The apolipoprotein E (apoE) epsilon 4 allele increases risk of Alzheimer's disease (AD), perhaps by accelerating plaque formation, or by impairing neuron repair. Considerable evidence supports both mechanisms. AD patients with epsilon 4 have more and earlier amyloid deposits than do patients without epsilon 4. The same is true of non-demented control subjects. In vitro, all apoE isoforms inhibit amyloid beta protein (A beta) aggregation, but apoE4 less effectively than apoE3. Transgenic amyloid-producing mice expressing apoE3 or apoE4 develop less A beta deposition than apoE knockout mice. These observations are consistent with an effect of apoE isoforms on A beta aggregation in AD. ApoE is important for neurite maintenance since apoE knockout mice lose neurites and suffer behavioral deficits with aging or treatment with excitotoxins. ApoE4 mice show similar defects, but apoE3 mice are normal. AD patients with epsilon 4 show more neuritic deficits than epsilon 3 carriers. ApoE epsilon 4 also worsens neurological impairment in head injury, stroke, and multiple sclerosis. Thus, apoE4 is less effective at neurite maintenance. Perhaps epsilon 4 increases AD risk by both mechanisms: allowing amyloid deposition and failing to repair neurites. In either case, introducing apoE3 or apoE2 into the brain, for example by gene therapy or cell grafts, might delay AD progression. (C) 2000 Wiley-Liss, Inc.