Alzheimer disease (AD) is a devastating neurodegenerative disease with complex and strong genetic inheritance. Four genes have been established to either cause familial early onset AD (APP, PSEN1, and PSEN2) or to increase susceptibility for late onset AD (APOE). To date approximately 80% of the late onset AD genetic variance remains elusive. Recently our genome-wide association screen identified four novel late onset AD candidate genes. Ataxin 1 (ATXN1) is one of these four AD candidate genes and has been indicated to be the disease gene for spinocerebellar ataxia type 1, which is also a neurodegenerative disease. Mounting evidence suggests that the excessive accumulation of Abeta, the proteolytic product of beta-amyloid precursor protein (APP), is the primary AD pathological event. In this study, we ask whether ATXN1 may lead to AD pathogenesis by affecting Abeta and APP processing utilizing RNA interference in a human neuronal cell model and mouse primary cortical neurons. We show that knock-down of ATXN1 significantly increases the levels of both Abeta40 and Abeta42. This effect could be rescued with concurrent overexpression of ATXN1. Moreover, overexpression of ATXN1 decreased Abeta levels. Regarding the underlying molecular mechanism, we show that the effect of ATXN1 expression on Abeta levels is modulated via beta-secretase cleavage of APP. Taken together, ATXN1 functions as a genetic risk modifier that contributes to AD pathogenesis through a loss-of-function mechanism by regulating beta-secretase cleavage of APP and Abeta levels.