NEW TECHNOLOGY

45 of 65 plaques (70%) were cleared 3 days after initial imaging. In the 16B5 group, only 9 of 45 plaques were cleared after 3 days (20%; χ2= 30.5, P< 0.001). This result demonstrates that dense-core amyloid-β deposits are reversed by 10D5 application. We next examined whether all immunodetectable forms of amyloid-β deposits could be cleared by immunotherapy by repeating the experiment using labeled 10D5 as the imaging agent at the initial imaging session. As expected, the labeled 10D5 revealed innumerable diffuse and dense-core amyloid-β deposits. The animals recovered without incident and three days later were re-imaged. Very little or no detectable fluorescence remained from the application of fluorescein-labeled 10D5 that had been administered three days before; labeled 10D5 was then re-applied directly to the cortex in both treatment groups. Repeat imaging with labeled 3D6, a monoclonal antibody that recognizes a distinct epitope on plaques (as assessed on cryostat sections), showed that few or none of the amyloid-β deposits that were present at the initial imaging remained. Amyloid angiopathy was still detected (Fig. 3). Thus, three days after a single application of anti-amyloid-β antibody, we observed dramatic resolution of amyloid-β deposits in the parenchyma. It is possible that the vascular amyloid deposits are less accessible, more stable or more rapidly replenished, as they appear largely unchanged, despite the labeling with anti-amyloid-β antibody. However, small changes in the amyloid surrounding blood vessels over three days cannot be ruled out. Replication of this experiment with 3-to 8-day delays after initial imaging in one or two sites in each of 6 animals showed nearly identical results. Sham experiments were carried out in five animals in which fluorescein-labeled antibody 16B5 was used in the initial imaging session. The initial imaging session, using antibody 16B5, did not image any amyloid-β at all. This was expected as the monoclonal antibody was not directed against an epitope present on senile plaques. Repeat imaging 3 to 5 days later using labeled 10D5 imaged numerous amyloid-β deposits that were indistinguishable from the initial imaging sessions of any of the 6 mice initially imaged with 10D5. Thus it does not appear that the surgical preparation, application of an irrelevant monoclonal antibody, or imaging per se led to resolution of amyloid-β deposits. As the in vivo immunofluorescence technique images both diffuse