Non-steroidal anti-inflammatory drugs (NSAIDs) have been associated with reduced risk for Alzheimer's disease (AD) and selected NSAIDs racemates suppress beta-amyloid (Abeta) accumulation in vivo and Abeta42 production in vitro. Clinical use of NSAIDs for preventing or treating AD has been hampered by dose-limiting toxicity believed to be due to cyclooxygenase (COX)-inhibition that is reportedly not essential for selective Abeta42 reduction. Profens have racemates and R-enantiomers were supposed to be inactive forms. Here we demonstrate that R-ibuprofen and R-flurbiprofen, with poor COX-inhibiting activity, reduce Abeta42 production by human cells. Although these R-enantiomers inhibit nuclear factor-kappaB (NF-kappaB) activation and NF-kappaB can selectively regulate Abeta42, Abeta42 reduction is not mediated by inhibition of NF-kappaB activation. Because of its efficacy at lowering Abeta42 production and low toxicity profile, R-flurbiprofen is a strong candidate for clinical development.