Computationally guided design of bioactive nanostructures for targeted clearance of amyloid-β aggregates in Alzheimer's disease.

Targeted clearance of pre-existing amyloid-β (Aβ) aggregation remains a central challenge in Alzheimer's disease (AD) therapy. Here we report a computationally guided protein-gold hybrid nanostructure, Aβ3FTnAu, that integrates Aβ-recognition motifs into self-assembled human ferritin nanocages containing structurally defined gold nanoclusters composed of 12 gold atoms with Au-Au distances of 2.4-4.5 Å, enabling the selective recognition and disassembly of aggregated human Aβ. Structural analysis, mutagenesis and molecular simulations identify key interactions between gold-coordinating residues within Aβ3FTnAu (H118, T122, C130) and the Met35 residue of Aβ, revealing a mechanism in which multivalent engagement destabilizes fibrillar interfaces and promotes progressive plaque disassembly. In 5 × familial AD transgenic mice, systemic administration of Aβ3FTnAu reduced cerebral amyloid burden, preserved synaptic integrity and improved cognitive performance. This work establishes a rationally designed bioactive nanomaterial for targeted remodelling of pathological protein aggregates.