A connection between Alzheimer's disease (AD) and Down syndrome (trisomy 21) is indicated by the fact that all Down syndrome individuals develop Alzheimer's disease neuropathology by the 4th decade of life. Previous studies have examined the frequency of aneuploidy and other chromosomal defects in cells from familial Alzheimer's disease (FAD) patients, with varying results. To investigate the possibility that a specific type of aneuploidy--trisomy 21 mosaicism--may contribute to Alzheimer's disease, we used quantitative fluorescence in situ hybridization to measure the number of trisomy 21 cells in primary fibroblast cultures from AD and unaffected subjects. The 27 AD cultures, including 15 that were derived from individuals carrying FAD mutations in presenilin 1 or 2, exhibited a significant approximately twofold increase in the number of trisomy 21 cells compared to 13 control cultures. A small double-hybridization experiment suggested that the aneuploidy in AD cells was not limited to chromosome 21 but extended at least to chromosome 18 as well. In a parallel study, the endogenous presenilin proteins in fibroblasts were localized to the centrosomes, the nuclear envelope, and its associated interphase kinetochores. Together these results indicate that the presenilin proteins may be involved in mitosis and that FAD mutations in the presenilin genes may predispose to chromosome missegregation (nondisjunction). The data reported here also suggest that trisomy 21 mosaicism may contribute to other forms of AD that are not caused by a presenilin mutation.