On the role of seizure activity and endogenous excitatory amino acids in mediating seizure-associated hippocampal damage.

Abstract

For many years, the brain damage associated with epilepsy was thought to be the result of anoxia believed to occur in brain regions susceptible to the effects of seizures (Meldrum and Corsellis, 1984). This was a reasonable supposition since anoxia causes neuron loss in some of the same brain regions affected in epilepsy (Meldrum and Corsellis, 1984). Two concurrent and related lines of research during the past fifteen years have significantly changed this view of epileptic brain damage. Meldrum and others have shown that although anoxia causes cell death, neither anoxia nor decreased blood flow occur in the affected brain regions during seizure activity (Meldrum and Corsellis, 1984). The second derives from the discovery by Olney (1978) that a number of endogenous excitatory amino acids are directly neurotoxic. This finding led to the ‘excitotoxic’ hypothesis which states that excitation and cell death are causally related and that, therefore, abnormal concentrations of endogenous excitatory amino acids could be responsible for the neuropathologie lesions seen in a number of neurologic disorders (Olney and de Gubareff, 1978). The subsequent discovery by Olney et al. (1974) of the neurotoxic and convulsant properties of the glutamate analog kainic acid served as a stimulus to research in experimental epilepsy in particular, since this compound rapidly and reliably produces a seizure state in normal animals that is associated with a pattern of brain damage (Nadler et al., 1978) similar to that seen in the brains of many chronic human epileptics (Meldrum and Corsellis, 1984). Our interest in this area of research began with the desire to understand how kainate caused hippocampal seizure activity and whether a direct neurotoxic effect of kainate or seizure activity per se causes hippocampal cell death. Olney et al. (1974) originally suggested that kainate produces excitation and cell death by an agonist action at glutamate receptors. Accordingly, the sensitivity of different hippocampal cell types to kainate (Nadler et al., 1978) was suggested to be due to differences in glutamate receptor density (Olney et al., 1979). On the basis of studies showing that transection of the mossy fiber pathway (Nadler and Cuthbertson, 1980) or pretreatment with diazepam (Ben-Ari et al., 1979) prevented kainateinduced hippocampal damage, it was alternately suggested that the hippocampal damage caused by kainate is due to seizure activity induced in the hippocampal granule cells by kainate (Ben-Ari et al., 1979; Nadler and Cuthbertson, 1980). According to this view, seizure activity in the mossy fiber pathway damages CA3 pyramidal cells by an unspecified mechanism that does not involve glutamate or glutamate receptors (Ben-Ari et al., 1979; Nadler and Cuthbertson, 1980). The results of studies by Olney et al., (1974), Nadler and colleagues (1980) and Crawford and Connor (1973) led us to form an hypothesis that accomodated both theories. Nadler et al. (1980) showed that among the hippocampal cells most sensitive to the neurotoxic effects of kainate were the cells of the dentate hilus. Some of these interneurons receive dense innervation from the granule cells (Amaral, 1978) and are believed to mediate recurrent inhibition in the granule cell layer (Andersen et al., 1966). We hypothesized at the time that if, as had been suggested by Crawford and Connor (1973), the granule cells use glutamate as a transmitter, then the inhibitory interneurons that receive dense innervation from the granule cells might possess the highest density of glutamate receptors. If, as suggested by Olney and colleagues (1974), kainate acts via glutamate receptors, then these inhibitory inter-neurons might be preferentially damaged by kainate. Since the loss of inhibition is associated with the onset of seizure activity (Roberts, 1980), we predicted that kainate injection might decrease inhibition first and cause granule cell seizure activity as a result. According to this scenario, granule cell seizure activity would release glutamate from the mossy fibers and cause damage to the CA3 pyramidal cells as a result. This hypothesis would explain how kainate initiates granule cell seizure activity and why transection of the mossy fiber pathway or diazepam protects the CA3 pyramidal cells. Since it was not known at the time if kainate affected inhibition or even if it caused hippocampal granule cell seizure activity, this seemed a worthwhile starting point. Our initial study with kainic acid, and experiments that were the logical extension of it, are reviewed in this chapter. They provide evidence that seizure activity per se causes neuronal damage and that the release of endogenous excitatory amino acids in high concentrations during seizures may mediate epileptic brain damage.