Glycine receptor antagonists have been proposed to have multiple therapeutic applications, including the treatment of stroke, epilepsy, and anxiety. The present study compared the biochemical and behavioral profiles of two strychnine-insensitive glycine receptor antagonists, MDL 100,458 (3-(benzoylmethylamino)-6-chloro-1H-indole-2- carboxylic acid) and MDL 102,288 (5,7-dichloro-1,4-dihydro-4-[[[4- [(methoxycarbonyl)amino]phenyl]sulfonyl]imino]-2-quinolinecarboxylic acid monohydrate). Both compounds potently inhibited [3H]glycine binding to rat cortical/hippocampal membranes (Ki = 136, 167 nM, respectively) without showing significant activity in 18 other receptor binding assays. In an in vitro functional assay, both compounds completely antagonized N-methyl-D-aspartate (NMDA)-stimulated cGMP accumulation in rat cerebellar slices. However, in contrast to their equipotency in the glycine receptor assay, MDL 100,458 was approximately 6-fold more potent than MDL 102,288 in the cGMP assay (IC50 values = 1.25, 7.8 microM, respectively). Behavioral tests demonstrated that MDL 102,288 and MDL 100,458 exhibited strikingly different in vivo profiles. MDL 100,458 antagonized audiogenic seizures in DBA/2J mice (ED50 = 20.8 mg/kg i.p.), whereas MDL 102,288 was without effect in the dose range tested (ED50 > 300 mg/kg i.p.). Central nervous system penetration did not appear to account for this difference. For example, MDL 102,288 was not active following direct intracerebroventricular administration (ED50 > 16 micrograms; vs. 0.78 microgram for MDL 100,458). In a test of anxiolytic activity, MDL 102,288 reduced separation-induced ultrasonic vocalizations in rat pups (ED50 = 6.3 mg/kg i.p.) whereas MDL 100,458 was only weakly active (ED50 = 80.8 mg/kg i.p.). Furthermore, the anxiolytic effect of MDL 102,288 was selective in that it occurred at doses that did not produce motoric disruption as measured by an inclined-plane test (ED50 > 160 mg/kg; therapeutic index > 25.4). In contrast, the anxiolytic activity of MDL 100,458 was non-selective in that it occurred at doses that also produced motoric disruption (ED50 = 57.7 mg/kg; therapeutic index = 0.7). Thus, two glycine receptor antagonists which have similar in vitro binding profiles as selective ligands for the strychnine-insensitive glycine receptor, demonstrate different in vitro and in vivo functional profiles. The reason for these differences is not clear, though one possibility could be that the compounds may act on different NMDA receptor subtypes. These data support the possibility that different glycine receptor antagonists may have different therapeutic targets.