Following reports that ascorbic acid (AA) blocks NMDA receptors, we examined its possible neuroprotective properties in vivo (gerbil bilateral carotid artery occlusion model: BCAO) and in vitro (ischaemia-induced dopamine (DA) release in brain slices). Five minutes of BCAO caused substantial cell loss of 90-95% and 40-50% in gerbil CA1 hippocampus and striatum, respectively, measured in haematoxylin and eosin-stained sections, 5 days post-insult. AA (500 mg kg(-1) day(-1) i.p. for 31/2 days, first dose 1 h before occlusion) significantly (P < 0.05) reduced striatal cell loss (from 40 to 13%) while only reducing CA1 cell loss from 95 to 88%. A lower dose (250 mg kg(-1) day(-1) i.p. for 31/2 days) was ineffective in either region. AA (750 mg kg(-1) day(-1) i.p, for 31/2 days) caused significant striatal protection (cell loss reduced from 49 to 20%) if treatment was initiated 1 h before occlusion. Initiation of treatment immediately post occlusion did not cause significant protection. Neither treatment regime protected CA1 hippocampus. In separate experiments we examined the effect of AA on DA release, monitored by voltammetry, in an in vitro model of striatal ischaemia. Four DA release variables were measured: T-on - time from initiation of ischaemia to the onset of DA release, T-pk - the time from onset of DA release to maximum, delta DA/delta t - the mean rate of DA release and [DA](max) - the maximum extracellular DA concentration. Control values in drug-naive slices were: T-on = 193 +/- 8 s, T-pk = 24 +/- 4 s, [DA](max) = 69 +/- 6 mu M and delta DA/delta t = 4.2 +/- 0.7 mu M s(-1) (means +/- S.E.M., n = 15). 21/2 h pretreatment with AA (0.4 to 10 mM) did not affect T-on or [DA](max) but increased T-pk and decreased delta DA/delta t (P < 0.05) with an EC50 of 1.66 mM. NMDA (100 mu M) shortened T-on. N-ethylmaleimide (20 mu M) had no effect on the response to AA but potentiated the action of NMDA on T-on. AA (2 or 10 mM) had no effect on the response to NMDA, We conclude that AA is neuroprotective against global ischaemia in the striatum and that some of this action may be due to attenuation of ischaemia-induced DA release. This action is mediated neither by blockade of the NMDA receptor nor modulation of its redox status. (C) 1999 Elsevier Science B.V. All rights reserved.