13 Jul 2000

Although there are many reports of neurotoxic effects of amyloid when applied to neurons in culture, the evidence for toxicity in vivo has been less compelling. Of particular surprise has been the finding that most transgenic mice in which abundant plaques form, show little evidence of neuronal loss. Bradley Hyman’s group has carried out much of these analyses. Is it possible that a subpopulation of plaques affect neurons in their immediate vicinity? This was the question addressed by Hyman’s group using a statistical model based on methods adapted from solid state physics. Transgenic mice (Tg2576 x PS1 M146L) were studied using double-immunofluorescence labeling of plaques and neurons. No global loss of neurons was found. However, when local analysis was carried out, local neuronal density was found to depend on the size and density of Aß deposits. For most plaques there was no effect on regional neuronal density. However, when subdivided on size and intensity of Aß staining, a subpopulation of plaques (about 2% of all plaques) was found to be associated with local decreases in neuronal density. These plaques stain with Thioflavin S, presumably due to dense fibrillar deposits of amyloid. In order to determine whether the decreased neuronal density was due to displacement of cells or neuronal loss, computer modeling was used. The results suggest that the plaques are not simplying "pushing" neurons aside but are truly causing cell loss. The reduced density is limited to the immediate region of the plaque, indicating that there is not a diffusible toxic signal emanating from the plaques. Whether these effects are indicative of a similar situation in human AD remains to be determined. However, the fact that some plaques in transgenic mice appear to cause local loss of neurons renews hope that these models will yet provide critical clues to understanding the pathological cascade in the Alzheimer brain.—Keith Crutcher

References: Hyman B. Assessing neuronal loss in transgenic models. World Alzheimer Congress 2000 Abstract 325.