The present study explores in vivo whether and how prostaglandin F(2alpha) (PGF(2alpha)), a membrane phospholipid hydrolysis product, causes neuronal death. The concentration of PGF(2alpha) measured by microdialysis sampling increased threefold immediately following impact injury to the rat spinal cord. Administration of PGF(2alpha) into the cord through a dialysis fiber caused significant cell loss, increased extracellular levels of hydroxyl radicals and malondialdehyde - an end product of membrane lipid peroxidation - to 3.3 and 2.3 times basal levels, respectively. This suggests that PGF(2alpha)-induced cell death is partly due to hydroxyl radical-triggered peroxidation. Generating hydroxyl radical by administering Fenton's reagents into the cord through the fibers significantly increased malondialdehyde production - the first direct in vivo evidence that hydroxyl radical triggers membrane lipid peroxidation. Methylprednisolone significantly reduced the release of PGF(2alpha) upon spinal cord injury and blocked PGF(2alpha)-induced hydroxyl radical and malondialdehyde production, but did not significantly reduce Fenton's reagent-induced malondialdehyde production, despite the production of more malondialdehyde by PGF(2alpha). This suggests that methylprednisolone may not directly scavenge hydroxyl radical, and that its 'antioxidant' effect is a consequence of blocking the pathways for producing toxic PGF(2alpha) and for PGF(2alpha)-induced hydroxyl radical formation, thereby reducing membrane lipid peroxidation.