Apolipoprotein E (APOE) is a true susceptibility polymorphism of the common form of Alzheimer's disease (AD). There are three APOE alleles (epsilon 2, epsilon 3, epsilon 4) that are universally distributed in the population with some variation in allele frequency due to racial and ethnic differences, and are associated with different risks and age of onset distributions. In multiple studies, the positive predictive value for symptomatic possible or probable AD patients who carry at least one epsilon 4 allele was consistently > 95%. Thus, early in the clinical course of dementia, when diagnoses are only 60-70% accurate, the presence of an epsilon 4 allele raises the diagnostic accuracy of AD to 95%. With the anticipation of a second major late-onset AD susceptibility locus on chromosome 12, a matrix of relative susceptibility risks in the population raises many ethical and social questions associated with preclinical prediction. The metabolism of apoE (protein) in the brain is a new and exciting area of neurobiology research made relevant by the association with AD. We have constructed transgenic animals using large human genomic fragments containing human APOE on an APOE-deficit mouse background as well as homologous recombination experiments replacing mouse APOE with human APOE promoter elements. The APOE tissue elements, NOT the human APOE gene coding sequence, is associated with the human pattern of intraneuronal apoE immunoreactivity.