine is a use-dependent (uncompetitive) N-methyl-D-aspartate (NMDA) antagonist. In an independent series of experiments, the effect of memantine on a preestablished morphine dependence was investigated. A residual dependence to morphine was present 3 days after cessation of morphine administration. Repeated administration of memantine (10 mg/kg, IP) or the competitive NMDA antagonist NPC 17742 [2R, 4R, 5S-(2-amino-4,5-(1,2-cyclohexyl)-7-phosphonoheptanoic acid)] (6 mg/kg, IP) during this 3-day period abolished subsequent naloxone-precipitated jumping. In contrast, when administered concurrently with morphine after dependence had already been well established, memantine (10 and 20 mg/kg, IP) did not affect the maintenance of morphine dependence. Based on these finding, NMDA antagonists appear to inhibit the maintenance of opioid dependence, an action distinct from their acute inhibitory effects on the expression of dependence. Nonetheless, these regimen-dependent effects of memantine indicate that the most efficacious use of NMDA antagonists would be in detoxified subjects, rather than in individuals with an established dependence who are currently abusing opioids.