A preferential loss of GABAergic, symmetric synapses in epileptic foci: a quantitative ultrastructural analysis of monkey neocortex

Previous immunocytochemical results from five monkeys with cortical focal epilepsy produced by alumina gel showed a severe decrease at seizure foci of axon terminals that contained glutamic acid decarboxylase (GAD), the synthesizing enzyme for the inhibitory neurotrasmitter, GABA. These data indicated a functional loss of GABAergic terminals but did not show: (1) whether this loss was caused by GABAergic nerve terminal degeneration or by a lack of GAD immunoreactivity within these terminals and (2) if this loss of GABAergic terminals was selective for only this terminal type. To resolve these issues, cortical tissue from three of the five monkeys used in the previous study was reexamined using electron microscopy, and a quantitative morphological analysis of cortical structures was made to compare profiles of terminals and glia in the nonepileptic cortex with those in the focus and parafocus. The following statistically significant changes were observed: (1) the number of axosomatic symmetric synapses with layer V pyramidal cells was decreased 80% at the focus and 50% at the parafocus, (2) in the neuropil adjacent to these pyramidal somata, the number of terminals forming symmetric synapses was reduced 50% at the epileptic focus but was unchanged at the parafocus, while the number of asymmetric synapses was reduced 25% at the focus and 15% at the parafocus, and (3) a 50% increase of glial profiles occurred at epileptic foci both in the neuropil and at sites apposed to pyramidal cell somata. The quantitative results also showed that terminals which form symmetric synapses had twice the number of mitochondria per terminal as those that form asymmetric synapses. Axon terminals which form symmetric synapses with somata and dendrites in the neocortex have been shown previously to contain GAD. Therefore, the large reduction in the number of symmetric synapses at epileptic foci and the increased gliosis indicate that the previously observed loss of GABAergic terminals at sites of focal epilepsy is caused by terminal degeneration. Since such terminals are reduced more severely at epileptic foci than other terminals, their loss could be the basis for seizure activity due to a preferential decrease of inhibitory function at epileptic foci. Hypoxia has been shown to cause a selective degeneration of terminals with the same morphology as GABAergic terminals in the cortex, and the basis for this loss could be related to higher physiological and/or metabolic activities of GABAergic cortical cells which may inhibit other cells tonically. The fact that increased numbers of mitochondria occur in GABAergic terminals supports this idea.