The neuroprotective effects of enhancing neuronal inhibition with a gamma-aminobutyric acid (GABA) uptake inhibitor were studied in gerbil hippocampus following transient ischemia. We used in vivo microdialysis to determine a suitable dosing regimen for tiagabine (NNC328) to elevate extracellular levels of GABA within the hippocampus. In anesthetized (normothermic) gerbils, tiagabine (45 mg/kg, i.p.) selectively elevated extracellular GABA levels 450% in area CA1 of the hippocampus. In gerbils subjected to cerebral ischemia via 5-min bilateral carotid occlusion, extracellular GABA levels increased 13-fold in area CA 1 returning to baseline within 30-45 min. When tiagabine was injected 10 min following onset of reperfusion, GABA levels remained elevated (200-470%) for 90 min. In addition, tiagabine significantly reduced the ischemic-induced elevation of glutamate levels in area CA1 during the postischemic period when GABA levels were elevated. There was no effect of postischemic tiagabine on aspartate or six other amino acids. Using the same dosing regimen, we evaluated the degree of neuroprotection in the hippocampus of gerbils 4 and 21 days after ischemia. Tiagabine decreased body temperature a maximum of 2.7 degrees C beginning 30 min into reperfusion and lasting 90 min. In untreated gerbils sacrificed 4 and 21 days after ischemia, there was severe necrosis (99%) of the pyramidal cell layer in area CA1. Whereas tiagabine significantly protected the CA1 pyramidal cell layer in ischemic gerbils at 4 days (overt necrosis confined to about 17% of area CA1), the protection diminished significantly 21 days postischemia. When normothermia was maintained both during and after ischemia in a separate group of tiagabine-treated animals, approximately 77% of the CA1 pyramidal cell layer was necrotic at 4 days. Based on these findings, we suggest that 1) tiagabine slows the development of hippocampal degeneration following ischemia, and 2) that mild, postischemic hypothermia is responsible, in large part, for the neuroprotective actions of this drug. We conclude that the histological outcome after administration of cerebral neuroprotectants should be assessed following long-term survival.