In vivo Neurotoxic effects of thioflavin S positive amyloid deposits

L. CRUZ, B. URBANC, R. LE, K. HSIAO-ASHE, K. DUFF,
H. E. STANLEY, M. C. IRRIZARRY, AND B. T. HYMAN

$^\ast$Center for Polymer Studies and Department of Physics, Boston University, Boston, MA 02215;

$^\dagger$Department of Neurology, Massachusetts General Hospital, Charlestown, MA 02219;

$^\ddagger$Department of Neurology, University of Minnesota, Minneapolis, MN 55455;

$^\S$Dementia Research, Nathan Kline Institute, Orangeburg, NY 10962.

ABSTRACT

Although amyloid-$\beta$ (A$\beta$) is clearly implicated in Alzheimer's disease pathogenesis as the major constituent of senile plaques, whether or not this peptide is directly neurotoxic is quite controversial. Transgenic mouse models of A$\beta$ deposition show extremely high amounts of immunoreactive A$\beta$ deposits in the cortex, but minimal or no neuronal loss as assessed by stereological methods. The vast majority of A$\beta$ does not appear to impact neuronal viability. We examined the hypothesis that only a subset of A$\beta$ deposits might affect neurons: by analyzing the neuronal landscape around each A$\beta$ deposit, we found that a subset of the total A$\beta$, accounting only for approximately 4% tissue coverage in PSAPP transgenic mice, is associated with local neuronal loss. This subset of A$\beta$ deposits corresponded to the biggest deposits containing the highest density of A$\beta$, and were also universally thioflavine-S positive, suggesting a $\beta$-pleated sheet conformation of the peptide. Computer simulations suggest that neurons are lost rather than simply being pushed out of the way by these dense A$\beta$ deposits. The results are consistent with a model in which A$\beta$ develops neurotoxic properties in vivo only when it adopts a densely packed $\beta$-pleated sheet conformation.