In vitro studies of Flemish, Dutch, and wild-type beta-amyloid provide evidence for two-staged neurotoxicity.

Mutations in the beta-amyloid (Abeta) sequence of the amyloid precursor protein gene (APP) present with variable disease phenotypes. While patients with the Dutch APP mutation (E693Q) have predominantly hemorrhagic strokes, Flemish APP (A692G) patients develop both strokes and Alzheimer's disease (AD). To determine whether these diverse clinical and pathological presentations are due to mutant Abeta or APP, we studied the effect of Flemish, Dutch, and wild-type Abeta/APP on phosphorylation of specific tau epitopes observed in AD. No effect was observed in differentiated SH-SY5Y cells either stably expressing APP or treated with synthetic Abeta(12-42). However, we did observe a paradoxical temporal difference in the neurotoxic potential of mutant and wild-type Abeta. While long 24-h incubation at physiological levels of Abeta (2 microM) showed a higher amount of apoptosis for Dutch Abeta, a short 2-h incubation showed elevated apoptosis for Flemish and wild-type Abeta. The altered aggregating properties of Abeta, with Dutch Abeta aggregating faster and Flemish Abeta slower than wild type, elucidated a discrete two-phase Abeta neurotoxicity. We propose here that, at least in vitro, Abeta might be neurotoxic in an initial phase due to its soluble oligomeric or other early toxic Abeta intermediate(s), which is perhaps distinct from the late neurotoxicity incurred by aggregated larger assemblies of Abeta.