RESUMEN
An imbalance between production and excretion of amyloid ß peptide (Aß) in the brain tissues of Alzheimer's disease (AD) patients leads to Aß accumulation and the formation of noxious Aß oligomers/plaques. A promising approach to AD prevention is the reduction of free Aß levels by directed enhancement of Aß binding to its natural depot, human serum albumin (HSA). We previously demonstrated the ability of specific low-molecular-weight ligands (LMWLs) in HSA to improve its affinity for Aß. Here we develop this approach through a bioinformatic search for the clinically approved AD-related LMWLs in HSA, followed by classification of the candidates according to the predicted location of their binding sites on the HSA surface, ranking of the candidates, and selective experimental validation of their impact on HSA affinity for Aß. The top 100 candidate LMWLs were classified into five clusters. The specific representatives of the different clusters exhibit dramatically different behavior, with 3- to 13-fold changes in equilibrium dissociation constants for the HSA-Aß40 interaction: prednisone favors HSA-Aß interaction, mefenamic acid shows the opposite effect, and levothyroxine exhibits bidirectional effects. Overall, the LMWLs in HSA chosen here provide a basis for drug repurposing for AD prevention, and for the search of medications promoting AD progression.