There is mounting evidence that inflammatory processes, including activation of microglia, are upregulated in Alzheimer's disease. The importance of this phenomenon is indicated by multiple epidemiological studies showing that patients taking non-steroidal anti-inflammatory drugs (NSAIDs) have a substantially reduced prevalence of Alzheimer's disease. The pharmacological actions of anti-inflammatory drugs in brain are still uncertain. As a step towards identifying key pharmacological targets, we developed a neurotoxicity assay based on the property of supernatant media from stimulated human monocytic THP-1 cells to cause human neuroblastoma cell death. Similar neurotoxicity was observed when postmortem human microglia were substituted for THP-1 cells, establishing the validity of the assay for simulating neurotoxicity in human brain. A combination of lipopolysaccharide and interferon-gamma was used to activate the THP-1 cells. NSAIDs were effective in inhibiting neurotoxicity by this assay, while steroidal anti-inflammatories and propentofylline had no effect. The neuroprotective potency of NSAIDs appeared to be unrelated to their selective ability to inhibit cyclooxygenase-1 (COX-1) or cyclooxygenase-2 (COX-2). It is suggested that inhibition of monocyte cytotoxicity might be responsible for the apparent beneficial effects of NSAIDs in Alzheimer's disease.