Although the concentration of drugs in brain homogenates is relatively easy to determine, such data are sometimes misleading due to accumulation in intracellular compartments. This is apparent for uncompetitive N-methyl-D-aspartate (NMDA) receptor antagonists where concentrations assessed in this manner are much higher than those sufficient to block the NMDA channel from the extracellular space. The aim of the present study was to determine whether free brain concentrations (extracellular fluid - ECF) of a new uncompetitive NMDA receptor antagonist MRZ 2/579 (1-amino-1,3,3,5,5-pentamethyl-cyclohexane hydrochloride) following administration of doses effective in animal models are sufficient to block NMDA receptors based on its potency in vitro. This issue was addressed using brain microdialysis corrected for in vivo recovery and patch clamp experiments.MRZ 2/579 blocked steady-state inward current responses of cultured hippocampal neurones to NMDA with an IC50 of 1.11 microM at -70 mV. Much higher concentrations of MRZ 2/579 blocked voltage-activated Ca2+ channels with an IC50 of 340 microM. MRZ 2/579 (10 microM) reduced peak inward current responses of neuronal nicotinic receptors only to 72.3% of control. MRZ 2/579 (10-100 microM) had little or no effect at AMPA and GABA(A) receptors. Following chronic s.c. infusion of MRZ 2/579 (40 mg/kg day for 7 days) brain ECF (2.15 microM) and cerebro-spinal fluid (CSF) levels (2.16 microM) were twofold lower than free plasma levels (4.3 microM). MRZ 2/579 showed pronounced accumulation in brain tissue compared to free plasma (28-fold) and ECF (58-fold). After acute i.p. administration (5, 10 and 20 mg/kg) peak concentrations in ECF were 0.70, 0.96 and 2.53 microM, respectively. In conclusion, MRZ 2/579 is indeed strongly accumulated in brain tissue compared to brain ECF, CSF and plasma. However, the brain ECF levels attained following administration of behaviourally effective doses are sufficient for selective NMDA receptor blockade.