The commonly used inhaled anesthetic isoflurane has been shown to be both neuroprotective and neurotoxic in various cell cultures and animal models. We hypothesize that, like cerebral ischemia, isoflurane is inherently neurotoxic. Limited exposure of isoflurane provides neuroprotection via induction of endogenous neuroprotective mechanisms (preconditioning), while prolonged exposure of isoflurane induces neurotoxicity directly by its inherent neurotoxic effects. To test this hypothesis, we treated rat primary cortical neurons at different days in vitro (DIV) and rat pheochromocytoma neurosecretory (PC12) cells with or without Alzheimer's mutated presenilin-1 (PS1) with 2.4% isoflurane for 24 h to induce cell damage determined by both MTT (3-(4,5-dimethyithiazol-2-yl)-2,5-diphenyl-tetrazolium bromide) reduction and LDH (lactate dehydrogenase) release assays. For isoflurane preconditioning, we treated the above cells with isoflurane at 0.6%, 1.2% and 2.4% for 60 min, 4 h prior to a prolonged exposure of 2.4% isoflurane for 24 h. One hour of preconditioning with isoflurane dose-dependently inhibited neurotoxicity induced by 2.4% isoflurane for 24 h in both primary cortical neurons and PC12 cells. This neuroprotection was most dramatically observed in matured cortical neurons (DIV 16) and PC12 cells with over expression of Alzheimer's mutated PS1 (L286V). Preconditioning L286V PC12 cells with equivalent two minimal alveolar concentrations (MAC) of halothane (1.5%), but not sevoflurane (4%), also abolished the neurotoxicity induced by 2.4% isoflurane for 24 h. Overall, these results suggest that isoflurane may be both neuroprotective and neurotoxic, depending on the exposure concentrations and duration.