On the probabilistic nature of excitotoxic neuronal death in hippocampal neurons.

In an attempt to distinguish hypothesized rapid and slow components, we have systematically studied the time course of hippocampal neuronal death in an in vitro model of excitotoxicity. In all paradigms involving glutamate, NMDA or AMPA as toxins, the population of trypan-blue excluding (live) neurons progressively declined over 48 hr. The percent survival over time could be fit mathematically using single exponential decay curves, implying that the death of any individual neuron was a stochastic event. One or two hours after glutamate exposure, prevention of further glutamate-receptor interactions by addition of MK-801 or MK-801 plus CNQX resulted in the survival of 60-80% of the original population at 24 hr. Thus delayed, continuous blockade of secondary glutamate receptor stimulation was protective, apparently interrupting the cyclic nature of the toxicity cascade. Twelve hours of MK-801 immediately following glutamate removal protected the majority of cells during the period of active receptor blockade. As soon as MK-801 was removed, the progressive decay in population size resumed, indicating that short term receptor blockade was insufficient to prevent expression of the initial injury. A kinetic model is proposed to place these experimental results into a framework for discussion and formulation of future experimentation.