During brain ischaemia, the run-down of transmembrane ion gradients caused by the fall of ATP levels occurring leads to a rise in extracellular glutamate and GABA concentrations. The rise of glutamate concentration triggers neuronal death. In simulated ischaemia of hippocampal slices, taken from rats humanely killed in accordance with UK animal use legislation, we have used receptors in whole-cell patch-clamped CA1 pyramidal cells to sense released glutamate and GABA. E_Cl was set to 0 mV, so that currents mediated by ionotropic GABA receptors were inward, and membrane current was recorded at -30 mV to allow glutamate sensing by NMDA and AMPA receptors.

On applying superfusion solution mimicking the energy deprivation occurring during severe ischaemia (no oxygen and glucose, cyanide and iodoacetate present), a slow small increase of inward current occurred over the first few minutes, followed by a sudden massive inward current (nanoamps) which then sagged back to a less inward plateau (Rossi et al. 2000). The massive inward current corresponds to the ‘anoxic depolarization’ known to occur in vivo, when [K⁺]_o and [glutamate] rise to ~60 mM and 200 mM, respectively. Before this anoxic depolarization current, there was an increase in spontaneous EPSCs and IPSCs. Applying bicuculline or GABAzine during the maintained plateau current generated an outward current shift, demonstrating the occurrence of GABA release induced by the ischaemia. None of these events were affected by blocking action potentials with tetrodotoxin.

With bicuculline present throughout, a similar sequence of current changes was observed (Rossi et al. 2000). Both the initial anoxic depolarization current and the maintained plateau current were greatly reduced by glutamate receptor blockers, and so are produced by glutamate release. The glutamate-mediated current was not affected by blockers of Ca²⁺-dependent transmitter release, but was blocked by preloading cells with a slowly transported glutamate analogue (PDC) to block glutamate release by reversal of plasma membrane glutamate transporters (Rossi et al. 2000). Blocking the glial glutamate transporter GLT-1 with dihydrokainate, or knocking it out in transgenic mice, had no significant effect on the rise of [glutamate]_o in ischaemia, suggesting that glutamate release is by a neuronal transporter and that glia do not take up significant glutamate in ischaemia (Hamann et al. 2002). Reversal of neuronal transporters may occur more easily than reversal of glial transporters because [glutamate]_i is normally lower in glia due to glutamate conversion to glutamine by glutamine synthetase.

One ATP-dependent mechanism that will be blocked in ischaemia is the Na/K pump. In non-ischaemic solution, blocking the Na/K pump with ouabain produced a large transient inward current like the glutamate-mediated current underlying the start of the anoxic depolarization in ischaemia, but produced less sustained glutamate release than did ischaemia. However, in the presence of bafilomycin to block the vesicular H⁺-ATPase, or methionine sulfoximine to block glial glutamine synthetase, ouabain produced a maintained [glutamate]_o rise more similar to ischaemia. This suggests that, following a fall of [ATP] in ischaemia, loss of glutamate from vesicles caused by inhibition of the H⁺-ATPase may raise the cytoplasmic glutamate concentration and thus promote the release of glutamate which occurs by the reversal of plasma membrane transporters when the Na/K pump is inhibited. Furthermore, inhibition of glutamine synthetase by the fall of ATP occurring in ischaemia may prevent a protective uptake of glutamate into glia in early ischaemia: with glutamine synthetase inhibited, uptake of only a little glutamate released from neurons may be sufficient to raise [glutamate]_i in glia sufficiently to inhibit further uptake.

This work was supported by The Wellcome Trust.

All procedures accord with current local guidelines.