The present study characterized the in vitro NMDA receptor antagonistic properties of novel amino-alkyl-cyclohexane derivatives and compared these effects with their ability to block excitotoxicity in vitro and MES-induced convulsions in vivo. The 36 amino-alkyl-cyclohexanes tested displaced [3H]-(+)-MK-801 binding to rat cortical membranes with K(i)s between 1.5 and 143 microM. Current responses of cultured hippocampal neurones to NMDA were antagonized by the same compounds with a wide range of potencies (IC50s of 1.3-245 microM, at -70 mV) in a use- and strongly voltage-dependent manner (delta 0.55-0.87). The offset kinetics of NMDA receptor blockade was correlated with equilibrium affinity (Corr Coeff. 0.87 P < 0.0001). As an example, MRZ 2/579 (1-amino-1,3,3,5,5-pentamethyl-cyclohexane HCl) had similar blocking kinetics to those previously reported for memantine (K(on) 10.67 +/- 0.09 x 10(4) M(-1) s(-1), K(off) 0.199 +/- 0.02 s(-1), K(d) = K(off)/K(on) = 1.87 microM c.f. IC50 of 1.29 microM). Most amino-alkyl-cyclohexanes were protective against glutamate toxicity in cultured cortical neurones (e.g. MRZ 2/579 IC50 2.16 +/- 0.03 microM). Potencies in the three in vitro assays showed a relatively strong cross correlation (all corr. coeffs. > 0.72, P < 0.0001). MRZ 2/579 was also effective in protecting hippocampal slices against 7 min. hypoxia/hypoglycaemia-induced reduction of fEPSP amplitude in CA1 with an EC50 of 7.01 +/- 0.24 microM. MRZ 2/579 showed no selectivity between NMDA receptor subtypes expressed in Xenopus oocytes but was somewhat more potent than in patch clamp experiments-IC50s of 0.49 +/- 0.11, 0.56 +/- 0.01 microM, 0.42 +/- 0.04 and 0.49 +/- 0.06 microM on NR1a/2A /2B, /2C and 2/D, respectively. In contrast, memantine and amantadine were both 3-fold more potent at NR1a/2C and NR1a/2D than NR1a/2A receptors. All Merz amino-alkyl-cyclohexane derivatives inhibited MES-induced convulsions in mice with ED50s ranging from 3.6 to 130 mg/kg i.p. The in vivo and in vitro potencies correlated indicating similar access of most compounds to the CNS. MRZ 2/579 administered at 10 mg/kg resulted in peak plasma concentrations of 5.3 and 1.4 microM following i.v. and p.o. administration respectively, which then declined with a half life of around 170-210 min. Analysis of A.U.C. concentrations indicates a p.o./i.v. bioavailability ratio for MRZ 2/579 of 60%. MRZ 2/579 injected i.p. at a dose of 5 mg/kg resulted in peak brain extracellular fluid (ECF) concentrations of 0.78 microM (brain microdialysates). Of the compounds tested MRZ 2/579, 2/615, 2/632, 2/633, 2/639 and 2/640 had affinities, kinetics and voltage-dependency most similar to those of memantine and had good therapeutic indices against MES-induced convulsions. We predict that these amino-alkyl-cyclohexanes, which all had methyl substitutions at R1, R2, and R5, at least one methyl or ethyl at R3 or R4 and a charged amino-containing substitution at R6, could be useful therapeutics in a wide range of CNS disorders proposed to involve disturbances of glutamatergic transmission.