The antagonist pharmacology of glutamate neurotoxicity was quantitatively examined in murine cortical cell cultures. Addition of 1- 3 mM DL-2-amino-5-phosphonovalerate (APV), or its active isomer D-APV, acutely to the exposure solution selectively blocked the neuroexcitation and neuronal cell selectively blocked the neuroexcitation and neuronal cell loss produced by N-methyl-D-aspartate (NMDA), with relatively little effect on that produced by either kainate or quisqualate. As expected, this selective NMDA receptor blockade only partially reduced the neuroexcitation or acute neuronal swelling produced by the broad-spectrum agonist glutamate; surprisingly, however, this blockade was sufficient to reduce glutamate- induced neuronal cell loss markedly. Lower concentrations of APV or D- APV had much less protective effect, suggesting that the blockade of a large number of NMDA receptors was required to acutely antagonize glutamate neurotoxicity. This requirement may be caused by the amplification of small amounts of acute glutamate-induced injury by subsequent release of endogenous NMDA agonists from injured neurons, as the “late” addition of 10–1000 microM APV or D-APV (after termination of glutamate exposure) also reduced resultant neuronal damage. If APV or D-APV were present both during and after glutamate exposure, a summation dose-protection relationship was obtained, showing substantial protective efficacy at low micromolar antagonist concentrations. Screening of several other excitatory amino acid antagonists confirmed that the ability to antagonize glutamate neurotoxicity might correlate with ability to block NMDA-induced neuroexcitation: The reported NMDA antagonists ketamine and DL-2-amino- 7-phosphono-heptanoate, as well as the broad-spectrum antagonist kynurenate, were all found to attenuate glutamate neurotoxicity substantially; whereas gamma-D-glutamylaminomethyl sulfonate and L- glutamate diethyl ester, compounds reported to block predominantly quisqualate or kainate receptors, did not affect glutamate neurotoxicity. The present study suggests that glutamate neurotoxicity may be predominantly mediated by the activation of the NMDA subclass of glutamate receptors--occurring both directly, during exposure to exogenous compound, and indirectly, due to the subsequent release of endogenous NMDA agonists. Given other studies linking NMDA receptors to channels with unusually high calcium permeability, this suggestion is consistent with previous data showing that glutamate neurotoxicity depends heavily on extracellular calcium.