Neurons treated with sub-lethal concentrations of amyloid-β1–42 developed phenotypic changes and selectively bound a CD14-IgG chimera; in co-cultures, microglia recognised and killed these amyloid-β1–42-damaged neurons. Pre-treatment with the platelet-activating factor (PAF) antagonists (Hexa-PAF, CV6209 or ginkgolide B) reduced CD14-IgG binding to amyloid-β1–42-damaged neurons, and the presence of PAF antagonists in co-cultures increased neuronal survival in a dose-dependant manner. PAF antagonists also protected neurons treated with HuPrP82–146, a peptide found in prion diseases. Second messenger studies demonstrated that the addition of PAF mimicked some of the effects of amyloid-β1–42 on neurons. PAF-damaged neurons bound CD14-IgG, and PAF-damaged neurons were killed by microglia in a CD14-dependent process. Neuronal death was inversely related to both the concentration of PAF, and the number of microglia added. The effects of PAF were reduced by an antagonist of the prostanoid D receptor (BWA868C) indicating that neuronal damage induced by PAF is partly mediated by prostaglandins. These observations are compatible with the hypothesis that sub-lethal concentrations of amyloid-β1–42 stimulate a cascade of second messengers including PAF and the prostaglandins. At nanomolar concentrations PAF induces a change in neuronal phenotype that activates microglia via the CD14 molecule, these activated microglia then kill the amyloid-β1–42 damaged neurons.