Warren J. Strittmatter, Karl H. Weisgraber, Michel Goedert, Ann M. Saunders, David Huang, Elizabeth H. Corder, Li-Ming Dong, Ross Jakes, Mark J. Alberts, John R. Gilbert, Seol-Heui Han, Christine Hulette, Gillian Einstein, Donald E. Schmechel, Margaret A. Pericak-Vance, Allen D. Roses

A genetic classification of Alzheimer disease(s) (AD) is presented. We describe a potential metabolic process in individuals who inherit apolipoprotein E-epsilon 4 (APOE4, gene; apoE4, protein) alleles, leading to increased risk and earlier age of onset of late-onset Alzheimer disease. Apolipoprotein E-epsilon 3 (apoE3) binds to tau protein, possibly slowing the initial rate of tau phosphorylation and self-assembly into paired helical filaments (PHFs); apoE4 does not bind tau. Tau promotes microtubule assembly and stabilizes microtubules; hyperphosphorylated tau does not bind, thereby destabilizing microtubules. Hyperphosphorylated tau may self-assemble into PHFs. Over time a bias toward destabilization of microtubules and the formation of neurofibrillary tangles may occur in individuals who inherit APOE4 alleles, leading to a shorter functional neuronal life span. This hypothesis focuses attention on two important aspects of AD research design: (1) Although the inheritance of APOE4 is associated with increased risk and decreased age of onset, apoE4 does not directly cause the disease. Our data point to the absence of an important function of apoE3 or apoE2 in individuals who do not inherit these alleles as the genetically relevant metabolic factor. This has important implications for design of experiments directed toward understanding the relevant neuronal metabolism. (2) Should this hypothesis be proven and confirmed, targets for pharmaceutical therapy designed to mimic the metabolic function of apoE3 or apoE2 become a realistic preventive strategy.