Amyloid-beta (A beta) peptides, the primary constituents of amyloid plaques in the brain in Alzheimer's disease (AD), may cause AD, but how they do so is not clear. A beta peptides spontaneously aggregate, or self-assemble, to generate several distinct macromolecular and morphological forms that can differ significantly in their effects on cells. We have compared different assembly forms of A beta(1-42) (A beta 42) for their ability to trigger apoptosis in cultured hippocampal neurons at a submicromolar concentration and for their binding to such neurons. Fibrillar A beta 42 caused both morphologicalchanges indicative of apoptosis and specific activation of caspase-3, a characteristic marker of neurodegeneration in AD, in hippocampal neurons, whereas other preparations tested did not do so under the same conditions. More aggregated forms of A beta 42, including both fibrils and a mixture of assembly forms termed A beta-derived diffusible ligands (ADDLs), bound to neurons much more extensively and at lower concentrations than preparations that contained smaller forms. Fibrillar A beta 42, in particular, bound to neurons at concentrations as low as 1 nM. Colocalization studies showed that fibrillar A beta 42 bound almost exclusively at nonsynaptic sites. These results show differences between assembly forms of A beta 42 in the ability to trigger apoptotic signaling in CNS neurons, and they directly demonstrate differences between assembly forms in the binding to CNS neurons, a possible first step in the pathogenesis of AD. These results suggest that fibrillar A beta 42 contributes to the pathogenesis of AD.