Senile plaques, the major neuropathological lesions of Alzheimer's disease (AD), are composed primarily of amyloid-beta (Abeta) peptide and contain high concentrations of iron (1.0 mM). We have previously shown that intracortical injections of 1.0 mM iron to adult rats produce significantly more neuronal loss than control injections of saline vehicle, whereas injections of Abeta do not. Because iron has been shown to increase the in vitro toxicity of Abeta, the present study was undertaken to determine whether iron can make Abeta neurotoxic in vivo. Abeta and 1.0 mM iron (as ferric ammonium citrate) were coinjected into rat cerebral cortex, and the neuronal loss was compared with that produced by pure Abeta or pure iron. The human and rat variants of Abeta(1-42) were compared to determine whether they produce the same amount of neuronal loss when combined with iron. Coinjection of iron with either Abeta variant caused significantly more neuronal loss than Abeta peptide alone, suggesting that iron may contribute to the toxicity associated with senile plaques. Rat Abeta(1-42) combined with iron was as toxic as iron alone, whereas iron combined with human Abeta(1-42) was significantly less toxic. This latter finding indicates that fibrillar human Abeta is able to reduce iron-induced neurotoxicity in vivo and raises the interesting possibility that senile plaques in AD may represent a neuroprotective response to the presence of elevated metal ions.